Report No. 20488 NI NICARAGUA POVERTY ASSESSMENT Challenges and Opportunities for Poverty Reduction Volume II: Annexes February 21, 2001 Poverty Reduction and Economic Management Sector Unit Latin America and the Caribbean Region U Document of the World Bank FLE COPY CURRENCY EQUIVALENTS US$1 = 12.60 C6rdobae (2000) US$1 = 10.58 C6rdobas (1998) FISCAL YEAR January I to December 31 ASDI Swedish Agency for International Development BCN Central Bank of Nicaragua CORNAP Coorporaci6n Nacional del Sector Publico DANIDA Danish International Development Agency DHS Demographic and Health Survey ENACAL Water and Sanitation Public Company ERP Effective Rate of Protection ESDENIC Encuesta Socio Demografica de Nicaragua FISE Emergency Social Investment Fund GDP Gross Domestic Product GON Government of Nicaragua HIPC Highly Indebted Poor Countries IDB Interamerican Development Bank IDF World Bank Institutional Development Fund IDG International Development Goals INIFOM Municipal Promotion Institute INSS Social Security Institute LSMS Living Standard Measurement Survey MAGFOR Ministry of Agriculture MECOVI Programa de Mejoramiento de Encuestas de Condiciones de Vida MIFAM Ministry of the Family MINSA Ministry of Health MITRAB Ministry of Labor NORAD Norway Agency for Development NQPE Nicaragua Qualitative Poverty and Exclusion Study NRP Net Rate of Protection PAD Project Appraisal Document PROTIERRA Project for Rural Municipalities PRSP Poverty Reduction Strategy Paper SAS Social Action Secretariat SETEC Technical Secretariat to the Presidency SNIP National System for Public Investment UNAM Autonomous University of Nicaragua UNFPA United Nations Population Fund UNDP United Nations Development Programme Vice President: David de Ferranti Country Director: Donna Dowsett-Coirolo PREM Director: Guillermo Perry Lead Economist: Ian Bannon Task Manager: Florencia T. Castro-Leal NICARAGUA POVERTY ASSESSMENT: CHALLENGES AND OPPORTUNITIES FOR POVERTY REDUCTION TABLE OF CONTENTS VOLUME 2: ANNEXES I. THE CONSUMPTION AGGREGATE, by Carlos Sobrado 2. MEASURING AND COMPARING POVERTY, PRE AND POST-MITCH, AND FUTURE POVERTY PROJECTIONS, by Florencia T. Castro-Leal and Carlos Sobrado 3. THE INCOME AGGREGATE, by Carlos Sobrado 4. SUMMARY STATISTICAL APPENDIX 5. STATISTICAL APPENDIX 6. ANALYSIS OF POVERTY COMPARISONS: CONFIDENCE INTERVALS SENSIVITY ANALYSIS, PRICE CHANGES AND DOMINANCE CONDITIONS, by Florencia T. Castro-Leal and Carlos Sobrado 7. INEQUALITY COMPARISONS, by Florencia T. Castro-Leal and Carlos Sobrado 8. COMPARING POVERTY FOR THE YOUNG AND THE OLD, by Carlos Sobrado 9. A SHORT GUIDE TO POVERTY AND INEQUALITY MEASURES, by Gabriel Demombynes 10. THE DETERMINANTS OF POVERTY IN NICARAGUA, by Carlos Sobrado 11. HEADSHIP, GENDER AND POVERTY IN NICARAGUA, by Nadeem Ilahi 12. A GROWTH ACCOUNTING APPROACH, by Ulrich Lachler 13. EDUCATION AND POVERTY IN NICARAGUA, by Gustavo Arcia 14. RURAL EDUCATION EXPERIENCES, by Gustavo Arcia and Vanessa Castro 15. RATES OF RETURN TO EDUCATION, by Diana Kruger 16. DETERMINANTS OF PRIMARY SCHOOL ATTENDANCE, by Diana Kruger 17. DEMAND FOR HEALTH CARE, by Mukesh Chawla 18. MALNUTRITION AMONG PRE-SCHOOL CHILDREN, by Mukesh Chawla 19. NICARAGUA POVERTY MAP FOR TARGETING THE EXTREME POOR, by Florencia T. Castro- Leal, Carlos Sobrado, Peter Lanjouw, Berk Ozier, Gabriel Demombynes, Carlos Lacayo, Juan Rocha, Dulce Maria Mayorga, and Luis Alaniz 20. NICARAGUA QUALITATIVE POVERTY AND EXCLUSION STUDY, by Mary Lisbeth Gonzalez 21. PUBLIC SOCIAL SPENDING AND PROGRAMS BY SECTOR IN 2000, by Ema Budinich 22. THE INCIDENCE OF PUBLIC EXPENDITURES ON HEALTH AND EDUCATION, by Julia Dayton. 23. BETWEEN POVERTY AND PROSPERITY: RURAL HOUSEHOLDS IN NICARAGUA, by Benjamin Davis and Rinku Murgai 24. REDISTRIBUTION EFFECTS OF AGRICULTURAL INCENTIVE POLICIES, by Diana Kruger 25. LABOR MARKETS AND POVERTY REDUCTION, by Nadeem Ilahi 26. FINANCIAL MARKETS, by Susana Sanchez 27. DETERMINANTS OF INFANT AND CHILD MORTALITY, by Maria Fernanda Mufiiz 28. DO PUBLIC HEALTH AND WATER SERVICES IMPROVE CHILD HEALTH?, by Julia Dayton The Poverty Assessment Team: Acknowledgements This report is the product of intense collaboration between the Government of Nicaragua, the World Bank, and the Nicaragua MECOVI program to contribute to the Nicaragua Poverty Reduction Strategy Paper (PRSP). The World Bank's task team included: Norman Hicks (Poverty Reduction Sector Manager), Ana-Maria Arriagada (Social Protection Sector Manager), Helena Ribe (Sector Leader), Ulrich Lachler (Nicaragua Resident Representative), Florencia T. Castro-Leal (Task Manager), Mukesh Chawla, Nancy Gillespie, Nadeem llahi, Kathy Lindert, Kalpana Mehra, Berk Ozler, Laura Rawlings, Susana Sanchez, Carlos Sobrado, and Diane Steele. Gustavo Arcia, Ema Budinich, Vanessa Castro, Benjamin Davis, Julia Dayton, Gabriel Demombynes, Mary Lisbeth Gonzalez, Diana Kruger, Maria Fernanda Mufiiz, and Rinku Murgai provided valuable background papers and notes. Diana Kruger, Kalpana Mehra and Carlos Sobrado provided excellent research assistance. Kathy Bain, Sara Calvo, Joaquin Caraballo, Miguel Angel Castell6n, Klaus Deininger, Alejandro Izquierdo, Emmanuel James, Isabel Lavadenz, Maurizio Guadagni, Judith McGuire, Ana Julia Moreno, Norman Piccioni, Dagoberto Rivera, Andrea Vermehren, and Torn Wiens also contributed. Aida Alvarado provided excellent administrative assistance and coordinated the production of the report. Peter Lanjouw, Giovanna Prennushi and Alberto Vald,s are Peer Reviewers. Maria-Valeria Junho Pena is Peer Reviewer of the Nicaragua Qualitative Poverty and Exclusion Study. The team also benefited from the guidance of Ian Bannon (Lead Economist). The task team from the Government of Nicaragua included: Mario De Franco (Advisor to the President), Luis Duran (SETEC Secretary), members of the SETEC team (Luis Alaniz, Mario Arana, Francisco Diaz, Carmen Largaespada, Matilde Neret and Carlos Sevilla), actual and former members of MAGFOR (Carlos Arce, Roberto Bendana, Oscar Neira, Horacio Rose and Diana Saavedra), members of BCN (Mario Flores, General Director; Jose de Jesus Rojas and Li Is Rivas), members of FISE (Carlos Noguera, President; Carlos Lacayo, Technical Director; Erasnio Vargas, Operations Director; and Transito Gomez and Joaquin Murillo, Poverty Map. The National Institute of Statistics and MECOVI-Nicaragua task team included: Luis Benavides (INEC General Director), Gonzalo Cunqueiro (MECOVI-Coordinator), Margel Beteta (INEC Director Surveys and Census), Martha Vargas (National Coordinator LSMS98&99, and Povert~ Map), Melva Bernales (Intemational Coordinator LSMS98&99), Irene Alvarez (Methodology), Javier Argefial (Data Processing), Rutilio Moreno (Field Operation), Elsa Maria Guti6rrez (Administrative Support). Francisco Arag6n, Denis Canizales, Silvia Chamorro, Thomas Gutierrez, Carol Herrera, Praxedis Hurtado, Elisa Lugo, Mayra Martinez, Arcides Menjivar, Yanett Narvaez, Rene Navarro, Javier Perez, Rolando Silva, Mario Solis, Juana Urroz y Cristina Zufiiga provided invaluable contributions. Juan Rocha and Dulce Maria Mayorga (Poverty Map). Edmundo Berumen and Alfredo Aliaga (Sampling). Vicente Merino, Alberto Valle, and Dominao Primante also contributed. Funding for this report, including the LSMS98&99 and the MECOVI-Nicaragua, was generously provided by the GON, the World Bank (including funding from FISE, the IDF and the Strategic Compact), ASDI, DANIDA, IDB, NORAD, UNFPA and UNDP. Annex 1, Page I Annex 1 - The Consumption Aggregate by Carlos Sobrado INTRODUCTION 1. A representative national and regional' Living Standard Measurement Survey (LSMS) was conducted in Nicaragua in 1998. The survey included all the questions necessary for constructing a consumption aggregate and for computing adjustment indexes that take family size, sampling design and geographical price differences into consideration. 2. A consumption aggregate was created for 4,040 households (22,423 persons) throughout Nicaragua: 370 households in the Atlantic Rural region (2,277 persons) and 843 households in the Pacific Urban region (4,360 persons). All information was collected between April and August 1998. THEORY 3. Since the consumption aggregate is a measure of welfare, the decision to include (or exclude) any specific item is related to the wellbeing of the household. An affirmative answer to the question "Does household condition/welfare improve with the consumption of the specific item?' indicates that such an item should be included as part of the consumption aggregate. 4. The final objective of the consumption aggregate is to allow for comparisons to be made between households, not to compute their individual welfare. The main objective is to rank the households from worst off to best off. This ranking allows us to identify common patterns or tendencies for different population groups. Any policy maker can use the observed characteristics to improve the planning and implementation of poverty reduction programs. 5. Compared to other measures of wellbeing, the consumption aggregate has a unique set of desirable characteristics. First, households tend to smooth out consumption over time (during a year and over the years). This attribute partially shields the consumption aggregate from temporary fluctuations that are not representative of the condition of the household2. Second, consumption information has proven to be more reliable than, for example, income data, due to concerns of household members about privacy and the difficulties in measuring informal eamings. Third, poverty lines can be derived from the same data sets. Fourth, the consumption aggregate is an objective measure that does not require a subjective selection and weighting of household characteristics. It reflects the preferences and choices made by each household, without assigning preconceived values to specific selections that may or may not reflect the utility of the individual household. Lastly, most data sets available for the construction of consumption aggregates contain the information required for computing other measures of wellbeing such as income aggregates and Basic Needs Indicators. We acknowledge the benefits of other measures of wellbeing and recommend their use as a complement to the preferred consumption aggregate. ' For the regions of Managua, Pacific Urban and Rural, Central Urban and Rural, and Atlantic Urban and Rural. 2 Christmas bonus and harvest season being the classical examples of temporary fluctuations within a year. Annex 1, Page 2 COMPONENTS OF TOTAL CONSUMPTION OVERVIEW 6. Total household consumption can be divided into two easily differentiated groups: food and non- food items. The food component includes not only items purchased and consumed at home, but also food from any other source, i.e. own production and donations and food consumed outside the home. The non- food component includes the value of housing, household services, health, education, as well as the use value of durable goods and spending on consumer goods and services. To value these components, irices from the same LSMS survey and the price questionnaire were used. To adjust for geographic price differences, a price index was constructed and applied to household consumption. Also, household size was used to convert total household consumption into per capita consumption. Finally, in producin.; any statistic the sampling design was taken into consideration (see below). Box 1 - Components of Total Consumption FOOD Food consumed at home: purchased, own production, donations, gifts Food consumed outside the home: restaurants, school, etc. NON-FOOD Housing: annual value of house and basic services Health Education Consumer goods and services: transport, communications, personal goods, leisure, etc. Use value of durable goods Food Components Food consumed at home 7. All data for this category was collected in Section 9.A of the LSMS household questionnaire (Expenditures and other household income Part A - food, beverages, and tobacco in the last 15 days). The annual purchase value of each food item is the product of the total value paid (question 6, variable GA06, items I through 62) multiplied by the yearly frequency reported in question 4, variable GA04, item, I through 62-3. Both non-purchased consumption and purchased food consumption increase the welfare of the household in the same way. Non-purchased food included own production and that which was received as payment household members from one's own business or received as a donation or gift. and that from any other source. Its total value was reported in question 10, variable GA10, items I through 62, and its frequency in question 8, variable GA08, items I through 62. 8. For 203 specific cases, no food value was provided (and question 3 or 7; variables GA03 oi GA07, indicated that consumption took place). For these cases and the identified outliers, the average 3 The values for question GA04 were transformed to represent the yearly frequency as follows: daily was re-coded to 365 (times per year), weekly was re-coded to 52, twice a month to 24, monthly to 12, every three months to 4, every six months to 2, and yearly to 1. Annex 1, Page 3 yearly per capita price, quantity, or value4 was used to estimate the consumption value5. A record was kept of the estimated values. 9. A few households did not report itemized food purchases. For these cases the yearly value was calculated using the entire value of food purchases over the last 15 days (question 24; variable GA24) and the corresponding yearly frequency (question 23; variable GA23A). Food consumed outside the home 10. Food consumed outside the home includes prepared food and beverages purchased outside the home (restaurants, cafeterias, etc.), and food consumed at school. 11. The weekly value of prepared foods and beverages purchased outside the home is reported in Section 9.B. I (Expenditures and other household income Part B. I - Expenditures last week). The total yearly value is the product of question 5 (variable GB 1405) multiplied by 52. The monthly value for food consumed at school was reported in section 4.B (Education: schooling for persons six years and older), question 19. Taking the length of the school year into consideration, the total yearly value for this category was calculated as the product of the value reported in question 19 (variable P4B 19) multiplied by 8.56. 12. The total food consumption in and out of the home is obtained by adding the yearly consumption values for food purchased and consumed at home, household production, gifts, donations, supermarket purchases, prepared foods purchased outside the home, and food consumed at school. Food consumption can be divided not only into food consumed in or out of the home, but also into purchased or non- purchased foods, or that which is produced at home or not. Different divisions do not change the value of the total consumption aggregate; they reflect the needs of the analyst. Non-Food Components Housing: use value of house 13. Since living in a house' increases the household member's welfare, the yearly value of the house must be included in the consumption aggregate. Housing data is collected in section 1 of the LSMS questionnaire (Household and housing characteristics). Of the households sampled, 4. 1% reported values for rental housing, 90.4% estimated the rental value of their own housing, and 5.5% did not report any real or estimated rent value data (Table 1). 4To compute the average values, 5% of the top and 5% of the bottom were truncated from the observed frequencies. 5 Use of average price or quantity was conditioned to the presence of other variable values in each case. If no information at all was reported, the average yearly per capita value was used. 6 The school year is eight and a half-months long. 7 By house we mean any living quarters including houses, apartments, and any other type of home. Annex 1, Page 4 Table A1.1 - House use values reported in the 1998 Nicaragua LSMS Households in the entire sample Information Numbers Percentage Households with monthly rent information (renters) 171 4.1% Households with monthly self-estimated value 3,806 90.4% (owners) Households without anv data 232 5.5% TOTAL 4,209 100.0% 14. House rental. Rent is a very good estimate of the use value of housing. For households renting their house, the reported rent was used to estimate the yearly value of living in the house. The annual rent value is calculated by multiplying the monthly rent value provided in question 14 (variable VB14) by 12. 15. Owned homes. Homeowners do not pay rent and no current market value is available. Homeowners were asked to estimate the monthly rental value of their house. The annual rent value was calculated by multiplying the answer to question 17 (variable VB 17) by 12. 16. Households with no data. Some households did not provide any data about the use value of their house. For those cases, the use value was estimated using household characteristics and expenditures. To impute the use value of these households, two steps were taken. First, we selected the households with rental information (real or self-reported5) and estimated a multi-variable linear regression between tlhe monthly rent value (dependent variable) and the selected household characteristics and expenditure-; (independent variables). A list of the variables and the estimated coefficients is provided in Table 2 Table A1.2 - Estimated coefficients from households with any type of rent information Variable Estimate t-value Definition Question Variabl significance e Constant -66.117 <0.0005 # of rooms (exclude kitchen, bathroom, 10 Vbl0 87.385 <0.0005 etc.) _ _ _ _ _ _ _ Urban or Rural household 5 105 -83.314 <0.0005 Water bill (last month) 22 VB22 1.683 <0.0005 Flush toilet in the house 28 VB28 109.608 <0.0005 Garbage disposal fee (last month) 37 VB37 14.244 <0.0005 Electric bill (last month) 40 VB40 1.248 <0.0005 Telephone bill (last month) 50 VB50 2.251 <0.0005 'For the Urban or Rural classification the question number and variable name are found in the general household ID section (first page). For all other variables, the question number and variable names are located in Section 1. 2 With an adjusted R2 of 0.36, an F value with an error probability below 0.001. Data source: Nicaragua 1998 LSMS It was necessary to include the households with self-estimated rent information due to the low number of households with real rent information. Annex 1, Page 5 17. Second, for households with no type of rental information, the rental value was imputed using the household characteristics and expenditure levels and the parameters estimated and reported in Table 2. 18. For all households, the yearly rental value was obtained by multiplying the reported, self- estimated, or imputed rental value by 12. Housing: value of services 19. All values for household services were collected in Section I of the LSMS questionnaire (Household and housing characteristics). Household services include water, garbage collection, electricity, cooking fuel, non-electric lighting, and telephone. Households with recently installed services were given a yearly use value of zerol'. If any services were included in the rent, the value assigned those services was also zero". For missing cases, a monthly estimated value was used. The estimates also took into consideration other household characteristics: for water, household size and water source; for garbage collection, urban or rural area and department; for electricity, household size and urban or rural area; for cooking fuel, urban or rural area and fuel type; and, for telephone, household size and urban or rural area. 20. The annual value for water consumed was obtained by multiplying the monthly payment reported in question 22 (variable VB22) by 12. There were 654 cases with no expenditure for water since the water consumed was from a river or creek. For these cases the minimum consumption value of C$2.37/month/household member was imputed'2. The yearly expenditure for garbage disposal was obtained by multiplying the monthly expenditure reported in question 37 (variable VB37) by 12. Yearly consumption values for electricity, non-electric lighting, cooking fuel, and telephone are the product of multiplying their corresponding monthly values in questions 40, 41, 48 and 50 (variables VB40, VB41, VB48, and VB50) by 12. 21. The total annual value for housing services is the sum of the yearly values for electricity, water, non-electric cooking fuel, non-electric lighting fuel, garbage disposal, and telephone. Health 22. Health expenditure data is included in section 3 of the LSMS questionnaire (Health), and section 9.B.4 (Expenditures and other household income Part B.4 -Expenditures in last 12 months). 23. From section 3.B (Children five years and younger), monthly expenditures in doctor fees for diarrhea treatment (question 20; variable P3B20B), and medicines (question 21; variable P3B21B) were used. The monthly value in total health expenditures was selected from section 3.C (Disease presence and control at any age) question 40 (variable P3C40). All the monthly values were added together and multiplied by 12 to obtain the yearly expenditure. The total health expenditure also includes the annual cost for private insurance from section 9.B.4 question 14 (variable GB4214). The sum of all the yearly values for household members gives the total annual health expenditure. 9 For eight cases it was necessary to set up a minimum often "cordobas" per month because the estimated value was negative. 'o Newly installed services imply zero consumption the previous year. " This shortcoming does not affect the consumption aggregate total value. 12 This is the minimum per capita monthly cost reported. The water source was a public supply center. Annex 1, Page 6 Education 24. Education expenditures are reported in section 4 (education) of the LSMS questionnaire. Section 4.A has the information for children five years and under (pre-primary) and section 4.B has the data for household members six years and above. Expenditure information concerning enrollment fees, uniforms, books, registration fees, and transport was collected for annual pre-primary expenses (questions 5 ard Sa; variables P4A05 and P4AO5A), and for students six years of age and older (questions 29, 30, 31, anc 32; variables P4B29, P4B30. P4B3 1, and P4B32). Households were also asked to report monthly expenditures for pre-primary students (question 4; variable P4A04) and for students six years of age and older (questions 24, 27, and 28; variables P4B24, P4B27, and P4B28). 25. The monthly values were multiplied by 8.5 (number of school months in one year) to obtain the annual expenditure. The total annual education expenditure was obtained by adding up the yearly values for all household members. Consumer goods expenditures 26. All consumer goods expenditures are in section 9.B (Expenditures and other household income) of the LSMS questionnaire. Among others, these expenditures include transport (taxi, bus), communications (public telephones, fax), personal items (deodorant, toothpaste), cleaning supplies (plastic gloves, brooms), leisure (movies, clubs, parties, gambling), clothing, footwear, cookware, services (haircuts, lawyers), furniture, small appliances (hair drier, electric shaver), and fines3. 27. The value for total personal expenditures includes question 4 items I through 4 (section 9.B 1: weekly expenditures), question 2 items I through 20, 22 and 23 (section 9.B.2: monthly expenditures), question 2 items I through 5 and 7 through 12 (section 9.B.3: expenditures every six months), and question 2 items 1, 4 through 7, 10, 11, and 15 through 17 (section 9.B.4: yearly expenditures). The total yearly expenditure is the sum of all the values multiplied by their annual frequency4. Use value of durable goods 28. Durable goods data are included in Section 9.E (Durable goods) of the LSMS questionnaire. which provides information for 25 durable goods, including radios, televisions, refrigerators, cars, etc.. The information collected for each item includes the number of goods in the household, as well as their age and current value. 29. Since durable goods are not consumed within a year, their current value or purchase cost does not reflect the benefit obtained by the household. For example, even if a household purchased a refrigerator during the time period of the survey, we expect the item to be used for more than a year. In the same way, a household will benefit from an old car without having paid any money during the time period of the survey. Using the purchase price of the durable goods implies that they are consumed completely in the first year and there is no further consumption. To accurately reflect the welfare that durable goods provide within a year, it is necessary to estimate the annual use value for each item. 30. To estimate the annual use value of each durable good (items I through 28), it is necessary lo know the age of the good (question 3; variable GE03), its current value (question 4; variable GE04). and its average lifetime (estimated). With the average lifetime of the durable good (Avg) and its present age 3 This is not an exhaustive list. A complete list of all items is included in the following paragraph no. 27. 14 For example, weekly expenditures were multiplied by 52 and monthly expenditures by 12. Annex 1, Page 7 (AGE), its remaining use life (Rem) can be calculated (Rem =Avg-AGE). For example, if computers are expected to have an average lifetime of eight years and we have a three-year old computer, the remaining useful lifetime would be Rem = 8 - 3 = 5. In other words, we expect the computer to work for the next five years. Dividing the current value of a durable good (VALUE) by its remaining useful lifetime (Rem) yields an estimate of its annual use value. In our example, if we estimate current sale value of the computer at $500, its estimated annual use value (YEAR) is YEAR = $500 / 5 = $100 per year. Looked at another way, the $500 computer will provide service worth $100 a year for the next five years. 31. Lastly, an estimate has to be made of the average lifetime for each durable good. To do this, we use the Hentschel and Lanjouw postulate "... the typical lifetime of such items (durable goods) was twice the estimated average age. This seems a reasonable hypothesis if we assume a given lifetime per item, and that household purchases of the item over time are relatively uniformly distributed." 15 Therefore, the estimated average lifetime for each of the 25 durable goods was estimated by calculating the average age reported in question 3 (variable GE03) and multiplying the result by 2. 32. The same procedure was used for each of the 25 durable goods. The annual use value for durable goods is obtained by adding the estimated values of all individual items. Total Yearly Consumption 33. Total yearly consumption per household is the sum of all the individual components described above. After the total yearly consumption was calculated, 129 households showed zero consumption for all the variables from section 9 and above'6. They were eliminated from the sample. Record was kept of the imputed values for each household. It was felt that consumption aggregates with more than 25% of the imputed value were not sufficiently reliable to be included. 39 households with imputed values representing more than 25% of their total consumption were eliminated from the sample. Finally, one household was eliminated because of incomplete information about its members. After making both selections, 4,040 households with consumption aggregates remained in the data set. ADJUSTMENTS TO TOTAL YEARLY CONSUMPTION 34. Three adjustments to the total consumption aggregate are needed to compare household welfare: (i) compute the per capita value, (ii) adjust the values for regional price differences, and (iii) consider the probability of selection. Per Capita Values 35. With the same total yearly consumption, smaller households (fewer members) are better off than large households'7. To compare the wellbeing of the households, total yearly consumption is divided by the number of members in each household. Household size was selected over adult equivalent scales because of the transparency and ease of understanding of the per capita concept. Adult equivalent scales primarily reflect calorie requirements, and not necessarily the cost of those calorie?8. Further, households tend to concentrate on protecting the younger more vulnerable members, making expenditures for health 15 Hentschel, J. and Lanjouw, P. Constructing an Indicator of Consumption for the Analysis of Poverty: Principles and Illustrations with Reference to Ecuador. World Bank LSMS working paper No. 124, Washington DC, March 1996. 6 These sections were covered in a second visit to the household (second round). 7 Using consumption as the measure of welfare. 18 For example, the cost per calorie tends to be higher for baby food than for adult food. Annex 1, Page 8 and education or spending more time caring for them. It is not clear that any equivalence scale can capture the actual needs or requirements of different household members. Regional Price Index 36. The same items have different prices in different regions of the country. To adjust for differences in the cost of living, price indexes were constructed for each Primary Sampling Unit (PSU) using th- information collected in section 9.A of the LSMS and data from the price questionnaire. The price index was based on food items and required two pieces of information: the average quantity consumed of 2ach item and the corresponding prices in each geographical unit and the entire country. 37. The national average quantity of food consumed (in pounds) was computed from section 9.A! of the LSMS questionnaire. The average per capita expenditure was obtained by adding up the national consumption values per item, as reported in questions 6 and 10 (GA06 + GA 10), and dividing the stjm by the number of persons. The average national consumption in pounds is the result of dividing the average per capita expenditure by the national average price for each food item. 39 items were used in this exercise, representing more than 85% of the total food consumed'9. At least 60 observations per food item were required to obtain the national average prices, with units expressed in pounds. 38. The LSMS (section 9.A) and the price questionnaires were used to obtain the prices for eacl, geographical area (PSU). For different levels of aggregation, a minimum number of observations (prices) were required: five for each PSU, ten for each urban/rural department, 50 for each urban or rural area, and 100 for national prices. The primary source of price data was the LSMS questionnaire. The price questionnaire was used when not enough price information was provided by the LSMS. With the price information, the purchase cost of the average national consumption in pounds for each PSU was estimated20. The purchase cost at the national level is the same as the average per capita expenditure computed previously. 39. The geographical index is the ratio of the purchasing cost of the average national consumption in pounds for each PSU over the average per capita expenditure. The computed index was found Lo vary between 0.88 and 1 .17 and to have a mean value of I. Probability of Selection 40. Because of the sampling design, not all households in the sample had the same probability of selection. Any statistic derived from the sample has to take this fact into consideration. The appropr ate index used is called the expansion factor, which is the inverse of the probability of selection. For example, if a household in the sample has a probability of selection of 0.01, the expansion factor is 1/0.001 = 100. A value of 100 implies that the household was selected from a total of 100 households1. The expansion factor was calculated for each household and is called "PESO2". 41. In addition, a second expansion factor was computed by multiplying the original expansion factor by the household size: PESO3 = PESO2 * HH size. The second expansion factor should only be used to obtain population statistics when working with a file containing one observation per household. The use of the proper expansion factor is determined first by the type of data file the analyst is working with (i.e. 19 Other items were a combination of more than one food product and therefore, quantities and prices were meaningless. 20 When not enough prices at the PSU level were available, the next geographical level of aggregation was used. 21 The expansion factor takes into consideration the households that were eliminated from the sample, as described in the "Total yearly consumption" section of this document. Annex 1, Page 9 one entry per household or one entry per person), and second by the desire inference level (i.e. population or households). See Table 3. Table A1.3 - Guide for the appropriate use of the expansion factors Data file: one entry per: Results to represent number of: Expansion factor Household Households PESO2 Person Persons PESO2 Household Persons PESO3 Source: Nicaragua 1998 LSMS TOTAL PER CAPITA CONSUMPTION IN NICARAGUA 1998 42. After adjusting total consumption, it was found that the average total per capita consumption in Nicaragua was C$6,645, ranging from C$1,331 for the poorest ten percent to C$23,231 for the richest ten percent of the population (See box 2). BOX 2 - Consumption level: Republic of Nicaragua, 1998 % of Annual per capita consumption values population for each ten percent of the population (C$') 100 23,231 90 9,828 80 7,262 70 5,792 60 4,812 50 4,056 40 3,347 30 2,696 20 2,103 10 1,331 Lowest level of total per capita consumption l All values are expressed in May 1998 c6rdobas. Source: Nicaragua 1998 LSMS Annex 2, Page I Annex 2 - Measuring and Comparing Poverty, Pre- and Post- Mitch, and Future Poverty Scenarios By Florencia T. Castro-Leal and Carlos Sobrado A. METHODOLOGY FOR DETERMINING POVERTY LEVELS IN NICARAGUA IN 1998 COMPARED TO 1993 1. In 1998, INEC (Nicaraguan Institute of Statistics and Census) conducted a Living Standard Measurement Survey (LSMS98). The survey was designed to be representative for the whole nation. covering rural and urban areas and seven regions making up the national territory.' 2. One objective of the LSMS98 is to estimate the levels of poverty in the country during the period under study in order to be able to identify characteristics of the poor groups that can help us design and apply policies and programs to reduce poverty. 3. Because a survey "similar"2 to the LSMS98 had been conducted in 1993, it was felt to be highlv necessary to compare the results of LSMS93 with LSMS98. 4. We have available to us a number of methods to obtain results for comparing the two surveys. The first and easiest is to prepare a consumption aggregate and poverty lines for the LSMS98 using the same techniques applied to the LSMS93. Unfortunately, the questions in both surveys were not sufficiently similar to be able to apply the '93 techniques to the LSMS98, and so the results would not be comparable. The second alternative consists in only taking the questions that are similar in both surveys, preparing a new consumption aggregate and determining the poverty lines utilizing a common methodology. Although technically this alternative is possible, it represents a change in the consumption aggregates, poverty lines, and percentages of poverty reported in 1993. It was felt that the confusion and difficulty of interpretation resulting from this alternative would be detrimental to the reliability and acceptability of results, and so it became necessary to maintain the figures reported initially in the LSMS93. The third alternative makes use of all the information gathered in the LSMS98 and in tum enables a comparison of the poverty levels to be made between the two surveys, without altering the original figures reported in the 1993 survey. The latter alternative was selected and is explained in more detail below. METHODOLOGY APPLIED 5. The selected methodology has two clearly defined parts: the construction of the consumption aggregate and the determination of the poverty lines4 Construction of the comprehensive consumption aggregate 6. Although being able to obtain results that can be compared to those of 1993 is one objective, it is not the only one. We also want to obtain a consumption aggregate that is the best ' Managua, Pacific Urban, Pacific Rural, Central Urban, Central Rural, Atlantic Urban, and Atlantic Rural. 2 Similar in terms of the number of themes covered, the methodology, the style of the questionnaire, and its application. 3 The World Bank, Republic of Nicaragua Povelty Assessment, Report No. 14038-NI, June 1995 4 The relation between these two components results in the poverty levels or the proportion of poverty. Annex 2, Page 2 possible representation of wellbeing in order to be able to order and group households in a way that allows us to determine the characteristics or behaviors of the different groups. 7. There is no doubt that the more complete or comprehensive the consumption aggregate. Fhe better it measures the wellbeing of the households. This in mind, the consumption aggregate based on the LSMS98 was constructed using all information available from the survey, and the norms and techniques currently applied in similar World Bank studies in other countries of Latin America and other parts of the world were followed, making adjustments in order to reflect conditions in Nicaragua. Components of the consumption aggregate Food 8. The food consumed by household members includes food purchased, produced by the houselhold, received as payment, from their own business, donations received, eaten in restauranrs, and other sources. The idea is to include any food that any household member consumed, regardless of origin. For other non0-purchased food. those interviewed made an estimate of its value. Housing 9. For housing, consumption was considered to be the payments for services and the use value of the home. The services included were water, garbage collection, electricity, non-electric lighting (kerosene), cooking fuel, and telephone. The use value of the home is the rental cost of a home f,r which rent is paid, the estimated value of those homes for which rent is not paid, and in some cases wvhere there was no information, the rental value was imputed using the characteristics of the horne to determine its use value. Education 10. This section used the enrollment costs, monthly payments, snacks, food, registration, registration fees, pre-registration fees, uniforms, books, school supplies, parents' association dues, and transport to school for all students, including preschoolers. Health 11. This includes the costs for doctors' visits and treatments received for diarrhea for children under six years of age, the cost of health care for all those in the household, and the costs for health and accident insurance. Household goods 12. For durable household goods, the use value for one year was estimated for each good, taking into account the age and estimated current value for each one. Transport 13. Expenditures for transport include taxi fares, urban and interurban bus fares (excluding school transport), fuel and lubricants, and private vehicle maintenance and repair. 5Consult the questionnaire to see the list of household goods included. Annex 2, Page 3 Personal and others 14. Included in this category are expenditures for periodicals, daily newspapers, magazines, personal items like soap, colognes, light bulbs, toilet paper. cooking pots, etc., repair services, domestic servants, legal expenses, entertainment, ceremonies, charitable donations, and clothing. 15. All components were calculated on an annual basis and the sum of these became a single indicator. Transformations 16. To convert household consumption into per capita consumption, the total consumption was divided by the number of household members. 1 7. In order to take into account the difference in prices in different regions of the country, the total consumption was transformed using a price index constructed on the basis of the data from the LSMS98. Since the prices of consumer goods vary from one zone to another in the country, the consumption aggregate must be modified in order to take into account this characteristic, making it possible to make comparisons between all the households in the country. 18. To prepare the geographic price index, it is necessary to: first, select the articles to be included; second, determine the quantities of the items selected for costing (similar to a shopping basket of items); third, estimate the prices of the selected articles in the different areas; fourth. calculate the acquisition costs for given quantities of articles in each area (amount by price); and fifth, select a basis on which all the adjustment factors for each area will be reported. a. Because there is no information for prices for all the regions of the country, an index was prepared based on data from the survey itself. The items to be included in the index should be relatively homogenous in all areas. The items considered as relatively homogenous, and for which we have information, are the foodstuffs reported in LSMS98 Section 9A and the information from the price questionnaire. b. After selecting the food items6, the average national per capita cost of the selected items was estimated. c. For each census segment, the average price for the items included was estimated, first of all using the prices for the items purchased by the households, and in their absence, the prices noted in the price questionnaire. When a census segment did not report a particular item and the information could not be found in the price questionnaire, the average for larger areas that included that census segment were used instead. d. The cost of purchasing the average national quantity was estimated in each census segment, using the prices from each segment (or an estimate). 6 This included: cornmeal/oatmeal, tortillas, nacatamales, corn, plain bread, sweetened bread, crackers/cookies, rice, pasta, coffee, beef, pork, bones (beef/pork), chicken, fish cutlets, tuna, sardines, canned goods, milk (pasteurized/from the cow), powdered milk, cheeses, butter/margarine, eggs, vegetable oil, pork lard, white onions, yellow onions, garlic, sweet peppers, tomatoes, potatoes, beans, cassava, sugar, salt, canned juice, liquor, beer, ice cream, and sherbet. Annex 2, Page 4 e. Although traditionally the area of the capital is selected, it was felt that for this consumption aacregate the basis should be the national average. The average national cost of purchasing the selected items was estimated, and this amount became the basis. The purchase costs iII each census segment were divided by the basis, resulting in an adjustment factor to be used for the purpose of division. 19. With the geographic price index available to us, we adjusted the per capita consumption, dividing it by the index and obtaining the comparable per capita comprehensive consumption aggregate for all households included in the survey. 20. Lastly. in order to report the results nationally, the use of expansion factors is required. The expansion factors are the inverse of the probability of selection for each household and are the number of households/persons in Nicaragua represented by each household/person in the sample. These expansion factors were calculated by INEC and were adjusted slightlv in order to take into account those houselholds for which the consumption aggregate could not be estimated for lack of data or because the data were not very reliable. DETERMINING POVERTY' LINES 21. The poverty lines for the LSMS98 were determined in such a way that, without changing the resuits reported in the '93 survey. it was possible to compare the percentages of the poor and extremely poor population for both years. In order to be able to elaborate poverty lines, it was necessary to create two new consumption aggregates. These should not be confused with the "comprehensive consumption aggregate" described earlier. Common consumption aggregates 22. In order to make a comparison between both surveys, two consumption aggregates were prepared, one with data from the LSMS93 and the other with data from the LSMS98. These consumption aggregates utilized only the questions considered to be "common" or similar in botl- surveys. It should be kept in mind that these consumption aggregates were 6onstructed with only one objective, to compare both surveys, and they should not be utilized individually. In order to differentiate these consumption aggregates they were called "common" consumption aggregates. 23. The common questions allowed for making a calculation of common foodstuffs7, a household consumption very similar to that described earlier, common education expensesd, and others, including ceremonies, club dues, lotteries. domestic services, shoes, sheets/blankets, towels, etc.. kitchen utensils, soap, detergent, transport, and communications. 24. For each year, all the components were added up to obtain a total common consumption aggregate for each household. 7Tortilla, cob corn, corn, plain bread, sweetened bread, rice, coffee, beef, chicken, fish cutlets, powdered milk, eggs, vegetable oil, plantains, cooking bananas, white onions, yellow onions, tomatoes, potatoes, beans, sugar, salt, cigarettes. 8 Including pre-school expenses (monthly fee, food, pre-enrollment registration, uniforms, books, fees, school supplies). For the rest of the students, only pre-enrollment, enrollment, uniforms, and books could be included. Annex 2. Page 5 Transformations 25. Similar to what was described earlier, the data were transformed utilizing the number of household members per household to obtain the common per capita aggregates. 26. To take into account the price differences by geographic area of the country. the same method described earlier was used, but this time Managua was chosen as the basis (fifth point in the previous description). 27. Furthermore, the common per capita aggregate for 1998 was modified to take into account the change in purchasing power of the c6rdoba between the LSMS93 and the LSMS98. For this transformation, the month of the interview was determined for each household, and the change in the price index9 during the period from LSMS9310 and the price index for the month prior to the LSMS98 interview were compared. 28. This work was done separately for each component of the 1998 common aggregate. The nine components were: I) food and beverages. 2) clothing and footwear, 3) housing, 4) furniture. accessories, household goods. 5) medical services and health care, 6) transport and communications, 7) leisure and recreation, 8) teaching, and 9) others. 29. Dividing the index for the month prior to LSMS98 by the average index for 1993 gives the "deflation" factor, with which the common consumption components for '98 will be modified. (The common consumption components will be divided by the "deflation" factor and the results for each household will be added up). Bv making this latter adjustment, we obtain the common per capita aggregate for 1998 in 1993 c6rdobas, allowing for comparisons to be made between both common aggregates. Povertj' Lines 30. In 1993 cordobas and in per capita annual terms, the extreme poverty line is C$1,233 and the overall poverty line is C$2,609. According to what was estimated on the basis of the LSMS93 in the World Bank Poverty Report, 19.4% of the population was classified as extremely poor and 50.3% as overall poor. 31. To calculate comparative poverty lines, we first ordered the values for the common per capita aggregate for '93 from greatest to least and we determined which are the values that would divide the population into groups of 19.4% (Value 1) and of 50.3% (Value 2). Figure 1 is a graphic depiction of how to carry out this process. 32. Second, we applied these two values (Value I and Value 2) to the common aggregate for 1998 converted to 1993 prices and ordered it from greatest to least. These values applied to the common aggregate for 1998 classify the population in 17.3% and 47.9%. These are the percentages for poverty in 1998 that we consider can be compared to the 1993 results. 33. The final step is to determine the value corresponding to the poverty lines to be applied to the 1998 comprehensive consumption aggregate (not the common aggregate). Here again we make 9 The price index used is the one for Managua published by the Central Bank. '° The survey was conducted between February and June 1993, but in order to take into account that the questions referred to earlier consumption, the price indexes for the previous month were used (January to May 1993). Annex 2, Page 6 a determination very similar to that made with the 1993 comnmon aggregate. The objective is to obtain the values for the poverty lines of the 1998 (per capita) comprehensive consumption aggregate that classify 17.3% and 47.9% of the population as extremely poor and overall poor, respectively. 34. This is achieved by ordering the values for the 1998 per capita comprehensive aggregate from greatest to least and determining the values for the poverty lines. Figure 2 represents this process, giving values of C$2.246 for extremely poor and C$4,259 for overall poor, whichl theretby classifies 17.3% as extremely poor and 47.9% as overall poor. INTERPRETATION AND PRECAUTIONS 35. After preparing statistical proofs between the levels for overall and extreme poverty nationally, it was found that both categories of poverty had been lowered statistically to a probability of 5% or less.]' On making the proofs at the level of rural areas and urban areas, reductions at the rural level were significant, but the change in poverty levels in urban areas was 0ot signifi canit. 36. Much care should be taken in interpreting these results since small changes in the assumptions in the technique utilized may change the conclusions in regards to the change of poverty in Nicaragua. If we make the comparison of the change of poverty nationally in a more rigorous way, for example to a 1% significant probability, the conclusion would be that no change can be detected in overall poverty or in extreme poverty. 37. At least three alternative methods have been tried in comparing the results from 1993 and 1998, and so far the results observed show the same tendencies reported, although to differing degrees. 38. Lastly, although the differences found in poverty nationally are at most marginal, it is felt that the methodology applied is based on solid technical know-how and represents as much as possible the situation of Nicaragua. " The degree of probability expresses the probability of being wrong when deternining that poverty has been reduced. Annex 2. Paze 7 FIGURE A2.1: DETERMINATION OF EXTREME AND GENERAL POVERTY LINES USING THE 1993 COMMONAGGREGATE Highest level of 1993 Percentage common consumption of population aggregate accumulated C$ 61,820 100% Given the percentages of the population of 50.3% and 19.4%. the values which would obtain those same percentages in the 1993 common aggregate are determined 50.4% Midpoint= Valuel 50.3% I9.5% Midpoint= Value 2 19.4% C$ 207 1% Lowest level of Percentage of common consumption population 1993 Accumulated Annex 2. Page 8 FIGURE A2.2 - DETERMINATION OF EXTREME AND GENERAL POVERTY LINES USING THE 1998 COMMON AGGREGATE Highest level of 1998 Percentage common consumption of population aggregate accumulated CS 177,095 100% Given the percentages of the population of 50.3% and 19.4%. the values which would obtain those same percentages are determined CS 4.260.6 48.0% Midpoint C$ 4,259 = overall poverty line C$4.258.6 47.9% C$ 2,246.5 17.4% Midpoint C$ 2,246 = extreme poverty line C$ 2,245.0 17.3% C$ 239 1% Lowest level of Percentage of common consumption population 1998 Accumulated Annex 2, Page 9 B. POVERTY IN NICARAGUA AFTER HURRICANE MITCH 39. To measure the poverty impact of Hurricane Mitch in Nicaragua, the households included in the 1998 LSMS located in "affected'2" areas were interviewed for a second time in May 1999 using a questionnaire containing the same sections needed to construct the consumption aggregate. These houselholds will be referred to as the "Post-Mitch" households and there is data for them from 1998 and 1999. 40. To measure the levels of poverty after Mitch, the results for the 1999 Post-Mitch households were substituted in the 1998 data set and new poverty rates were calculated. The resulting poverty rates are aii estimate of poverty in Nicaragua in 1998 if Mitch had happened before the 1998 survey. In other words, we combine the conditions in 1998 for the households in areas not "affected" by Mitch with the conditions in 1999 for the households in areas "affected" by Mitch. 41. In order to substitute the Post-Mitch household results'3 in the 1998 data set, several adjustments were needed. First, it was necessary to select or pair the Post-Mitch households from 1998 and 1999. Second. new expansion factors were calculated to take into consideration mismatcies. Third. the 1999 results were adjusted to take into consideration the change in the purchasing power of the c6rdoba (inflation). Fourth, the results were adjusted by the regional or geographic price index. ADJUSTMENTS Pairing the Post Mitch households 1998 and 1999 surveys 42. The Post-Mitch survey included 595 households. Of these, 505 were paired with the 1998 data without any problem. 26 households surveyed after Mitch did not have a consumption aggregate in 1998, 16 households had incomplete interviews in 1999, and 48 were "split" households from 1998. 43. The "split" households were cases in which the original 1998 household members split up into two or three households. The rule for selecting the Post-Mitch household to substitute in the 1998 data set was to use that Post-Mitch household in which the houselhold head (from 1998) lived. or in the absence of the household head, the Post-Mitch household with more of the original members from 1998. By applying these criteria, 23 Post-Mitch households were selected for substitution in the 1998 data set and 25 were removed. 44. In the end, 528 of the new interviews were used to estimate the new poverty rates. Since the original data set had 583 households in "affected" areas, there were 55 Post-Mitch-households without data for l999'4 Computing the new expansion factors 45. The expansion factors were adjusted only for households in "affected" areas. For the remaining households the expansion factors did not change. This was necessary because some 1998 12 "Affected" areas in Nicaragua were determnined at the segment level by INEC in Managua, Nicaragua. 13 Expressed by the consumption aggregate. 14 Some of these were households with incomplete interviews in 1999, but most correspond to households not found in 1999. Annex 2, Page 10 households in "affected" areas were not located in 1999 and some households hiad an increase or decrease in the number of members. 46. The adjustment was performed at the Primary Sampling Unit (PSU) level. The PSUs can he identified by a combination of the Department (iOI 5), Municipality (iO2), Area (iO5), and Segment (iO4) variables. The adjustment is made in such a way that each Post-Mitch PSU did expand to tl-ie same number of persons in both years. 47. To calculate the new expansion factor, first the number of persons sampled in each 1998 Post-Mitch PSU was computed (from 583 households). Second, the number of persons sampled in each 1999 Post-Mitch PSU was computed (from 528 households). The "adjustment factor" for each PSU is obtained by dividing the first result by the second. To obtain the new expansion factors. tie old expansion factors were multiplied by the "adjustment factor.' Adjusting for inflation between survey periods 48. Since a 1999 c6rdoba does not have the same value as a 1998 c6rdoba, all expenditures had to be adjusted to take this into consideration. 49. The Consumer Price Index (CPI) published by the Central Bank of Nicaragua was used to adjust the 1999 survey results. An "inflation" factor was calculated for each category published nv the Central Bank and each component in the 1999 consumption aggregate was modified using th-ie corresponding "inflation" factor. The different categories and the reported CPI values are shown in Table A2.1. Table A2.1 Nicaragua CPI by category and "inflation factor" Category May 1998 May 1999 Factor CP1 CPI Food 156 159 1.021 Clothing 108 123 1.136 Housing 190 249 I 1.307 Furniture and accessories 151 169 1.120 Health 134 140 1.042 Transport and 150 159 1.058 communications Leisure 144 153 1.060 Education 126 147 1.171 Personal Expenditure 122 130 1.064 | The 1999 values were divided by this number Adjusting for different prices in different areas in Nicaragua 50. To take into consideration the different prices in different regions of the country, the 1999 values were adjusted using the same factors calculated for 199816 POVERTY RATES 51. Since the 1999 consumption aggregates were adjusted to the time of the 1998 survey, the final poverty rates were calculated by replacing the 583 households (in the 1998 data set) in "affected" areas with the new data from 1999 (529 households). Then, the same poverty lines 15 The variable names correspond to the 1998 data set. 16 See the document XXXXX (how the consumption aggregate for 1998 was constructed). Annex 2. Page I I calculated for the original data set (C$2,246 for the extreme poverty line and C$4,259 for the overall poverty line) were applied. The new poverty rates were calculated using the adjusted expansion factors (Table A2.2). Table A2.2 Nicaragua-Pre- and Post-Mitch Poverty by region, 1998 & 1999 INCIDENCE OF AND CHANGE IN INCIDENCE OF AND CHANGE IN EXTREME POVERTY POVERTY Headcount Index (%) HeadCount Index (%) Region 1998 1999 Change 1998 1999 Change National 17.3 17.3 -0.1 ns 47.9 47.9 0.1 ns Urban 7.6 7.5 -0. I ns 30.5 30.3 -0.2 ns Rural 28.9 28.9 0.0 ns 68.5 69.0 0.4 ns Managua 3.1 3.1 0.0 X 18.5 18.5 0.0 X Pacific Urban 9.8 9.6 -0.3 ns 39.6 39.0 -0.6 ns Rural 24.1 20.6 -3.6 ** 67.1 63.1 -4.0 ** Central Urban 12.2 12.1 -0.1 ns 39.4 39.4 0.0 ns Rural 32.7 35.7 2.9 ** 74.0 77.6 3.6 ** Atlantic Urban 17.0 17.0 0.0 X 44.4 44.4 0.0 X Rural 41.4 40.6 -0.8 ns 79.3 80.6 1.3 ns Source: Nicaragua LSMS 1998 and LSMS 1999. Table notation:ns = non-significant at p<=lO%; ** = significant at p<=-l% and. IX= Managua and Atlantic Rural did not have any Post-Mitch surveyed households. C. FUTURE POVERTY SCENARIOS 52. Growth in mean per capita consumption could have a greater impact on reducing extreme than overall poverty. By estimating consumption elasticities of poverty measures (headcount, poverty gap, and Foster-Greer-Thorbecke indices) we can analyze their responsiveness to growth of mean per capita income.17 Intuitively, assuming that the consumption distribution remains the same as in 1998, a I percentage point increase in mean per capita consumption decreases the headcount ratio of the extremely poor by 2 percentage points in but of the overall poor by only 1.14 percentage points (Table A2.3). Note that the elasticity of each poverty index is greater for extreme poverty than for overall poverty in both urban and rural areas. This means that extreme poverty is more responsive to increases in mean per capita consumption. '' The estimations used POVCAL. Headcount means the proportion of the population that is poor. Annex 2, Page 12 Table A2.3 Nicaragua - Consumption Elasticities of Poverty Indices National Urban Rural Poverty Index Overall Extreme Overall Extreme Overall Extreme povertv poverty poverty povertv poverty poverty Head Count -1.14 -1.96 -1.44 -2.88 -0.72 -1.44 Povert Gap -1.61 -2.87 -2.02 -4.70 -1.19 -1.95 FGT2 -2.00 -3.77 -2.56 -6.52 -1.50 -2.42 Source: Nicaragua LSNIS 1998. 53. However. rural poverty is less responsive to economic growth and thus its reduction requires more comprehensive strategies. The response to growth in mean per capita consumption will not be the same in all areas of the country. Both overall and extreme poverty will decrease twice as fast n urban than in rural areas due to higher urban elasticities: a one percentage point increase in mean per capita consumption decreases the headcount ratio of the urban poor by 1.44 percentage points and of the rural overall poor by 0.72 percentage points. It decreases the headcount ratio of urban extreme poverty by 2.88 percentage points and of rural extreme poverty by 1.44 percentage points. 54. An optimistic scenario of growth in mean consumption of 2 percent annually reduces extreme poverty by half and overall poverty by one-third by 2015 (Table A2.3). On the basis of the elasticities in Table A2.2, it is possible to generate future poverty projections in Nicaragua under alternative scenarios of mean consumption growth. Between 1993 and 1998, Nicaragua's GDP experienced an average real growth rate of 4.3 percent per year; however, fast population growth brought the average real growth rate of GDP per capita to only 1.2 percent per year." Assuming an unchanged distribution of income,l9 poverty rates will decline substantially by 2015. However. fertility rates among the poorest are much higher than among the non-poor; live births per womat. in the poorest quintile (6.6) are more than three times those of the highest (1.9). This factor alone, even under distributionally neutral policies, will worsen the distribution of income. Thus, assuming distributionally neutral growth of mean household consumption might well be an optimistic scenario.20 This clearly points towards the need for a more vigorous stance on population policies focusing strongly on the poor, to reduce poverty and ultimately to distribute the fruits of growth more evenly. 18 The source for GDP data is the Central Bank of Nicaragua (BCN); and for population growth rates, the Institute of Statistics and Census (INEC). '9 We use consumption, rather than income, as a measure of welfare in this report. See Annex I for an explanation of the consumption aggregate. However, they are used interchangeably in this section. 20 Scenarios with higher annual increases in mean household consumption (such as 2 or 3 percent in Table A2.3). necessarily assume either high rates of economic growth (4.7 percent to 5.7 percent by adding a 2.7 percent constant rate of population growth) or declining rates of population growth (between 2.3 percent tco 1.3 percent by subtracting a 4.3 percent constant rate of economic growth such as that of 1993-1998). Annex 2, Page 13 Table A2.3 Projected Levels of Povertv with different scenarios for mean household consumption (assuming no change in income distribution) 1 .2% annual 2% annual increase 3% annual increase Poverty Index increase in mean in mean household in mean household household consumption consumption consumption 1998 2005 2010 2015 2005 2010 2015 2005 2010 2015 NATIONAL Overall Poverty: Head Count 47.9 43.5 40.6 37.9 40.8 36.3 32.4 37.6 31.6 26.5 Poverty Gap 18.3 16.0 14.5 13.1 14.6 12.4 10.5 12.9 10.1 7.9 FGT2 9.3 7.8 7.0 6.2 7.0 5.7 4.7 6.0 4.4 3.3 Extreme Povertv: Head Count 17.3 15.7 14.7 13.7 13.1 10.7 8.8 11.3 8.4 6.2 Poverty Gap 4.8 4.2 3.8 3.5 3.8 3.2 2.8 3.4 2.7 2.1 FGT2 2.0 1.6 1.5 1.3 1.5 1.2 1.0 1.3 0.9 0.7 URBAN Overall Poverty: HeadCount 30.5 27.0 24.7 22.7 24.8 21.5 18.6 22.4 18.0 14.4 Poverty Gap 9.9 8.3 7.3 6.5 7.4 6.0 4.9 6.4 4.6 3.4 FGT2 4.5 3.6 3.1 2.6 3.1 2.4 1.8 2.6 1.7 1.2 Extreme Poverty: Head Count 7.6 6.8 6.2 5.7 5.0 3.7 2.8 4.1 2.6 1.6 Poverty Gap 1.9 1.6 1.4 1.2 1.0 0.6 0.4 0.7 0.3 0.1 FGT2 0.7 0.5 0.5 0.4 0.3 0.1 0.1 0.1 0.0 0.0 RURAL Overall Poverty: Head Count 68.5 64.5 61.7 59.1 61.9 57.6 53.5 58.8 52.7 47.2 PovertvGap 28.3 25.6 23.8 22.2 23.9 21.2 18.8 21.9 18.3 15.3 FGT2 14.9 13.1 12.0 11.0 12.1 10.4 8.9 10.8 8.6 6.8 Extreme Poverty: Head Count 28.9 27.2 26.0 24.9 23.6 20.4 17.6 21.2 17.0 13.7 Poverty Gap 8.3 7.5 7.0 6.5 6.3 5.1 4.2 5.4 4.0 3.0 FGT2 3.5 3.1 2.8 2.5 2.5 1.9 1.5 2.0 1.4 1.0 Source: Nicaragua LSMS 1998. Annex 2, Page 1 Annex 2 - Measuring and Comparing Poverty, Pre- and Post- Mitch, and Future Poverty Scenarios B l Florencia T Castro-Leal and Carlos Sobrado A. METHODOLOGY FOR DETERMINING POVERTY LEVELS IN NICARAGUA IN 1998 COMPARED TO 1993 1. In 1998, INEC (Nicaraguan Institute of Statistics and Census) conducted a Living Standard Measurement Survey (LSMS98). The survey was designed to be representative for the whole nation, covering rural and urban areas and seven regions making up the national territory. 2. One objective of the LSMS98 is to estimate the levels of poverty in the country during the period under study in order to be able to identify characteristics of the poor groups that can help us design and apply policies and programs to reduce poverty. 3. Because a survev 'similar"2 to the LSMS98 had been conducted in 1993, it was felt to be highil necessary to compare the results of LSMS93 with LSMS98. 4. We have available to us a number of methods to obtain results for comparing the two surveys. The first and easiest is to prepare a consumption aggregate and poverty lines for the LSMS98 using the same techniques applied to the LSMS93. Unfortunately, the questions in both surveys were not sufficiently similar to be able to apply the '93 techniques to the LSMS98. and so the results would not be comparable. The second alternative consists in only taking the questions that are similar in both surveys, preparing a new consumption aggregate and determining the poverty lines utilizing a common methodology. Although technically this alternative is possible, it represents a change in the consumption aggregates, poverty lines, and percentages of poverty reported in 1993. It was felt that the confusion and difficulty of interpretation resulting from this alternative would be detrimental to the reliability and acceptability of results, and so it became necessary to maintain the figures reported initially in the LSMS933. The tilird alternative makes use of all the information gathered in the LSMS98 and in turn enables a comparison of the poverty levels to be made between the two surveys, without altering the original figures reported in the 1993 survey. The latter alternative was selected and is explained in more detail below. METHODOLOGY APPLIED 5. The selected methodology has two clearly defined parts: the construction of the consumption aggregate and the determination of the poverty lines4 Construction of the comprehensive consumption aggregate 6. Although being able to obtain results that can be compared to those of 1993 is one objective, it is not the only one. We also want to obtain a consumption aggregate that is the best ' Managua, Pacific Urban, Pacific Rural, Central Urban, Central Rural, Atlantic Urban, and Atlantic Rural. 2 Similar in terms of the number of themes covered, the methodology, the style of the questionnaire, and its application. 3 The World Bank, Republic of Nicaragua Poverty Assessment, Report No. 14038-NI, June 1995 4 The relation between these two components results in the poverty levels or the proportion of poverty. Annex 2, Page 2 possible representation of wellbeing in order to be able to order and group households in a way that allows us to determine the characteristics or behaviors of the different groups. 7. There is no doubt that the more complete or comprehensive the consumption aggregate, the better it measures the wellbeing of the households. This in mind, the consumption aggregate baszd on the LSMS98 was constructed using all information available from the survey, and the norms and techniques currently applied in similar World Bank studies in other countries of Latin America and other parts of the world were followed, making adjustments in order to reflect conditions in Nicaragua. Components of the consumption aggregate Food 8. The food consumed by household members includes food purchased, produced by the houselhold, received as payment, from their own business, donations received, eaten in restauranLs. and other sources. The idea is to include any food that any household member consumed, regardless of origin. For other non-purchased food. those interviewed made an estimate of its value. Housing 9. For housing, consumption was considered to be the pavments for services and the use va lue of the home. The services included were water, garbage collection. electricity, non-electric light ing (kerosene), cooking fuel, and telephone. The use value of the home is the rental cost of a home for which rent is paid, the estimated value of those homes for which rent is not paid, and in some cases where there was no information, the rental value was imputed using the characteristics of the horne to determine its use value. Education 10. This section used the enrollment costs, monthly payments, snacks, food, registration, registration fees, pre-registration fees, uniforms, books, school supplies, parents' association dues, and transport to school for all students, including preschoolers. Health 11. This includes the costs for doctors' visits and treatments received for diarrhea for childrzn under six years of age, the cost of health care for all those in the household, and the costs for health and accident insurance. Household goods 12. For durable household goods, the use value for one year was estimated for each good, taking into account the age and estimated current value for each one.' Transport 13. Expenditures for transport include taxi fares, urban and interurban bus fares (excluding school transport), fuel and lubricants, and private vehicle maintenance and repair. 5 Consult the questionnaire to see the list of household goods included. Annex 2, Page 3 Personal and others 14. Included in this category are expenditures for periodicals, daily newspapers, magazines, personal items like soap, colognes, light bulbs, toilet paper, cooking pots, etc., repair services, domestic servants, legal expenses, entertainment, ceremonies, charitable donations, and clothing. 15. All components were calculated on an annual basis and the sum of these became a single indicator. Transformations 16. To convert household consumption into per capita consumption, the total consumption was divided by the number of household members. 17. In order to take into account the difference in prices in different regions of the country, the total consumption was transformed using a price index constructed on the basis of the data from the LSMS98. Since the prices of consumer goods vary from one zone to another in the country, the consumption aggregate must be modified in order to take into account this characteristic, making it possible to make comparisons between all the households in the country. 18. To prepare the geographic price index, it is necessary to: first, select the articles to be included; second, determine the quantities of the items selected for costing (similar to a shopping basket of items); third, estimate the prices of the selected articles in the different areas: fourth, calculate the acquisition costs for given quantities of articles in each area (amount by price); and fifth, select a basis on which all the adjustment factors for each area will be reported. a. Because there is no information for prices for all the regions of the country, an index was prepared based on data from the survey itself. The items to be included in the index should be relatively homogenous in all areas. The items considered as relatively homogenous, and for which we have information, are the foodstuffs reported in LSMS98 Section 9A and the information from the price questionnaire. b. After selecting the food items6, the average national per capita cost of the selected items was estimated. c. For each census segment, the average price for the items included was estimated, first of all using the prices for the items purchased by the households, and in their absence, the prices noted in the price questionnaire. When a census segment did not report a particular item and the information could not be found in the price questionnaire, the average for larger areas that included that census segment were used instead. d. The cost of purchasing the average national quantity was estimated in each census segment, using the prices from each segment (or an estimate). 6 This included: cornmeal/oatmeal, tortillas, nacatamales, corn, plain bread, sweetened bread, crackers/cookies, rice, pasta, coffee, beef, pork, bones (beef/pork), chicken, fish cutlets, tuna, sardines, canned goods, milk (pasteurized/from the cow), powdered milk, cheeses, butter/margarine, eggs, vegetable oil, pork lard, white onions, yellow onions, garlic, sweet peppers, tomatoes, potatoes, beans, cassava, sugar, salt, canned juice, liquor, beer, ice cream, and sherbet. Annex 2, Page 4 e. Although traditionally the area of the capital is selected, it was felt that for this consumption aggregate the basis should be the national average. The average national cost of purchasing the selected items was estimated, and this amount became the basis. The purchase costs i I each census segment were divided by the basis, resulting in an adjustment factor to be used for the purpose of division. 19. With the geographic price index available to us, we adjusted the per capita consumption, dividing it by the index and obtaining the comparable per capita comprehensive consumption aggregate for all households included in the survey. 20. Lastly. in order to report the results nationally, the use of expansion factors is required. Trhe expansion factors are the inverse of the probability of selection for each household and are the number of households/persons in Nicaragua represented by each household/person in the sample. These expansion factors were calculated by INEC and were adjusted slightly in order to take into account those houselholds for which the consumption aggregate could not be estimated for lack of data or because the data were not very reliable. DETERMINING POVERTY LINES 21. The poverty lines for the LSMS98 were determined in such a way that, without changing the results reported in the '93 survev, it was possible to compare the percentages of the poor and extremely poor population for both years. In order to be able to elaborate poverty lines, it was necessary to create two new consumption aggregates. These should not be confused with the "comprehensive consumption aggregate" described earlier. Common consumption aggregates 22. In order to make a comparison between both surveys, two consumption aggregates were prepared, one with data from the LSMS93 and the other with data from the LSMS98. These consumption aggregates utilized only the questions considered to be "common" or similar in botl surveys. It should be kept in mind that these consumption aggregates were 6onstructed with only one objective, to compare both surveys, and they should not be utilized individually. In order to differentiate these consumption aggregates they were called "common" consumption aggregates. 23. The common questions allowed for making a calculation of common foodstuffs. a household consumption very similar to that described earlier! common education expensesd, and others. including ceremonies, club dues, lotteries, domestic services, shoes, sheets/blankets, towels, etc., kitchen utensils, soap, detergent, transport, and communications. 24. For each year, all the components were added up to obtain a total common consumption aggregate for each household. Tortilla, cob corn, corn, plain bread, sweetened bread, rice, coffee, beef, chicken, fish cutlets, powdered milk, eggs, vegetable oil, plantains, cooking bananas, white onions, yellow onions, tomatoes, potatoes, beans, sugar, salt, cigarettes. 8 Including pre-school expenses (monthly fee, food, pre-enrollment registration, uniforms, books, fees, school supplies). For the rest of the students, only pre-enrollment, enrollment, uniforms, and books could be included. Annex 2. Page 5 Transformations 25. Similar to what was described earlier, the data were transformed utilizing the number of household members per household to obtain the common per capita aggregates. 26. To take into account the price differences bv geographic area of the country. the same method described earlier was used, but this time Managua was chosen as the basis (fifth point in the previous description). 27. Furthermore, the common per capita aggregate for 1998 was modified to take into account the change in purchasing power of the cordoba between the LSMS93 and the LSMS98. For this transformation, the month of the interview was determined for each household, and the change in the price index9 during the period from LSMS93'0 and the price index for the month prior to the LSMS98 interview were compared. 28. This work was done separately for each component of the 1998 common aggregate. The nine components were: 1) food and beverages. 2) clothing and footwear, 3) housing, 4) furniture, accessories, household goods. 5) medical services and health care, 6) transport and communications, 7) leisure and recreation, 8) teaching, and 9) others. 29. Dividing the index for the month prior to LSMS98 by the average index for 1993 gives the "deflation" factor, with which the common consumption components for '98 will be modified. (The common consumption components will be divided by the "deflation" factor and the results for each household will be added up). By making this latter adjustment, we obtain the common per capita aggregate for 1998 in 1993 c6rdobas, allowing for comparisons to be made between both common aggregates. Poverty Lines 30. In 1993 c6rdobas and in per capita annual terms, the extreme poverty line is C$1,233 and the overall poverty line is C$2,609. According to what was estimated on the basis of the LSMS93 in the World Bank Poverty Report. 19.4% of the population was classified as extremely poor and 50.3% as overall poor. 3 1. To calculate comparative poverty lines, we first ordered the values for the common per capita aggregate for '93 from greatest to least and we determined which are the values that would divide the population into groups of 19.4% (Value 1) and of 50.3% (Value 2). Figure 1 is a graphic depiction of how to carry out this process. 32. Second, we applied these two values (Value I and Value 2) to the common aggregate for 1998 converted to 1993 prices and ordered it from greatest to least. These values applied to the common aggregate for 1998 classify the population in 17.3% and 47.9%. These are the percentages for poverty in 1998 that we consider can be compared to the 1993 results. 33. The final step is to determine the value corresponding to the poverty lines to be applied to the 1998 comprehensive consumption aggregate (not the common aggregate). Here again we make 9 The price index used is the one for Managua published by the Central Bank. '0 The survey was conducted between February and June 1993, but in order to take into account that the questions referred to earlier consumption, the price indexes for the previous month were used (January to May 1993). Annex 2. Page 6 a determination verv similar to that made with the 1993 common aggregate. The objective is to obtain the values for the povertv lines of the 1998 (per capita) comprehensive consumption aggregate that class,fy 17.3% and 47.9% of the population as extremely poor and overall poor, respectivelv. 34. This is achieved by ordering the values for the 1998 per capita comprehensive aggregate from greatest to least and determining the values for the poverty lines. Figure 2 represents this process. giving values of C$2.246 for extremely poor and C$4,259 for overall poor, whicih thereby classifies 17.3% as extremely poor and 47.9% as overall poor. INTERPRETATION AND PRECAUTIONS 35. After preparing statistical proofs between the levels for overall and extreme poverty nationally, it was found that both categories of poverty had been lowered statistically to a probability of 5% or less.'' On making the proofs at the level of rural areas and urban areas, reductions at the rural level were significant. but the change in poverty levels in urban areas was riot significanit. 36. Much care should be taken in interpreting these results since small changes in the assumptions in the technique utilized may change the conclusions in regards to the change of poverty in Nicaragua. If we make the comparison of the change of poverty nationally in a more rigorous way, for example to a I% significant probabilitv, the conclusion would be that no chang,- can be detected in overall poverty or in extreme poverty. 37. At least three alternative methods have been tried in comparing the results from 1993 ancl 1998. and so far the results observed show the same tendencies reported, although to differing degrees. 38. Lastly, although the differences found in poverty nationally are at most marginal, it is felt that the methodology applied is based on solid technical know-how and represents as much as possible the situation of Nicaragua. " The degree of probabilitv expresses the probability of being wrong when deternining that poverty has been reduced. Annex 2. Page 7 FIGURE A2.1: DETERMINATION OF EXTREME AND GENERAL POVERTY LINES USING THE 1993 COMMONAGGREGATE Highest level of 1993 Percentage common consumption of population aggregate accumulated C$ 61,820 100% Given the percentages of the population of 50.3% and 19.4%. the values which would obtain those same percentages in the 1993 common aggregate are determined 50.4% Midpoint= Valuel 50.3% 19.5% Midpoint= Value 2 19.4% C$207 1% Lowest level of Percentage of common consumption population 1993 Accumulated Annex 2. Paw:e 8 FIGURE A2.2 - DETERMINATION OF EXTREME AND GENERAL POVERTY LINES USING THE 1998 COMMON AGGREGATE Highest level of 1998 Percentage common consumption of population aggregate accumulated CS 177,095 100% Given the percentages of the population of 50.3% and 19.4%, the values which would obtain those same percentages are determined C$ 4,260.6 48.0% Midpoint CS 4,259 = overall povertv line C$4.258.6 47.9% CS 2,246.5 17.4% Midpoint CS 2,246 = extreme poverty line CS 2,245.0 17.3% CS 239 1% Lowest level of Percentage of common consumption population 1998 Accumulated Annex 2, Page 9 B. POVERTY IN NICARAGUA AFTER HURRICANE MITCH 39. To measure the poverty impact of Hurricane Mitch in Nicaragua, the households included in the 1998 LSMS located in "affected12" areas were interviewed for a second time in May 1999 using a questionnaire containing the same sections needed to construct the consumption aggregate. These houseliolds will be referred to as the "Post-Mitch" houseliolds and there is data for them from 1998 and 1999. 40. To measure the levels of poverty after Mitch. the results for the 1999 Post-Mitch households were substituted in the 1998 data set and new poverty rates were calculated. The resulting poverty rates are an estimate of poverty in Nicaragua in 1998 if Mitch had happened before the 1998 survey. In other words. we combine the conditions in 1998 for the households in areas not '"affected" by Mitch with the conditions in 1999 for the households in areas "affected" by Mitch. 41. In order to substitute the Post-Mitch household results, in the 1 998 data set, several ad justmets were needed. First, it was necessary to select or pair the Post-Mitch households from 1998 and 1999. Second. new expansion factors were calculated to take into consideration mismatches. Third, the 1999 results were adjusted to take into consideration the change in the purchasing power of the c6rdoba (inflation). Fourth, the results were adjusted by the regional or geographic price index. ADJUSTMENTS Pairing the Post Mitch households 1998 and 1999 surveys 42. The Post-Mitch survey included 595 households. Of these, 505 were paired with the 1998 data without any! problem. 26 households surveyed after Mitch did not have a consumption aggregate in 1998, 16 households had incomplete interviews in 1999, and 48 were 'split" households from 1998. 43. The "split"l households were cases in which the original 1998 household members split up into two or three households. The rule for selecting the Post-Mitch household to substitute in the 1998 data set was to use that Post-Mitch household in which the household head (from 1998) lived, or in the absence of the household head, the Post-Mitch household with more of the original members from 1998. By applying these criteria. 23 Post-Mitch households were selected for substitution in the 1998 data set and 25 were removed. 44. In the end. 528 of the new interviews were used to estimate the new poverty rates. Since the original data set had 583 households in "affected" areas, there were 55 Post-Mitch-households without data for 199914. Computing the new expansion factors 45. The expansion factors were adjusted only for households in "affected" areas. For the remaining households the expansion factors did not change. This was necessary because some 1998 t2 "Affected" areas in Nicaragua were determined at the segment level by INEC in Managua, Nicaragua. 3 Expressed by the consumption aggregate. 14 Some of these were households with incomplete interviews in 1999, but most correspond to households not found in 1999. Annex 2, Page 10 households in "affected" areas were not located in 1 999 and some households had an increase or decrease in the number of members. 46. The adjustment was performed at the Primary Sampling Unit (PSU) level. The PSUs can le identified by a combination of the Department (iO 115), Municipality (i02), Area (i05). and Segment (i04) variables. The adjustment is made in such a way that each Post-Mitch PSU did expand to the same number of persons in both years. 47. To calculate the new expansion factor, first the number of persons sampled in each 1998 Post-Mitch PSU was computed (from 583 households). Second. the number of persons sampled n each 1999 Post-Mitch PSU was computed (from 528 households). The "adjustment factor" for each PSU is obtained by dividing the first result by the second. To obtain the new expansion factors. the old expansion factors were multiplied by the "adjustment factor." Adjusting for inflation between survev periods 48. Since a 1999 c6rdoba does not have the same value as a 1998 c6rdoba, all expenditures had to be adjusted to take this into consideration. 49. The Consumer Price Index (CPI) published by the Central Bank of Nicaragua was used to adjust the 1999 survev results. An "inflation" factor was calculated for each category published by the Central Bank and each component in the 1999 consumption aggregate was modified usilg the corresponding "inflation" factor. The different categories and the reported CPI values are shown in Table A2.1. Table A2.1 Nicaragua CPI by category and "inflation factor" Category [ May 1998 May 1999 Factor a Food j 156 159 1.021 Clothing 108 123 1.136 Housing 190 249 1.307 Furniture and accessories 151 169 1.120 Health 134 140 1.042 Transport and 150 159 1.058 communications Leisure 144 153 1.060 Education 126 147 1.171 Personal Expenditure 122 130 1 .064 The 1999 values were divided by this number Adjusting for different prices in different areas in Nicaragua 50. To take into consideration the different prices in different regions of the country, the 1999 values were adjusted using the same factors calculated for 199816 POVERTY RATES 51. Since the 1999 consumption aggregates were adjusted to the time of the 1998 survey, the final poverty rates were calculated by replacing the 583 households (in the 1998 data set) in "affected" areas with the new data from 1999 (529 households). Then, the same poverty lines I5 The variable names correspond to the 1998 data set. 6 See the document XXXXX (how the consumption aggregate for 1998 was constructed). Annex 2, Page 11 calculated for the original data set (C$2,246 for the extreme poverty line and C$4,259 for the overall poverty line) were applied. The new poverty rates were calculated using the adjusted expansion factors (Table A2.2). Table A2.2 Nicaragua-Pre- and Post-Mitch Poverty by region, 1998 & 1999 INCIDENCE OF AND CHANGE IN INCIDENCE OF AND CHANGE IN EXTREME POVERTY POVERTY Headcount Index (%) Headcount Index (%) Region 1998 1999 Change 1998 1999 Change National 17.3 17.3 -0.1 ns 47.9 47.9 0.1 ns Urban 7.6 7.5 -0.1 ns 30.5 30.3 -0.2 ns Rural 28.9 28.9 0.0 ns 68.5 69.0 0.4 ns Managua 3.1 3.1 0.0 X 18.5 18.5 0.0 X Pacific Urban 9.8 9.6 -0.3 ns 39.6 39.0 -0.6 ns Rural 24.1 20.6 -3.6 ** 67.1 63.1 -4.0 ** Central Urban 12.2 12.1 -0.1 ns 39.4 39.4 0.0 ns Rural 32.7 35.7 2.9 ** 74.0 77.6 3.6 ** Atlantic Urban 17.0 17.0 0.0 X 44.4 44.4 0.0 X Rural 41.4 40.6 -0.8 ns 79.3 80.6 1.3 ns Source: Nicaragua LSMS 1998 and LSMS 1999. Table notation:ns non-significant at p<=-10%; significant at p<=1%: and. X= Managua and Atlantic Rural did not have anv Post-MiHch surveyed households. C. FUTURE POVERTY SCENARIOS 52. Growth in mean per capita consumption could have a greater impact on reducing extreme than overall poverty. By estimating consumption elasticities of poverty measures (headcount, poverty gap, and Foster-Greer-Thorbecke indices) we can analyze their responsiveness to growth of mean per capita income.'7 Intuitively, assuming that the consumption distribution remains the same as in 1998, a I percentage point increase in mean per capita consumption decreases the headcount ratio of the extremely poor by 2 percentage points in but of the overall poor by only 1. 1 4 percentage points (Table A2.3). Note that the elasticity of each poverty index is greater for extreme poverty than for overall poverty in both urban and rural areas. This means that extreme poverty is more responsive to increases in mean per capita consumption. 17The estimations used POVCAL. Headcount means the proportion of the population that is poor. Annex 2, Page 12 Table A2.3 Nicaragua - Consumption Elasticities of Poverty Indices National Urban Rural Poverty Index Overall Extreme Overall Extreme Overall Extreme poverty povertv poverty poverty poverty poverty Head Count -1.14 -1.96 -1.44 2.88 -0.72 -1.44 Poverty Gap -1.61 -2 87 -2.02 -4.70 -1.19 -1.95 FGT2 -2.00 -3.77 -2'56 -6.52 -1.50 -2.42 Source: Nicaragua LSNIS 1998. 53. However, rural poverty is less responsive to economic growth and thus its reduction requires more comprehensive strategies. The response to growth in mean per capita consumption will not be the same in all areas of the country. Both overall and extreme poverty will decrease twice as fast in urban than in rural areas due to higher urban elasticities: a one percentage point increase in mean per capita consumption decreases the headcount ratio of the urban poor by 1.44 percentage points and of the rural overall poor by 0.72 percentage points. It decreases the headcount ratio of urban extreme povertv by 2.88 percentage points and of rural extreme poverty by 1.44 percentage points. 54. An optimistic scenario of growth in mean consumption of 2 percent annually reduces extreme poverty by half and overall poverty by one-third by 2015 (Table A2.3). On the basis of the elasticities in Table A2.2, it is possible to generate future poverty projections in Nicaragua under alternative scenarios of mean consumption growth. Between 1993 and 1998, Nicaragua's GDP experienced an average real growth rate of 4.3 percent per year; however, fast population growth brought the average real growth rate of GDP per capita to only 1.2 percent per year.18 Assuming an unchanged distribution of income,19 poverty rates will decline substantially by 2015. However. fertility rates among the poorest are much higher than among the non-poor: live births per woman in the poorest quintile (6.6) are more than three times those of the highest (1.9). This factor alone, even under distributionally neutral policies, will worsen the distribution of income. Thus, assuming distributionally neutral growth of mean household consumption might welJ be an optimistic scenario.2" This clearly points towards the need for a more vigorous stance on population policies focusing strongly on the poor, to reduce poverty and ultimately to distribute the fruits of growtth more evenly. 18 The source for GDP data is the Central Bank of Nicaragua (BCN); and for population growth rates, the Institute of Statistics and Census (INEC). 19 We use consumption, rather than income, as a measure of welfare in this report. See Annex I for an explanation of the consumption aggregate. However, they are used interchangeably in this section. 20 Scenarios with higher annual increases in mean household consumption (such as 2 or 3 percent in Table A2.3), necessarily assume either high rates of economic growth (4.7 percent to 5.7 percent by adding a 2.7 percent constant rate of population growth) or declining rates of population growth (between 2.3 percent to 1.3 percent by subtracting a 4.3 percent constant rate of economic growth such as that of 1993-1998). Annex 2. Page 13 Table A2.3 Projected Levels of Poverty with different scenarios for mean household consumption (assuming no change in income distribution) 1.2% annual 2% annual increase 3% annual increase Poverty Index increase in mean in mean household in mean household household consumption consumption consumption 1998 2005 2010 2015 2005 2010 2015 2005 2010 2015 NATIONAL Overall Poverty: Head Count 47.9 43.5 40.6 37.9 40.8 36.3 32.4 37.6 31.6 26.5 Poverty Gap 18.3 16.0 14.5 13.1 14.6 12.4 10.5 12.9 10.1 7.9 FGT2 9.3 7.8 7.0 6.2 7.0 5.7 4.7 6.0 4.4 3.3 Extreme Povertv: Head Count 17.3 15.7 14.7 13.7 13.1 10.7 8.8 11.3 8.4 6.2 Poverty Gap 4.8 4.2 3.8 3.5 3.8 3.2 2.8 3.4 2.7 2.1 FGT2 2.0 1.6 1.5 1.3 1.5 1.2 1.0 1.3 0.9 0.7 URBAN Overall Poverty: Head Count 30.5 27.0 24.7 22.7 24.8 21.5 18.6 22.4 18.0 14.4 Poverty Gap 9.9 8.3 7.3 6.5 7.4 6.0 4.9 6.4 4.6 3.4 FGT2 4.5 3.6 3.1 2.6 3.1 2.4 1.8 2.6 1.7 1.2 Extreme Poverty: Head Count 7.6 6.8 6.2 5.7 5.0 3.7 2.8 4.1 2.6 1.6 Poverty Gap 1.9 1.6 1.4 1.2 1.0 0.6 0.4 0.7 0.3 0.1 FGT2 0.7 0.5 0.5 0.4 0.3 0.1 0.1 0.1 0.0 0.0 RURAL Overall Poverty: Head Count 68.5 64.5 61.7 59.1 61.9 57.6 53.5 58.8 52.7 47.2 Poverty Gap 28.3 25.6 23.8 22.2 23.9 21.2 18.8 21.9 18.3 15.3 FGT2 14.9 13.1 12.0 11.0 12.1 10.4 8.9 10.8 8.6 6.8 Extreme Poverty: Head Count 28.9 27.2 26.0 24.9 23.6 20.4 17.6 21.2 17.0 13.7 Poverty Gap 8.3 7.5 7.0 6.5 6.3 5.1 4.2 5.4 4.0 3.0 FGT2 3.5 3.1 2.8 2.5 2.5 1.9 1.5 2.0 1.4 1.0 Source: Nicaragua LSMS 1998. Annex 3, Page I Annex 3 - The Income Aggregate by Carlos Sobrado INTRODUCTION 1. In 1998, a representative national and regional' Living Standard Measurement Survey (LSMS) was conducted in Nicaragua. The survey included all the questions necessary for constructing an income aggregate using all incomes obtained by each households in a year and for computing adjustment indexes that take into consideration family size, sampling design, and geographical price differences. 2. To make comparisons between the consumption aggregate results -including poverty classifications- and the income aggregate, the latter was calculated for the same 4,040 households used to obtain the consumption aggregate. Also, the same adjustment indexes used for consumption were applied to the income aggregate. COMPONENT OF TOTAL INCOME Overview 3. Total income can be divided into two mayor components: labor and non-labor income. Labor income can be derived from agricultural activities (wage or own production) or from non-agricultural activities (wage or own business). Non-labor income includes equipment and property leases (including owned housing), interest from savings and investments, pensions, and donations, transfers, and gifts received in cash or kind. The data sets were reviewed beforehand to identify outliers and to estimate or re- codify missing data. BOX 1 - Components of Total Income LABOR INCOME From agriculture: wage labor or own production From non-agricultural activities: wage labor or own business NON-LABOR INCOME Own house use value Food received as gift Remittances received Charity received Returns to capital: equipment, property, savings, and investments. Pensions Other income 'For the regions of Managua, Pacific Urban and Rural, Central Urban and Rural, and Atlantic Urban and Rural. Annex 3, Page 2 Labor Income 4. Labor earnings were derived from sections 5 (economic activity), 9 (expenditures and other, income) and 10 (independent agricultural activities) of the LSMS questionnaire. Although the Nicaragua poverty report organizes labor income into agricultural and non-agricultural sources (and wage and non- wage within each one), this report is presented following the LSMS questionnaire sections. The allocation of the variables is obvious and no further explanation is provided. Economic activity 5. This section has labor income from up to three specific jobs and for all other jobs (after the third). The variables included and the reference time period were not the same for the different jobs. Incomie variables with reference time periods other than "in the last 12 months" were first converted to a yearly basis and then adjusted by the length of the job (time worked in the specific job during the last 12 months). Table I provides a guide of the variables for each job and their characteristics. Table A3.1 - Labor question number for different jobs Information 15'job 2nl job 3rd job Other jobs Length ofthejob' 15A&B 35A&B 51 A&B Net Income Value 19 A 38 A 54 A 5(B Frequency 19 B 38B 54 C 59 2 TOTAL 42 B' 58 B-1 Food 23 B4 In kind Income Housing 24 B4 Transport 26 B4 Uniforms: Value 25 B Frequency 25 C Tips, overtime, etc. 21 B4 40 B4 56B 4 13t' month (Christmas bonus) 22 B 41 B 57 B To classify the job: "branch" and "position" 13 & 20 33 & 39 49 & 55 '260 working days per year (five per week). 2 Number of times in the last 12 months. 3 In the last 12 months. 4 In the last month. 6. In the "branch" variable agricultural jobs were coded with a value of 1; non-agricultural jobs have other values. In the variable "position" wage earners have values of 1, 2, or 7, while self-employed people have values from 3 to 6. From this classification, non-agricultural earnings (wages and business) and agricultural wage incomes were selected. Expenditures and other income 7. The value of food produced and consumed at home was computed from section 9.A of the LSMS questionnaire. A value of I in question 7 classifies food values as "own production". The total annual value of food produced and consumed at home was calculated from the consumption frequency reported in question 8 and the total value of consumption in question 10. Annex 3, Page 3 independent agricultural activities 8. Income from agricultural activities was calculated without differentiating between farming, forestry, animal production, or by-products. The net income is the result of subtracting the production costs from the gross income". Production costs 9. The first production cost is the rent paid for land used. This value was reported on an annual basis in Section 1 O.A, question 30.The second cost is for agricultural inputs. Annual input costs for farming are in Section I O.C.2 question 69; for animal production, costs are in section 1 O.D question I I 1. Third, labor costs for all activities were computed from section IO.F.1, questions 119 through 124. Total yearly values for day laborers, permanent workers, and lump sum payments were included. Fourth, annual expenditures in transport, storage, fuel, machinery repairs, sub-products elaboration, and machinery and animals rented were calculated from section I O.F.2, question 126 (excluding items 1, 2, and 3). Fifth, the annual use value of owned equipment and machinery was calculated from sections I O.F.3 and 1 O.F.4. The method used to compute the annual use value was the same as that used for the household durable goods and can be expressed with the formula: Yij = Vij, [2 (Avgj AGEij) - AGEij] were: Yij = Yearly use value for item i, in household j V~j = Present value of item i, in household j Avgj = Function: average for all j households (from j=l toj=n) AGE,j= Age of item i, in household j 10 The total annual production cost is the sum of the five components for each household. Gross income 11. The first gross income is revenues from rented land from section I O.A. 1, question 13. The second is the sale value of produced crops recorded in section I O.C. 1, question 63. The third is the annual value of animals sold, computed by adding the reported values in questions 96.B and 1 06.C from section 1 O.D. The fourth is the total value of sub-products sold computed by adding questions 46 and 47 in section l0.B and 116.B in section IO.E. 12. The total annual gross income from agriculture is the sum of the four components for each household and the value of food produced and consumed at home (computed from section nine). 2 Gross income includes the value of food produced and consumed at home described previously. Annex 3, Page 4 Non-Labor Income Own house use value 13. The use value of owned property is an income3. Most homeowners imputed themselves the rental value of their property shown in question 17, section I (House characteristics). Some households di i not provide any data about the use value of their house. In these cases an estimate was made using their household characteristics and expenditures. To impute the use value for these households, two steps were taken. First, we selected the households with rental information (real or self-reported), and estimated a multi-variable linear regression between the monthly rent value (dependent variable) and the selected household characteristics and expenditures (independent variables). A list of the variables and the estimated coefficients is provided in Table 2. Second, for households with no type of rental information, the rental value was imputed using their household characteristics and expenditure levels and the parameters estimated and reported in Table 2.5 Table A3.2 - Estimated coefficients from households with no type of rent information Variable Estimate 2 t-value Definition Question Variabl significance e Constant -66.117 <0.0005 # of rooms (exclude kitchen, bathroom, 10 VblO 87.385 <0.0005 etc.) Urban or Rural household 5 105 -83.314 <0.0005 Water bill (last month) 22 VB22 1.683 <0.0005 Flush toilet in the house 28 VB28 109.608 <0.0005 Garbage disposal fee (last month) 37 VB37 14.244 <0.0005 Electric bill (last month) 40 VB40 1.248 <0.0005 Telephone bill (last month) 50 VB50 2.251 <0.0005 For the Urban or Rural classification the question number and variable name are found in the general household ID section (first page). For all other variables, the question number and variable names are located in Section 1. 2 With an adjusted R2 of 0.36, an F value with an error probability below 0.001 Data source: LSMS Nicaragua 1998 14. The annual use value of the house was calculated only for homeowners by multiplying the self- reported figure or the one imputed by 12. Food received as gifts 15. The value of food received as gifts or donations was reported in Section 9.A, question 10 (and a value of 4 in variable 7). The annual values are the result of multiplying question 10 by the corresponding frequency reported in question 8. Food received at school was also considered as income. The monthly value of food consumed at school was reported in section 4.B (Education), question 19. Taking into 3 One way to look at this point is to imagine that the homeowner is renting the property to him/herself. 4It was necessary to include the households with self-estimated rent information due to the low number of households with real rent information. 5For eight cases it was necessary to set a minimum of ten c6rdobas per month because the estimated value was negative. Annex 3, Page 5 consideration the length of the school year, the total yearly value for this category was calculated as the product of the value reported in question 19 multiplied by 8.. Remittances received 16. Remittances from friends and family members were reported in section 9.D.1, item 4. To obtain the annual value, the reported quantities were multiplied by 12. Chiarity received 17. Charity received from institutions in kind or cash was reported in section 9.D.2, item 8. The quantities were reported on an annual basis and no transformation was necessary. Capital gains 18. Capital gains were included in section 9.D. I, items I (rent received from properties) and 2 (rent received from machinery and equipment). For the annual value, the reported values were multiplied by 12. Also, section 9.D.2 had capital gains information reported on an annual basis: interest from savings (item 1), interest from loans (item 2), and stock dividends (item 5). Pensions 19 Pensions were reported in section 9.D. I (items 5, 6, and 7) and employment compensation in section 9.D.2, item 4. The pension information was multiplied by 12 to obtain the yearly value. Othler income 20. Other income included student scholarships (section 9.D. 1, item 3) and that reported in section 9.D.2: insurance compensation, lottery winnings, job accident compensation, inheritances, and other income (items 3, 6, 7, 9, and 10). The value of student scholarships was multiplied by 12 to obtain the yearly value. ADJUSTMENT TO TOTAL INCOME 21. Total income is the sum of all the individual income components for each household. The same adjustment indexes computed for total consumption were used for total income. Total income was reported in per capita terms, adjusted by geographical price differences. Any statistic produced takes into consideration the probability of selection by means of the expansion factor. COMPARISON OF INCOME AGGREGATES AGRICULTURE SOURCES OF INCOME 22. Table A3.3 below shows a comparison for the Agriculture Sources of Income between the Income Aggregate used in this study and that generated by FAO.7 Both income aggregates for agriculture 6 There school year is eight and a half months long. 7Corral, Leonardo and Thomas Reardon. 1999. Rural non-farm and farm incomes in Nicaragua: evidence from the 1998 LSMS. Draft. (Leonardo.corralai),fao.org). FAO/RLC and Michigan State University. Annex 3, Page 6 sources of income are very similar and at the national level the difference is only around 1 percent. The FAO income aggregate does the following adjustments: (a) eliminates 166 households without consumption aggregate; (b) adjusted by geographical price differences; and, (c) the used the appropriate expansion factors. Table A3.3 - Comparisons of Income Aggregates - Agriculture Income Only (Annual per capita May 1998 Cordobas) Income Aggregate Source Urban Rural Total Nicaragua Wage agriculture 158 659 .87 Poverty Assessment Self-employment agriculture 213 958 ' 53 Total 371 1617 940 Wage agriculture 223 706 4 44 FAO Self-employment agriculture 178 891 . 03 Total 401 1597 'i47 Total value 30 -20 7 Difference Total percentage 7% -1% I % Annex 5 - Statistical Appendix List of Tables PART A - POVERTY AO1 Nicaragua 1993 - 1998 Contribution to Poverty A02 Nicaragua 1993 - 1998 Headcount Rates A03 Nicaragua 1993 - 1998 Poverty Gap & FGT P2 A04 Nicaragua 1993 - 1998 Extreme Poverty Gap & FGT P2 A05 Nicaragua 1998 - Poverty Populations by Areas and Regions A06 Nicaragua 1998 - Poverty Populations by Regions A07 Nicaragua 1998 - Poverty % (Regions) A08 Nicaragua 1998 - Poverty Population by Areas A09 Nicaragua 1998 - Poverty % (Areas) AIO Nicaragua 1993 - 1998 LSMS - Significant Head-Count Ratios All Nicaragua 1993 - 1998 LSMS - Significant Head-Count Ratios: Original-Final Urban A 12 Nicaragua 1993 - 1998 LSMS - Significant Head-Count Ratios: Original-Final Rural A 13 Nicaragua 1993 - 1998 LSMS - Confidence Intervals A14 Nicaragua 1998 - LSMS Calories Provided by C$2,246 for Three Different Population Groups A 15 Nicaragua 1993 - 1998 LSMS - Comparing the Dollar Value of Poverty Lines A 16 Nicaragua 1993 - 1998 LSMS -Consumer Price Index: 1994=100% A 17 Nicaragua 1993 - 1998 Comparing the Cordoba Value of Poverty Lines A 18 Poverty measures and Change from 1998 to Post Mitch A 19 Measures and Change from 1998 to Post Mitch for Extreme Poverty A20 Nicaragua 1998 - 1999 LSMS Post Mitch Segment A21 Nicaragua 1998 - 1999 LSMS Post Mitch Segment A22 Nicaragua 1998 - 1999 LSMS Post Mitch Segment PART B - CONSUMPTION BO] Nicaragua 1998 - LSMS Consumption Patterns (Comprehensive) B02 Nicaragua 1998 - LSMS Annual Average Value of Consumption Per Capita (Comprehensive) B03 Nicaragua 1998 - LSMS Annual Average Value of Consumption Per Capita (Comprehensive) by Poverty Group and Area B04 Nicaragua 1993 - 1998 - Self-Consumption from Agriculture Only PART C - INCOME COI Nicaragua 1998 - LSMS Average Income C02 Nicaragua 1998 - LSMS Average Income C03 Nicaragua 1998 - LSMS Average Income C04 Nicaragua 1998 - LSMS Average Income PART D - DEMOGRAPHIC CHARACTERISTICS DOI Nicaragua 1993 - Demographic Characteristics D02 Nicaragua 1998 - Demographic Characteristics by Poverty Group D03 Nicaragua 1998 - Demographic Characteristics, Urban and Rural Area D04 Nicaragua 1998 - Demographic Characteristics by Quintile D05 Nicaragua 1998 - Demographic Characteristics by Zone PART E - EDUCATION EOI Nicaragua 1998 - Gross Enrollment Rates - Preschool, Primary and Secondary by Gender E02 Nicaragua 1993 - 1998 Gross Enrollment Rates for 1993 and 1998 / Primary and Secondary by Gender E03 Nicaragua 1993 - 1998 Net Enrollment Rates for 1993 and 1998/ Primary and Secondary by Gender E04 Nicaragua 1998 - Net Enrollment Rates - Preschool, Primary and Secondary by Gender E05 Nicaragua 1993 - Reason for not Attending School E06 Nicaragua 1998 - Reason for not Attending School E07 Nicaragua 1993 - 1998 Percent not attending School E08 Nicaragua 1993 - 1998 Preschool Attendance of Children 0-3 and 4-6 Years Old E09 Nicaragua 1998 - Primary School Repetition Rates, Percent with no Books and Mean Number c f Days Absent. E10 Nicaragua 1998 - Secondary School Repetition Rates, Percent with no Books and Mean Number of Days Absent. El I Nicaragua 1993 - School Attendance of Children 6-18 Years Old E12 Nicaragua 1998 - Household Expenses on Education by Educational Level and Poverty Group (public school enrollments only) PART F - HEALTH FOI Nicaragua 1993 - Fertility by Quintile, Poverty Status and Urban/Rural (Women 15-49 Years of Age) F02 Nicaragua 1998 - Fertility by Quintile, Poverty Status and Urban/Rural (Women 15-49 Years of Age) F03 Nicaragua 1993 - DPT and Polio immunization by Quintile, Poverty Status and Urban/Rural (% of 12-23 months of age with card) F04 Nicaragua 1998 - DPT and Polio Immunization by Quintile, Poverty Status and Urban/Rural (°/, of 12-23 months of age with card) F05 Nicaragua 1998 - DPT and Polio Immunization by Quintile, Poverty Status and Urban/Rural (°/0 of 12-23 months of age) F06 Nicaragua 1993 - Incidence of Diarrhea F07 Nicaragua 1998 - Incidence of diarrhea and type of care of those reporting diarrhea (Children under 6 years of age) F08 Nicaragua 1998 - Reason for not seeking care of those reporting Diarrhea last month (Children under 6 years of age) F09 Nicaragua 1998 - Of those consulting for Illness, time spent waiting for medical attention by facility. FIO Nicaragua 1998 - Of those consulting for Illness, cost of round trip transportation for last consultation, by facility. F l I Nicaragua 1998 - Of those consulting for Illness, total cost of last consultation, by facility F 12 Nicaragua 1998 - Of those consulting for Illness, other health expenditures last time, by facility F13 Nicaragua 1998 - Of those consulting for Illness, total cost of last consultation, by facility F14 Nicaragua 1998 - Reason for not seeking care of those ill last month F15 Nicaragua 1998 - Maternal Health by Poverty and Region F 16 Nicaragua 1998 - Place of Consultation by Poverty Group, Quintile and Geographic Areas (include all ill and children under 6 reporting diarrhea). F17 Nicaragua 1993 - Place of Consultation by Poverty Group, Quintile and Geographic Area (include all ill and children under 6 reporting diarrhea) Fl 8 Nicaragua 1998 - Place of Consultation by Poverty Group, Quintile and Geographic Area F19 Nicaragua 1998 - Place of Consultation by Poverty Group, Quintiles and Geographic Area (include all ill excluding children under 6 reporting diarrhea) F20 Nicaragua 1993 - 1998 Place of Consultation by Poverty Group (include al ill and children under 6 reporting diarrhea) F21 Nicaragua 1998 - Health Services by Poverty Group, Quintiles and Geographic Area (includes all ill and children under 6 reporting diarrhea) F22 Nicaragua 1993 - Distance to Health Post/Center F23 Nicaragua 1993 - Adults Consulting when Ill (%) F24 Nicaragua 1998 - Household Expenses on Health by Level and Poverty Group (public facilities only) PART G - MALNUTRITION GO1 Nicaragua 1998 - Prevalence of Malnutrition by Quintile, Region and Poverty Status (children under 5 years of age) G02 Nicaragua 1998 Prevalence of Malnutrition by Age Group (children under 5 years of age) G03 Nicaragua 1998 Percent of Children (0-59 months) classified as malnourished by poverty and region G04 Nicaragua 1998 Percent of Children (0-59 months) classified as malnourished by poverty and age PART H - HOUSING & BASIC SERVICES HOI Nicaragua 1998 - Access to Services/Housing by Poverty Group H02 Nicaragua 1998 - Access to Services/Housing by Quintile H03 Nicaragua 1998 - Access to Services/Housing by Zone H04 Nicaragua 1998 - Access to Services/Housing by Urban/Rural Area H05 Nicaragua 1993 - Households (%) with inadequate walls, floor, ceiling, housing, overcrowding H06 Nicaragua 1998 - Household (%) with inadequate walls, floor, ceiling, housing, overcrowding H07 Nicaragua 1993 - Main Source of Water by Poverty Group H08 Nicaragua 1993 - Main Source of Water by Rural/Urban Area H09 Nicaragua 1993 - Main Source of Water by Quintile H 10 Nicaragua 1993 - Main Source of Water by Zone HI] Nicaragua 1998 - Main Access Road Condition since 1993 Annex 5. Page I Annex 5- Statistical Appendix PART A - POVERTY A.01-Contribution to poverty: Nicaragua 1993-1998 LSMS Extreme poverty All poverty Year Change Year Change 1993 1998 from 93 1993 1998 from 93 All Nicaragua 100.0% 100.0% 0.0% 100.0% 100.0% 0.0% Urban 22.0% 23.9% 2.0% 36.9% 34.6% -2.3% Rural 78.1% 76.1% -2.0% 63.1% 65.4% 2.3% Managua 7.2% 4.7% -2.6% 16.5% 10.1% -6.4% Pacific Urban 6.5% 9.5% 3.0% 11.0% 13.9% 2.8% Pacific Rural 23.0% 21.7% -1.3% 19.8% 21.8% 2.1% Central Urban 8.6% 7.4% -1.2% 10.7% 8.7% -2,0% Central Rural 45.6% 39.3% -6.3% 31.3% 32.2% 0.9% Atlantic Urban 1.8% 4.9% 3.2% 3.0% 4.7% 1.7% Atlantic Rural 7.3%. 12.5% 5.2% 7.8% 8.7% 0.9% A.02 Headcount rates: Nicaragua 1993-1998 LSMS Extreme poverty All poverty Year Change Year Change 1993 1998 from 93 1993 1998 from 93 All Nicaragua 19.4% 17.3% -2.1% 50.3% 47.9% -2.5% Urban 7.3% 7.6% 0.3% 31.9% 30.5% -1.4% Rural 36.3% 28.9% -7.4% 76.1% 68.5% -7.6% Managua 5.1% 3.1% -2.0% 29.9% 18.5% -11.4% Pacific Urban 6.4% 9.8% 3.5% 28.1% 39.6% 11.5% Pacific Rural 31.6% 24.1% -7.5% 70.7% 67.1% -3.6% Central Urban 15.3% 12.2% -3.1% 49.2% 39.4% -9.8% Central Rural 47.6% 32.7% -14.8% 84.7% 74.0% -10.7% Atlantic Urban 7.9% 17.0% 9.0% 35.5% 44.4% 8.9% Atlantic Rural 30.3% 41.4% 11.1% 83.6% 79.3% -4.3% Annex 5. Page 2 A.03 Poverty gap & FGT P2: Nicaragua 1993-1998 LSMS Poverty Gap FGT P2 Year Change Year Change 1993 1998 from 93 1993 1998 from 93 All Nicaragua 21.8% 47.9% 26.1% 12.1% 18.3% 6.2',b Urban 10.9% 30.5% 19.6% 5.1% 9.9% 4.8c1, Rural 37.1% 68.5% 31.5% 21.9% 28.3% 6.4t'o Managua 9.5% 18.5% 9.0% 4.2% 5.1% 0.91'o Pacific Urban 9,4% 39.6% 30.1% 4.2% 12.6% 8-4% Pacific Rural 32.3% 67.1% 34.9% 18.7% 26.0% 7,3Cvo Central Urban 19.4% 39.4% 20.0% 10.0% 14.3% 4.3% Central Rural 45.3% 74.0% 28.7% 28.1% 30.9% 2.8c)o Atlantic Urban 12.1% 44.4% 32.3% 6.0% 17.5% 11.5% Atlantic Rural 35.9% 79.3% 43.4% 18.8% 37.3% 18.5%'o A.04 Extreme poverty gap & FGT P2: Nicaragua 1993-1998 LSMS Poverty Gap FGT P2 Year Change Year Change 1993 1998 from 93 1993 1998 from 93 All Nicaragua 5.9% 4.8% -1.1% 2.6% 2.0% -0.7To Urban 1.6% 1.9% 0.3% 0.6% 0.7% 0.1` Rural 11.8% 8.3% -3.5% 5.50/0 3.5% -2.0`%t Managua 1.1% 0.6% -0.4% 0.4% 0.2% -0.2/o Pacific Urban 1.2% 2.3% 1.2% 0.4% 0.8% 0.5'/o Pacific Rural 10.0% 6.0% -4.1% 4.6% 2.3% -2.3:/o Central Urban 4.0% 3.5% -0.5% 1.6% 1.4% -0.2"%o Central Rural 16.3% 9.8% -6.5% 7.8% 4.0% -3.8% Atlantic Urban 2.4% 4.2% 1.8% 1.0% 1.4% 0.4%1/o Atlantic Rural 7.4% 13.6% 6.2% 2.5% 6.6% 4.1'/c Annex 5. Page 3 Table A.05 Nicaragua 98 Poverty Populations by areas and regions (thousands) Extreme Poor Urban Rural TOTAL REGION Urban Rural Poor No Poor Poor No Poor Managua 17.2 21.5 171. 88. _ 67 1. 1,253.8 Pac. Urb. 79.3 00 319.1 47200.0 806.3 Pac. Rur. 0.0 180.8 0.0 0.0 503.2 246.3 749.5 Cent. Urb. 61.~9 00 199.3 306.6 0.0 0.0 505.9 Cent. Rur. 0.0 328.2 0.0 0.0 741.7 260.6 1,0=. Alt. Urb. 41.21 0.0 10 7.9 135.2 0.0T 0.0 243.1 Atl. Rur. 0.0 104.5 0.0 0.0 200.4 52.2 252.6 TOTAL 199.6 63b.01 YT.4 8t 8.4 - 1,506.0= 691.7 4,813.5 Table A. 06 Nicaragua 98 Poverty Populations by regions (thousands) Pobre Pobre No Pobre TOfAL REGION extremo Managua 38.7 21 1,. 1,253.8 Pacifico Urbano 479. 319.1 4 8 Pacifico Rural 180.8 503.2 246.3 749.5 Central Urbano 61.9 199.3 306.6 505.9 Central Rural 328.2 741.7 260.6 1,002.3 Atlantico Urbano 412 107.9 1- W. *-743.1 Atlantico Rural 104.5 200.4 52.2 252.6 {TOTAL T 834.63 ---T=T ,T11 ,3 Table A.07 Nicaragua 98 Poverty % (Regions) Pobre Pobre No TOTAL REGION extremo Pobre Managua 4.6 10.1 --TU-7o -47 26.1% Pac. Urb. 5 1 9. 16.7 /o Pac. Rur. 21.7% 21.8% 9.8% 15.6% Cent. Urb. 74/ 8.7% 12.2% 10.5% Cent. Rur. 39.3% 32.2% 10.4% 20.8% Alt. Urb. 4T T-/c 4.7% 5.4% 5.1% Atl. Rur. 12.5% 8.7% 2.1% 5.2% TOTAL 100/c 100.0% 1 0 T% o Table A.08 Nicaragua 98 Poverty Populations by areas (thousands) Extreme Poor No Poor I lAL REGION Poor Urban 19. 797. T1,815 Rural . 1,506.0 691.7 2,197.7 TOTAL 2,5101 Annex 5, Page 4 Table A. 09 Nicaragua. 98 Poverty % (Areas) Extreme Poor Non TOiAL REGION Poor Poor Urban 23.9% 34.6% 72.4% 54.3% Rural 76.1% 65.4% 27.6% 45.7% TOTAL 0 1010.0%o 100.0% o Table A. 10 Nicaragua LSMS 93-98 - Significant Head-Count Ratios Head count Sample size Extreme P. General P, 1993 . /0 .Z 4,20 1998 17.3% 47.8% 4,040 Difference 2 1% 2.5% Standard deviation * 0.0110 0.0085 t value * 2.27 2.46 Signiticance level 24b-/o 1.54% M. tavallion Poverty comparisons: A Guiue to Co ncepts and Methods LMS N o. 88 p. 49 (9) Table A. 11 Nicaragua LSMS 93-98 Significant Head-Count Ratios: Original-Final Urban Head count Sample size Extreme P. General P. 1993 7.3% 31T0 2 1998 7.6% 30.5% 2,187 Difference -0.3% 14%T Standard deviation * 0.0137 0.0078 t value ' 1.02 -0.39 Signi-icance level >5o >5% 1 M. Ravallion Poverty comparisons: A Guide to Concepts and M ethods LSMS No. 88 p. 49 (9) Table A. 12 Nicaragua LSMS 93-98 Significant Head-Count Ratios: Original-Final Rural Head count Sample size Extreme P. General P. 1993 36. 3- - 761%,8 1998 28.9% 68.5% 1,853 DifFerence 7.4% 76 Standard deviation * 0.0148 0.0155 t value * 5.13 4.77 Significance level <1 1% 1% M. Ravallion Poverty comparisons: A Guide to Cncepts and thods L oMS . 88 p. 49 (9) Annex 5, Page 5 Table A.13 Nicaragua LSMS 93-98 Confidence Intervals Bajo * Medio Alto General 1993 -48.8°/% --3 5T.8% General 1998 50/ 46.3% 47.8% 49.3% Extrema 93 c T82° T9.47o 20.6% Extrema 98 16.1°/ 17.3% 18.5% General 1993 48.3% 50.3% 52.3% General 1998 1O 45.8% 47.8% 49.8% Extrema 93 - 7.8°7 19Tg4° 21.0% Extrema 98 15.8% 17.3% 18.8% * M. Ravallion Poverty comparisons: A Guide to Concepts and Metfiods LSS o. 88 p.49 (10) Annex 5. Page 6 abeA14 AzWq7Be 18LSIVS Q*eM cdYSMcft dfE]P " yp 0 Yqxrdcncj o(a) LUb-a prrr2 LAo pxr n-e Ktpcr rnEs Rrxrdn da Gaies f( rixjxlo (b) pcIibra pc I Gt pacdmr- (c) Gippdiaiad (c) Gipdiax3rd (C) Gippdaiaicr (c) &~400pia3orE (c) pc (a) utial 1u4YL/c U-Y/ (J1UY 1Y/. aU~~Y/O U-1UY 1LJ4Jyo EMTY/o MMLLY 14CF/o LU-jY/c (J-lLxYl. 1LI4JYc. IRiY/o O-lUY CXdta irt1ilI&Ae $9.14 1 1.4%/ 1 .3/ l.6 28 0.27 03 $ 255 $ 246$ 30?7 468 <2 u O7/o Q6% 091 183 otila $ 313 11.50/c 10. 91/o 870/c 689 6a51 5T $ 21.57 $ 223) $ 160 6,442 6Cff 4,86 9.01/o 64% 7f8f!' 29 acaarde $ 504 0.4%/ 0.4%/ 0.91/ 016 0.16 0.3 $ 0OW $ 083 $ 1.72 88 91 193 0.1% 0.1% O.3Y/ 110 azmEngau $ 1.43 1 7.01/c 7.4%/ 4.70/ 918 9.69 6.1( $ 1317 $ 139) $ 8.75 15,051 15,8E 9,99 21.01/o 21.91/o 16.01/( 1,14 am srr1e $ 283 12 38%/ a38% 4.9D/c 253 254 3 $ 7.15 $ 7.19 $ 85) 3j41 3160 3,733 4.4%/ 44%/ 69OP/ 440 J182S ~~$ 1627 21- 0.2?/o 0.2Yo Q50/( 0.02 0.03 0.O $ 0.35 S 041 $ 094 47 55 125 0.1% 0.1% 0.2'/ 133 rrrp, en gmS a26 1 16ff,/c 16f!0 1370/ 9.17 9.16 7.a $ 298 $ 293J $ 2577 14,981 14,937 1290 20.9/ MT70/ 2ie &/ sn 2stfi aln-atjes$ 668 168 014% 0.4%/ 0.60/c 0.12 0.11 01V $ 0.78 $ 076 $ 117 193 18 m 9 03%/ 0137/o 0.50/c 25 & ~~~~$17.79 10 4.91/ 40f/c 3 40c 0.42 0.43 0.3 $ 7.45 $7.57 $6.42 43D 437 370 06%/ Q9D/c 0.&/C 83I ar~~~~~ c~~~~~ r31 10E 24% 24%/ 5 2D/ (13 03 . $ 4.43 $ 4.55 S 9.82 374 O. 3. 1 77 8 aCLcatb $12739831 091'/0 09'/a 0.91/c 00C7 007 O1$ 0.86 $0.95$1.68 67 73 129 0.1% 0.1% 0.20/ 77 LEocb resCtk)c $ 46D443 09/c 08%/ 0. 8/c 0 0.25 03 $ 1.03 $ 1.18 $ 157 97 114 1M 0.1% 0.20/o 0.20/ 97 aedgall~itra/pojlo S 11.47 9E 280/6 31% 5.103' 046 0.51 0. $ 5.2B $ 589 $ 9.59 449 5) 816 0 91/o 0.70/o 1.3Y1 85 hetia d ecaJDk $ 6.81 9 06% 0.40/o 0981' 012 003 0.1 $ 1.04 $ 083 $ 142 58 47 83 0.1% 0.1% 0.1 0/c 55 tLfl sadrm $ 7025, 2/ 0.2%1 0.37 . $ .2 6 2 0 7 11 Of/ -ND-T 21 rrtadics $ 1382 112 0 0% 0ff!0 040/c 003 0.01 0.O $ O0C2 $ 003B $ 083 2 7 68 0ff01/ 0.1/o 0.1 0/c 81 edepg Wa $ 1.6522 4ff/c 4,20/c 40f/( 4.48 4.71 4. $ 7.41 $ 7.78 $ 7. 1,017 1,06 1,03 1.4%1 1. 9/a 1.91/ 13) e?4eEn pclvo $ 2945 I5 1.70/c 1.73'o 200/4 011 0.11 01 $311l $ 318$ 377 1657 171 203 0.20/c 020/c 0 30/ 54 LeSiY.JIIO $ 1361 119 21% 220/o 361/c G29 03) 0. $ 39.3 $ 4.06 $ 6.83 346 357 58 05%F/ 091/o 0 90/ 85 aI.Ia'ejlUervir lQwm $ 11.94 2 0.70/c 091/cl 1.494 012 0.13 0I $ 1.38 $ 1.83 $ 261 310 35) 58 Q 40/ 091/c 09O/ 225 tasc ira . 4.7% 4.5% 4.317 TM 39 --- $T 88 aa s 4 $7.99 82 779 737 = -777T77 - cnite vge $ 583 4012 6.37/c 6.37/o 5.91/( 203 203 1.8 $ 11.74 $ 11.77 $ 1083 aim 8154 7,2M 11.37/c 11.37/o 11.703' 69 a-ec IBCErcCHb $541 391 011% 03%D/ 0.13'X 003 0.03 0. $OZ$ 029 $0.22 191 213 19) 0.30/c 0.30/o 0 30/ 738 el -~~~ $282 3 1.4%/ 1.4%/ 1.37/( 095 0.91 0.& $268 $ 255 $235 194 185 171 0. 30/c G 03/o 30/c 72 6 (ila $a9 3 A8 02D/o 0.37/o G40/c 010 0.12 02 $ 0.41 $ 0.48 $0,83 21 25 41 09D/c, 0901/a 0 1/ OX 5 T-W $la37 0.37/c 0.37/0ac T 0.3 .0 HE $ 056 $ 53$0159 3 31 35 OOVc ( 091/o 0M13'59 hltmna $585 141 0910/c 0. 1k 0.70c 017 0.17 0 $ 0.9B$098 $1.37 24 22 33 09D/C 0901/c 013'/ 24 uTEe $3o3 9 .91/% 1.910/c 1.80/c 1.03 0.83 1.11 S am s278$3a41 95 85 103 0.134/ 0.1% 0.20/ 31 apes sa$363 3a 0/oa8%09/o 1. 13'/ 03)040 0. 5 $ 1.51 $1.57 $913 13) 143 195 0234,0.2%/ 037/ 91 rijd engu $ 574 1M 1223% 11.70ko 9.80/c 383 383 a31 $ ma $ 21.83 $i1&2 6,227 5972 4,985 &70/c VI0/ aff! 272 um ~ ~$1.47 54 (091k 0.91/c 070/c 1.0 1.21 0. $ 1.57 $1.77 $1.34 59 653 503 0.8%/ 0.91/c 080/c 374, 71 $2- lN ~~~~7.91/o 7.91/o 6rae/ 4.6 4.2 $Z 138 la $ 14.16 $ 1231 a471 5R574 7,473 Th W7 .7/-.M 60 ai ~~~$0835 C 91/o 1.09/o 809/ 1.85 1.97 1.51 $1.77 $ 1.87 $1.44 GO- - 91 O91/ 091O/c - inEe $472 C 01% 0.1% 0.20/ 004 0.04 O0 $ 019$0Q17$0(33 - - 0.91k 091D/c  91/( - u:,lEraakos $9.8 24 Q91/a 0901/c 0.20/ 001 0.01 0.0 $00.7 $ 05 $037 2 1 9 091kD 091/c 09U1/ 25 icues ~~$15971I 0.20/o 020D/c 0.37/ D02 O0C2 0. $ OM$ 32$0.54 21 21 3) (D, 01 (09/c 0. 13 O/ 3 ERIEa $ .11 ffi 01%1% % 037/-= U0 .m 0 $ 0.22 $ 017$0.62 6 5 16 U,-yr , 2 eias,sabtes $ 4.83 0.37/. 91/ 0.40/, 011 OC3 0.1 $ 0.53 $0a43 $ 070 94 77 124 0.1%/ 01% 0.20/ 177 U I /c =N&ns hub/ IUWI 1UYI, R7Y T.3 49.4 $ Th/ $P Thf $ 18/ fl,db /Z4&L db6~tb( lUJYa lWY IUYI, UIA7 12=XPM~ - tu6 bwQ $ 4246 $ 424b $ eZ2M -TfM /4 7bLJ,aIj OTALI/ 32 = Oaio 1.183 1.84 1. $ 6.24 $ 624 $ 6a241 Z39 2413 Z08________ Annex 5, Page 7 Table A.15 Nicaragua LSMS 93-98 Comparing the dollar value of Poverty Lines Povery Lines in Cordobas Exchange rates Povery Lines in Dollars Change 1993 1998 1993 (a) 1998 (b) 1993 1998 Dollars O General C$2,574 C$4,258 6.C 10.43 $ 429 1 $ 408 -$20.81 -4.8% Extreme C$1,216 C$2,246 6.0 10.43 23 $ 215 + $12.7 6.3% (a) For 1993 $1.00 = C$ 6.00 (used in the 1993 report) (b) The 1998 exchange rate correspond to May 1998 (Banco Central de Nicaragua). NOTE: Since the 1993 study was carried out from February throgh June, the correspondig exchange rates should be used, but no information was provide in the Banco Central de Nicaragua Web Page ( http://www.bcn.gob.ni/) for the 93 monltly exchange rates. Fron December 91 to December 92 the official exchange rate was $1.00 = C$5.00 Table A.16 Nicaragua LSMS 93-98 Consumer Price Index: 1994=100% L otal Food Housing ransport. I Educaion | Health Furniture 1323 92.1 91.7 81.3 -3.4 99.5 98.9 Mayo 98 151.9 156.2 190.8 150.0 125.6 134.3 151.2 Change 165% 170% 209% 180%j 126% 136% T680/ % consump tion 798 88 00 4. 7% 4.5% 5.2% 3.7% 83.1% 41.7% 8.5% 5.7% 7.0% 6.2% % consumption 93 51.1% 20.9% 2.6% 4.6% 3.2% 4.3% 87.0% 43.7% 4.7% 5.8% 4.3% 7.2% Personal-x i,Esparcim.-x lClothFx Average-x 1993 - 97.4 86.4 97.6331 93. Total May 98 122.1 144.5 107.9417 124.8 y Change 125% 167% 111 % 1 Suma % consump3tion 98 132% TZ£Yo 17.7% 169.9% % consumption 93 13.3% 100% 17.9% 170.6% Tabla A. 17 Nicaragua LSMS 93-98 Comparing the Cordoba value of Poverty Lines Povery Lines in Corcobas Change Price Index Povery Lines in 1993 value Change 1993 1998 All I Food 1993 I 1998(a) ICordobasl % General CTM4 C$4,258 165%| CS2,574+ 0 ..reme +- (a) The General poverty line for 1998 was deflated ussing the percentage change in the Average CPI. (a) The Extreme poverty line for 1998 was deflated ussing the percentage change in the Average FOOD CP Annex 5, Page 8 Table A. 18 P Measures and chage from 1998 to Post Mitch Region Post Mitch , Pre Mitch 1998 Change % points % change P-0 P-1 P-2 P-0 P-1 P-2 P-0 P-1 P-2 P-0 P-1 P-2 All Nicaragua 48% 18% 9% 47.9% 18% 9% 0.10% ns 0% 0% 0% 0% 0%7. Urban 30% 10% 4% 30.5% 10% 4% -0.2% ns 0% 0% -1% -1% -1 '4 Rural 69% 29% 15% 68.5% 28% 15% 0.4% ns 0% 0% 1% 1% 0'/0 Managua 18% 5% 2% 18.5% 5% 2% 0.0% X 0% 0% 0% 0% 0%/o Pacific Urban 39% 12% 6% 39.6% 13% 6% -0.6% ns 0% 0% -2% -1% -1 '40 Pacific Rural 63% 23% 11% 67.1% 26% 13% -4.0% * * -3% -2% -6% -10% -14/o Central Urban 39% 14% 7% 39.4% 14% 7% 0.0% ns 0% 0% 0% -1% -1 /O Central Rural 78% 33% 18% 74.0% 31% 17% 3.6% * * 3% 1% 5% 8% 9 /o Atlantic Urban 44% 17% 9% 44.4% 18% 9% 0.0% X 0% 0% 0% 0% 0 ' Atlantic Rural 81% 37% 21% 79.3% 37% 22% 1.3% ns 0% 0% 2% 0% -1 z ^ 1998 values with the 1999 Post Mitch households data substituted Significant at p < 1 % ns = Non Significant at p<=10% X Altantic Urban and Managua did not have any Post-Mitch households P-Os Head count Index, P-1: poverty Gap, P-2: Foster-Greer-Thorbecke Table A.19 P Measures and chage from 1998 to Post Mitch for Extreme Poverty Post Mitch ' Pre Mitch 1998 Change % points % change | P-0 P-1 P-2 _P- P-1 P P- PT P2 All Nicaragua 17.3% 3M 27o T3% 1 2% -0 . % ns D0 0% 7 -4 ':/ Urban 7.5% 2% 1% 7.6% 2% 1% -0.1% ns 0% 0% -1% 0% 10'/ Rural 28.9% 8% 3% 28.9% 8% 3% 0.0% ns 0% 0% 0% -4% -6"/o Managua FTW ---u _7TW. _77o --- Wo X~ 0% W07 --Wo 0%FT - Pacific Urban 9.6% 2% 1 % 9.8% 2% 1% -0.3% ns 00/ao 0% -3% 2 % 4% Pacific Rural 20.6% 5% 2% 24.1% 6% 2% -3.6% * * -1% -1% -15% -23% -291Y/o Central Urban 12.1% 3% 1% 12.2% 3% 1% -0.1% ns 0% 0% -1% -2% -3%/o Central Rural 35.7% 10% 4% 32.7% 10% 4% 2.9% * * 1% 0% 9% 5% 5% Atlantic Urban 17.0% 4% 1% 17.0% 4% 1% 0.0% X 0% 0% 0% 0% 0'/o Atlantic Rural 40.6% 13%1 6%1 41.4% 14% 7% -0.8% ns -1%j 0% -2% -5% -6/o * 1998 values with the 1999 Post Mitch households data substituted ** Significant at p < 1% ns = Non Significant at p-1 0% X Altantic Urban and Managua did not have any Post-Mitch households P-0: Head count Index, P-1: poverty Gap, P-2: Foster-Greer-Thorbecke Annex 5, Page 9 Table A.20 Nicaragua LSMS 1998-1999 Post Mitch Segments 1999 1998 Cordobas % change Consumption group Post-mitch Pre-mitch diference Food consumed 2430.51 2423.43 7 0.3 House use value 289.18 330.17 -41 -14.2 House services 130.11 196.93 -67 -51.4 Education 181.22 111.54 70 38.5 Health 353.45 217.06 136 38.6 Transport 157.77 163.7 -6 -3.8 Durable goods 54.55 57.61 -3 -5.6 Other incl. personal 322.56 428.53 -106 -32.9 Total 3919.341 3928.96 -10 -0.3 # of household expanded 762008.81 762008.81 The values are the national averages Table A.21 LSMS Nicaragua 1998-99 Post Mitch segments 1Y99 19 9 % points Percentage Consumption group pecentages pecentages change change Food consumed 62.0 - 61.6 0. House use value 7.38 8.4 -1.0 -12.2 House services 3.32 5.01 -1.7 -33.8 Education 4.62 2.84 1.8 62.9 Helth 9.02 5.52 3.5 63.2 Transport 4.0 4.7 -.1 -3.4 Durable goods 1.39 1.47 -0.1 -5.1 ,Other incl. personal | 8.23 10.91 -2.7 -24.5 Table A.22 LSMS Nicaragua 1998-99 Post Mitch segments 1999 1998 % points Percentage Consumption group pecentages pecentages change change Food consumed 65.36 .7 2.2%/ House use value 8.21 8.8 -0.6 -6.7% House services 2.86 5.17 -2.3 -44.7% Education 4.77 2.9 1.9 64.5% Health 6.47 4.99 1.5 29.7% Transport 34.5% Durable goods 0.94 1.03 -0.1 -8.7% Other incl. personal 8.18 9.76 -1.6 -16.2% Annex 5, Page 1O PART B - CONSUMPTION Table B.1- Nicaragua 98 LSMS Consumption Patterns (Comprehensive) REGION QUINTIL Managua Pacific Pacific Urbano Central Atlantic Atlantic 1 2 3 or or Central Rural Urban Rural Urban Rural GROUP Food 50.69 53.96 62.26 53.19 67.00 59.27 66.60 64.70 63.59 59.'0 Housing 13.46 9.96 8.85 10.12 8.79 10.52 10.23 10.24 8.64 9.38 Water, Electricity, etc 7.16 8.96 5.03 7.62 3.94 6.99 4.35 6.60 5.95 6.36 Education 5.37 4.62 3.26 4.69 2.51 3.75 1.64 3.14 3.46 3.71 Health 3.22 4.89 5.89 6.16 4.46 4.99 5.49 4.16 4.63 4.35 Personal and others 11.86 11.63 9.88 11.77 9.67 10.27 9.97 8.73 9.70 10.28 House Equipment 3.11 2.74 1.10 2.69 0.75 2.26 0.41 0.60 1.00 1.132 Transportation 5.09 3.22 3.72 3.75 2.87 1.95 1.29 1.83 3.03 3.f1 Tranferences 0.04 0.02 0.01 0.01 0.00 0.00 0.03 0.00 0.01 0.00 QUINTIL Area Poverty (2) Poverty (2) by area 4 5 Urban Rural All Non- Rural Urban poor poor All Non- All No-6 poor poor poor poor GROUP Food 56.68 45.22 52.89 64.01 63.57 52.83 66.05 59.58 58.89 50.26 Housing 10.08 13.72 11.40 9.46 9.42 11.52 9.42 9.52 9.40 12.28 Water, Electricity, etc 6.32 6.53 7.88 4.53 6.24 6.46 4.75 4.05 9.05 7.37 Education 4.43 5.13 4.84 2.93 3.39 4.50 2.82 3.17 4.47 5.D1 Health 4.47 5.55 4.37 5.08 4.46 4.91 4.56 6.20 4.27 4.41 Personal and others 11.77 13.84 11.73 9.84 9.42 12.19 9.28 11.07 9.70 12.62 House Equipment 2.09 4.56 2.87 0.99 0.90 3.04 0.69 1.66 1.28 3 57 Transportation 4.16 5.38 3.99 3.15 2.60 4.52 2.42 4.74 2.95 4 44 Tranferences 0.00. 0.07 0.02 0.01 0.00 0.03 0.01 0.01 0.00 0 03 Poverty (3) Poverty (3) by area Tot Extreme No Extre Non Rural Urban Poverty Poverty Poor Extreme Non- Non- Extreme Non- Non- ext. p. poor ext. p. poor GROUP Food 64.59 62.99 52.83 66.48 65.73 59.58 58.58 59.00 50.26 57.97 Housing 10.53 8.78 11.52 10.44 8.68 9.52 10.83 8.92 12.28 10.51 Water, Electricity, etc 6.69 5.98 6.46 5.47 4.23 4.05 10.56 8.54 7.37 6.35 Education 3.17 3.52 4.50 2.71 2.91 3.17 4.64 4.41 5.01 3.97 Health 4.13 4.64 4.91 4.16 4.86 6.20 4.05 4.34 4.41 4 69 Personal and others 8.58 9.90 12.19 8.60 9.77 11.07 8.50 10.10 12.62 10 87 House Equipment 0.57 1.08 3.04 0.52 0.82 1.66 0.75 1.46 3.57 202 Transportation 1.74 3.10 4.52 1.62 3.00 4.74 2.09 3.24 4.44 3 60 Tranferences 0.00 0.01 0.03 0.00 0.01 0.01 0.00 0.00 0.03 0 02 Values are the average of each household percentages Others includes: Cerominies, clubs fees, lottery, domestic service, shoes, linens, cookware, soap, detergent, transport, comunications, Annex 5. Page 11 Table B.2 Nicaragua 93 LSMS Annual Average Value ot consumption Per Capita (Comprehensive) Total FIood Housing Water f R Hat Equip- Ierans- I rans- erso- elect cation ment portation feren- nal and etc ces others Cordobas Cordobas Crdobaa . Cordobai CTrdobas Co rdoba- r as Cordobas Cordobas Cordobas Area Urbano 8,441 3,753 1,224 614 429 415 377 426 5 1,198 Rural 4.069 2,421 450 177 132 234 69 161 0 424 Group Total 6,445 3,145 871 415 293 332 236 305 3 845 REGION Managua 10,440 4,273 1,830 695 597 440 483 593 7 1,523 Urban Pacific 6,699 3,294 741 566 310 363 268 276 5 875 Rural Pacific 4.039 2,419 361 193 131 266 62 160 0 445 Urban Central 6,956 3,311 790 505 330 451 276 340 1 951 Rural Central 3,617 2,294 345 125 82 187 48 159 0 376 Urban Atlantic 6,407 3,338 822 465 213 381 314 155 0 719 Rural Atlantic 3,169 2,081 329 114 49 182 16 60 1 338 Group Total 6,445 3,145 871 415 293 332 236 305 3 845 Poverty (3 groups) Extreme Poor 1,623 1055 166 106 52 66 9 29 0 141 Not Extreme Poor 3.207 2,009 285 192 115 150 36 100 0 320 No poor 9,943 4,505 1,448 648 478 527 429 517 5 1,387 Group Total 6,445 3,145 871 415 293 332 236 305 3 845 PoDr and No Poor All Poors 2,633 1,663 242 161 92 119 26 74 0 255 No poors 9.943 4,505 1,448 648 478 527 429 517 5 1.387 Group Total 6,445 3.145 871 415 293 332 236 305 3 845 Cordobas (May 1998) Table B.3 Nicaragua 98 LSMS Annual Average Value of Consumption Per Capita (Comprehensive) by Poverty Groups and Area Total Food Housing Water Edu- Health Equip- iran TI rans- Perso- elect. caton ment portation feren- nal and etc. ces others Cordobas Cordobas Cordobas Coroobas 2ordobas CoraS Cordobas Cordobas Cordobas Cordobas Urban POVERTY (3 GROUPS) Extreme Poor 1,688 986 184 181 77 69 13 36 0 143 No Extreme Poor 3,282 1,929 295 279 147 142 50 107 0 333 No Poor 10,878 4,657 1,644 772 560 542 524 574 7 1,599 Group Total 8,441 3,753 1,224 614 429 415 377 426 5 1,198 Urban POOR AND NON POOR All Poors 2,883 1,693 267 254 129 124 40 90 0 285 Non Poor 10,878 4,657 1,644 772 560 542 524 574 7 1,599 Group Total 8,441 3,753 1,224 614 429 415 377 426 5 1,198 RURAL Poverty(3 groups) Extreme Poor 1,602 1,076 160 82 44 65 8 27 0 140 No Extreme Poor 3,155 2,064 278 132 92 155 26 96 0 311 No Poor 7,483 4,104 932 322 262 488 178 367 1 828 Group Total 4,069 2.421 450 177 132 234 69 161 0 424 RURALi POOR AND NON POOR All Poors 2,500 1,648 228 111 72 117 19 67 0 239 Non Poor 7,483 4,104 932 322 262 488 178 367 1 828 Group Total 4,069 2,421 450 177 132 234 69 161 0 424 Cordobas (May 1998) Annex 5, Page 12 Table B.4 Nicaragua 93-98 Self-consumption from Agriculture only Total National consumption C$ (thousands) Share of Self-Consumption 1993 1998 1993' I99F- Poor (all) Non Poor 13,982 24,965 2.4% 3.60% Poor 3,071 6,065 15-0% 11.o7°,,, txtreme Poor Not Extreme Poor 16,366 29,675 5.9% 6.1° Extreme Poor 687 14 19.8% 14.1%5sa Area Urban 13,113 22,087 1.3% 1.6% Rural 3,940 8,943 18.6% 14.50/, QUINTIL.. _ 1 720 1,652 19.4% 13.8%,/, 2 1,358 2,875 13.7% 11.0% 3 2,156 4,303 6.9% 7.0%/'v 4 3,396 6,278 2.7% 4.2%/ D 5 9,421 15,921 1.0% 1.6° REGIlONI Managua 6,916 13,097 1.2% 1.5%0) Urban Pacific 4,322 5,402 1.0% 1.2% Rural Pacific 1,474 3,027 10.9% 10.0% Urban Central 1,804 3,519 2.5% 2.70/% Pacific Central 1,321 3,625 25.0% 17.4°,s Urban Atlantic 835 1,558 4.3% 4.1°) Rural Atlantic 378 800 28.2% 23.70,% Reported Auto- consumption No Auto - consumption 12,252 19,577 0.0% 0.00,%z Auto-consumption 4.801 11,453 18.5% 16.0%h Area X Poor Urban Not Poor 11,794 19,788 0.6% 1.1% Urban Poor 1,318 2,299 3.3% 2.8% Rural Not Poor 2,187 5,177 9.9% 10.3c,i Rural Poor - 1,753 3,765 21.8% 16.4'ij Area X Extreme Poor Urbano No Extreme Poor 12,944 21,750 0.9% 1.3%'i Urbano Extreme Poor 168 337 8.2% 5.6%0o Rural No Extreme Poor 3,421 7,925 16.4% 13.6°'i Rural Extreme Poor 518 1,017 23.1% NATIONAL 17,053 31,031 8.4%1o 7I.b' 1993 Agricultural and non-agricultural self-consumption 1998 observed relationship between agricultural to non-agricultural self-consumption 2 Observed 1998 values for agricultural only self-consumption Annex 5, Page 13 PART C - INCOME Table C. Nicaragua 1998 LSMS Income Pa erns Area REGION Poor (3 croups) Poor or no Total Urban Rural Mana Pacific Central Itantic- Ext. Non Ex No A1l No gua _Urb5a_n R`u_ral a u_ U Rural Ub . oor PPoorPoor Poor Poor Work I T Wage Agricultural 50% 20.2% 26% 6.0% 17.9% 8.6% 25.5% 101% 141% 250% 15.7% 5.4% 19.1% 5.4% 11.9% Wage Non-agricultural 42 3% 18 4% 46 5% 41 5% 23.3% 36.0% 13.5% 35 9% 5.9% 16.9% 29.2% 37.5% 24 8% 37.5% 31.4% Self-emp. Non-agnc 22 7% 8.9% 20.0% 24.0% 12.2% 24.4% 6.6% 19.6% 6 4% 6.6% 12.1% 22.2% 10.1% 22.2% 16 4% Self-emp. Agricultural 3 4% 26.5% 2.9% 2.6% 18.2% 6.3% 30.7% 8.1% 46 3% 22.8% 18.8% 8.1% 20.3% 8.1% 13.9% Undetermined 1 2% 1.2% 1.3% 1.6% 2.2% 1.2% 0.5% 0 3% 0.8% 2.3% 1.3% 0.9% 1.6% 0 9% 1.2% Imputed rent 14 ou/o 13. 7 7 5 13 4/. 12.S/:. 1 14.o I J.U 17/. 12A% T33G 2. 15.2%o 14 1ul Education transfers 04% 08% 0.3% 0.2% 0.5% 0.5% 1.1% 0.5% 0.7% 1 1% 0.8% 0.2% 0.9% 0.2% 0.5% Food Gifts received 1 5% 3.2% 1.4% 1 4% 3.2% 1.9% 3,3O/ 21% 4.0% 4 1% 2.9% 1.3% 3.4% 1.3% 2.3% Remitances received 4.0% 4 4 % o 8 T,'o 6. - ,o 227/l 6.3 / J 4 8U/° 36 V/o b6,4% 3.9 * o . Charities received 0. 0% 0 0% 0 0% 0 0% 0 0% 0 0% 0.0% 0.0% 0.0% 0.0% 0 0% 0 0% 0.0% 0 0% 0.0% Returns from capital 0 =o -7.o / u. b¶To u.3 io 0.7o 2o O 0A4% 04%0 Pensions 1.4% 04% 1 8% 1.1% 05% 0.6% 0.4%/6 1.7% 0.0% 04% 0.6% 14% 0.5% 1 4% 1.0% Others 0.8% 3 0% 0 7% 0 7% 2.5% 1.1% 3.8% 1.2% 2.3% 3.8% 1.9% 1.0% 2.6% 1.0% 1.8% Total yearty per capitia income 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Values are the average of each household percentages. Table C.2 Nicaragua 1998 LSMS income Patterns by Poverty Group and Area URBAN RURAL Poor (3 groups) Non Poor Tota Poor (3 Groups Non Poor Total Ext. No.1Ex No All No Ext. No. Ex No All No Poor Poor Poor Poor Poor Poor Poor Poor Poor Poor Work Wage Agricultural 18.0% 8.0% 2.6% 10.5% 2.6% 5.0% 27.2% 21.0% 12.7% 23.6% 12.7% 20.2% Wage Non-agricultural 35.7% 43.4% 42.7% 41.4% 42.7% 42.3% 11.0% 19.6% 23.8% 15.9% 23.8% 18.4% Self-emp. Non-agric. 13.4% 20.2% 24.6% 18.5% 24.6% 22.7% 4.5% 6.5% 16.0% 5.6% 16.0% 8.9% Self-emp. Agricultural 9.3% 4.4% 2.4% 5.6% 2.4% 3.4% 27.1% 28.7% 23.2% 28.0% 23.2% 26.5% Undetermined 2.1% 1.9% 0.9% 1.9% 0.9% 1.2% 2.3% 0.8% 0.7% 1.5% 0.7% 1.2% Imputed rent I T12b 5.7o I2% 1 4o Z/o .6 7b/o 14U 12./o 14.8 1 o 13.2T3 Education transfers 0.9% 0.8% 0.1% 0.9% 0.1% 0.4% 1.1% 0.8% 0.4% 1.0% 0.4% 0.8% Food Gifts received 3.1% 2.7% 1.0% 2.8% 1.0% 1.5% 4.4% 3.1% 2.3% 3.6% 2.3% 3.2% Remitances received . Z7o 4. 7 1 7A z 7 6 7 3 6% T. T Charities received 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% eturns rom capita u 0U/ OJ U. 0 / /0 0. /0 U. % U. US7 0 = 0 Pensions 1.0% 0.8% 1.7% 0.8% 1.7% 1.4% 0.1 % 0.6% 0.5% 0.4% 0.5% 0.4% Others 0.9% 0.7% 0.8% 0.7% 0.8% 0.8% 4.8% 2.8% 1.5% 36% 15% 3.0% Total yearly per capitia income 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Values are the average Ot each nousehc percentages - _ _ Annex 5. Page 14 Table C.3 Nicaragua 9S LSMS Average Income Area REGIONPoor (3 Groups) Poor or Not Tai- Urban Rural Mana Pacific Central Atantic Ex NoExl NoN Wrgua U?5an Rural Urban Rural Urban Rurba P. P. P Poor Poor Work__ . __ Wage Agricultural 223 706 131 228 769 413 833 379 298 548 560 341 556 341 4.i 4 Wage Non-agrcultural 3,516 949 4,430 2,722 1,139 2,786 624 2,317 326 499 1,144 3,660 910 3,660 2.344 Self-emp. Non-agnc. 2.096 542 2,283 1,830 612 2,013 339 1,680 431 155 465 2,335 352 2,335 1.316 Seif-emp. Agricultural 178 891 123 155 558 395 1,108 310 1,349 330 488 570 431 570 5: 3 Undetermined 86 33 88 112 54 69 13 22 11 62 64 61 63 61 S 2 Imputed rent 1,045 393 930 680 329 77 306 662 31 161 262 . 7 47 Education transfers 1 0 14 11 6 12 13 19 9 8 14 17 8 16 8 '2 Food Gifts received 62 84 65 59 91 79 78 42 87 63 69 77 67 77 72 Remitances received 3 T 93 72 246 425 98 1355 212 6 631 9 631 6 TT Charities received. 1 1 2 0 0 0 1 0 0 0 1 1 1 1 1 FReturns from capital =1 13 307 44 1T 79 - 7 3 1 4 194 3 1 1 Pensions 112 20 164 67 33 42 15 65 2 6 2 3 118 17 118 7 0 Others 56 70 56 43 65 48 70 152 59 57 45 74 49 74 f. 2 Total yearly I per capitia income 8,129 3,876 9,552 6,520 3,9211 7,049 3,510 7,149 3,119 1,965 3.272 9,298 2,798 9,298 6,1E8 Cordobas (May 1998) Table c.4 Nicaragua 98 LSMS Average Income by Groups and Area URBAN IRURAL Poor (3 groups) Poor or Not FToa Poor (3 Groups) Poor or Not I Toa Et A Ext. NNo All No Ext. ExFN No All N P P Poor Poor Poor P. P Poor Poor Poor Work Wage Agricultural 489 243 187 305 187 223 566 778 745 689 745 706 Wage Non-agricultural 974 1,585 4,430 1432 4,430 3,516 349 841 1,636 634 1,636 949 Self-emp. Non-agric. 369 772 2,721 671 2,721 2,096 87 254 1,320 184 1,320 542 Self-emp. Agricultural 144 133 196 136 1 96 178 388 731 1554 587 1554 891 Undetermined 91 118 74 1 11 74 86 53 26 24 38 24 33 Imputed rent 179 277 1,393 252 1,393 tO45 15 21 791 21T1 791 = 39 Educationtransfers 14 17 7 16 7 10 14 17 11 16 11 14 Food Gifts received 45 62 64 58 64 62 69 73 111 72 111 84 Remitances received 1 i5 T 7U4 1T2T 7F 56 75 1T22 282 102 -= T57 Charities received 0 0 1 0 1 1 1 1 2 1 2 1 Returns from capital O 1 256 1 5 7 2 6 33 4 33 Pensions 20 19 152 19 152 112 2 26 29 16 29 20 Others 10 18 73 16 73 56 72 63 78 66 78 70 Total yearly - - T T 61 per capitia income 2,385 3,390 10,318 3,139 10,318 8,129 t,833 3,190 6,615 2618 6,615 3,876 Cordobas (May 1998) Annex 5, Page 15 PART D - DEMOGRAPHIC CHARACTERISTICS D.01 Nicaragua - 1993 Demographic Characteristics Total 1998 number of Women Men adults Children Children Children (5 Demographic people in adults (above (above 16 (16 yrs and (5 yrs and yrs and Characteristics household 16 yrs) yrs) under) under) under) All 3T~5.5 1. . . 1. 09 - Extreme Poor 7.4 1.5 1.5 4.4 1.6 1.7 Poor 6.5 1.5 1.4 3.6 1.4 1.4 Non-poor 4.7 1.4 1.3 2.0 0.7 0.6 Urban 5.3 1.5 1.3 2.4 0.8 0.8 Extreme Poor 8.0 1.8 1.6 4.6 1.8 1.6 Poor 6.8 1.7 1.5 3.7 1.4 1.4 Non-poor 4.8 1.5 1.3 2.0 0.7 0.6 Rural 5.8 1.3 1.3 3.2 1.2 1.1 Extreme Poor 7.2 1.4 1.5 4.3 1.6 1.7 Poor 6.4 1.4 1.4 3.6 1.4 1.4 Non-poor 4.5 1.2 1.2 2.1 0.7 0.7 Poorest 7.3 1.5 1.5 4.3 1.6 1.7 11 6.3 1.5 1.3 3.5 1.3 1.2 III 5.8 1.6 1.4 2.8 1.5 1.0 IV 5.2 1.5 1.3 2.3 0.8 0.7 Richest 4.0 1.3 1.1 1.6 0.5 0.4 Managua 5.2 1.5 1.3 2.4 0.8 0.7 Pacific - Urban 5.4 1.6 1.4 2.4 0.8 0.8 Pacific- Rural 6.1 1.4 1.4 3.3 1.2 1.0 Central - Urban 5.5 1.6 1.3 2.6 0.9 0.7 Central - Rural 5.7 1.3 1.3 3.1 1.2 1.1 Atlantic- Urban 5.0 1.4 1.1 2.5 1.0 1.2 Atlantic - Rural 5.6 1.2 1.2 3.2 1.3 1.4 Annex 5. Page 16 D.02 - Nicaragua 1998 Demographic Characteristics by Poverty Group Extreme Demographic Characteristics All Poor Poor Non-poor Total number ot people in household 574 7.7 6T7 4.6 Women adults (16 and above) 1.5 1.6 1.5 1.4 Men adults (16 and above) 1.4 1.6 1.5 1.3 Children (under 16) 2.6 4.5 3.7 1.9 Mean age of household head 45.3 46.2 44.8 45.7 Mean age of head's spouse 37.7 38.8 37.3 38.1 Dependency ratio' 0.8 1.2 1.1 0.7 Dependency ratio2 0.9 1.2 1.1 0.7 D.03 - Nicaragua 1998 Demographic Characteristics, Urban and Rural Area Urban Rural Extreme Extreme Demographic Characteristics All Poor Poor Non-poor All Poor Poor Non-poc, Total number of people in household 5.2 7.7 6.8 4.7 5.7 7.7 6.6 4.4 Women adults (16 and above) 1.6 1.8 1.7 1.5 1.4 1.5 1.5 1.2 Men adults (16 and above) 1.3 1.5 1.4 1.3 1.5 1.6 1.5 1 4 Children (under 16) 2.3 4.5 3.7 1.9 2.9 4.5 3.6 1.8 Mean age of household head 45.7 47.3 45.7 45.6 44.9 45.9 44.4 45.7 Mean age of head's spouse 38.2 40.2 38.0 38.3 37.2 38.5 37.0 37.6 Dependency ratio' 0.8 1.1 1.1 0.7 0.9 1.2 1.1 0.7 Dependency ratio2 . 08 1.2 1.1 0.7 1.0 1.3 1.1 0.8 D.04 - Nicaragua 1998 Demographic Characteristics by Quintile Demographic Characteristics AllI Poorest C | IlI IV Richest Total number of people in household 5.4 7.6 6.2 5.5 5.0 4.0 Women adults (16 and above) 1.5 1.6 1.5 1.5 1.5 1.4 Men adults (16 and above) 1.4 1.6 1.5 1.4 1.3 1.2 Children (under 16) 2.6 4.4 3.3 2.7 2.1 1.4 Mean age of household head 45.3 46.0 44.0 44.8 45.7 46.0 Mean age of head's spouse 37.7 38.4 36.6 37.7 37.1 38.9 Dependency ratio' 0.8 1.2 1.0 0.9 0.7 0.5 Dependency ratio2 0.9 1.2 1.1 1.0 0.8 0.6 Annex 5. Page 17 D.05 - Nicaragua 1998 Demographic Characteristics by Zone Pacific - Pacific - Central - Central - Atlantic- Atlantic - Demographic Characteristics All Managua Urban Rural Urban Rural Urban Rural Total number of people in household 5.4 5.0 5.2 5.9 5.2 5.6 6.1 6.2 Women adults (16 and above) 1.5 1.5 1.6 1.5 1.6 1.3 1.6 1.3 Men adults (16 and above) 1.4 1.3 1.3 1.5 1.2 1.4 1.3 1.4 Children (under 16) 2.6 2.1 2.4 2.9 2.3 2.9 3.2 3.5 Mean age of household head 45.3 45.7 46.1 46.7 45.8 43.7 44.3 42.6 Mean age of head's spouse 37.7 38.3 38.6 39.0 38.6 36.0 36.9 35.1 Dependency ratio' 0.8 0.7 0.8 0.9 0.8 1.0 1.0 1.2 Dependency ratio2 0.9 0.7 0.8 1.0 0.8 1.0 1.0 1.2 Annex 5, Page 18 PART E - EDUCATION E.01 Nicaragua 1998 - Gross Enrollment Rates - Primary and Secondary by gender Pre-school' Primary 2 Secondary3 Total Male Female Total Male Female Total Male Female All 4.0 4.0 4.1 107.3 106.9 107.6 57.0 49.9 63.7 Rural 3.9 3.7 4.1 100.9 101.0 100.9 30.0 23.7 36.5 Urban 4.2 4.3 4.1 113.1 112.4 113.8 79.4 73.5 84.6 Poorest 4.3 5.7 3.2 92.4 90.0 95.2 13.2 10.1 16.9 11 4.1 3.8 4.4 108.9 109.3 108.6 31.6 25.4 37.6 III 4.0 4.6 * 110.0 111.9 108.3 62.0 58.9 64.9 IV 5.4 * 120.0 119.8 120.1 80.2 76.1 83.6 Richest * 109.8 109.7 109.9 103.3 95.2 109.9 Managua * 113.8 112.7 114.9 90.3 86.5 93.5 Pacific- Urban 4.1 5.9 114.7 117.7 111.7 69.0 61.8 76.4 Pacific - Rural * * 117.0 115.9 118.0 33.2 27.4 39.1 Central - Urban * 108.6 102.5 114.9 79.2 65.4 90.2 Central - Rural 5.4 4.1 6.5 92.2 95.1 89.2 23.8 18.5 29.4 Atlantic- Urban 6.1 * 115.7 115.7 115.8 56.3 49.2 63.7 Atlantic- Rural 6.3 * 81.5 77.4 86.2 10.7 13.1 8.5 Total Extreme Poor 5.1 6.8 3.7 89.1 87.3 91.1 11.3 8.6 14.8 Poor 4.3 4.9 3.7 102.3 101.7 103.0 26.4 21.9 31.3 Non-poor 3.7 2.8 4.5 113.1 113.3 112.9 86.6 81.5 90.9 Urban Extreme Poor 6.5 * * 95.6 94.3 96.9 21.5 17.8 26.1 Poor 5.9 6.7 5.1 110.1 106.7 113.7 44.3 45.1 43.5 Non-poor 3.0 2.7 114.8 115.8 113.9 94.7 87.7 100.5 Rural Extreme Poor 4.7 5.7 3.9 86.9 85.2 89.1 8.0 5.7 11.0 Poor 3.5 3.9 3.1 98.3 99.0 97.6 17.2 10.5 24.8 Non-poor I 108.5 106.8 110.2 62.6 63.1 62.2 1 number in pre-school/ number of 4-6 yrs old 2number in elementary school/ number of 7-12 yrs old 3number in secondary school/ number of 13-17 yrs old . n<10 Annex 5, Page 19 E.02 - Gross Enrollment Rates 1993 and 1998 - Primary and Secondary by gender 1993 1998 Primary2 Secondary3 Primary2 Secondary3 Total Male Female Total Male Female Total Male Female Total Male Female All 118.1 115.4 121.0 21.6 22.0 21.3 107.3 106.9 107.6 57.0 49.9 63.7 Rural 93.4 90.2 96.7 9.7 9.9 9.5 100.9 101.0 100.9 30.0 23.7 36.5 Urban 138.6 136.1 141.3 30.3 31.3 29.3 113.1 112.4 113.8 79.4 73.5 84.6 Poorest 76.0 73.3 78.7 4.2 2.7 6.0 92.4 90.0 95.2 13.2 10.1 16.9 11 107.9 103.5 112.8 15.0 14.9 15.2 108.9 109.3 108.6 31.6 25.4 37.6 III 130.1 125.2 135.4 26.1 26.4 25.7 110.0 111.9 108.3 62.0 58.9 64.9 IV 144.9 145.2 144.6 29.0 30.1 27.9 120.0 119.8 120.1 80.2 76.1 83.6 Richest 151.6 146.6 157.8 36.8 42.8 31.6 109.8 109.7 109.9 103.3 95.2 109.9 Managua 132.9 131.1 134.9 34.0 38.3 29.8 113.8 112.7 114.9 90.3 86.5 93.5 Pacific- Urban 137.4 139.8 135.0 33.1 33.3 32.9 114.7 117.7 111.7 69.0 61.8 76.4 Pacific - Rural 98.5 100.0 97.0 16.3 15.5 17.3 117.0 115.9 118.0 33.2 27.4 39.1 Central - Urban 150.8 143.1 159.1 17.1 13.9 19.8 108.6 102.5 114.9 79.2 65.4 90.2 Central - Rural 81.2 76.7 86.1 2.6 2.3 2.9 92.2 95.1 89.2 23.8 18.5 29.4 Atlantic- Urban 145.9 137.5 155.0 26.2 24.7 27.5 115.7 115.7 115.8 56.3 49.2 63.7 Atlantic - Rural 85.8 74.7 101.6 81.5 77.4 86.2 10.7 13.1 8.5 Total Extreme Poor 75.9 72.7 79.1 4.0 2.7 5.5 89.1 87.3 91.1 11.3 8.6 14.8 Poor 97.3 93.9 100.9 12.4 11.6 13.2 102.3 101.7 103.0 26.4 21.9 31.3 Non-poor 144.9 142.6 147.5 31.8 34.3 29.5 113.1 113.3 112.9 86.6 81.5 90.9 Urban Extreme Poor 90.5 89.9 91.1 10.3 95.6 94.3 96.9 21.5 17.8 26.1 Poor 122.6 119.8 125.8 21.5 19.8 23.4 110.1 106.7 113.7 44.3 45.1 43.5 Non-poor 147.7 145.6 149.9 34.9 38.0 32.1 114.8 115.8 113.9 94.7 87.7 100.5 Rural Extreme Poor 71.9 68.3 75.6 2.2 . 3.7 86.9 85.2 89.1 8.0 5.7 11.0 Poor 83.5 79.1 88.0 6.7 6.6 6.8 98.3 99.0 97.6 17.2 10.5 24.8 Non-poor 134.1 131.6 137.2 19.7 21.0 18.3 108.5 106.8 110.2 62.6 63.1 62.2 2number in elementary school/ number of 7-12 yrs old 3number in secondary school! number of 13-17 yrs old * n<10 Annex 5, Page 20 E03 - Net Enrollrret Rates 1993 and 1998 - Prirnary and Secondary by genar 1993 1998 Prinary' Secondary3 Priniar Secndary Toal l iFae Toal be FIemale Toal We F-ernale I otal Nale eniale Ail 75.6 74.4 76.8 15.5 15.8 15.2 79.6 77.9 81.4 36.8 31.0 42.3 Rural 65.5 62.4 68.7 7.2 7.2 7.1 74.8 72.5 77.2 17.2 11.0 23.7 Urban 84.0 84.3 83.6 21.5 22.5 20.6 84.1 82.9 85.3 53.1 49.1 56.7 Poorest 55.4 52.2 58.5 3.7 2.5 5.3 67.9 65.0 71.2 9.1 6.7 12.0 11 72.8 69.7 76.2 11.2 10.8 11.5 80.2 80.3 80.0 23.7 18.8 28.3 III 82.8 81.3 84.5 19.2 20.3 18.1 83.8 82.2 85.2 37.0 32.3 41.4 IV 88.6 89.0 88.2 20.2 21.0 19.5 84.3 84.4 84.3 53.2 49.8 56.1 Richest 87.0 88.3 85.5 25.0 29.4 21.2 85.9 82.4 88.9 64.2 56.9 70.2 Managua 82.6 83.0 82.1 24.6 26.9 22.5 84.7 81.7 87.8 57.2 54.8 59.3 Pacfic- Urban 83.9 84.3 83.5 22.0 22.8 21.2 85.9 86.3 85.5 49.2 44.2 54.3 Pacific- Rural 71.6 70.0 73.2 12.3 12.5 12.0 84.8 83.3 86.3 23.6 18.4 28.8 Central - Urban 88.1 88.2 88.0 14.7 12.2 16.7 80.0 79.2 80.8 52.3 43.7 59.2 Central - Rural 57.8 54.5 61.5 1.7 1.3 2.0 69.9 67.5 72.3 11.4 4.0 19.3 Atlantio- Urban 84.7 83.3 86.2 16.6 20.0 13.7 86.5 85.8 87.1 35.3 29.6 41.2 Atiantic- Rural 58.1 53.9 64.0 * 58.0 57.3 58.9 6.5 8.6 4.6 Tota' ExtrerTe Poor 55.3 51.7 58.8 3.6 2.5 4.9 65.1 62.7 67.9 8.1 5.8 11.1 Poor 66.6 64.0 69.4 9.4 8.7 10.2 75.6 73.5 77.8 18.8 14.5 23.6 Nan-poor 87.2 87.7 86.6 23.2 244 20.2 84.4 83.2 85.5 54.3 49.6 58.1 Extrere Poor 64.7 63.9 65.5 9.9 * 68.1 65.9 70.5 15.1 13.0 17.7 Poor 77.8 77.9 77.8 16.0 14.7 17.5 80.4 79.0 81.9 30.8 29.6 32.1 Non-poor 87.5 88.1 86.8 24.4 27.1 22.0 86.2 85.2 87.1 62.8 58.7 66.1 Extreme Poor 52.7 48.7 569 1.8 2.9 64.1 61.7 66.9 5.9 3.5 8.9 PMor 60.5 56.0 650 5.4 5.1 5.7 73.1 70.5 75.7 128 7.2 19.1 Nzn-poor 86.0 86.2 85.8 13.3 14.5 12.0 79.7 78.1 81.4 28.8 22.4 34.0 2nuTber 7-12 yr odrs in elerentary schoo/ nurber of 7-12 yrs old 3nurrber 13-17 yrods in secodary school/ nLrrber of 13-17 yrs old n<10 Annex 5, Page 21 E.04 - 1998 Net Enrollment Rates - Pre-school, Primary and Secondary by gender Pre-school' Primary2 Secondary3 Total Male Female Total Male Female Total Male Female All 2.9 2.4 3.4 79.6 77.9 81.4 36.8 31.0 42.3 Rural 3.0 2.7 3.2 74.8 72.5 77.2 17.2 11.0 23.7 Urban 2.8 2.1 3.6 84.1 82.9 85.3 53.1 49.1 56.7 Poorest 2.9 3.3 2.6 67.9 65.0 71.2 9.1 6.7 12.0 11 2.7 1.8 3.8 80.2 80.3 80.0 23.7 18.8 28.3 III 3.1 3.2 * 83.8 82.2 85.2 37.0 32.3 41.4 IV 4.2 * * 84.3 84.4 84.3 53.2 49.8 56.1 Richest * * 85.9 82.4 88.9 64.2 56.9 70.2 Managua 84.7 81.7 87.8 57.2 54.8 59.3 Pacific- Urban 3 1 3.9 85.9 86.3 85.5 49.2 44.2 54.3 Pacific - Rural 84.8 83.3 86.3 23.6 18.4 28.8 Central - Urban 80.0 79.2 80.8 52.3 43.7 59.2 Central - Rural 3.9 2.3 5.3 69.9 67.5 72.3 11.4 4.0 19.3 Atlantic- Urban 4.0 86.5 85.8 87.1 35.3 29.6 41.2 Atlantic - Rural 4.3 * 58.0 57.3 58.9 6.5 8.6 4.6 Total Extreme Poor 3.4 3.9 3.0 65.1 62.7 67.9 8.1 5.8 11.1 Poor 2.8 2.6 3.1 75.6 73.5 77.8 18.8 14.5 23.6 Non-poor 3.0 2.1 3.8 84.4 83.2 85.5 54.3 49.6 58.1 Urban Extreme Poor 3.5 * * 68.1 65.9 70.5 15.1 13.0 17.7 Poor 3.8 2.6 5.0 80 4 79.0 81.9 30.8 29.6 32.1 Non-poor 2.2 1-.8 * 86.2 85.2 87.1 62.8 58.7 66.1 Rural Extreme Poor 3.4 3.6 3.2 64.1 61.7 66.9 5.9 3.5 8.9 Poor 2.4 2.6 2.2 73.1 70.5 75.7 12.8 7.2 19.1 Non-poor I - I 79.7 78.1 81.4 28.8 22.4 34.0 ' number 4-6 yr olds in pre-school/ number of 4-6 yrs old 2 number 7-12 yr olds in elementary school/ number of 7-12 yrs old 3number 13-17 yrolds in secondary school/ number of 13-17 yrs old * nc10 Annex 5. Page 22 E.05 - Reason for Not attending School - 1993 (7 - 12 years old only) Extreme Poor Poor Non-poor Urban Rural Urban Rural Urban Rural Total Male Vacation,sick 0.0 0.0 13.5 0.0 0.0 0.0 3.2 Expensive 41.7 25.3 22.9 26.4 0.0 0.0 20.8 To work or help 0.0 23.7 10.3 22.1 8.9 0.0 16.7 Far 0.0 4.9 0.0 5.3 0.0 0.0 3.1 Repeated Grades 0.0 0.0 0.0 0.0 17.8 0.0 2.8 Not worth it 13.3 0.0 5.2 0.0 11.0 74.8 4.6 No Teacher 0.0 7.2 6.7 8.1 0.0 0.0 6.3 Home Reasons 9.1 7.6 3.5 4.5 14.4 0.0 5.7 Other 35.9 31.4 37.9 33.7 47.9 25.2 36.8 Female Vacation,sick 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Expensive 29.4 21.0 11.7 20.1 0.0 20.8 15.8 To work or help 0.0 6.6 0.0 0.0 0.0 4.6 3.0 Far 33.6 8.6 13.4 0.0 0.0 11.0 9.1 Repeated Grades 0.0 2.3 23.6 0.0 0.0 1.6 4.5 Not worth it 0.0 0.0 0.0 0.0 0.0 2.3 1.5 No Teacher 0.0 9.9 0.0 59.8 9.8 10.2 10.3 Home Reasons 11.3 19.9 17.8 20.1 15.0 18.6 17.9 Other 25.7 31.7 33.5 0.0 75.1 30.8 37.9 Annex 5. Page 23 E.06 - Reason for Not attending School - 1998 (7 - 12 years old only) Extreme Poor Poor Non-poor Urban R Urban R-ura Urban Rural Total Male Age 0.0 2.7 0.0 4.1 0.0 0.0 2.4 Economic Problems 83.7 48.0 75.1 43.7 65.4 23.4 51.9 Rural Activities 0.0 5.7 1.1 7.0 5.7 3.7 5.4 Domestic Duties 0.0 0.0 1.8 0.6 0.0 0.0 0.7 Not interested 10.3 6.3 8.0 6.6 4.7 0.0 6.1 Distance 0.0 20.3 1.4 20.8 0.0 25.2 14.0 Illness 3.7 1.6 7.9 3.5 3.9 0.0 4.1 Not avaialable 0.0 0.0 0.0 0.0 4.2 0.7 0.7 Level not offered 0.0 1.4 0.0 1.1 0.0 0.0 0.6 Insufficient Teachers 0.0 7.9 0.5 5.7 0.0 13.2 4.3 Insufficient Security 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Insufficient Textbooks 0.0 0.8 0.0 0.5 0.0 0.0 0.3 Handicapped 0.0 0.6 3.1 1.0 13.0 0.0 3.2 Other 2.3 4.7 1.2 5.5 3.2 33.8 6.3 Femaie Age 1.7 3.3 0.9 2.4 18.4 0.0 2.9 Economic Problems 84.9 48.9 68.4 42.2 31.1 23.3 45.9 Rural Activities 0.0 1.9 0.0 1.1 0.0 0.0 0.7 Domestic Duties 2.1 4.1 2.7 2.5 10.0 2.7 3.0 Not interested 9.2 3.9 16.2 6.9 12.6 0.0 8.8 Distance 2.2 26.2 1.1 27.2 0.0 34.1 21.3 Illness 0.0 1.0 3.8 0.6 18.6 0.0 2.5 Not avaialable 0.0 0.0 1.1 0.0 0.0 0.0 0.2 Level not offered 0.0 1.6 0.0 1.3 0.0 0.0 0.8 Insufficient Teachers 0.0 3.8 0.0 5.6 0.0 6.2 4.0 Insufficient Security 0.0 1.0 0.0 1.1 0.0 0.0 0.7 Insufficient Textbooks 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Handicapped 0.0 1.6 4.4 2.1 0.0 5.7 2.8 Annex 5, Page 24 E.07 - Nicaragua 1993 and 1998 Percent not Attending School 1993 1998 7 - 12 13 - 18 7 - 18 7- 12 13- 18 yrs yrs yrs yrs yrs 7- 18 yrs All 21.5 43.0 31.0 14.3 43.5 28.2 Extreme Poor 44.7 71.8 55.7 32.4 70.0 49.1 Poor 32.3 57.6 43.0 21.3 60.7 39.1 Non-poor 7.4 26.9 16.5 6.1 26.7 16.4 Urban 11.3 26.6 18.3 9.0 28.1 18.3 Extreme Poor 35.3 57.7 44.5 28.5 61.7 42.9 Poor 19.9 37.6 27.8 15.9 48.5 30.3 Non-poor 6.3 20.8 13.1 5.1 19.4 12.4 Rural 33.7 65.6 47.2 20.1 61.9 39.5 Extreme Poor 47.3 75.8 58.8 33.6 72.6 51.1 Poor 39.0 69.9 51.8 24.1 66.7 43.7 Non-poor 11.7 51.2 30.0 8.7 49.2 28.1 Quintiles Poorest 44.6 71.6 55.6 29.6 68.5 46.8 11 25.1 52.4 36.9 16.1 57.5 35.3 liI 14.0 38.7 25.6 10.7 42.2 25.0 IV 7.8 27.0 16.4 6.9 29.6 18.4 Richest 5.2 20.4 12.4 1.6 16.8 9.6 Region Managua 12.7 27.2 19.2 6.5 23.9 15.0 Pacific- Urban 9.7 25.5 16.8 8.5 31.4 19.5 Pacific- Rural 27.1 61.3 41.0 10.8 53.3 30.7 Central - Urban 10.3 31.8 20.9 12.1 34.1 23.1 Central - Rural 41.8 75.6 56.3 24.5 67.6 44.7 Atlantic- Urban 13.5 23.5 18.0 11.1 39.6 24.0 Atlantic - Rural 41.9 68.1 53.2 40.0 73.4 54.2 Annex 5. Page 25 E.08 - Nicaragua 1993 & 1998 Preschool attendance of Children 0-3 and 4-6 year old (using data from section 4a in both years) 1993 1998 0-3 years 4-6 years 0-3 years 4-6 years CDI, CDI, Pre-school' INSSBI Pre-school' INSSBI Pre-school2 CDI Pre-school2 CDI All 2.7 0.6 31.3 1.3 3.4 1.2 44.8 0.3 Urban 4.3 0.9 48.5 1.8 5.1 1.2 57.1 0.5 Rural 1.0 0.3 13.0 0.8 1.9 1.2 33.6 0.1 Extreme Poor 0.4 0.3 8.3 0.2 2.4 1.4 25.4 0.0 Poor 1.0 0.6 17.4 1.2 2.0 0.8 33.8 0.3 Non-poor 5.4 0.6 50.9 1.5 5.5 1.9 60.3 0.2 Quintiles Poorest 0.3 0.3 9.1 0.2 2.2 1.2 26.8 0.0 11 1.1 0.3 20.6 0.5 1.9 0.4 35.4 0.8 III 2.1 1.1 32.2 4.1 1.8 0.4 49.4 0.0 IV 5 1 0.7 47.8 1.7 5.8 0.8 57.0 0.6 Richest 7.9 0.8 63.0 0.4 9.3 5.3 75.8 0.0 Region Managua 6.1 0.6 49.3 2.1 4.5 1.3 62.5 0.0 Pacific - Urban 2.8 0.9 43.3 1.9 5.9 0.3 59.0 0.0 Pacific - Rural 1.0 0.0 14.9 0.0 2.5 0.6 45.8 0.0 Central - Urban 2.8 1.8 45.4 3.3 4.5 4.0 47.4 1.7 Central - Rural 0.6 0.1 8.2 0.5 1.8 1.2 26.7 0.0 Atlantic- Urban 4.3 0.5 42.3 0.0 3.1 1.9 43.8 1.3 Atlantic - Rural 0.0 . 1.3 8.7 0.0 1.7 0.5 15.3 0.5 'includes codes 1,2,3 for q2 in section 3c 2includes codes 1,3 for ql in section 4a Note: 538 respondents in pre-school/CDI in 1993 and 727 respondents in pre-school/CDI/school in 1998 Annex 5. Page 26 E.09 - 1998 Primary school repetition rates, percent with no books and mean number of days absent No books Repetition (%) (%) Days absent All 11.2 10.4 5.1 Rural 12.1 8.8 5.8 Urban 10.4 11.7 4.4 Poorest 14.4 13.5 5.3 11 14.3 10.0 5.0 IlIl 9.6 9.8 5.4 IV 11.2 9.4 4.5 Richest 5.1 8.9 5.3 Managua 8.2 15.0 4.6 Pacific- Urban 13.8 5.7 4.2 Pacific - Rural 14.3 4.3 5.0 Central - Urban 10.1 10 .5 4.6 Central - Rural 11.9 10.7 6.4 Atlantic- Urban 9.1 13.6 5.6 Atlantic - Rural 8.0 20.1 7.7 Total Extreme Poor 14.7 14.2 5.4 Poor 13.5 10.9 5.3 Non-poor 8.7 9.8 4.8 Urban Extreme Poor 13.9 17.0 4.5 Poor 14.5 13.2 4.6 Non-poor 8.3 10.9 4.4 Rural Extreme Poor 15.0 13.1 5.7 Poor 12.9 9.6 5.8 Non-poor 9.9 6.5 5.8 Note Official repetition rate MECD9B Repetition rate LSMS98 Annex 5, Page 27 E.10 - 1998 Secondary school repetition rates, percent with no books and mean number of days absent Repetition No books (%) (%) Days absent All 8.7 37.0 4.4 Rural 7.4 30.5 3.8 Urban 9.1 39.0 4.7 Poorest 8.2 42.3 4.8 11 7.7 34.9 4.9 III 9.2 41.5 4.0 IV 10.4 43.4 4.6 Richest 7.6 29.2 4.4 Managua 10.1 38.1 3.7 Pacific- Urban 9.1 35.8 5.4 Pacific - Rural 8.8 34.0 5.2 Central - Urban 8.1 44.2 4.4 Central - Rural 2.5 24.2 3.3 Atlantic- Urban 4.9 41.0 5.7 Atlantic - Rural 29.1 26.9 9.4 Total Extreme Poor 8.4 43.4 4.9 Poor 8.4 37.8 4.6 Non-poor 8.8 36.7 4.4 Urban Extreme Poor 0.0 41.0 9.1 Poor 8.7 44.2 5.7 Non-poor 9.2 37.9 4.5 Rural Extreme Poor 15.8 45.5 2.3 Poor 8.0 29.5 3.2 Non-poor 7.0 31.3 4.1 Official repetition rate MECD98 Repetition rate LSMS98 E.11 - Nicaragua - 1993 School Attendance of Children 6-18 Years Old Attending Not attending Inapplicable* Total 13.9 Extreme Poor 42.1 18.7 39.2 Poor 55.0 16.4 28.7 Non-poor 82.2 10.9 6.9 Urban 80.3 10.6 9.1 Rural 51.0 17.9 31.1 *Coded as 7 which is 'inap' in the codebook Annex 5, Page 28 E12 - Nicaragua 1998: Household expenses on education byeducational lesel and povety group (public school enrolmiert only Household expenses on education Tctal~ Pe b( e Tdag FH stu*int stucent education education spending espenses eX1rneS as as a seare as a share a shae of of:)er Educatiora Net Share of net Schdo Scihoo Sd-cd Text- of Ih non- per-cpta capita non level & perety ErTolment Enrdirrents fees Tuition Registration b sfcood knifonTrs Supplies boos food non-food food group rate (O/N) (q26) (q27) (q29) (q28) (q30) (q31) (q32) Tdat expenses expenses' expe(ses PRNIAARY_ All 68,2 1)00 9.3 0.2 2.2 21 4 40.6 23 3 3.0 100.0 5.3 18.8(23.4) 4c 5 Poor 72.9 58 7.7 0.3 21 17.1 42.8 26.5 3.5 100.0 6.3 21.9(41.3) 6c8 Ncr-poor 62.6 42 10.3 0.1 2.2 24.2 39.3 21.3 2.6 100.0 43 14.7(18.2) 212 SECONDARY Ail 261 100 163 0.3 37 32.0 25.9 18.3 3.5 100.0 67 29.8(28.8) 177 Poor 159 30 19.8 0.3 47 21.8 28.7 21.6 3.1 1000 7.9 43.2(52.3) 3! 9 Noa-pmr 3680 70 15.5 0.3 35 34.3 25.2 17.6 3.6 1000 6.2 24.8(26.1) 12 3 TERTIARY' 24 100 16.1 0.1 13.6 47.0 20 10.1 111 100.0 7.8 39.1(30.6) 44`c.9 Poor 0.6 11 129 0.0 14.3 49.8 2.1 11.9 9.0 100.0 108 721(813) 11031 Non-poDr 37 89 16.3 0.2 13.5 46.7 20 10.0 113 100.0 7.4 351(293) 37'.2 Source LSMVS 1998 ' This category indudes Lniversity and techrical 'The nrtnbr in parenthesis is the ratio of the surn of expenditures Nde: Shares were done orgy for househocdds wth at Ieast one dhid in pnrmiy schd Annex 5, Page 29 PART F - HEALTH F.01 Nicaragua - 1993 Fertility by Quintile, Poverty Status and Urban/Rural (Women 15 49 years of age) T otai Fertility (Births Quintile per woman) All 2.8 Poorest 4.2 11 3.4 III 2.9 IV 2.3 Richest 1.9 Managua 2.5 Pacific- Urban 2.3 Pacific - Rural 3.7 Central - Urban 2.5 Central - Rural 3.6 Atlantic- Urban 2.6 Atlantic - Rural 4.0 Poverty Status- Total Extreme Poor 4.2 Poor 3.6 Non-poor 2.2 Poverty Status- Urban 2.4 Extreme Poor 3.6 Poor 3.1 Non-poor 2.1 Poverty Status- Rural 3.6 Extreme Poor 4.4 Poor 3.9 Non-poor 2.8 Annex 5, Page 30 F.02 - Nicaragua 1998 Fertility by Quintile, Poverty Status and Urban/Rural (Women 15-49 years of age) Quintile Total Fertility (Births per woman) All 2.5 Poorest 3.8 11 3.1 III 2.5 IV 2.0 Richest 1.6 Managua 2.1 Pacific- Urban 2.2 Pacific - Rural 2.9 Central - Urban 2.1 Central - Rural 3.0 Atlantic- Urban 2.8 Atlantic - Rural 3.7 Poverty Status-Total Extreme Poor 3.9 Poor 3.3 Non-poor 1.9 Poverty Status-Urban 2.2 Extreme Poor 3.5 Poor 3.2 Non-poor 1.9 Poverty Status-Rural 2.9 Extreme Poor 4.0 Poor 3.4 Non-poor 2.1 Annex 5, Page 31 F.03 - Nicaragua 1993 DPT and Polio immunization by Quintile, Poverty Status and Uraban/Rural (% of 12-23 months of age with card) Times Immunized DPT Times Immunized Polio Quintile 0 1 2 3+ 0 1 2 3+ All 1.6 6.8 9.8 81.9 0.7 3.6 6.0 89.6 Poorest 0.9 6.4 7.8 84.9 1.7 2.2 6.0 90.1 11 2.4 8.8 10.8 78.1 0.5 6.0 10.3 83.3 III 0.5 4.6 10.9 84.1 0.0 4.4 6.6 89.1 IV 2.3 9.4 13.0 75.3 0.8 5.2 5.5 88.5 Richest 2.0 4.2 6.7 87.1 0.0 0.0 0.0 100.0 Managua 0.0 9.3 15.1 75.6 0.0 3.5 7.0 89.5 Pacific- Urban 5,4 7.1 5.4 82.2 0.0 3.6 5.4 91.1 Pacific - Rural 0.0 4.1 2.0 93.9 0.0 2.0 2.0 95.9 Central - Urban 1.2 5.6 4.5 88.7 1.2 0.8 3.7 94.4 Central - Rural 1.5 6.7 8.2 83.6 1.6 5.1 6.0 87.3 Atlantic- Urban 0.0 5.9 23.5 70.6 0.0 5.9 11.8 82.4 Atlantic- Rural 3.0 6.1 30.3 60.6 3.0 6.1 15.2 75.8 Poverty Status Total Extreme Poor 0.9 6.5 7.9 84.6 1.8 2.2 6.1 89.9 Poor 1.4 6.8 9.3 82.6 1.0 3.8 7.4 87.8 Non-poor 1.8 6.7 10.6 80.9 0.4 3.4 4.0 92.3 Urban 2.0 7.6 10.5 79.9 0.3 3.4 5.7 90.6 Extreme Poor 0.0 5.2 1.7 93.1 0.0 0.0 5.2 94.8 Poor 1.4 6.5 10.6 81.6 0.0 2.5 8.7 88.9 Non-poor 2.3 8.4 10.5 78.8 0.5 4.0 3.8 91.7 Rural 1.1 5.8 9.1 84.0 1.2 3.9 6.3 88.6 Extreme Poor 1.2 6.9 9.5 82.5 2.2 2.8 6.4 88.7 Poor 1.4 7.0 8.5 83.1 1.5 4.6 6.8 87.2 Non-poor 0.0 1.5 11.1 87.4 0.0 1.5 4.6 93.9 Annex 5, Page 32 F.04 - Nicaragua 1998 DPT and Polio immunization by Quintile, Poverty Status and Uraban/Rural (% of 12-23 months of age with card) Times Immunized DPT Times Immunized Polio Quintile 0 1 2 3 0 1 2 3 All 0.3 3.4 6.2 90.1 0.1 3.0 4.9 92.1 Poorest 0.0 6.5 8.3 85.2 0.0 5.9 5.5 88.6 11 0.9 3.8 5.4 89.9 0.2 2.3 8.0 89.5 III 0.0 0.0 3.3 96.7 0.0 2.5 3.0 94.4 IV 0.4 3.7 3.2 92.8 0.0 0.7 2.3 97.0 Richest 0.0 0.0 13.8 86.2 0.0 2.3 2.3 95.4 Managua 0.0 2.7 5.9 91.3 0.0 2.7 2.7 94.6 Pacific- Urban 0.0 0.0 3.7 96.4 0.0 1.4 2.5 96.1 Pacific- Rural 0.0 5.8 4.1 90.1 0.0 4.9 3.2 91.9 Central - Urban 0.0 3.5 0.0 96.5 0.0 0.0 0.0 100.0 Central - Rural 0.7 3.5 5.4 90.5 0.0 3.0 5.6 91.4 Atlantic- Urban 0.0 4.3 15.5 80.2 0.0 3.6 12.2 84.1 Atlantic- Rural 1.7 6.0 17.9 74.4 0.9 5.3 14.5 79.4 Poverty Status Total Extreme Poor 0.0 7.2 5.7 87.1 0.0 6.5 5.1 88.4 Poor 0.4 4.6 7.1 87.9 0.1 3.6 6.6 89.7 Non-poor 0.2 1.6 4.7 93.5 0.0 2.2 2.0 95.8 Urban 0.0 2.2 5.8 92.0 0.0 2.1 3.0 95.0 Extreme Poor 0.0 3.3 10.1 86.6 0.0 2.3 10.1 87.6 Poor 0.0 2.0 6.3 91.8 0.0 0.5 7.6 91.9 Non-poor 0.0 2.4 5.5 92.1 0.0 3.2 0.8 96.0 Rural 0.6 4.4 6.5 88.5 0.1 3.8 5.8 90.2 Extreme Poor 0.0 8.1 4.7 87.2 0.0 7.5 4.0 88.6 Poor 0.6 5.7 7.5 86.2 0.2 5.0 6.2 88.7 Non-poor 0.5 0.0 3.1 96.4 0.0 0.0 4.6 95.4 Annex 5, Page 33 F.05 - Nicaragua 1998 DPT and Polio immunization by Quintile, Poverty Status and Urban/Rural (% of 12-23 months of age) Times Immunized DPT Times Immunized Polio Quintile 0 1 2 3 0 1 2 3 All 4.7 7.8 5.7 81.8 5.2 5.7 5.3 83.8 Poorest 3.2 12.9 7.3 76.6 3.2 7.8 5.7 83.2 11 3.7 10.9 5.3 80.1 3.2 7.9 7.0 82.0 III 4.7 0.0 3.7 91.6 4.7 2.2 2.5 90.7 IV 4.4 6.5 2.9 86.3 6.2 4.9 6.4 82.5 Richest 11.6 4.5 11.8 72.1 14.9 2.8 3.1 79.2 Managua 0.0 13.9 3.7 82.4 3.4 10.2 5.2 81.1 Pacific- Urban 3.5 0.9 3.6 92.0 3.5 1.0 2.8 92.8 Pacific - Rural 4.5 10.4 5.1 80.1 4.5 6.4 2.3 86.8 Central - Urban 2.7 4.4 5.4 87.5 2.7 2.0 2.6 92.8 Central - Rural 6.8 4.7 4.5 84.1 6.2 4.4 5.9 83.5 Atlantic- Urban 0.0 8.6 16.3 75.2 0.0 9.0 11.0 80.0 Atlantic - Rural 17.8 12.0 11.7 58.5 17.3 6.5 12.0 64.3 Poverty Status Total Extreme Poor 3.6 14.1 5.5 76.8 3.6 8.4 5.7 82.3 Poor 3.8 10.5 6.4 79.2 3.6 7.0 6.1 83.3 Non-poor 5.9 4.0 4.7 85.4 7.5 3.8 4.1 84.6 Urban 1.6 8.3 5.9 84.3 3.1 6.4 5.1 85.4 Extreme Poor 0.0 11.5 10.5 78.0 0.0 5.6 7.1 87.4 Poor 0.0 11.3 5.6 83.1 0.0 7.3 5.7 87.0 Non-poor 2.7 6.1 6.1 85.2 5.3 5.7 4.7 84.3 Rural 7.5 7.5 5.6 79.5 7.1 5.1 5.4 82.4 Extreme Poor 4.5 14.7 4.4 76.5 4.5 9.1 5.3 81.1 Poor 5.8 10.1 6.9 77.3 5.4 6.9 6.2 81.4 Non-poor 12.3 0.0 2.0 85.8 11.9 0.0 2.9 85.1 Annex 5. Page 34 F.Ub - Nicaragua - 1i99 inciaence of Diarrhea Quintile Had Diarrhea? (%) All 18.3 Poorest 18.0 11 20.6 III 17.6 IV 17.7 Richest 16.8 Managua 13.5 Pacific- Urban 13.3 Pacific - Rural 20.4 Central - Urban 20.9 Central - Rural 22.2 Atlantic- Urban 18.9 Atlantic - Rural 27.7 Poverty Status- Total Extreme Poor 17.7 Poor 18.5 Non-poor 18.0 Poverty Status- Urban . 15.4 Extreme Poor 18.2 Poor 15.5 Non-poor 15.3 Poverty Status- Rural 21.5 ExTreme Poor 17.6 Poor 20.1 Non-poor 26.9 Annex 5. Page 35 F.07-1998 ncidence ofDia and Typeof Care Fthosereporting Lha (Chiliren u r 6years oF age) Of those reporing Dianflea: Type of treatrrent (%/6) Consulted (%/.) WMan Dianriea last HOne Saline/Oral Consulted consultations Quinlile month (/) Mediaciel vkey Medicine None soniTeone (%/6) Doctor Nurse Other (nurrber) AJI 19.6 14.9 26.4 53.7 5.0 54.7 81.7 11.7 6.6 1.5 RLural 23.1 16.9 29.4 48.7 5.0 50.9 74.3 18.8 6.9 1.4 Urban 15.9 12.1 21.7 61.3 4.9 60.5 91.4 2.6 6.0 1.7 Poorest 23.9 20.8 24.9 48.3 6.0 45.2 66.6 22.7 10.7 1.4 11 21.4 16.5 29.4 48.7 5.4 54.8 86.4 12.0 1.6 1.6 III 16.8 9.0 27.4 60.5 3.1 63.3 82.5 10.5 7.0 1.5 IV 19.0 11.0 22.5 61.3 5.2 51.7 91.8 2.8 5.4 1.7 Richest 12.9 7.8 28.0 60.6 3.6 79.3 91.4 0.0 8.6 1.6 Managua 10.8 2.4 20.7 66.8 10.1 55.7 93.4 0.0 6.6 1.9 Pacific- Urban 17.6 18.6 26.5 51.5 3.4 59.3 88.3 2.6 9.1 1.6 Pacific-Rural 19.3 17.1 31.3 47.7 3.9 47.2 82.9 11.0 6.1 1 6 Central - Urban 17.8 14.2 23.4 61.0 1.4 64.9 95.9 4.1 0.0 1.6 Central - Rural 264 15.1 31.5 49.4 4.0 57.2 77.2 18.6 4.2 1.4 Atlantic- Urban 21.1 11.6 17.3 64.3 6.8 62.7 89.2 4.5 6.3 1.6 Atlantic - Rural 29.4 22.9 17.1 52.3 7.7 38.9 44.2 35.1 20.7 1.5 Total Extreme Poor 24.4 23.0 21.7 48.5 6.8 46.1 66.7 22.7 10.6 1.4 Poor 21.7 18.4 26.3 49.7 5.6 49.1 76.5 16.2 7.3 1.5 Non-poor 16.7 8.8 26.4 61.0 3.8 65.1 89.1 5.5 5.4 1.6 Urtan Extreme Poor 19.1 24.3 29.4 38.2 8.1 49.4 71.2 10.4 18.4 1.4 Poor 18.1 17.0 22.4 52.9 7.7 53.7 86.7 6.2 7.1 1.6 Non-poor 14.4 7.8 21.1 68.7 2.4 66.5 94.8 0.0 5.2 1.7 Rural Extrenme Poor 25.9 22.7 20.1 50.7 6.5 45.4 64.8 26.1 9.1 1.4 Poor 23.6 18.8 27.9 48.5 4.8 47.3 71.9 20.7 7.4 1.5 Non-poor 21.6 10.2 34.3 49.6 5.9 62.9 80.3 14.2 5.5 1.3 indudes the other category Annex 5, Page 36 F.08 - Nicaragua 1998 Reason for not seeking care of those reporting Diarrhea last month (Children under 6 years of age) Economic Quintile Not Serious No time Too far Bad care problem Other' All 23.5 6.2 18.9 7.1 19.1 25.2 Rural 20.8 5.3 27.5 6.0 17.2 23.2 Urban 28.7 7.8 2.3 9.0 22.8 29.4 Managua 21.9 5.4 0.0 11.7 34.1 26.9 Pacific- Urban 22.6 16.4 4.3 7.3 17.9 31.5 Pacific - Rural 14.7 14.9 21.3 3.0 18.2 27.9 Central - Urban 44.1 0.0 4.1 3.5 15.2 33.1 Central - Rural 24.6 1.1 27.2 8.1 17.7 21.3 Atlantic- Urban 32.8 0.0 0.0 11.7 26.1 29.5 Atlantic - Rural 22.3 2.4 38.4 6.3 11.9 18.7 Poorest 21,4 6.7 23.5 9.4 17.9 21.2 11 13.0 6.7 23.3 3.4 33.3 20.3 III 38.0 6.9 12.8 0.8 14.1 27.4 IV 28.3 3.7 11.0 12.7 6.4 37.9 Richest 39.6 5.6 0.0 5.8 13.0 36.0 Poverty Status Total Extreme Poor 24.2 6.9 23.9 8.2 18.8 18.0 Poor 19.3 6.1 23.2 6.2 23.5 21.7 Non-poor 34.9 6.3 7.1 9.4 7.3 35.0 Urban Extreme Poor 30.4 12.1 5.5 6.0 13.8 32.2 Poor 19.3 7.0 4.2 5.4 35.0 29.1 Non-poor 40.3 8.9 0.0 13.5 7.8 29.5 Rural Extreme Poor 23.0 5.9 27.5 8.6 19.8 15.2 Poor 19.3 5.9 29.8 6.5 19.5 19.0 Non-poor 27.8 2.9 16.6 3.9 6.8 42.0 includes categories - too expensive, long wait, lack of personnel/medicine and other Annex 5, Page 37 F.09 - Nicaragua 1998 Of those consulting for Illness, time spent waiting for medical attention, by facility Time spent waiting for medical attention last time (hours) First-aid Health Hospital Polyclinic Private Private All station center MINSA INSS Hospital Pharmacy Clinic Poverty Status Total 8.5 7.4 9.7 8.2 18.0 9.2 3.5 4.4 Extreme Poor 11.9 8.6 13.4 13.0 * * * * Poor 9.2 8.0 9.4 10.6 10.7 * 8.3 Non-poor 8.0 6.0 10.0 6.5 19.8 12.0 2.9 3.6 Urban 7.3 0.9 10.0 8.4 10.4 2.9 3.7 Extreme Poor 9.0 14.2 * * * * Poor 7.5 1.3 7.8 8.5 * * 70 Non-poor 7.2 0.6 11.4 8.3 97 * 3.2 Rural 9.8 9.8 9.4 8.0 24.3 13.8 3.8 6.2 Extreme Poor 12.7 9.5 13.1 18.0 * * * * Poor 10.0 9.3 10.0 12.4 * * * 9.8 Non-poor 9.4 12.0 7.8 2.8 * * 4.6 Poorest 10.4 7.6 12.5 10.2 * * 45 11 9.2 5.2 9.2 9.4 * 10.8 III 8.1 12.2 6.7 10.2 * 74 IV 7.5 6.8 10.1 8.2 * * 39 Richest 8.2 * 12.6 5.3 23.1 2.8 Managua 5.9 * 9.7 4.9 7.8 * 2.4 Pacific- Urban 8.6 1.1 9.8 8.8 16.6 * * 6.5 Pacific - Rural 11.7 14.8 12.7 5.5 * * * 4.9 Central - Urban 5.5 * 6.4 7.5 * * * 3.2 Central - Rural 9.8 6.0 8.4 9.8 * * * 7.7 Atlantic- Urban 9.6 * 13.9 11.3 * * * 1.1 Atlantic - Rural 7.5 5.6 5.9 19.0 6.4 * N < 10 Annex 5, Page 38 F.10 - Nicaragua 1998 Of those consulting for Illness, cost of round trip transportation for last consultation, by facility Cost of round trip transportation for last consultation First-aid Health Hospital Polyclinic Private PrivateT All station center MINSA INSS Hospital Pharmacy Clinic: Poverty Status Total 9.3 1.0 3.0 21.3 14.4 9.8 2.6 15.3 Extreme Poor 4.1 1 4 3.8 8.5 * 8.4 Poor 5.2 1.2 2.6 15.4 15.8 8.4 3.9 16.4 Non-poor 12.1 0.6 3.4 24.9 14.0 10.2 2.2 9.9 Urban 10.3 1.1 1.7 20.7 10.1 9.7 1.8 15.6 Extreme Poor 2.2 3.9 1.4 2.2 * * t t Poor 4.6 1.8 1.1 14.5 * * 2.0 6.7 Non-poor 12.2 0.4 2.1 23.4 10.6 10.3 1.7 16.5 Rural 83 1 0 3.9 22.2 19.5 10.2 3.4 14.8 Extreme Poor 4.7 1 0 4.5 10.1 * * * Poor 5.6 1.1 3.3 16.2 * . 5.9 12.1 Non-poor 12.0 0 8 5.2 28.3 20.1 2.7 16.0 Poorest 4.1 1.3 3.6 7.7 * 8.3 11 63 1.4 26 19.9 . 3.0 9.2 III 55 07 2.2 13.1 10.2 8.4 2.2 12.6 IV 82 07 39 15.4 11.0 11.4 2.5 11.7 Richest 18.0 0.6 3.0 39.2 16.1 10.9 2.8 19.2 Managua 8.6 0.8 3.5 18.8 8.3 10.5 4.5 9.3 Pacific- Urban 7.9 0.3 1.2 14.0 15.6 * 1.2 13.5 Pacific - Rural 7.3 0.9 3.6 18.6 * * 5.7 14.8 Central - Urban 5.9 0.9 1.0 16.9 * 7.2 Central - Rural 7.6 0.8 3.8 18.9 * * 13.4 Atlantic- Urban 34.7 5.5 3.2 47.3 * * 103.2 Atlantic- Rural 14.2 1.8 4.3 38.1 * . 52.7 * N < 10 Annex 5, Page 39 F.11 - Nicaragua 1998 Of those consulting for Illness, cost of last consultation, by facility Cost of last consultation by appointment last month First-aid Health Hospital Polyclinic Private Private All station center MINSA INSS Hospital Pharmacy Clinic Poverty Status Total 15.6 1.5 1.4 3.9 0.0 14.7 3.8 50.7 Extreme Poor 2.2 1.0 1.1 2.3 * * 30.2 Poor 4.8 1.3 1.2 4.5 0.0 2.9 6.2 29.0 Non-poor 23.0 2.0 1.6 3.5 0.0 17.7 3.1 54.9 Urban 20.5 1.9 1.3 4.1 0.0 17.8 3.6 52.5 Extreme Poor 1.2 0.6 0.4 3.0 * * * * Poor 6.0 1.2 1.2 4.5 * * 0.0 32.2 Non-poor 25.3 2.6 1.4 2.6 0.0 18.8 4.7 54.7 Rural 10.3 1.4 1.5 3.6 0.0 4.3 4.1 47.0 Extreme Poor 2.5 1.0 1.2 2.2 * . * * Poor 4.2 1.3 1.3 4.5 * 12.8 26.9 Non-poor 18.4 1.7 1.9 2.6 0.0 1.7 55.3 Poorest 2.9 0.9 1.1 2.0 26.0 11 5.6 1.8 1.6 4.1 2.8 29.6 III 7.5 1.6 1.2 4.9 0.1 0.3 6.5 36.6 IV 18.0 2.4 1.8 2.2 0.0 23.0 1.0 50.9 Richest 33.5 1.0 1.6 5.3 0.0 18.5 8.1 59.1 Managua 22.2 5.3 1.9 2.4 0.0 15.7 6.3 55.4 Pacific- Urban 17.5 0.7 0.9 4.6 0.0 3.0 48.5 Pacific - Rural 9.9 0.6 1.2 2.8 6.8 44.8 Central - Urban 25.3 0.0 1.7 6.3 * * 58.8 Central - Rural 7.4 1.6 1.5 2.2 . * * 34.0 Atlantic- Urban 14.6 2.8 1.0 1.6 * * * 67.3 Atlantic - Rural 9.9 1.1 2.0 15.0 58.5 * N < 10 Annex 5, Page 40 F.12 - Nicaragua 1998 Of those consulting for Illness, other health expenditures last time, by facility Other Costs of last consultation last time First-aid Health Hospital Polyclinic Private Private All station center MINSA INSS Hospital Pharmacy Clinic; Poverty Status Total 104.2 20.2 39.0 123.4 60.4 691.2 52.0 178.3 Extreme Poor 26.3 18.6 17.6 65.7 * t * 54.1 Poor 40.2 17.9 26.8 78.2 24.2 44.8 63.1 90.8 Non-poor 147.7 25.3 54.5 151.6 68.9 856.8 48.6 195.1 Urban 136.4 29.4 46.1 135.9 33.7 895.7 61.7 178.6 Extreme Poor 29.1 42.1 17.7 62.1 Poor 44.7 27.3 27.8 82.3 75.1 64.1 Non-poor 167.4 31.5 58.5 159.4 36.0 958.8 57.4 190.9 Rural 68.9 17.1 33.4 104.4 93.4 25.4 42.8 177.6 Extreme Poor 25.5 14.9 17.6 66.6 * * * Poor 38.0 15.9 26.4 74.5 50.4 108.9 Non-poor 110.2 21.0 48.7 135.0 127.4 * 40.8 206., Poorest 27.4 21.8 17.7 60.5 * 61.5 11 44.6 14.3 29.6 55.3 * 38.4 94.7 III 59.7 16.8 40.4 86.3 13.9 47.7 62.4 132.7 IV 75.6 26.2 56.6 90.2 21.4 63.5 47.6 113.0 Richest 243.1 27.8 60.5 236.0 102.1 1745.5 52.3 242.0 Managua 158.8 18.4 37.9 88.1 56.9 925.5 71.1 154.9 Pacific- Urban 120.4 29.8 50.2 195.4 48.7 * 61.8 191.7 Pacific - Rural 77.4 29.8 38.7 143.9 47.7 188.1 Central - Urban 119.4 29.0 47.3 119.5 * f i 199.6 Central - Rural 59.4 9.3 28.0 90.0 * 159.3 Atlantic- Urban 97.0 18.8 44.8 129.3 * * * 204.7 Atlantic - Rural 48.9 7.6 46.8 41.6 . 199.8 * N < 10 Annex 5, Page 41 F.13 - 1998 Of those consulting for Illness, total cost of last consultation, by facility Total cost of last consultation First-aid Health Hospital Polyclinic Private Private All station center MINSA INSS Hospital Pharmacy Clinic Poverty Status Total 129.2 22.7 43.3 148.6 74.8 715.7 58.4 244.3 Extreme Poor 32.6 21.0 22.5 76.5 * * 92.7 Poor 50.2 20.4 30.7 98.1 40.0 56.1 73.1 129.7 Non-poor 182.8 28.0 59.5 180.1 83.0 884.7 54.0 266.4 Urban 167.2 32.4 49.2 160.7 43.8 923.3 67.1 246.7 Extreme Poor 32.5 46.7 19.5 67.3 * * * * Poor 55.3 30.3 30.1 101.2 * * 77.0 103.0 Non-poor 204.9 34.4 62.1 186.8 46.5 987.9 63.8 262.1 Rural 87.5 19.6 38.8 130.1 113.0 39.9 50.3 239.5 Extreme Poor 32.7 17.0 23.4 78.9 * * * Poor 47.8 18.3 31.0 95.2 * 69.0 147.9 Non-poor 140.6 23.5 55.7 165.9 147.5 * 45.2 277.6 Poorest 34.3 23.9 22.4 65.0 * * 95.8 11 56.5 17.4 33.8 110.3 * 44.3 133.6 III 72.7 19.1 43.8 103.8 24.2 56.4 71.0 182.0 IV 101.8 29.2 62.3 107.8 32.4 98.0 51.1 175.6 Richest 294.6 29.3 65.1 280.5 118.1 1774.9 63.2 320.4 Managua 189.6 24.5 43.3 109.2 65.3 951.7 81.9 219.5 Pacific- Urban 145.9 30.7 52.3 214.0 64.3 65.9 253.8 Pacific- Rural 94.7 31.2 43.5 165.3 * 60.2 247.7 Central - Urban 150.6 29.8 50.0 142.6 * * * 265.5 Central - Rural 74.5 11.8 33.2 111.1 * * * 206.8 Atlantic- Urban 146.2 27.1 49.0 178.2 * * 375.2 Atlantic- Rural 73.0 10.5 53.1 94.6 311.0 * N < 10 Annex 5, Page 42 F-.14 - 1991 Reason tor not seeKing care ot those ill last month Not -Economic Had Know Serious No time Too far Bad care problem medicine disease Other' All 28.3 5.3 7.7 3.6 17.8 3.9 17.9 15.6 Rural 26.2 3.9 13.6 3.9 19.3 3.0 16.1 14.1 Urban 30.6 6.8 1.0 3.3 16.1 5.0 20.0 17.2 Managua 29.1 7.3 1.4 3.9 15.6 7.6 15.6 19.4 Pacific- Urban 29.7 6.7 0.7 2.4 19.6 3.5 19.8 17.7 Pacific - Rural 24.6 5.6 7.4 3.4 24.3 2.9 20.5 11.3 Central - Urban 35.0 4.2 2.4 2.9 14.8 4.3 23.6 12.7 Central - Rural 31.7 3.6 14.4 2.9 16.7 2.1 13.0 15.6 Atlantic- Urban 24.4 4.9 1.2 10.6 17.5 3.2 23.8 14.4 Atlantic- Rural 14.5 2.3 31.2 5.2 12.9 2.2 17.6 14.2 Poorest 22.3 3.8 14.5 2.6 24.3 2.4 15.3 14.9 11 26.9 4.3 10.6 5.1 22.2 3.6 11.9 15.5 Ill 34.0 5.4 4.5 3.5 16.8 2.9 17.7 15.2 IV 29.5 6.3 4.7 3.1 14.6 5.5 20.9 15.4 Richest 29.7 7.1 1.9 4.2 8.3 5.8 25.9 17.3 Poverty Status Total Extreme Poor 21.8 3.9 15.2 2.4 24.6 2.5 14.4 15.2 Poor 25.9 4.6 11.6 3.7 22.2 3.1 14.1 14.9 Non-poor 30.9 6.0 3.3 3.5 13.1 4.8 22.0 16.3 Urban Extreme Poor 29.1 5.1 3.6 3.6 24.7 4.0 17.9 12.0 Poor 28.5 6.1 .1.6 4.6 23.7 3.7 14.6 17.2 Non-poor 31.8 7.2 0.6 2.6 12.2 5.6 22.8 17.3 Rural Extreme Poor 19.8 3.6 18.5 2.1 24.5 2.1 13.4 16.0 Poor 24.8 3.9 16.1 3.3 21.5 2.7 13.9 13.8 lNon-poor 29.3 3.9 8.4 5.2 14.7 3.4 20.7 14.6 'includes categories - too expensive, long wait, lack of personnel/medicine, difficult schedule and other Annex 5, Page 43 F.15 - 1998 Maternal Health by poverty and region Fteceive . pre-natal care (%) Where gave Birth: Birth attended by. Health Hospital Private Private Midwite's PatientS N -O center MINSA Hospital Clinic House House Clinic Other Doctor Midwife Nurse Other All 80 7 7.8 57.1 37 2.9 2.5 25.2 0.2 0.6 70 4 21.1 34 5.1 Rural 76.9 9.8 45.6 1.5 0.5 3.5 38 5 0.2 0.4 55.8 31.9 44 7.8 Urban 84.4 5.8 68.3 5.8 5.3 1.6 12.2 0.2 07 84.6 10 6 24 2.4 Poorest 675 7.7 40.5 20 0.2 3.3 45 1 07 0.4 47.9 36.0 53 10 9 11 75.7 11.0 54.3 1.9 25 0 26 29.2 0.1 0.5 66 7 25.2 2.6 5.5 III 85 4 7.1 64.4 39 2.8 2.9 18.8 0.0 0.2 75.5 18.2 4.2 2 1 IV 87 7 85 63 3 2.5 39 1.9 18.9 0.3 0.7 80.2 13.7 27 35 Richest 919 3.0 68.5 104 10.0 1 6 54 00 1.1 91.2 5.7 1.4 1 8 Managua 83.0 73 67.1 7.3 6.8 1.9 8.6 0.0 1.0 87.2 8.7 1.3 2.9 Pacific- Urban 87.9 3.9 74.6 2.5 4.6 2.5 11.6 0.0 0 3 85.6 10.6 2.0 1.9 Pacific - Rural 81 8 8.5 54 5 1.3 0.2 3.9 31.6 0.0 0.0 64.9 28.5 2.8 3.8 Central - Urban 90 3 8 1 63.3 6.9 4 0 1.6 15.6 0.6 0.0 81.5 13 4 3.9 1.2 Central - Rural 79.0 12.1 46.0 2 0 0.1 2.0 37.1 0.3 0.5 57.6 27 5 5 7 9.3 Atlantic- Urban 72.8 5 0 52.3 4.3 2.1 2.0 31.5 1.2 1.7 62 3 23.2 5.6 8.9 Atlantic - Rural 46.7 4.2 14 9 0.7 0.8 5.1 72.3 0.7 1.5 18 4 64.7 4.9 12.0 Total Extreme Poor 66 5 7 9 40.1 2.3 0.3 3 7 44.5 0.8 0.5 47 8 36 1 5.3 10.8 Poor 73.7 9.7 50.0 2.2 0 6 3.1 33.7 0.3 0.4 60.1 28.5 4 4 7.1 Non-poor 88.8 5.6 65.6 5.5 5.7 1.9 15.0 0.1 0.7 82.7 12.4 2.2 2.7 Urban Extreme Poor 68 5 10 6 54 2 3.2 1.1 3.2 26.3 0.9 0 6 66.9 19.8 4.9 8.4 Poor 77.9 8 6 64.1 3.0 1 7 2.6 19.5 0.3 0.3 76.6 17 1 2 5 3.8 Non-poor 88.2 4.2 70.9 7.6 7.5 0.9 7.8 0.2 1.0 89 5 6 7 2.3 1.6 Rural Extreme Poor 65.8 7.0 35.4 2.1 0 0 3.9 50.5 0.7 0.4 41 5 41.5 5 4 11.6 Poor 71.4 10.3 42.3 1.7 0.0 3 4 41.5 0.3 0.5 51 1 34 6 5 4 8.9 Non-poor 900 8.7 53.8 0.8 1.6 4.0 31 1 0.0 0 1 67.5 25.2 2.0 5.2 Annex 5, Page 44 F.16 - 1998 Place of consultation by Poverty Group, Quintiles and Geographic area (includes all ill and children under 6 reporting diarrhea) First-aid Station/Health Hospital Polyclinic Patient's Center MINSA INSS Private' Other2 House All 47.7 12.6 3.4 27.4 7.5 1.4 Rural 59.5 10.4 3.0 18.2 8.5 0.4 Urban 36.6 14.8 3.8 36.0 6.6 2.3 Poorest 69.2 11.3 1.0 6.3 11.9 0.3 11 65.5 13.8 2.3 12.8 5.3 0.4 II) 58.8 11.9 2.3 17.4 8.2 1.4 IV 39.3 13.4 3.7 34.4 7.9 1.4 Richest 18.7 12.6 6.3 53.7 5.8 2.9 Managua 28.4 13.3 7.7 40.1 8.1 2.4 Pacific- Urban 39.4 13.6 4.1 32.3 8.5 2.2 Pacific- Rural 59.5 10.8 1.6 20.1 7.3 0.7 Central - Urban 43.8 14.3 0.6 36.9 2.4 2.0 Central - Rural 62.4 8.6 3.1 17.4 8.2 0.2 Atlantic- Urban 49.3 24.8 0.2 21.0 3.8 0.9 Atlantic- Rural 61.6 13.1 0.0 6.9 16.9 1.5 Total Extreme Poor 70.4 10.6 1.2 3.7 13.8 0.3 Poor 67.1 11.8 1.5 11.2 8.2 0.3 Non-poor 33.3 13.2 4.8 39.4 7.0 2.3 Urban Extreme Poor 71.3 11.9 2.1 7.1 7.6 0.0 Poor 59.4 17.5 1.0 14.1 7.6 0.3 Non-poor 28.5 13.8 4.8 43.8 6.2 3.0 Rural Extreme Poor 70.1 10.3 1.0 2.9 15.3 0.4 Poor 70.7 9.1 1.7 9.8 8.5 0.2 Non-poor 42.7 12.2 4.9 30.8 8.7 0.8 'includes private clinic, private hospital and work place 2includes pharmacy, community health worker, medicine man's house and other Annex 5. Page 45 F.17 - Nicaragua 1993 Place of consultation by Poverty Group, Quintiles and Geographic area (includes all ill and children under 6 reporting diarrhea) First-aid Station/Health Hospital Patient's All Center MINSA Private' Other2 House All Public Private All 53.6 19.3 19.1 6.2 1.9 72.9 27.1 Rural 64.4 14.6 12.7 7.3 1.1 78.9 21.1 Urban 47.7 21.9 22.5 5.6 2.3 69.6 30.4 Poorest 73.3 14.6 3.2 7.8 1.1 87.9 12.1 11 67.6 20.4 5.8 4.8 1.4 88.0 12.0 IlIl 61.9 16.6 12.7 7.0 1.8 78.5 21.5 IV 51.1 20.5 21.4 5.5 1.5 71.6 28.4 Richest 34.0 21.3 35.6 6.4 2.7 55.3 44.7 Managua 48.4 20.1 23.5 6.6 1.4 68.5 31.5 Pacific- Urban 40.1 23.6 26.1 6.5 3.8 63.7 36.3 Pacific - Rural 59.4 15.9 15.9 7.6 1.2 75.3 24.7 Central - Urban 69.9 14.4 13.1 1.7 1.0 84.3 15.7 Central - Rural 76.5 11.3 7.1 3.8 1.3 87.8 12.2 Atlantic- Urban 46.4 35.5 13.8 4.4 0.0 81.9 18.1 Atlantic - Rural 41.7 23.3 11.7 21.7 1.7 65.0 35.0 Total Extreme Poor 72.5 14.9 3.3 8.2 1.1 87.4 12.6 Poor 68.4 18.2 6.5 5.5 1.5 86.6 13.4 Non-poor 44.9 20.0 26.5 6.6 2.1 64.9 35.1 Urban Extreme Poor 80.7 11.0 2.5 4.0 1.9 91.6 8.4 Poor 69.3 19.9 6.5 2.6 1.7 89.2 10.9 Non-poor 41.2 22.5 27.4 6.5 2.4 63.7 36.3 Rural Extreme Poor 69.3 16.5 3.7 9.8 0.8 85.7 14.3 Poor 67.8 17.0 6.4 7.5 1.3 84.8 15.2 Non-poor 58.6 10.6 23.0 7.0 0.8 69.2 30.8 'includes private hospital, work place and other private 2includes pharmacy and other F.18 - 1998 Place of consultation by Poverty Group, Quintiles and Geographic Area Percent ist-id Hospital Plolyclinic Patient's PretIIStation/Health MINSA INSS Piae1Ohr2House Center Private 1 Other 2 All 39 4 45.4 12.9 3.6 28. 7 1.6 Rural 44.1 57.2 10.8 3.4 19.1 9.1 0.5 Urban 35.5 34.9 14.7 3.7 37.5 6.6 2.5 Poorest 39.2 67.3 11.9 1.3 6.3 12.9 0.3 Ii 40.1 63.7 13.9 2.6 13.7 5.7 0.4 III 41.3 57.1 11.5 2.6 18.3 9.0 1.6 IV 37.6 39.1 13.9 3.0 34.3 8.1 1.6 Richest 38.7 17.9 12.9 6.6 54.1 5.5 3.0 Managua 31.8 27.2 13.0 7.6 41.4 8.4 2.6 Pacific Urban 40.1 38.2 14.2 4.0 33.0 8.3 2.4 Pacific Rural 43.1 58.3 10.8 1.7 20.9 7.6 0.7 Central Urban 40.5 40.9 14.9 0.7 39.0 2.5 2.1 Central Rural 43.2 60.0 9.0 3.7 17.9 9.0 0.2 Atlantic Urban 38.9 50.0 23.8 0.0 21.5 3.6 1.1 Atlantic Rural 47.7 56.4 14.4 0.0 8.0 19.4 1.9 Total Extreme Poor 38.9 68.6 11.2 1.5 3.2 15.1 0.4 Poor 40.3 65.4 12.2 1.7 11.7 8.7 0.3 Non-poor 38.6 31.8 13.3 4.9 40.4 7.2 2.4 Urban Extreme Poor 34.3 73.0 11.7 2.5 6.2 6.6 0.0 Poor 36.3 58.2 18.2 1.0 14.3 8.0 0.4 Non-Poor 35.0 27.3 13.6 4.7 45.1 6.2 3.2 Rural Extreme Poor 40.4 67.6 11.1 1.3 2.5 17.1 0.5 Poor 42.3 68.9 9.4 2.0 10.4 9.1 0.2 Non-poor 47.8 40.8 12.7 5.3 31.1 9.2 0.9 Annex 5, Page 47 F.19 - 1998 Place of consultation by Poverty Group, Quintiles and Geographic Area (includes all ill excluding children under 6 reporting diarrhea) First-aid HsiaPoylncPatient's Percent Ill StationlHealth Hospital Polyclinic Private 1 Other 2 House Center MNA IS os All 39.4 45.4 12.9 3.6 28.8 7.8 1.6 Rural 44.1 57.2 10.8 3.4 19.1 9.1 0.5 Urban 35.5 34.9 14.7 3.7 37.5 6.6 2.5 Poorest 39.2 67.3 11.9 1.3 6.3 12.9 0.3 11 40.1 63.7 13.9 2.6 13.7 5.7 0.4 ill 41.3 57.1 11.5 2.6 18.3 9.0 1.6 IV 37.6 39.1 13.9 3.0 34.3 8.1 1.6 Richest 38.7 17.9 12.9 6.6 54.1 5.5 3.0 Managua 31.8 27.2 13.0 7.6 41.4 8.4 2.6 Pacific Urban 40.1 38.2 14.2 4.0 33.0 8.3 2.4 Pacific Rural 43.1 58.3 10.8 1.7 20.9 7.6 0.7 Central Urban 40.5 40.9 14.9 0.7 39.0 2.5 2.1 Central Rural 43.2 60.0 9.0 3.7 17.9 9.0 0.2 Atlantic Urban 38.9 50.0 23.8 0.0 21.5 3.6 1.1 Atlantic Rural 47.7 56.4 14.4 0.0 8.0 194 1.9 Total Extreme Poor 38.9 68.6 11.2 1.5 3.2 15.1 0.4 Poor 40.3 65.4 12.2 1.7 11.7 8.7 0.3 Non-poor 38.6 31.8 13.3 4.9 40.4 7.2 2.4 Urban Extreme Poor 34.3 73.0 11.7 2.5 6.2 6.6 0.0 Poor 36.3 58.2 18.2 1.0 14.3 8.0 0.4 Non-Poor 35.0 27.3 13.6 4.7 45.1 6.2 3.2 Rural Extreme Poor 40.4 67.6 11.1 1.3 2.5 17.1 0.5 Poor 42.3 68.9 9.4 2.0 10.4 9.1 0.2 Non-poor 47.8 40.8 12.7 5.3 31.1 9.2 0.9 1 includes private clinic, private hospital and work place 2 include pharmacy, community health worker, medicine man's house and other Annex 5, Page 48 F.21- Nca.a 1993&19 -1 RamdCraitba'ai bP tyGup(irIiLdsaI ill d diIchu rdr6 rpu1ing cdant . _ 5t~~~~~~~~ah I\kicne Firstid fflM Pdyd iniic FViyEe V\brk Fhiwle (miy Myls PSiErtS aicn1 (Ete MNSA lNS9 Htq3t Ran FRanay aind3 v\i2 eIEL he-b ciw 199B Al 45 49.1 19.3 35 24 32 14.8 1.9 28 B LFbr 62 %2 149 00 Q5 22 28 1.1 59 Pccr 7.2 61.2 182 1.0 1.3 27 42 1.4 28 \h4m 29 420 2Q0 45 28 33 19.2 21 33 1996 Al 89 3R8 126 34 20 1.2 29 24.2 1.0 1.0 1.4 27 Bbu eP r 21.1 49.3 106 1.2 G2 00 1.6 36 55 24 Q3 44 Par 145 527 11.8 1.5 1.1 04 1.6 9.7 22 22 Q:3 23 Nbr-~~ 1 47 235 132 48 27 1.7 38 349 01 Q2 2:3 30 Ha h Ft in 1993 2l'b a1&cAeyin 1993 'RSive C#-ffin 1993 Annex 5, Page 49 F.21 - 1998 Health services by Poverty Group, Quintiles and Geographic area (includes all ill and children under 6 reporting diarrhea) Consulted (%) Place of consultation Consultation Coverage Concentration Rates' Rates2 Rates3 Doctor Nurse Other Public Private All 0.66 42.7 1.54 89.2 5.3 5.5 62.9 37.1 Rural 0.58 38.9 1.44 83.1 9.0 7.9 73.2 26.8 Urban 0 74 44.3 1.63 94.7 2.0 3.3 53.6 46.4 Poorest 0.44 31.6 1.41 75.1 14.1 10.8 86.1 13.9 11 0.59 40.1 1.46 88.1 6.5 5.4 80.6 19.4 IlIl 0.67 42.9 1.56 90.9 4.8 4.3 71.7 28.3 IV 0.71 44.9 1.57 89.8 3.9 6.4 56.1 43.9 Richest 0.88 54.5 1.62 95.3 1.7 3.0 37.5 62.5 Managua 0.73 44.1 1.63 93.8 2.8 3.5 47.7 52.3 Pacific- Urban 0.70 43.5 1.62 94.0 1.6 4.4 56.6 43.4 Pacific - Rural 0.63 41.1 1.53 84.3 9.6 6.1 71.5 28.5 Central - Urban 0.75 49.0 1.53 94.8 3.4 1.8 56.9 43.1 Central - Rural 0.59 39.9 1.41 84 6 7.8 7.6 75.0 25.0 Atlantic- Urban 0.75 45.4 1.60 96.5 0.7 2.8 74.0 26.0 Atlantic - Rural 0.41 27.4 1.38 60.4 17.4 22.2 78.6 214 Total Extreme Poor 0.43 31.1 1.39 72.3 15.2 12.5 87.8 12.2 Poor 0.54 37.1 1.45 83.9 8.8 7.4 81.7 18.3 Non-poor 0.77 48.1 1.60 92.9 3.0 4.2 50.0 50.0 Urban Extreme Poor 0.41 26.1 1.42 83.2 11.7 5.2 87.2 12.8 Poor 0.61 36.4 1.59 91.9 3.9 4.2 77.8 22.2 Non-poor 0.80 47.7 1.64 95.6 1.4 3.0 45.7 54.3 Rural Extreme Poor 0.44 *29.4 1.38 69.7 16.0 14.3 87.9 12.1 Poor 0.51 35.2 1.38 80.0 11.1 8.9 83.5 16.5 Non-poor 0.71 45.8 1.52 87.4 6.1 6.5 58.8 41.2 'Consultancy rate=number of visits/number of sick 2Coverage rate=number of people with at least one visit/number of sick 3Concentration rate=number of visits/number of people with at least one visit Annex 5, Page 50 F.22 - Nicaragua - 1993 Distance to Health Post/Center Quintile Oistance (km) Time (hours All 2.9 0.4 Poorest 5.9 0.8 11 3.2 0.6 iII 2.5 0.4 IV 1.7 0.3 Richest 2.9 0.3 Managua 3.3 0.4 Pacific- Urban 0.7 0.2 Pacific - Rural 3.6 0.6 Central - Urban 0.2 0.2 Central - Rural 6.1 0.8 Atlantic- Urban 3.4 0.4 Atlantic - Rural 6.9 0.9 Poverty Status- Total Extreme Poor 6.1 0.8 Poor 3.8 0.6 Non-poor 2.2 0.3 Poverty Status- Urban 1.1 0.2 Extreme Poor 0.8 0.3 Poor 0.4 0.2 Non-poor 1.5 0.3 Poverty Status- Rural 5.6 0.8 Extreme Poor 8.6 1.0 Poor 6.4 0.9 Non-poor 4.3 0.5 F.23 - Nicaragua 1993 Adults consulting when ill (%) Consulting (%/o) Total 41.8 Extreme Poor 2. .5 Poor 31.7 Non-poor 50.0 Urban .5 Extreme Poor 31.2 Poor 40.5 Non-poor 52.7 Rural . Extreme Poor 21.4 Poor 27.3 Non-poor 41.6 Annex 5, Page 51 F.24 - Mca,agia 1998: Househoidepese on heath bylevel andpoaverty 9 ,(pudic facilites oniy) IHaiehd exenses on health Plubic per Total Hh Ta per visit visit health health sperding expenses expenses as as a share Share of as a share a share of of per RPblic Heath utilizabon Cors-lt- Aiibonal of Hh non- per-capta capta non- Care Fadlity- & (wthn Share of abon Traspcort- Iediares services Hosptaliz- food rn-fcod food poverty group group- /o) uilizabon(/) (q32) atbo (q30) (q36) (q37) abom (q39) Tdal epenses evenses expenses FIRST-AID STATION Al 13.4 100 4.6 6.7 83.8 4.9 0.0 10D20 7.6 37.1 7.2 Poor 17.2 69 6.0 6.4 81.6 6.0 0.0 100.0 7.7 38.9 9.1 Non-poor 8.9 31 22 7.3 87.4 3.1 0.0 100.0 7.5 33.1 30 HEALTH CENTER Al 646 100 6.8 3.3 79.6 88 1.5 100.0 13.9 53.5 5.2 Poor 67.1 56 8.6 4.1 78.1 8.2 1.0 1020 142 60.9 7.3 Nr-poor 61.6 44 5.6 27 80.7 9.2 1.8 1020 136 43.9 25 HOSPITAL Al 220 10) 14.3 26 61.5 20.1 1.5 1000 25.3 119.8 30.5 Poor 15.7 39 15.7 4.5 68.7 7.0 4.1 1000 34.8 189.3 54.5 Nr-1pc0r 29.5 61 13.8 20 59.1 24.5 0.6 100.0 19.7 76.6 15.5 Soaroe: LSI\S 1998 * Ths category indudes oSy codes 1-3 of q23 Annex 5, Page 52 PART G - MALNUTRITION G.01 - 1998 Prevalence of Malnutrition by Quintile, Region and Poverty Status (Children under 5 years of age) Underwe ig ht_(weigh t-for-age) Stunting (height-for-age) Wasting (weight-for-height) Malnourished Severe Moderate Total Severe Moderate Total Severe Moderate Total ThDtal Poorest 38 149 187 13.2 21.0 34.2 0.8 26 34 330 11 1 8 10 0 11.8 7.5 15.8 23.3 0.1 4 1 4.2 219 3 III 1.8 6.3 81 3 5 9.3 12.9 0.4 2,9 3.3 1 D 0 IV 1.2 47 5.9 34 77 11.0 04 2.0 24 '3 1 Richest 0.0 3 9 39 1.7 1.8 3.6 0.5 2.2 2 7 4 Managua 07 6.2 6.9 3.8 5.9 9.7 0 0 3.7 37 1 3 4 Pacific- Urban 1.8 8.2 10.0 5.1 12.8 17.9 0.4 1.4 1.8 2D 3 Pacific - Rural 32 9.0 12.2 6.8 13.1 19.9 0.2 2.3 2,5 22 5 Central - Urban 2.3 9.2 11.5 7.3 14.7 22.0 0A4 1.3 1.7 25 3 Central - Rural 3.2 9 7 12.9 10.9 17.6 28.4 0.7 3 7 4 5 .2 7 Atlantic- Urban 0,5 9.3 9 8 5 0 114 16.4 0.0 4.7 4 7 23 5 Atlantic - Rural 0.6 14.9 15.5 7.1 16.4 23.5 1.5 3.3 4.8 2i0 9 Total 2.0 90 11.0 6.8 12.9 197 04 2.9 3.3 ,3.6 Extreme Poor 4.2 15 3 19.5 13.9 21.8 35.7 0.9 3.0 3 9 2 9.5 Poor 2.8 11.4 14.2 9.6 17.2 26.7 05 3.4 3.9 : 1.3 Non-poor 0.8 5A4 6.2 2.8 6.5 9.3 0.3 2.1 2.4 12.2 Urban 1 4 8.1 9.5 5.3 10.5 15.8 0.2 2.5 2.7 19.5 Extreme Poor 3.4 13 1 16.5 12.0 23.7 35.7 0.0 2.2 2.2 29.3 Poor 24 11.0 13.4 8.7 17.2 25.9 0.2 3.3 3.5 :0.6 Non-poor 0.6 5.9 6.5 2.8 5.5 8.2 0.2 1.8 2.0 1 1 Rural 26 9.8 12.4 8.3 15.1 23.4 0.6 3.3 3.9 7.5 Extreme Poor 44 16.0 20.4 14.5 21.2 35.7 1.1 3.3 4.4 :9.5 Poor 3.0 11.6 14.7 10.0 17.1 27.2 0.7 3.5 41 :1.7 Non-poor 1.2 4.2 5.4 2.9 8.9 11.7 05 28 3.3 4.5 Note: Severe values are less than -3 z-score and moderate values are -2 to -3 z-score. Malnourished is defined as either underweight, stunted or wasted GOZ 1990 Pweenece d NPiimftnbcn byAgee g3(atom aude 5ysct ae) OSeve kbeae IV lToZa bae tdeEte Trdai Sie Mxere I U2a 0-5 nrmts 0.0 0.6 0.6 0.4 0.4 0.7 03 0.6 0.9 1.9 6-11 nrrts 23 5.9 82 3.6 5.5 9.1 1.5 7.8 9.3 19.9 12-23 nrrmt 3.0 11.5 14.5 6.9 11.1 18.0 1.0 6.0 7.0 24.9 24-35 rrrith 25 9.7 122 6.6 11.5 18.1 0.2 20 21 21.2 3-59 nrrths 1.6 9.8 11.5 9.0 18.7 27.8 01 1.2 1.3 29.5 Nbte Sfv ase s Is ie-less - -3 z-se rad Tx eveue s ae -2 to-3 z-s Annex 5, Page 53 G.03 - 1998 Percent of Children (0-59 months) classified as rnainourshed by poverty and region Level Stunting (height-for-age) Underweight (Weight-for-age) Wasting (weight-for-height) Poverty Very poor Poor Non-poor Very poor Poor Non-poor Very poor Poor Non-poor Managua 15.0 13.8 7.7 0.0 7.7 6.6 0.0 6.0 2.5 Pacific- Urban 37.7 28.8 7.3 17.2 13.5 6.5 1.0 1.7 1.9 Pacific- Rural 27.0 22.1 13.0 17.1 12.8 10.2 3.3 2.1 3.7 Central - Urban 52.2 39.3 5.5 21.9 19.6 3.8 2.0 2.1 1.2 Central - Rural 45.5 32.3 13.6 24.1 15.8 2.1 4.6 5.7 0.0 Atlantic- Urban 17.7 18.9 13.9 15.0 11.3 8.3 5.5 3.2 6.1 Atlantic- Rural 28.3 25.9 13.7 20.1 18.0 5.8 6.5 4.5 6.1 Note: ?nounshed is defined as less than -2 z-score. G.04 - 1998 Percent of Children (0-59 months) classified as malnourished by poverty and Age Level Stunting_(height-for-age) Underweight (weight-for-age) Wasting (weight-for-height) Poverty Very poor Poor Non-poor Very poor Poor Non-poor Very poor Poor Non-poor 0- 5 nonths 1.6 1.2 0.0 0.0 0.6 0.1 1.2 1.5 0.0 6- 11 mnonths 19.6 10.3 7.3 16.0 9.6 6.2 5.2 11.0 6.8 12 - 17 months 23.2 18.6 10.1 22.2 18.7 9.1 9.6 6.6 6.5 18 - 23 months 41.0 27.1 12.2 27.2 16.8 10.9 8.8 11.3 2.0 24- 35 months 32.6 24.2 9.4 16.6 15.3 7.6 3.4 2.0 2.3 36 - 47 months 47.9 37.1 9.1 25.3 15.7 4.3 2.0 1.1 0.0 48 - 59 months 48.6 40.4 12.6 21.5 16.1 5.0 2.1 1.7 2.3 Note: Malnourished is defined as less than -2 z-score. Annex 5, Page 54 H.01 - Nicaragua 1998 Access to Services/Housing by Poverty Group Access to Services/Housing All Extreme Poor Poor Won-poor Own house with title 34.5 3 Main source of water Pipes inside 27.0 3.0 7.0 39.7 Pipes outside 33.8 16.8 30.0 36.2 Public source 4.5 10.8 8.3 2.1 Public or private well 18.3 33.9 28.4 11.9 River/stream 9.1 26.6 16.8 4.2 Truck/oxcart 0.3 0.0 0.5 0.2 From another house 6.4 8.1 8.3 5.3 Other 0.6 0.7 0.8 0.5 Type of sanitary service Latrine/lavatory 61.7 58.5 66.7 58.5 Toilet discharges into sewers 16.9 0.9 2.8 25.9 Toilet discharges into cesspool 5.5 0.6 1.5 8.0 Toilet discharges into river/stream 0.1 0.0 0.0 0.1 There is none 15.9 40.0 29.0 7.5 Garbage Disposal Collected by truck 31.1 5.3 11.3 43.6 Burned 46.8 54.8 57.9 39.8 Buried 3.4 3.1 3.8 3.2 Made into fertilizer 0.6 0.7 0.8 0.5 Dumped into river/stream 15.7 36.2 25.7 9.4 Authorized dump 2.4 0.0 0.5 3.7 Type of lighting Electric 68.7 24.4 43.3 84.8 Generator 0.3 0.2 0.6 0.2 Kerosene/gas 28.9 69.4 52.7 13.9 Other 1.7 5.1 2.9 0.9 None 0.3 1.0 0.5 0.2 Mean paymet for electricity 105.4 36.3 45.0 123.2 Fuel for cooking Firewood 66.7 98.6 94.3 49.0 Butane/Propane 29.3 0.4 4.5 45.2 Kerosene 1.8 1.0 0.8 2.5 Coal 1.2 0.0 0.3 1.9 Electricity 1.0 0.0 0.2 1.5 Households where at least one member has public hith insurance 20.6 4.0 9.7 27.6 Households where at least one member has private hith insurance 2.0 0.6 0.6 2.9 Distance to health post/center (kms) 2.4 4.7 3.7 1.5 Distance to elementary school (kms) 0.9 1.7 1.3 0.8 Minutes to health post/center 34.7 66.3 51.9 23.8 Minutes to elementary school 16.1 25.7 21.7 12.6 Principal access road is paved 22.3 7.4 11.0 29.6 Households requesting loans last year (%) 19.0 8.5 11.5 23.7 Average amount of loan 8,290 2,444 2,141 9,971 Households buying on credit last year (%) 9.0 4.0 6.3 10.8 Average amount of credit 5,987 1,474 1,135 7,881 Annex 5, Page 55 PART H - HOUSING & BASIC SERVICES H.02 - Nicaragua 1998 Access to Services/Housing by Quintile Access to Services/Housing Al I Poorest 1F III IV Richest Own house with title 45.0 35 4. 427 47. 52. Main source of water Pipes inside 27.0 3.0 6.3 20.6 29.0 55.9 Pipes outside 33.8 19.5 34.5 41.5 41.0 29.3 Public source 4.5 10.1 8.6 3.5 2.9 0.9 Public or private well 18.3 32.3 27.5 18.0 15.3 7.8 River/stream 9.1 25.9 12.9 7.2 5.3 2.3 Truck/oxcart 0.3 0.0 0.6 0.5 0.1 0A4 From another house 6 4 8.4 9.2 7.7 5.8 3.3 Other 0.6 0.9 0.4 1.0 0.6 0.2 Type of sanitary service Latrine/lavatory 61.7 61.2 67.7 71.3 68.5 45.6 Toilet discharges into sewers 16.9 1.0 3.3 10.1 17.7 38.3 Toilet discharges into cesspool 5.5 0.6 2.1 2.5 4.8 12.8 Toilet discharges into river/stream 0.1 0.0 0.0 0.0 0.0 0.2 There is none 15.9 37.2 26.8 16.1 9.0 3.1 Garbage Disposal Collected by truck 31.1 6.2 11.6 25.0 36.0 57.0 Burned 46.8 54.8 62.7 53.3 45.3 29.0 Buried 3.4 3.6 3.5 4.4 3.4 2.5 Made into fertilizer 0.6 0.8 1.0 0.4 0.6 0.4 Dumped into river/stream 15.7 34.7 21.2 15.2 11.6 5.8 Authorized dump 2.4 0.0 0.1 1.8 3.0 5.3 Type of lighting Electric 68.7 25.7 49.4 71.2 80.2 92.6 Generator 0.3 0.2 1.0 0.2 0.3 0.1 Kerosene/gas 28.9 68.4 47.2 26.8 18.1 6.8 Other 1.7 4.9 2.0 1.5 1.3 0.3 None 0.3 0.9 0.4 0.2 0.1 0.2 Mean paymet for electricity 105.4 38.2 42.1 58.9 76.7 175.5 Fuel for cooking Firewood 66.7 98.2 93.2 81.7 60.6 26.5 Butane/Propane 29.3 0.5 5.8 15.1 32.7 67.6 Kerosene 1.8 1.0 0.6 1.7 3.0 2.3 Coal 1.2 0.3 0.2 1.0 2.5 1.6 Electricity 1.0 0.0 0.3 0.6 1.3 2.0 Households where at least one member has public hith insurance 20.6 5.0 10.7 16.3 27.9 32.6 Households where at least one member has private hith insurance 2.0 0.7 0.7 1.6 1.8 4.1 Distance to health post/center (kms) 2.4 4.5 3.5 2.1 1.7 1.3 Distance to elementary school (kms) 0.9 1.6 1.2 0.9 0.7 0.5 Minutes to health post/center 34.7 65.1 47.4 33.1 25.3 19.5 Minutes to elementary school 16.1 25.2 20.8 15.6 13.7 10.6 Principal access road is paved 22.3 7.7 12.8 17.6 21.3 40.5 Households requesting loans last year (%) 19.0 9.4 10.8 17.8 24.0 26.0 Average amount of loan 8,290 2,169 1,893 2,065 5,110 15,494 Households buying on credit last year(%) 9.0 3.7 7.4 7.2 11.2 12.5 Average amount of credit 5,987 1,485 1,135 1,329 14,090 4,703 Annex 5, Page 56 H.03 - Nicaragua 1998 Access to Services/Housing by Zone Pacific - Pacific - Central - Central - Atlantic- Atlantic - Access to Services/Housing All Managua Urban Rural Urban Rural Urban Rural Own house with title 45.0 42.3 45.9 33.8 62 7 46.5 52.6 36.2 Main source of water Pipes inside 27.0 49.0 40.4 3.8 35.5 4.3 16.6 1 8 Pipes outside 33.8 39.5 45.9 29.0 43.2 23.8 18.0 3.8 Public source 4.5 0.0 0.4 5.9 3.1 12.2 11.9 5.1 Public or private well 18.3 4.9 4.6 40.6 6.4 28.9 44.4 39.4 River/stream 9.1 2.4 0.4 8.4 1 0 23.6 1.7 48.3 Truckloxcart 0.3 0.3 0.0 1.2 0.6 0 1 0.0 0.0 From another house 6.4 3.9 8.4 10.7 9.8 4 9 5.5 1.2 Other 0.6 0.0 0.1 0.5 0.4 2.1 0.0 0.4 Type of sanitary service Latrine/lavatory 61.7 50.3 65.1 74.5 62.1 64.4 80.3 46.8 Toilet discharges into sewers 16.9 39.7 18.5 0.2 22.4 0.2 0.6 0.0 Toilet discharges into cesspool 5.5 6.6 12.0 1.5 8.3 0.5 6.8 0.0 Toilet discharges into river/stream 0.1 0.0 0.0 0.0 0.5 0.0 0.0 0.0 There is none 15 9 3.4 4.4 23.8 6.7 34.9 12.3 53.2 Garbage Disposal Collected by truck 31.1 50.1 52 4 0 8 57.0 2.1 23.3 0.7 Burned 46.8 35.4 35 4 80.1 34.5 56.2 50.7 39.8 Buried 3.4 2.9 1.7 2.2 2.3 4.9 8.9 7.4 Made into fertilizer 0.6 0.0 0.1 0.9 0.4 1.7 0.8 0.7 Dumped into river/stream 15.7 4.8 9.2 15.0 4.3 35.1 15 4 51.4 Authorized dump 2.4 6.7 1.1 1.0 1.5 0.1 0.9 0 Type of lighting Electric 68.7 95.8 90.4 48.8 83.8 31.7 68.6 10.2 Generator 0.3 0.0 0.1 0 7 0.0 0.5 0.3 2.0 Kerosene/gas 28.9 4.2 8.2 48.2 12.3 64.1 29.5 83.7 Other 1.7 0.0 1.0 1.9 3.5 3.3 1.6 3.1 None 0.3 0.0 0.3 0.5 0.5 0.4 0.0 1.0 Mean paymetforelectricity 105.4 124.7 111.3 65.6 97.4 46.6 131.2 105.3 Fuel for cooking Firewood 66.7 35.5 62.0 94.2 60.6 94.6 50.7 94.1 Butane/Propane 29.3 57.7 35.3 4.8 37.6 5 1 24.0 3.1 Kerosene 1.8 3.3 1.9 0.8 1.6 0.0 4.6 1.5 Coal 1.2 1.1 0.2 0.0 0.0 0.3 17.7 1.3 Electricity 1.0 2.4 0.6 0.2 0.2 0.0 3.0 0.0 Households where at least one member has public hith insurance 20.6 37.2 23.4 12.9 16.8 5.7 23.2 5.3 Households where at least one member has private hith insurance 2.0 4.6 2.6 0.3 0.8 0.3 1.7 0.4 Distance to health post/center (kms) 2.4 1.2 0.7 3.5 0.8 4.9 1 6.6 Distance to elementary school (kms) 0.9 0.4 0.4 1.1 0.4 1.7 0.5 2.8 Minutes to health posUcenter 34.7 20.8 13.9 49.7 14.5 64.1 19.2 87.4 Minutes to elementary school 16.1 10.6 8.4 20.0 8.1 27.7 10.2 40.0 Principal access road is paved 22.3 32.6 37.8 11.0 19.2 12.1 13.6 0.2 Households requesting loans last year (%) 19.0 24.2 23.3 17.7 21.2 14.1 4.2 4.9 Average amount of loan 8,290 5,953 9,247 14,217 9,805 6,084 7,593 7,268 Households buying on credit last year (%) 9.0 10.0 12.2 8.9 11.5 6.2 6.1 0.5 Average amount of credit 5,987 14,993 1,942 1,840 3,683 1,653 3,385 607 Annex 5, Page 57 I able H.U4 Urban Rural Extreme Exlreme Access to Services/Housing All Poor Poor Non-poor All Poor Poor Non-poor Own house with title 77Y 42.1 4 . T5 41 321 35. Main source of water Pipes inside 42.5 9.9 16.5 50.4 6.7 0.9 2.0 13.4 Pipes outside 41.4 38.3 51.1 38.4 23.8 10.0 19 1 30.6 Public source 1.7 7.2 5.1 0.6 8.2 12.0 10.0 5.5 Public or private well 6.8 24.6 13.3 4.9 33.3 36.8 36.2 29.1 River/stream 0.4 3.3 1.1 0.2 20.5 33.9 24.9 14.0 Truck/oxcart 0.3 0.0 0.3 0.2 0.4 0.0 0.6 0.2 From another house 6.8 16.7 12.5 5.1 6.0 5.5 6.1 5.8 Other 0.1 0.0 0.1 0.1 1.2 1.0 1.1 1.3 Type of sanitary service Latrine/lavatory 57.2 75.5 77.9 50.9 67.6 53.2 60.9 77.3 Toilet discharges into sewers 29.5 3.3 8.1 36.0 0.4 0.1 0.0 0.9 Toilet discharges into cesspool 8.1 1.7 3.7 9.5 2.0 0.3 0.4 4.3 Toilet discharges into river/stream 0.1 0.0 0.0 0.1 0.0 0.0 0.0 0.0 There is none 5.1 19.5 10.4 3.5 30.0 46.4 38.7 17.5 Garbage Disposal Collected by truck 53.7 21.3 32.9 60.0 1.4 0.3 0.2 3.2 Burned 32.7 60.3 50.0 27.4 65.3 53.0 61.9 70.2 Buried 2.5 3.6 4.2 2.0 4.6 3.0 3.6 5.9 Made into fertilizer 0.2 0.0 0.2 0.2 1.2 0.9 1.1 1.2 Dumped into river/stream 6.9 14.9 11.5 5.6 27.2 42.9 33.0 18.7 Authorized dump 4.0 0.0 1.2 4.9 0.4 0.0 0.1 0.8 Type of lighting Electric 90.7 52.5 72.9 96.0 40.0 15.5 28.1 57.3 Generator 0.1 0.0 0.2 0.0 0.7 0.2 0.8 0.5 Kerosene/gas 8.0 40.4 23.3 3.4 56.4 78.5 67.8 39.8 Other 1.1 6.5 3.1 0.5 2.5 4.7 2.8 2.0 None 0.2 0.6 0.5 0.1 0.5 1.1 0.5 0.4 Mean paymet for electricity 118.5 40.9 49.0 133.2 61.8 31.3 39.2 77.1 Fuel for cooking Firewood 47.1 98. 1 87.1 34.8 92.2 98.7 97.9 83.7 Butane/Propane 47.1 1.3 11.0 58.1 6.2 0.1 1.2 13.5 Kerosene 2.4 0.7 1.1 2.7 1.2 1.1 0.7 1.9 Coal 2.0 0.0 0.4 2.5 0.3 0.0 0.2 0.3 Electricity 1.5 0.0 0.4 1 9 0.2 0 0 0.0 0.5 Households where at least one member has public hlth insurance 28.7 9.5 15.8 32.7 10.0 2.3 6.5 15.0 Households where at least one member has private hlth insurance 3.2 0.5 0.7 4.0 0.5 0.6 0.6 0.4 Distance to health post/center (kms) 0.8 1.1 0.9 0.8 4.4 6.1 5.0 3.2 Distance to elementary school (kms) 0.4 0.6 0.4 0.4 1.5 2.1 1.7 1.2 Minutes to health post/center 15.8 21.5 19.0 14.9 59.5 80.2 68.9 45.9 Minutes to elementary school 8.9 11.4 9.7 8.6 25.6 30.2 27.8 22.3 Principal access road is paved 31.5 11.4 15.0 36.5 10.5 6.1 8.9 12.8 Households requesting loans last year (%) 22.5 7.6 14.7 24.8 14.4 8.8 9.8 21.0 Average amount of loan 7,726 1,727 1,827 8,684 9,664 2,633 2,410 14,438 Households buying on credit last year (%) 10.8 4.5 8.9 11.4 6.6 3.9 4.9 9.1 Average amount of credit 7,788 867 611 9,558 2,073 1,692 1,623 2,459 Annex 5, Page 58 H.05 - Nicaragua 1993 Household (%) with inadequate walls, floor, ceiling, housing, overcrowding All households Rural households Only Households (%) with Households (%) with Inadequate Inadequate Over- Walls' Floor2 Ceiling3 Housing4 crowding5 Ceiling6 Housing7 All 36.1 43.7 6.8 55.5 23.9 3.3 81.2 Extreme Poor 54.8 82.0 16.2 88.3 54.0 4.0 91.1 Poor 51.1 69.5 12.7 79.6 41.3 3.8 87.6 Non-poor 25.1 24.9 2.5 37.9 11.2 2.0 66.8 Urban 25.3 25.4 1.5 38.9 16.3 Extreme Poor 41.8 68.7 3.0 77.1 52.3 Poor 39.1 50.8 3.0 65.1 35.6 Non-poor 20.7 17.0 1.0 30.2 9.9 Rural 52.8 71.9 15.0 81.2 35.7 Extreme Poor 58.1 85.3 19.5 91.1 54.5 Poor 57.8 79.7 18.0 87.6 44.5 Non-poor 41.6 54.5 8.4 66.8 16.0 Quintiles Poorest 54.4 81.0 15.9 87.7 53.7 4.0 91.3 11 52.9 67.8 13.1 78.9 38.4 4.1 86.4 III 41.1 48.3 5.0 63.4 25.4 2.3 79.7 IV 29.6 29.3 3.2 44.1 12.6 3.3 67.8 Richest 16.9 15.7 1.9 26.3 6.0 0.5 55.1 Region Managua 29.3 29.1 1.9 41.4 20.7 Pacific- Urban 19.4 23.8 1.3 31.7 19.7 Pacific- Rural 47.3 71.4 12.4 78.7 40.8 2.2 78.4 Central - Urban 10.9 37.0 0.7 39.9 11.5 Central - Rural 51.0 78.2 16.5 85.0 32.0 3.7 85.5 Atlantic- Urban 73.6 8.8 0.8 76.3 8.1 Atlantic- Rural 88.4 67.7 30.5 91.7 33.1 5.8 91.5 'Codes 5,6,8,10 and 11 in q. 2 2Codes 1 and 7 in q. 3 3Codes 4,5 and 7 in q. 4 41f inadequte walls or floors or ceiling (defined in 3) 5more than 3 persons per room 6Codes 4 and 7 in q. 4 71f inadequte walls or floors or ceiling (defined in 6) Annex 5, Page 59 H.06 - Nicaragua 1998 Household (%) with inadequate walls, floor, ceiling housing, overcrowding All households Rural households Only Households (%) with Households (%) with Inadequate Inadequate Over- Walls' Floor2 Ceiling3 Housing4 crowding5 Ceiling6 Housing7 All 33.5 48.8 5.6 56.2 38.8 3.0 75.5 Extreme Poor 58.8 77.7 16.3 85.8 73.5 5.8 87.6 Poor 49.3 73.7 10.7 80.1 60.7 4.5 83.9 Non-poor 23.4 32.7 2.4 41.1 24.7 0.9 63.5 Urban 25.2 33.2 3.3 41.5 34.3 Extreme Poor 49.7 68.8 13.3 80.0 72.1 Poor 43.3 65.6 7.3 72.7 65.3 Non-poor 19.7 23.2 2.1 32.0 24.8 Rural 44.2 68.9 8.6 75.5 44.7 Extreme Poor 61.7 80.6 17.3 87.6 73.9 Poor 52.3 77.8 12.4 83.9 58.3 Non-poor 32.3 55.9 3.0 63.5 24.6 Quintiles Poorest 58.7 77.8 15.9 85.7 72.8 5.5 88.1 11 46.8 74.8 8.2 81.0 57.6 4.2 84.8 III 35.2 56.6 4.3 62.4 44.2 1.1 69.6 IV 29.2 38.9 2.7 48.5 30.3 1.8 68.1 Richest 13.6 18.5 1.6 26.5 11.4 0.5 51.1 Region Managua 25.0 33.0 2.7 40.8 34.4 Pacific- Urban 21.1 35.7 5.0 38.1 35.1 Pacific - Rural 45.5 72.9 9.6 73.2 52.3 1.9 73.2 Central - Urban 15.4 38.2 1.5 40.7 32.0 Central - Rural 36.2 76.9 5.4 78.2 41.2 3.8 78.2 Atlantic- Urban 59.1 20.4 2.0 66.1 30.3 Atlantic - Rural 97.0 48.0 26.8 96.8 51.8 6.0 96.7 'Codes 5,6,9 and 10 in q. 4 2Codes 5 and 6 in q. 5 3Codes 4,5 and 6 in q. 6 41f inadequte walls or floors or ceiling (defined in 3) 5more than 3 persons per room 6Codes 5 and 6 in q. 6 71f inadequte walls or floors or ceiling (defined in 6) Annex 5. Page 60 H.07 - Nicaragua 1998 Main Source of Water by Poverty Group Main Source of Water All Extreme Poor Poor Non-poor Pipes inside 51.0 14.4 27.9 68.0 Pipes outside 10.6 9.0 11.0 10.3 Public standpipe 3.9 7.3 5.9 2.4 Truck 0.7 0.2 0.5 0.8 Public or Private well 19.4 33.8 27.2 13.7 River, etc. 7.7 20.7 15.8 1.8 Spring 4.2 12.2 8.2 1.3 Other 2.5 2.4 3.6 1.7 H.08 - Nicaragua 1993 Main Source of Water by RurallUrban Area Urban Rural Extreme Extreme Main Source of Water All Poor Poor Non-poor All Poor Poor Non-poor Pipes inside 74.0 42.6 58.2 79.1 15.8 7.2 11.2 26.1 Pipes outside 11.3 21.0 15.7 9.8 9.6 6.0 8.4 12.2 Public standpipe 1.9 3.4 3.5 1.4 6.9 8.2 7.2 6.2 Truck 0.4 0.0 0.5 0.4 1.1 0.3 0.4 2.5 Public or Private well 8.8 16.9 13.0 7.4 35.8 38.0 35.1 37.3 River, etc. 0.8 9.4 2.7 0.2 18.3 23.5 23.0 7.7 Spring 0.3 2.8 1.2 0.1 10.2 14.6 12.1 6.0 Other 2.5 3.8 5.2 1.6 2.5 2.1 2.6 2.1 H.09 - Nicaragua 1993 Main Source of Water by Quintile Main Source of Water All Poorest | II II IV Richest Pipes inside 51.0 15.1 30.3 30.3 62.1 79.0 Pipes outside 10.6 8.9 11.6 14.7 12.3 6.8 Public standpipe 3.9 7.2 5.1 4.5 3.2 1.4 Truck 0.7 0.3 0.3 0.8 1.4 0.4 Public or Private well 19.4 33.4 27.6 18.2 16.1 9.8 River, etc. 7.7 20.6 14.5 7.8 1.9 0.7 Spring 4.2 11.9 6.7 3.9 1.8 0.6 Other 2.5 2.7 3.9 3.8 1.3 1.5 H.10 - Nicaragua 1993 Main Source of Water by Zone Pacific- Pacific - Central - Central - Atlantic- Main Source of Water All Managua Urban Rural Urban Rural Urban Atlantic - Rura. Pipes inside 47.8 22.9 5.9 68.7 8.6 73.6 77.2 23.9 Pipes outside 10.0 9.2 4.0 9.0 5.0 14.9 10.1 12.2 Public standpipe 4.4 3.5 4.9 8.8 5.9 2.0 0.9 8.0 Truck 0.8 0.0 0.0 2.2 0.0 0.6 0.0 3.4 Public or Private well 20.5 59.9 42.9 6.2 36.0 3.0 6.7 42.1 River, etc. 8.7 0.9 35.0 1.8 26.3 1.6 0.1 2.3 Spnng 4.5 0.0 6.5 0.1 16.7 1.2 0.3 2.5 Other 3.3 3.5 0.8 3.3 1.7 3.2 4.9 5.7 Note: used weights given by Cailos for Regional breakdown Annex 5, Page 61 H.11 - Nicaragua 1998 - Main Access Road Condition since 1993 Since 1993, main access road has: Did not Improved Worsened Same live here All 16.8 10.5 60T.T 12.7 Extreme Poor 13.5 11.0 63.6 12.0 Poor 14.5 11.1 61.4 13.0 Non-poor 18.2 10.1 59.2 12.5 Urban 17.7 8.4 60.7 13.3 Extreme Poor 16.6 11.2 59.7 12.5 Poor 15.2 8.6 61.5 14.7 Non-poor 18.4 8.3 60.4 12.9 Rural 15.6 13.2 59.3 11.9 Extreme Poor 12.5 10.9 64.9 11.8 Poor 14.2 12.4 61.4 12.1 Non-poor 17.7 14.4 56.4 11.6 Quintiles Poorest 13.3 11.5 63.7 11.5 11 16.2 11.9 58.7 13.2 III 17.4 8.2 62.2 12.2 IV 16.2 9.8 59.7 14.3 Richest 19.0 11.1 57.8 12.1 Region Managua 14.9 9.4 62.4 13.3 Pacific - Urban 19.1 9.6 60.2 11.1 Pacific - Rural 11.6 15.7 64.1 8.7 Central - Urban 19.0 8.5 58.1 14.4 Central - Rural 21.1 11.2 53.6 14.2 Atlantic- Urban 19.6 8.1 58.1 14.2 Atlantic - Rural 8.5 7.7 68.2 15.7 Annex 6, Page I Annex 6 - Significance Tests, Confidence Intervals, Sensitivity Analysis, Price Changes and Dominance Conditions By Florencia T. Castro-Leal and Carlos Sobrado A. 1993 AND 1998 SIGNIFICANCE TESTS 1. Poverty comparisons for overall and extreme poverty nationally between 1993 and 1998 indicate that the reductions are statistically significant at a level of 5% or less (Table A6.1). If more stringent criteria (significance level of 1%) are applied, then the differences are not significant. Consequently, and although relatively modest, we can assert that the reductions in overall and extreme poverty nationally level are statistically significant. Table A6.1 Significance Tests for Poverty Indicators, 1993-98 Head count Sample size Extreme Poverty Overall Poverty 1993 19.4% 50.3% 4,201 1998 17.3% 47.8% 4,040 Difference 2.1 % 2.5% Standard deviation * 0.0110 0.0085 t value * 2.27 2.46 Significance level 2.46% 1.54% M. Ravallion. Poverty Comparisons: A Guide to Concepts and Methods LSMS No. 88 p. 49 (9) 2. For urban and rural areas, poverty comparisons for overall and extreme poverty between 1993 and 1998 indicate that the changes are not statistically significant for urban areas but for rural areas they are significant at a level of 1% level or less (Tables A6.2 and A6.3). These results suggest that we cannot state that poverty changed in urban areas. However, reductions in poverty in rural areas are indeed considerable and significant, even when applying the most stringent criteria. Table A6.2 Significance Tests for Poverty Indicators - URBAN, 1993-98 Head count Sample size Extreme Poverty General Poverty 1993 7.3% 31.9% 2,397 1998 7.6% 30.5% 2,187 Difference -0.3% 1.4% Standard deviation * 0.0137 0.0078 t value * 1.02 -0.39 Significance level >5% >5% M. Ravallion Poverty comparisons: A Guide to Concepts and Methods LSMS No. 88 p. 49 (9) Table A6.3 Significance Tests for Poverty Indicators - RURAL, 1993-98 Head count Sample size Extreme Poverty General Poverty 1993 36.3% 76.1% 1,804 1998 28.9% 68.5% 1,853 Difference 7.4% 7.6%°/ Standard deviation * 0.0148 0.0155 t value * 5.13 4.77 Significance level <1% < 1% M. Ravallion Poverty comparisons: A Guide to Concepts and Methods LSMS No. 88 p.49 (9) Annex 6, Page 2 B. 1993 AND 1998 CONFIDENCE INTERVALS FOR HEAD-COUNT RATIOS 3. Poverty results in Nicaragua are based on a representative sample of households, therefore, they are an approximation of the true values with a margin of error. Taking this into account, the rate of poverty can be expressed within a range of minimum and maximum values with a confidence level of 95%. At this confidence level, the rates of poverty in 1998 and 1993 with their confidence intervals are shown in Table A6.4 and Figure A6. 1: Table A6.4 Nicaragua: Poverty Confidence Intervals 5% Year Point estimate 95% confidence interval 1993 50.3 48.8 a 51.8 1998 47.9 46.3 a 49.3 Source: LSMS98 and LSMS93 following M. Ravallion Poverty comparisons: A Guide to Concepts and Methods LSMS No. 88 p.49 (10) Figure A6.1 Nicaragua 1993-98 Powrty rates - 95% Confidence Intervals 53% 52%. 51.8 0 51%.. 50%~~~ 49 50% 48.81 ._ 48% 47.9 . 47% a 46% 46.3 45% _ 1993 1998 4. Similarly, the incidence of extreme poverty in 1998 and 1993 with its confidence intervals is in Table A6.5 and Figure A6.2: Table A6.5 Nicaragua: Extreme Poverty -Confidence Intervals 5% Year Point estimate 95% confidence interval 1993 19.4 18.2 a 20.6 1998 17.3 16.1 a 18.5 Source: LSMS98 and LSMS93 following M. Ravallion Poverty comparisons: A Guide to Concepts and Methods LSMS No. 88 p.49 (10) Annex 6, Page 3 Figure A6.2 Nicaragua 1993-98 Extreme Powrty rates - 95% Confidence Intervals 25% 24% @ 23% ; 22% 21% 20.6 19% 20% 19.41 *E 18% 18.2 18.5 17% 17.3 16% 16.1 15% j_ 1993 1998 5. The rate of overall and extreme poverty at a confidence level of 1 % with its confidence intervals is shown in Table A6.6: Table A6.6 Nicaragua Confidence Intervals Low * Point Estimate High* General 1993 1% 48.3% 50.3%l 52.3% General 1998 45.8% 47.8% 49.8% Extreme 93 17.8% 19.4% 21.0% Extreme 98 15.8% 17.3% 18.8% Source: LSMS98 and LSMS93 following M. Ravallion Poverty comparisons: -Al Guide to Concepts and Methods LSMS No. 88 p.49 (IO) 6. One must be careful when interpreting confidence intervals. The fact that confidence intervals cross does not imply that estimated values cannot be differentiated. It is a common mistake to think that for confidence intervals to be different they should not have a common area, but this assessment is false. Crossing confidence intervals indicates that two probabilities are crossing and the fact that both occur at the same time is not the average of the two of them. C. SENSITIVITY ANALYSIS - POVERTY CHANGES DUE TO INCREASES AND DECREASES OF POVERTY LINES, 1993 AND 1998 7. To test the robustness of poverty rates calculated for both 1993 and 1998 Nicaragua LSMS, we carried out a sensitivity analysis by increasing and decreasing the poverty lines (extreme and overall) for both periods and calculating the new poverty rates. The poverty lines' absolute values were increased and decreased by 5% and 10% for both years. The resulting values are reported in Table A6.7. To estimate the new poverty rates, the new poverty lines were compared to the old consumption aggregates; the rate changes are reported in Table A6.8. There are two ways to look at the resulting poverty rates: the difference in rates between LSMS 93 and LSMS 98 and the differences in 1998 alone. 8. Differences in poverty rates between LSMS 93 and LSMS 98. The difference in poverty rates between 93 and 98 does not change significantly with the new poverty lines. For example, in 1998 the extreme poverty rate in rural Nicaragua was 7.4 percentage points lower than in 1993, Annex 6, Page 4 and with the new poverty lines the percentage decrease ranges from 7.1 to 7.6 percentage points. Under any scenario -for all Nicaragua or for urban or rural groups- the signs observed with the original calculations do not change. If the original calculations showed a decrease in poverty between LSMS 93 and LSMS 98, the new calculations also show a decrease. The consistency i:. the differences is a sign of robustness in the comparisons between the two periods. Any conclusion about the evolution of poverty between LSMS 93 and LSMS 98 will not be affected by changes in the poverty lines. 9. Differences in poverty rates in 1998 due to changes in poverty lines. Increasing or decreasing the extreme poverty line by 5% will change the extreme poverty rate by 1.7 percentage points (-2.1 percent). The same 5% increase or decrease in the overall poverty line will change the overall poverty rate by 2.8 percentage points (-2.9%). A 10 % increase or decrease of the poverty lines has an effect of similar magnitude. Changing the extreme poverty line by 1% will change the national extreme poverty rate by 0.37 of a percentage point, and changing the overall poverty line by 1% will change the overall poverty rate by 0.58 of a percentage point.' 10. The extreme poverty rates are less sensitive for urban households, where a 1% change in the line will change the extreme poverty rate by 0.23 of a percentage point compared with a change of 0.51 percentage points for rural households. On the other hand, overall poverty rates are more sensitive in urban households, where a 1% change in the line will change the overall poverty rate by 0.66 of a percentage point compared with a change of 0.45 of a percentage point for rural households. 11. In general, changing the extreme poverty line by 5% will change the extreme poverty rate by 1.9 percentage points, and changing it by 10% will change it by 3.7 percentage points. Changing the overall poverty line by 5% will change the overall poverty rate by 2.8 percentage points and changing it by 10% will change it by 5.9 percentage points. Urban households are more sensitive to changes in the overall poverty line and rural households are more sensitive to changes in the extreme poverty line. The magnitude of all the changes is considered reasonable and does not show an unexpected concentration of households with aggregate consumption values above or below the poverty lines. Table A6.7 Nicaragua 93-98 LSMS Sensitivity analysis: Poverty lines values Poverty Original Increase Decrease Lines * Year 5% 10% 5% 10% Nicaragua Overall 1998 4,259 4,472 4,685 4,046 3,833 Extreme 2,246 2,358 2,516 2,134 2,021 Overall 1993 2,574 2,702 2,831 2,445 2,316 Extreme 1,2161 1,2771 1,362 ,155 1,094 Source: LSMS98 and LSMS93. * The 1998 Values are in May 1998 cordobas, and the 1993 are in February., June 1993 cordobas (yearly per capita) This applies to increases or decreases in the poverty lines between 5 and 10 percent. Annex 6, Page 5 Table A6.8 Nicaragua 93-98 LSMS Sensitivity analysis: Poverty rates Poverty rates Rates change (% points) Original Increase Decrease Increase Decrease 5% 10% 5% 10% 5% 10% 5% 10% Nicaragua Extreme 1993 19.4% 21.1% 22.7% 17.6% 15.8% 1.7% 3.3% -1.8% -3.6% Poverty 1998 17.3% 19.0% 21.2% 15.3% 13.9% 1.7% 3.8% -2.1% -3.5% 98-93 -2.1% -2.0% -1.5% -2.3% -2.0% 0.0% 0.5%1 -0.2% 0.1% Overall 1993 50.3% 52.7% 54.8% 48.1% 45.7% 2.3% 4.4% -2.2% -4.6% Poverty 1998 47.9% 50.7% 53.5% 45.0% 41.8% 2.8% 5.7% -2.9% -6.1 % 98-93 -2.5% -2.0% -1.3% -3.1% -3.9% 0.5% 1.2% -0.6% -1.4% Urban Extreme 1993 7.3% 8.3% 9.3% 6.2% 5.1% 1.0% 2.0% -1.2% -2.2% Poverty 1998 7.6% 8.8% 10.0% 6.8% 6.0% 1.1% 2.4% -0.9% -1.6% 98-93 0.3% 0.4% 0.8% 0.6% 0.9% ° 0.1% 0.4% 0.3% 0.6% Overall 1993 31.9% 34.6% 36.9% 30.0% 27.5% 2.8%° 5.1% -1.9% -4.4% Poverty 1998 30.5% 34.0% 36.7% 27.5% 24.5% 3.5% 6.2% -3.0% -6.0% 98-93 -l.4°%6 -0.7% -0.3% -2.5% -3.0% 0.7% 1.1% -1.1% -1.6% Rural Extreme 1993 36.3% 38.9% 41.5% 38.9% 30.9%i 2.6% 5.2% 2.6% -5.4% Poverty 1998 28.9% 31.3% 34.4% 31.3% 23.2% 2.4% 5.5% 2.4% -5.7% 98-93 -7.4% -7.6% -7.1% -7.6% -7.7% -0.2% 0.3% -0.2% -0.3% Overall 1993 76.1% 77.8% 79.7% 73.5% 71.2% 1.7% 3.5% -2.6% -4.9% Poverty 1998 685% 70.6% 73.6% 65.9% 62.4% 2.0% 5.0% -2.7% -6.1% . 98-93 -7.6% -7.3% -6.1 % -7.7 -8.8 0.3% 1.5% 0.0% -1.2% Source: LSMS98 and LSMS93 D. PRICE CHANGES IN NICARAGUA BETWEEN 1993 AND 1998 12. The following paragraphs detail the analysis of price increases for food and beverages in Nicaragua between February-June 1993 and April-August 1998. 13. Data sources. In order to assess whether price changes for food items have been disproportionate, one needs to compare the differences in prices for food items in the periods detailed above with respect to the difference in prices of non-food items and/or services during the same period. 14. The only source of information for both periods that includes food and non-food items is that generated by the National Institute of Statistics and Census (INEC) and published by the Central Bank of Nicaragua (BCN) (accessible on the Internet http://www.bcn.gob.ni/) to calculate the Consumer Price Index (CPI). The 1993 and 1998 LSMS also contain appropriate information but only for food items. 15. At a certain point, we considered using the LSMS information for food items and the CPI information for non-food items thinking that the LSMS might be a more reliable source of information. Unfortunately, the combination of these two data sources is always a source of immeasurable variation due to differences in collection and processing methods and in informants. Additionally, if we use the CPI information for non-food items (there is no other Annex 6, Page 6 option), we are implicitly accepting the quality of this data and it would be difficult to substantiate that the CPI information for food items is unreliable. 16. Methodology: With the CPI information, products were divided into food and non-food items. The CPI was recreated for the periods detailed above for these two groups, and then the change in the indexes for food items was contrasted with the change in the indexes for non-food items. The four indexes use the same relative weights per article as that used by the BCN in the construction of the CPI. 17. Results. The four indexes and their comparison are in Table A6.9. Prices for food items increased 20 % more than prices for non-food items (76% vs. 63%). Tabla A6.9 Change in prices for food and non-food items' February to April to % change % change over non- June 93 August 98 food items Food 41.20 72.66 76.3% 120.4% Non-food 48.64 79.48 63.4% 100.0% 'Using the same values as the CPI, but divided into food and non-food items 18. Potential Effects on Consumption: A price increase for food items that is proportionall) higher than the price increase for non-food items can have two potential effects on consumption. First, there is a propensity to decrease the share of total consumption dedicated to food. Second, there is also a propensity to substitute the consumption of relatively more expensive food items for relatively inexpensive ones. Consequently, the share of food consumption dedicated to relatively inexpensive food items (such as those rich in carbohydrates) increases, while the share of food consumption dedicated to relatively more expensive food items (such as those rich in protein) decreases, with an overall effect of impoverishing the nutritional content of the diet. E. FIRST-ORDER DOMINANCE CONDITIONS IN 1998 AND 19932 19. The First-Order Dominance Conditions can be interpreted as a test that answers the question: Are the poverty results and comparisons robust in relation to the' choice of poverty lines and measures? This sort of test is performed by plotting a pair of cumulative distribution functions and observing if one distribution lies entirely above (dominates) the other. Then the condition is said to hold. In such an event, one can conclude that all well-behaved poverty measures and all possible poverty lines will show an unambiguous decrease in aggregate poverty from one distribution to the other. 20. 1993 consumption (common) and 1998 consumption (common):3 The 1993 common consumption aggregate line lies above the 1998 common consumption aggregate line for all poor households, with the exception of the first data point on the graph. Thus, poverty in 1993 is lower than in 1998, although not by much. This first result generally implies that poverty comparisons between 1993 and 1998 using the LSMS93 and LSMS98 confirm that poverty declined. In addition, the poverty line does not lie in a particularly steep part of the distribution functions (see charts A6.3 and A6.4). This second result implies that moving the poverty line up or down would not change significantly the incidence of poverty or percentage of poor estimated in either 1993 or 1998. Consequently, poverty estimates for either 1993 or 1998 with selected poverty lines are statistically robust. 2 See Martin Ravallion POVERTY COMPARISONS: A Guide to Concepts and Methods. LSMS Study Working Paper No. 88, p57-62 3See Annexes I and 2 for an explanation of consumption aggregates and poverty lines Annex 6, Page 7 F Figure A6.3 - Ist Order Stocastic Dominance: common consumption aggregates 9,000 u 8,000 g 7,000 c 6,000 5,000 1993 e 4,000 -1998 I e 3,000 c 2,000 1,000 0 20 40 60 80 100 % cumulative population Figure A6.4 Ist Order Stocastic Dominance: common consumption aggregates - poor household only 2,000 ~69 1,600 E 1,200 199 800 1 1998 U 400 /. 20 40 60 80 100 % cumulative population Annex 6, Page 8 21. Urban and rural consumption, 1993: The rural consumption line lies above the urban consumption line for all poor households. Thus, urban consumption, and poverty as a consequence, is strictly higher than rural consumption (poverty) in 1993 (see charts A6.5 and A6 .6). Figure A6.5 - Ist Order Stocastic Dominance: comprehensive consumption aggregate, 1993 30,000 e 25,000 2000 e 15,000 =_ b ! 10,000 5,000 20 40 60 80 100 % cumulative population Figure A6.6 - Ist Order Stocastic Dominance: comprehensive consumption aggregate - poor households only, 1993 3,000 u 2,5000 c2,000 ; ran E ~~~mli rbano . 1,500 - - ural 0 Ur~ -1,000 500 0 20 40 60 80 100 % cumulative population Annex 6, Page 9 22. Urban and rural consumption, 1998: The rural consumption line lies above the urban consumption line for all poor households. Thus, urban consumption, and poverty in consequence, is strictly higher than rural consumption (poverty) in 1998 (see charts A6.7 and A6.8). Figure A6.7 - Ist Order Stocastic Dominance: comprehensive consumption aggregate, 1998 30,000- . Y-25,000 ~j20,000 E: -,oo -- . -. : - ., /Urbano ] < ~15,000- 3-- Rua 0 Rujral 1~ 10,000 > 5,000 0 20 40 60 80 100 % cumulative population Figure A6.8 - Ist Order Stocastic Dominance: comprehensive consumption aggregate - poor households only, 1998 5,000 4,500 u. 4,000 3,500 E 3,000 E 2,500 °°°- j _Urbano -Rural 2,000 7 ;- 1. 15004 500 0 20 40 60 80 100 % cumulative population 23. Other comparisons are not feasible because the test requires comparable consumption aggregates and a single measure for the poverty line. For this reason dominance can be tested for common consumption aggregates but not for comprehensive consumption aggregates (see Annexes 1 and 2 for an explanation of consumption aggregates and poverty lines). Annex 7, Page 1 Annex 7 - Inequality Comparisons By Florencia T Castro-Leal and Carlos Sobrado 1. Because of comparability issues,' it is not possible to examine the trends in income distribution between 1993 and 1998. For 1998, however, consumption inequality in Nicaragua appears to be about equal to the Latin American median (Table A7.1). The poorest quintile in Nicaragua consumes 5.3 percent of the total, while the richest quintile consumes 51.3 percent, rendering consumption in Nicaragua less unequal than in Brazil, Mexico and Panama, for example, but more unequal than Bolivia and Jamaica. The consumption ratio of the richest to the poorest is 9.6 This inequality level is already high by international standards. Table A7.1 - International Comparisons of Inequality using the Gini Index Consumption Gini Year Latin America: median 45 Brazil 55 1974 Mexico 50 1992 Panama 49 1997 Nicaragua 45 1998 Peru 45 1994 Colombia 43 1972 Ecuador 43 1998 Bolivia 42 1990 Jamaica 38 1993 Africa: median (various countries/years) 48 Middle East/North Africa: median (various countries/years) 39 Eastern Europe: median (various countries/years) 36 Asia: median (various countries/years) 32 Western Europe: median (various countries/years) 32 Source: Nicaragua 1998 LSMS and World Bank Panama Poverty Assessment. 2. International evidence suggests that highly unequal distribution of wealth and access to basic services can perpetuate slow and unequal growth and can translate into higher poverty. This conclusion is robust to different types of frameworks, such as credit constraint, asymmetric information, and political economy models?2 Cross-country evidence also indicates that the channel from high inequality to slower growth and higher poverty is through lower investment in physical and human capital. Fertility rates in Nicaragua among the poorest are more than three times than among the richest, which raises solid concerns about potential further increases in inequality and thus slower progress in poverty reduction ' See paragraph no. 3 for an explanation. 2 Bruno, Ravallion and Squire (1995) Annex 7, Page 2 3. The aggregate consumption measure obtained from the 1998 LSMS is more comprehensive than that obtained from the 1993 LSMS. According to Lanjouw & Lanjouw (1997), more comprehensive aggregates are better measures of welfare. The authors explain that after corrections to allow comparability of data sets from different surveys and definitions of consumption have been applied, only robust comparisons of poverty are possible. As the consumption aggregate expands, however, the effect on measures which proxy for inequality is ambiguous. Thus, it follows that the consumption Gini Index estimated for Nicaragua using the 1998 LSMS is not statistically comparable to the Gini Index reported by the 1995 World Bank Nicaragua Poverty Assessment (Table A7.2) and neither are the Lorenz curves (Figures A7. 1 and A7.2). Table A7.2 GINI Nicaragua 1993-1998 LSMS GINI Year Aggregate Source Coefficient 1993 Comprehensive Consumption LSMS 1993 and World Bank 50.4 Nicaragua Poverty Assessment (1995) 1993 Common Consumption LSMS1993 39. 1998 Common Consumption LSMS 1998 39.4 1998 Comprehensive Consumption LSMS 1998 45.2 Annex 7, Page 3 Figure A7.1 Lorenz Curves Nicaragua 98-93 Comprehensive Consumption (non-common) 0D 1.0 . ._. ..-.-...... 'i a0.9 _ ; 0.8 _ _ - _ _ _- 0.7 - ----- i Ch l 0.5 ---------- C o n s u m p t o n _ 1 - 4 5 0 d 0.4 __ __ _ _ _ ___ _ _ co 75 0.3 Comprehensive E 0.2 ___ - ------.-Consumption 1993 L, 0.1 _ -- ---- Comprehensive i < 0.0 Consumption 1998 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 -- Acumulated population % WARNING: These consumption aggregates are not comparable Figure A7.2 Lorenz Curves Nicaragua 98-93 Common Consumption Aggregates 0 1.0 Y' 0.9 _ _! tm 0.8 ___ _ ___ 0) 0.7 -____ < 0.6 __ . _- aaZ 0.5/ 45o _ 0 ___3- - _ . Common Consumption : 0.2 _ _ > .z 1993 u > Q.1 - / .- - - Common Consumption < 0.0 ,,^~_ 1998 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Acumulated g0opulation % Waming: Although these consumption agg ga tes are comparable, neither the Lorenz curves nor the Gini accurately reflect inequality in Nicaragua Annex 7, Page 4 4. POVCAL Projections (Table A7.3) applying a 1.2 percent growth in mean household consumption3 to LSMS93 show almost no difference to actual estimated results using the LSMS98 - indicating that inequality did not worsen between 1993 and 1998. We had to use this indirect method to estimate changes in inequality since the LSMS93 and LSMS98 are not comparable for inequalitv comparisons. Table A7.3 NATIONAL: LSMS1993 POVCAL Projections to 1998 with 1.2% Annual Growth o-fi Mean Household Consumption Projected Levels of Overall Poverty given a 1.2% Annual Growth in Mean Household Consumption _ Povert Consumpti 1993 1994 1995 1996 1997 1998- Projected 1998- Actual y on Elast Projected Change | Actual Change Index l H -0.915 50.3 49.7 49.2 48.7 48.1 47.6 -2.7 47.9 -2.4 PG -1.296 21.8 21.5 21.1 20.8 20.5 20.2 -1.6 18.3 -3.5 FGT2 -1.594 12.1 11.9 11.6 11.4 11.2 11.0 -1.1 9.3 -2.8 Projected Levels of Extreme Poverty given a 1.2% Ainnual Growth in Mean Household Consumption Povert IConsumpti 1993 1994 1995 1996 1997 1998- Projected 1998- 1998- on Elast Projected Change Actual Actual H -1.603 19.4 19.0 18.7 18.3 18.0 17.6 -1.8 17.3 -2. PG -2.312 5.9 5.7 5.6 5.4 5.3 5.1 -0.8 4.8 -1.1 FGT2 | 3.001 2.6 2.5 2.5 2.4 2.3 2.3 -0.3 | 2.0 -0.6 3 Nicaragua's GDP experienced an average real growth rate of 4.3 percent per year; however, fast population growth brought the average real growth rate of GDP per capita to only 1.2 percent per year. Annex 8, Page I Annex 8 - Comparing poverty for the young and the old by Carlos Sobrado SOURCE OF INFORMATION I. To compare the state of relative poverty for the different population age groups (0-5, 0-18, and over 60), six tables were constructed. The first three tables (35, 36 and 37) are the extreme poor and overall poor percentages for each group and the deviation of these percentages from the national averages (in percentage points and percentage terms). The second three tables (38, 39 and 40) are numbers of the extreme poor and overall poor numbers for each group and their population and poverty shares. Each set of tables starts with figures for the entire country, followed by a breakdown for urban and rural households. 61 YEARS AND OLDER GROUP 2. In all the tables, this group shows a lower rate of extreme and overall poverty than the national average. Nationally, the rate of extreme poverty for the "elderly" group is 27% lower (five percentage points higher) and the rate for overall poverty is 20% lower (ten percentage points higher) than the national average. Another way to look at this is to note that the "elderly" group represents 5. 1% of the Nicaraguan population, but only 3.5% (4.1 %) of the extreme (overall) poor. The lower rates for the urban and rural level show the same pattern as the national results. 3. There are three main reasons for a lower poverty rate in the "elderly" group. First, there is the effect of capital accumulation over time. Second, this group tends to live in smaller households', and third, persons living in poverty tend to have lower life expectancies. 0 TO 5 YEAR OLD GROUP 4. The younger group is always associated with a higher rate of poverty (extreme and overall). Nationally, this group has a 33% higher extreme poverty rate (six percentage points higher), and a 25% higher overall poverty rate (12 percentage points higher). Children from zero to five years old represent 13.7% of the population, but they account for 18.2% (17. 1 %) of the extreme (overall) poor. The higher poverty rate is more severe in urban households2 when comparing the group's averages, but in absolute terms rural children have a much higher probability of being poor-. ' In our case the household size for persons 61 years and older was 5.59 members and for younger persons was 6.82 members per household. 2 The only statistic with higher incidence increases for Rural kids is the percentage point increase in extreme poverty where Urban kids show a three percentage points increase and Rural kids have a six percentage points increase compared to each group average. The same increases represented as a proportion to the extreme poverty rates in each region are higher for the Urban kids (37%) than the Rural kids (21%). 3 This is a characteristic of Rural households. Annex 8, Page 2 0 TO 17 YEAR OLD GROUP 5. The zero to seventeen-year old group shows an increase in the poverty rate increase very similar to that of the zero to five-year old group, but to a lesser degree. Representing 50. 1% of the population, this group includes 61.1% (57.3%) of the extreme (overall) poor. The same pattems are also observed for the urban and rural sub-groups. A8.1 - Nicaragua 98 Poverty for Selected Groups With Respect to the "Total" Groups Extreme Poor Overall poor % points diference % diference Extreme Non-Ext. Overall Years of Age Poor Poor poor Non-Poor Extreme Overall Extreme Overall 0-5 23.0 77.0%-= 59.7% 40.3% 6% 12%7 33%_= 25'=T 0-17 21.2% 78.8% 54.7% 45.3% 4% 7% 23% 14'/o 61 and More 12.7% 87.3% 38.1% 61.9% -5% -10% -27% -20%1/ Total 17.3% 82.7% 47.8% 52.2% 0% 0% 0% 0%1/o A8.2- Nicaragua 98 Urban Poverty for Selected Groups With Respect to the "Total" Groups Extreme Poor Overall poor % points diference % diference Extreme Non-Ext. Overall Years of Age Poor Poor poor Non-Poor Extreme Overall Extreme Overall 0-5 1 0.4%- 89o 42.2% 57.8% % 12% o 37% 38%o 0-17 9.8% 90.2% 36.5% 63.5% 2% 6% 29% 20%1/o 61 and More 5.9% 94.1% 23.4% 76.6% -2% -7% -22% -23% Total 7.6% 92.4% 30.5% 69.5% 0% 0% 0% 0% A8.3- Nicaragua 98 Rural Poverty for Selected Groups With Respect to the "Total" Groups Extreme Poor Overall poor % points diference % diference Extreme Non-Ext. Overall Years of Age Poor Poor poor Non-Poor Extreme Overall Extreme Overall 0-5 "To9% 65.1% 76-2%o 23.8% 6 8%1 11 1(/ 0-17 33.4% 66.6% 74.2% 25.8% 5% 6% 16% 8% 61 and More 21.7% 78.3% 57.4% 42.6% -7% -11%1 -25% -1 6% Total 28.9% 71.1%0 - 68.5% 31.5% UoT 0%| 0%| 0f A8.4 - Nicaragua 98 Poverty Number for Selected Groups l_________ l Population Poverty Share Years of Age Extre Poor Overall Poor Number Share Extreme Overall 10-5 | 152,05| 394.90U 661,22 13.-T7| - 18.2| 17.1| 0-17 510,591 1,319.80 2,411.40 50.1 61.1| 57 3 61 and 31,551 94,54 247,91 5.1 3.81 4.1 Total | 834,981 2,303.851 4,814.80| 100| 1001 100 Annex 8, Page 3 A8.5 - Nicaragua 98 Poverty Numbers for Urban Groups Population Poverty Share Year of Age Extreme Poor Overall Poor Numbers Share Extreme Overall 0-5 3372 135,627 321,012 12.3% 1 17 0-17 121,533 455,011 1,245,922 47.6% 60.8% 57.0% 61 and more 8,265 32,879 140,464 5.4% 4.1% 4.1% Total 199,799 797,677 2,616,703 100.0% 100.0% 100.0% A8.6 - Nicaragua 98 Poverty Numbers for Rural Groups Population Poverty Share Year of Age Extreme Poor Overall Poor Numbers Share Extreme Overall 0-5 118,683 259,273 340,203 5 18.7% 17.2% 0-17 389,058 864,799 1,165,480 53.0% 61.3% 57.4% 61 and more 23,293 61,670 107,447 4.9% 3.7% 4.1% Total 635,183 1,506,175 2,198,105 100.0% 100.0%0 100.0% A8.7 - Nicaragua 98 Poverty Numbers for Urban and Rural Groups l l I Population Poverty Share Year of Age Extreme Poor Overall Poor Numbers Share Extreme Overall Urban 199,799 797,677 2,616,73 54.3% 23.9% 34.6%h Rural 635,183| 1,506,175 2,198,105 45.7% 76.1% 65.4% Total 4, ,3, 4, 100.0% 100.0% Annex 9, Page I Annex 9 - A Short Guide to Poverty and Inequality Measures by Gabriel Demombynes A. POVERTY MEASURES 1. The three measures of poverty employed most frequently are those of the Foster-Greer- Thorbecke class (FGT). The FGT indices can be applied to either income or consumption. In the following paragraphs, the indices are discussed in the context of an application to consumption. 2. Consider M households, where the number of individuals in each household is given by mh, for Al a total of N = E mh individuals in the population. Each household has a per capita consumption hi=1 equal to y,, and P of these households have levels of consumption below the poverty line, z. An FGT poverty index is calculated as follows: FGT(a)= = I mhh (_Z__ 3. The choice of alpha (a) determines the specific FGT measure. Typical choices for alpha are 0, 1, or 2. The extent ofpoverty (headcount ratio) 4. The FGT index with alpha equal to zero has a very simple form: 1 " *Z-Yh' 1 1 FGT(0) = - Y mh * = - ) mh N h =1 ( . h=1 This is simply the percentage of persons living in households with per capita consumption below the poverty line. The depth ofpoverty (poverty gap) 5. If alpha equals 1, the measure takes the following form: FGT(l) = Mh * (Z IY N h=1 mhyz) This measure increases not only if the number of poor increases, but also if the poor become poorer. 6. The overall poverty gap is the average of the "individual poverty gaps" over the whole population. For households with levels of consumption below the poverty level, the individual poverty gap is the difference between the poverty line and each household's per capita income (or consumption), as a fraction of the poverty line. For households with incomes above the poverty level, this individual poverty gap is zero. The severity of poverty 7. If alpha is equal to 2, the measure has the following form: FGT(2) =- imh * (Yi ) Annex 9, Page 2 8. This is similar to the poverty gap, but the squared factor gives more weight to the poorest of the poor, i.e. those who have levels of consumption far below the poverty line. The severity of povertv is greater if there is more inequality among the poor. B. GRAPHIC DEPICTION OF POVERTY MEASURES 9. The FGT measures can be understood graphically with "TIP" curves, which show the cumulative distribution of consumption in the entire population. The following figure shows the headcount ratio (FGT for alpha=O). Maximum consumption 0 E Cl) u line (z) 0 0 0 0 0 Headcount ratio 1.0 Cumulative percentage of the population 10. The poverty gap (FGT with alpha equal to 1) is determined by the average "individual poverty gaps" between household consumption and the poverty line; graphically represented in the following figure: Annex 9, Page 3 Maximum consumption E n Poverty O line (z) 4-, 0 0 0 Headcount ratio 1.0 Cumulative percentage of the population Annex 9. Page 4 1 1. Two communities may have the same headcount ratios but different poverty gaps. In the following figure, Community 2 has a larger poverty gap, but the headcount ratio is the same for the two communities: Maximum Consumption 0 0. E Poverty ---- : line (z) 0 0) 0 0 headcount ratio 1.0 Cumulative percentage of the population - - - - COMMUNITY 1 COMMUNITY 2 12. It is slightly more difficult to depict poverty severity (FGT with alpha equal to 2) graphically. Consider two communities with identical headcount ratios and poverty gaps, as in the following figure. The average gap between the poverty line and consumption by the poor is equal in the two communities. There is an important difference, however. In Community 1, almost all the poor have the same level of consumption. But in Community 2, there are some poor who are extremely pooT and other poor who are the "well-off' poor, with consumption levels just below the poverty line. There is much inequality between the poor in Community 2. Since the poverty severity measure gives greater weight to the extreme poor, Community 2 has a level of poverty severity greater tha:n that of Community 1. Maximum Consumption 0. E o line (z) 0 -J~~~ 0)~~~ 0 headcount ratio 1.0 Cumulative percentage of the population - - - - COMMUNITY 1 COMMUNITY 2 Annex 9, Page 5 C. INEQUALITY MEASURES There are a number of inequality measures. Two groups of inequality measures commonly employed are those of the Generalized Entropy and Atkinson classes.Measures of tlte Generalized Entropy class. 13. Particular measures of the Generalized Entropy class are specified by the choice of parameter alpha. If alpha is not 0 or 1: I I [n Yi0 GE(a) = a2 -[n Lj ] If alpha is equal to 0, the measure has a special form and is equal to the mean log deviation: GE(O) =-I log n=1 Y, If alpha is equal to I, the measure has a special formn and is known as the Theil index: GE(1) =-E Y' log Yi nl 1y y If alpha is equal to 2, the measure is equal to half the coefficient of variation squared. 14. The value of a Generalized Entropy measure ranges from zero to infinity, where higher values represent greater levels of inequality. The alpha parameter determines the weight that the measure gives to differences in the different parts of the distribution of consumption. For low alpha values, the measure is more sensitive to changes in distribution among the poor. For high alpha values, the measure is more sensitive to changes in distribution among the well off. The most common values are 0, 1, and 2. If alpha is equal to 0, the measure is more sensitive to differences between consumptionlevels in the lower portion of the distribution. If alpha is equal to 1, the measure gives equal weight to inequality for the entire distribution. And if alpha is equal to 2, the measure gives more weight to differences in the higher portion of the distribution. Measures of thze Atkinson class 15. The measures of this class have the following formn: The epsilon parameter corresponds to "aversion to inequality." The higher epsilon is, the more sensitive the measure is to inequality. Epsilon ranges in value from 0 to infinity. Annex 9, Page 6 Any particular Atkinson measure has a range of values from 0 to 1. Zero represents no inequality and 1 corresponds to total inequality. There is a correspondence between Atkinson and Generalized Entropy measures. Specifically if alpha and epsilon are chosen so that alpha = I - epsilon, an Atkinson measure with parameter epsilon is equivalent to a Generalized Entropy measure with parameter alpha, in terms of the rankings they produce. Annex 1o, Page 1 Annex 10 -The Probability of Being Poor by Carlos Sobrado I. Poverty tends to be concentrated in groups with similar conditions and can be related to many characteristics within a household. For example, 78 percent of the Nicaraguan poor live in rural areas', and 76 percent of poor families have five or more members. These figures are useful in understanding the distribution of poverty but can be misleading because they do not consider the relationship among variables. In this case, rural households tend to have more members than urban households, which affects the way we associate poverty with these characteristics. 2. To relate household characteristics to the probability of being poor, we developed multivariate models using data from the Living Standards Measurement Survey 1998 (LSMS 98). By including all household characteristics in one single model, we estimated the effect of each variable independently on poverty. The resulting model allows us to see the relative role of factors in determining poverty and to assess the potential impact of policies on the probability of being poor. The multivariate models were limited by the variables available in LSMS 98 and do not include macroeconomic characteristics, social conditions such as crime, weather, or the dynamic impact of certain factors over time. Furthermore, the statistical analysis shows evidence of a relationship between the variables and poverty, but causality can run both -ways. We base our interpretations of these statistical relationships on theory. 3. A probabilistic model2 was developed using a dummy variable for poverty as the dependent variable (with a value of I for poor households and 0 for non-poor households) and the household characteristics as the independent variables. With the household as the base unit, we calculated regressions using a two-stage sampling design weighted according to population. This paper examines the results of a regression for the entire country and compares these national results with individual regressions for urban and rural households. DETERMINANTS OF POVERTY FOR ALL NICARAGUA 4. This sections analyzes the Probit model results for all Nicaragua. The selected variables, estimated marginal effects, and significance levels are reported in Table Al 0.1. Marginal effects were calculated for all selected variables. These effects can be interpreted as the expected change in the probability of being poor from a change of one in the corresponding household variable. For categorical variables the interpretation is the change in the probability of being poor if we change from the excluded category to the corresponding one; for example, for "geographical location: rural excluding Managua," the marginal effect is the expected probability difference for a household in a rural area compared to a household in Managua. The results of calculating marginal effects (df/dx) for the entire sample are presented in Table AIO. 1. The reported marginal effects apply only to the observed range of values on the independent variables. For example, the electricity variable can take values of 0 or 1, but the years of education can range from 0 to 17. To compare two or more marginal effect estimates it is important to know the possible values that can be assigned the parameters. A reference household illustrates the possible changes in probabilities when household conditions change.3 The changes in the probability of being poor were calculated for specific changes in selected variables. 'Nicaragua 1998 LSMS 2 See Appendix 1 for a detail technical explanation of the Probit model. 3 Annex 10, Page2 5. Compared to households in Managua, both rural and urban households in all other parts of the country have a much higher probability of being poor. Table A10.1 - Statistical Significance and Marginal Values for the Probability of Being I_oor Variables df/dx a Observed poverty rate (Headcount) 47.85 percent Geographical Location Urban excluding Managua 0.224*** Rural excluding Managua 0.260*** Number of household members by age group 0-5 years old 0.108*** 6-12 years old 0.071 *** 13-18 years old 0.057*** 19-24 years old 0.028* 25-64 years old 0.019Ns 65 years old or older 0.028Ns Maximum years of education of household head or companion -0.026*** Percentage of females in the household -0.151 ** Somebody speaks an Indigenous language in the household -0.220** Land Worked (excluding landless) 0. I to 2 ha. 0.1 13*** 2 to 5 ha. -0.OOONs 5 tol5 ha. -0.132** >15 ha. -0.209*** Labor Number of persons working last weekb over total household size -0.369*** Agriculture provides more than 25 percent of total income 0.161 Own farmn provides more than 25 percent of total income -0.137*** House amenities and size Water inside the house or on property -0.070** Electricity in the house (excluding generators) -0.158*** Telephone in the house -0.301 *** Rooms per Person 0.027*** Sum of bad housing characteristics (type, walls, floor, roof, roads) One 0.065Ns Two 0.218*** Three 0.279* * * Four 0.268 * * * Five 0.441 *** Owns car, motorcycle or boat -0.337*** House tenancy No registered title (vs. registered) 0.067** Does not own house (vs. registered) 0.05ONs Constant Source: Nicaragua LSMS 1998 Marginal values were calculated at the observed poverty rate of 47.85 percent using only significant variables or variables in a g -oup with at least one significant variable. Several variables lacked significance and were excluded: household head present, female hoilsehold head, household head with a partner, languages other than Spanish and indigenous, and participation in community organizations bIncludes persons receiving a pension, on vacation, or on medical leave Probability 15 ha. -0.121Ns -0.259*** Labor # of persons working last week' over total household size -0.281*** -0.396*** Agriculture provides more than 25 percent of total income 0.084Ns 0.156*** Own farm provides more than 25 percent of total income -0. 103Ns -0.142*** House amenities and size Water inside the house or property -0.069Ns -0.085** Electricity in the house (excluding generators) -0.200*** -0.144*** Telephone in the house -0.242*** Excluded Rooms per Person 0.021** 0.028*** Bad housing materials (type, walls, floor, roof, roads) One 0.029Ns 0.055Ns Two 0.229*** 0.158** Three 0.308*** 0.234*** Four 0.398*** 0.157Ns Five 0.479*** 0.317*** Owns car, motorcycle, or boat -0.247*** -0.438*** House tenancy No registered title (vs. registered) Excluded 0.163*** Does not own house (vs. registered) Excluded 0.106Ns Constant * Source: Nicaragua LSMS 1998 Marginal values were calculated at observed poverty rates of 30.48 percent for urban households and 68.52 percent for rural ouseholds using only significant variables or variables in a group with at least one significant variable. Several variables lacked ignificance and were excluded: household head present, female household head, household head with a partner, and participation ir ommunity organizations Significance level. Probability 60) by Headship Traditional Rural Urban Total Measure Male-Headed 0.45 0.40 0.42 Female-Headed 0.48 0.43 0.45 Table A11.10 - Share of Dependents Those <15 & >60) By New Headship New Measure Rural Urban Total Male-Fern 0.43 0.38 0.40 Fem Only 0.59 0j49 0.53 Male Only 0.39 0.34 0.37 Ot-h-er 0.94 0.96 0.95 7. Are female headed households larger? The answer from both sets of results is a resounding no! In fact, female-headed households are significantly smaller than male-female households. Table Al1.11 - Household Size by Headship Traditional Rural Urban Total Measure Male-Headed 5.78 5.33 5.55 Female-Headed 5.54 4.99 5.15 Table All.12 - Household Size by New Headship New Measure Rural Urban Total Male-Fem 6.27 | 5.72 5.97 Female Only 4.31 4.12 4.17 Male Only 3.07 2.71 2.89 |; Other X | 1.99 | 1.63 1.79 C. HEADSHIP, POVERTY AND FEMALE OWNERSHIP OF AND ACCESS TO ASSETS (RURAL ONLY) Ownership of Property by Gender 8. Because it is very difficult to get gender-disaggregated information on urban assets, only rural households are included in the rest of this appendix. Four types of rural assets are analyzed-land, movable agricultural machinery, livestock, and patio livestock. * Land. The proportion of rural Nicaraguan households in which women own land is quite low. These households do not differ by poverty status. However, a much larger proportion of female-headed households have female-owned property than do male-headed households. Table Al1.13 - Property Ownership by Women Land Poverty Status Traditional Extreme Overall Non-Poor Total Measure Poor Poor Male-Headed 0.02 0.02 0.02 0.02 Female- 0.11 0.14 0.08 0.11 Headed _____ Total 0.04 0.04 0.03 0.04 * Agricultural Machinery. Again the proportion of households in which women own agricultural machinery is quite low. However, there are vast and significant differences by poverty status-I 0 percent of extremely poor households have female-owned agricultural machinery, whereas only 3 percent of non-poor households do. One in five female-headed households have female-owned agricultural machinery, while only 2 percent of male-headed households do. Annex I1. Page 4 Table Al1.14 - Property Ownership by Women-immovable Agricultural Proverty Property Status Traditional Extre Overall Non-Poor Total Measure Poor I Poor Male-headed 0.04 0.02 0.01 0.02 Female-Headed 0.31 0.19 0.11 0.19 Total 0.10 0.05 0.03 0.05 * Livestock. About one in three rural Nicaraguan households have female-owned livestock. These numbers do not vary much by headship, but they do vary by poverty status. A significantly larger proportion of extremely poor households have female-owned livestock than do non- poor households. Table A11.15 - Property Ownership by Women-Livestock Poverty Status Traditional Extre Overall Non-Poor Total Measure me Poor Poor _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Male-Headed 0.04 0.31 0.25 0.31 Female-Headed 0.38 0.37 0.21 0.30 Total 0.40 j 0.32 0.24 0.31 * Patio Livestock. Less than a fifth of all households have female-owned patio livesiock. Female-headed households are more than twice as likely as male-headed households to have female- owned patio livestock. The incidence of female-ownership of patio livestock falls with povertv. Table A11.16 - Property Ownership by Women-Patio Livestock Poverty Status Traditional Extre Overall Non-Poor Total Measure me Poor Poor Male-Headed 0.09 0.17 0.15 0.15 Female-Headed 0.30 0.31 0.39 0.34 Total 0.14 0.19 0.20 0.18 Women's Participation in Agricultural Sales 9. Overall, few women participate in agricultural sales. There is no correlation between poverty and female participation in agricultural activities. This participation does. however. vary by headship. About 13 percent of female-headed households report female participation in agricultural sales whereas only 2 percent of male-headed households do. Table A11.17 -Women's Participation in Agricultural Activities- Sales Poverty Status Traditional Extre Overall Non-Poor Total Measure Poor Poor Male-Headed 0.02 0.01 0.04 0.02 Female-Headed 0.09 0.19 0.10 0.13 Total 0.03 0.04 0.05 0.04 Incidence of Credit by Gender 10. Two types of credit activities are considered here-loans and purchases on credit. * Loans. Consistent with women's low level of participation in agricultural activities, women's use of credit is also quite rare. The incidence of loans by gender varies only slightly by Annex 11, Page 5 headship. Hardly any women in extremely poor households have taken out loans. On the other hand, in 11 percent of non-poor female-headed households, women have taken out loans. Table Al1.18 - Women's Access to Credit-Loans Poverty Status Traditional Extre Overall Non-Poor Total Measure me Poor Poor _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Male-Headed 0.00 0.02 0.07 0.04 Female-Headed 0.01 0.04 0.11 0.06 Total 0.01 0.02 0.08 0.04 Purchases on Credit. Another source of credit is that provided by suppliers, but the pattern does not differ significantly from that of loans. Table A11.19 - Women's Access to Credit-Purchase on Credit Poverty Status Traditional Extreme Overall Non-Poor Total Measure Poor Poor Male-Headed 0.01 0 02 0.03 0.03 Female-Headed 0.03 0.04 0.07 0.05 Total 0.01 0 03 0.04 0.03 D. SCHOOL ATTENDANCE 11. Is school attendance bv poverty status and headship different? According to expectations, school attendance is significantly lower among the poor and extremely poor households than it is among the non-poor. Contrary to expectations. school attendance is higher among female-headed households than among male-headed households. This is another indication that female-headed households in Nicaragua are not poorer than their male counterparts. Table Al 1.20 - Proportion of Children (6-14) in School by Poverty Status and Headship Poverty Status Traditional Extreme Overall Non-Poor Total Measure Poor Poor Male-Headed 0.61 0.80 0.94 0.84 Female-Headed 0.71 0.83 0.94 0.88 Total 0.63 0.81 0.94 0.85 Annex 12, Page I Annex 12 - A Growth Accounting Approach By Ulrich Lachler I. In the last 3 years, Nicaragua's real GDP growth has been averaging over 5 percent, and over 2 percent in per-capita terms. This is relatively high by world standards, and above average for developing countries as a group. To see how much scope there is for raising output growth beyond recent performance levels, it is useful to apply a growth accounting framework that decomposes economic growth into that portion attributable to input growth and the portion due to productivity growth. 2. Assume a stable relationship, Q = AF(K. (LH)), between output (Q), and the inputs consisting of physical capital (K), workers (L). their human capital (H), and total factor productivity (A). The evolution of per capita output can then be expressed as: GQ - GL = a(GK/L) + b(GH) + GA where Gx refers to the growth rate of variable X, and (a,b) refer to the income shares of capital and labor. Table A 12.1 compares the evolution of each of these terms in Nicaragua with that in several other regions. Table A12.1 Sources of Growth in Selected Regions (in percent) Region/Period Contribution made by: Output per Physical Total Factor Worker Capital Education Productivity NICARAGUA 1960-70 3.8 2.0 0.1 1.6 1971-80 -2.5 0.8 0.2 - 3.5 1980-86 -3.6 - 0.2 0.9 - 4.3 1987-92 -5.6 - 1.3 0.8 - 5.1 1993-98 1.3 0.3 0.6* 0.4 Other Central America 1960-70 2.5 1.2 0.2 1.1 1971-80 1.9 1.6 0.4 - 0.1 1980-86 -3.1 - 0.3 0.4 - 3.3 1987-92 0.5 0.0 0.4 0.1 Latin America 1960-70 3.2 1.3 0.2 1.6 1971-80 2.6 1.6 0.3 0.7 1980-86 -1.4 0.4 0.4 - 2.2 1987-92 0.1 -0.0 0.4 - 1.0 East Asia (excl. China) 1960-70 3.8 2.5 0.5 0.8 1971-80 4.8 3.4 0.4 0.9 1980-86 2.7 2.5 0.6 - 0.4 1987-92 5.1 2.6 0.6 1.8 Source: Barry Bosworth, Susan M. Collins and Yu-chin Chen, "Accounting for Differences in Economic Growth". December 1995; and U.Lachler for figures on Nicaragua in 1993-98. * = educated guess. 3. Nicaragua exhibited the same pattern of growth as other Latin American countries during the 1970s and early 1980. Instead of recovering in the late 1980s, however, Nicaragua's growth performance worsened, both on account of less investment in physical capital and a decline in total factor productivity growth. Only on the human capital front, proxied by education Annex 12. Page 2 attainment, was Nicaragua able to keep up, not only with its Latin American neighbors, but even with the fast growing East Asian economies. This situation turned around in the mid- I 990s, when Nicaragua's growth performance surpassed that of its neighbors. The process of physical capital de-cumulation that had occurred over the previous decade was halted and gradually reversed, which contributed to this turn-around. The main contributing factor, however, was the reversal of performance oil the productivity front. The key to this story, however, is not that total factor productivity growth became highly positive, but rather that it stopped being highly negative. This reversal is attributable to the major advances made on both the political and economic fronts since 1990: the cessation of civil war, macroeconomic stabilization and fiscal reform, trade policy and pricing reforms that opened up the economy and eliminated domestic policy distortions, public sector reforms that slimmed down the size of government, the privatization of state-owned enterprises, and financial sector reform. These reforms eliminated major distortions that had led to systematic resource misallocations, which are reflected in the negative TFPG rates of past decades. 4. To increase output growth beyond the average reported in Table A12.1 Nicaragua either would have to make a greater savings effort, which translates into faster physical and human capital accumulation. or improve its productivity, which includes improving the quality of factor inputs. This has in fact been happening: in 1996-97, per-capita income growth averaged 2.1 percent and total investment rose to almost 30 percent of GDP, which ranks among the highest investment rates in Latin America. We estimate that a further increase of investment by 5 percentage points of GDP would translate into a growth increase of less than one percent. At the same time. it would mean a decline in domestic consumption by the full five percent. Also, increased investment spending would entail higher recurrent expenditures for operation and maintenance, which would require even further cuts in consumption elsewhere. A more promising alternative under these circumstances may be to focus on improving the quality of physical and human capital investment, which would become reflected in higher total factor productivity growth. Looking at the historical comparisons in Table 1.3, this source of growth could add another 0.5 to 1.0 percentage points to per-capita GDP growth on a long-term basis. Annex 13, Page 1 Annex 13. Education and Poverty in Nicaragua by Gustavo Arcia INTRODUCTION 1. Public education in Nicaragua is the main vehicle for increasing the quality and stock of human capital. As such, it is considered to be a key component in the country's social policy. The objective of this paper is to assess the education sector within the context of poverty in 1998, and compare it with the state of the sector in 1993. This assessment is based on the analysis of data from the results of the 1998 Living Standards Measurement Survey (LSMS) and data from the Ministrv of Education and other sources. The purpose of this paper is to identify key education issues that need to be addressed by the Government and donors in order to increase the human capital of the poor. 2. The paper is organized as follows: Section 2 presents a summary of problems affecting the education sector, particularly public education, and the actions taken by the Government to address them. Section 3 presents empirical evidence from the 1993 and 1998 LSMS on results in education and their relation to the different levels of poverty. There is also a review of key indicators for education, such as enrollment, repetition, and dropout rates by poverty level, as well as the economic returns from basic education. Section 4 presents the LSMS results on education financing, emphasizing the private costs of public education and the impact of these costs on family expenditures at different levels of poverty. This section also reviews Government spending on basic education in order to be able to make an evaluation of possible policy changes in education financing, such as targeting family subsidies to improve the demand for public education among the poor. Section 5 presents the main conclusions and recommendations. OVERVIEW OF THE EDUCATION SECTOR 3. Nicaragua's education sector has substantial problems. i) Inadequate school coverage, especially in rural areas. Less than 35% of preschool-age children attend preschool. Although primary school coverage from grades I to 3 is 92%, this rate drops to less than 50% by grade 6. In rural areas a large proportion of primary schools do not give classes beyond the fourth grade. ii) Inadequate and incomplete infrastructure. Despite substantial gains in the rehabilitation and expansion of school infrastructure -especially in rural areas- about 35 percent of public schools still need to be repaired and/or expanded. iii) Low teacher salaries. The sector has problems in attracting and retaining good teachers, mainly because of low salaries. However, the absence of a good system for teacher accountability makes a salary increase a weak instrument for improving teaching quality. iv) Low level of teaching quality and learning. Poor teaching quality is cited as the most visible problem for the quality of education in Nicaragua. It stems from the poor quality of pre-service teacher training, the high incidence of non-accredited teachers in the system, and the low rates of retention of good teachers. v) Weak connection between school autonomy and a centralized MECD, which acts as a centralized ministry outside of the financial autonomy program. vi) Low managerial capacity at the central and local levels. Despite progress in budget planning and in the financial management of fiscal transfers to autonomous schools, the managerial performance of the MECD's Program Directors remains low because of a lack of training, deficient managerial information in the system, inadequate statistics, and unclear roles and responsibilities. Also, most school directors lack managerial and human resource management skills. vii) Weak planning, research, evaluation, and statistical capacity due to a lack of technical personnel. viii) No feedback mechanisms to know about the impact of policy on decision making at the central and local level and low capacity for communicating system performance to decision-makers and parents in ways that promote accountability & participation. Addressing the Problems of Public Education: Government Goals and Strategies 4. The Government is aware of the above problems and has defined two main goals in education: increase the coverage of basic education and improve education quality. To achieve these goals the MECD has defined a strategy that combines the following components: 5. Expansion and consolidation of school autonomy. The objectives of school autonomy are to: i) increase community participation in school management; ii) increase local participation in school finance; Annex 13, Page, 2 and, iii) increase local accountability. Autonomous schools receive a fiscal transfer that is administered by a local school council in which parents are a majority. This transfer is based on a formula that takes into account the number of students, the location of the school, and the school record for repetition and dropouts. The MECD selects the school Principal from a short list submitted by the school council. The school council meets at least once a month, and most decisions are made by simple majority. The entire autonomy program is run on an ad hoc basis because there is no legislation to back it up yet. The Government believes that school autonomy will lead to steady improvements in educational performance due to better local accountability and more parental support for learning activities. MECD officials expect that all secondary school students, and 70 percent of primary school students, will be covered by school autonomy by the year 20001. The success of school autonomy has been mixed. Overall student achievement has improved in those autonomous schools where parents and teachers have a clear understanding of the program. the rules of autonomy are well understood, and thle Principal exercises good leadership2,3. To improve school finances, autonomous schools are allowed to ask for voluntary contributions from parents. However, a recent evaluation indicates that parent support and acceptance of autonomy was positivelv linked to their ability to pay additional fees to increase school revenues. 6. Rehabilitation and replacement of school infrastructure. In collaboration with the Emergency Social lnvestment Fund (FISE), the MECD has expanded the quality of primary school infrastructure in rural areas, including those affected by Hurricane Mitch. Schools to be rehabilitated or rebuilt are selected with the help of a poverty' map based on the 1993 LSMS. This map is being revised using 1998 LSMS data. Little funding has been earmarked for the rehabilitation and replacement of secondary schools. 7. Modernization of central administration and planning. The MECD is in the process of modernizing central administration and planning through two mechanisms: i) a system for measuring and reporting student achievement; and, ii) an Integrated Management Information System containing education statistics, administrative and budgetary information, and information on student achievement. 8. Expansion of preschool coverage. The MECD is expanding preschool access through communit) preschools, hoping to cover close to 30% of preschool-age children by 2001. The community prescho-Jol system includes the training of mothers as preschool caregivers, nutritional assistance, and preventive health care. Since community preschools seem to be more cost-effective4 than iformal preschools stafled by teachers, the MECD is phasing out the latter. 9. Pedagogical improvements. The MECD has modified its primary school curriculum to make it less encyclopedic and more oriented to basic skills in math, reading, and writing. It has also expanded the provision of textbooks and school materials with external funds. Books are rented to students at a modest fee to encourage better care of them. However, textbooks are still in short supply in more isolated areas. lIt is better to report autonomy coverage in terms of the percentage of students. rather than in terms of the number of schools covered by autonomy. Itt is a better indicator of impact on students and their families. Reporting the number of schools covered by autonomy would tend to underestimate its impact, since the largest number of schools are in rural areas with relatively few students. 2 Fuller, Bruce, and Magdalena, Rivarola, 1998. "Nicaragua's Experiment to Decentralize Schools: Views of Parents, Teachers, and Directors." Paper No. 5, Working Paper Series on Impact Evaluation of Education Reforms. Development Economics Research Group, The World Bank, Washington DC. 3King, Elizabeth, Laura Rawlings, Berk Ozler, Patricia Callejas, Nora Gordon, and Nora Mayorga de Caldera, 1996. "Nicaragua's School Autonomy Reform: A First Look." Paper No. 1, Working Paper Series on Impact Evaluation of Education Reforms. Policy Research Department, The World Bank, Washington D.C. King, Elizabeth, and Berk Ozler, 1998. "What's Decentralization Got To Do With Learning? The Case of Nicaragua's School Autonomy Reform." Paper No. 9, Working Paper Series on Impact Evaluation of Education Reforms. Development Research Group, The World Bank, Washington DC. 4Arcia, Gustavo, and Vanessa Castro, 1999. Evaluaci6n de medio termino del Proyecto PAININ. Consulting report presented to the Interamerican Development Bank, Managua. Annex 13, Page, 3 The Ministry is revising the curriculum and teaching methods in teacher training schools and improving in-service training. 10. Linking teacher incentives to teacher and student attendance. The MECD is implementing a program whereby teacher salaries are linked to student and teacher attendance. If attendance by a teacher and her/his students is perfect, the supplement can increase a salary by 50 percent. Half the supplement depends on the teacher's attendance and the other half on the attendance of her/his students. The salary supplement applies only to teachers in autonomous schools as a way to promote the voluntary incorporation of more schools to the autonomy program. 11. Long tern planning. The MECD has prepared a long-term plan for public education that reflects the views of most stakeholders in the sector and discusses some new areas that had not been covered in previous strategic efforts5. The Plan, as it is known informally, emphasizes eight basic principles, which are then translated into strategic objectives. These principles and strategic objectives are presented in Table Al in the Annex. The Plan is ambitious. but it follows the trends in education reform in the rest of Central America and the Caribbean. 12. So far, the Government strategy has served to increase the supply of schools in rural areas, increase the participation of parents in school affairs, and improve the financial management of the MECD. Although other elements of the strategy are underway and/or their effects will be felt in a few years. a remaining concern now is education quality. The MECD and parents recognize that teacher quality, the quality of the classroom environment, and the quality of teaching must improve. However, there are no direct indicators or measurements of quality, as described above. This is a serious shortcoming, which may hinder the MECD's ability to monitor the effect of its policies on teaching quality and learning in the near future. EDUCATION AND POVERTY: A REVIEW OF SECTOR PERFORMANCE, 1993-1998 13. The results of the 1998 LSMS show that some key output indicators related to human capital accumulation have improved since 1993, while others remain unchanged. Below is a summary of the most important changes in educational outputs (Table A13. 1). 14. Illiteracy has declined significantly among the poor, and especially among the extreme poor. At the national level, the proportion of illiterate persons in extreme poverty has declined from 44.5% to 37.6%, for a net decline of 6.9 percentage points over the five-year period. Unfortunately, illiteracy is highly correlated with poverty. The results from 1998 indicate that illiteracy among the urban extreme poor is four times more prevalent than among the urban non-poor. In rural areas the rate of illiteracy among the extreme poor is twice the rate of the non-poor. Among the extreme poor, almost three of every ten people in urban areas and four of ten people in rural areas are illiterate. Still, the rate of illiteracy among the non- poor is still high: around 20% in rural areas and 7% in urban areas. 15. School attendance has also improved significantly since 1993. Among the poor, attendance by preschool-age children has increased by more than 15 percentage points -the same rate as among non- poor children. More significantly, school attendance among rural children 4-6 years of age has increased by more than 20 percentage points between 1993 and 1998. These gains show progress, but the proportion of preschool-age children attending school is still very low among the extreme poor (25%) and among the poor (33%). In comparison, 60% of non-poor children of preschool age attend school. 5 The MECD was extremely successful in reaching consensus among a diverse group of stakeholders with different -and sometimes conflicting- interests. Close to 3,000 people had an opportunity to give an opinion on the Plan. In part, this success was due to the inclusive nature of the dialogue, in which everyone's voice was reflected in the interim documents. Another part of the success was the generation of discussion without a budget constraint,. This in turn resulted in the support of all stakeholders for presentation of the Plan. Ostensibly, the MECD will now begin the second phase of the Plan: setting priorities to configure them to the constraints of the education budget. Annex 13, Page, 4 Table A13.1 Comparison of key educational indicators, 1993 and 1998 Group Ililiterate % 4-6 % Not Net Net Gross % Not Mean Mean Mean Distance Distance (10 or attending attending Enrollment Enrollment Enrollment attending Y'ears i'ears Years Primary Primary more) pre- 7-12 Primarv Primary Primary 13-18 Schooling Schooling Schooling (kms) (mins) years school Boys Girls Total Male Female All Nicaragua 1993 21.5 31.3 21.5 77.9 81.4 79.6 43.0 4.5 4.5 4.6 1998 18.8 44.8 14.3 77.9 81.4 107.3 43.5 4.9 4.7 5.0 0. 9 16.1 Point Change -2.7 13.5 -7.2 0.0 0.0 27.7 0.5 0.4 0.2 0.4 Extreme Poor 1993 44.5 8.3 44.7 71.8 1.9 1.8 2.0 1998 37.6 25.4 32.4 62.7 67.9 89.1 70.0 2.3 2.2 2.3 1., 25.7 Point Change -6.9 17.1 -12.3 -1.8 0.4 0.4 0.3 Poor 1993 33.8 17.4 32.3 57.6 2.8 2.7 2.9 1998 29.6 33.8 21.3 73.5 77.8 102.3 60.7 3.1 3.0 3.2 1. 21.7 Point Change -4.2 16.4 -11.0 3.1 0.3 0.3 0.3 Non-Poor 1993 10.6 50.9 7.4 26.9 6.1 6.1 6.0 1998 10.4 60.3 6.1 83.2 85.5 113.1 26.7 6.2 6.2 6.3 0.6 12.7 Point Change -0.2 9.4 -1.3 -0.2 0.1 0.1 0.3 Urban (All) 1993 10.6 48.5 11.3 82.9 85.3 84.1 26.6 5.9 5.9 5.8 1998 10.2 57.1 9.0 82.9 85.3 113.1 28.1 6.2 6.2 6.1 0.4 8.9 Point Change -0.4 8.6 -2.3 0.0 0.0 29.0 1.5 0.3 0.3 0.3 Urban Extreme Poor 1993 29.0 20.2 35.3 65.9 70.5 68.1 57.7 3.0 3.0 3.0 1998 28.4 32.3 28.5 65.9 70.5 95.6 61.7 3.0 3.1 3.0 0.6 11.4 Point Change -0.6 12.1 -6.8 0.0 0.0 27.5 4.0 0.0 0.1 0.0 Urban Poor 1993 17.8 31.7 19.9 79.0 81.9 80.4 37.6 4.1 4.1 4.1 1998 19.6 43.8 15.9 79.0 81.9 110.1 48.5 4.0 4.1 4.0 0.4 9.7 Point Change 1.8 12.1 -4.0 0.0 0.0 29.7 10.9 -0.1 0.0 -0.1 Urban Non-Poor 1993 7.5 59.1 6.3 85.2 87.1 86.2 20.8 6.6 6.7 6.5 1998 6.7 66.2 5.1 85.2 87.1 114.8 19.4 6.9 7.1 6.9 0.4 8.6 Point Change -0.8 7.1 -1.2 0.0 0.0 28.6 -1.4 0.3 0.4 0.4 Rural (All) 1993 38.4 13.0 33.7 72.5 77.2 74.8 65.6 2.5 2.4 2.6 1998 29.8 33.6 20.1 72.5 77.2 100.9 61.9 3.2 3.0 3.4 1.5 25.6 Point Change -8.6 20.6 -13.6 0.0 0.0 26.1 -3.7 0.7 0.6 0.8 Rural Extreme Poor 1993 48.8 5.5 47.3 65.9 70.5 68.1 75.8 1.6 1.5 1.6 1998 40.6 23.6 33.6 61.7 66.9 86.9 72.6 2.0 1.9 2.1 2.1 30.2 Point Change -8.2 18.1 -13.7 -4.2 -3.6 18.8 -3.2 0.4 0.4 0.5 Rural Poor 1993 43.6 10.1 39.0 79.0 81.9 80.4 69.9 2.0 1.9 2.1 1998 35.0 28.7 24.1 70.5 75.7 98.3 66.7 2.6 2.5 2.7 1.7 27.8 Point Change -8.6 18.6 -14.9 -8.5 -6.2 17.9 -3.2 0.6 0.6 0.6 Rural Non-Poor 1993 23.6 24.6 11.7 85.2 87.1 86.2 51.2 3.8 3.8 3.9 1998 20.2 47.8 8.7 78.1 81.4 108.5 49.52 4.3 4.1 4.6 1.2 22.3 Point Change -3.4 23.2 -3.0 -7.1 -5.7 22.3 -1.68 0.5 0.3 0.7 Source: LSMS 1993 and 1998. 16. Similar progress is observed in the proportion of children not attending school. At the national level the proportion of children not attending school declined by 7.2 percentage points -from 21.5% to 14.3%. Most of the increase in school attendance was observed among the extreme poor, where the proportion of children not attending school declined by 12.3 percentage points, almost double the figure nationallx. Moreover, the decline in the proportion of children not attending schools was more pronounced in rural Annex 13, Page, 5 areas, particularly among the rural poor, with a decline in non-attendance more than double the decline observed in urban areas. 1 7. Despite progress in school attendance, about one third of extremely poor children in rural areas between 7 and 12 years of age do not attend school -more than three times the rate of the rural non-poor. This situation is even worse in urban areas, where 28.5% of extremely poor children between 7 and 12 do not attend school, versus 5. 1 % of the non-poor children in the same age group -almost a 6-to- I ratio. 18. Another area where significant progress was made from 1993 and 1998 is the gross enrollment ratio in primary school, increasing by almost 30 percentage points in urban areas -regardless of poverty level -and around 20 percentage points in rural areas. 19. The progress in literacy, preschool enrollment, school attendance, and gross enrollment show that the additional funding targeted to the preschool and primary school population has paid off. However, there is plenty of room for more progress. Although the gains observed in literacy, school attendance, and school enrollment are significant, other key indicators did not change much during the 1993-98 period. As Table I shows, neither the net enrollment ratio, nor secondary school attendance increased very much over the five-year period. On the other hand, there seems to be an upward trend in educational achievement, which may signal a better prognosis for the poor. Figure I shows the educational achievement of the population by age cohorts in 1998. It clearly shows that people between 20 and 29 years of age have more schooling than people above 30, and significantly more schooling than people 40 or older. This suggests that -if the current upward trend in per-student expenditures is maintained- the cohort between 10 and 19 years old may continue increasing their schooling at the same rate as the 20-29 year old cohort. A Detailed Look at the Indicators for Educational Outputs and Poverty 20. Total Enrollment. The MECD figures for total enrollment between 1993 and 1998 show a mixed record. On the positive side, preschool and secondary enrollment has increased at a rate of 9% per year (Figure 2 and Table 2). However, the same official figures for primary school enrollment show an increase of 1% per year, less than half the rate of population increase. The LSMS gains in the gross enrollment rate show a discrepancy that may reflect significant problems in the way the MECD collects, registers, and verifies enrollment figures, or problems in the reporting of enrollment statistics from private schools, which have little incentive to send correct figures. In the case of university education, the data are difficult to analyze, since enrollment figures do not reflect full-time equivalencies. Annex 13, Page, 6 Fig. A13.1 Mean Years of Schooling by Age Cohort and Poverty Level, 1998 7 C6 S ____ . __- -_ ___ _- 0 Cohort of Cohort of Cohort of Cohort of Cohort of 10-19 yrs 20-29 yrs 30-39 yrs 40-49 yrs 50t yrs o Extreme Poor * Poor * Non-poor Source: LSMS 1998 Fig. A13.2 Total School Enrollment 1991-99 900000 800000 ___ 700000 ____ _ 600000 m 500000 - 400000 _ ___ 300000 ___ . __ 200000 __ _ __L. 1000000 0 1991 1992 1993 1994 1995 1996 1997 1998 1999 +... Pre-school a Primary education .; Secondary education Source: MECD Annex 13, Page, 7 Table A13.2 - Total Enrollment by level, 1993 and 1998 Level 1993 1998 5-yr. growth rate (%)6 Total 1,003,741 1,190,120 4 Preschool 79.543 143,677 9 Primary 737,476 783,090 Secondarv 186,722 287,21 7 9 Special education 3,168 3,176 1 Adult education 64,625 59,007 Pre-service teacher training 8,907 5.649 -7. University' 30.000 48,000 n.a. Source: MECD 21. Gains in school attendance among children ofpre-school age. Between 1993 and 1998 preschool attendance increased. especially in rural areas, thanks to the creation of Community Preschools by the MECD. Table 3 shows that, overall, the proportion of 4-6 year olds in school has increased by 12.5 percentage points between 1993 and 1998. This increase, shown in Table 3, has been more significant in rural areas, where the proportion of 4-6 year olds attending school has more than doubled, from 13.8% to 33.7%. Table A13.3 - School attendance of children 0-3 and 4-6 years old 1993 1998 0-3 years 4-6 years 0-3 years 4-6 years All 3.3 32.6 4.6 45.1 Urban 5.2 50.3 6.3 57.6 Rural . 1.3 13.8 3.1 33.7 Extreme Poor 0.7 8.5 3.8 25.4 Poor 1.6 18.6 2.8 34.1 Non-poor 6.0 52.4 7.4 60.5 Quintiles Poorest 09.3 3.4 26.8 11 1.4 21.1 2.3 36.2 III 3.2 36.3 2.2 49.4 IV 5.8 49.5 6.6 57.6 Richest 8.7 63.4 14.6 75.8 Region Managua 6.7 51.4 5.8 62.5 Pacific - Urban 3.7 45.2 6.2 59.0 Pacific - Rural 1.0 14.9 3.1 45.8 Central - Urban 4.6 48.7 8.5 49.1 Central - Rural 0.7 8.7 3.0 26.7 Atlantic- Urban 4.8 42.3 5.0 45.1 Atlantic - Rural 1.3 8.7 2.2 15.8 Source: LSMS 1993 and 1998 22. Gross and net enrollment rates. Nationally, the primary gross enrollment rate declined by about 10 percentage points due to a lower enrollment rate in urban areas higher than the increase in the gross 6 The growth rate is calculated from current enrollment figures. 7 The number of students in universities is approximate, and most are only part-time. There are no estimates of the number of full-time equivalent students. Annex 13, Page, 8 enrolment rate in rural areas (Table 4). In contrast, the gross enrollment rate in secondary school went up significantly, from 21.6% to 57%, due to a three-fold increase in the gross enrollment rate in rural areas and a four-fold increase in urban areas. In general, the gross enrollment rate increased among all income groups, with the exception of gross primary enrollment among the non-poor, which declined. The overall increase on the gross enrollment rate suggests a corresponding decrease in the dropout rate, which is good news for sector performance. Table A13.4 - Gross Enrollment Rates - Primary and Secondary by Gender, 1993-1998 1993 1998 Primary Secondary Primary Secondary Total |Male IFemale Total MeFemale Total Male Female Toal e IFemale All 118.1 115.4 121.0 21.6 22.() 21.3 107.3 106.9 107.6 57.0 49.9 63.7 Rural 93.4 90.2 96.7 9.7 9.9 9.5 100.9 101.0 100.9 30.0 23.7 36.5 Urban 138.6 136.1 141.3 30.3 31.3 29.3 113.1 112.4 113.8 79.4 73.5 84.6 Total Extreme 75.9 72.7 79.1 4.0 2.7 5.5 89.1 87.3 91.1 11.3 8.6 14.8 Poor Poor 97.3 93.9 100.9 12.4 11.6 13.2 102.3 101.7 103.0 26.4 21.9 31.3 Non-poor 144.9 142.6 147.5 31.8 34.3 29.5 113.1 113.3 112.9 86.6 81.5 9(.9 -Ur ba-n Extreme 90.5 89.9 91.1 10.3 * * 95.6 94.3 96.9 21.5 17.8 26.1 Poor Poor 122.6 119.8 125.8 21.5 19.8 23.4 110.1 106.7 113.7 44.3 45.1 43.5 Non-poor 147.7 145.6 149.9 34.9 38.0 32.1 114.8 115.8 113.9 94.7 87.7 100.5 Rural Extreme 71.9 68.3 75.6 2.2 * 3.7 86.9 85.2 89.1 8.0 5.7 11.0 Poor Poor 83.5 79.1 88.0 6.7 6.6 6.8 98.3 99.0 97.6 17.2 10.5 24.8 Non-poor 134.1 131.6 137.2 19.7 21.0 18.3 108.5 106.8 110.2 62.6 63.1 62.2 Source: LSMS 1993 and 1998 * = n< 10. Gross enrollment rate for primary is defined as (number in elementary school/number of 7-12 years old). For secondary is (number in secondary school/number of 13-17 years old). 23. A gross enrollment rate above 100% indicates there are more children in school than the total number of children in the appropriate school age group. As a result, a gross enrollment rate of 107% indicates that a.significant proportion of students is over (or under) age. For secondary school the gross enrollment rate shows that the total number of children in school (including those who are over-age) is almost half the total number of 13-18 year olds, indicating that a very large number of children of secondary school age do not attend school. 24. In general the gross enrollment rate is inversely related to poverty: a lower gross enrollment ratz is associated with a higher level of poverty. In the case of primary school, the difference in the gross enrollment rate between urban and rural areas in 1998 is about 9 percentage points overall among the extreme poor, and 6 percentage points among the non-poor. For secondary school this gap widens considerably. The difference in the gross enrollment rate between urban and rural areas among the extreme poor is 13.5 percentage points, and among the non-poor this difference is of 32.1 percentage points. 25. In the case of the net enrollment rate, there was a significant improvement among the rural poor from 1993 to 1998 and a decline among the non-poor, especially in rural areas. Overall, the strength and magnitude of the improvement among the poor helped generate an overall increase in the net enrollment rate at the national level. Annex 13, Page, 9 26. The 1998 net enrollment rate for primary education -defined as the ratio of students between 7 and 12 to the total number of 7-12 year olds- is 79.6% (Table 5). This percentage means that four out of five primary school students are in the appropriate school level for their age. For secondary school, the results show that only 36.8% of students are in the appropriate age bracket. Again, the higher the level of poverty, the lower the net enrollment rate. Moreover, there are very large differences between urban and rural areas, which are worse in secondary school. Both the gross enrollment rate and the net enrollment rate show that females do better than males. Table A13.5 - Net Enrollment Rates - Primary and Secondary by Gender, 1993-1998 1993 1998 Primary Secondary Primary Secondary Total |Maale lFemale TaMaleJiemale Total IMale |Female Total |Male Female All 75.6 74.4 76.8 15.5 15.8 15.2 79.6 77.9 81.4 36.8 31.0 42.3 Rural 65.5 62.4 68.7 7.2 7.2 7.1 74.8 72.5 77.2 17.2 11.0 23.7 Urban 84.0 84.3 83.6 21.5 22.5 20.6 84.1 82.9 85.3 53.1 49.1 56.7 Total Extreme 55.3 51.7 58.8 3.6 2.5 4.9 65.1 62.7 67.9 8.1 5.8 11.1 Poor Poor 66.6 64.0 69.4 9.4 8.7 10.2 75.6 73.5 77.8 18.8 14.5 23.6 Non-poor 87.2 87.7 86.6 22.2 24.4 20.2 84.4 83.2 85.5 54.3 49.6 58.1 Urban Extreme 64.7 63.9 65.5 9.9 * * 68.1 65.9 70.5 15.1 13.0 17.7 Poor Poor 77.8 77.9 77.8 16.0 14.7 17.5 80.4 79.0 81.9 30.8 29.6 32.1 Non-poor 87.5 88.1 86.8 24.4 27.1 22.0 86.2 85.2 87.1 62.8 58.7 66.1 Rural Extreme 52.7 48.7 56.9 1.8 * 2.9 64.1 61.7 66.9 5.9 3.5 8.9 Poor Poor 60.5 56.0 65.0 5.4 5.1 5.7 73.1 70.5 75.7 12.8 7.2 19.1 ,Non-poor 86.0 86.2 85.8 13.3 14.5 12.0 79.7 78.1 81.4 28.8 22.4 34.0 Source: LSMS 1993 and 1998 * = n< 10. The net enrollment rate for primary is defined as (number of 7-12 yrs. old in elementary school/number of 7-12 years old). For secondary it is (number of 13-17 years old in secondary school/number of 13-17 years old). 27. In the case of net enrollment rates8, the difference between poverty groups is more marked at both primary and secondary levels. As shown in Figure 3, net enrollment rates increase significantly with the level of family per capita expenditures. Among the extreme poor, less than 70% of primary students are within the appropriate age range for primary school, while the rate for the non-poor is around 85%. The difference suggests a higher number of over-age students among the extreme poor, which in turn indicates late entry and/or higher levels of repetition. In secondary school the difference in net enrollment rates by poverty group is dramatic, especially between urban and rural areas. In general, the secondary net enrollment rate among the non-poor is six times the net enrollment rate of the extreme poor. Moreover, the secondary net enrollment rate in urban areas is twice the secondary net enrollment of rural areas, even among the non-poor. 28. As shown in a previous section, the proportion of over-age primary students in rural areas is substantially larger than in urban areas. This difference -due in part to late entry and in part to higher repetition rates in the first and second grades- may explain the variation in net enrollment rates between urban and rural areas. In secondary school, however, the difference is not as easily explained. It may combine several factors, including earlier entry into the labor market by rural children, earlier acquisition s The net enrollment rate is obtained by dividing the number of 7-12 year old primary school students by the total number of 7-12 year olds in the country. Annex 13. Page, 10 of adult responsibilities (i.e.: spouse and children) in rural areas, inappropriate curricula for rural children, significantly lower access to secondary schools in rural areas, and a significantly higher financial btrden for rural families. All these factors -which may be intertwined- may account for the very large differences in net enrollment between rural and urban areas. The results of the 1998 LSMS may help explain the main reasons, as discussed in the section on family expenditures. Fig. A13.3 Net Enrollment Rates 1998 100.0 90.0 80.0 _ _ 70.0 _____ c 60.0 _ = -- - Primarn ? 50.0 - 3 40.0 __ }I Secondary 30.0--_ - - _ - 20.0 _ 10.0 --- Urban Urban Urban Rural Rural Rural Ext. Poor Non- Ext. Poor Non- Poor poor Poor poor Source: LSMS 1998 29. Given the high repetition rates in Nicaragua, it could be considered as being within the acceptable range to have children in a given grade who are one year older than they should be. Even with this flexibility in the interpretation of the net enrollment rate, the number of chiildren who are not in their ideal grade is very large. After allowing for some flexibility in the determination of the appropriate age for a grade, the problem of over-age affects around 30% of the children in primary school. Although over.age is related to high repetition, late enrollment could also be a problem, especially in rural areas. Accordiing to a 1995 survey on repetition and dropouts9, only 40% of 7-year olds were enrolled in school, while in urban areas, 77% of 7-year olds were enrolled in school. 30. Repetition Rates. Repetition continues to be a problem. LSMS estimates show repetition rates somewhat higher than official averages'° (Table 6). The pattern of repetition is typical of other count^ies in the region. First grade repetition is the highest, declining steadily until 6"' grade. Repetition in secondary school follows a similar pattern, but to a lesser degree. The 1997 MECD repetition rate for 7t1 grade is 8.91, declining steadily until the I 1th grade, where the rate is 2.2 at the end of secondary school. The MECD repetition rates for the I" and 2nd grades are unusually high since first- and second-graders are supposed to be automatically promoted to the next grade. However, official figures suggest that the automatic promotion program may not be practiced, or that there are serious problems with the collection and verification of repetition data. Table A13. 6 - MECD and LSMS Repetition Rates for Primary School Grade MECD 1997 LSMS 1998 Ist 21.5% 25.1% 2ni 12.6% 14.3% 3 r 10.4% 8.8% 4th 8.4% 9.3% 5th 6.4% 6.2% 6th 3.5% 4.0% Source: MECD, 1999; LSMS, 1998 9 Gargiulo, Carlos, and Luis Crouch, 1995. Nicaragua: Schooling, Repetition '° MECD figures on repetition are taken from MECD, Direcci6n de Estadisticas Educativas, 1999. Estadisticas de la Educacion en Nicaragua, 1997. Managua, pp. 31-32. Annex 13, Page, 11 31. LSMS 98 estimates indicate that the repetition rate for rural first-graders repeat is about one-fifth more than for urban first-graders: 29.1% versus 22.1% (Table 7). This is a marked improvement from 1995, when the first grade repetition rate was 25% in urban areas and 40% in rural areas"'. Still, the figures show a serious problem of inefficiency in resource use. This must be addressed by putting into place policies oriented to the reduction of the I"5 and 2nd grade repetition rate, such as the placement of the best teachers in the first two grades and the expansion of preschool programs across the country. Clearly, the current policy of automatic promotion is not working. The overall repetition rate for urban areas is 11.5%, while for rural areas it is 12.7%. As in the 1995 survey, girls repeated less than boys: 10.3% versus 13.9 %. Table A13.7 - LSMS 98 - Primary School Repetition for Primary School by Area and Gender Grade Urban Rural Male Female jst 29.1% 22.9% 25.9% 24.2% 21U 14.1% 14.5% 17.5% 11.1% 3rd 8.9% 8.7% 10.4% 7.2% 4Xi 10.4% 7.6% 12.1% 6.8% 5thl 6.0% 6.6% 6.7% 5.7% 6th 0 4.7% 2.8% 2.5% 5.3% |Total 11.5% 12.7% 13.9% 10.3% Source: LSMS 1998. 32. Repetition rates among the poor. Repetition rates for poor children in primary school are almost double the rates of the non-poor (Table 8). For secondary school, urban areas show similar repetition rates, except among the extreme poor, where the data is so scant as to be inconclusive. In rural areas, the secondary repetition rate among the extreme poor is almost double that for the rest of the poor and non- poor. Table A13.8 - Repetition rates by poverty level, 1998 Primary Secondary Repetition Repetition (%) (%) Total Extreme Poor 14.7 8.4 Poor 13.5 8.4 Non-poor 8.7 8.8 Urban Extreme Poor 13.9 Poor 14.5 8.7 Non-poor 8.3 9.2 Rural Extreme Poor 15.0 15.8 Poor 12.9 8.0 Non-poor 9.9 7.0 Source: LSMS 1998 33. MECD figures for the first grade in 1997 indicate a 14.6% dropout rate in urban areas and a 30% rate in rural areas"2 in the first grade. These high rates underscore a methodological problem. According to the l Gargiulo and Crouch, op. cit. p. 6. 12 The LSMS does not allow for an exact determination of dropouts or the dropout rate. Annex 13, Page, 12 MECD definition, a child is a dropout if he/she withdraws from school before the end of the school vear. However, if that child returns to school the following year and enrolls in thie same grade, he/she is not considered a repeater, although the child is using the same resources twice. As a consequence, the dropout rate tends to be overestimated. Factors Associated with School Enrollment 34. The impact of economic need on children not attending school is very clear. It remains to be seen if economic need has any impact on those already attending school. Econometric estimates by KrugeH indicate that the strongest factor affecting school attendance is the educational level of the mother. Just having a literate mother increases the probability of a child attending school by 20%. The econometric results also indicate that poor children are less likely to attend school than non-poor children. Moreover. the higher the level of family expenditures, the larger the probability of attending. Family wealth also has a similar effect. 35. Urban or rural location does not seem to have an impact on the probability of attending school, but the age of siblings does. Children in families with younger siblings (between I and 5 years of age) have a lower probability of attending school, suggesting that the use of siblings for childcare is a deterrent to school attendance. Moreover, this finding indicates that the provision of preschool education could have a significant positive effect on school attendance by other siblings. Girls have a higher probabilitv of attending school than boys, perhaps because of the opportunity cost of a boy's labor for the family, and distance to school has a small negative effect on attendance. 36. The proportion of children not attending school has decreased significantly for primary school. and remained relatively unchanged for secondary school during the last five years. In 1993. 21.5% of 7-1: year olds did not attend school, while in 1998 this proportion dropped to 14.3%. For 13-18 year olds. the proportion not attending school actually increased by 0.5 of a percentage point between 1993 and 1998 (Table 9). Most of the increase in the proportion of children not attending school was observed in urban areas, particularly in the Pacific-Urban, Central-Urban, and Atlantic-Urban regions. In rural areas, only the Atlantic-Rural region shows an increase in the proportion of children not attending school during the 1993-1998 period. Table A13.9 - Percentage of children not attending school 1993 1998 7 -12 yrs 13 -18yrs 7-18yrs 7 - 12 yrs 13 -18 yrs 7 - 18 yrs All 21.5 43.0 31.0 14.3 43.5 28.2 Extreme Poor 44.7 71.8 55.7 32.4 70.0 49.1 Poor 32.3 57.6 43.0 21.3 60.7 39.1 Non-poor 7.4 26.9 16.5 6.1 26.7 16.4 Urban 11.3 26.6 18.3 9.0 28.1 18.3 Extreme Poor 35.3 57.7 44.5 28.5 61.7 42.9 Poor 19.9 37.6 27.8 15.9 48.5 30.3 Non-poor 6.3 20.8 13.1 5.1 19.4 12.4 Rural 33.7 65.6 47.2 20.1 61.9 39.5 Extreme Poor 47.3 75.8 58.8 33.6 72.6 51.1 Poor 39.0 69.9 51.8 24.1 66.7 43.7 Non-poor 11.7 51.2 30.0 8.7 49.2 28.1 Source: LSMS 1993 and 1998 13 Kruger, Diana, 2000. "A Model for School Attendance." Background paper for the Nicaragua Poverty Assessment. World Bank, Washington D.C. Annex 13, Page, 13 37. Overwhelmingly, the majority of children between 7 and 12 years of age cited economic problems as a reason for not attending school. Among extremely poor children in urban areas, almost 85% do not attend for this reason. In rural areas, almost one-half of extremely poor children do not attend school for economic reasons. The difference between urban and rural areas is due mostly to the longer distances to schools in rural areas, a significant barrier to school attendance for about a quarter of all children in rural areas, regardless of poverty level (Table 10). Although the questions asked in 1993 were not identical, the same themes were cited as deterrents to school attendance. The issue of distance to school is actually an issue of personal safety (14% among boys and 21% among girls). Rural families are reluctant to send tiheir young children, especially the girls, over long distances through scarcely populated areas because of the high level of citizen insecurity in the country. Table A13.10 - Reason for not attending school among, 7-12 year olds, 1998 Extreme Poor Poor Non-poor Urban Rural Urban Rural Urban Rural Total Male Age 0.0 2.7 0.0 4.1 0.0 0.0 2.4 Economic Problems 83.7 48.0 75.1 43.7 65.4 23.4 51.9 Rural Activities 0.0 5.7 1.1 7.0 5.7 3.7 5A4 Domestic Duties 0.0 0.0 1.8 0.6 0.0 0.0 0.7 Not interested 10.3 6.3 8.0 6.6 4.7 0.0 6.1 Distance 0.0 20.3 1.4 20.8 0.0 25.2 14.0 Illness 3.7 1.6 7.9 3.5 3.9 0.0 4.1 Not available 0.0 0.0 0.0 0.0 4.2 0.7 0.7 Level not offered 0.0 1.4 0.0 1.1 0.0 0.0 0.6 Insufficient Teachers 0.0 7.9 0.5 5.7 0.0 13.2 4.3 Insufficient Security 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Insufficient Textbooks 0.0 . 0.8 0.0 0.5 0.0 0.0 0.3 Handicapped 0.0 0.6 3.1 1.0 13.0 0.0 3.2 Other 2.3 4.7 1.2 5.5 3.2 33.8 6.3 Female Age 1.7 3.3 0.9 2.4 18.4 0.0 2.9 Economic Problems 84.9 48.9 68.4 42.2 31.1 23.3 45.9 Rural Activities 0.0 1.9 0.0 1.1 0.0 0.0 0.7 Domestic Duties 2.1 4.1 2.7 2.5 10.0 2.7 3.0 Not interested 9.2 3.9 16.2 6.9 12.6 0.0 8.8 Distance 2.2 26.2 1.1 27.2 0.0 34.1 21.3 Illness 0.0 1.0 3.8 0.6 18.6 0.0 2.5 Not available 0.0 0.0 1.1 0.0 0.0 0.0 0.2 Level not offered 0.0 1.6 0.0 1.3 0.0 0.0 0.8 Insufficient Teachers 0.0 3.8 0.0 5.6 0.0 6.2 4.0 Insufficient Security 0.0 1.0 0.0 1.1 0.0 0.0 0.7 Insufficient Textbooks 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Handicapped 0.0 1.6 4.4 2.1 0.0 5.7 2.8 Other 0.0 2.7 1I.5 5.0 9.4 28.1 6.4 - Source: LSMS 1998 38. To reduce the negative impact of the economic and work needs of students on primary school enrolment, the Government has initiated a pilot project as part of its Social Safety Net. The project will give families in extreme poverty with eligible children a monthly stipend to help them defray the cost of education, and also help them improve their nutrition. It is expected that each participating child would represent about US$10/month in cash transfers to extremely poor households. This amount would cover Annex 13, Page. 14 all household expenditures in education per eligible child, with some left over for household use. The pilot project has begun activities in six municipalities and it is expected to begin disbursements. However, future expansion will depend on the availability of donor funding. EDUCATION EXPENDITURES AND POVERTY 39. The evidence from the review of the sector's performance clearly indicates a strong negative lInk between poverty, enrollment, and performance. Although public schooling is free, parents are asked to contribute with some fees -which are supposed to be optional- and to cover the cost of registration, uniforms, school snacks, and transportation. Overall, the average annual expenditure (excluding food at school) per primary student per family was C$438 (approx. US$41), ranging from C$240 (approx. US$23) per student per year in rural areas, to C$600 (approx. US$57) in urban areas (Table 11). Considering that in 1998 the Government spent approximately US$41 per year per primary school student and US$32 per year per secondary school student, the overall average family expenditures represent about half the total cost of sending a child to a public school'4. Clearly, such a cost sharing arrangement ha: a negative impact on enrollment, as evidenced by the overwhelming proportion of those who do not attend, citing economic need as the main cause for non-attendance. 'Table A13.11 - Annual Education Expenditures 1998- Primary School Unweighted Total in Cordobas (excludes food at school)' All 438.4 Rural 239.8 Urban 600.5 Total Extreme Poor 81.8 Poor 147.3 Non-poor 746.8 Urban Extreme Poor 125.5 Poor 206.9 Non-poor 811.4 Rural Extreme Poor 65.8 Poor 111.8 Non-poor 564.6 Source: LSMS 1998. (a) Includes fee, tuition, registration and "other" but excludes food at school (annualized monthly expenditures using 8.5). Other include: transportation, school lunch, uniforms, school materials, books and other school-related expenditures. (b) The unweighted total refers to average expenditures without using expansion factors. (c) Averages are only for those responding with a positive value for expenditures. 40. In general, urban families spend at least twice the amount spent by rural families. In urban areas, the amount spent by the non-poor is six times the amount spent by the extreme poor. Within rural regions non-poor families spend almost nine times the amount spent by extremely poor families. 41. In secondary school the same pattern of expenditures emerges. However, the amount spent by all families, urban and rural, does not differ by much (Table 12). Overall, the average annual expenditure per secondary student is C$ 1162 (US$109), 3.4 times the US$32 per student spent by the Government in 14 Details on Government per-student expenditures are covered in the next section. Annex 13, Page, 15 1998. Even among extremely poor families, per student expenditures were C$554 (US$52) or 1.6 times the amount spent by the Government. This expenditure pattern helps explain why school enrollment drops precipitously after children finish primary school. Table A13.12 - Annual Education Expenditures - Secondary School Unweighted Total in Cordobas (excludes food at school)' All 1161.8 Rural 999.2 Urban 1212.9 Total Extreme Poor 554.1 Poor 628.4 Non-poor 1318.3 Urban Extreme Poor 594.6 Poor 593.2 Non-poor 1338.2 Rural Extreme Poor 518.9 Poor 673.9 Non-poor 1228.5 Source: LSMS 1998. (a) Includes fee, tuition, registration and "other" but excludes food at school (annualized monthly expenditures using 8.5). Other include: transportation, school lunch, uniforms, school materials, books and other school-related expenditures. (b) The unweighted total refers to average expenditures without using expansion factors. (c) Averages are only for those responding with a positive value for expenditures. 42. The biggest educational expenditure made by families is for school uniforms, accounting for 40% of annual per student expenditures (Table 13). Nearly 80% of per student expenditures are for bus transportation, uniforms, and school supplies. Annex 13, Page, 16 Table A13.13 - Shares of education expenses by educational level and poverty group, public schools only, 1998 Percent of household expenses on education Educational Share School TuitionlRegistratI School Unifor School Text- Total Total Hh Total per level & of net fees ion bus/food ms Suppliesl hooks education student povertv group Enroll expenses education ment as a share expenses ;1s of Hh a share cf non-food per-capita expenses non-foot: l_____________________________ _ lexpense Primary All 100 9.3 0.2 1.2 21.4 40.6 23.3 i3.0 100.0 5.3 18.8 Poor 58 7.7 0.3 2.1 17.1 :42.8 26.5 3.5 100.0 6.3 21.9 Non-poor 42 10.3 0.1 ,2.2 24.2 39.3 '21.3 2.6 100.0 4.3 T4.7 Secondary All 1() 16.3 0.3 3.7 :32.0 25.9 118.3 3.5 10((.( 6.7 29.8 Poor 30 19.8 10.3 4.7 21.8 28.7 21.6 3.1 100.0 7.9 43.2 Non-poor 70 15.5 10.3 3.5 34.3 25.2 1 7.6 3.6 100.0 6.2 24.8 Higher* 100 16.1 !0.1 113.6 47.0 20 10.1 11.1 1 00.() 7.8 39.1 Poor 11 12.9 0.0 114.3 49.8 2.1 111.9 9.0 100.0 10.8 72.1 Non-poor 89 16.3 0.2 1l3.5 46.7 2.0 10.0 11.3 100.0 74 35.1 Source: LSMS 1998. * Includes university and technical education. Budget shares are for households with at least one child in school. 43. One of the most politically sensitive issues in school autonomy is the payment of voluntary fees. These are raising teacher salaries and improving school facilities. Judging from news reports and public opinion, fees are very unpopular. The results of the LSMS 98 appear to indicate that among the poor. fees represent almost 10% of total expenditures for primary students. For secondary students, fees represent 20% of total expenditures. Although both are significant expenditures, their magnitude is not proportional to the level of dissatisfaction expressed by parents. This suggests that fees may be unpopular because they were unexpected and their use is not well accounted for by many schools. At the primary school level approximately half the parents interviewed indicated that fees were not voluntary, as did almost three- quarters of parents of secondary students. 44. The cost of public education for poor families is very significant. The total cost of sending a child to a primary public school represents about 20% of the per capita non-food expenditures of poor households. For secondary students, the education share of per capita non-food expenditures is 43% in poor households and 24% among non-poor households. The higher education share of per capita non-food expenditures among poor households is even larger: 72%. Given this scenario, it is pertinent to ask if public expenditures in education are pro-poor. This topic is examined later in this section. The Returns from Education 45. The issue of private costs in public education is important because even when education is a good investment- the poor lack the collateral necessary to find investors willing to lend them the money to pay for education. Therefore, it is important to look at the returns from education in Nicaragua to determine if the main issue is one of profitability or collateral. Clearly, if the issue were to be one of' Annex 13, Page, 17 profitability, the poor would behave as a rational consumer and not invest in education. However, if education proved to be profitable, then collateral for the poor becomes a matter of public policy. 46. The results of the LSMS indicate that the private return from primary education is 7.3% and for secondary education it is close to 10% (Table 14). For university education the private return is close to 12%. Women receive a substantially lower return than men, perhaps because of wage discrimination, reduced opportunities for advancement, or the impact of the childbearing years on the accumulation of work experience. Still, if the poor could finance their education and the opportunity cost implied by the foregone work of children, their educational achievement would increase. ITable A13.14 - Rates of Return from Education: Private and Social| Dependent Variable: Log hourl/ wages I l Males l Females Urban Rural All IPrivate I I Rates of Return: l I Averaae j 8.6% 7.7% 8.4% 1 7.9% 1 8.1 % Primarv 6.3% 3.2% 8.1% 3.9% 0 5.2% 8.9% 12.5% 10.3% 12.6% 10.1% University 11.9% 24.6% 19.2% 15.8% 17.3% Social Rates' of Return: _ _ _ _ Average 8.0% 7.0% 7.8% 7. F % 7.5% Primary 6.1% 3.1% 7.8% 3.7% 5.0% ! Secondary 8.7% T 12.2% 10.1% 12.3% 9.9% University 9.2% 15.3% 14.9% 9.9% [ 12.7% | (1)F University 9.8% 16.9% 15.8% 10.9% 13.6% (2 ) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ liniversity 10.4% 18.9% 1 16.8% - 12.2%. 14.6% (1) The costs per university student are based on the assumption of 15,000 full time equivalent students. This is the most conservative estimate. (2) and (3) are estimated with assumptions of 20,000 and 30,000 full time equivalent students, respectively. Source: Kruger, 200015. 47. The results also show that rural primary education does have the highest private return within basic education ( 13.6%) and that rural secondary education has a very low private return (5.2%). This low rate of return may help explain the low secondary enrollment of rural youth. In both urban and rural areas, private returns from university education are highest: between 11.9% for males and 24.6% for females. Private returns from university education in rural areas are about 19%. With these rates of return, rural dwellers enrolling in secondary education may do so with the expectation that it is only a stepping stone to university enrollment. 48. For the nation, the returns from education are also positive. Because public investment in education is relatively low (with parents covering almost half the total cost), the social rate of return is almost as high as the private rate of return. However, because economic need is cited by the non-poor almost as often as by the poor, targeting education subsidies must be done with extreme care. If, for example, there were scholarship programs for those in need, the economic need of the family must be evaluated in terms of their per capita expenditures and poverty level, and not on their stated needs. Relying only on the latter 15 Kruger, Diana, 2000. "Rates of Return to Education in Nicaragua." Background paper for the Nicaragua Poverty Assessment, World Bank, Washington D.C. Annex 13. Page, 18 is highly subjective, since it confuses the family's own assessment of economic need with the quality of schooling. which may be wortlh the family's effort. PUBLIC FINANCING OF EDUCATION 49. Total funding for the MECD, and for education in general, has increased in dollar terms since 1 996 (Tablel5). The increase was particularly noticeable in 1999, when the MECD budget went from a to-al of US$74 million to US$101 million, an increase of 36% from one year to the next, most of it going to higher teacher salaries. For 2000 the MECD has submitted a budget of US$108 million, an increase of 7% over the 1999 budget. 50. Between 1993 and 1998 the MECD was able to spend between US$12 and US$15 million per year on activities funded by external donors. These funds increased to US$25 million for 1999. These amounts represent between 20% and 25% of the MECD budget. In addition, donor-funded infrastructure implemented by the FISE adds another US$15 million a year to the total amount spent on education. As a consequence, donor funding for recurrent and capital investment has represented around 50% of Nicaragua's total funding expenditures in education. Table A13.15 - MECD budget 1991-2000 MECD Total 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Expenditures I MECD (CSmillions) 242 274 298 416 433 515 678 781 1.186 1413 Exchange rate 4.50 5.00 6.12 6.72 7.53 8.44 9.45 10.58 11.77 13 01 jMECD ('OOOs of USS) 54 55 49 62 57 61 72 74 1101 10 Source: Ministry of the Treasury and Public Credit. Exchange rate: Central Bank. Does not include FISE expenditures in education infrastructure. 5 1. Although budget allocations to education have increased, funding has been relatively stable as a share of the Gross Domestic Product (GDP). As shown in Figure 3, Nicaragua allocated between 4.3%"o and 5% of the GDP to education over the past decade, except in 1999, when it rose to 6.1 %. This share of the GDP is greater than the share allocated by other countries in the region16, such as Guatemala (1.7`%o); El Salvador (2.6%), Honduras (4.1 %) and the Dominican Republic (2.6%). 52. The education sector in Nicaragua has also fared well within the social sector in the last few years, showing an increased proportional allocation of resources compared to health and other social expenditures (Figure 4). Since 1993, each education sub-sector has received significant budgetary allocations as a share of social sector expenditures. This trend shows that the Government is making a substantial effort to improve the per capita allocation to public education. 16 This information comes from the following sources: Guatemala (1996): UNESCO Statistical Yearbook, 1999; El Salvador: estimated from data from the Central Reserve Bank of El Salvador, 2000; Dominican Republic: estimated from data from the Banco Central, 2000 and National Budget Office, 2000; and Honduras: Policy Analysis Unit (UDAPE), 1997. Honduras 2005: Construyendo nuestro progreso. Propuesta. Tegucigalpa. Annex 13, Page, 19 Fig. A13.4 - Education as a percentage of GDP 7.00% 6.00% 5.00% -... - 4.00% ___ ____ _ 3.00% _*__ _ - _ _ 1.00% _ _ _ _ _ _ __ _ _ _ _ 0.00% 1991 1992 1993 1994 1995 1996 1997 1998 1999 _-. MECD Budget -_- Universities All educational institutions /a Source: Ministry of the Treasurv and Public Credit. Fig. A13.5 - Education as a percent of total social expenditures 60.0% 50.0% _ 40.0% - 30.0% * 20.0% 1 0.0%. 0.0% Primary ed. Secondary MECD Universities All ed. Budget educational institutions la 3 1993 j 1998 Source: Ministry of the Treasury and Public Credit. 53. Because the per capita GDP of neighboring countries is much larger than Nicaragua's, actual expenditures'7 per student still remain low. In 1998 the Government spent US$41 per primary school student and US$32 per secondary school student. In 1999 these amounts improved significantly: US$58 per primary school student, and US$36 per secondary school student'8 (Figure 5). These expenditures refer only to recurrent items, such as salaries and materials, and do not include capital expenditures, such as those made for school buildings. In comparison, recurrent expenditures in El Salvador and the Dominican Republic, with a much larger per capita GDP than Nicaragua, hover around US$130 per primary school student and about US$100 per secondary school student. 54. Per student recurrent expenditures have varied significantly by level (Fig. 6). Per student recurrent expenditures for primary school have increased steadily, going from US$40 in 1994 to US$58 in 1999. For secondary school the trend has been the opposite, dropping from US$40 in 1994 to US$36 in 1999. 17 Expenditures for Nicaragua only include MECD budgetary resources. " Technical Secretariat for Presidency, 1999. Supplementary Social Fund (FSS). Protecting priority social services. Managua, pp. 8-9. Annex 13, Page. 20 Although total secondary recurrent expenditures have increased, an above-average increase in secondary enrollment has caused a reduction in per student recurrent expenditures. So far, there is no clear rea ,on why the rate of enrolLment in secondary school has been much higher than the primary school rate. Fig. A13.6 - Trends in recurrent expenditures per student by level (in US$) 140 +__ _ _ _ _ _ _ _ _ 120 _ education - Primary education 100 _ _ 8o ___ Secondary l education VI60 . Pre-school. primarv & 40 ____ secondarx i -MECD Budget divided by all 20 s _ _ __ _ students _s s _ -_. N, , N \q '11 'q \q -I, Source: Ministry of the Treasury and Public Credit. 55. Recurrent public expenditures in education 1999 represented 61% of total sector expenditures. The remaining 39% accounted for public investment in infrastructure, most, but not all of which is covered by external funds administered by the FISE, the body in charge of rehabilitating and expanding education infrastructure in poor areas. These expenditures are still lower than those in the rest of the region, blut they are increasing faster than other social sector expenditures. Overall, the FISE allocates 60% of its financial resources to the education sector. From June 1998 to June 1999, the FISE invested US$27.4 million in the education sector. This amount represents an additional 21% on top of the total amount budgeted by the MECD, resulting in per student expenditures of US$70.18 for primary school and US$43.56 for secondary school. If the cost of the infrastructure financed by all external donors is included, total expenditures per primary and secondary student are approximately US$91 and US$59 respectively. 56. Education Fundingfrom External Sources. A significant portion of Nicaragua's public expenditures in education comes from external sources. Some of these funds come from large basic education projects (i.e.: APRENDE I and II with World Bank funding; and BASE I and II with USAID funding), secondary education, and adult education. A summary of these projects and their sources of funding are presented in Table 16. 57. A temporary source of additional external funds is the Supplementary Social Fund (SSF), created for the purpose of obtaining bridge funding for the social sector to increase recurrent expenditures to the levels proposed during the Consultative Group Meetings in Geneva in 1998 and in Stockholm in 1949. Operating within the macroeconomic restrictions of the International Monetary Fund adjustment program, the SSF has channeled about US$12 million to recurrent expenditures for school autonomy and for the provision of a school backpack to students living in poverty, increasing per student recurrent expenditures from US$46 to US$58 in 1999, 21% more than in 1998. The SSF will operate for three years, until the end of 2001. At that time Nicaragua may start receiving the benefits of the initiative for the Heavily Indebted Poor Countries (HIPC). After the three-year period the Government is supposed to increase education funding and replace the funds administered by the SSF with national funds. Annex 13, Page, 21 Table A 13.16 - External funding sources for the MECD Project Objective Funding Agency Project Funding Improve coverage and quality in World Bank APRENDE I. (closing Total funding: US$52.5 preschool and primary education in rural date: Dec 31, 99), and million areas. and increase the institutional APRENDE 11 (closing Funding Balance: capacity of the Ministrv of Education. date: Dec. 2002) US$52.5 million Improve school infrastructure IDB. StDA. KfW. World FISEIL FISEII; FISE 111t. Total funding for 1999: Bank Ongoing. US$27 million. Funding Balance: US$3 million. Improve the qualitv of basic education USAID Basic Education Project Total funding: US$16.9 through textbooks, teacher training. and (BASE I and 11). Closing million improved institutional management. date of phase II: Funding September 2003 Balance:US$16.4 million Adult education coverage and Spanish Cooperation. Nicaraguan Literacy and Total funding:US$4.5 effectiveness. Government of Spain. Basic Adult Education million Program (PAEBANIC). Funding Balance: US$1.8 Closing date: 2003. million Increase student retention: provision of European Union Support to the Total funding:US$18.2 school infrastructure and school Nicaraguan Educative million backpacks and materials. Sector (ASEN). Closing Funding Balance: date: 2001. US$14.2 million Improve the quality of pre-service Luxembourg Teaching Colleges I and . Total funding:USS3.2 teacher training. 11. Closing date: 2000 million Funding Balance: US$0.25 million Need to increase student nutrition and World Food Program Integral Program for Total funding: US$14.4 retention. School Nutrition (PINE). million Closing date: Dec. 2000 Funding Balance: US$4.7 million Education reform. IDB Distance education: Total funding: US$9.4 preschool education million Funding Balance: US$9.4 million The incidence of public educational expenditureP' 58. On average, at the primary level the poorest three income quintiles of the population receive the largest share of public education expenditures (Fig.7). At the public primary school level enrollment comprises mostly poorer children, since families with higher income tend to send their children to private schools. At the secondary level, however, this pattern is reversed slightly, with the children from middle and upper income quintiles accounting for most of the enrollment in public schools. Clearly, poorer children have lower relative enrollment rates at this level. Other factors that help to explain the larger share of public education funding for the poor is that in primary school enrollment is thrice that for secondary school. This means that poorer children capture most of the funding and that government expenditures per primary school student are higher than those for secondary school students. '9 This section is taken from Dayton, Julie M., 2000. "The Incidence of Public Expenditures on Health and Education in Nicaragua, 1998." Background paper for the Nicaragua Poverty Assessment. World Bank, Washington D.C. Annex 13, Page, 22 Fig. A13.7 - Per capita government expenditure by level of education and income quintile, 1998 120 '; 100 _ __ __ < 80 _ - _ _ < . . " Secondary D. 60 _ _Primary 0 40__ .) 20 Quintile Source: Dayton. 2000. 59. Within secondary school the poorest quintile captures only 5 c6rdobas of Government spendino while the richest 40% of the population captures four times that. In summary, government spending at the primary level is very progressive, since it benefits the poor disproportionately. At the secondary education level government expenditures are somewhat regressive. The implications of the incidence of public educational expenditures are clear. During the past decade the Government has been successful in channeling funds to primary education, with consequent benefits for the poor, to the point where per student expenditures for primary school now exceed per student expenditures in secondary school. Although this policy makes sense for the poor, government expenditures in secondary school are mostly captured by the non-poor, since secondary school enrollment is positively associated with income. THE IMPACT OF SCHOOLING ON THE RURAL POOR 60. Nicaragua's social sector policy is based on the reactivation of agricultural production in rural areas, investing in the human capital of the poor, and providing a social safety net for the extreme poor. Within this framework, the Government has been trying to define what kind of education policy is more appropriate for rural areas, especially rural youth. 61. The analysis of LSMS 1998 data shows that per capita rural consumption is largely determined by access to land, education, access to production finance, and access to the economic infrastructur&0. In terms of impact, education is the second most important determinant of household consumption leve Is. Moreover, with higher levels of education, there are increasing levels of household consumption. The results also indicate that the relative impact of education is greater for poorer households, suggesting that the completion of primary education in rural areas would have important positive consumption effects among the rural poor. 62. Is a specialized curriculum necessary for rural areas to accelerate educational attainment and agricultural production? The experience in Nicaragua and other Latin American countries suggests that a specialized program is not necessary21. A review of experiences with different types of rural education in Nicaragua and Latin America indicates that investing in primary education produces better results than 20 Davis, Banjamin, and Rinku Murgai, 2000. "Between Prosperity and Poverty: Rural Households in Nicaragua." Background paper for the Nicaragua Poverty Assessment. World Bank, Washington D.C. 21 Arcia, Gustavo and Vanessa Castro, 1999. "Experiencias en educaci6n rural: ,que lecciones puede aprender Nicaragua?" Background paper for the Nicaragua Poverty Assessment. World Bank, Washington D.C. Annex 13, Page, 23 investing in vocational education, agricultural education, or specialized rural training. The most effective innovations in rural education have tended to increase coverage and improve basic cognitive skills (mathematics, reading, writing, and logical thinking). 63. Distance learning targeted at rural areas -through the use of video or radio- are interesting innovations that have had success in Mexico (Tele-secondary) and in Nicaragua (Radio Education). However, it can be argued that their success can be extended to urban areas as well. The use of video is particularly important to take into account because it seems to improve academic achievement significantly, using well-planned video segments and being executed by good teachers. Given the cost of video production (or the cost of adapting existing material) it is to the country's advantage to extend coverage to urban areas and reduce costs by taking advantage of the implicit economies of scale. 64. Past experience with agricultural education and vocational education in Nicaragua indicates that coverage is a problem. Given the high cost per student, the total investment needed is very high -which does not bode well given the limited resources of the sector- and its impact is not that great, since the acquisition of specialized skills by students does not guarantee they will stay in rural areas. On the other hand, experiences with improving the quality of basic rural education (such as Escuela Nueva in Colombia) show better impacts at a lower cost. CONCLUSIONS AND POLICY RECOMMENDATIONS 65. Nicaraguas public education sector has made remarkable progress in improving primary education among the poor. Additional progress has been made in the area of scholastic achievement in rural areas and the reduction of in rural areas. Nevertheless, questions of education quality and access still present large problems.. 66. Supply side issues. In the area of quality there is a pressing need to measure student achievement and to report this measure in order so that there is more awareness of geographical or group priorities and to complement the climate of local accountability framed by the school autonomy program. Along the same line, the MECD must make efforts to expand the scope and quality of training of the school autonomy program so that autonomy is more accessible to the poor. Also, the MECD must begin to look at the issue of teacher quality and begin a process of measuring teacher quality, however imperfect, in order to enable a monitoring of the effectiveness of its methods for training, selecting, and rewarding teachers. 67. Demand side issues. In terms of access, a look should be taken at the gaps in primary grades in rural areas, where a large number of schools do not offer a program that goes beyond the 3.d or the 4"' grade. Although the MECD has expressed interest in reinforcing the quality of multi-grade programs and multi- grade teachers, there is a need for more resources to be implemented more broadly. Last but not least, the MECD must conduct a review of its policy regarding school uniforms -apparently a substantial expenditure- thereby becoming a yet another barrier to increased enrollment. 68. The following matrix (Table 17) summarizes the main problems associated with the supply and demand for public education, and the possible solutions that could be implemented by the MECD Annex 13, Page, 24 Table A13.17 - Summary of Supply and Demand Problems and Strategies in Public Education Supply Problems Main Determinants Proposed Strategy Low access. especially beyond the * Low per student expenditures in a Expand multi-grade teacher 4th grade in rural areas, and teachers and materials funding inadequate infrastructure * High proportion of defective a Increase budget for materials infrastructure a Continue FISE work in * High population growth in rural education areas Low teacher and teaching qualities * Low salaries * Create a new a Low incentives tied to salary/perforrnance career track perfornance U Improve leadership and * Lack of a teacher evaluation, managerial training of schcol selection, and retention system directors * New curricula still not * Improve quality and funding of operational pre-service teacher training to attract better entrants Low managerial and technical * Central MECD still * Use school autonomy as the capacity of MECD "centralized" pivot for Central MECD re rorm * Weak connection between the * Consolidate use of financial, MECD and school autonomy statistical, and management * Low managerial and technical inforrnation systems capacity at municipal and school * Establish staff performance levels measures and incentives * Low capacity for education policy analysis in the MECD Demand Problems Main Determinants Proposed Strategy High dropout rates above the 4'n * Need to work n Scholarship -programs foT grade: high non-attendance * Inadequate curricula children in extremely poor * Low academic achievement households * Educational expenses * Provide of school materials * Consolidate new curricula * Broaden school food progrims Low parent participation * School autonomy program still a Strengthen school autonomy weak training at local level * Institutional inefficiency a Implement system of academic * Weak family structures after a achievement measurement and decade of conflict reporting * Strengthen leadership training of school Principals 69. More specific recommendations: i. More training in "SchoolAutonomy." The step and the scope of training among parents, Principals, teachers, and students has to pick up. Although the APRENDE project has allocated significant resources to training, these do not cover secondary schools. Furthermore, training, should be continuous, in line with its being based on the concepts of local governance and accountability, both of which are new to education stakeholders. ii. Invest in school Principals. School autonomy calls for leadership and managerial ability. School Principals are the main link between parents and teachers and can make... or break... school autonomy. So far, autonomy works well so long as it is complemented by a good Principal. Inversely, it does not work well should the Principal lack leadership qualities. Therein, the MECD should set itself as one target, among others, a faster process in order to Annex 13. Page, 25 identify and training teachers who possess with leadership qualities more suitable for the position of Principal. For this: a. Train Principals in management and communication so as to provide them with tools for working with parents, b. Maintain and reinforce the current process in use for selecting School Principals, in such a way as to have it so schools wouldn't be affected bv politics or favoritism. iii. Revise school systenm supervision. Support from MECD's Municipal Delegations can be very important for autonomy to function properly. The APRENDE project allocated funds for a review of the current system, seeking to make it over into a mid-level office with technical and managerial expertise accessible to School Directors, by way of helping them to boost performance. 70. Regarding the long-term plan and current MECD efforts in Education Reform, the following is recommended: a. Place less emphasis on adult literacy. The MECD must be careful not to assign too much funding to literacy work, meanwhile neglecting funding for primary schooling. Should too much attention be paid adult literacy work, the MECD could run the risk having illiteracy rise among the young. only for these youths to be illiterate adults later on. b. Revise the policy on vocational education. Currently, vocational education is handled by INATEC. This government agency has been criticized for consuming a disproportionate amount of resources relative to results. More and more, the Government has been considering the privatization of INATEC and the reduction or elimination of the 2% tax surcharge that formal businesses pay, ostensibly to support vocational education. INATEC works independently of the MECD, so this matter should be looked into carefully so as to have a balanced allocation of funds in the near future. c. Begin to pay more attention to secondary education. For almost 10 years, bilateral and multi- lateral donors have supported additional funding for primary school. This still makes sense, since only half the children that enter 1 grade, finish 6h . However, those able to finish primary school five years ago are now faced with entry into low-quality, low-access secondary schooling. d. Reexamine Higher Education. Currently, 6% of the national budget is allocated Constitutionally to higher education. This thorny political issue might be dealt with in two parts. The first part might relate to expanding the offer of post-secondary options for 2ndary graduates (e.g.: junior colleges, tech-schools), The second might have to do with control of 6% funding. The second may also have to do with the use made of that funding. Options to grant Merit Scholarships and develop education credit markets and a student voucher system could be explored. e. Implement the measurement and reporting of student achievement. The MECD's strategic plan wholeheartedly supports the establishment of a system for measuring and reporting academic achievement. This activity should be of utmost priority, since it gives substantial leverage for other reforms proposed by the plan, such as improved teacher training and the reinforcement of school councils. Annex 13, Page, 26 f. Annex Table Al. Principles, Policies and Strategies of the National Education Strategy Principles Policies Strategies 1. Education is an obligation of the State The State will promote the demand 1. Expand and maintain literacy proL:rams and a fundamental human right. This for education and it will diversifv and projects. implies universal access and opportunity to educational services to fulfill 2. Expand education coverage. espe.ially in all. demand needs. rural areas. all the way to sixth grad . 3. Improve the quality and relevance of education. 2. Education fosters social. environmental. The State, in collaboration with civil I . Reinforce the formation of values and the ethical, civic, human, and cultural values society, must find the resources to participation of families and the communitv. that shape national identity and respect of guarantee an integral education 2. Transform schools into centers fot the religious, political, ethnic, cultural. leading to human development. practice of democracy'. psychological. and gender diversity. 3. Education, a fundamental investment, Orient education programs towards I . Strengthen and expand technical and seeks to achieve human. economic. forming human capital and vocational education. scientific, cultural, and technological increasing available and equitable 2. Reform the secondarv school curriculum development resources assigned to them. so that it relates to the job market and Higher Education. 3. Reform Higher Education. 4. Promote the idea that education funding should be 6% of the GDP 4. Parents and civil society have the right Create favorable conditions for the I. Strengthen parent's councils. student and obligation to participate in the process participation of parents and society govemment, and local organizations so they of their children's education. in education. can foster their own involvement in education. 2. Consolidate academic and administrative decentralization and school autononry. 5. Teachers are keys to the education State and Civil Society will promote I. Update teacher's skills with in-service process; they have a right to working and teacher well-being, improving training at Teaching Colleges. living conditions in accordance with their professional standards, salaries, and 2. Reform the Law for the Teaching Career role. working and living conditions. to improve teacher's welfare. 3. Implement projects for monetary incentives for teachers. 6. Students are in control of their own Organize curricula, learning I . Improve and expand school infrastructure learning. methods, teaching materials and so it is more conducive to learning. forms of evaluation to better assist 2. Reform curricula to allow so there is more students in their learning process. student involvement in the learning process. 3. Implement educational standards that promote education quality and equity. 4. Implement a national system for measuring and evaluating academic achievement. 5. Implement an information system to assist students in an evaluation of career aptitudes. Provide textbooks and school materials to all primary students in public schools. 7. Education is an integrating process. All Promote development of an I. Create a commission for the planning and education sub-systems should be integrated system that articulates evaluation of curricula in the three educational articulated in order to form a single curricular and operational links subsystems. national system. among sub-systems. 8. Education must function under a system Promote the creation of 1. Modemize the entire education system. of administration and management that is organizational structures that link all 2. Develop an Information System that coordinated, decentralized, participatory, the participatory processes in the ensures an efficient administration and efficient, and transparent. development of education policy and management. accountability. Source: National Education Council, 1999. "National Strategy for Education," Managua Annex 14, Page I Annex 14 - Experiences in Rural Education: What Lesson can Nicaragua Learn? Bv Gustavo Arcia and Vanessa Castro INTRODUCTION I. The objective of this report is to review some experiences with rural education that are relevant to the Nicaraguan situation in order to help in defining alternatives for the design of programs focusing on rural youth. The purpose of the report is to provide a point of departure for the design of programs that help resolve the problem of students leaving school in rural zones, where public education is lacks quality and the relevance required by the student body. 2. Over these last two decades, Latin America has executed some innovations in the service of rural education. In general, these innovations seek to improve the quality and relevance of rural education and to make basic teaching universal, broadening the coverage of education. Studies made in the last 15 years indicate that the State, some international organizations, and local non- governmental organizations -at times acting in isolation- have been responsible for promoting these innovations in the field of rural education, including new modalities for delivery, curriculum adaptations, and new texts (Gajardo, 1988; Chesterfield, 1992). 3. Unfortunately. there is little systematized information about rural education and its impact on the education sector and even less on its role in the configuration of education reform. In part, this is because the evaluations made of these innovative experiences have not been sufficiently rigorous (Gajardo. op. cit.), and so -with the exception of some experiences like the New School in Colombia and EDUCO in El Salvador- there has been poor systematization of the lessons coming out of these rural education programs. 4. These limitations in mind, this report summarizes information about some projects considered relevant to the situation of Nicaragua, organizing it in such a way as to help the education authorities know about experiences in rural education in other countries so they can make decisions about programs that could be applied in the rural sector of the countrv. 5. The report is organized as follows. The second section gives a summary of three primary education programs oriented towards rural areas of El Salvador, Colombia, and Honduras. The third section deals with the theme of televised education, and the fourth section discusses vocational education. The fifth section gives a summary of the Nicaraguan experiences, including education by radio, which should be taken into account as important points of departure. The sixth and final section presents the conclusions and recommendations. Annex 14, Page 2 EXPERIENCE IN PRIMARY EDUCATION IN THE RURAL SECTOR: EL SALVADOR, COLOMBIA, AND HONDURAS 6. One way to deal with the low academic level in rural zones is by reinforcing the systems of basic education in rural areas, giving more pedagogical support and adapting the system to the circumstances of production. The general idea behind this approach is that a student needs a minimum base in mathematics, reading, and writing in order to be able to learn the practical aspects of rural living. It is also argued that learning the concepts and expertise associated with greater rural productivity depends on the initial command of that minimum base. Because of this, in some countries much more emphasis has been put on having greater access to that minimum base than on providing more specialized education. The EDUCO and New School programs are the most outstanding in this category, although the program in Honduras is also discussed because of the lessons it contains in regards to political support. The EDUCO program in El Salvador 7. EDUCO is an education program for rural areas established in 1990. This program introduced two innovations in Salvadoran primary education: community participation in school organization and affairs, and the use of pedagogical adaptations and improvements that are more appropriate for the rural setting. Community participation includes transferring the administration of the school service to its beneficiaries, giving control over the school to those who are most interested in the quality of the service, while pedagogical innovation includes substantial changes in the formation of teachers and in the training of parents. 8. Early in the decade, El Salvador had around 500,000 children outside of the education system. Most of these students lived in rural areas, the majority of them in the areas of conflict where governmental presence was minimal. This lack of access for such a large group of children was of great concern to the government, and so it decided to establish targets for the period from 1989 to 1999: an increase of 15% in the rate for pre-school enrollment, 7% for the first and second grades in order to have 90% coverage, and 14% for students older than that. Institutional aspects 9. The EDUCO program was organized as a government response to the efforts of parents in the areas of conflict in El Salvador, where because of the war, the presence of the State was weak. In these zones, the communities had organized themselves in order to contract, with their own funds, one or a number of teachers to give classes to the children in the locality. These experiences were very interesting to the Ministry of Education (MED). It saw that the teachers paid by the community demonstrated greater commitment to their work than those paid with public funds. With support from UNESCO, the MED made a study of the needs of the population, especially that located in the areas of conflict. This research was based first of all on the analysis of the self- managed education experiences in the war zones, and then it centered on three core areas of analysis: i) the analysis of the experiences in similar countries; ii) the analysis of direct interviews with students, teachers, and parents; and, iii) a Size Census, by means of which the most disadvantaged were located. 10. Based on the experience of the schools administered by the communities, and taking into account the needs of the population, the MED formalized the system found in the field by means of a support strategy that proposed the formal transfer of control over the school to the parents and the community and the transfer of operative decisions to the local level. Annex 14. Page 3 H1. This change in control over the school would attack the main problems found in the public system administered by the MED, among them: slowness in appointing teachers for the rural communities; a lack of commitment of the teachers to the needs of those receiving education and disrespect; and, the lack of involvement of parents in the process of educating their children. Similarly, the strategy would correct the excessive use of methods to teach by rote, the lack of national content related to the local reality, the excessive amount of material in the curriculum, and, in general, an education without ludic sense. Operative aspects 12. The instrumentation of the strategy was done through a pilot project that, after being expanded later, became EDUCO. It added value to the local efforts with the complete State financing of primary education and the provision of more teacher training and better teaching materials. Legally, the Ministry investigated and transformed the laws for education in order to be able to give legal back up to the direction and execution of EDUCO. Organizationally, the MED created a small National Coordinating Body, assisted by offices for regional coordination. The National Coordinating Body did not replace the already existing units, but rather complemented them. Thereby, the general directorates of the MED fulfilled their original functions, but adaptin. T them to the needs of EDUCO. 13. The Ministrv of Education finances the salaries of the EDUCO teachers and all labor benefits, such as the Christmas bonus, insurance, and a housing fund. The MED also finances the materials and expenses for the education facilities, providing the EDUCO schools with a legal reserve and labor fund. The MED is also responsible for giving training to the members of the community associations and the teachers, and for supervising and giving orientations to all participating school centers. 14. The Community Associations for Education (ACEs, in Spanish) are responsible for the administration of the EDUCO school. Parents who are selected at a Community General Assembly make up the ACEs. The principal functions of these associations are: i) obtain funds for education labor of benefit to the community; ii) contract the teachers and guarantee them social insurance, a Christmas bonus, and other rights; and, iii) administer the funds that the State transfers in order to guarantee that education is free of charge. Pedagogical aspects 15. In relation to pedagogy, EDUCO reinforces the schools by means of: * extra training for the teachers, in which the EDUCO teachers receive approximately 40 initial hours more than the rest of the teachers in the education system that are not in the program. Further, the teachers receive monthly modular-style training that covers themes like: zones of work in the classroom, student attendance, community visits, evolutionary characteristics of children, teaching materials, etc. * new study programs with appropriate teaching guides * schools for parents * follow up or supervision, according to the modular training program being carried out monthly Principal results of the EDUCO program 16. The principal results of the EDUCO program are summarized in Table 1. In general, EDUCO has managed to increase the attendance of teachers by 56% (from 3.2 to 5 days a week). The value of the class materials provided by the Government has increased from US$37 per section Annex 14. Page 4 to US$137.5, or by 3.7 times, and the net teaching time increased from 10 to 18.75 hours a week per section. The EDUCO teachers have to take into account the parents, who are in fact their bosses. This creates an environment that propitiates dialogue and -in terms of the agent/principal relationship- it helps to better define the expectations the parents have of the work of the teachers and vice versa. The increase in class hours and in the amount of teaching inputs has so far resulted in only a modest increase in academic results in mathematics, where the first and second grade students in the EDUCO schools show better results than the first and second grade students in traditional schools. In the other areas, like language, no significant difference has been noted between EDUCO and the traditional schools. The lack of improved academic results in the EDUCO schools' suggests that a sustained increase in learning would only happen after appropriately combining the pedagogical inputs with family participation and with the quality of the teacher. So far, EDUCO has managed to improve the first two actors, but it has no direct influence over the third factor, which takes years to form. On the other hand, the participation of the parents has resulted in greater political support to the MED and a better climate of governability, which are very important for the development of the rural sector of El Salvador. Table A14.1 - Differences in results between EDUCO and traditional schools Activity Traditional School EDUCO Teacher attendance 3.2 days per week 5 days Teaching Materials A basket of materials The basket costs 1,1 00 costing approximately colones per section (US$ 300.00 colones per section 137.50). (US$37.00). Teaching time 10 hours weekly 18. 75 hours weekly Standing in mathematics Better than that of the EDUCO first and second grade students Parent/Teacher Nil Active, transcends the Interrelation interests of the school and benefits the community 17. In 1993, the enrolment in EDUCO schools represented 10% of rural enrolment in grades 1- 3. That year, more than 80% of the EDUCO students were from rural municipalities with the lowest indicators for nutrition and education (World Bank, 1994). In 1995, EDUCO enrolment was over 100.000 children, covering from pre-school to 5th grade. From 1990 to 195, the program grew from 263 sections to 3,554 (MED, 1995:36). An interesting aspect of EDUCO is the level of formal education of the parents and their involvement in the education of their children. Almost 40% of the parents participating in the ACEs had a 4'11 grade education, and 25% of them have two grades or less of formal education. The monthly income of the parents in EDUCO schools ranged from US$72 to US$79. Most of the presidents (80%) and treasurers (66%) of the ACEs are men. Financial aspects 18. Although the MED is the principal source of funds for the EDUCO program, the community assumes extraordinary activities. In the case of the traditional schools, the parents pay for enrolment, uniforms, books, and school supplies, while in the EDUCO schools, those costs are ' A recent study concluded that among 3rd grade students, there are no differences in the academic standing of the EDUCO students and students in the traditional schools. See: Jimenez, Emmanuel and Yasuyuki Sawada. 1998. "Do Community-Managed Schools Work? An evaluation of El Salvador's EDUCO program." Annex 14, Page 5 covered by the MED. On the other hand, the EDUCO school parents contribute a great deal of tine to work on school affairs and on the interrelationship with the teachers. 19. The annual recurrent coste per student in the traditional primary school is estimated to be US$67 for preschoolers and US$73 for primary students. These costs combine urban centers with rural centers: it is assumed that the costs in rural areas would be higher, but there is no information about this. For primary schools, the operating costs (salaries, materials) represent 80% of the total cost, while administration at the central level absorbs 14.5% of the total cost. The SABE Project (sponsored by USAID), which reinforces the coverage and quality of basic education outside of EDUCO, pays the remaining 5%. 20. For the MED, the annual recurrent cost per student in an EDUCO primary school is US$T 1. The preceding recurrent cost includes the costs of promotion, the training of teachers and members of the ACEs. and the teachers' salaries. If the direct cost to the MED is added to the implicit cost of family participation, the annual recurrent costs per students in an EDUCO primary school is US$35, or 16% more than the average for a traditional school. 21. Obviously for the rural community. the attraction of participating in EDUCO is based on conserving the community control there was before the program, combined with a greater availability of resources that were not there before the program. What remains to be seen is if the program will eventually produce greater learning consistently through all the grades. The New School program in Colombia 22. New School represents an innovation based on the application of new pedagogical principles in the rural sector. The program began in the local sphere in 1975 and was later adopted by the central government (McEwan, 1999; Volvamos a la Gente, Internet, 1999). It presently covers about 9,000 schools (more than 50% of the rural schools in Colombia) through a system that integrates curriculum strategies, teacher training, and community and administrative activities. New School promotes a process of active, cooperative, and personalized learning centered on the student, as well as a close relationship with the community and flexible systems for evaluation and promotion. 23. The objectives of New School are: * expand coverage in the rural areas, offering this population complete primary schooling * increase the academic standing of the students, introducing a qualitative improvement in low- income schools * improve the flow of students, lowering repetition rates * reinforce creativity, self-esteem, and civic behavior among the students * promote the formation of democratic values and behavior and attitudes of cooperation and solidarity in the school and the community Institutional aspects 24. The project began in 300 schools, expanding rapidly after receiving US$1 million in financing from the central government of Colombia (60%) and USAID (40%). These funds helped the Ministry of Education (ME) to convert the local experiences of New School to an expanded pilot project. That project was organized in five basic components: i) technical assistance; ii) 2 Recurrent costs are based on 40 students per classroom in the EDUCO schools and 36 students per classroom in the traditional schools. Annex 14, Page 6 teacher training, iii) curriculum transformation: iv) evaluation and research; and. v) training for leaders. 25. At first, foreign development agencies (USAID, Interamerican Development Bank, UNICEF, World Bank) and the local organizations had a preponderant role in the development of New School. But once the success of the program was seen, the Ministry of Education (ME) joined in the effort, as did some universities and local non-governmental organizations. Once the program took on a more defined form, the ME developed it as an official program of the new curriculum for students. Operative aspects 26. The ME gave initial training for the pilot phase to all the participants. This training was based on the use of a pedagogical guide and the management of the new curriculum. During the following three years of the pilot phase, systematic short course training was given to the participants. Study materials adapted to the new curriculum for the students were also prepared. 27. As a result of the training, the teachers had a great deal of leeway to administer the teaching process. The compulsory activities in the curriculum are few, and so the creativity of the teacher plays a central role in the school. The students are also major players in the program, participating in directing it through participation in a governing council in each school. The communities do not administer the school. 28. In short, New School operates as a special program of the ME within the traditional centralized set up. The operative emphasis is on pedagogical aspects and not on local control or rendering accounts. Operationally, instead of focusing efforts on the family. New School focuses on the teachers by means of pedagogic innovations. The pedagogic innovation of New School has three main components: the curriculum, the communitv, and training and follow up. This includes the demonstration schools. 29. Curriculum component. This includes the interactive textbooks, the learning guides for the students, the learning corners, and the classroom library. The textbooks promote cooperative and active learning and individualized learning, and so students make progress at different rates. The teaching guides are used for curriculum planning and adaptation. The learning corners facilitate the handling of concrete materials to promote comprehensive, non-rote learning, and the classroom library complements and supports the activities of learning. 30. The community component. In order for New School to be set up, the community must prepare a family index card for each student, along with a socio/economic profile of the community. Second, the teachers prepare a farming calendar as an input for organizing the school calendar and a learning tool for teachers and students. Third, the parents help with the construction, maintenance, and furnishing of the school, using a manual guide for this. The school also forms a student government that uses books of minutes, self-control of attendance, books of participation, and suggestion boxes. 31. The component of basic training and in-service follow up. This component prepares the teacher as someone who guides, orients, and evaluates the learning process. The training is done in local workshops with participatory methodologies very similar to those that the teachers will use later with the students. The workshops are more oriented towards practice than theory. Annex 14, Page 7 32. The demonstration schools, which could be the equivalent of quality circles in firms, are part of the system for training andfollow up workshops and are called micro-centers. At these workshops, the teachers interact, learning from other teachers with more experience, reflecting or their own practices, and resolving concrete problems. All this effort promotes a new role for the teacher as someone who orients and facilitates. Results of New School 33. The principal results of New School are in the academic area (McEwan, 1999). The most outstanding achievement has been the expansion of the offer of education to complete primary education in rural schools with I or 2 teachers, where only the I 5, 2nd, and 3rd grades of primary schooling were offered before. This was done by reinforcing the training of teachers in multi-grade teaching, crucial for expanded access to education in rural zones. Another important result is the improved academic standing of the children in the EDUCO schools. These achievements are greater than those among children not participating in the program. In particular, the 3rd grade students achieved a higher academic standing in mathematics and reading. The good results of New School are generally attributable to better training and follow up of the rural teachers, who make better use of the good qualitv materials produced by the program. Lastly, an important result has been the introduction of adjustments to the Law for Education, taking into account the most salient characteristics of the program in order to thereby improve the rest of the rural school system. Financial aspects 34. Although there is no study of the unit costs, the differences in costs between New School and the traditional schools are found in the use of interactive textbooks, the classroom library, ancd the teacher training. Although these costs are high at first, their effect is spread over a number of years. For example, the textbooks or learning guides are used over a period of four or five years b y a host of students (Volvamos a la Gente, 1999). According to a UNESCO study , the program costs 5% to 10% more than that of a traditional Colombian public school. This estimate seems optimistic given the organizational complexity of the program and the implicit cost of community participation. Project to improve the efficiency of primary schooling in Honduras 35. This project is presented because it reflects the problems brought about by the lack of political support for a program oriented towards rural primary schooling, like the two previous. The objective of this project is to increase the quality and efficiency of basic education in the rural areas of Honduras. Its specific objectives are: substantial improvements in academic standings, reduced repetition and desertion, and an increase in the percentage of students completing primary school from 28% to 45%. To achieve these objectives, the program bases itself on the modernization of rural school affairs, the provision of better teaching materials, the construction of better rural school infrastructure, and the creation of a training system, the pivot of which is the national center for teacher training (Chesterfield, 1992). 3Schiefelbein, Ernesto, 1991. "In Search of the School of the XXI Century, Is the Colombian New School the Right Pathfinder?" UNESCO/UNICEF, Santiago, 1991, cited in: Psacharopoulos, et al, 1992. "Achievement Evaluation of Colombia's New School. Is Mutligrade the Answer?" Annex 14, Page 8 Institutional aspects 36. This project came about with funds from USAID, but under the responsibility of the Ministry of Education, which executed it in coordination with a private non-profit association called AVANCE and with the Ministry of Agriculture. AVANCE is responsible for administering a radio station and a periodical for small farmers, while the Ministry of Education is responsible for the other six components: i) design and printing of new textbooks; ii) training and supervision of teachers; iii) research into education policies; iv) creation and administration of a system of managerial information; v) design and execution of a system for evaluating academic standings; and, vi) construction of school classrooms and organization of activities for community development. The Ministry of Agriculture provides technical advice. 37. The project began before the new government took office, and so the government of Honduras did not accept it officially. Because of this, USAID had to make a special effort to "Honduranize" the program. This meant making efforts to convince the ministerial authorities about the national priorities for education and about the need for the project. 38. In short. the program started late, with little political support. Organically, it is a centralized program that emphasizes improved quality and quantity of inputs for the production of education. Operative aspects 39. The project emphasizes the construction of classrooms and the redesign and massive distribution of new textbooks for the 1It, 2nd, and 3rd grades in rural areas. However, the latter component was carried out with a lot of delay. The Ministry of Education was not interested in changing the school textbooks that had been used since the '60s. USAID felt that it was not enough to distribute recently-printed textbooks, but rather, these had to be modernized in order to have a positive impact on the quality of education. 40. Teacher training was affected by the delays in printing the textbooks and their corresponding guides. Consequently, in the first years of the program the teachers.were receiving training without having the textbooks they would use, and as well, the school year began without the corresponding materials. The systems for evaluation and information suffered from design problems and by 1992, there was only information about the results of some pilot experiences. 41. The periodical called "El Agricultor" (The Farmer) still functions, and although it has not been well received by small farmers, teachers use it. This instrument has been useful for teachers because of its general interest articles about Honduras and because of its posters, which serve as complementary teaching material. The radio station broadcasts 50% of its programming in Spanish and the rest in Miskito and Garifuna. Originally the program was to replicate the radio project to teach mathematics that had been carried out in Nicaragua. However, this could not be done since the Honduran mathematics textbooks were not compatible. Results of the program to improve the efficiency of primary schooling 42. Although the project has not been systematically evaluated, a study made in 1989 indicates that there were some improvements among the target student population. There was a 5% increase in students completing primary and repetition was reduced slightly from 27.3% in 1982 to 26.9% in 1987, with the biggest reduction in the 2nd grade, going from 16.2% to 15.6% during the same period. Most significant was the drop in the rate of desertion, falling from 16% to 12% over those Annex 14, Page 9 years (Chesterfield 1992; San Giovanni. R., et al. 1989). Nonetheless, it is unclear whether these improvements in the indicators are attributable to the project. 43. This project, despite its excellent design, was not executed well because the national counterpart did not have the necessary political will during the start up phase so that the project would go ahead according to what was programmed. It is clear that the needs of the population olf the rural areas of Honduras are great and that small investments like the periodical "El Agricultor", textbooks even if delivered late. and teacher training -though not optimal- have a certain positive impact on problems such as repetition and desertion. MEXICO: TELE-SECONDARY 44. A constant concern among governments is the lack of coverage of secondary education in the rural sector. Although primary education can be adapted under a multigrade system, this is mcre difficult for secondary education. On the other hand, empirical evidence shows that the investment in secondary education in Nicaragua is highly profitable for both the individual and for society (Belli and Ayadi, 1998). Because of this, if the question at hand is to improve the productivity of rural youth, it is necessary to consider wide-coverage secondary education with low requirements for human resources as one alternative to explore. It should also be remembered that in Nicaragua, approximately 50% of those entering the Ist grade finish the 6'h grade, and only a little more than half of these students enter secondary. In the rural sector, a big problem is the availability of secondary education centers, and so it is worthwhile to explore better alternatives for access. 45. A program that has captured the attention of multilateral agencies is the Tele-secondary program being realized in Mexico for almost 30 years. This enthusiasm has been transferred to Nicaragua, where in the last two years there has been talk of adopting the program in order to improve the quality and access to education in rural areas. In fact, the plan for the secondary education project of the Interamerican Development Bank (IDB) includes distance education and :t considers Tele-secondary as one of its components. Institutional aspects 46. The concept of Tele-secondary in Mexico began with the objective of increasing access to education for more than 200,000 rural communities with a population less than 2,500. In 1968, the Secretariat for Public Education (SEP, in Spanish) began to broadcast a televised program aimed at 6,500 students in seven Mexican states (Calderoni, 1998). In 1998, the program reached 800,000 secondary students and 12,700 communities. This coverage is equal to 16% of the student population in grades 7 to 9. 47. To receive the program, the communities must have a minimum enrolment of 15 students and a classroom for study. The SEP and the corresponding Secretariat for Education of the State provide the rest of the resources. These include a teacher, a television, a signal decoder, a satellite dish, the textbooks and instruction manuals, and teacher training. Operative aspects 48. Each lesson consists of a 15-minute televised broadcast followed by a 35-minute guided discussion session run by the teacher. Each subject has more than 100 lessons for each grade. The Tele-secondary curriculum is designed for the rural student, but also includes elements of the basic curriculum in order to meet the minimum national requirements. Annex 14, Page 10 49. Each student receives four volumes of textbooks on Basic Concepts that provide detailed explanations of the televised lessons. Each student also receives a learning guide to be used for group exercises and for the discussions. The teacher receives a guide with the objectives of each lesson, teaching strategies, and techniques for organizing the group for discussions. The televised segments are well prepared, very similar to children's television programs like Sesame Street in terms of animation and production, that is, they are more than a simple narrative or a professor in front of a blackboard. Results of Tele-secondary 50. Almost 75% of the students starting Tele-secondary in the 7"' grade finish the 9"' grade, but only 21% of those leaving the 9th grade continue on to the IOt'. This difference shows the importance of the relevance of education via Tele-secondary to the productive life of the rural students, since those who do not continue on into the 1 oth grade go directly to the labor market. In terms of academic achievement, the Tele-secondary students have the same standing as the students in regular schools. Financial aspects 51. Tele-secondary is costly at first. Each original 15-minute segment costs between US$30,000 and US$50,000 and takes 20 days to produce. There needs to be a script, a director, a musical director, actors, editors, and television and computer technicians. About 3,850 segments are in use at present, although more than 6,500 have been produced over the three decades of the program's existence. Equipping each classroom to receive the signal costs approximately US$2,000 for a capacity of 15 students. As a result, the annual recurrent cost per student is US$431 and the total annual cost, including the cost of investment, is US$554 per student. 52. The Government of Mexico has offered Nicaragua the televised segments free of charge. Nicaragua would commit itself to financing the costs of adapting the language and the costs for equipment and transmission. The IDB program plans to serve 135 secondary schools and a total of 5,400 students during the pilot phase at a cost of US$4.42 million, resulting in a unit cost of US$775 during this phase. It is expected that the unit cost in a mature program fluctuates between US$150 and US$300 per student per year. For the purpose of comparison, the total unit cost (governmental cost and family cost) for a secondary student in Nicaragua is approximately US$90/year. This difference in cost should be weighted with the difference in quality between the normal Nicaraguan secondary schooling and Tele-secondary in order to make an effective assessment. VOCATIONAL EDUCATION IN COLOMBIA AND COSTA RICA 53. A constant concern of government in regard to the rural student population is the relevance of the education to their productive lives. According to evidence from the unit for studying poverty of the Ministry of Agriculture, Ranching, and Forestry (MAGFOR) of Nicaragua, many rural students need a kind of education that has an immediate impact on their productivity. Some countries, like Colombia, have opted for diversifying their strategy with vocational programs in agriculture and agro-industry that complement the efforts in rural primary and secondary schools. This section describes some of the experiences in the area of vocational education. Annex 14, Page I1 54. A study published in 19914 by the International Labor Organization and the Inter-American Center for Vocational Training CINTERFOR analyzed vocational education in nine countries of Latin America in order determine which innovative experiences were being applied. The study determined: i) that the vast majoritv of vocational institutes are not connected to the formal education systems that the Ministries of Education are in charge of in each country; ii) that the teachers in these institutes are technicians and engineers before being teachers: and, iii) that the vocational technical institutes give detailed teaching for occupations that are not in line with the needs of the labor market. As a general tendency, many of the vocational programs are outdated and/or encyclopedic in nature, seeking to replicate the contents of secondary education. Regarding the teaching methodology, the use of the "workshop" is predominant, in which the teacher is a facilitator of the teaching process. 55. There is limited attention to the Vocational Institutes in rural areas. In 1986, rural enrolment represented only 10% of the total number of students. However, three of the nine countries where the study was made (Brazil, Costa Rica, and Colombia) have developed special organizational structures for offering services to disadvantaged sectors in the rural areas. The National Service for Rural Learning (SENAR-Brazil), the National Service for Learning (SENA-Colombia). and the National Institute for Learning (INA-Costa Rica) offer education to small farmers, as well as offering -as in the other six countries- attention for agro-industry workers. Rural vocational education in Colombia 56. The objective of the rural education program of the National Service for Learning (SENA) is to contribute to modernization, in terms of the productive and business capacity of this population, seeking to increase their income through greater productivity. The training offered by SENA focuses on agricultural production, the organization of agricultural labor, and rural administration. 57. Using production as a frame of reference, the students identify the needs and problems of the region. Based on this situation analysis, problems are identified that the campesino population should attend to in order to raise their living standards: e.g. energy, aqueducts. garbage disposal systems, etc.. The training methodology is based on learning by doing and implies the organization of seminars. courses, and educational trips, complemented with the use of technological packets., audio/visual means, and materials to take home after class. 58. The SENA also makes use of distance education strategies and gives courses in micro- centers for training. It also uses the agricultural calendar to prepare the training plans and calendars, along with flexible training modules. Rural vocational education in Costa Rica 59. Rural vocational training is provided by the National Institute for Learning (INA) througi the Program for Support to the Rural Sector (PROASER), which is directed by small rural firms with a low level of organization. The main mechanisms for instruction are the pubic workshops and the mobile rural actions. The public workshops offer completely practical training, acting as an "incubator" for the activities carried out later by the students to earn a living. The workshop simulates the physical setting of the campesino firm which is receiving training, although there is a tendency to be oriented towards too much of an urban profile. 4 The study was made in Argentina, Brazil, Colombia, Costa Rica, Chile, the Dominican Republic, Peru, Ecuador, and Uruguay. Annex 14, Page 12 Financial aspects 60. Vocational teaching is generally offered by autonomous Vocational Technical Institutes (VTIs) financed with: i) taxes levied on the firms that, supposedly, benefit from the training offered by the VTIs; ii) outside funding (donations and loans); iii) contracts with firms for the training of their staff; iv) fees paid by the students; and, v) sales of the work done by the students during the workshops. Each VTI has its own method for estimating costs. 61. From the above it can be deduced clearly that the VTls act like subsidized firms that sell products that are not always in line with the needs of the market. A large part of the problems in adapting the technological offer are due precisely to the negative incentive implied by State employment on the efforts of the officials to keep themselves up to date in order to continue being relevant to the needs of the market. RURAL EDUCATION IN NICARAGUA 62. The research into experiences with rural education in Nicaragua gathered information about four very interesting experiences: i) Teaching Mathematics by Radio, ii) the Rural Schools for Education and Work (ERETs in Spanish), iii) the Distance Education programs, and, iv) the Campesino-to-Canmpesino program. Teaching mathematics by radio 63. This project started in 1974 under the auspices of Stanford University which went ahead with it through the Ministry of Education with financial support from USAID (Newman, Rawling, Gertler, 19945). The project was based on a radio program that gave daily one-hour lessons in mathematics for the whole school year with aid from teachers who assisted the students in doing the exercises. The project was interrupted in 1979 by the war situation in the country. The objective of the project was to develop and execute a system for the interactive teaching of mathematics by radio for primary students. Institutional aspects 64. The project design had three phases: research, piloting, and expansion. Each phase was evaluated on the basis of a broad evaluation program, which was very important for the design. The participating schools had to have at least four grades and 15 students enrolled in the first grade. The student population was categorized into groups, first by grade and then by the area in which the school was located since the project was interested in including both urban and rural 15' to 4"' grade students. 65. The evaluation system combined quantitative and qualitative methods. The quantitative evaluation had an experimental design and was organized to enable those in charge of the project to measure its impact by means of standardized exams. These exams were applied every once in awhile to the participating students and a control group. The qualitative evaluation was made on a class basis in order to assess the teaching methods used by the teachers and the reactions of their students. 5 The summary of this project is based on an article published by John Newman, Laura Rawlings, and Paul Gertler in the "The World Bank Research Observer." Annex 14, Page 13 66. The first two years of the project (1974-1976) were for organization. During that time, there was detailed preparation of the radio lessons and a baseline was established to serve as a reference for future evaluations. This baseline was established by making evaluations of the learning of mathematics in Masaya and California. The program began in 1977 in a group of schools locatecl in Masaya, Carazo. Granada, and later the Rio San Juan (in 1978). 67. Those in charge of the project made great efforts to obtain political support for the program. As a result of that work. two Advisory Committees were constituted. They included officials from the Ministry of Public Education and representatives from the participating schools. The other protagonists were the teachers who supported the project in the selected classes. There is no information about the role of the principals of the participating schools. Results of teaching mathematics by radio 68. Using tests of academic achievement, the quantitative evaluation showed that the students showed more improvement in learning mathematics than in the regular schools. The first grade students learning over the radio scored 65.5 in the math tests, while the control group only scored 38.8. For the second grade, the test results were 66.1 for the radio students and 58.4 for the control group. The difference was even greater in the third grade. with the control group students scoring 43.2 and those of the project scoring 51.7. 69. The observations of the class showed that the students were attentive and could keep up with the work when they were doing the exercises included in the work sheet. The teachers expressed their satisfaction with the program and noted that their workload had been reduced and that with the support of the project, they were able to introduce new concepts to help in the teach ing of their students. 70. The program was considered so successful that it was copied in Bolivia, Costa Rica, the Dominican Republic, Ecuador, Guatemala, Honduras, Lesotho, Nepal, and Thailand. Besides expanding its geographic coverage, the contents were broadened, and radio programs to teach science, Spanish, English, and health began. Financial aspects 71. Although no data are easily accessible for the case of Nicaragua, there are data for other countries whose programs were modeled on the basis of the Nicaraguan system. In general, the radio education projects showed declining costse to the extent that the audience expanded. In Honduras, the radio mathematics program broadcast in 1990 had an additional annual cost per student of US$2.94 the first year, later dropping to US$ 1.01 per student. Of these costs, the mosi expensive was transmission. In Bolivia, a cost study in 1991 gave the figure of US$1.51 in additional costs per student for the radio mathematics program, and the results were similar in Lesotho and South Africa. Rural Schools for Education and Work (ERET, in Spanish) 71. The ERET schools were founded in 1981 under the auspices of the Ministry of Agricultural Development. The ERET project consisted in creating a set of schools to which land was allocated in order to develop a program of standard teaching linked to agricultural practices. The ERETs 6This section is based on Bosch, Andrea, 1999. "Interactive Radio Instruction: Twenty-three Years of Improving Educational Quality." Annex 14, Page 14 were, according to UNESCO, a pilot curriculum experience that sought to provide an excellent setting for pedagogical innovation (Arrien and Matus, 1988:486). The ERETs covered preschool, primary, and secondary education. Seven years after being created, in 1988, there were 12 ERETs under the direction of the Ministry of Education with an enrolment of 5,000 students. At that time, this student population represented 0.5% of the total enrolment of 257,464 students from the rural area (Ibid. p. 487). 72. The objectives of the ERETs were: i) educate by integrating theory and practice, manual and intellectual labor; ii) promote the pre-vocational formation of the students, especially in technical and productive areas; iii) develop the concept of the self-management of the school among the students; and, iv) develop a dynamic linkage between the school and the community. Institutional aspects 73. Once they were transferred to the MED, the ERETs became part of a UNESCO project called "Support for the development of Basic Education in rural areas" (FMR/UNESCO, 1991:2-3). This project began in 1984, three years after the first ERETs had begun to function on the State farms, known as State Production Units. that produced coffee, tobacco, and cattle. The general components of this new project were: i) strengthen the workshops of the ERET schools: ii) create an in-service training system for the directors, supervisory staff, and teachers; and, iii) design and produce educational materials that are better adapted to the integrated rural education promoted by the ERETs. 74. The ERETs were always an experimental project that was received well by the participating schools. At first they only functioned on State Production Units, but later they were extended to farming cooperatives. In the productive part of the program, they developed agricultural projects for coffee and tobacco, irrigation for vegetables, farming, and ranching. All the ERETs had a dining hall for children. Two ERETs were very successful, one in Jalapa and another in Managua. The one in Jalapa attended to a student population of 200 children. Those students worked in the tobacco fields, received classes based on an adapted curriculum, and had access to food and health services. The ERET in Managua worked on transforming the curriculum, creating Integrated Learning Units in which traditional teaching was integrated with production around central didactic pivots (Arrien and Matus, 1988:486). 75. The pedagogical elements of the ERETs are described fairly vaguely in the different documents there are, although it can be said that they were based on work in the field combined with formal interdisciplinary schooling. The ERETs had their own textbooks, prepared with contributions from the students, and a training system to support the teachers and supervisors of the project. Results of the ERET schools 76. According to an investigation by UNESCO in 19907, it was found that the practical farm work was developed independently from the study activities, creating a gap between the productive educational work and the formative educational work. This was attributed to an insufficient command of the teachers of the theoretical aspects of production. On the other hand, the community/school relationship was minimal, and the so the system was not very effective because its vertical design did not facilitate feedback. 7 Apochryphal document provided by the UNESCO office to one of the authors. Annex 14, Page 15 Financial aspects 77. There are no data about the costs of the ERET project. It is easy to conclude that it was a high-cost project since the students lived in residence and received a number of services besides education. However, if the current costs for a similar system are calculatedc, the unit cost runs between US$800 and US$ 1,200 per year for interned students and US$20 per year for outside students. Distance education and accelerated primary schooling 78. Distance Education was started in 1981 to extend education services to rural areas. Thirougi this service. attention was provided for groups whose work did not allow them access to a regular school. The schooling was developed using self-study and by means of weekly gatherings that had two phases: one for orientation and the other for control. The teachers played the role of orienter and tutor. Institutional aspects 79. The Ministry of Education was the principal organizer and financer of this service. Distarce Education had the following components: * a technical team based in the central offices and located in the Adult Education sub-system that prepared plans, profiles, and methodological orientations for developing the curriculum for this project. The same team was responsible for executing the training of the teachers and the supervising the educational work (Arrien and Matus, 1988:544). * a set of methodological guides on the study topics and textbooks. - a number of centers in different zones of the country that imparted courses of Accelerated Primary to students, generally those in residence. The primary education went for three semesters for the residence students and three years for extemal students. They imparted compacted I"s and 2nd grades in the First level, 3rd and 4th in the Second level, and 5"' and 6"' ir1 the Third. - a distance secondary institute called "Filem6n Rivera." This institute imparted classes on Saturday afternoons. The students were organized in study circles that functioned autonomous,ly during the week. Results of accelerated primary schooling and distance education 80. The results of this program have not been evaluated. Arrien reports (1988:545) that 5,503 students, most of them adults, were covered by the program in 1988. A good number of the students were members of the Ministry of the Interior and the Army. Distance Education facilitated the access of a large number of campesino students, leaders of farming cooperatives, and ranchers to education. Financial aspects 81. The costs of the program were shared among the participating institutions. Students only paid for the study guides and textbooks. The residence students received food and lodging and when they were heads of households, there was a scholarship to cover the deficit created by the temporary absence from the work force of the person who brought the most income to the household. The s The calculation is made as follows: cost of food per students: US$30/month; lodging in school dornitories: US$30/month; cost of teachers and materials (25 students per section): US$15/month. Annex 14, Page 16 modality for gatherings only included the cost of teaching materials and food. Under the present conditions, the unit cost9 for residence students would be US$600/year, while for the external student, it would be US$1 54/year. Multigrade schools 82. Multigrade schools have existed in Nicaragua since the 1950s. However, starting in 1996 through an AID Base Project executed by AED, a process of modernizing a group of multigrade schools began. The innovation sought to: * reduce the repetition of grades, and thereby the expenses stemming from this problem * raise the quality of the educational offer, and * increase scholastic coverage, offering complete primary in this group of schools. Institutional aspects 83. The principal components of this innovation have been: i) the renewal of educational materials: ii) training for principals and teachers; and, iii) adaptation of the curriculum to the reality of the rural area. In regard to the latter, it has sought to have a greater linkage between the school and the community, and so the new curriculum contains topics like handling agricultural machinery, techniques for treating water, techniques for the preservation of animal species, and sex education. Results of the innovation in 26 multigrade schools 84. According to the AED-Aid document cited previously, in 1997 the following achievements and problems were noted. ACHIEVEMENTS PROBLEMS There is cooperative work. Teaching of spelling/grammar Good planning based on the use of index cards Shortage of materials for teachers Good start on community participation in the Student governing bodies are incomplete educational project Good start on evaluation as feedback Lower grades have more problems and require more support Time is used with more flexibility Training is still day to day 85. The officials of the Base Project feel that this model has good chances for success, since it has been "constructed by the protagonists themselves: the schoolteachers and the education authorities." However, they consider that the Base Project should give more systematized support and the advice should be adapted to the reality of these schools. Financial aspects 86. No data could be obtained about the costs of this innovation. However, from the interviews it can be deduced that it is a high-cost project. Its current functioning depends on financial support from the AID and technical support from the AED, which is executing the Base Project. 9 The cost for student in accelerated primary is calculated to be US$60/yeasr for teachers and materials, plus US$96/year for the food provided at the weekly gatherings. Annex 14. Page 17 Campesino-to-Campesino Program 87. The Campesino-to-Canipesino program is a flexible program for adults. Its main components are: * participatory research, conceived of in such a way as to make the campesinos active subjects in the education process, starting with an analysis of their production problems. For this., the idea of the program was to define the most appropriate technological alternatives to counteract limitations to production in a collaborative fashion, and to adapt the teaching curriculum to their productive needs and develop their own plans for experimentation and training (Castillo and Castro, 1993:9). * the promotion and organization of networks of campesino experimenters who participate voluntarily in the program as implementers of alternatives technologies. These networks start on small farms and/or cooperatives in the zones with the most production problems and degradation of the natural resources. * facilitating technical learning by means of one or two promoters financed by the NGO. Institutional aspects 88. This program started in Mexico and was then extended to Central America in the 1980s. In Nicaragua. it was sponsored initially by UNAG (National Union of Farmers and Ranchers), whicl sought to improve the living conditions of poor campesinos in the dry tropical areas. In the late 1980s., the program was sponsored by a number of non-govemmental organizations, which adopted it to the wet tropics in order to improve the living conditions of the rural population and contribute to the conservation of the natural resources by means of more appropriate farming techniques in order to reduce soil erosion and improve income levels. Operative aspects 89. Given the lack of systematized information about this program, an experience from the wet tropics is presented as a reference in order to give a brief description of the Campesino-to- Campesino program. The experience was in Rama after the area had been hit by Hurricane Joan il 1988. The campesino population living in very poor distant communities had low-yield production systems based on the classic cycle of slash and burn. 90. Based on a self-analysis, the participating campesinos selected fertilizing beans (Mucuna l as a technological alternative. The beans were planted in small areas of the production parcels of the participants in 1992. The project provided the seed and left the farmers free to use the bean crop as they saw fit. The technician/promoters began the process of providing attention to the parcels of the experimenters. In 1993, the results were evaluated and validated. The validation was done in workshops in which all the experimenters participated. As a result of the workshops, the planting techniques for Mucuna beans were improved by finding out which methods had produced the best results in terms of weed control. Results of the Campesino-to-Campesino Program 91. Following the previous example, after eight years the project in Rama had 210 farmers participating. It began in three communities and has been extended to 22, covering approximately 800 families. Standing out from among the achievements of the project is the recovery of productive areas and increased yields for beans, maize, and rice. A high percentage of the farmers went from producing eight hundredweight of beans per manzana to 12. Rice yields increased from 35 to 70 hundredweight per manzana and corn yields rose by 30%. Annex 14, Page 18 92. Besides the direct results on production over the short term, the program helped with the inclusion of biological pest control techniques in order to avoid polluting the river and with the use of seeds that are appropriate for the climate and soils of the wet tropics. Many of these seeds are being produced locally. This contributed to the start of construction of silos for storing the production so it can be sold at a better price. Financial aspects 93. The unit cost of the program is approximately US$200/year. These costs are derived directly from the data for execution of the pilot project. CONCLUSION AND RECOMMENDATIONS 94. The principal conclusion of this review of experiences is that the investment in rural primary education appears to produce better results than the investment in vocational education or in agricultural training. The innovations in rural basic education have been most effective in increasing the coverage and -given the positive experiences that show increased learning- the quality of schooling than vocational education and agricultural education. 95. The cases of basic distance education -Tele-secondary in Mexico and Teaching Mathematics by Radio in Nicaragua- are interesting innovations that warrant a serious evaluation for their adaptation outside of the rural sphere. Tele-secondary is particularly important to take into account in light of the fact that the increases in academic standing appear to be linked to the use of high-quality televised segments. This is a strong indication that the quality of education in Nicaragua would gain more if it adapted Tele-secondary to the whole country, instead of relegating it only to the rural sector. 96. The experiences in vocational education and agricultural education have had little coverage, were very costly, and had a fairly modest impact nationally. Experiences in basic education 97. Both EDUCO and New School are considered to be successful programs because they managed to increase the coverage and quality of education in rural zones at a moderate unit cost. Although both programs involved the education community -EDUCO directly and New School in a collaborative fashion- community participation has not been an end in itself, but rather a very important element for having the system be governable. The experience derived from decentralization in general indicates that the primordial objective of community participation in the provision of public services is for there to be more closeness to the final client, thereby seeking better operative efficiency and a better climate of responsibility. Both EDUCO and New School have achieved this coming closer, and as a result have obtained better schooling and an eventual increase in coverage. In both programs, the pedagogical innovations and the climate of responsibility towards the parents helped improve the relevance and efficiency of the education, and to a lesser extent, the learning, resulting in a better perception of the quality of education within the community. This suggests that the gains in quality of the education supply as perceived by the community result in students staying in school and more coverage, even when the real increase in learning are modest. 98. In the case of EDUCO, the existing community interest was primordial for setting up the State program. The eagerness of the parents to participate in the education of their children was key Annex 14, Page 19 to having the investment in the Parents' School, in the specialized training of teachers, and in the production of more appropriate textbooks yield better results than the projects in which the community was not taken into account, like the case of Honduras. 99. In New School. community participation in the administrative affairs has been less. Although the families contributed with labor and materials -showing their commitment to the education of their children- there has not been a good local control of the resources, thereby reducing the capacity to adapt the curriculum even more to the rural needs and to render accounts, to the parents. However, the New School model contributes significant pedagogical innovations, such as the use of learning guides and specialized corners in the multigrade classrooms. 100. School autonomy in Nicaragua is an adaptation of EDUCO, but with less depth in the pedagogical aspects than New School. This suggests that a good model for rural basic education would combine the autonomy of EDUCO with the teaching materials and methods of New School.'° 101. The case of Honduras shows that the lack of political support greatly reduces the possibilities for success. In order for Nicaragua to successfully carry out a project that combines the characteristics of EDUCO with those of New School, it needs to improve the political support to the areas that would constitute de facto decentralization -e.g. local control over the appointment of :he principal, adaptation of school hours and the school calendar to the community's agenda for farm work, effective supervision of administrative and pedagogic functioning, and training of the principals in leadership and communication. Experiences in vocational education 1 02. The experiences in vocational education show a structure of agent-principal, in which the Government (acting as principal) is not clear about the market failure to be corrected, or the nature of the public good that should be provided. As a consequence, the vocational institutions (acting as agents) do work that is imperfect, but which clearly responds to the bureaucratic incentives imposed on them by the Government. As a result, the experience of the vocational institutes in other countries confirms the anecdotal evidence there is about INATEC (National Institute for Technical Education and Training) in Nicaragua. This indicates a system overloaded with administrative bureaucracy, with teachers and a curriculum that is not in contact with the real labor market of its students, and.with problems in the design of the mode for generating and absorbing information about the relevance of its programs to the demand for labor in the country. Therefore, it is not surprising to see such a lack of quantity and quality of vocational programs oriented towards the rural sector, since the failing in the agent-principal relationship leads to locating the vocational programs in areas that are more convenient (physically and pedagogically) to the bureaucratic system administering them. Because of this, there is a clear trend in Latin America to privatize vocational education services and to solve the problem of access and equity by allocating coupons to subsidize the payments by the users for private vocational services. Experiences in agricultural education 103. Specialized agricultural education -through informal training, short courses, or agricultural diploma courses- is an improved variant of the vocational institutes since they are specially designed to attend to the rural school age population. The results obtained in terms of 0 In fact, some components of New School are in the process of being adapted to Nicaragua since they are included in the Aprende 11 project of the World Bank. Annex 14, Page 20 training and cost effectiveness vary from very good to very bad, depending on the funds, their continuity, the managerial capacity, and the projection of the programs nationally. 104. The experience of the programs with residence students is problematic since the unit costs are at least five times higher. Furthermore, the residence student programs have much more complex requirements in terms of the management and administration of the center and in the quality of the school principal, and these are difficult to find in Nicaragua. Consequently, the coverage of the agricultural residence student programs is low. 105. In the case of agricultural programs for outside students who live at home, the programs show a great range in variation. The younger the student, the more integration there is between the agricultural component and the basic education. If the agricultural component is too specialized, coverage is lost, and if it is not specialized at all, it does not help make the education relevant. in some way, the success of New School is partly due to the ability to combine basic education with elements of agriculture. 106. Lastly, to complicate the scenario even more, the training of teachers for agricultural programs is a strong limitation since an expanded coverage of agricultural/basic education requires training programs that still do not exist. As shown by the experience in Honduras (and to a certain extent, Costa Rica), the teachers are not well trained in agricultural labor, and this detracts from the utility of the program. Experiences in distance education 107. The positive experience with teaching mathematics by radio in Nicaragua indicates that the program should be looked at again, especially in order to increase the coverage and quality of primary education. The low unit cost, plus the possibility of having greater coverage for other subjects at a low cost, warrants a review and improvement of the experience in Nicaragua in the 1970s. This is very important if one wishes to improve the performance of the multigrade classrooms (also know as one-teacher schools) at a low cost. 108. The broad presence of radios in rural zones -or the low cost of providing radios free of charge to potential users- means that instruction over the radio is a very attractive option for filling the void that currently exists in the zones where the State schools do not reach the 4"'t or 5"' grade. This means a major investment in the design and production of radio programs, student guides, and special textbooks and exercise books that do not now exist in Nicaragua, but which could be adapted from those countries, which inspired by the experience in Nicaragua (!), currently make use of instruction via radio on a large scale. 109. Tele-secondary is also a very promising medium for education that would complement primary education via radio. Empirical evidence from Mexico indicates that keeping academic standings at the same level as the regular schools is the result of having a well-planned televised class, with the best teachers as protagonists. This is a big incentive for extending this mode of instruction to the urban zones of Nicaragua where there are few State secondary schools with quality teaching. 110. A problem with Tele-secondary is the high unit cost for the televised segments. This cost is reduced significantly if the Mexican segments (or those used in El Salvador, which are adaptations of the Mexican segments) are adapted. Mexico has promised to give the segments to Nicaragua free of charge but will not pay for adapting them to Nicaraguan language. Even so, the costs of adapting them are high. Obviously, Tele-secondary benefits from economies of scale: the greater the Annex 14, Page 21 coverage of the program, the lower the unit cost. This indicates that if the coverage of the program is expanded to the whole country, the earnings that could be obtained in academic achievement and in the national scholastic level would justify the additional expense. 111. Another problem with Tele-secondary is the administration, maintenance, and security of the equipment for receiving the television signal. The cost of a satellite dish and TV equipment 'is high, but not prohibitive. However, given the poor conditions for storing it in the schools, the irregular electricity service, and the problems of theft in the schools, risks are created that shouid be taken into account in terms of the cost. The pilot program that the MECD (Ministry of Education. Culture, and Sports) will begin with IDB financing will help in finding solutions to these problemns. 112. In short, the adaptation of innovations in rural basic education -through programs like New School within a context of school autonomy- is more efficient than the other programs in terms of coverage and improved leaming. Programs like Tele-secondary and education via radio are promising. but they have to be adapted to Nicaraguan reality, and this will take more time since t depends on greater technical and managerial capacity, something there is a shortage of in the MECD. The agricultural and vocational programs can help a lot to improve the technical capacitv of campesino youth. but their low coverage, high unit cost, and high requirements for training of teachers are an obstacle to solving the problem of low rural academic levels in the short or medium term. Annex 14. Page 22 REFERENCES AED -AID, 1997, "Propuesta de Educaci6n multigrado para Nicaragua" Unpublished internal document, Base Project - Substantive Area. Arrien, Juan B. and Matus Lazo, Roger 1989, Nicaragua Diez Afios de Educaci6n en la Revoluci6n, MED publications, Managua, Nicaragua. Belli Pedro and Mohammed, Ayadi, 1998. Returns to Investment in Education, the Case of Nicaragua. Working paper, The World Bank, Washington D.C. Bosch, Andrea, 1999. "Interactive Radio Instruction: Twenty-three Years of Improving Educational Quality." Education and Technology Technical Note Series, Educational Development Center, The World Bank, Washington D.C Castillo, Nelly and Castro. Vanessa, 1993, "El Diagn6stico Participativo: Dos experiencias de Trabajo." IPADE Publication, Managua, Nicaragua. Calderoni, Jose. 1998. "Telesecundaria: Using TV to Bring Education to Rural Mexico." Vol. 3, No. 2, Education & Technology Technical Notes Series. Education and Technology Team, Human Development Network-Education Group, The World Bank, Washington D.C. CINTERFOR-ILO, 1991, Vocational Training on the Threshold of the 1990s. Volume 1. Education and Employment Division Population and Human Resources Department, The World Bank, Washington. Chesterfield Ray, USAID, 1992. Basic Education, Review of Experience. Bureau for Latin America and the Caribbean, Office of Development Resources, Education and Human Resources Division. FMR/UNESCO, 1991, "Apoyo al Desarrollo de la Educaci6n Basica en las Areas Rurales, Resultados y Recomendaciones del Proyecto." San Jose, Costa Rica. Gajardo, Cecilia, 1988, Ensefianza Basica en las Zonas Rurales. Experiencias Innovadoras. UNESCO. Jimenez, Emmanuel y Yasuyuki, Sawada, 1998. "Do Community-Managed Schools Work? An evaluation of El Salvador's EDUCO program." Working Paper Series on Impact Evaluation of Education Reforms No. 8. The World Bank, Washington D.C. Ministerio de Educacion de El Salvador, 1995, "EDUCO Una Experiencia en Marcha." Internal document. McEwan, Patrick, 1999, "Escuela Nueva" Revista de la Federaci6n Nacional de cafeteros de Colombia, no 13. Newman, John; Rawlings, Laura; Gertler, Paul, 1994, "Using Randomized Control Designs in Evaluating Social Sector Programs in Developing Countries" The World Bank Research Observer, Vol. 9, (pp. 181-201). Annex 14, Page 23 Padgett, Tim, 1998, "The Light of Learning" Time, Special Reports/Summit of the Americas, Santiago, Chile, April, 20, Vol. 151.15. Psacharopoulos, George; Rojas, Carlos and Velez. Eduardo, 1992. "Achievement Evaluation of Colombia's Escuela Nueva. Is Multigrade the Answer?" WPS 896. Policy Research, Human Resources Technical Department, Latin America and the Caribbean Region. The World Bank, Washington D.C. UNESCO, undated, "Desarrollo de la Practica Educativa en las ERET." Unpublished. The World Bank, 1994. "Community Education Strategy: Decentralized School Management." Report No. 13502-ES, Country Department 11, Human Resources Operations Division, Latin America and the Caribbean Regional Office. Washington D.C. Other sources: Interviews with Oscar Mogoll6n, curriculum specialist for the AED-AID Base Project; interview with the coordinator of the Rama Project, Canipesino-to-Canipesino, Oveida Morales. The two interviews were conducted in May 1999. Volvamos a la Gente, 1999, documents from the Internet. Annex 15, Page 1 Annex 15. Rates of Return to Education in Nicaragua by Diana Kruger INTRODUCTION I. The objective of this paper is to estimate the returns from education in Nicaragua based on the 1998 Living Standards Measurement Survey. It will include estimates of average private rates of return, rates of return by schooling level, and social rates of return. 2 . The rate of return to education, both average and by schooling level, reveals the profitability of investments in education. This information can be of use to policy makers in evaluating public investment and expenditures in the education sector. It also indicates how the labor market rewards the education of workers in the form of higher wages after controlling for other factors. 3. The paper is organized as follows: Section 2 is a brief review of the theory on returns from education and the methodology used in this paper and Section 3 provides a description of the data. Section 4 contains results, including a discussion about private and social rates of return, the estimation procedures, variation in results by gender, area, and regions, and a comparison to the results from the 1993 LSMS. Section 5 presents conclusions and policy implications.' BACKGROUND AND METHODOLOGY Average Rates of Return 4. The literature on rates of return from education originated with the theory and earnings function derived by Mincer (1974). The simplest form of the model estimates the return to wages from an additional year of education: lnw=a+3 S+02Exp+j33Exp2 (1), where In w = log of hourly wages; S = number of years of schooling completed, Exp and Exp2 are post-schooling work experience and its squared term, respectively. The average rate of return to education is captured by the coefficient ,B,: an individual's investment in an additional year of schooling has a marginal rate of return of P3I on her/his earnings. In Mincer's model, the rate of return is assumed to be the same for all levels of schooling, implying that PI is also the average rate of return. 5. Mincer's original work has been extended vastly in the literature, both theoretically and empirically. This paper estimates the original model (1), as well as extensions with additional controls for geographic location, gender, and poverty grouping. The results are found in Appendix 1, Tables Al through A4. Rates of Return by Level of Schooling ' The paper contains a statistical Appendix 1, which has all regression results (part A) and analytical tables (part B). Annex 15, Page 2 6. One of the more useful extensions to (1) is to drop the assumption of a constant rate of return across different education levels. This permits an estimation of the rate of return for primary, secondary, and higher education, and allows for a more detailed analysis regarding the impact of different education levels on wages. 7. The specific methodologv used to estimate the rate of return from education at different schooling levels follows Lachler (1998).2 A detailed description of the methodology used is found in Appendix 2. THE DATA 8. The data used in this paper is from the Living Standards Measurenment Survey (LSMS) carried out in Nicaragua in 1998. The data is nationally representative at the urban/rural and regional aggregation levels. The survey gathered information from over 4,600 households and almost 23.000 individuals. The questionnaire covered topics such as housing infrastructure, demographic composition, health, education. employment, fertility, and household income/expenditures, among others. These different data modules allowed for the construction of income and consumption aggregates, as well as wealth proxies that permit the study of the effec: of education on wages. The Dependent Variable 9. The dependent variable used in this paper is the log of average hourly real wages, composed of cash and in-kind payments received in remuneration for employment. Thus, the sample is restricted to wage earners and does not capture the segment of the population that derives their income from other sources, including agricultural production, which is not sold on the market. Wages have been adjusted for spatial/regional price differences. The Explanatory Variables 10. The variables used to explain variation in wages are grouped into four general categories: human capital, household-level characteristics, poverty categorization, and individual-specific characteristics. 11. Human capital: As mentioned above, the LSMS contained a section that gathered the education history of each individual interviewed. Each person was asked: 1) "What is the highez;t education level and grade achieved?" and 2) "Are you enrolled for the current school year?" From questions I) and 2), it was possible to construct a continuous variable of the total years of education attained by each respondent. 12. In the construction of this variable (years of education), total efficiency in educational attainment was assumed, i.e., that individuals completed primary school in 6 years, secondary school in 5 years, and university in 5 years. The LSMS questionnaire captures the highest level and years completed, but it does not capture past grade repetition.3 It is important to point out that this efficiency assumption can lead to inaccuracies when estimating private rates of return, and 2 The method used to estimate the variation in returns from education by education level will follow Lachler: "Education and Earnings Inequality in Mexico." World Bank Policy Research Working Paper N o. 1949. July 1998. 3The LSMS asks those who are enrolled in school if it is their first time in the current grade, so it is possible to find repetition for the current year only. The sample used is persons who are not enrolled in school, so the efficiency assumption was made. Annex 15, Page 3 especially social rates of return, because it underestimates the cost of education. This is a potentially serious problem in Nicaragua, because the average repetition rates are very high - 25% and 14% for the 15' and 2nd grades, respectively (Arcia 2000).4 13. In addition to the continuous variable of years of education, seven educational categories were constructed based on the total number of years studied: no education (0 years), primary incomplete (I -5 vears), primary complete (6 years), secondary incomplete (7-1 0 years), secondary complete (I I), university incomplete (12-15 years), and university complete and higher ( 16 years or more). 14. Other human capital variables include experience in the labor market. Following the literature, the variable "total experience" was constructed as: age of person minus 6 minus years of education. Tenure captures the number of years of experience at the current job. 15. Regional: Dummy variables were constructed to measure the effect of household location on wages. i.e.. to see if there is regional variation in wages. The variables include a dummy for urban location (it equals I if the household is located in an urban area, 0 if it is rural). and a set of seven dummy variables for the seven regions of the country (in the regressions, Managua was the left-out dummy). 16. Poverty: Welfare is proxied by annual per capita consumption in the Nicaragua Poverty Assessment. Households were grouped into three poverty groups (extreme poor, poor, and non poor) depending on whether or not their total annual consumption was sufficient to obtain minimum caloric requirements (extreme poverty line) and basic items (poverty line). The regressions in this paper include dummy variables that rank households into consumption quintiles. 17. Individual: The estimations include a dummy variable for females (to test for gender discrimination in wages), as well as other factors that may affect the self-selection of women into the labor force, such as marital status and number of children aged 0-5, 6-9, and 10-14. The Sample 18. The sample used in the estimations was selected following criteria in Pessino (1993) to allow comparison with the rates of return from education that were estimated from the 1993 LSMS. The regressions were estimated for individuals who: 1) were between 25-64 years of age, 2) were not currently attending school, and 3) were employed and received positive wages (including self-employment).5 4 This assumption has been addressed in the literature, see for instance, Behrman & Deolalikar. 5 Any disparity between the sample and the number of observations in the regressions is due to missing variables. Annex 15, Page 4 Sample Description -O L S Heckman - Criteria Males Females Total Males Females Total Total Sample 9.096 9.438 18.534 9.096 9.438 18.534 Aged between 25-64 years 3.515 3.892 7.407 3,515 3.892 7.407 Not enrolled in School 3,363 3,690 7,053 3.363 3.690 7.053 Employed 2,853 1,559 4,412 n/a n/a n/a Received positive wages 2,853 1,559 4.412 n/a n/a n/a Source: 1998 LSMS. Note: The Heckman regressions include observations of both employed and unemployed. 1 9. The distribution of educational attainment for gender and location is illustrated in the table below. Over half of the sample did not complete primary education, which is not surprising given that the average educational achievement in Nicaragua in 1998 was 4.5 years of schooling. A greater proportion of men relative to women did not complete primary school, and an even greater disparity exists along urban-rural lines: while 41% of urban dwellers didn't finish a basic education, 79% of rural adults did not complete primary school! These gender and location divisions will be explored further. Table A15.1 - Educational Attainment Education Level Male Female Urban Rural Total No education 817 29% 326 21% 368 15% 775 40% 1,143 26% Primary not completed 932 33% 453 29% 635 26%/'o 750 39% 1,385 31% Primary complete 343 12% 217 14% 381 15% 179 9% 560 13%/o Secondary not completed 400 14% 246 16% 524 21% 122 6% 646 15% Secondary complete 157 6% 148 9% 256 10% 49 3% 305 7% University not completed 96 3% 101 6% 158 6% 39 2% 197 4% University completed 108 4% 68 4% 164 7% 12 1% 176 4% Total 2,853 100% 1.559 100% 2,486 100% 1,926 100% 4,412 100% Tabulations include the entire sample. Annex 15, Page RESULTS A) Private Rates of Return Ordinary Least Squares 20. The results of OLS estimations of the average rates of return and the returns from different education levels are shown in Table A2 (the main findings are summarized below). The average private rate of return to education is 8.3% for the total sample.6 8.4% for males and 8.0% for females. Heckman Regressions 21. On average, labor force participation rates are 72% for males and 35% for females.7 The rates of return obtained from OLS are likely to be biased, since it is probable that persons with higher "ability" choose to obtain more education and are more likely to be employed. To correct for this phenomenon, the Heckman (two-step) procedure was applied.8 The first regression estimated the probability of being employed as a function of age, marital status, and number of children in the ranges of 0-5 years, 6-9 years, and 10-14 years of age. The second regression estimates the returns from education. Regression results are found in Appendix 1, Table A3. 22. The resulting rates of return are summarized below. Henceforth, this paper will consider the Heckman rates of return from education. The main findings are summarized below: Table A15.2 - Private Rates of Return from Education-OLS and Heckman Estimations OLS Heckman Total Males Females Total Males Females Average 8.3% 8.4% 8.0% 8.1% 8.6% 7.7% Primary (vs. primary not completed) 6.6% 9.7% 0.9% 5.2% 6.3% 3.2% Secondary (vs. secondary not completed) 10.4% 8.3% 12.7% 1 10.1% 8.9% 12.5% University (vs. university not completed) 17.5% 11.6% 25.6% 17.3% 11.9% 24.6% 23. The average rate of return from education after correcting for self-selection is 8. 1% for the population as a whole, 8.6% for males, and 7.7% for females. This means that on average, an additional year of education brings an 8.6 and 7.7% increase in earnings for males and females, respectively. 24. The impact of an additional year of education varies at different schooling levels. The rates of return for the different levels should be interpreted as follows. Men who complete primary education earn 6.3% more in hourly wages than those who did not complete primary and if they complete secondary education they earn 8.9% more than those who do not complete secondary. 6 The first column of Table A2 resulted in a statistically significant (negative) sign on the dummy variable for females, implying wage discrimination. I proceeded to split the sample along gender lines and ran separate regressions. 7Estimates from the Labor Market background paper (in progress) by Nadeem Ilahi. 8 The Heckman procedure is basically a two-step method where first, the probability of being in the labor force is estimated for the entire sample (based on some "selection criteria"), and then a the wage regression is run, accounting for the results of the first step. The selection criteria are variables that affect the probability of being in the labor force but not the level of wages, such as: age, being single, number of children. Annex 15, Page 6 Completing university has a marginal effect of 11.9% on hourly wages compared to not completing university. The same interpretation applies to women. 25. At first glance. the total private rate of return of 17.3% from university education might seem high. However, it is noteworthy that the rate of return from university studies in Nicaragua is comparable to other countries: this rate of return for Panama was 24.0% in 1998 and for Mexico'° it was about 20.0% in 1994. 26. In Nicaragua, the average hourly wage increases by about 50% when a person completes university studies (relative to incomplete university), from 19.1 to 31.7 c6rdobas (Appendix 1, Table B5). Tlhus, the labor market in Nicaragua pays a very high wage premium to those who complete university studies. Nonetheless, it is important to consider that only 5% of the sample reported completion of a university degree or more (Table B 1). 27. Education has a positive effect on earnings because it creates opportunities to access economic sectors paying higher wages. Tables B12 and B13 summarize the distribution of economic activity bv education level for men and women. Persons with higher educational levels are employed in the economic sectors paying higher wages. which explains the variation of rates of return from education across gender. 28. The return from primary schooling is lower for women than men. Note from Table B 1 2 that men who complete primary education participate less in agriculture (29%) than those that have no education (76%) or did not complete primary (49%). Completing primary school moves men out of agriculture -the economic sector paying the lowest hourly wages- and they obtain large wage increases for completing this level of education (relative to women, Table B 10). 29. Women, on the other hand, are concentrated in the commerce sector and have lower education levels (Table B 13). Completing primary school does not change the distribution of employment by economic activity, and thus does not bring them as large an increase in wages as it does for men (Table B 10). This is reflected in women's low rate of return to primary education. On the other hand, completion of higher education (secondary and university education) has a somewhat stronger impact on women's hourly wages. As the achievement of higher education increases, more women are employed in higher paying sectors (Table B 13). For example, 13% cf women who obtain a university degree participate in the Financial Services sector (compared to only 4% of women who do not complete their university studies), which pays a very high wage premium to womenll (Table BI 1). So even though women on average earn lower hourly wages than do men, completing secondary and university education has a comparativelv higher impacl on their wages. B) Social Rates of Return'2 30. The ideal estimation of social rates of return would capture total public benefits and all costs of education. Total benefits from education include the externalities captured from having a more highly educated population. These externalities are difficult to quantify, so in practice, estimation procedures consider only private benefits from education (as measured by higher wages). It is important to point out that not including the positive externalities from education 9 World Bank, Panama Poverty Assessment, Volume 2, Annex 14. June 1999. '° Lachler 1998. " The high wages reported for women in Construction and Transportation represent a small portion of the sample and can be considered outliers. 12 For a description on the methodology of how social rates of return were estimated, see Appendix 2. Annex 15, Page 7 leads to understating social rates of return, especially for primary schooling: the economic literature on returns from education stipulates that the greatest positive externalities occur at the primary level. 31. Following the literature, the social rates estimated in thispaper consider the social (private and public) costs of education. Table A15.3 - Average Public Costs of Education, 1998 Cordobas lJS$ Average 1.129 107 Primary 555 52 Secondary 358 34 University ' 23.237 2,196 University 2 17.427 1.647 University 3 11,618 1.098 Annual cost per student 32. Government subsidies to university education in Nicaragua are very high. By law, the National Council of Universities (CNU) receives 6% of the government budget (ordinary revenues). In 1998, this figure was C$349 million (about US$33 million).'" 33. Three estimates are presented for the cost of university education. It is important to point out that the exact number of full-time equivalent (FTE) university students is available to the public, so each cost estimate above is based on a different assumption about the number of FTE students. The most conservative estimate, University- I, assumes a total enrollment of 15.000 FTE students, in which case the annual per-student subsidy would be C$23,237. The other two estimates, University-2 and University-3, are based an FTE enrollment of 20,000 and 30,000 students, respectively, resulting in per-student annual subsidies of C$17,427 and C$ 11,618, respectively. The estimates of public costs (above) are taken from Arcia (1998)." 34. These three different assumptions yield the corresponding social rates of return from university education noted in the table below: Table A15.4 :Social Rates of Return to Education-Heckman Regression Total Males Females Average 7.5% 8.0% 7.0% Primary (vs. primary not completed) 5.0% 6.1% 3.1% Secondarv (vs. secondary not completed) 9.9% 8.7% 12.2% University I (vs. university not completed) 12.7% 9.2% 15.3% University 2 13.6% 9.8% 16.9% University 3 14.6% 10.4% 18.9% Rates are statistically significant at the 501o confidence level. 35. Note that once the public subsidy to education is considered, the average rate of return for the total sample falls from 8.1 (private) to 7.5%; the rates of return from each level of schooling fall to 5.0, 9.9 and 12.7%. 13 Arcia, Gustavo. "El Financiamiento de la Educaci6n en Nicaragua". Consultant Report to the Interamerican Development Bank. November 1998. 14 The author suggests that the conservative estimate is the most realistic. Annex 15, Page 8 36. The differences in the private and social rates of return reflect the degree of public subsidv. of education (Psacharopoulos 1995). There is very little difference between private and social rates of return from primary and secondary schooling. This is due to the low level of per-student public subsidy. The largest difference between private and social rates occurs at the university level (due to the large subsidies provided by the State to university students). C) Urban v. Rural Rates of Return 37. The rates of return from education vary not only across gender, but also across geographic location (see Appendixl, Table A4). Below is a summary of rates of return for urban and rural areas: Table A15.5 - Private and Social Rates of Return by Area All F Urban Rural Private Rates of Return Average 8.1% 8.4% 7.9% Primaxy 5.2% 8.1% 3.9% Secondary 10.1% 10.3% 12.6% University 17.3% 19.2% 15.8% Social Rates of Return Average 7.5% 7.8% 7.1% Primary 5.0% 7.8% 3.7% Secondarv 9.9% 10.1% 12.3% University (1) 12.7% 14.9% 9.9% University (2) 13.6% 15.8% 10.9% University (3) 14.6% 16.8% 12.2% (1) The costs per university student are based on the assumption of 15.000 full time equivalent (FTE) students. This is the most conservative estimate. 2) and 3) are estimated with assumptions of 20.000 and 30.000 FTE students, respectively. 38. Average private rates of return to education and the return from primary education are slightly higher in urban areas than they are in rural areas. This implies that completing rural primary compared to urban primary is not as profitable, probably due to the fact that rural residents are mostly employed in the agriculture sector (Table B7 in Appendix 1), which pays very low wages regardless of the level of education. 39. We also observe that for secondary education, the rates of return (both private and social) are higher in rural relative to urban areas. The trend is reversed for university studies.15 Achieving higher education commands a proportionately higher premium in urban areas. This, again, is due to the nature of the urban and rural labor markets: the sectors paying the higher premiums for education -financial services and transportation- are concentrated in urban areas. Regional Differences 40. Regional wage premiums exist, which is demonstrated by the significant (negative) coefficients on the regional dummy variables in Tables A2 and A3 (Managua is the left-out variable). These imply that not living in the capital city has a negative effect on wages due to location after we control for other variables. " It is important to point out that there are few observations in the sample of individuals completing university studies in rural areas (see table on p.3), so care must be taken when interpreting these results. Annex 15. Page 9 41. The premium is correlated to the fact that Managua's economy has the least participation of the agriculture sector, where the lowest wages are paid, and has the highest concentration of high-wage sectors (Table B6 and B7). D) Returns from Education in 1993 42. The background paper on Labor Markets'6 for the 1993 Nicaragua LSMS included a section on returns from education. Tables 12 and 18 of that study contain the results of the different regression specifications that were estimated. The only directly comparable specification with this study is the basic Mincerian regression (see footnote 1 7). Table A15.6 Rates of Return from Education: 1993-1998 1993 1998 Men Women Men Women OLS "Mincer" Average 14.1 6.2 14.2 11.6 Heckman ** Average 10.9 5.9 8.6 7.7 Primary 10.7 4.6 6.3 3.2 Secondary 9.1 5.0 8.9 12.5 University 14.1 13.8 11.9 24.6 "*Note: Comparisons must be made with caution since it was impossible to replicate the regressions from 1993. Source for 1993: Tables 12, 16. 17, 18. 19 and 20 in Pessino (1994). 43. The only perfectly comparable measures included in 1993 and 1998 estimations are results of applying the basic "Mincer" equation, summarized in the first line above. Based on this figure, the average rate of return for an additional year of education for men increased slightly from 14.1 to 14.2 percent; however, the return from an additional year of school of women increased from 6.2 to 11.6%. 44. In 1998 it was impossible to exactly replicate the other estimations of the 1993 study.' However, certain assertions can be made: if we compare the two Heckman regressions that are the most "similar," we obtain the trends observed above. Returns from primary education probably remained the same for women but decreased for men; secondary returns were constant across time for men but increased for women, who also experienced a large increase in returns from university education. This could be due to the fact that the labor force participation rate of women increased slightly to 35%, but it is likely due to improved access to economic sectors that pay higher wage premiums for education. 45. It is possible that the higher monetary benefits from education are due to structural changes in the labor market that occurred between 1993 and 1998: a period of economic transition, when the economy became more open internationally and competitive markets continued to develop. In 1998, higher education receives relatively larger wage premiums compared to five years before. 16 Cf. 3. '' The 1993 LSMS contained a section on training, which the 1998 LSMS did not include. Every regression specification in the previous study (except the basic Mincerian equation) included training as explanatory variables. Annex 15. Page 10 However, despite a higher return on their investment, the labor market continues to discriminatc against women via lower wages. CONCLUSIONS AND POLICY LESSONS 46. The immediate conclusion one can draw from this exercise is that investing in education is profitable for Nicaragua: it has positive rates of monetary returns (wages) and positive non- monetary returns in the form of externalities to society, which have not been quantified. The main findings and policy lessons are summarized below. a. The average private rate of return"8 for the entire population is 8.1%; for men and women these are 8.6 and 7.7%, respectively. The highest private rate of return for the entire sample occurs at the university level at 17.3%, with marginal returns of 10.1 and 5.2 % for secondary and primary, respectively. b. The average social rate of return is 7.5% for the sample: the social rates of return by schooling level are 5.0% for primary, 9.9% for secondary, and 12.7% for university education. c. Although the highest marginal rates of return are from university education, caution must be exercised in drawing policy conclusions strictly from the estimated rates of return. Policy guidelines must necessarily weigh efficiency and equity considerations. It is true that a university education almost doubles average hourly wages, yet only 5% of the sample completed a university degree. Government policy makers should ask themselves in regards to the allocation of resources to the different education levels: Do differential rates of return across levels justify large differences in subsidies? Using the most conservative figure (University-I), the ratio of university to primaly per-student government spending in Nicaragua is 42:1, and the ratio of university to secondary education spending is 65:1. Yet it is alarming that on average, 53% of the sample did not complete even primary schooling and the proportion is even higher -74%- for rural residents. d. The issue of retaining students in school until they complete at least primary is the most urgent priority when one considers that positive externalities -improved health, reduced infant mortality, fewer teenage pregnancies, and the education of future sons and daughters- are obtained at the primary school level. h Given that the majority of the population would immediately benefit from this achievement, policies should focus on school retention at the primary level. Note that for men, the social rate of return from primary schooling -even without including the non-quantifiable, positive externalities from primary education- is almost as high as the rate of return from completing university. e. The literature on the economics of education maintains that higher positive social externalities are obtained from primary education, and that the benefits from higher education a,re mainly private'9. A policy that incorporates equity considerations would shift resources from university to primary education. This implies that the charging of higher fees to attend 18 Recall these are rates of return from the Heckman regressions. 19 See the argument in Pscharopoulos 1995, pp.9-10. Annex 15, Page 11 university and reducing government subsidies to higher education are socially efficient policy 20 implications . 20 According to a paper by Belli and Ayadi on Returns from Education in Nicaragua, approximately 17 primary school students can be educated for every university student, which implies a net present value to society ten times greater if the 17 primary school students are educated (Belli p.8). Annex 15, Page 12 Appendix A15.1 - Tables Part A - Rates of Return from Education Part B - Various Table A 15.1.AI1 NSicaragua LSMS 1998 Mincer Regressions (OLS) Dependent l ariable: Log hourhZ, wages All Men Women yrsed 0.132622** 0.142005** 0. 116022** exp 0.0471 * 0.050276** 0.038619-* expsq -0.00048-* -0.00053 -0.00036* female -0.18731 * constant -0.08084 -0.17799 -0.0417 R-squared 0.216 0.229 0.1872 No.Obs. 3774 2390 1384 ** and * represent statistical significance at the 5% and 10% level, respectively Annex 15, Page 13 Table A15.1.A2 Nicaragua LSMS 1998 Rates of Return from Education: Average and Schooling levels (OLS) Dependent Variable: Log hourly wages All Males Females yrsed 0.0833** 0.0845** 0.0796** ed2 0.1402** 0.1175* 0.1891** ed3 0.3207** 0.3824** 0.2125** ed4 0.3681 ** 0.4605** 0.2241** ed5 0.6538** 0.6943** 0.5624** ed6 0.9017** 1.0134** 0.7150** ed7 1.4776** 1.3884** 1.5704** xp 0.0331 * 0.0345** 0.0394** 0.0414** 0.0236** 0.0260** expsq -0.0003 - -0.0004** -0.0004 ** -0.0005 -0.0002 -0.0004 ** tenure 0.0138 ** 0.0 133 ** 0.0069 0.0065 0.0290** 0.0275 ** tenuresq -0.0004 ** -0.0004 ** -0.0003 * -0.0002 -0.0007 ** -0.0006 ** urban -0.0054 0.0162 0.1235 0.1408 -0.2110 -0.1919 regATLUR -0.0052 -0.0153 0.0899 0.0804 -0.1660* -0.1568* regATLRU -0.3380** -0.3657** -0.2711 -0.2932 -0.3101 -0.2990 regCENUR -0.3097** -0.3241** -0.3689** -0.3754** -0.2269** -0.2441** regCENRU -0.5387** -0.5635** -0.4588** -0.4813** -0.6590** -0.6545** regPACUR -0.2306** -0.2371** -0.2866** -0.2846** -0.1671** -0.1682** regPACRU -0.2588** -0.2702** -0.2611 -0.2628 -0.2142 -0.2379 quint2 0.2348** 0.2505** 0.3062** 0.3153** 0.0594 0.0825 quint3 0.3484** 0.3928** 0.4224** 0.4491** 0.1613 0.2494** quint4 0.4400** 0.4969** 0.5608** 0.5937** 0.1701* 0.2819** quint5 0.7958** 0.8273** 0.9207** 0.9353** 0.5325** 0.6176** female -0.2371** -0.2377** -cons 0.2507 0.3984** 0.0193 0.1372 0.4639 0.6041** R-squared 0.2999 0.3043 0.3318 0.3318 0.2556 0.2789 No.Obs. 3,774 3,774 2,390 2.390 1,384 1,384 ** and * represent statistical significance at the 5% and 10% level, respectively Annex 15, Page 14 Table A15.1.A3 Nicaragua LSMS 1998 Return from Education: Average and Education levels (Heckman regressions) Dependent Variable: Log hourlv wages All Males Females Urban Rural Regression: Yrsed 0.081 ** 0.086** 0.077** 0.084** 0.079** ed2 0.133** 0.120** 0.180** 0.072 0.165** ed3 0.310** 0.380** 0.194** 0.287** 0.274** ed4 0.352** 0.466** 0.214** 0.328** 0.308** ed5 0.629** 0.714** 0.547** 0.601 ** 0.666** ed6 0.871** 1.030** 0.706** 0.801** 1.074** ed7 1.440** 1.414** 1.529** 1.420** 1.583 Exp 0.028** 0.029** 0.040** 0.042** 0.016* 0.020** 0.030** 0.034** 0.031** 0.031 Expsq 0.000** 0.000** 0.000** 0.000** 0.000 0.000** 0.000** 0.000** 0.000** 0.000* Tenure 0.013** 0.013** 0.007 0.007 0.030** 0.028** 0.020** 0.019** 0.001 -0.002 Tenuresq 0.000** 0.000** 0.000* 0.000* -0.001 -0.001 ** 0.000** 0.000** 0.000 0.000 Urban -0.006 0.019 0.111 0.129 -0.192 -0.165 Female -0.233 ** -0.233 ** -0.316** -0.321 ** -0.062 -0.060 RegATLUR -0.028 -0.039 0.095 0.085 -0.168 -0.167 -0.014 -0.030 RegATLRU -0.369** -0.391 ** -0.280** -0.305** -0.328 -0.305 -0.308** -0.301 RegCENUR -0.310** .40.325** -0.368** -0.375** -0.214** -0.237** -0.315 ** -0.331 ** RegCENRU -0.536** -0.558** -0.471 ** -0.495** -0.674 ** -0.655** -0.511 * * -0.506** RegPACUR -0.228** -0.235 ** -0.292** -0.291** -0.187** -0.188** -0.237** -0.239** RegPACRU -0.263** -0.270** -0.269** -0.271** -0.221* -0.233 * -0.276** -0.259** quint2 0.258** 0.272** 0.307** 0.317** 0.056 0.082 0.273** 0.301 ** 0.243** 0.253 quint3 0.382** 0.424** 0.427** 0.455** 0.178** 0.266** 0.321 ** 0.379** 0.420** 0.455 quint4 0.485** 0.539** 0.573** 0.608** 0.198** 0.305** 0.433** 0.505** 0.477** 0.517** quint5 0.836** 0.866** 0.944** 0.960** 0.539** 0.629** 0.825** 0.874** 0.734** 0.744** -cons -0.179 -0.033 0.231 0.356** -0.049 0.174 0.517** 0.667** -0.349 -0.233 Selection Criteria: Age -0.007** -0.007** -0.009** -0.009** 0.013 ** -0.013 ** -0.009** -0.009** -0.003 -0.003 Single 0.026 0.026 -0.383** -0.383** 0.375** 0.390** -0.103** -0.099** 0.071 0.071 chle5 -0.049** -0.048** 0.010 0.010 -0.086** -0.087** 0.019 0.015 -0.023 -0.021 ch69 -0.038** -0.038** 0.015 0.015 -0.062** -0.062** -0.002 -0.003 -0.036 -0.036 chIO14 -0.035** -0.037** 0.028 0.030 -0.032 -0.030 -0.002 -0.002 -0.044** -0.044** -cons 0.472** 0.483** 0.964** 0.954** 0.246** 0.247** 0.618** 0.619** 0.113 0.125 Log Likelihood: -9904.7 -9892.3 -5211.2 -5211.2 -4196.1 -4176.1 -5555.1 -5537.4 -4260.1 -4260.9 No. Observations: 7052 7052 3363 3363 3689 3689 3833 3833 3219 3219 Chi-squared: -59.71 -53.19 72.03 77.05 50.61 51.67 115.45 119.8 -118 -111.6 ** and * represent statistical significance at the 5% and 10% level, respectively Annex 15, Page 15 Table A15.1.A4 Nicaragua LSMS 1998 Rates of Return from Education (Heckman Regressions): Private and Social Dependent Variable: Log hourliv wages All Males Females Urban Rural Private Rates of Return Average 8.1% 8.6% 7.7% 8.4% 7.9% Primary 5.2% 6.3% 3.2% 8.1% 3.9% Secondary 10.1% 8.9% 12.5% 10.3% 12.6% University 17.3% 11.9% 24.6% 19.2% 15.8% Social Rates of Return Average 7.5% 8.0% 7.0% 7.8% 7.1% Primary 5.0% 6.1% 3.1% 7.8% 3.7% Secondary 9.9% 8.7% 12.2% 10.1 % 12.3% University (1) 12.7% 9.2% 15.3% 14.9% 9.9% University (2) 13.6% 9.8% 16.9% 15.8% 10.9% University (3) 14.6% 10.4% 18.9% 16.8% 12.2% (I) The costs per university student are based on the assumption of 15,000 full time equivalent (FTE) students. This is the most conservative estimate. 2) and 3) are estimated with assumptions of 20,000 and 30,000 FTE students. respectively Annex 15, Page 16 Table A15.I. BI Distribution of Educational Attainment by Area and Regions Ed Level Urban Rural AtlUr AtlRu CenUr CenRu Mga. PacUr PacRu Total No Education 13% 36% 27% 50% 16% 42% 10% 13% 32% 23% Primary Incomplete 24% 38% 22% 32% 26% 39% 22% 29% 41% 30% Primary Complete 16% 11% 13% 6% 16% 1 0% 18% 15% 11% 14% Secondary Incompl. 23% 6% 19% 6% 20% 4% 25% 20% 9% 16% Secondary Compl. 10% 4% 12% 2% 11% 3% 9% 9% 4% 7% University Incomplete 7% 3% 4% 3% 4% 2% 9% 8% 2% 5% University Complete 7% 1% 3% 0% 7% 0%so 8% 6% 1% 5% rotal 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% Table A15.1.B2 Number of Observations: Education Level Achieved by Geographical Area and Regions Ed Level Urban Rural AtlUr AtIRu CenUr CenRu Managua PacUr PacRu Total o Education 331 670 53 92 83 344 117 109 201 1000 Primary Incomplete 629 704 44 59 135 322 263 250 260 1333 Primarv Complete 409 201 24 11 85 79 210 129 71 610 Secondary Incomplete 596 117 36 12 104 31 297 173 60 713 Secondary Complete 245 71 22 4 57 27 105 78 24 316 Universitv Incomplete 175 64 9 5 22 20 107 66 12 240 University Complete 186 14 7 0 37 3 91 54 8 200 Total 2571 1841 195 183 522 826 1190 860 636 4412 Annex 15. Page 17 Table A15.1.B3 Participation of Education Levels in Economic Activity Gas/Elec/ Financial Community Ed Level Agric. Mining Manuf. Water Constr Comm. Transp. Svces. Svces. Total No Education 62% 1% 4% 0% 3% 18% 2% 1% 10% 100% Primary Incomplete 36% 1% 11% 1% 4% 26% 4% 2% 15% 100% Primary Complete 19% 0% 14% 0% 7% 30% 6% 3 % 20% 100% Secondary Incomplete 10% 1% 12% 1%oj 6% 35% 7% 4% 24% 100% Secondary Complete 9%/o 1% 10% 3% 5% 25% 4% 4% 39% 10()% University Incomplete 8% 0% 8% 2% 5% 21% 9% 5%/o 41% 100% University Complete 6% 0% 5% 1/ 5% 16% 2% 15% 53% 100% Total 31° 1% 10%/ 1% 5% 25% 5% 3% 21% 100% Table A15.1.B4 Distribution of Economic Activity by Education Level Gas/Elec/ Financial Community Ed Level Agric. Mining Manuf. Water Constr Commerce Transp. Svces. Svces. Total No Education 46% 20% 10% 11% 15% 16% 8% 6% 11% 23% Primary Incomplete 36% 32% 33% 27% 27% 31% 29% 21% 21% 30% Primary Complete 9% 4% 21% 8% 20% 16% 18% 14% 13% 14% Secondary Incomplete 5% 28% 21% 14% 21% 22% 24% ' 20% 19% 16% Secondary Complete 2% 12% 8% 27% 7% 7% 6% 10% 14% 7% University Incomplete 1% 4% 5% 11% 6% 4% 10% 10% 11% 5% University Complete 1% 0% 2% 3% 4% 3% 1%°so 21 % 12% 5% (Total 100% 100% 100% 100% 100% 100% 99% 101% 100% 100% Annex 15, Page 18 Table A15.1.B5 Average hourly wage (Cordobas) by Education Level, Area and Regions: Ed Level Urban Rural Managua PacUr PacRu CenUr CenRu AtlUr AtIRu Total No Education 5.9 3.9 6.7 5.7 5.0 4.5 2.6 6.4 5.9 4.7 Primary Incomplete 6.4 5.5 6.6 5.5 5.1 5.6 5.5 12.0 6.3 5.9 Primary Complete 9.0 5.7 10.6 6.3 5.3 7.8 4.5 13.6 4.1 8.0 Secondary Incomplete 8.5 5.8 9.4 7.0 6.6 7.4 5.0 10.9 5.7 8.1 Secondary Complete 12.3 7.6 14.3 8.6 6.1 15.1 4.5 12.7 3.5 11.2 Universitv Incomplete 19.4 18.1 32.3 9.1 5.5 8.7 6.5 10.0 10.0 19.1 University Complete 32.0 27.7 42.5 23.4 37.6 19.6 8.6 19.7 31.7 Total 10.6 5.9 13.9 7.6 5.8 8.3 4.3 10.9 6.0 8.8 Table A15.1.B6 Average hourly wage (Cordobas) by Economic Activity, Area and Regions Activity Urban Rural AtlUr AtlRu CenUr CenRu Nlanagua PacUr PacRu Total Agriculture 6.9 5.4 12.9 5.7 5.1 7.1 4.2 10.3 5.1 5.7 Mining 12.3 5.7 7.7 4.3 4.1 6.7 15.0 5.0 9.0 Manufacture 8.2 5.9 10.3 6.2 7.2 6.4 3.8 8.8 5.2 7.6 Gas/Elec/Water 8.2 5.3 5.6 8.6 5.5 10.4 5.0 7.7 7.4 onstruction 13.0 5.4- 17.4 10.2 4.4 6.8 5.7 18.7 6.3 10.6 Commerce 8.9 6.1 10.3 6.7 6.7 8.9 2.8 11.0 11.4 8.3 Transportation 20.0 10.2 28.5 9.8 12.1 12.7 3.1 20.6 14.9 18.2 Financial Svces. 23.8 8.0 26.9 8.7 3.2 12.9 16.1 . 8.5 5.8 20.5 Community Svces. 10.1 6.3 12.4 8.8 5.0 8.3 4.2 8.2 5.1 9.1 Total 10.6 5.9 1 13.9 7.6 5.8 8.3 4.3 10.9 6.0 8.8 Table A15.1.B7 Distribution of Economic Activitv by Area and Regions (in percent) Activity Urban Rural AtlUr AtlRu CenUr CenRu Managua PacUr PacRu Total Agriculture 13 66 23 83 19 76 7 10 47 36 Mining 1 1 4 1 1 0 0 0 1 1 Manufacture 11 5 4 2 10 5 13 14 8 9 Gas/Elec/Water 1 1 I 0 2 0 1 1 1 I Construction 5 3 5 1 7 3 5 5 4 4 Commerce 33 12 31 8 30 7 35 33 19 24 Transportation 6 2 4 1 4 0 7 6 4 4 Financial Svces. 3 1 1 0 3 0 7 2 1 2 Community Svces. 27 10 27 4 25 8 25 29 14 20 Total loo loo 1 100 100 100 100 100 100 100 100 Annex 15, Page 19 Table A15.1.B8 Distribution of Economic Activity by Gender Activity Male Female Total Agriculture 43% 9% 31% Mining 1% 0% 1% Manufacture 9% 10% 10% Gas/Elec/Water 1% 0% 1% Construction 8% 0% 5% Commerce 17% 40% 25% Transportation 6% 1% 5% Financial Svces. 4% 2% 3% Community Svces. 11% 37% 21% Total 100% 100% 100% Table A15.1.B9 Distribution of Educational Attainment by Gender Ed Level Males Females Total No Education 25% 18% 23% Primary Incomplete 31% 29% 30% Primary Complete 13% 15% 14% Secondary Incomplete 16% 17% 16% Secondary Complete 6% 10% 7% University Incomplete 5% 7% 5% University Complete 4% 5% 5% Total 100% 100% .100% Table A15.1.B10 Average hourly wage (Cordobas) by Education Level and Gender Ed Level Males Females Total No Education 5.0 3.9 4.7 Primary Incomplete 6.2 5.4 5.9 Primary Complete 9.2 6.1 8.0 Secondary Incomplete 9.5 5.8 8.1 Secondary Complete 13.6 8.8 11.2 University Incomplete 27.3 8.6 19.1 University Complete 32.9 29.9 31.7 Total 9.8 7.2 8.8 Annex 15, Page 20 Table A15.1.BI I Average hourly wage (Cordobas) by Economic Activity and Gender Activity Males Females Total Agriculture 5.9 3.6 5.7 Mining 9.1 6.7 9.0 Manufacture 7.8 7.2 7.6 as/Elec/Water 7.7 5.6 7.4 Construction 10.2 69.9 10.6 Commerce 11.1 6.2 8.3 Transportation 17.1 28.5 18.2 Financial Svces. 16.7 35.0 20.5 Community Svces. 13.9 6.7 9.1 Total 9.8 7.2 8.8 Table A15.1. B12 Distribution of Economic Activity by Educational Attainment: Males Gas/Elec/ Financial Commun. Ed Level Agric. Mining Manuf. Water Constr. Commerce Transp. Svces. Svces. Total No Education 76% 1% 3% 1% 5% 8% 2% 1% 3% 100% Primary Incomplete 49% 1% 11% 1% 7% 17% 6% 2% 5% 100% Primarv Complete 29% 0% 14% 1 % 11% 20% 10% 4% 9% 100% SecondarvIncomplete 15% 1% 11% 1% 10% 27% 10% 6% 18% 100% econdary Complete 16% 2% 12% 6% 9% 15% 6% 7% 28% 100% University Incomplete 14% 1% 13% 2% 9% 25% 11% 6% 19% 100% University Complete 8% 0% 6% 1% 7% 10% 1 % 15% 52% 100% Total 43% 1% 9% 1% 8% 17% 6% 4% 11% 100% Table A15.1.B13 Distribution of Economic Activity by Educational Attainment: Females Gas/Elee/ Financial Commun. Ed Level Agric. Mining Manuf. Water Constr. Commer Transp. Svces. Svces. Total No Education 26% 0% 7% 0% 0% 41% 0% 0% 26% 100/ Primary Incomplete 11% 0% 10% 0% 0% 43% 1% 2% 32% 100% Primary Complete 3% 0% 15% 0% 0% 44% 0% 1% 36% 100% Secondary Incomplete 3% 0% 15% 0% 0% 46% 1% I % 34% 100% Secondary Complete 1% 0% 9% 1% 0% 34% 3% 2% 50% 100% University Incomplete 2% 0% 3% 1% 0% 16% 6% 4% 68% 100% University Complete 2% 0% 4% 0% 1% 23% 3% 13% 54% 100% Total 9% 0% 10% 0% 0% 40% 1% 2% 37% 100% Annex 15, Page 21 Appendix A15.2 Methodology on Estimating Rates of Return to Schooling Levels (Private and Social) The "Mincerian" method was used to estimate the rates of return from years of education (Mincer 1974). The method used to estimate the rate of return to education levels is from Lachler 1998 (both methods are outlined in Psacharopoulos 1995.) The methods assume that the annual earnings of a person with education level t, denoted by E,, are equal to the earnings she could have obtained with education level t-l (foregone earnings) plus the cost of obtaining the additional education, Ct. times r,. the rate of return on that "investment". If we let K, be the ratio of cost of schooling of level t to (foregone) earnings of level t-1, that is, K,=C,/E,-,, we can say the following: E, =Et-, + r, *C, = E,.,(lI+ r, *C, /E,,I) =E,I( I +r,*Kt)=Eo* H(l +rj*Ki)fori=l. t Ln(E,)=LnEo + Ln(I +r* Ki) (1) The "Mincerian" specification is the following: Ln(E, )= Ln Eo + (rK) S (2) In the limit, equation (1) is the same as (2). In the basic Mincerian equation, r and K don't vary by education level and S = number of years of education. Now let us assume that r and K vary by education level. Let us denote the schooling levels in Nicaragua as follows: 0 No education, I = primary incomplete, 2 = primary complete, 3 = secondary incomplete, 4 secondary complete, 5 = university incomplete, and 6 = university complete (and beyond). We can extend equation (2) as follows: Ln(E,) = Ln Eo + (r,KI) SI + (r2K2) (SI + S2) + (r3K3) (SI + S2 + S3) + (r4K4) (S1+S2+S3+S4) + (r5K5) (S]+S2+S3+S4+S5) + (r6K6) (S1+S2+S3 +S4+S5+S6) =LnE0+ f31DI +P32D2+P3D3+ P4D4+35D5+ 36D6 (3) Where Si are the number of years of moving from schooling level i- I to level i, the Di are dummy variables for having completed schooling level i, and pi are the coefficients of regression (3). To interpret these coefficients and derive the average rates of return of each level, note that Pi = r,K1 , so ri _ / (K Si). To find the marginal rates of return form moving from one schooling level to the next, denoted by mi, the average rate is expressed as the weighted sum of marginal rates so that the unadjusted marginal rate of return of investing in education level i is: m1K1 = (pi - Pi.) / Si (4) Annex 15. Page 22 Private and Social Rates of Return Private and social rates of return to education are derived from (4) by dividing by the appropriate cost/earnings ratios, Ki. These can be expressed as: Private Ki = K1P = ( Ej, + Cip) / E1-l Social K = Ks = ( Ei-l + Cip + Cig) / E-l = K p + (Cig / Es ) where CjP denotes private out-of-pocket expenses of education, Cil denotes government's expenses per student per year at level i. To estimate private rates of return, Ki is assumed to be equal to one, so from (4), the marginal rate of each schooling level i is mi = (p - pi-,) / Si . To estimate social rates of return, Kig is calculated from estimating the ratio (C,l /Ei-1) ard adding one. then by adjusting equation (4) correspondingly. To estimate this ratio. the annual per-student government spending was used, as well as tie average annual earnings per education level for each sub-sample. These average earnings are as follows: Total Male Female Urban Rural AtlUr AtIRu CenUr CenRu Managua PacUr PacRu Average 17.480 20.230 12,650 22,829 10.007 27,121 8.517 17.142 7,812 29.230 16,015 9.921 Primarv 17.101 19.773 12.908 17,918 15.435 32.504 5.289 16.626 14,837 21.626 11,852 12.888 Secondary 23.384 30.533 16.353 26,318 13,298 51,941 4,438 28.532 9.786 22.957 19.518 16.964 University 69.523 85.216 46.543 70.472 56.683 32.413 44,122 52.802 96.039 48.199 65.717 69.523 In cordobas per year In this study of Nicaragua, it was assumed that persons completed primary school in 6 years, secondary in 5 years, and university in 5 years'-. Based on this assumption, the following average Si were estimated for the different sub-samples: Level: Total Males Females Urban Rural AtlUr AtlRu CenUr CenRu Managua PacUr PacRu Total 0 0 0 0 0 0 0 0 0 0 0 0 0 c 1 3.21 3.26 3.30 3.31 3.12 3.43 2.85 3.24 3.07 3.47 3.26 3.29 3.21 2 6 6 6 6 6 6 6 6 6 6 6 6 6 3 8.29 8.199 8.34 8.35 8.02 8.4 8.39 8.35 8.3 8.16 8.47 7.76 8.29 4 11 11 11 11 11 11 11 11 11 11 11 I1 11 5 12.79 12.92 12.85 12.88 12.44 12.92 13 12.57 12.38 13 12.91 12.14 12.79 6 16.19 16.16 16.19 16.21 16 16.18 16.23 16 16.19 16.21 16 16.19 16.19 Equations (2) and (4) were estimated across gender, area and regions. Private and social rates of return were estimated. The regression results are found in Tables Al-A5 of the Appendix. The estimated rates of return are reported in Tables A6-A7. 21 Setting these years to completion of education levels implicitly assumes 100% efficiency in the sample. i.e., that persons did not repeat any grades. Annex 15, Page 23 References Arcia, Gustavo, 1998. "El Financiamiento de la Educaci6n en Nicaragua". Consultant Report to the Interamerican Development Bank. Washington, D.C. Arcia, Gustavo, May 2000. "-" World Bank, Background Paper for the 1998 Nicaragua Poverty Assessment. Behrman, Jere R. and Anil B. Deolalikar, 1991. "School Repetition, Dropouts, and the Rates of Return to Schooling: The Case of Indonesia." Oxford Bulletin of Economics and Statistics, pp. 467-480. Belli, Pedro and Mohamed Ayadi, 1998. "Returns from Investment in Education, the case of Nicaragua." Working Paper (Draft). Chiswick, Barry R., 1997. "Interpreting the Coefficient of Schooling in the Human Capital Earnings Function." Policy Research Working Paper No. 1790. The World Bank, Washington, D.C. Lachler, Ulrich, 1998. "Education and Earnings Inequality in Mexico." Policy Research Working Paper No.1949. The World Bank, Washington, D.C. Mincer, Jacob, 1974. Schooling, Experience, and Earnings. National Bureau of Economic Research and Columbia University Press, New York. Pessino, Carola, 1994. "Labor Markets in Nicaragua after the Stabilization Plan." Background Paper for the 1995 Nicaragua Poverty Assessment. The World Bank, Washington, D.C. Psacharopoulos, George, 1995. "The Profitability of Investment in Education: Concepts and Methods." Working Paper, The World Bank, Washington, D.C. World Bank, 1999. Panama Poverty Assessment, Volume 2, Annex 14. Washington, D.C. Annex 16. Page I Annex 16 - Determinants of Primary School Attendance in Nicaragua: An Analysis of Household Demand by Diana Kruger INTRODUCTION 1. Education is a key component to economic development. There is a large of literature relating education and economic growth. At the micro level, increased educational attainment brings private economic benefits to individuals -such as higher wages, better and more employment opportunities- as well as positive social externalities. Increased education levels, especially of women, has positive effects on children's health, nutrition, and education outcomes. among other things. 2. Enrollment rates in elementary education vary by region of the world in which a person lives. A smaller percentage of school-aged children are enrolled in primary school in developing nations relative to high-income countries, as Table I reveals. Table A16. 1 - Net Enrollment Ratio' (%) of Primary School - by World Regions, 1996 East Asia and Pacific 101 Europe/Central Asia 92 Latin America and the Caribbean 91 Nicaragua 78 Costa Rica 91 El Salvador 78 Honduras 90 Middle East and N. Africa 87 South Asia, Sub-Saharan Africa n.a. High Income 97 Source: World Bank, 1999 World Development Indicators 3. This paper will analyze the determinants of primary school enrollment in Nicaragua. The net enrollment rate in primary schooling in Nicaragua is below the Latin American average, and far from the international development goal of providing "universal primary education," which is to be achieved by all countries by the year 20152. 4. The paper is organized as follows: section 2 provides some background about education enrollment and outcomes in Nicaragua. Section 3 describes the data and methodology; in section 4 the estimation exercise and main findings are presented, section 5 compares rural-urban outcomes and finally section 6 provides concluding comments and policy implications. 'This paper follows the World Bank definition of net enrollment rates, which is the ratio of the number of children of official school age (as defined by the education system) enrolled in school to the number of children of official school age in the population. A net enrollment ratio greater than 100 indicates discrepancies between estimates of the school-age population and the reported enrollment data. 2 See World Bank website: http://www.worldbank.org/data/dev/devgoals.html Annex 16, Page 2 EDUCATION IN NICARAGUA 5. In Nicaragua, primary school consists of 6 years, and secondary school takes an additional 5 years to complete. Table 2 presents the average years of education obtained by Nicaragua's adult population in 1998, broken down by region of residence and gender. Educational attainment in rural areas is at critically low levels, and substantially below the achievements by urban dwellers. Women in this sample, obtained slightly less education than men. Table A16.2 Average Years of Education Attainment Men Women Total Urban 6.2 5.5 5.8 Rural 2.7 2.7 2.7 Total 4.6 4.4 4.5 Source: Nicaragua LSMS. 1998. Expansion factor applied. Persons aged 25 and older. not currently enrolled in school 6. Deeper analysis of the sample of persons aged 25 and older reveals a bleak picture: about 35 percent had no education, and another 31 percent did not complete primary school, so 61.2% of adults did not complete even primary school. Table A16.3 - Sample Educational Attainment Level of Education Completed N % None 2,531 31.9 Primary not completed (I to 5) 2,330 29.3 Primary completed (6) 978 12.3 Secondary not completed (7 to 10) 1,076 13.6 Secondary completed (I 1) 448 5.6 Higher education (12 and more) 577 7.3 rOTAL 7,941 100.0 Source: Nicaragua LSMS. 1998. Expansion factor applied. Persons aged 25 and older, not enrolled in school 7. These low education outcomes are originated at the beginning of the education chain, with low net enrollment rates in the schooling system. Net enrollment rates in Nicaragua again vary greatly by area of residence: a smaller proportion of rural school-aged children attend primary school relative to urban ones. Table A16.4 - Nicaragua: Net Enrollment Rates (Percentage of primarv school-aged children attending school) Males Females Total Urban 88.1 89.6 88.8 Rural 72.8 79.5 76.0 otal 80.6 84.8 82.7 Children Aged 6-12. following official school ages. Expansion factor applied. 8. The objective of this paper is to identify the determinants of school attendance in Nicaragua, with special interest on the effect that poverty has on education outcomes. Increasing elementary education enrollment and completion is one of the most relevant education policy issues in Annex 16, Page 3 Nicaragua. In fact, in a recent Ministry of Education (MED) strategy paper, the two final goals of the MED were defined as providing school access to all school-age children in Nicaragua and generating the highest possible schooling outcomes (Arcia and Gillespie, 1998). 9. In recent years, the government has focused on increasing the supply and quality of primarx schools and teachers (especially in rural areas) to increase primary school attendance'. In the past months, a pilot program of income transfers to families conditional on attendance was implemented to address demand-side issues of enrollment. DATA AND METHODOLOGY 10. The data used in this paper is from the Living Standards Measurement Survey (LSMS), carried out in Nicaragua by the Instituto de Estadisticas y Censo (INEC), the World Bank, and other donors in 1998. The survey, which is nationally representative, gathered information from 4,040 households and almost 22,800 individuals. The questionnaire included detailed data modules on housing infrastructure, demographic composition, health, education, employment, fertility, household income/expenditures, agricultural activities, and other topics. These differem modules allowed for the construction of household income and consumption variables, as well as wealth proxies. to study the effect of income/wealth on educational attainment. 11. The official primary school ages are 7 to 12 years; however, due to problems in enforcement, late entry and grade repetition, the relevant age group for this analysis will be defined as children aged 6 to 14 years (these add to 5,835 observations). Creating the dependent variable 12. The LSMS contained a section that gathered the education history of each individual interviewed. Each person was asked: (]) "what is the highest education level and grade achieved?" and (2) "are you enrolled in the current school year?". From questions (1) and (2) it was possible to construct a dummy variable "attend" that took the value of I if the person currently attends school or has attended in the past, and 0 (zero) otherwise, in order to estimate the probability of school attendance4. The regression approach was applied to specify the relationships between the discrete dependent variable ("attendance") and a set of explanatory regressors, with the use of a binary Probit model5. DETERMINANTS OF SCHOOL ATTENDANCE 13. The probability of school attendance was estimated using the following reduced-form specification: Prob(attendance = 1) = a +, * Y + * Z + 0 * X +Y*cost 3Analyzing determinants of enrollment does not necessarily produce insights into education completion. In fact, the education system in Nicaragua has serious repetition and dropout problems, which are separate topics of research. 4Unlike developed countries like the United States, in developing countries, mandatory school attendance is not usually enforced, either because governments lack the institutional capacitv or the political will. Thus, school attendance is a "decision" variable. 5 The selection of a binary choice model -probit or logit- depends on the assumption of the underlying distribution of the disturbances. The probit model assumes normally distributed error terms, while the logit assumes logistic distribution. It is difficult to justify the choice on theoretical grounds, and the results are usually similar (Greene 1997). Annex 16, Page 4 where Y and Z are vectors of the household's income/wealth and demographic characteristics, respectively, and X is a vector of child-specific variabled. The variable "cost" is a proxy for average school fees in the department where a person was located to capture the effect of school costs on enrollment decisions. The variables in the different categories are described in the paragraphs below. Income variables 14. Two proxies for income were built from aggregate consumption data: a continuous per- capita annual consumption aggregate of all household expenditures ("consum", in thousands of cordobas), and a discrete variable "quintX", which is a dummy variable classifying households into five consumption quintiles from "consum" (X takes on values 1-5 for the different quintiles; x=1 is the poorest quintile and was the left-out variable). The indirect proxies for household wealth were a dummy variable for ownership of home ("title"), and a dummy variable "water" for having piped water within the household's premises8. Demographtics 15. The variables to capture household demographic characteristics came from the section of the survey covering housing conditions, household composition, and the education history of the parents of the person interviewed. The variables used were a dummy variable for rural location ("rural"). the number of small children aged 5 years or younger in the household ("nchO_5"), the number of other primary school-aged children in the home ("nch6_14"), and the distance to the nearest primary school in minutes ("distmin"). The dummy variable "institu" (which equals one if a member of the household is involved in a community institution) was included to capture parents' interest in civic activities9. Age, age squared, and gender of the child were the variables for child-specific characteristics. 16. The construction of the variable for education level of the child's parents was not straightforward. The head of each household was responsible for answering the survey, and many homes were composed of three generations of family members. The education section of the questionnaire asked for the detailed education history of each member of the household, thus, by identifying the children in the sample, their parents' education could be constructed. Exact child- parent identification was only possible for children in primary school age who were the son or daughter of the head of the house, so observations with a different relationship to the head of the household were dropped. Two dummy variables: "litfath" and "litmoth" were constructed, which equaled I if father and mother were literate, respectively. Doing this reduced the sample to 3,639 children. 6 The selection of the explanatory variables drew on previous studies for Nicaragua -Kinnon Scott's similar analysis for Nicaragua (using the 1993 LSMS data) and King and Ozler (1998)- and also from a paper on Indian education enrollment by Filmer and Pritchett (1998). 7 The assumption held in the development literature that consumption data is not as subjective as income data and tend to be more reliable was implicitly accepted. 8 "Water" is equal to I if the household has piped water inside the housing unit or outside the unit but within its terrain. 9 The variable "institu" equals one if a member of the household belongs to one or more of the following: neighborhood committee, municipal development committee, women's organization, PTA, savings and loan cooperative, sports club, professional association, religious organization, or "other". Annex 16, Page 5 School costs 17. Each child has to pay a certain amount of fees to enroll in school. Since the data on fees paid was available only for children who attended school, it was not possible to use the exact fee that each child paid as a regressor for all children in the sample, since those who did not attend did not pay for school costs. It was possible to obtain the average fees paid by children in each geographical department. A proxy for school costs was constructed, "fee," which is the ratio of average annual fees paid (by geographical department) to per capita household income. This is an imperfect measure that attempts to capture the average level of costs relative to income incurred by families in each region. The average annual school fees in each department of the country are included in Table 5. Table A16. 5 - Average school fees per child, by geographical Department Ave. Annual School Ave. Annual School Department Fee (CS) Department Fee (CS) Nueva Segovia 128.53 Masaya 72. 12 Jinotega 64.27 Chontales 211.20 Madriz 79.94 Granada 113.33 Esteli 109.75 Carazo 62.85 Chinandega 92.45 Rivas 113.88 Leon 109.49 Rio San Juan 115.19 Matagalpa 74.25 RAAN 81.45 Boaco 82.34 RAAS 106.28 Managua 138.11 Total Average = CS109.47 Source: LSNMS 1998. Hypotheses 18. A priori, we should suspect that higher income and home ownership, i.e., higher wealth, have a positive effect on the probability of school attendance. Better living conditions such as piped water should also affect attendance in a positive way. The sign of the "rural" dummy variable should be negative, due to a less supply of schools in remote areas and/or to the higher opportunity costs for rural (usually agricultural) households of sending a child to school. Parents with higher education should also increase the attendance probability. The number of infants aged 5 or younger may have a negative effect on the school attendance of older kids, because older children may be expected to stay home and help take care of young siblings. The presence of other primary school-aged children should also have a negative effect on attendance because more children are competing for the household's scarce resources. Distance to nearest school should have a negative impact on attendance, especially since most people walk to school in rural Nicaragua. Finally, higher costs of schooling should hinder school attendance. Results 19. The results of the estimation are summarized below in Table 6, which reports the marginal effects (evaluated at the mean) of the explanatory variables on the probability of school attendance for all children ages 6 to 14 (Appendix I contains variables' summary statistics). Regression specification (1) uses income quintiles, while (2) uses the continuous income variable "consum".1° The critical chi-squared statistics for 17 and 14 degrees of freedom are 27.59 and 23.69, respectively. The hypothesis that all coefficients are jointly equal to zero is rejected. ' Standard errors were corrected for heteroskedasticity and clustering about the household. Annex 16. Page 6 Table A16.6 - Marginal effects on the probability of primary| school attendance (1) ~~~(2) Quintile2 # 0.0269 Quintile3 # 0.0479 Quintile4 # 0.0548 Quintile5 # 0.0748 * Consum 0.0049 Title (owns home) # 0.0318 **0.0334 ** ubed water # 0.0389 ** 0.0430 ** Rural # -0.0181 -0.0219 Literate father # 0.0282 ** 0.0298 ** Literate mother # 0.0646 ** 0.0679 ** No. children aged 0-5 -0.0149 ** -0.0159 *e No. children aged 6-14 -0.0027 -0.0055 Distance to school (min) -0.0009 ** -0.0009 ** Institution # 0.0237 ** 0.0244 ** Ate 0.0876 ** 0.0883 ** Age-squared -0.0038 **-0.0038 ** Female # 0.0208 **0.0215 ** Fee/Consum -0.1986 -0.6076 ** Observed P 0.8742 0.8742 Predicted P (at X-bar) 0.9295 0.9301 Log likelihood -1029.43 -1037.93 No. of observations 3,639 3,639 Chi-squared 298.13 272.34 Pseudo R-squared 0.2521 0.2459 *and ** represent statistical significance at the I() and 5 percent level, respectivel3 (#) dF/dx is for discrete change of dummy variable from 0 to 1. Expansion factor was applied. 20. Poorer children are less likely to attend school with both income measures. The results indicate that children from households with higher levels of income have a greater probability of attending school and that the effect is stronger at higher levels: the probability that children from the highest quintile attend school is 7.5 percentage points higher than children from the poorest quintile (left out variable), whereas the probability of attendance for children in the second quintile is only 2.7 percentage points higher than that of the poorest quintile. In specification 2, the continuous income variable has the expected positive sign implying that higher per capita income increases probability of school attendance, but it is not statistically significant. The other proxy for wealth, ownership of home, (significantly) increases the probability of school attendance in both regression specifications by 3 percent relative to families that don't own their homes. 21. An indirect measure of poverty is the quality of housing that families can afford. The probability that a child attends school is about 4 percentage points higher if her home has piped water. This may be explained by the fact that having piped water decreases the time that children would participate in the household chore of gathering water from a well or river, leaving them Annex 16. Pag-e 7 more time for schooling activities". In sum, households in higher income quintiles and with higher wealth are more likely to send their children to school. 22. The geographic location of the household-rural or urban-does not significantly affect the probability of school attendance, although it has the expected negative sign. 23. Parents' literacy has a significant and large effect on the probability of a child attending primary school-and the effect is more than two times as great if the mother is literate than if the father is. Having a literate father increases the probability of attending school by about 3 percentage points. Literacy of a child's mother has a very strong effect on school attendance, In both regression specifications, having a literate mother increases the probability of school attendance by almost 7 percentage points. 24. The presence of infants in the household reduces the probability of a child attending school by about 1.5 percentage points. The presence of other primary school-aged children in the home does not have a significant effect. This seems to indicate that primary school attendance is jeopardized if there are small siblings that need care, but their schooling is not affected by the presence of children of older ages that can take care of themselves. 25. As expected, the distance to the nearest primary school (in minutes) reduces the probability of attendance, although the effect is very small: for every ten minute increase in'travel time, attendance is reduced by 0. 1 percentage point. 26. An interesting result is that parents' involvement in one or more community institutions significantly increases a child's probability of attendance by 2.4 percentage points. This implies that after controlling for their literacy and income levels, parents who are actively involved in communitv or other organizations are more concerned about their children's education. 27. Older children are more likely to attend school: an additional year increases the probability of attendance by almost 9 percentage points. (Note that this effect is positive at a decreasing rate. since the sign of the coefficient for age-squared is negative and significant)' 28. Unlike some developing countries in Asia and Africa, in Nicaragua there is no gender bias against girls in school attendance in Nicaragua. On the contrary, being female significantly increases the probability of attendance by about 2 percentage points in both regressions2. 29. Higher school fees as a proportion of income has a large and negative effect on school attendance, although it is only significant in the second specification. An increase in this ratio reduces the probability of school attendance by 60 percentage points. As mentioned earlier, the variable constructed is an imperfect proxy and ue to the way the variable was constructed it is subject to high measurement error. A more conservative interpretation of the coefficients on the "Fee/Consum" variable is that school costs seem to be negatively related to a parent's decision to send their children to school, but we are unsure of how large this effect is or whether it is statistically significant. " This result is consistent with the results from a study on the determinants of child labor in Nicaragua (llahi 1999): boys and girls aged 10 to 14 are 7 and 5% less likely to work if they have in-house tap water. So, we can conclude that access to tap water has a positive effect on the probability of school attendance because it reduces the demand for child labor in gathering water and frees up children's time to attend school. 12 This, again, is consistent with the results on the child labor study which finds that boys have higher employment rates than girls, thus lowering their attendance rates (see also the results on attendance rates presented in Table 3, above). Annex 16, Page 8 POVERTY AND SCHOOLING 30. The results from the previous section reveal that economic factors such income, wealth and schooling costs play a large and significant role in schooling. Are the determinants of school attendance different for poor families? 31. A categorical variable for poverty groups was created using the poverty line defined for the Nicaragua 1998 LSMS.) All children who fell below the poverty line were classified as "poor" and those who did not were classified as "non-poor." The sample was then split along this poverty classification and Probit regressions were run separately for each group to see if the determinants of attendance are different for poor vs. non-poor children. For ease of exposition, only the results for poor and non-poor children will be presented here. Appendix 2a and 2b contains the results for comparisons across area (urban/rural) and three poverty categories (extreme poor, all-poor and non-poor). In general, the main results are the same when comparing two poverty categories instead of three. 32. From a comparison of the summary statistics for each poverty group (Appendix 3), we observe that about 81% of poor school-aged children attend school while 96% who are non-poor attend. The demographic composition of households differs along poverty categories: more persons, and specifically, more children in the age groups 0-5 and 6-14 live in each average poor household. The average per-capita annual income in poor homes was C$2.574 (about US$243), whereas in poor homes it was C$9,150 (US$865). Other proxies of wealth also differ for the two sub-groups: 78% of poor homes legally own their house, whereas 82% of non-poor persons do, and 36% of poor children in the sample have piped water, compared to 78% non-poor. The average distance to school in minutes is greater for poor relative to non-poor children. Parents of non-poor families are more often literate than poor ones. School fees per child as a proportion of per capita income are slightly higher for poor families. 33. Many of these results could be due to the fact that poor families are more often living in rural areas -68% of the poor live in rural vs. 28% of non-poor. 34. The results of the regressions for each sub-group follow in Table 7. Interesting differences in the determinants of attendance appear when the sample is split. The predicted probability of attending school is about 13 percentage points less for poor relative to non-poor children (85% to 98%). Income is a determining factor of school attendance for poor children (with both discrete and continuous variables), but not for the non-poor. Having piped water has a much greater effect on poor children's attendance probability-by almost three times compared to non-poor. Home ownership has an ambiguous effect, since it increases the probability of attendance of poor children but REDUCES the likelihood for the non-poor. Income proxies, as well as the indirect proxies for wealth, are more relevant in poor families' schooling decisions, which indicates that the poor are liquidity constrained. 13 The full poverty line for Nicaragua in 1998 is C$4,259 (about US$403) per person per year. A person is considered poor if his/her level of annual expenditure is below the poverty line. Annex 16. Page 9 Table A16.7 - Marginal effects on the probability of primary school attendance, by poverty group All-Poor Non-Poor (1) (2) (1) (2) uintile2 4 0.0545 . Quintile3 4 O.1011 ** -0.0042 Quintile4 . 0.0003 Quintile5 # Consum 0.0556 ** -0.0003 Title (owns home) 0.0924 ** 0.0929 ** -0.0177 ** -0.0178 Tubed water # 0.0567 * 0.0576 ** 0.0202 ** 0.0208 ** Rural # -0.0283 -0.0240 -0.0066 -0.0072 Literate father # 0.0368 0.0361 0.0186 * 0.0195 * Literate mother 0 .0985 ** 0.0997 ** 0.0241 * 0.0256 ** No. children aged 0-5 -0.0177 * -0.0164 0.0100 ** -0.0104 ** No. children aged 6-14 0.0012 0.0005 0.0055 -0.0058 Distance to school (min) -0.0016 ** -0.0016 * Lo0.0003 * -0.0003 ** Institution 0 .0358 0.0365 10.0112 0.0108 Acre 0.1619 ** 0.1624 **K0263 ** 0.0262 ** gAe-squared -0.0069 ** -0.0070 ** -0.0012 ** -0.0012 ** Female 0 (.0343 0 (.0342 ** i.0061 0.0061 Fee/Consum -0.3563 0.0125 -0.3229 -0.4902 Observed P 0.8101 0.8101 0.9603 0.9603 Predicted P (at X-bar) 0.8543 0.8540 0.9836 0.9834 Log likelihood -897.78 -897.81 -169.35 -169.33 No. of observations 2.306 2.306 1.333 1.333 Chi-squared 202.78 202.78 73.56 70.62 Pseudo R-squared 0.199 0.199 10.2391 0.2392 . variable dropped due to roilti-collinearitv. * and ** represent statistical significance at the 10 and 5 percent level, respectively. Expansion factor applied. # dF/dx is for discrete change of dummy variable from 0 to I. 35. Literate fathers increase the probability of attendance of non-poor kids, but they do not affec t schooling in poor homes. Literacy of a child's mother increases the chances of schooling for both poor and non-poor, but note that the effect is more than four times as great for the poor compared to non-poor homes. This implies that when homes are poor and face tight budget constraints, educated mothers are more appreciative of the benefits of education and are more willing to make the necessary sacrifices so that their child goes to school. 36. The presence of children aged 0-5 significantly decreases the likelihood of attendance of a primary-aged boy or girl by 1-2 percent, regardless of poverty category. The effect is stronger in non-poor households, where it is more likely that both mother and father work, implying the need for older children to help with childcare. A puzzling result is that the presence of other primary aged children decreases the likelihood of attendance in non-poorfamilies, the opposite of what would be expected. 37. Finally, it is interesting to note that poor girls are more likely to go to school than poor boys, which points to discrimination in demand for labor of boys that is taking them away from primary schooling. This effect is not significant in non-poor households. This result is corroborated in Appendix 3, which lists the reasons for children not attending school: 9.6% of poor boys who did not attend school cited "Work/Agricultural labor" and 4.2% of poor girls said Annex 16, Page 10 they had to help with domestic jobs-but no non-poor boys or girls cited any type of work as a reason for not attending'4. CONCLUSIONS AND POLICY LESSONS 38. The most relevant conclusion from this study is that poverty and school attendance are negatively related: the predicted probability of attending primary school is lower for poor relative to non-poor children by about 13 percentage points. Income and wealth variables, measured directly and indirectly, have a significant effect on poor households but not on non-poor ones. These results indicate that economic resources are significant determinants of poor children's education. It is probably the case that in subsistence agricultural settings (most of the rural poor in Nicaragua), and in dire urban poverty, children's time is a valuable input to household production and income, so that a child's time dedicated to schooling has a very high opportunity cost. 39. In a study on poverty and the education sector, Arcia (2000) finds that private per-student expenditures for primary education represent 22 percent of poor households' per-capita non-food expenses, compared to 15 percent for the non-poor. Thus, private schooling costs per child represent a large proportion of families' expenditures. Of total costs, uniforms, school supplies, and transportation/food represent 41. 23, and 21 percent of total costs, respectively. 40. The demographic composition of households affects poor families differently from non- poor ones: literacy of fathers affects only the non-poor. Younger siblings, and siblings in the same age-cohort significantly reduce likelihood of schooling in non-poor households. Finally, only in poor homes are girls more likely to attend school than boys. Mother's literacy increases the probability of attendance of both poor and non-poor, and as mentioned above, the effect is more than four times greater in poor homes. 41. Poor children's school attendance is lower due to two general categories of determinants on the demand side": adverse economic situations of families and the lower human capital stock of fathers and mothers (see Appendix 3)-furthermore, these two are likely to be inter-related. It seems that in Nicaragua, the poor are stuck in a "bad Nash equilibrium" that perpetuates itself across generations: education levels of the poor are low- uneducated people can't find better employment opportunities and thus cannot afford to send their children to school e children from poor families don't obtain an education and grow up limited to low productivity jobs - children from poor families remain in poverty. 42. If Nicaragua is to reach the international goal of universal primary school enrollment in rural areas by the year 2015, policies must aggressively address the two channels that are detaining school attendance. Some policy lessons are summarized below. a. Primary scitooling must be economically affordable to all households. This should address both demand and supply sides of the question. On the demand side, a pilot program has been recently implemented by the GON of income transfers to extremely poor families conditional on children's school enrollment. This should have a positive effect on enrollment of poor children because their economic possibilities increase with this subsidy. An education policy aimed at boosting enrollment should consider the fact that more than 80 percent of private education costs for primary school go to uniforms, school 14 Appendix 3 lists "illness" as a more common reason for non-attendance among poor children compared to non-poor. '5 Serious problems exist in the provision of adequate school supplies as well; the focus of this paper has been on variables affecting the household-level decision to send children to school. Annex 16. Page 11 supply, and transportation expenses, perhaps by reducing or eliminating value-added taxes on these items or subsidizing part of these costs for the most poor. Mother's education is a significant determinant in sch0ool enrollment, especially of poorfamilies. Although this variable is a human capital "stock", adult education programs such as existing PAEBANIC'6 should be sought out and promoted to increase the education levels of at least two generations of families. Adult education has a potentially immediate effect on family income by the wider employment opportunities t provides adults, and long-term effects through the increased the likelihood of children staying in school if their parents are literate. b. Chiildcare of young children may have an important effect on primary school attendance. This issue can be addressed through increased pre-school enrollment, which not only provides "care" for children aged 2-5 years and thus "frees up" time of the older siblings in primary school age- it probably also increases the likelihood (and performance) of these toddlers attending primary school. A policy along these lines has been followed in Nicaragua in the recent past and slhould be continued. The MEDC has organized community preschools (which are more effective than formal preschools) by training mothers as preschool caregivers, training them in nutritional assistance and preventive health care (from Arcia 2000). The strategy has paid off: the percentage of children aged 4-6 enrolled in pre-school increased from 3 to 45 percent between 1993 to 1998; among the extreme poor, this increase was more than threefold, from 8.3 to 25.4 percent. The proportion of 0-3 year olds in preschool is still very low (Arcia 2000). c. Boys in poor homes are particularly at risk of not attending school. This is related to economic conditions and the need for the additional income of boys or for their labor in household production. This problem is partially addressed by income transfers conditional cn school enrollment of the type described above. 16 "Programa de Alfabetizaci6n y Educaci6n Basica de Adultos de Nicaragua", funded by Cooperacion Espanola, Govemment of Spain. Closing date is in the year 2003. Source: Arcia 2000. 17 This is consistent with results from the background paper on labor markets. See Ilahi 1999. Annex 16, Page 12 REFERENCES Arcia, Gustavo, May 2000. "Education and Poverty in Nicaragua: Evidence from the 1993 and 1998 Surveys on Living Standards." Consultant Report, The World Bank. Washington, D.C. Arcia, Gustavo and Nancy Gillespie, August 1998. "Estrategia a Mediano Plazo del Ministerio de Educaci6n de Nicaragua 1998-2003." Consultant Report to the Ministry of Education of Nicaragua. Filmer, Deon and Lant Pritchett, 1998. "Estimating Wealth Effects without Expenditure Data-or Tears: An application to Educational Enrollments in States of India," mimeo. The World Bank. Washington, D.C. Filmer, Deon and Lant Pritchett, 1998. "The Effect of Household Wealth on Educational Attainment Around the World: Demographic and Health Survey Evidence," mimeo. The World Bank. Washington, D.C. Gargiulo, Carlos A. and Crouch, Luis A. 1995. "Nicaragua: Schooling, Repetition, Dropouts. Results of a National Study." Policy Paper Series, Research Triangle Institute. Research Triangle Park, N.C. Green, William H., 1997. Econometric Analysis. Third Edition. New Jersey: Prentice- Hall, Inc. Ilahi, Nadeem, 1999. "An Analysis of Labor Markets and Time use in Nicaragua," draft, background paper for the 1998 Nicaragua Poverty Assessment. The World Bank. Washington, D.C. King, Elizabeth and Berk Ozler, 1998. "What's Decentralization Got To Do With Learning? The Case of Nicaragua's School Autonomy Reform." Working Paper Series on Impact Evaluation of Education Reforms, The World Bank. Washington, D.C. Scott, Kinnon, 1995. "The Determinants Of Educational Attainment In Nicaragua," mimeo, PRD/PH. The World Bank. Washington, D.C. Annex 16. Page 13 Appendix A16.1 - Summary Statistics (Sample of children ages 6-14) Variable Obs Mean Std. Dev. Min Max Attend 4,453 0.88 0.33 0 1 No. persons in household 4,453 6.96 2.65 1 19 No. children aged 0-5 4,453 1.14 1.12 0 7 No.childrenaged6-14 4,453 2.14 1.12 0 7 Quintile2 4,453 0.19 0.39 0 1 Quintile3 4,453 0.22 0.41 0 1 Quintile4 4,453 0.20 0.40 0 1 Quintile5 4,453 0.19 0.39 0 1 Consum' 4.453 5.537 7,213 239 155.691 Owns home 4.453 0.80 0.40 0 1 Tubed water 4.453 0.55 0.50 0 1 Distance to school (minutes) 4.439 17.41 28.64 0 540 Rural 4,453 0.50 0.50 0 ] Female 4-453 0.50 0.50 0 1 Litfath 3.789 0.76 0.43 0 1 Litmoth 4,312 0.72 0.45 0 1 Fees 4,453 0.03 0.03 0.00 0.45 ,Source: LSMS 1998. Expansion factor applied. lUnits: cordobas. Annex 16, Page 14 Appendix A16.2a: Marginal effects on the probability of primary school attendance (Consumption quintiles as income proxy) National Urban Rural Ext.Poor All-Poor Non-Poor quint2 # 0.0269 -0.0109 0.0638 ** 0.1287 * 0.0545 * quint3 # 0.0479 ** 0.0002 0.0950 ** 0.1011 ** 0.0042 quint4 # 0.0548 ** 0.0214 0.0851 ** -0.0003 quint5 # 0.0748 ** 0.0331 0.1115 ** title # 0.0318 ** -0.0169 0.0913 ** 0.1519 ** 0.0924 ** -0.0177 tubwater f 0.0389 ** 0.0049 0.0941 ** 0.0452 0.0567 * 0.0202 rural # -0.0181 -0.1470 ** -0.0283 0.0066 litfath # 0.0282 ** 0.0089 0.0502 ** 0.0320 0.0368 0.0186 litrnoth # 0.0646 ** 0.0152 0.1247 ** 0.1860 ** 0.0985 ** 0.0241 nchO_5 -0.0149 ** -0.0024 -0.0278 ** -0.0213 -0.0177 * 0.0100 nch6_14 -0.0027 -0.0015 -0.0039 0.0027 0.0012 -0.0055 distmin -0.0009 ** -0.0001 -0.0015 ** -0.0016 ** -0.0016 ** 0.0003 institu # 0.0237 ** 0.0214 ** 0.0338 0.0264 0.0358 0.0112 age 0.0876 ** 0.0377 ** 0.1530 ** 0.1694 ** 0.1619 ** 0.0263 agesq -0.0038 ** -0.0016 ** -0.0065 ** -0.0068 ** -0.0069 ** -0.0012 female f 0.0208 ** 0.0157 ** 0.0240 0.0337 0.0343 ** 0.0061 fee -0.1986 -0.4821 -0.0078 -0.6198 -0.3563 -0.3229 Observed P 0.8742 0.9476 0.8094 0.7113 0.8101 0.9603 Predicted P (at X-bar) 0.9295 0.9677 0.8716 0.7440 0.8543 0.9836 Lo, likelihood -1029.43 -280.15 -745.48 -493.25 -897.78 ]169.35 No. of observations 3,639 1.597 2.042 987 2,306 1,333 Chi-squared 298.13 69.28 240.67 104.24 202.78 73.56 Pseudo R-squared 0.2521 0.1464 0.2504 0.1684 0.199 0.2391 * and ** represent statistical significance at the 10 and 5 percent level, respectively (#) dF/dx is for discrete change of dummy variable from 0 to 1. "." Means the variable was dropped due to multicollinearity. Annex 16, Page 15 Appendix A16.2b: Marginal effects on the probability of primary school attendance (Consumption as income proxy) National Urban Rural Ext.Poor All-Poor Non-Poor consum 0.0049 0.0025 0.0119 0.2803 ** 0.0556 ** -0.0003 title # 0.0334 ** -0.0161 0.0921 ** 0.1469 ** 0.0929 ** -0.0178 tubwater # 0.0430 ** 0.0056 0.0977 ** 0.0478 0.0576 ** 0.0208 rural # -0.0219 -0.1248 ** -0.0240 -0.0072 litfath # 0.0298 ** 0.0096 0.0509 ** 0.0330 0.0361 0.0195 litmoth # 0.0679 ** 0.0191 0.1252 ** 0.1790 ** 0.0997 ** 0.0256 nch0_5 -0.0159 ** -0.0036 -0.0295 ** -0.0120 -0.0164 -0.0104 nch6_14 -0.0055 -0.0027 -0.0085 0.0050 0.0005 -0.0058 distmin -0.0009 ** 0.0000 -0.0015 ** -0.0018 ** -0.0016 ** -0.0003 institu # 0.0244 ** 0.0215 ** 0.0345 0.0254 0.0365 0.0108 age 0.0883 ** 0.0372 ** 0.1567 ** 0.1765 ** 0.1624 ** 0.0262 agesq -0.0038 ** -0.0016 ** -0.0067 ** -0.0071 ** -0.0070 ** -0.0012 female # 0.0215 ** 0.0151 * 0.0267 * 0.0287 0.0342 ** 0.0061 'fee -0.6076 ** -0.5867 ** -0.5822 2.0229 * 0.0125 -0.4902 Observed P 0.8742 0.8742 0.8094 0.7113 0.8101 0.9603 Predicted P (at X-bar) 0.9301 0.9301 0.8729 0.7493 0.8540 0.9834 Log likelihood -1037.93 -284.56 -752.88 -480.32 -897.81 - 1 69.33 No. of observations 3,639 1,597 2,042 987 2,306 1.333 Chi-squared 272.34 56.68 223.88 117.07 202.78 70.62 Pseudo R-squared 0.2459 0.133 0.243 0.1902 0.199 0.2392 * and ** represent statistical significance at the 10 and 5 percent level, respectively (#) dF/dx is for discrete change of dummy variable from 0 to 1. Annex 16, Page 16 Appendix A16.3 - Summary Statistics by Poverty Category (Sample of children ages 6-14) Poor Non-Poor Obs Mean Std. Min Max Obs Mean Std. Min Max Attend 2,734 0.81 0.39 0 1 1,719 0.96 0.20 0 1 No. persons in household 2.734 7.97 2.71 1 19 1,719 5.73 1.97 1 16 No. children aged 0-5 2,734 1.51 1.20 0 7 1,719 0.69 0.82 0 6 No. children aged 6-14 2,734 2.43 1.14 0 7 1,719 1.77 0.97 0 5 Quintile2 2,734 0.35 0.48 0 1 1,719 0.00 0.00 0 0 Quintile3 2,734 0.28 0.45 0 1 1,719 0.13 0.34 0 1 Quintile4 2734 0.00 0.00 0 0 1,719 0.45 0.50 0 1 Quintile5 2,734 0.00 0.00 0 0 1,719 0.42 0.49 0 1 Consum 2V,734 2,574 920 239 4,259 1,719 9,150 9,524 4.261 155,691 Owns home 2.734 0.78 0.42 0 1 1,719 0.82 0.38 0 1 Tubed water 2,734 0.36 0.48 0 1 1,719 0.78 0.41 0 1 Distance to school (minutes) 2,725 21.17 35.52 0 540 1,714 12.82 15.59 0 210 Rural 2,734 0.68 0.47 0 1 1,719 0.28 0.45 0 1 Female 2.734 0.49 0.50 0 1 1,719 0.52 0.50 0 1 Litfath 2,386 0.65 0.48 0 1 1,403 0.90 0.30 0 1 Litmoth 2,658 0.59 0.49 0 1 1,654 0.89 0.32 0 1 Fees 2734 0.04 0.03 0.01 0.45 1,719 0.02 0.01 0.00 0.05 ~ource: LSMS 1998. Expansion factor applied. Units: c6rdobas. Annex 16, Page 17 Appendix A16.4 - Reasons for Children not Attending School (Sample of children ages 6-14) Number of Responses Poor Non-Poor TOTAL hfale Female Total Mfale Female Total Male Female Total Age 21 11 32 2 6 7 23 17 39 Economic problem 153 127 279 4 6 40 187 133 319 WorkWAg. Labor 31 4 35 0 0 0 31 4 35 Domesticjobs I I ] 12 0 0 0 1 I 1 12 Not interested 29 16 45 5 4 8 34 20 53 No school nearby 54 60 113 10 8 18 64 68 131 illness 4 4 8 I 1 1 5 5 9 No capacity in school 0 1 1 3 0 3 3 4 1 Grade notoffered 2 3 4 1 0 1 3 4 4 Lackofteachers 12 12 25 4 2 6 16 14 31 School is unsafe I 1 2 0 0 0 1 1 2 Lack of books in school 1 0 1 0 0 0 1 0 1 Handicapped 5 7 11 9 1 10 14 8 21 Other 9 5 14 8 4 12 17 9 26 Total 322 261 583 76 31 107 398 292 690 Percent of Total Poor Non-Poor TOTAL Male Female Total Male Female Total mlale Female Total Age 6.5 4.2 5.5 2.6 19.4 6.5 5.8 5.8 5.7 Economic problem 7.5 48.7 47.9 44.7 19.4 37.4 47.0 45.5 46.2 Work/Ag. Labor .6 1.5 6.0 0.0 0.0 0.0 7.8 1.4 5.1 Domestic jobs 0.3 4.2 2.1 0.0 0.0 0.0 0.3 3.8 1.7 Not interested 9.0 6.1 7.7 6.6 12.9 7.5 8.5 6.8 7.7 Noschool nearby 16.8 23.0 19.4 13.2 25.8 16.8 16.1 23.3 19.0 Illness 1.2 1.5 1.4 1.3 3.2 0.9 1.3 1.7 1.3 o capacity in school 0.0 0.4 0.2 3.9 0.0 9.7 0.8 1.4 0.1 Grade not offered 0.6 1.1 0.7 1.3 0.0 3.2 0.8 1.4 0.6 Lack ofteachers 3.7 4.6 4.3 5.3 6.5 5.6 4.0 4.8 4.5 School is unsafe 0.3 0.4 0.3 0.0 0.0 0.0 0.3 0.3 0.3 Lack of books in school 0.3 0.0 0.2 0.0 0.0 0.0 0.3 0.0 0.1 Handicapped 1.6 2.7 1.9 11.8 3.2 9.3 3.5 2.7 3.0 Other .8 1.9 2.4 10.5 12.9 11.2 4.3 3.1 3.8 Total 100.0 100.0 100.0 1]00.0 100.0 100.0 100.0 100.0 100.0 Annex 16, Page 18 Appendix A16.5 - Education Costs, 1993 - 1998 Category 1993 1994 1995 1996 1997 1998 1993-1 998 Total Education 99.5 100.0 102.9 104.2 112.8 126.8 27.3 Tuition, Fees, and Other 98.9 100.0 102.0 103.3 112.9 128.4 29.5 Primary school tuition 100.0 100.0 100.0 100.0 109.0 124.6 24.6 Primary school - monthly fees 89.9 100.0 119.2 131.0 145.9 160.6 70.7 School Materials 103.8 100.0 109.3 111.3 112.1 114.3 10.5 Notebooks 102.0 100.0 120.5 118.3 107.4 102.2 0.2 Pencils 104.8 100.0 92.0 101.3 118.9 117.2 12.5 Pens 106.0 100.0 101.2 105.9 115.6 131.2 25.2 General CPI 92.1 100.0 110.9 123.9 135.2 152.9 60.7 Index: 1994=100. |Source: BCN and INEC. Annex 17, Page I Annex 17 - Demand for Health Care in Nicaragua, 1998 by Mukesh Chawla EXECUTIVE SUMMARY i. The Nicaraguan poor experience higher levels of illness, are less likely to seek care. and spend a greater proportion of their income on health care. Demand for health care responds to prices, and demand for health care by the poor is more responsive compared to the non-poor. This finding suggests that a price increase is likely to reduce use of health care by the poor more than by the country as a whole. ii. Knowledge of household demand for health care in developing countries is important for a variety of reasons. Such knowledge can improve allocation of scarce resources in the health sector; the organization of health care systems; the government's role in the production, finance, delivery, and management of health services: the private sector's role and governmental policies encouraging its growth; cost-sharing goals for publicly provided medical services; and decisions regarding facility design, scope of services, intensity of services and location. iii. This paper explores the determinants of health-care-seeking behavior and choice of provider in Nicaragua. The decision to seek care is modeled jointly with the choice of provider in a nested decisionmaking framework, for the country as a whole and for the seven administrative regions. Price and income elasticities are computed for different providers and across income groups. iv. The most important findings are: a. A little more than a third of Nicaraguans reported illness in the month before the study, but less than half of them seek health care. Disparities between the poor and non-poorare vast, and even though there are no significant differences in the incidence of illness, the non-poor are about 30 percent more likely to seek care when ill than are the overall poor, and 50 percent more likely than are the extremely poor. The disparities are even more pronounced at extreme ends of the income distribution, with the richest 10 percent twice as likely to seek care as the poorest 10 percent. The likelihood of seeking care increases rapidly with small increments of income at lower income levels, but at middle-income levels, marginal increments in income increases this likelihood only slightly. b. Rural-urban disparities in health care are also quite pronounced. The incidence of illness among the rural population is significantly higher in rural areas, yet significantly fewer seek care compared to the ill in urban areas. Assuming that all those who seek care recover from their illness, at any time the rural population is 12 percent more likely to have untreated ill people than is the urban population. c. Regional disparities are significant as well. The incidence of illness is highest in the Atlantic Rural region, yet the proportion of people in the Atlantic Rural region seeking treatment when ill is lowest. Similarly, the incidence of illness is high in the Central Rural and Pacific Rural regions, and both have significantly lower levels of treatment. In general, Managua and the urban regions report lower levels of illnesses and higher levels of treatment. Disparities in health care are most significant across poverty groups within regions. The extremely poor in Annex 17, Page 2 the Atlantic Rural region are less than half as likely to seek care when ill compared to the non-poor in Atlantic Urban region. d. To some extent, disparities in health care across regions can be explained bv the general economic and health status of people living in these regions. Managua has the highest average real household and per capita incomes, highest food, health, and education expenditures. and lowest incidence of illness. These indicators are worst in the Atlantic Rural and Central Rural regions. e. Disparities in health care between males and females are not pronounced. While more females report an event of illness compared to males, more women also seek care when ill compared to men, so that women's general health status is no likely to be worse off compared to men. Extremely poor males appear to be the worst off, and less than a third of them seek care when ill. In contrast, females in non-poor households are almost twice as likely to seek care. There appear to be no differences in illness and treatment according to household headship. f. In general, people in Nicaragua prefer to visit trained physicians when ill and seeking treatment. However, differences are significant across regions and poverty groups. For instance, those ill and seeking treatment in the Atlantic Rural region are only half as likely to visit trained physicians, preferring instead nurses, nurse assistants, community health workers, and even untrained physicians. Most of those ill prefer going to health centers and private clinics for most types of illness, the only exception being that accident cases are more likely to be treated in hospitals than elsewhere. Private clinics are used significantly more in urban areas than in rural areas, so much so that patients in the Central Urban region are over five times more likely to visit private clinics compared to those in the Atlantic Rural region. g. Disparities in health care use by provider type and poverty groups are also vast, with the non- poor much more likely to visit trained physicians than the extremely poor. Thus, not only di: members of extremely poor households report higher levels of illness, they seek treatment less often. Of those that do, a large number go to nurses, community health workers and untrained physicians. h. The most important reason that the ill do not seek care is that the illness is considered too slight. Cost is important to some, mostly to the extremely poor, who are twice as likely to not seek care than the non-poor. Distance is an important impediment to the extremely poor and the overall poor, but not to the non-poor. Distance prevents a very large number of the ill from getting treatment in the Atlantic Rural regions and to a smaller extent in the other rural regions. i. Health insurance does not play a significant role in the health care system, with a little over 7 percent of households reporting any kind of insurance. As would be expected, insurance coverage is higher among the employed than the unemployed and is predominantly higher among the rich than the poor. j. More than one-fourth of all ill and seeking treatment do not incur any health care expenditure at all. Of the others, visits to the physician cost the most, followed by pharmacies. Private clinics are the most expensive institutions, followed by hospitals. Health care following accidents cost the most, followed by visits for illnesses other than measles, coughs and colds, and diarrhea. Annex 17, Page 3 v. As a proportion of income, the poor report high out-of-pocket expenditures on health care. Out-of-pocket expenditure is regressive across income levels, and the non-poor spend a considerably smaller proportion of their income than the overall poor and the extremely poor. Annex 17. Page 4 INTRODUCTION 1. Knowledge of household demand for health care in developing countries is important for a variety of reasons. First, allocation of scarce resources in the health sector can be made more effectively if the structure of demand for health services is known. Second, critical choices about the organization of health care systems depend crucially on consumer preferences. Third, the role of the government in the production, finance, delivery and management of health services can be appropriately defined only if the factors affecting the demand for health care are known and understood. Fourth, the role of the private sector and the government policies towards encouraging and supporting its growth can be appropriately defined only if consumer decisions about medical treatment and choice of providers are known. Fifth, knowledge of the demand for health care is important to set cost sharing goals for the publicly provided medical services. Finally, decisions regarding facility design, scope of services, intensity of services and location would be better informed if the patterns of utilization of health services were known. 2. This paper explores the determinants of health-care-seeking behavior and choice of provider in Nicaragua. The decision to seek care is modeled jointly with the choice of provider in a nested decision-making framework, for the country as a whole and for the seven administrative regions. Price and income elasticities are computed for different providers and across income groups. Finally, implications for household willingness to pay for health care are drawn.' 3. To focus on the health status and utilization of health services by the poor, we analyze trends in the incidence of illness, treatment-seeking behavior, choice of providers, and burden of out-of-pocket expenditure on health by poverty classification. To classify households according to poverty groups, the income ranges are set at less than C$2,246 for extremely poor, at less than C$4,259 for overall poor, and more than C$4,259 for non-poor.2 Using these cutoffs, 17 percent of the sample population of households is extremely poor, 48 percent is overall poor, and 52 3 percent is non-poor. 4. The rest of the report is organized as follows. A discussion on modeling the demand for health care is placed in section 2. Section 3 contains a review of the literature, followed by a description of the data sources in section 4. Results of the estimation are contained in section 5, and the paper ends with a discussion in section 6. All tables and figures are placed in the appendix. MODELING THE DEMAND FOR HEALTH CARE 5. Demand for health care can be defined as the quantity of a particular type of service thatpeople are willing to obtain over a given period of time. More important than quantity of health care, however, is the discrete phenomenon of seeking care. In this specification, the values taken by the dependent variable are merely a coding for some qualitative outcome, where the mutually exclusive choices may be "seek treatment from providerf' and "seek treatment from provider k." The choice of provider would naturally be conditional on the decision to "seek ' A previous study (Chawla, 1999) focuses on the incidence of illness, health utilization patterns, insurance coverage, household spending on health care, and the determinants of seeking health care in Nicaragua. 2 Note that the 'extremely poor' category is a subset of the 'overall poor' category. 3Individual and household level expansion factors (weights), as applicable, are used in computing the results reported here and in the remainder of the analysis. Annex 17, Page 5 treatment", which in turn would be conditional on being ill. Consumer decisions are based on maximizing utility, which depends on the individual's health status after consumption of the health good as well as on consumption of other goods. Estimation of demand thus takes the form of estimating these marginal and conditional probabilities.4 6. Formally, let the expected utility conditional on receiving care from providerj be defined 5 as Uj = U(Hj, Cj) (1) where H. is the expected health status after receiving treatment from providerj and Cj is the consumption net of the cost of receiving care from providerj. As Gertler, Locay, and Sanderson hiave shown, income can influence the choice only if the conditional utility function allows for a non-constant marginal rate of substitution of health consumption6 Following Gertler and van der Gaag (1 990) we use a functional form in which utility is linear in health and quadratic in consumption.7 Specifically, we express the utility function for the "seeking care" alternatives as: Ujj = a0Hj + QI(Yi - Pj) + ct2 (Yi - Pj)2 + Ei (2) When no care is sought, (2) reduces to UO = c0Hj + CCIYi + °c2 yi2+ Eo (3) 7. As Alderman and Gertler (1997) note, the parameters in the equations 2 and 3 are identified only when the values of expected health and consumption vary across the alternatives 8. The quality of health care providers is not introduced so far in the model. We do so by defining quality of providerj as the difference betwveen expected health outcome from the jth provider and self-care, and express quality as Q( = Hj - Ho (4) Substituting into (2) yields 4 In the theoretical and empirical framework that follows, we restrict ourselves to determining the probability of seeking care conditional on an event of illness. To this extent, therefore, the estimated price elasticities may be considered to be short-term elasticities that may differ from long-term elasticities if the probability of reporting an illness is responsive to prices. Dow (1995) presents a case in which the short- tern and long-term responses are not significant, and shows that there is in fact no sample selection bias ir using a sample conditional on illness. Other researchers have also used such a framework (see, for instance, Gertler and van der Gaag, 1990, Lavy and Quigley, 1993, and Lavy and Germain, 1994). For a specification that uses the full sample in a sequential decisionmaking framework, see Chawla and Ellis, (2000). 5The framework that we adopt for the analysis closely follows the models developed in Gertler, Locay, and Sanderson (1987), Gertler and van der Gaag (1990), Lavy and Germain (1994), and Alderman and Gertler (1997), 6 This is also consistent with the notion of health being a normal good. 7Other functional forms that have been used are the translog indirect utility function (Gertler, Locay, and Sanderson, 1987), and the Cobb-Douglas function (Lavy and Germain, 1994). Annex 17, Page 6 Uij = cLo(Qj + Ho) + C1(Y - Pj) + a2 (Yi - Pi)2 + Ej (5) Normalizing quality of self-care to zero, the utility from the self-care alternative reduces to: UO = aooHo + ciY + a2Y, + o (6) 9. Estimating (5) poses the problem that quality is not directly observable. We address this issue by letting quality of health care providerj depend on the characteristics of providerj as well as on the characteristics of those seeking treatment, insofar as their personal ability to implement the recommended treatment affects the quality of care they obtain. Defined thus, quality is a function of such parameters as age, gender, education, marital status, and family size. 10. Following Alderman and Gertler (1997), we define a reduced-form model of the utility from quality as: cLOQJ = Eoj+ PIX + P1Zj + nj (7) where X is a vector of demographic variables, Zj is a vector of characteristics that do not enter the budget function, and im is a random disturbance term with mean zero and finite variance. Substituting (7) into (5) produces: UJ = V + nj + £j (8) where V, = aOIO3 + cLOI3Xj + aOa2jZj + a(Y, - Pi) + a2 (Yi - pj)2 (9) IH. This model can be estimated if the stochastic distribution of the error term is known. For the purposes of this study we assume that the error terms take on a nested multinomial logit form. In this specification, the probability of not seeking care is defined as: 12. Prob (no treatment) = exp (Vo)/{exp (Vo)+[Ejo> exp (Vi/CT)]A} (10) and the probability of seeking care from providerj is defined as: (I 1) Prob (seeking care from providerj) = [l-Prob (no treatment)][exp (Vj/a)] / Fj,O exp (Vj/a) where a is a coefficient of dissimilarity between the "no treatment" and "seeking care from providerj" altematives. As demonstrated by McFadden (1981), this coefficient must be between 0 and I for the model to be consistent with utility maximization. If cy1, all alternatives, including that of no treatment, are treated as equally close substitutes, and the nested aspect of the model disappears. Annex 17. Page 7 REVIEW OF THE LITERATURE 13. Previous studies on household and individual demand and utilization of health care in developing countries indicate that factors such as age, gender, income, education, location, and price of health care affect the reporting of an illness, seeking of health care, and choice of provider (see, for instance, Ii, 1996, de Buthane et al, 1989, Mwabu et al, 1991, Huber, 1993). While some studies find the demand for health care to be fairly price-inelastic, others report that price increases significantly lower demand for health care (see, for instance, Litvack and Bodart. 1993). Factors that increase utilization include convenience or ease of physical access to health services, improvements in health care quality, and household income. Some studies have distinguished between factors that influence the decision to report an illness and those that influence the decision to seek care. They find that the probabilitv of reporting an illness increases with income and education and decreases with family size, but the probability of seeking treatment increases principally with perception of quality, ease of access, and income (see, for instance, Chawla and Ellis, 2000). 14. Previous studies on the demand for health care have typically found small price effects. In a study of demand for health care in rural Malaysia, Heller (1982) found that total annual medical visits were not significantly influenced by prices. Similarly, Akin et al. (1984) concluded that prices had little effect on demand for medical care in a rural region of the Philippines. In a study analyzing the determinants of demand for health care in urban Bolivia. Ii (1996) found tha-: though demand for medical care is responsive to changes in prices, the price elasticities tend to he very low. 15. Other studies have found that cost recovery decreases use of health services. In a study of the effect of a price increase in health facilities in Zaire, de Bethune et al (1989) found a decrease in utilization after an abrupt increase in prices. Waddington and Enyimavew (1989) found a general fall in utilization of health facilities following a price increase in the Ashanti-Akim district in Ghana. Mwabu et al (1991) estimated that utilization of health facilities dropped by 38 percent during 1989-90 in the four facilities they examined, though Huber (1993) attributes mucl of the drop in demand to insufficient use of exemptions. However, there is some evidence of improvement in access to health facilities following increases in both costs and quality of care, as found by Litvack and Bodart (1993) in Cameroon. DATA SOURCES 16. In 1998, the Government of Nicaragua conducted its second Living Standards Measurements Survey (LSMS 98), the Encuesta Nacional de Hogares Sobre Medicion de Niveh;s de Vida. Administered by the Instituto Nacional de Estadisticas y Censos (INEC), the survey collected information from a nationally representative sample of 4,656 households on a variety .of household and individual characteristics, and compiled data on such variables as demographic, education, economic, health care utilization, employment, time-use, consumption, and agriculture. The present report uses data generated by LSMS 98 to document the incidence of illness, patterns of utilization of and expenditure on health care, and to identify and evaluate theeffect of factors influencing the demand for health care in Nicaragua. Annex 17, Page 8 ESTIMATION AND RESULTS 17. Two statistical methods are used: univariate comparisons of sample statistics across regions and over time, and a multivariate nested logit model of the decision process underlying treatment seeking and choice of provider. (i) Univariate ComnparisonsA 18. Of the 22.786 individuals for whom relevant information is available from LSMS 98 data, 8.294 (36 percent) report an illness event in the recall period of the previous month (Figure 1). Illness events (hereafter called simply "illness") include, for example, cough, cold, or other respiratory infections; measles or other eruptive diseases; accidents; diarrhea, both infant and adult; and other diseases. 19. The incidence of illness varies by income and poverty classification. Among the overall poor, 37 percent report an incidence of illness, significantly greater than the 36 percent of the non-poor in Nicaragua that report an illness (p = 0.0056). Among extremely poor households, the incidence of illness is 36 percent, insignificantly different from the incidence if illness among both the overall poor (p = 0.1324) and the non-poor (p = 0.2456). The incidence of illness varies with income deciles, and the general trend is toward lower levels of illness at higher levels of income. The incidence of illness is the highest for the median income category (C$3.229 - C$3,966), with 39 percent reporting an illness. A gradual decrease in the incidence of illness is observed in the upper half of the income segment, and only 35 percent of the richest decile report an illness. 20. Of those reporting illness, 3,551 (43 percent) report seeking some form of medical advice for the reported illness. The medical providers include trained physicians, nurses, nurse assistants, pharmacies, midwives, untrained physicians (quack doctors), community health workers, and other providers. Among the overall poor, 37 percent seek treatment when ill, significantly lower than the 48 percent of the non-poor Nicaraguans who seek care when ill (p < 0.0001). Among extremely poor, 31 percent seek treatment, significantly lower than those in the other two poverty groups (p < 0.0001 in both cases). 21. Individuals with higher incomes are more likely to seek treatment compared to those with lower incomes, and this trend is consistent across all income deciles (figure 16). Among the poorest 10 percent, only 29 percent seek care, compared to 34 percent in the second decile, 42 percent in the median decile and 54 percent in the highest income decile. 22. There is a significant difference in the proportion of the population from rural and urban areas that seek treatment. Compared to 45 percent of the urban population, only 38 percent of the rural population seeks treatment (p < 0.0001). Similarly, more people in urban regions than rural regions seek treatment. The proportion of people seeking treatment is the lowest in the Atlantic Rural region (30 percent), followed by the Pacific Rural region (41 percent) and the Central Rural 8 Readers interested in more details of the results presented in this section are referred to Chawla, Mukesh, Health Care In Nicaragua: An Analysis of the Incidence of Illness, Health Care Utilization and Out-of- Pocket Spending, Report prepared for LSMS Poverty Study on Nicaragua, draft, 1999. Annex 17. Page 9 region (42 percent). Among urban areas, the proportion of people seeking treatment is highest in the Central Urban region (49 percent) followed by the Atlantic Urban region (48 percent) and Managua (44 percent). Within regions. the differences in treatment-seeking behavior is the highest between the Atlantic Urban and Rural regions followed by the Central Urban and Rural regions, and to a smaller extent, between the Pacific Urban and Rural regions. 23. Health-care-seeking behavior varies by gender as well. Among females, 44 percent seek care, significantly greater than 41 percent of males who do (p = 0.0009). However, there is no statistical difference in such behavior between male-headed and female headed households, witlh 43 percent of those in male-headship households seeking treatment when ill compared to 42 percent of those in female-headship households (p = 0.3254). 24. Education affects the decision to seek health care significantly: the more educated are more likelv to seek care. Care-seeking is highest among those with a university education (58 percent) and lowest among those with only preschool education (36 percent). On average. individuals with technical, high school, or university education are significantly more likely to seek care than individuals with lower levels of education. 25. An overwhelming majority of those ill and seeking care prefer trained physicians (88 percent), followed distantly by nurses (6 percent), nurse assistants (I percent), pharmacies (2 percent), midwives (0.02 percent), untrained physicians (0.8 percent), community health workers (I percent) and others (0.4 percent). Of those ill and seeking care, most prefer health centers (39 percent), followed by private clinics (24 percent), hospitals (12 percent) and first-aid stations, or puestos de salud (9 percent). Patients also visited INSS polyclinics (3 percent), pharmacies (3 percent) and private hospitals (2 percent) are the other institutional providers. 26. The choice of provider does not vary significantly with type of illness, and most patients prefer trained physicians to other health care providers. Among all types of illnesses, more than 85 percent of patient-provider contacts are with physicians, followed distantly by nurses. Other providers are used sparsely, with community health workers playing a somewhat important role among patients with adult diarrhea (3 percent), cough and cold (1.8 percent) and measles and other eruptive diseases (1.7 percent). Most patients with coughs and colds visit health centers (4:2 percent), followed by private clinics (22 percent). Patients with measles and other eruptive diseases follow a similar pattern; 42 percent visit health centers, and 26 percent visit private clinics. Most accident cases go to hospitals (53 percent), followed by health centers (15 percent) and private clinics (13 percent). Adult diarrhea patients prefer health centers (25 percent), pharmacies (1 7 percent), private clinics (16 percent), hospitals (15 percent) and first-aid stations (12 percent), in that order. 27. Trained physicians are the preferred first contact across all regions as well, though there are wide variations between urban and rural regions. Thus, while 95 percent of all those ill and seeking care visit physicians in the Atlantic Urban region, significantly fewer (56 percent, p < 0.0001) of those in the Atlantic Rural region do so. Similarly, while 94 percent in the Central Urban region and 93 percent in the Pacific Urban region go to physicians, significantly fewer. 83 percent in the Central Rural and 84 percent in the Pacific Rural regions (p < 0.0001 in both cases), go to physicians. 28. In regions where physician contacts were low, nurse and nurse assistant contacts are high. For instance, in Atlantic Rural region, 21 percent of all contacts are with nurses, followed by 8 percent with nurse assistants. Community health workers (8 percent) and untrained physicians (3 Annex 17, Page 10 percent) also substituted for trained physicians in Atlantic Rural region and Central Rural region (2 percent for community health workers and 1.5 percent for untrained physicians). 29. Among urban areas, 94 percent of those ill and seeking care visit trained physicians, significantly greater than the 82 percent of those in rural areas. As noted above, significantly more patients in rural areas (10 percent) seek care from nurses compared to patients in urban areas (2 percent, p< 0.0001). Similarly, patient visits with untrained physicians and community health workers are significantly greater in rural areas than in urban areas. 30. Across poverty levels, of those ill and seeking care, 93 percent of the non-poor visit trained physicians, significantly greater than 83 percent of the overall poor and 70 percent of the extremely poor (p < 0.0001 in both cases). Among the extremely poor, 17 percent of patients visit nurses, 6 percent visit community health workers and 2 percent visit untrained physicians, and the utilization of all three types of providers is significantly greater than in other poverty categories. A comparatively large number of overall poor patients visit nurses (10 percent), while some visit untrained physicians (1 percent). 31. The average9 expenditure on health care per individual seeking care is C$28. However, 28 percent of those seeking care do not incur any expenditure. Among all those who report some expenditure, the average expenditure is C$60. Across households, 66 percent report some expenditure on health care; their average expenditure is C$65. 32. Among all providers and all those seeking care, the maximum average expenditure is for visits to the pharmacies (C$40), followed by visits to trained physicians (C$32) and untrained physicians (C$24). Excluding those reporting no expenditure on health care, the average expenditure is C$70 for physician visits, C$40 for pharmacy visits and untrained physicians, and C$17 for visits to nurses and nurse assistants. 33. There are significant variations in expenditure also by institution. The maximum average expenditure on health care visits is in private clinics (C$160), followed by MINSA hospitals (C$40), pharmacies (C$40), and private hospitals (C$33). Excluding those who report no expenditure, the average expenditure was C$180 for private clinic visits, C$160 for private hospitals, C$100 for INSS polyclinics, C$80 for workplace medical centers, C$70 for MINSA hospitals, and C$40 for pharmacies. 34. Across all types of illness and across all individuals seeking care, the maximum average expenditure is for visits related to accidents (C$75), followed by visits related to other illnesses (C$52), and measles and other eruptive diseases (C$40). Excluding those who did not have any expenditure, the average expenditure is C$200 for accidents, followed by visits related to other illnesses (C$100) and measles and other eruptive diseases (C$66). Treatment of coughs and colds costs patients C$40, while adult diarrhea costs C$30. (ii) Multivariate Nested Logit Model Results 35. The provider choice model is estimated for the country as a whole, as well as separately for each region. Patients have a choice among three providers: physicians, nurse (including nurse assistants and midwives), and pharmacies (including community health workers). Various factors 9 In this section, all references to "average" imply the median. Annex 17. Page I I are deemed to affect provider choice: age, gender, marital status. employment status, place of residence (urban or rural), headship, type of illness (cold and cough, measles, diarrhea) and education. In one variant of the model, dummies for highest grade completed (elementary school, high school, or university) are included; in another variant. one dummy representing any education is included. Estimation results are presented in Table 1. 36. The value of C in all the models is between 0 and 1, indicating that the model is consistent with utility maximization. Individuals view the provider choices as closer substitutes for each other than no treatment or self care. The coefficients on income net of prices'° and the square of this term are significantly different from zero, and positive for the income variable and negative for the squared term, implying that the utility function is concave in income. Prices enter the model via the income terms, and the fact that the income terms are significant implies that the relevant prices are relevant to the choice of provider. However, since prices and income enter the model in a highly nonlinear fashion through the income terms, it is hard to judge the magnitude of their effects by merely examining the coefficients. We examine them in detail in the next section. in which we compute and present arc elasticities of prices and income. 37. Utilization of health care in Nicaragua increases with age. as is indicated by the positive and significant coefficient on age for physician and nurse use, and positive though not significantly differently from zero coefficient for pharmacy use. Gender does not seem to affect provider choice or the decision to seek care other than self care. Headship gender also does not seem to affect the decision to seek care or the choice between providers. 38. The positive and significant coefficients on the dummy representing whether the individual is employed indicates that the employed are more likely to seek care than not to seek care. This is an expected result, and probability reflects an income effect as well, since the employed are likely to have higher incomes than the unemployed. 39. Education seems to affect the decision to seek care as well as the choice of provider. Those with higher education are more likely to seek care than not to seek care and are more like lv to seek care from physicians and nurses than pharmacies. Individuals with university level education are significantly more likely to seek care from a trained physician than from other providers. 40. Family size does not significantly affect the decision to seek care. However, households with larger family size are less likely to seek care from trained physicians, preferring nurses and pharmacies instead. Married people are much less likely to seek care. a result for which there is no obvious explanation. 41. People with colds and coughs are more likely to seek care from all types of providers than not to seek care. Surprisingly, measles and diarrhea are not significant factors in the individual's decision; patients with these illnesses view self care as a close substitute for seeking care from a provider. 42. The results in the models for the different regions (Tables 2-5) are fairly similar to the overall country situation. 10 For purpose of the estimation, the variable is divided by 10,000 in all the results presented in this paper. Annex 17, Page 12 (iii) Price Elasticities 43. To assess the direction and magnitude of the effect or price and income on demand for healtlh care from a specific provider, we estimate arc price elasticities of the demand for physician. nurse, and pharmacy care by poverty groups. Following Train (1986), Gertler and van der Gaag (1990), and Chawla and Ellis (2000), the arc elasticities are obtained by sample enumeration. Within the specified price range, the probability of an individual choosing an alternative is predicted for every individual, holding all characteristics constant at their mean values, except price and income. The percentage change in the probability of choosing an alternative is divided by the percentage change in price to yield the arc price elasticity. In other words, an arc price elasticity of-1.0 implies that a 10 percent increase in price will result in a 10 percent reduction in demand; an arc price elasticity of-2.0 implies that a 10 percent increase in price will result in 20 percent reduction in demand, and so on. 44. Arc price elasticities of the demand for physician, nurse, and pharmacy care are calculated for C$20 intervals in the range of C$0 to C$200. The poverty groups are defined as extremelv poor, overall poor, and non-poor. 45. Table 6 describes the arc price elasticities calculated for the three alternatives. The price elasticities along a demand curve are read moving down a column holding income constant. Price elasticity across demand curves is read moving across a row, holding price constant. The results show that price elasticity of demand for physician care, nurse care, and pharmacy care fall with income. Price elasticities for all three alternatives are high for the extremely poor and the overall poor and low for the non-poor. The price elasticity of the demand for nurse and pharmacy care is higher than the price elasticity of the demand for physician care. Reading down a column, we see that the demand for health care is relatively inelastic in the lower price ranges and becomes more elastic in the higher price ranges. Comparing across the alternatives. the demand for nurse care and pharmacy care become elastic at relatively smaller price increases compared to the demand for physician care. 46. These results indicate that an increase in health expenditures will reduce the utilization of health care the extremely poor and the overall poor substantially more than for the non-poor. Therefore, increases in user fees, for example, will be regressive, affecting the poor adversely while having little impact on the non-poor. Similarly, increases in expenditures for care provided by nurses and pharmacies will reduce demand for these providers substantially more than for physicians. DISCUSSION 47. The main findings of the multivariate analysis are that demand for health care responds to prices and that demand for health care by the poor is more responsive than by the non-poor. Because the poor are much more responsive to price changes than the rich, a price increase is likely to reduce utilization by the poor more than by the country as a whole. " Many conclusions reported here are drawn from the univariate analysis, details of which can be obtained from Chawla, Mukesh, Health Care In Nicaragua: An Analysis of the Incidence of Illness, Health Care Use and Out-of-Pocket Spending, Report prepared for LSMS Poverty Study on Nicaragua, draft, 1999. Annex 17, Page 13 48. A little more than a third of Nicaraguans report illness in the month before the study, but less than half of them seek health care. Disparities between the poor and non-poor are vast, and even though there are no significant differences in the incidence of illness, the non-poor are about 30 percent more likely to seek care when ill than are the overall poor, and 50 percent more likely than are the extremely poor. The disparities are even more pronounced at extreme ends of the income distribution, with the richest 10 percent twice as likely to seek care as the poorest 1 0 percent. The likelihood of seeking care increases rapidly with small increments of income at lower income levels, but at middle-income levels, marginal increments in income increases this likelihood only slightly. 49. Rural-urban disparities in health care are also quite pronounced. The incidence of illness among the rural population is significantly higher in rural areas; yet significantly fewer seek care compared to the ill in urban areas. Assuming that all those ill who seek care recover from their illness, the rural population is 12 percent more likely to have ill-not-treated people at any point of time compared to the urban population. 50. Disparities in health care are significant across regions as well. The incidence of illness is highest in the Atlantic Rural region, yet the proportion of population in the Atlantic Rural region seeking treatment when ill is lowest. Similarly, the incidence of illness is high in the Central Rural and Pacific Rural regions, and both have significantly lower levels of treatment. In generai, Managua and the urban regions report lower levels of illnesses and higher levels of treatment. Disparities in health care are most significant across poverty groups across regions. The extremely poor in the Atlantic Rural region are less than half as likely to seek care when ill compared to the non-poor in Atlantic Urban region. 51. To some extent, the disparities in health care across regions can be explained bv the general economic and health status of people living in these regions. Managua has the highest average real household and per capita incomes, highest on food, health, and education expenditures, and the lowest incidence of illness is lowest in Managua. These indicators are worst in the Atlantic Rural and Central Rural regions. 52. Disparities in health care between males and females are not pronounced. While more females report an event of illness compared to males, more women also seek care when ill compared to men, so that women are no likely to be worse off compared to men as far the general health status is concerned. Extremely poor males appear to be the worst off as far as health care is concerned, and less than a third of them seek care when ill. In contrast, females in non-poor households are almost twice as likely to seek care. There appear to be no differences in illness and treatment according to household headship. 53. In general, people in Nicaragua prefer visiting trained physicians when ill and seeking treatment. However, differences are significant across regions and povertv groups. For instance.. those ill and seeking treatment in the Atlantic Rural region are only half as likely to visit trained physicians, preferring instead nurses, nurse assistants, community health workers, and even untrained physicians. Most of those ill prefer going to health centers and private clinics for most types of illness, the only exception being that accident cases are more likely to be treated in hospitals than elsewhere. Private clinics are used significantly more in urban areas than in rural areas, so much so that patients in the Central Urban region are over five times more likely to visit private clinics compared to those in the Atlantic Rural region. 54. Disparities in health care use by provider type and poverty groups are also vast, with the non-poor much more likely to visit trained physicians than the extremely poor. Thus, not only do Annex 17. Page 14 members of extremely poor households report higher levels of illness, they seek treatment less often. Of those that do, a large number go to nurses, community health workers and untrained physicians. 55. The most important reason why many of the ill do not seek care is that the illness is considered too slight. Budget issues are important to some, mostly to the extremely poor, who are twice as likely to not seek care compared to the non-poor. Distance is an important impediment to the extremely poor and the overall poor, but not to the non-poor. Distance prevents a very large number of the ill from getting treatment in the Atlantic Rural regions and to a smaller extent in the other rural regions. 56. Health insurance does not play a significant role in the health care system, with a little over 7 percent of households reporting any kind of insurance. As would be expected, insurance coverage is higher among the employed than the unemployed, and is predominantly higher among the rich than the poor. 57. More than one-fourth of all ill and seeking treatment do not incur any health care expenditure at all. Of the others, visits to the physician cost the most, followed by pharmacies. Visits to private clinics are the most expensive, followed by visits to hospitals. Health care following accidents cost the most, followed by visits for illnesses other than measles, coughs and colds, and diarrhea. 58. As a proportion of income, the poor report high out-of-pocket expenditure on health care. Out-of-pocket expenditure is regressive across income levels, and the non-poor spend a considerably smaller proportion of their income than the overall poor and the extremely poor. Annex 17, Paoe 15 Fig. A17.1: Incidence of Illness and Treatment Seeking TOTAL SAMPLE 22,786 (100%) ILL NOT ILL 8,294 14,492 (36.40%) (63.60%) NOT TREAT TREAT 3,551 O-* 4,753 (42.81%) (57.19%) Annex 17, Page 16 Table A17.1: NMNL Model of Provider Choice Estimates for Nicaragua Model I Model 2 Variables Coefficient t-ratios Coefficient t-ratios Income 1.08 14.61 1.08 14.78 Income Squared -0.13 -9.26 0.13 -9.36 Sigma 0.19 18.04 0.19 18.03 Physician Constant 2.18 7.63 3.19 7.70 Age 0.02 3.40 0.02 3.41 Urban -0.02 -0.09 -0.02 -0.10 Male 0.16 0.74 0.16 0.76 Employed 0.83 3.18 0.82 3.18 Male Household Head -0.24 -0.84 -0.24 -0.83 Married -1.15 -3.81 -1.18 -3.92 Separated/Divorced -0.79 -1.58 -0.82 -1.65 Cold And Cough 2.52 10.54 2.52 10.59 Measles -0.14 -0.25 -0.14 -0.25 Diarrhea -0.15 -0.42 -0.16 -0.46 Family Size -0.03 -0.59 -0.03 -0.59 Elementary School 1.03 2.70 High School 0.62 2.08 University 1.34 1.01 Education 0.76 2.99 Nurse Constant 0.75 1.54 0.81 1.67 Age 0.02 2.51 0.02 2.65 Urban -1.63 -5.89 -1.64 -5.96 Male -0.01 -0.01 0.02 0.06 Employed 0.72 2.27 0.63 2.01 Male Household Head 0.22 0.66 0.21 0.63 Married -1.15 -2.99 -1.30 -3.39 Separated/Divorced -1.13 -1.87 -1.31 -2.17 Cold And Cough 2.77 9.92 2.78 9.97 Measles -0.21 -0.29 -0.23 -0.32 Diarrhea 0.48 1.25 0.42 1.21 Family Size 0.04 0.76 0.04 0.76 Elementary School 1.54 3.53 High School 0.12 0.32 University -13.86 0.00 Education 0.67 2.21 Source: Nicaragua LSMS 1998 Annex 17. Page 17 Table A17.1(continued): NMNL Model of Provider Choice Estimates for Nicaragua Model I Model 2 Variables Coefficient t-ratios Coefficient t-ratios Pharmacy, Constant -0.58 -0.94 -0.59 -0.95 Age 0.01 1.34 0.01 1.34 U rban -0.67 -2.06 -0.66 -2.02 Male 0.75 2.48 0.75 2.55 Employed 1.19 3.14 1.19 3.21 Male Household Head -0.39 -1.02 -0.38 -0.99 Married -0.75 -1.61 -0.76 -1.64 Separated/Divorced -0.16 -0.24 -0.16 -0.24 Cold And Cough 3.37 9.75 3.37 9.80 Measles 0.85 1.12 0.84 1.11 Diarrhea 0.08 0.17 0.08 0.17 Family Size 0.00 0.01 0.00 0.00 Elementarv School 0.19 0.28 High School 0.07 0.18 University 1.48 0.96 Education 0.19 0.49 Source: Nicaragua LSMS 1998 Annex 17, Page 18 Table A17.2: NMNL Model of Provider Choice Estimates (different regions) Atlantic Urban Atlantic Rural Variables Coefficient t-ratios Coefficient t-ratios Income 2.01 5.84 0.00 0.00 Income Squared -0.33 -3.39 0.14 0.68 Sigma 0.16 3.95 0.33 5.93 Phinsician Constant 3.18 1.53 2.38 2.26 Age 0.04 0.74 0.01 1.21 Male -0.87 -0.78 0.25 0.64 Employed 1.57 1.07 -0.05 -0.09 Male Household Head 0.33 0.29 -0.75 -0.81 Married 0.09 0.04 -0.79 -1.32 Separated/Divorced -1.26 -0.43 -1.04 -1.01 Cold And Cough 4.22 2.70 1.00 2.56 Diarrhea 0.64 0.27 -0.60 -1.12 Family Size 0.02 0.09 0.08 1.04 Education 0.96 0.90 1.19 0.90 Nurse Constant -3.01 -0.71 1.27 1.19 Age 0.07 1.08 -0.02 -1.18 Male -1.29 -1.03 0.15 0.35 Employed 0.13 0.06 -0.15 -0.26 Male Household Head 0.24 0.18 -0.54 -0.57 Married -0.29 -0.08 0.65 0.09 Separated/Divorced - 1.76 -0.37 -0.72 -0.60 Cold And Cough 6.74 1.71 1.55 3.40 Diarrhea 3.94 0.91 -0.50 -0.82 Family Size 0.19 0.77 0.14 1.66 Education 1.41 1.01 1.35 0.99 Source: Nicaragua LSMS 1998 Annex 17, Page 19 Table A17.3: NMNL Model of Provider Choice Estimates (different regions) Central Urban Central Rural Variables Coefficient t-ratios Coefficient t-ratios Income 1.27 6.37 2.88 6.16 Income Squared -0.17 -3.66 -1.20 -4.41 Sigma 0.13 4.71 0.36 9.49 Physician Constant 3.86 2.64 3.05 6.19 Age 0.07 2.20 0.00 0.33 Male 0.10 0.14 0.05 0.19 Employed -1.60 -1.56 0.65 2.04 Male Household Head 0.04 0.05 0.54 1.79 Married -1.95 -1.56 -0.52 -1.58 Separated!Divorced -0.38 -0.18 0.08 0.16 Cold And Cough 3.49 3.35 0.27 1.00 Diarrhea 1.41 0.71 -0.56 -1.19 Family Size -0.15 -0.67 -0.07 -1.53 Education 1.18 1.47 0.51 1.25 Nuirse Constant -1.52 -0.78 0.43 0.66 Age 0.07 1.54 0.00 0.39 Male -0.07 -0.07 -0.04 -0.10 Employed -1.70 -1.32 0.73 1.57 Male Household Head 0.23 0.20 0.69 1.72 Married -0.72 -0.42 -0.62 -1.08 Separated/Divorced -0.04 -0.02 -0.83 -0.83 Cold And Cough 4.07 3.03 0.74 1.88 Diarrhea 1.57 0.68 0.52 0.92 Family Size 0.14 0.57 0.00 0.07 Education 1.22 1.17 0.21 0.37 Source: Nicaragua LSMS 1998 Annex 17, Page 20 Table A17.4: NMNL Model of Provider Choice Estimates (different regions) Pacific Urban Pacific Rural Variables Coefficient t-ratios Coefficient t-ratios Income 1.10 4.81 0.86 2.66 Income Squared -0.18 -2.55 -0.09 -0.73 Sigma 0.23 7.04 0.20 5.69 Phy sician Constant 3.72 5.27 3.13 3.55 Age 0.02 1.77 0.02 1.35 Male -0.48 -1.57 -0.41 -1.02 Employed 0.69 1.71 0.85 1.42 Male Household Head -0.04 -0.09 -0.18 -0.34 Married -1.50 -2.36 -1.07 -1.71 Separated/Divorced -1.78 -2.77 -0.24 -0.22 Cold And Cough 0.76 2.16 0.55 1.50 Diarrhea -0.62 -1.05 0.29 0.49 Family Size -0.03 -0.33 -0.01 -0.15 Education 0.43 1.31 1.06 2.09 Nurse Constant 2.05 0.89 -0.67 -0.62 Age 0.04 0.87 0.03 1.92 Male -1.38 -1.48 -0.17 -0.35 Employed 0.97 0.60 -0.14 -0.19 Male Household Head 0.06 0.06 1.27 1.69 Married -3.73 -2.05 -1.82 -2.32 Separated/Divorced 3.69 -1.43 -1.66 -1.25 Cold And Cough -0.28 -0.31 0.45 0.93 Diarrhea 0.22 0.22 -0.07, -0.09 Family Size -0.39 -0.84 0.08 2.90 Education -0.04 -0.03 1.73 Source: Nicaragua LSMS 1998 Annex 17, Page 21 Table A17.5: NMNL Model of Provider Choice Estimates (different regions) Managua Variables Coefficient t-ratios Income 0.83 4.34 Income Squared -0.09 -3.41 Sigma 0.19 4.64 Physician Constant 4.61 2.07 Age 0.02 0.68 Male 0.36 0.57 Employed -0.64 -0.63 Male Household Head -1.31 -1.40 Married -0.20 -0.15 Separated/Divorced -1.37 -093 Cold And Cough 1.68 2.58 Diarrhea -0.52 -0.54 Family Size -0.12 -0.56 Education 0.71 1.16 Nurse Constant 0.43 0.18 Age 0.01 0.25 Male 0.31 0.27 Employed -0.18 -0.09 Male Household Head -0.42 -0.32 Married 0.52 0.20 Separated/Divorced -0.50 -0.14 Cold And Cough 1.36 1.38 Diarrhea -0.76 -0.41 Family Size -0.02 -0.08 Education 0.02 0.01 Source: Nicaragua LSMS 1998 Annex 17, Page 22 Table A17.6: Arc Price Elasticities for Nicaragua Physician very poor poor non-poor 0-20 -0.04 -0.04 -0.02 20-40 -0.09 -0.08 -0.04 40-60 -0.14 -0.12 -0.06 60-80 -0.18 -0.16 -0.08 80-100 -0.23 -0.20 -0.10 100-120 -0.27 -0.25 -0.12 120-140 -0.32 -0.29 -0.14 140-160 -0.36 -0.33 -0.17 160-180 -0.41 -0.37 -0.19 180-200 -0.46 -0.41 -0.21 Nurse very poor poor non-poor 0-20 -0.17 -0.16 -0.11 20-40 -0.34 -0.32 -0.23 40-60 -0.50 -0.48 -0.34 60-80 -0.67 -0.64 -0.46 80-100 -0.84 -0.80 -0.57 100-120 -1.01 -0.97 -0.69 120-140 -1.18 -1.13 -0.80 140-160 -1.35 -1.29 -0.92 160-180 -1.52 -1.45 -1.04 180-200 -1.69 -1.62 -1.15 Pharmacy very poor poor non-poor 0-20 -0.19 -0.18 -0.14 20-40 -0.38 -0.37 -0.28 40-60 -0.57 -0.55 -0.43 60-80 -0.76 -0.73 -0.57 80-100 -0.95 -0.92 -0.71 100-120 -1.14 -1.10 -0.85 120-140 -1.33 -1.29 -1.00 140-160 -1.52 -1.47 -1.14 160-180 -1.71 -1.66 -1.28 180-200 -1.90 -1.84 - 1.43 mean income 1567 2544 8980 Source: Nicaragua LSMS 1998 Annex 17, Page 23 REFERENCES 1. Akin, J., Griffin, C., Guilkey, D., Popkin. B.1984. The Demand for Primary Health Care Services in the Bicol Region of the Philippines. Economic Development and Cultural Change 34(4): 755-82. 2. Alderman, H. and Paul Gertler. 1997. Family Resources And Gender Differences In Human Capital Investments: The Demand For Children's Medical Care In Pakistan, in Haddad, Hoddinot and Alderman (ed): Intrahousehold Resource Allocation in Developing Countries. Baltimore: John Hopkins University Press. 3. Bitran, R. 1993. HFS Project: Major Applied Research in Niger Under the Cost Recovery Pilot Tests: Research Goals, Objectives, and Methods, HFS Project paper. 4. Chawla, Mukesh. 1999, draft. Health Care In Nicaragua: An Analysis of the Incidence of Illness, Health Care Use and Out-of-pocket Spending, Report prepared for LSMS Povertv Study on Nicaragua. 5. Chawla, Mukesh and Randall P. Ellis. Forthcoming. Demand-Side Analysis of Health Care Financing and Delivery Innovations in Niger. Health Policy and Planning. 6. de Bethune, et al/ 1989. The Influence of an Abrupt Price Increase on Health service Use: Evidence from Zaire. Health Policy and Planning 4:76-81. 7. Diop, Francois P., Abdo Yazbeck, Ricardo Bitran. 1995. The Impact of Alternative Cost Recovery Schemes on Access and Equity in Niger. Health Policy and Planning 10(3):223- 241. 8. Diop, Francois Pathe. 1993-94. Long Term Technical Assistance Pilot Tests on Cost Recovery in the Non-Hospital Sector: Quarterly Report, Technical and Financial Activities, Reports for the period Oct-Dec 1992, Aug-Oct 1993, and Nov 1993-Jan 1994. Washington, D.C.: USAID. 9. Ellis, R.P., McInnes, D.K., Stephenson, E.H. 1994. Inpatient and Outpatient Health Care Demand in Cairo, Egypt. Health Economics 3: 183-200. 10. Ellis, R.P., Mwabu, G. 1993. The Demand for Outpatient Medical Care in Rural Kenya, Unpublished Working Paper, Economics Department, Boston University. 11. Gertler, P., L. Locay and W. Sanderson. 1987. Are User Fees Regressive? The Welfare Implications of HealthCare Financing Proposals from Peru, Journal of Econometrics 36:67- 88. 12. Gertler, Paul and Jacques van der Gaag. 1990. The Willingness to Payfor Medical Care. Evidencefrom Two Developing Countries. Baltimore: John Hopkins University Press. 13. Heller, P. 1982. A Model for the Demand for Medical and Health Services in Peninsular Malaysia. Social Science and Medicine 16: 267-84. Annex 17, Page 24 14. Huber, J., 1993. Ensuring Access to Health Care with the Introduction of User Fees: A Kenyan Example. Social Science and Medicine 36: 485-494. 15. Ii, M. 1996. The Demand for Medical Care: Evidence from Urban Areas in Bolivia, LSMS Working Paper No. 123. Washington, D.C.: The World Bank. 16. Lavy, Victor, Jean-Marc Germain. 1994. Quality and Cost in Health Care Choice in Developing Countries, LSMS Working Paper No. 105. Washington, D.C.: The World Bank. 17. Litvack, J., Bodart, C. 1993. User Fees Plus Quality Equals Improved Access to Health Care: Results of a Field Experiment in Cameroon. Social Science and Medicine 37(3): 369-83. 18. McFadden, Daniel Little. 1981. Econometric Models of Probabilistic Choice, in Charles Manski and Daniel Little McFadden (ed): Structural Analysis of Discrete Data with Econometric Applications. Cambridge, MA: The MIT Press. 19. Mwabu, G., J.K Wang'ombe, V.N.Kimani. 1991. Health Service Pricing Reforms and Health Care Demand in Kenya, Paper Presented at the 4th Annual Meeting of the IHPP in Lyon, Switzerland.. 20. Train, K. 1986. Qualitative Choice Analysis. Cambridge, Massachusetts: The MIT Press. 21. Waddington, C.J., Enyimayew. K.A. 1989. A Price to Pay: The Impact of User Charges in Ashanti-Akim District, Ghana. International Journal of Health Planning and Management 4: 17-47. 22. Wouters, Annemarie, Anthony Kouzis. 1994. Quality of Health Care and its Role in Cost Recovery with a Focus on Empirical Findings About Willingness to Pay for Quality Improvements, Major Applied Research Paper No. 8, HFS Project, Report submitted to USAID. Annex 18, Page I Annex 18 - Malnutrition among Preschool Children hv Mukesh Chawla EXECUTIVE SUMMARY i. This paper examines the prevalence of malnutrition or undernutrition among preschool children in Nicaragua. Using data generated by the Living Standards Measurement Surveys, we identify the chief determinants of nutritional status as well as the population most at risk of undernutrition. Several variables are explored to explain variations in HFA, WFA, and WFH z- scores. ii. Our analysis indicates that the levels, extent, and nature of undernutrition among preschool children have not changed significantly since 1993, when the first Nicaraguan LSMS took place. Patterns of undernutrition vary by age, vulnerability to diseases, household poverty, maternal literacy, and residence. iii. The results of our analysis have several interesting policy implications. First, undernutrition is significantly related to low income and poverty; steps taken to improve economic status will improve the overall nutritional status of preschool children in Nicaragua and will specifically improve stunting, underweight, and wasting. iv. Second, individual nutritional status is significantly linked to morbidity, with children suffering from illnesses the most vulnerable to undernutrition. Changes in health care and improvement in treatment of diarrhea reduce malnutrition. v. Third, households treating their drinking water have better nourished children than households that do not. Improvements in treated piped water supply reduce undernutrition. vi. Fourth, children of literate mothers have significantly better nutritional status than children of illiterate mothers. Improvements in maternal education, therefore, reduce undernutrition. vii. Finally, children living in rural areas in general, and the Central regions in particular, have poorer nutritional status than others; these areas and regions require extra attention. Annex 18, Page 2 INTRODUCTION 1. In 1998, the Government of Nicaragua conducted its second Living Standards Measurements Survey (LSMS 98), the Encuesta Nacional de Hogares Sobre Medicion de Niveles de Vida. Administered by the Instituto Nacional de Estadisticas y Censos (INEC), the survey collected information from a nationally representative sample of 4,656 households on a variety of household and individual level characteristics, and compiled data on demographic, education, economic, health care, employment, time-use, consumption, and agricultural variables. The present report uses data generated by LSMS 98 to document the prevalence of undernutrition and identify and evaluate its determinants among preschool children in Nicaragua. 2. This paper was prepared as background for the World Bank's Poverty Assessment for Nicaragua. Its objectives are to profile undernourished children age 5 and younger; to determine the relationship between undernutrition and selected child, maternal, and household characteristics; and to discuss implications of our findings for nutrition policy in Nicaragua. 3. We use anthropometric measurements of age, weight, and height to assess individual nutritional status. The most widely used anthropometric bodily indices for nutritional assessment are wasting, stunting, and underweight. Wasting measures deficiencies in terms of weight for height (WFH), while stunting measures deficiencies in terms of height for age (HFA). Underweight, or weight for age (WFA), combines wasting and stunting. 4. It is generally accepted that wasting indicates acute food loss from a state of emergency, environmental disaster, or other situation that limits the family food supply, making the child too thin for a given height. Stunting, on the other hand, indicates chronic malnutrition, or past nutrition, and can diminish intellectual capacity and impair work performance later in life. Stunting indicates poor overall economic conditions, chronic or repeated infections, and abnormal intake of nutrients. 5. To improve their interpretation, we report these anthropometric indicators in relation to international reference values defined by the U.S. National Center for Health Statistics (NCHS). The most common methods of analyzing differences between the sample population's nutritional level and that of the NCHS are as a position within the reference percentile distribution, as a percentage of the reference median, or as a standard deviation score (z-scores). We chose to use z-scores. 6. This report consists of six sections. Section 2 contains an account of prevalence of undernutrition in Nicaragua by age, gender, urban or rural location, region, and economic status of the household. Section 3 compares nutritional status between 1993 and 1998. Section 4 summarizes the prevalence of undernutrition in other Latin American countries. Section 5 analyzes the determinants of undernutrition. The report ends with a brief discussion of policy implications in section 6. All tables and figures appear in the appendix. Annex 18. Page 3 PREVALENCE OF UNDERNUTRITION AMONG PRESCHOOL CHILDREN IN NICARAGUA 7. Data. The data used in this analysis come from LSMS 1998. To calculate the z-scores, children's age was computed as the difference between the measurement date and the child's birth date. On the basis of the age so determined. z-scores were computed for 2,805 preschool children belonging to 1,865 households. Computed z-scores were evaluated for biological plausibility using criteria established by the World Health Organization (and stated in the ANTHRO software). As a result, 34 height-for-age z-scores and 20 weight-for-age z-scores greater than +6.00 or less than -6.00 were set to missing, and 49 weight-for-height z-scores greater than +6.00 or less than -4.00 were set to missing. The final sample, therefore, has HFA z- scores for 2,771 children, WFA z-scores for 2,785 children, and WFH z-scores for 2.756 children. 8. To classify households according to poverty groups, the income range for extremely poor was set at less than CS2.246, for overall poor at less than C$4,259, and for non-poor at equal to or greater than C$4,259.1 Using these cutoffs, 17 percent of the sample population was classified as extremely poor, 48 percent as overall poor, and the remaining 52 percent as non-poor. The sample was evenly divided between boys (50.5 percent) and girls (49.5 percent), and almost evenly divided between urban (48 percent and rural (52 percent) area. Regional distribution varied more, ranging from 9 percent in Managua to 23 percent in the Central Rural region. In ascending order of representation, the Atlantic Urban (10 percent), Central Urban (11 percent), Atlantic Rural (14 percent), Pacific Rural (17 percent), and Pacific Urban (17 percent) lay in between. Eight percent of the sample comprised children 0-6 months old; 10 percent children 6- 12 months old; 9 percent, children 12-18 months old; 10 percent, children 18-24 months old; 21 percent, children 24-36 months old; 20 percent, children 36-48 months old; and 23 percent, children 48-59 months old.2 9. Individual and household-level expansion factors (weights), as applicable, were used in computing all results. In reporting the differences in mean z-scores by variable category, p-values are computed to assess significance. P-values below 0.05 imply statistical significance. 10. Results. Our analysis indicates that almost 20 percent of all preschool children in Nicaragua are stunted, 11 percent are underweight, and 3 percent are wasted (Tables I and 2).3 Approximately 34 percent of all children (7 percent of the preschool population) who are classified as low HFA are severely stunted. Similarly, 20 percent of all children classified as low WFA are severely underweight, and 14 percent of all children classified as low WFH are wasted (all measured by z-scores <-3, not shown in tables) . Note that the 'extremely poor' category is a subset of the 'overall poor' category. 2 The age categories are inclusive of the lower levels but exclusive of the upper levels. Thus, the 0-6 months category implies '0 and more but less than 6 months.' Similar interpretations hold for all other categories, here and throughout the text and the tables. 3 In the remainder of the text and the tables also, 'stunting' implies HFA z-score < -2, 'underweight' implies WFA z-score < -2, and 'wasting' implies WFH z-score < -2. Annex 18, Page 4 Variation by age and gender 1I. The distribution of scores by percentile is summarized in Table 3. One-tenth of the preschool population has HFA scores less than or equal to -2.83, WFA scores less than or equal to -2.16, and WFH z-scores less than or equal to -1.32. Half of the population has z-scores for HFA less than -0.86, WFA less than 4.68, and WFH of less than -0.1. Ninety percent of the preschool population has z-scores under 1.2 for HFA. 1.1 for WFA, and 1.1 for WFH. The highest 10 percent, however, record z-scores for HFA between 1.2 and 5.91, WFA between 1.1 and 5.4, and WFH between 1.07 and 5.8. 12. The incidence of stunting tends to increase with age, with over 30 percent of all children 48 to 59 months old reporting z-scores for HFA of less than -2. The incidence of low weight also increases with age till the first two years, affecting over 15 percent of the preschool population, but then declines and levels off to 12 percent in the 48 to 59 month age group. Wasting is the highest in the age group of 6-12 months, falling off to less than 2 percent at 48 to 59 months. 13. The percentage of undernourished children is summarized by gender in Table 4. The gender differences are not statistically very significant across any indicator (though weakly significant for WFA z-scores < -2), and very similar proportions of girls and boys are classified as malnourished. 14. Within age groups, gender differences in the prevalence of stunting are evident at 12-36 months, in which significantly more boys than girls are undernourished (Table 5). The incidence of underweight is also significantly greater among boys across the 12-36 month group. The pattern for wasting is mixed: boys have significantly higher scores than girls at 6-18 months, but significantly lower scores at 18-24 and 48-59 months. Variation by region 15. Table 6 summarizes the prevalence of malnutrition by region. Overall, the level of malnourishment is lowest in Managua. The prevalence of stunting is highest in the Central Rura region, followed by the Atlantic Rural region. The Atlantic Rural region also has the highest proportion of underweight children, followed by the Central Rural region, and the highest proportion of wasting, followed by the Atlantic Urban region. Overall, malnutrition is significantly greater in rural than urban areas (Table 7). Almost one-fourth of all preschool children in rural areas are undernourished and fare worse than urban children across the other categories as well. 16. The regional differences in children's nutritional status can partly be explained by the general economic and health status of people living in these regions. Average real household and per capita incomes are highest in Managua (C$9,912 and C$4,725 respectively) compared to other regions, and are lowest in the Atlantic Rural (C$3,060 and C$1,269 respectively) and Central Rural (C$3,716 and 1,658 respectively) regions. Expenditures on food, health, and education are also highest in Managua and lowest in the Rural Atlantic and Central regions. The incidence of reported illness is lowest in Managua (29 percent) and highest in Atlantic Rural (44 percent) and Central Rural (40 percent) regions. Recall that the nutritional status of children in Annex 18, Page 5 Managua is significantly superior to that of children living in the Atlantic Rural and Central Rural regions. Variation by poverty classification 17. There is a strong correlation between the poverty status of the household and the nutritional status of the child (Table 8). A significantly larger proportion of overall poor children than non-poor children are stunted (p < .0001). Similarly, the a significantly larger proportion of overall poor children than in non-poor children are underweight (p < .0001). The incidence of underweight is highest among children in extremely poor households, significantly greater than that of children from non-poor households. Wasting is also significantly greater among children in poor households than among children in non-poor households (p < .001). 18. When poverty status is further disaggregated to show urban and rural classification, differences are not very significant for stunting (Table 9). Among extremely poor households, the percentages of stunted urban and rural children are not significantly different. The different proportions of underweight children in extremely poor urban and rural households are weakly significant (p = .058). Wasting is significantly more prevalent among children in extremely poor rural households than in extremely poor urban households (p < .001). Among the overall poor and non-poor, the differences in nutritional status of urban and rural children are not significant. 19. The percentage of children classified as malnourished by poverty group and region is summarized in Table 10. Among the extremely poor, the prevalence of stunting is lowest in Managua and highest in the Central Urban region, followed by the Central Rural region. Among the overall poor, these regions also fare the worst. Among the non-poor the prevalence of stunting is greatest in Atlantic Urban region, followed by the Central Rural region. 20. The percent of children classified as underweight among the extremely poor is highest in the Central Rural region, followed closely by the Central Urban region. Significantly, there is no incidence of underweight, nor of wasting, among the extremely poor in Managua. Among overall poor households, the Central Urban region has the highest proportion of underweight children, followed by the Atlantic Rural region. Non-poor children in the Central Rural region show no incidence of wasting, perhaps because this region is so heavily agricultural. However, more than 10 percent of non-poor children in the Pacific Rural region are underweight, and more than 12 percent are stunted. 21. The prevalence of stunting increases with age among both extremely poor and overall poor households (Table 11), reflecting the sum total of deviations from normal growth over the life span of the child. Stunting affects less than 2 percent of infants in extremely poor households but rises dramatically to 23 percent by 6 to 12 months and again to 42 percent by 18 to 24 months. Similar trends are seen in overall poor households. The prevalence of stunting also increases with age among the non-poor, but the trend reverses as the child grows older. 22. Across all poverty groups, the prevalence of underweight is also lowest among infants, but highest among children 18 to 24 months old. The incidence of underweight falls at 24 to 36 months, but rises again in the extremely poor and overall poor categories, significantly so in extremely poor households. Annex 18, Page 6 23. The prevalence of wasting is also highest among children 12-18 months from extremely poor households but falls precipitously after 24 months and declines significantly by 48-59 months. Among overall poor households, the prevalence of wasting shows an erratic trend, falling from a high of 11 percent among children 6-12 month old to 7 percent among children 12- 18 months old and then rising again to 11 percent in the 18-24 age group. Wasting is not widelI prevalent in any children 24 months or older. 24. Malnourishment by gender and poverty group is shown in Table 12. Girls are more undernourished than boys among extremely poor households but less so among the overall poor; however, these differences are not statistically significant. The prevalence of underweight is consistently higher among boys than among girls across all poverty groups. Among the extremely poor, girls are significantly more wasted than boys (p < .001). Among the overall poor and non- poor households, the incidence of wasting is greater among girls, but the differences are not statistically significant. Severe undernutrition 25. Severe stunting affects 7 percent of preschool children in Nicaragua. Severe underweight affects 2 percent of these children, and severe wasting affects 0.5 percent. The distribution of severe stunting, severe underweight, and severe wasting across age groups, regions, gender, and poverty classification is fairly similar to the trends described above. The nutritional status of preschool children in 1993 and 1998 26. Comparisons of LSMS 93 data4 and LSMS 98 data indicate a general improvement in the nutritional status of Nicaraguan school children. In 1998, 20 percent of preschool children were stunted, significantly lower than the 24 percent of 1993 (p <.001). Similarly, 11 percent were underweight in 1998, significantly lower than 12 percent in 1993, but weakly so. The prevalence of stunting among rural children improved but still remained high, falling from 33 percent in 1993 to 24 percent in 1998. 27. Nutritional status is strongly related to household poverty status in Nicaragua. In 1998, slightly more than 39 percent of extremely poor children suffered from some form of undernutrition, compared to 33 percent in overall poor households and 12 percent in non-poor households. These figures reflect an almost 10 percent improvement since 1993, when 43 percent of extremely poor, 36 percent of poor, and 16 percent of non-poor children suffered from some form of malnutrition. 28. LSMS 98 shows that the prevalence of stunting increases with age, a finding similar to that of the LSMS 93 study. Stunting affects about I percent of children less than six months old, but increases almost monotonically to 30 percent of all children 48 to 59 months old. The prevalence of underweight is also very low in children less than six months old and is highest among all children 18-24 months. Wasting is generally low among infants and in children older 4Results of the 1993 study are based on Ellen G. Piwoz: Underrutrition in Nicaraguan Preschool Aged Children: Prevalence, Determinants and Policy Implications, The World Bank, 1995. Annex 18, Page 7 than two years. Very similar findings were reported in LSMS 93. Within poverty groups, the effect of age on nutritional status is also very similar in both reference years. 29. There is little change in the proportion of school children suffering from severe stunting, with approximately 8 percent of preschool population reporting very low z-scores in 1993 and 7 percent in 1998. Severe underweight and severe wasting continue to be relatively less common; severe underweight affects 2 percent in 1998 and less than 2 percent in 1993. Severe wasting affects 0.5 percent of all children in 1998 and 0.2 percent in 1993. Prevalence of malnutrition in selected countries 30. Table 13 presents the prevalence of malnutrition among preschool in selected Latin American countries, including Nicaragua in both 1993 and 1998. The indicators for countries other than Nicaragua refer to 1990-97 and are based on the same definition and method of calculation as the indicators for Nicaragua. Stunting is the highest in Guatemala, followed by neighboring Honduras. Underweight is highest in Haiti and Guatemala; Nicaragua ranks seventh. Wasting is the highest in Guyana, followed by Haiti, Mexico, Bolivia, and Nicaragua 1998. Determinants of nutritional status 3 1. Various factors influence individual nutritional status; this section attempts to identify and explain them. We begin by presenting a conceptual model of malnutrition, followed by a discussion of the methodology used in analyzing the data and a note on how the variables were obtained. After presenting the results, we compare them with those of LSMS 1993. 32. The model. A review of the relevant literature (see, for example, Hlarris-White, 1997, Ryan et al, 1984. Behrman and Deolalikar, 1989, 1990, Behrman, 1988, Harriss, 1990) shows that many variables can affect a child's nutritional status. We examine 17 such variables in four broad categories: child-specific demographic details, socioeconomic characteristics of the household, education characteristics of the mother, and regional characteristics. Nutritional status of children 33. The nutritional status of children is measured by z-scores for HFA (stunting), WFA (underweight), and WFH (wasting). Similarly, stunting, underweight, and wasting are measured by HFA, WFA and WFH z-score values below -2. The relationship between these indicators and age, gender, economic status, and location was examined in the previous section. Child-specific characteristics 34. Age andgender: Nutritional status is likely to vary with age and gender for a variety of reasons. The prevalence of undemourishment, for instance, is likely to increase with age, Annex 18, Page 8 reflecting cumulative distortions from the norm. The incidence of underweight and wasting are also likely to vary with age if there are times that make the child especially vulnerable to infections and low dietary intake. Similarly, differences in nutritional status may depend on the child's gender, reflecting some combination of feeding practices, gender bias in family resource allocations, and physiological responses. 35. Vaccinations: Children vaccinated against tuberculosis, diphtheria, measles, and polio are likely to be less susceptible to illnesses and thus have better nutritional status compared to others. 36. Morbidity: The nutritional status of children is also likely to be affected by their morbidity histoTy, with children suffering from recent diarrhea or some other illness having lower z-scores compared to the children who are not sick. 37. Exclusive breastfeeding: Children breastfed exclusively beyond the age of 6 months are likely to have very low intake of non-breastmilk foods, and to this extent are likely to have lower z-scores. Household characteristics 38. Income: Children in richer households are expected to have higher levels of consumption and greater z-scores than children from poorer households. 39. Food expenditures: As with income, higher expenditures on food are also likely to produce higher z-scores. 40. Energy availability: Households with larger availability of calories per person are expected to have children with better z-scores. 41. Living density: Children living in crowded conditions are expected to have lower z-scores because of the higher probability of morbidity due to transmission of infectious diseases. Further. such households are likely to be less able to provide adequate nutrition for their children. 42. Dependency Ratio: Children in households with fewer income-earning members per child are expected to have poorer nutritional status because there are fewer adults to devote the required attention or expenditure to the child. 43. Household Headship: A review of the literature indicates that the internal distribution of resources in female-headed households is more child-oriented than in male-headed households. Such households consume foods of higher nutritional quality and spend a larger share of their income on food and child goods and a significantly smaller share on alcohol (see, for instance, Thomas, 1997, Horton and Miller, 1989 and Engle, 1991). Children in female-headed householis are thus expected to have better nutritional status. 44. Toilet at home: The presence of a toilet at home is a measure of household hygiene and available sanitation, and is hypothesized to improve nutritional status. Annex 18, Page 9 45. Drinking water: Water delivered via pipes inside the house is likely to contain fewer pathogens than water drawn from wells and rivers. To this extent, children living in houses with a piped water supply, or those who boil the water before drinking, are expected to have better nutritional status than other children. 46. Street Access: Families living in houses on paved streets enjoy better access to health care facilities, education, and nutritional programs of the government, and their children are likely to have higher z-scores than children in houses on unpaved streets. Regional characteristics 47. Rural versus urban residence: Children living in rural areas are likely to belong to poorer families and have lesser access to health care than children in urban households, and are thus more likely to be malnourished. 48. Administrative regions: Differences in regional characteristics, particularly availability of health care, health infrastructure, political environment, ethnicity, and tradition are likely to affect the child's nutritional status. Maternal characteristics 49. Maternal literacy: Women who can read and write are more likely to know the importance of good nutrition for their children. Accordingly, children of literate mothers are expected to have better nutritional status than children of illiterate mothers. 50. Maternal employment: The effect of maternal employment on a child's nutrition depends on whether the mother's working conditions adversely affect the care she can provide, for the child, as well as on the value of the earnings that she brings to the family. Depending on the relative weights of these factors, the net effect may be positive or negative. 51. Methodology. To identify and assess the impact of the determinants of individual nutritional status as measured by the three anthropometric indicators, ordinary least squares regressions corrected for heteroskedasticity are carried out by regressing the set of potentially influencing variables (the independent variables) on the HFA, WFA and WFH z-scores (the dependent variables). Note that the data may experience such problems as arise when the measured variables are too highly intercorrelated to allow precise analysis of their individual effects, or when there are measurement errors. For instance, the income variable and such household characteristics as crowding, use of processed drinking water, material covering the floor surface, availability of toilet, maternal employment status, etc., are likely to be highly intercorrelated. Variables such as calorie availability per capita, proportion of income spent on food, expected cost of health care and expected waiting time for health care are vulnerable to measurement errors. In order to address these problems due to multicollinearity and badly measured variables, several different variants of the model are estimated for each of the anthropometric indicators. These include estimating different variants of the model without the Annex 18, Page 10 possibly correlated variables, and using instrumental variables to account for possible correlation of the income variable with the error term. 52. In addition, maximum likelihood logistic regressions were conducted to evaluate the impact of various variables on the probability that a child is severely stunted , severely underweight, or severely wasted The coefficients from the logistic regression models are reported as odds-ratios, a frequently used means to express the relative effect of the independent variable on the dependent variable in a logistic regression model. If the odds-ratio for an independent variable is less than one, the odds of the dependent variable occurring decrease with higher values of the independent variable; if the odds-ratio is greater than one, the odds of the dependent variable occurring increase. 53. Data. The LSMS 98 data on 2,576 preschool children allow us to examine the influence of various factors on stunting (HFA z-scores). Similarly, data on 2,588 children are used to evaluate the impact on underweight (WFA z-scores) and data on 2,561 children used to evaluate the impact on wasting (WFH z-scores). 54. Most of the individual child characteristics are described in dummy variables that take on the value of I if an attribute is present and of 0 if that attribute is absent. Thus, gender is described as a dummy variable =I for males and =0 for females. Similarly, the dummy variable on exclusive breastfeeding equals I if the child is reported to be exclusively breastfeeding and 0 if otherwise. Dummies on tuberculosis, measles, diphtheria, and polio indicate whether the chilcl has been vaccinated against these diseases. Finally, the dummy on diarrhea denotes diarrhea in the last 30 days and is used as a measure of morbidity. 55. The age of the child is measured in months in models that use the continuous representation of age. In models that use age categories, category-specific dummy variables are used instead of absolute age. One category is dropped in the regression to avoid the dummy variable trap. 56. Among household characteristics, log of income per capita is used instead of absolute income values, to generate a normally distributed variable. Per capita expenditure on food, computed in terms of adult equivalence, is also measured in log values. Household density is computed by dividing the total number of individuals living in the house by the number of rooms in the house. Dependency ratio is expressed as the number of children below 14 years of age divided by the number of persons above 14 years of age in the household. Headship gender is represented as a dummy variable signifying the gender of the oldest living person in the household. Household characteristics such as presence of toilet at home, access to paved streets. and access to drinking water are described as dummy variables that take on the value of I if the attribute in question is present. 57. Total caloric intake is estimated by computing the total calorific value of all food that the household reports obtaining, either through purchase or gift, and then dividing this by the number of adults in the household. Total calorific value of each item is calculated according to Mexican food consumption tables. Adult equivalence is computed on the basis of caloric requirements by age and gender, so that in a household of two adults each requiring 2,900 calories per day and one child requiring 1,450 calories per day, the number of adult equivalents are 2.5. The age- and gender-specific caloric requirement is calculated on the basis of recommendations of the Food and Nutrition Board of the Institute of Medicine, National Academy of Sciences, Standing Annex 18. Page 1 1 Committee on the Evaluation of Dietary Reference Intakes (Table 14). The per capita caloric availability values are log-transformed to create a normally distributed variable. Note that this variable is likely to have a lot of noise, both because of the nature of the recall as well as because of the implicit assumption that caloric availability per capita is equivalent to caloric consumption. Accordingly, some variants of the regression models are estimated without including per capita caloric availability as a regressor. 58. Regional characteristics are described by dummy variables indicating urban or rural residence and region of residence. Expected health care costs and expected waiting times for health care are computed as average costs and times reported by those seeking care in each region. 59. Finally, maternal characteristics of literacy and employment are described by dummy variables that take on a value of I if the child's mother can read and write and I if the mother is employed. 60. Results. Overall, the models explain variations in HFA and WFA z-scores better than variations in the WFH z-scores, as is indicated by the low R-squared in all the models for the latter. Generally speaking, the results are very similar across all variants (Tables 15-20). BINARY OUTCOMES Child-specific characteristics 61. Among all the child-specific characteristics, the coefficients on age, exclusive breastfeeding, and diarrhea are negative and statistically significantly different from zero in all variants of the models for HFA and WFA z-scores, suggesting a strong negative association of these variables with the child's chronic nutritional status (HFA) and underweight (WFA). Results of the two-stage least squares model (estimated with the ratio of dependents to adults, maternal employment, and education as instruments for income, and using child-specific age categories instead of absolute age as regressors) indicate that older children have significantly lower HFA and WFA z-scores compared to infants, and that the differences increase with increasing age. 62. Exclusive breastfeeding and diarrhea in the previous 30 days significantly worsen stunting and underweight scores. Exclusive breastfeeding significantly lowers HFA and WFA scores when age is included as a continuous variable in the list of regressors, but is not a significant determinant when age categories are used as independent variables, though it continues to retain its negative coefficient value.5 Diarrhea significantly lowers HFA and WFA z-scores in all variants. 63. Variables indicating whether the child has been vaccinated against tuberculosis (BCG), diphtheria, pertussis, and tetanus (DPT), polio, and measles are also included in one of the variants of the model. The signs of the coefficients on these are surprisingly negative in the OLS 5 In other words, the coefficients on age groups reflect some impact of breastfeeding as well, indicating a positive association of breastfeeding with higher HFA scores in the younger age groups. Annex 18, Page 12 models, significantly so for DPT and measles for WFA z-scores. However, since most children report having had these vaccinations (96 percent for BCG, 94 percent for DPT, 94 percent for polio and 83 percent for measles), there is little difference between the malnourished and others for these indicators. Thus, these variables are not very useful determinants of nutrition.6 64. Few child-specific variables attain significance in the models for WFH z-scores. The OLS models indicate a negative and significant effect of diarrhea on WFH z-scores, and the two- stage least squares model shows that children 6 to 24 months old have lower WFH z-scores compared to infants, but the difference is not significant at higher ages. Household characteristics 65. Total household income per capita, measured by expenditure per capita, significantly improves nutritional status for all anthropometric indicators. Higher per capita income improve; stunting and underweight, and to a lesser extent it improves wasting. 66. Surprisingly, per capita expenditures on food significantly worsen children's stunting and to a lesser extent their weight for age. Similarly, per capita caloric availability does not significantly affect a child's nutritional status in any of the models. Both these results are counterintuitive. In all likelihood, the relationships indicated by the regression result point to errors in measurement and reporting of food expenditures, and to the inadequacy of using per capita caloric availability as an indicator of actual consumption.7 67. Higher dependency ratios worsen children's nutritional status for all anthropometric indicators across all models, as expected. 68. Of the other household characteristics affecting nutritional status, a male-headed household , a toilet at home, and processed drinking water significantly but weakly improve children's stunting. The relationship of these variables with other nutritiohal indicators is not significant.8 Regional characteristics 69. In assessing the effect of region on nutrition, the Pacific Rural region is taken to be the control variable. The results do not show any general trend across regions, with wide variations across variants of the model for different anthropometric indicators. Broadly speaking, children The negative values of the coefficients probably reflect the negative intercept in the regression, as can be seen by comparing the constant term in models that do not include vaccination indicators with those that do. 7 Even if individual caloric intake is measured with precision, it is not clear whether this is a useful indicator of relative welfare. Bouis and Pena (1997) argue that caloric intake, which they define as a necessity, is more likely to be equitably distributed within households. Thus, this would be a rather insensitive empirical measure compared with consumption of foods richer in micronutrients. 8 Note, however, that many household characteristics, such as density, use of solid floor materials, toilet facilities, etc., are likely to be highly collinear with household income. Different variants of the model, therefore, drop these variables in estimating the individual effects of these determinants. Annex 18, Page 13 living in the Central regions, both urban and rural, have higher incidence of stunting than the other regions. However, children from these regions, as well as those from Atlantic Rural, Managua, and Pacific Urban regions, have much lower incidence of wasting. 70. Of the other regional characteristics, the cost of obtaining health care and the waiting time associated with it significantly worsen current nutritional status, but weakly so. However, a higher expected cost of health care is correlated with long-term nutritional status. This curious result underscores the manner in which the variables of expected cost of health care and waiting time are created, using the actual costs and time of those who use health care. The obvious selection bias thus introduced may also make these variables highly collinear with income. To this extent, these variables are not used in some variants of the models. Maternal characteristics 71. Maternal literacy reduces the child's stunting, suggesting that education benefits the child's nutritional well-being. Maternal literacy diminishes incidence of underweight as well, although the relationship is significant in only one variant of the model. Maternal employment is not associated with any nutritional indicator. 72. Binary Outcomes. Adjusted odds-ratios for the determinants of stunting, underweight, and wasting are presented in Tables 2 1-23. Z-SCORES Child-specific characteristics 73. The probability that a child is stunted increases significantly with age. Similarly, older children are more likely to be underweight than infants, but to a lesser degree. Children 6-24 months old are more likely to be wasted than other age groups. 74. Of the other childhood characteristics, girls are less likely to be stunted, underweight, or wasted, but the differences are not significant. Similarly, children exclusively breast-fed and children reporting the incidence of diarrhea in the recent past are also insignificantly more likely to be stunted, underweight, or wasted. Household Characteristics 75. As noted previously, the economic status of the household significantly affects individual nutritional status. Children belonging to extremely poor households are almost four times as likely to be stunted, and two and half times as likely to be underweight as the non-poor. Similarly, children in poor9 households are twice as likely to be stunted and 60 percent more ' Note that for purposes of this regression, the category of 'poor' is equivalent to the category of 'overall poor' minus 'extremely poor;' that is, those with income between C$2246 and C$4259. Since the 'extremely poor' category is a subset of the 'overall poor' category, both cannot be used in the regression at the same time. Annex 18, Page 14 likely to be underweight as children of non-poor households. Across all models for stuntilg and underweight, the probability of severe undemourishment decreases with rising income. 76. Among other household characteristics, children in crowded households are significantly more likely to be stunted and underweight. Children belonging to households which treat their water are significantly less likely to be stunted. Finally, children in households with a male heacl are less likely to be stunted or underweight. 77. None of the household characteristics explain well the likelihood of wasting, though all of them have signs similar to the above. Regional Characteristics 78. As noted previously, the risk of being stunted, underweight, or wasted varies across regions. Broadly speaking, children in the Central regions are almost twice as likely to be stunted and underweight as children from other regions. Children in the Atlantic Rural region are likely to be two and a half times more wasted as children from other regions. Maternal Characteristics 79. Children of illiterate mothers are 35 percent more likely to be stunted than children of literate mothers. The relationship between matemal illiteracy and underweight or wasting is not significant. DETERMINANTS OF NUTRITIONAL STATUS IN 1998 AND 1993 80. Econometric analysis of LSMS 98 data indicates a strong positive association between income and individual nutritional status of children. Maternal literacy and processed drinking water are also found to improve nutritional status. Older children report significantly more stunting, as do children who are exclusively breastfed and those living in crowded conditions. Very similar determinants of nutritional status were reported in 1993, the main exception being that girls reported significantly higher z-scores than boys. 81. Higher income improves wasting, as does maternal literacy. Older children are more likely to underweight. Exclusive breastfeeding and rural residence also worsen underweight scores. These results are also very similar to the findings of LSMS 93. 82. The chief determinants of wasting in 1998 are income and incidence of diarrhea. In general, the model does not explain the WFH z-scores very well, as is also indicated by the low R-squared values. Similar results are reported in the analysis of the 1993 data. Policy Recommendations 83. Our analysis indicates that the level, extent, and nature of malnutrition among preschool children have not changed significantly since 1993, when the first Nicaraguan LSMS took place. Annex 18, Page 15 Patterns of undernutrition vary by age, vulnerability to diseases, household poverty, maternal literacy, and residence. 84. The results of our analysis have several clear policy implications. First, undernutrition is significantly related to low income and poverty; steps taken to improve economic status will improve the overall nutritional status of preschool children in Nicaragua and will specifically improve stunting, underweight, and wasting. 85. Second, individual nutritional status is significantly linked to morbidity, with children suffering from illnesses the most vulnerable to malnutrition. Changes in health care and improvement in treatment of diarrhea reduce malnutrition. 86. Third, households treating their drinking water have better nourished children than households that do not. Improvements in treated piped water supply reduce malnutrition. 87. Fourth, children of literate mothers have significantly better nutritional status than children of illiterate mothers. Improvements in maternal education, therefore, reduce malnutrition. 88. Finally, children living in rural areas in general, and the Central regions in particular, have poorer nutritional status than others; these areas and regions require extra attention. Annex 18, Page 16 Table A18.1 - Average z-scores for children (0-59 months), by age group Height For Age'° Weight For Age Weight For Height Age Group 1 Mean SD N Mean SD N Mean SD N 0-6 1.10 1.48 215 1.27 1.44 217 0.14 1.01 216 6-12 -0.18 1.53 266 -0.28 1.44 266 -0.19 1.20 262 12-18 -0.39 1.86 282 -0.72 1.39 282 -0.56 1.03 277 18-24 -0.85 1.54 260 -0.74 1.33 264 -0.26 1.19 257 24-36 -0.84 1.39 562 -0.59 1.31 562 -0.04 1.07 560 36-48 -1.05 1.45 563 -0.67 1.11 567 0.07 0.90 562 48-59 -1.35 1.41 623 f -0.81 1.18 627 -0.02 1.10 622 All ages -0.73 1.63 2771 -0.50 1.39 2785 -0.08 1.06 2756 Source: Nicaragua LSMS 1998 Table A18.2 - Percent of children (0-59 months) classified as undernourished, by age group Level of malnourishment Age Group Chronic Underweight Acute 0-6 0.77 I 0.66 0.93 6-12 10.37 8.6 9.39 12-18 15.56 14.68 6.94 18-24 20.77 15.14 7.04 24-36 17.49 11.58 2.38 36-48 1 25.49 10.82 0.79 48-59 30.27 11.79 1.89 All ages 19.94 10.95 3.32 Source: Nicaragua LSMS 1998 '° Height for age measures stunting, or chronic malnutrition. Weight for age measures wasting, or acute malnutrition. Weight for height measures underweight. Annex 18. Page 17 Table A18.3 - Average z-scores for children (0-59 months): Percentile distribution. Percentiles Height For Age Weight For Age Weight For Height 0-10 -2.83 -2.16 -1.32 11-20 -2.11 -1.66 -0.92 21-30 -1.61 -1.33 -0.65 31-40 -1.23 -1.01 -0.35 41-50 -0.86 -0.68 -0.10 51-60 -0.48 -0.35 0.13 61-70 -0.11 0 0.38 71-80 0.38 0.44 0.66 81-90 1.19 1.12 1.07 91-100 5.91 5.94 5.80 Source: Nicaragua LSMS 1998 Table A18.4 - Percent of children (0-59 months) classified as malnourished, by gender Level of mainourishment Gender Chronic Underweight Acute Male I120.22 11.75 3.31 Female i 19.66 10.14 3.52 Source: Nicaragua LSMS 1998 Table A18.5 - Percent of children (0-59 months) classified as malnourished, by age and gender Level of malnourishment Age in months Chronic Underweight Acute Male Female Male Female Male Female 0-6 1.40 0 0.67 0.64 0.89 0.98 6-12 10.55 10.15 9.62 7.27 5.64 14.39 12-18 17.56 13.32 16.00 13.17 4.77 9.45 18-24 22.64 18.90 18.78 11.36 9.06 5.02 24-36 18.44 16.69 13.64 9.85 2.53 2.25 36-48 24.93 26.02 9.64 11.92 0.93 0.66 48-59 3038 30.16 12.27 11.33 2.62 1.14 Source: Nicaragua LSMS 1998 Annex 18, Page 18 Table A18.6 - Percent of children (0-59 months) classified as malnourished, by region Level of malnourishment Region Chronic Underweight Acute Atlantic Urban 16.96 10.16 4.65 Atlantic Rural 24.56 15.57 5.17 Central Urban 22.12 11.98 1.65 Central Rural 29.09 13.29 4.38 Managua 10.01 6.91 3.64 Pacific Urban 18.07 9.97 1.81 Pacific Rural ! 19.41 11.39 3.02 Source: Nicaragua LSMS 1998 Table A18.7 - Percent of children (0-59 months) classified as malnourished, by location Level of malnourishment Geographic Location Chronic Underweight Acute Urban 16.14 9.56 2.69 Rural 23.61 12.29 4.12 Source: Nicaragua LSMS 1998 Table A18.8 - Percent of children (0-59 months) classified as malnourished, by poverty classification Level of malnourishment Poverty Classification Chronic Underweight Acute Extremely poor 35.77 19.17 4.19 Overall poor 27.09 14.13 4.07 Non-poor 9.48 6.29 | 2.45 Source: Nicaragua LSMS 1998 Annex 18. Page 19 Table A18.9 - Percent of children (0-59 months) classified as malnourished, by poverty groups and geographic location Poverty Classification Extremely poor Overall poor Non-poor Geographic Location L Urban Rural Urban Rural Urban Rural Chronic 36.46 35.56 26.18 27.61 8.56 11.51 Underweight 17.28 19.72 13.52 14.46 6.58 5.64 Acute 2.15 4.79 3.54 4.37 2.04 3.34 Source: Nicaragua LSMS 1998 Table A18.10 - Percent of children (0-59 months) classified as malnourished, by poverty groups and region Level of mal- Chronic Underweight Acute nourishment Poverty Extr. Overall Non- Extr. Overall Non- Extr. Overall Non- Classification poor poor poor poor poor poor poor poor poor UAtrbantic 17.33 18.85 15.17 16.62 11.23 9.13 5.54 3.16 6.13 Atlantic Rural 29.15 27.09 14.11 19.79 17.89 5.99 7.36 4.91 6.24 CentralUrban 53.25 39.75 5.45 21.91 19.93 4.46 2.01 2.16 1.19 Central Rural 1 45.48 33.19 13.78 23.69 16.15 2.61 4.63 5.56 0 Managua 14.99 13.93 8.14 0 7.81 6.48 0 6.04 2.48 Pacific Urban 38.37 29.11 7.09 18.34 13.75 6.20 1.03 1.74 1.89 Pacific Rural 25.53 21.63 12.36 15.44 11.72 10.37 4.05 2.79 3.75 Source: Nicaragua LSMS 1998 Note: Extr. poor: Extremely poor Annex 18, Page 20 Table A18.11 - Percent of children (0-59 months) classified as malnourished, by poverty groups and age groups Age Chronic Underweight Acute Groups i Extr. Overall Non- Extr. Overall Non- Extr. Overall Non- | poor poor poor poor poor poor poor poor poor 0-6 1.65 1.31 0 0 0.62 0.72 1.22 1.59 0 6-12 22.74 12. 36 7.23 15.99 9.46 7.28 5.24 10.86 7.06 12-18 24.13 18.92 10.08 22.79 18.81 8.88 11.23 7.19 6.57 18-24 42.26 27.91 11.65 32.09 18.58 10.06 9.00 10.79 2.03 24-36 30.86 23.50 8.96 1 14.29 14.37 7.62 3.45 2.44 2.29 36-48 48.03 37.44 9.06 24.92 15.51 4.33 2.73 1.37 0 48-59 48.54 40.29 13.59 20.05 15.78 5.20 2.17 1.66 2.26 Source: Nicaragua LSMS 1998 Note: Extr. poor: Extremel/ poor Table A18.12 - Percent of children (0-59 months) classified as malnourished, by gender and povertv classification Level of malnourishment Poverty Grouping Chronic Underweight Acute Male Female Male Female Male Female Extremely poor 34.49 36.99 , 19 95 18.41 3.27 5.08 Overall poor 26.88 27.33 14.80 13.43 3.94 4.22 Non-poor 9.99 8.99 f 7.08 5.53 2.34 2.55 Source: Nicaragua LSMS 1998 Annex 18, Page 21 Table A18.13 - Recommended dietary allowance (Energy kcal) Age in years Male Female 0-1 750 750 1-3 1300 1300 4-6 1800 1800 7-10 2000 2000 11-14 2500 2200 15-18 3000 2200 19-24 2900 2200 25-50 2900 2200 51+ 2300 1900 Source: Recommended Dietary Allowance, National Academy Of Sciences. National Academy Press, Washington D.C. 1997. Table A18.14 - Percent of children under 5 classified as malnourished: selected countries, 1990-1997 Country Chronic l Undernourished i Acute Bolivia 28 | 16 4 Brazil 11 6 2 Chile 2 1 0 Colombia 15 8 1 Ecuador 34 17 -2 El Salvador 23 11 1 Guatemala 50 27 3 Guyana 10 12 12 Haiti 32 28 8 Honduras 40 18 2 Jamaica 6 10 4 Mexico 22 14 6 Nicaragua (1993) 24 12 2 Nicaragua (1998) 22 12 3 Panama 9- 7 1 Paraguay 17 4 0 Peru ! 26 8 1 Uruguay 8 5 1 Venezuela 13 5 3 Source: statistical tables- UNICEF (as obtained from Demographic and Health Surveys, multiple Indicator Cluster surveys. WHO and UNICEF). Annex 18, Page 22 Table A18.15 - OLS Regression (Dependent Variable: z-scores for HFA) Model I Model 2 Model 3 Variable Coeff SE Coeff SE Coeff SE Individual Clhild Clharacteristics Male -.017 .072 -.032 .072 -.045 .074 Age -.044* .003 -.039* .004 -.045* .003 Exclusive breastfeeding -.417* .104 -.375* .108 -.459* .107 TB - - .025 .242 - DPT - - -.516 .385 - - Polio -.049 .337 - - Measles - - -.095 .142 - - Diarrhea -.233* .109 -.209* .108 -.228* .09 Household Characteristics Income 1.03* .136 .853* .142 .291* .(,48 Food expenditure per capita -.736* .132 -.561* .129 Calorie Availability .055 .053 - Headship .103 .081 .113 .084 .148 .J86 Density - - -.021 .014 -.030 () 14 Dependents -.075 .042 -.079 .044 -.082 .('44 Processed Drinking Water - - .105 .101 .101 .102 Paved Street - - -.115 .119 -.068 .16 Solid Floor Material - - .030 .089 .050 .(86 Toilet at home - - .181 .131 .227 .134 Maternal Characteristics Literacy .303* .094 .345 .099, .321* .100 Employment - - -.006 .092 .043 .193 Regional Characteristics Atlantic Urban -.076 .120 -.043 .145 -.139 .128 Atlantic Rural .093 .121 .002 .143 -.052 .123 Central Urban -.236* .117 -.351 * .120 -.266* .123 Central Rural -.158 .128 -.226 .143 -.319* .125 Managua -.013 .126 -.297* .143 .037 .136 Pacific Urban -.196* .100 -.172 .107 -.136 . 09 Expected cost of health care - - .502* .150 Expected wait time - - -.251 * .130 Constant -2.147 .636 -1.869 1.014 -.947 .392 R Squared .232 .254 .231 N 2693 2581 2581 Source: Nicaragua LSMS 1998 Annex 18, Page 23 Table A18.16 - OLS Regression (Dependent Variable: z-scores for WFA) Model I Model 2 Model 3 Variable Coeff SE Coeff SE Coeff SE Individual Child Characteristics Male -.042 .060 -.047 .059 -.056 .062 Age -.031 * .003 -.020* .003 -.031 * .003 Exclusive breastfeeding -.442 * .099 -.306* .102 -.450* .102 TB - - .218 .206 - - DPT - - -.998* .297 - Polio - - .134 .268 - - Measles - - -.303* .124 - Diarrhea -.286* .079 -.233* .078 -.279* .081 Household Characteristics Income .744* .117 .616* .120 .227* .041 Food expenditure per capita -.502* .119 -.377* .113 - Calorie Availability -.005 .047 - - - Headship .089 .076 .100 .074 .131 .077 Density - - -.006 .012 -.011 .012 Dependents -.111* .036 -.127* .037 -.124* .038 Processed Drinking Water - - .080 .081 .070 .084 Paved Street - - -.119 .093 -.103 .094 Solid Floor Material - - .052 .071 .053 .070 Toilet at home - - .195 .120 .242 .123 Maternal Characteristics Literacy .160* .069 .203* .071 .156* .073 Employment - - .045 .069 .065 .070 Regional Characteristics Atlantic Urban -.012 .109 -.076 .124 -.054 .115 Atlantic Rural .199 .108 .029 .127 .138 .113 Central Urban -.011 .102 -.155 .104 -.041 .107 Central Rural .009 .095 -.068 .104 -.083 .094 Managua .111 .114 -.046 .125 .148 .121 Pacific Urban -.004 .091 -.037 .096 -.003 .099 Expected cost of health care - - .167 .114 - Expected wait time - - -.208 .110 - Constant -1.062 .531 -.402 .782 -.747 .352 R Squared .158 .194 .159 N 2706 2593 2593 Source: Nicaragua LSMS 1998 Annex 18, Page 24 Table 18.17 - OLS Regression (Dependent Variable: z-scores for WFH) I Model I Model 2 Model 3 Variable I Coeff SE Coeff SE Coeff SE Individual Child Characteristics Male -.0]] .050 -.007 .051 -.014 .0:1 Age .003 .002 .006* .003 .003 .002 Exclusive breastfeeding -.123 .086 -.064 .087 -.115 .086 TB - - .051 .145 - - DPT - - -.519* .255 - - Polio - - .391 .234 - - Measles - - -.157 .101 - - Diarrhea -.117 .062 -.099 .063 -.106 .062 Houisehold Characteristics Income .234* .098 .244* .099 .104* .038 Food expenditure per capita -.115 .103 -.133 .096 - Calorie Availability -.046 .040 - - - Headship .076 .063 .078 .064 .086 .065 Density - - .004 .010 .001 .010 Dependents -.074* .036 -.084* .038 -.086* .0218 Drinking Water - - .001 .073 -.008 .0,'3 Paved Street - - -.022 .078 -.024 .0'9 Solid Floor Materials - - .051 .059 .045 .04 9 Toilet at home - - -.050 .103 -.029 .1 3 Maternal Characteristics Literacy -.002 .063 .016 .065 -.007 .067 Employment - - .001 .057 .010 .04 Regional Characteristics Atlantic Urban .038 .089 -.080 .100 .020 .0"4 Atlantic Rural .122 .087 .013 .102 .136 .091 Central Urban .178* .085 .124 .090 .177 .089 Central Rural .172* .075 .127 .085 .168 .078 Managua .131 .094 .173 .111 .149 .098 Pacific Urban .153* .070 .114 .073 .123 .075 Expected cost of health care - - -.160 .098 - - Expected wait time - - -.141 .085 - Constant -.441 .430 -.443 .647 -.817 .31 R Squared .038 .045 .036 N 2678 2566 2566 Source: Nicaragua LSMS 1998 Annex 18, Page 25 Table A18.18 - Two-Stage Least Squares Regressions (Dependent variable: z-scores for HFA) Variable Coefficient Standard Error Individual child's characteristics Male -.033 .075 Age 6-12 months - 1.271 * .174 Age 12-18months -1.451* .242 Age 18-24 months -1.966* .178 Age 24-36 months - 1.952* .146 Age 36-48 months -2.290* .198 Age 48-59 months -2.518* .194 Exclusive breastfeeding -.076 .138 Diarrhea -.116 .108 Household Characteristics Income .545* .063 Headship .142 .079 Drinking Water .107 .101 Toilet at home .126 .132 Maternal Characteristics Literacy .266* .105 Regional Characteristics Atlantic Urban -.148 .122 Atlantic Rural -.049 .113 Central Urban -.280* .119 Central Rural -.259 .131 Managua -.035 .131 Pacific Urban -.139 .105 Constant -3.497 .597 R Squared .242 N 2581 Source: Nicaragua LSMS 1998 Annex 18, Page 26 Table A18.19 - Two-Stage Least Squares Regressions (Dependent variable: z-scores for WFA) Variable Coefficient Standard Error Individual child's characteristics Male -.048 059 Age 6-12 months -1.528* .162 Age 12-18months -1.991* .171 Age 18-24 months -1.986* .149 Age 24-36 months -1.871 * .130 Age 36-48 months -2.123* .175 Age 48-59 months -2.208* .174 Exclusive breastfeeding -.150 .133 Diarrhea -.121 .078 Household Characteristics Income .422* .1051 Headship .134 .172 Drinking Water .068 .179 Toilet at home .130 .1 14 Maternal Characteristics Literacy .108 .A074 Regional Characteristics Atlantic Urban -.074 .07 Atlantic Rural .079 . 03 Central Urban -.079 .099 Central Rural -.057 .096 Managua .067 . 13 Pacific Urban -.007 .092 Constant -2.228 .469 R Squared .219 N 2593 Source: Nicaragua LSMS 1998 Annex 18, Page 27 Table A18.20 - Two-Stage Least Squares Regressions (Dependent variable: z-scores for WHF) Variable Coefficient Standard Error Individual child's characteristics Male -.005 .052 Age 6-12 months -.318* .140 Age 12-18 months -.657* .112 Age 18-24 months -.386* .133 Age 24-36 months -.188 .107 Age 36-48 months -.154 .142 Age 48-59 months -.173 .143 Exclusive breastfeeding -.074 .107 Diarrhea -.052 .066 Household Characteristics Income .184* .045 Headship .108 .068 Drinking Water -.004 .071 Toilet at home -.052 .104 Maternal Characteristics Literacy -.028 .063 Regional Characteristics Atlantic Urban .001 .089 Atlantic Rural .101 .086 Central Urban .170 .087 Central Rural .169* .076 Managua .125 .098 Pacific Urban .137 .073 Constant -1.460 .332 R Squared .049 N 2566 Source: Nicaragua LSMS 1998 Annex 18, Page 28 Table A18.21 - Logistic Regression (Dependent Variable: Dummy indicating HFA z-score<-2) Model I Model 2 Model 3 Odds Odds Odds Variable Ratio t Ratio t Ratio t Individual Child Characteristics Male 1.137 1.027 1.131 .980 1.136 1.075 Age 6-12 months 17.019* 3.723 16.897* 3.722 13.901* 3.509 Age 12-18 months 24.967* 4.340 24.494* 4.318 21.975* 4.189 Age 18-24 months 42.858* 5.089 40.657* 5.007 41.558* 5.)92 Age24-36months 29.179* 4.623 28.779* 4.062 26.911* 4.552 Age 36-48 months 73.071* 5.396 72.824* 5.376 81.566* 5.510 Age 48-59 months 80.169* 5.544 78.478* 5.496 93.306* 5.716 Exclusive breastfeeding 1.476 1.118 1.476 1.100 1.740 1.514 Diarrhea 1.169 1.073 1.175 1.089 1.091 .522 Household Characteristics Income .578* 5.128 .579* 4.698 Extremely Poor 3.341* 6.304 Poor 1.989* 4.035 Calorie Availability .902 1.259 .901 1.248 Headship .746 1.947 .737 1.989 .770 1.779 Density 1.034 1.427 1.027 1.174 1.078* 3.526 Drinking Water .694 2.060 Paved Street .975 1.460 Solid Floor Materials .804 .116 Toilet at home .803 .643 Maternal Characteristics Literacy .622 3.244 .675 2.655 .563* 4.200 Employment .971 .219 Regional Characteristics Atlantic Urban 1.165 .581 1.300 .946 1.302 1:)48 Atlantic Rural 1.139 .544 1.196 .719 1.414 1.570 Central Urban 2.044* 3.055 2.216* 3.319 2.030* 3.240 Central Rural 1.799* 2.890 1.803* 2.904 2.022* 3.716 Managua .927 .251 .995 .016 .796 .798 Pacific Urban 1.373 1.456 1.435 1.596 1.324 1.379 Pseudo R Squared .138 .143 .153 N 2578 2576 2583 Source: Nicaragua LSMS 1998 Annex 18, Page 29 Table A18.22 - Logistic Regression (Dependent Variable: Dummy indicating WFA z-score<-2) Model I Model 2 Model 3 Odds Odds Odds Variable Ratio t Ratio t Ratio T Individual Child Characteristics Male 1,309 1.775 1.317 1.788 1.295 1.806 Age 6-12 months 13.525* 2.354 13.252* 3.318 14.430* 3.489 Age 12-18 months 24.998* 4.232 23.927* 4.166 23.035* 4.168 Age 18-24 months 28.800* 4.391 28.063* 4.331 27.906* 4.424 Age 24-36 months 17.921* 3.828 17.379* 3.774 17.493* 3.866 Age 36-48 months 18.422* 3.520 18.109* 3.479 15.830* 3.367 Age 48-59 months 19.197* 3.608 18.599* 3.549 16.792* 3.494 Exclusive breastfeeding .998 .006 1.013 .035 .874 .360 Diarrhea 1.132 .710 1.122 .650 1.125 .689 Household Characteristics Income .679* 2.978 .675* 2.995 Extremely Poor 2.717* 4.617 Poor 1.573* 2.243 Calorie Availability .907 .941 .914 .872 Headship .745 1.646 .729 1.732 .757 1.628 Density 1.046* 1.660 1.06* 1.700 1.076* 3.068 Drinking Water .732 1.473 Paved Street .969 .125 Solid Floor Materials 1.061 ..347 Toilet at home .825 .471 Maternal Characteristics Literacy .769* 1.471 .796* 1.247 .703* 2.146 Employment .885 .715 Regional Characteristics Atlantic Urban 1.099 .302 1.035 .106 1.179 .560 Atlantic Rural 1.444 1.310 1.328 .959 1.616 1.865 Central Urban 1.814* 2.109 1.823* 2.084 1.731* 2.079 Central Rural 1.379 1.362 1.346 1.241 1.359 1.418 Managua .874 .388 .872 .397 .926 .232 Pacific Urban 1.167 .578 1.147 .497 1.145 .549 Pseudo R Squared .080 .082 .083 N 2590 2588 2595 Source: Nicaragua LSMS 1998 Annex 18, Page 30 Table A18.23 - Logistic Regression (Dependent Variable: Dummy indicating WHF z-score<-2) Model I Model 2 Model 3 Odds Odds Odds Variable Ratio T Ratio t Ratio T Individual Child Characteristics Male 1.207 .614 1.226 .670 ..994 .0^ 0 Age 6-12 months 5.632* 2.173 5.761* 2.197 11.335* 3.2;>9 Age 12-18 months 3.812* 1.664 3.909* 1.680 6.722* 2.5',2 Age 18-24 months 5.149* 1.902 5.199* 1.903 8.110* 2.556 Age 24-36 months 1.435 .453 1.456 .472 2.302 1.154 Age 36-48 months 1.207 .192 1.217 .201 1.207 .2(08 Age 48-59 months 2.244 .878 2.255 .883 2.486 1.056 Exclusive breastfeeding 1.539 .897 1.514 .860 1.274 .553 Diarrhea 1.155 .448 1.154 .458 1.278 .778 Household Characteristics Income 1.162* .756 1.205* .911 Extremely Poor 2.125 1.7,5 Poor 2.058 1.612 Calorie Availability 1.023 .12 3 1.010 .056 Headship 1.036 .085 1.139 .331 .935 .155 Density 1.096 1.956 1.088 1.693 1.063 1.574 Drinking Water 1.141 .297 Paved Street .747 .597 Solid Floor Materials .839 .513 Toilet at home 1.217 .293 Maternal Characteristics Literacy 1.239 .622 1.152 .408 1.256 .386 Employment 1.642 1.437 Regional Characteristics Atlantic Urban 1.926 1.192 2.195 1.394 2.328 1.657 Atlantic Rural 2.747 1.956 3.173 2.150 2.279 1.761) Central Urban .843 .272 .811 .320 .989 .0 19 Central Rural 2.119 1.528 2.205 1.533 1.786 1.242 Managua 1.287 .403 1.284 .384 2.190 .1.23 7 Pacific Urban .722 .591 .711 .627 ..799 .449 Pseudo R Squared .108 .108 .116 N 2563 2561 2568 Source: Nicaragua LSMS 1998 Annex 19, Page I Anex 19 - Mapa de Pobreza Extrema de Nicaragua I. ANTECEDENTES 1. El objetivo del Mapa de Pobreza Extrema de Nicaragua es ordenar de acuerdo a su nivel de pobreza extrema a diferentes zonas geograficas, como las regiones,' los departamentos y los municipios,. Este ordenamiento da una idea confiable de la distribuci6n de la pobreza en todo el pais y se considera una excelente herramienta para la planeaci6n de politicas y programas para priorizar y asignar eficazmente los recursos que contribuyen a la reducci6n de la pobreza. A este proceso se le conoce como "focalizaci6n," y su objetivo es incrementar la eficacia de los recursos en atender a los mas pobres. 2. En este documento se detalla la elaboraci6n y resultados del Mapa de Pobreza Extrema de Nicaragua usando el Censo de Poblaci6n y Vivienda de 1995 (Censo95) y la Encuesta Nacional de Hogares sobre Medici6n de Nivel de Vida de 1998 (EMNV98).2 Asi mismo, se proporcionan ejemplos para el uso del Mapa de Pobreza Extrema como herramienta de focalizaci6n para decisiones de programaci6n de recursos a nivel sectorial. La combinaci6n de datos de los censos con los de encuestas de hogares (conteniendo informaci6n sobre el consumo de las personas) es una manera de aprovechar la amplitud del censo y la informaci6n detallada de las encuestas de hogares. 3. Este Mapa de Pobreza Extrema es una actualizaci6n del Mapa de Pobreza elaborado por el FISE, el cual utiliz6 la EMNV93 y el Censo95. En un contexto de escasos recursos, el Mapa de Pobreza Extrema ha sido elaborado clasificando a los municipios del pais con base en la Brecha de Pobreza Extrema, lo que implica un cambio de metodologia, pues el mapa previo fue elaborado sobre la Brecha de Pobreza General. Por esta raz6n, y debido a la utilizaci6n de una fuente mas reciente de informaci6n a nivel de hogares como es la EMNV98 y a la creaci6n de nuevos municipios, la actualizaci6n del S se refiere a las siete macro regiones geograficas del pais diferenciadas por zonas urbano y rural para las cuales los resultados de la Encuesta Nacional de Medici6n de Nivel de Vida 1998 (EMNV98) son significativos. Los departamentos por macro regiones son: la Regi6n Managua, que incluye Managua Urbano y Managua Rural; la Regi6n Pacifico, que incluye Chinandega, Le6n, Masaya, Granada, Carazo y Rivas; la Regi6n Central, que incluye Nueva Segovia, Jinotega, Madriz, Esteli, Matagalpa, Boaco y Chontales; y la Region Atlantico, que abarca la RAAN (Regi6n Aut6noma del Atlantico Norte), la RAAS (Regi6n Aut6noma del Atlantico Sur) y Rio San Juan. 2 Para este trabajo se utiliz6 la metodologia de Hentschel et. al. (2000), ya que nos permite calcular diferentes medidas de pobreza y a la vez determinar la precisi6n y confiabilidad de los resultados asociadas con cada una de las estimaciones. Annex 19, Page 2 Mapa de Pobreza Extrema implica un nuevo ordenamiento de los municipios por nivel de pobreza. Los tomadores de decisi6n tendran que considerar un nuevo esquema de priorizaci6n de esfuerzos y asignaci6n de recursos a nivel municipal. Esta publicaci6n es la primera de una serie que busca generalizar el conocimiento y uso del Mapa de Pobreza Extrema en la asignacion de los recursos a nivel municipal. En una siguiente publicaci6n se presentaran los resultados de la elaboraci6n conjunta con los sectores de los indicadores prioritarios y metas para cada sector, asi como una elaboraci6n detallada de los criterios para la aplicaci6n del Mapa para el disefio y focalizaci6n de programas para la reducci6n de la pobreza. 4. Con estas publicaciones se espera generalizar el uso del Mapa de Pobreza Extrema para priorizar las areas con mayores niveles de pobreza, y ampliar las capacidades para asignar recursos buscando un balance 6ptimo entre la equidad y la eficiencia que logre el mayor impacto posible con los recursos disponibles para la disminuci6n de la pobreza y la pobreza extrema. Se espera que estas publicaciones se conviertan en herramientas para la planeaci6n mediante procesos amplios de capacitaci6n, asi como para la implementaci6n y monitoreo de la Estrategia Reforzada de Reducci6n de la Pobreza (ERRP). 5. Este primer trabajo se ha realizado como una colaboraci6n conjunta de la Secretaria Tlcnica de la Presidencia (SETEC), del Programa para el Mejoramiento de las Encuestas de Condiciones de Vida en Nicaragua (MECOVI), del Instituto de Estadisticas y Censos de Nicaragua (INEC), del Fondo de Inversi6n Social de Emergencia de Nicaragua (FISE), y del Banco Mundial (BM). 2 Annex 19, Page 3 II. INTRODUCCION 6. El Mapa de Pobreza Extrema de Nicaragua es una herramienta para focalizar, ya que describe detalladamente y de manera confiable la distribuci6n espacial de la pobreza. El Mapa de Pobreza Extrema contiene informaci6n sobre pobreza a nivel regional, departamental y municipal que puede utilizarse como una herramienta para priorizar esfuerzos y asignar recursos proporcionalmente a las areas en donde la pobreza es mas profunda y en donde hay un mayor numero de pobres. 7. Otros tipos de informaci6n acerca de estas areas/municipios se puede sobreponer al Mapa de Pobreza Extrema para la toma de decisiones. Por ejemplo, el Mapa de Pobreza Extrema se puede combinar con un Mapa de Cobertura de Servicios para educaci6n, posibilitando la focalizaci6n de los recursos en aquellos municipios donde la pobreza extrema es mas profunda y donde los vacios de asistencia educativa son mas altos. Por ultimo, estos mapas combinados deben utilizarse conjuntamente con la participaci6n de los beneficiarios y de las autoridades locales en la asignaci6n de los beneficios de los programas a nivel local. III. FUENTES DE DATOS 8. Las fuentes de datos son el Censo95 y la EMNV98. Para la elaboraci6n del Mapa se utilizaron s6lo aquellas preguntas y variables del Censo95 y de la EMNV98 que son iguales 6 muy similares. Las preguntas/variables del Censo95 que no se encontraban en la EMNV98 no se incluyeron, como tampoco preguntas de la EMNV98 que no estaban en el Censo95. Tambien fue necesario utilizar el calculo del agregado de consumo y lineas de pobreza realizado por la SETEC con apoyo tecnico del Banco Mundial. 9. Los paquetes estadisticos utilizados fueron el Statistical Package for the Social Sciences (SPSS) y el Statistical Analysis Software (SAS). Tambien se hizo uso de un programa en SAS, desarrollado por el BM, para el cdlculo final de las medidas de pobreza. Todos las instituciones participantes tuvieron acceso sin restricci6n alguna a todas las fuentes de datos y programas desarrollados durante esta tarea. IV. BREVE DESCRIPCION METODOLOGICA 10. La metodologia utilizada para generar este Mapa de Pobreza Extrema nos permite contar con una descripci6n espacial de la pobreza estimada a nivel municipal, la cual no se puede calcular utilizando solamente los censos de poblaci6n y vivienda como el 3Estrategia Reforzada de Reducci6n de la Pobreza del Gobiemo de Nicaragua (2000) y Reporte de Pobreza del Banco Mundial (2000). 3 Annex 19, Page 4 Censo95 6 solamente con la encuestas de niveles de vida como la EMNV98. Por un lado, un censo de poblaci6n y vivienda no proporciona informaci6n suficiente sobre gastos del hogar4 a nivel municipal, y por otro lado las encuestas de niveles de vida no proporcionan informaci6n representativa a nivel municipal. 11. La unica base de datos disponible en Nicaragua con informaci6n sobre hogares a nivel de municipio es el Censo95. Por tanto, el Censo95 es la base que permite el ordenamiento de municipios para focalizar las intervenciones para la reducci6n de la pobreza. En Nicaragua la encuesta representativa de hogares a nivel nacional con informaci6n de pobreza mas reciente es la EMNV98. Por tanto, la mejor manera para focalizar es combinar estas dos fuentes de informaci6n para construir un Mapa de 5,6 Pobreza Extrema para Nicaragua.' 12. El Mapa de Pobreza Extrema de Nicaragua es la herramienta mas confiable que existe al momento para prop6sitos de determinar con razonable precisi6n: (i) cuantos son los pobres y pobres extremos en el pais; (ii) en que regiones, departamentos y municipios se encuentran ubicados; (iii) que porcentaje de la poblaci6n total y rural municipal representan; y, (iv) cual exactamente es la proporci6n que su consumo per capita se encuentra por debajo de la linea de pobreza, concepto definido como la brecha de la pobreza. V. MAPA DE POBREZA EXTREMA DE NICARAGUA 13. El Mapa de Pobreza Extrema de Nicaragua se presenta en el Grafico 1. En los resultados de la primera etapa del Mapa de Pobreza Extrema es importante hacer notar que en todos los 151 municipios de Nicaragua hay pobres generales y pobres extremos. Con fines de facilitar la focalizaci6n, en la primera etapa el mapajerarquiza a los 151 municipios que integran a Nicaragua en cuatro diferentes estratos o niveles (severo, alto, medio y menor), con base en los valores de la brecha de la pobreza extrema en cada municipio. En una segunda etapa, el mapa calcula para los 151 municipios la brecha de pobreza extrema municipal (en US$) como proporci6n de la brecha de pobreza extrema total a nivel nacional (en US$). Este mapa proporciona a los tomadores de decisi6n una 4Imposibilitando el calculo del agregado de consumo para todos los hogares del Censo95, y por tanto de medidas de pobreza a partir del Censo95. 5El Apendice I contiene una descripci6n de las principales medidas de pobreza utilizadas, el Apendice 2 contiene una descripci6n de las variables utilizadas comunes al Censo95 y a la EMNV98, el Apendice 3 contiene una descripci6n completa de la metodologia utilizada, y el Apdndice 4 contiene las pruebas estadisticas que validan el Mapa de Extrema Pobreza. 6 El Mapa de Extrema Pobreza define pobreza en base a la metodologia del agregado de consumo y las lineas de pobreza. Informaci6n mas detallada sobre el por que se prefiere esta metodologia y no otras corno la de Necesidades Basicas Insatisfechas se puede encontrar en Hentschel et al (2000), Elbers et al (2000) y Alderman et al (2000). 4 Annex 19, Page 5 herramienta de focalizaci6n que jerarquiza 6 prioriza a los municipios (etapa 1), y que permite asignar recursos para la reducci6n de la pobreza de una manera objetiva y proporcional a la profundidad de la pobreza extrema en cada municipio (etapa 2). 14. En un primer momento (etapa) el Mapa de Pobreza Extrema genera cuatro estratos basados en los niveles de la brecha de pobreza extrema.7 Los municipios con una brecha de 12.23 por ciento y mas son considerados en el estrato de Pobreza Severa, aquellos municipios con un rango entre 9.5 y 11.99 por ciento son considerados en Pobreza Alta, los municipios con un rango entre 7 y 9.2 por ciento son considerados en Pobreza Media, y los municipios con una brecha menor al 6.6 por ciento son considerados en Pobreza Menor. 7Los cuatro estratos se separaron utilizando un analisis de promedios por conglomerados 6 K-means Cluster Analysis. 5 Annex e 6 Grafico A19.1 El Mapa de Pobreza Extrema de Nicaragua NICARAGUA MAPA DE POBREZA E)CTREMA- Honduras ~~~~~~~~~~~~~~~~~~~~~~~~~I Pacffico I __ Mar Caribe Rangos de la Brecha de Pcbreza Extrema - Pobreza Severa _Pobreza Alta | Pobreza Media rAAD - Pobreza Menor ca 6 Annex 19, Page 7 15. En una segundo momento (etapa), el mapa cuantifica en dMlares la suma de las brechas de la pobreza con respecto a la linea de pobreza de cada individuo extremadamente pobre por area geografica. Estos montos en d6lares se utilizan para calcular la proporci6n de recursos necesarios para cerrar la Brecha de la Pobreza Extremna en cada area geografica con respecto a la Brecha de la Pobreza Extrema Total Nacional. Asi por ejemplo, si el total de recursos necesarios para cerrar la Brecha de la Pobreza Extrema Total a Nivel Nacional es 100 por ciento, entonces cada regi6n, departamento y/o municipio recibiria proporcionalmente lo que le corresponde con base en la contribuci6n a la profundidad de la pobreza en cada area geografica al total nacional. Por tanto, las areas geograficas donde la pobreza extrema es mas profunda y donde el nuimero de pobres extremos es mayor recibirian proporcionalmente mayores recursos dedicados a la reducci6n de la pobreza. 16. Es fundamental hacer notar que la cuantificaci6n en dMlares de la profundidad de la pobreza general y extrema representa la diferencia del consumo de los pobres y pobres extremos con respecto a las lineas de pobreza general y extrema. Estas brechas de la pobreza son indicativas de la proporci6n total de los recursos que deben dedicarse a cada area geografica a traves de diversos programas y no se pueden extrapolar para concluir que con ellos se solucionaria la pobreza. Ademas, las brechas de la pobreza por area geografica no incluyen ninguin costo administrativo de los programas ya que se basan en la diferencia entre el consumo individual y la linea de pobreza, por lo que tampoco es posible concluir que el monto total de las brechas de pobreza representa el costo de los programas para combatirla. Los programas y proyectos dedicados a la reducci6n de la pobreza intentan mejorar las condiciones de vida de los pobres y extremadamente pobres, mas alla de mejorar su nivel de consumo. 17. Los Cuadros 1, 2 y 3 presentan los resultados del Mapa de Pobreza Extrema de Nicaragua por regi6n, departamento y municipio, respectivamente. Estos cuadros incluyen los indicadores del (1) al (4) para la primera etapa del Mapa de Pobreza Extrer na y los indicadores del (5) y (6) para la segunda etapa del Mapa de Pobreza Extrema por area geografica.8 A continuacion se presenta una definici6n de los indicadores incluidos en los cuadros: (1) Extensi6n de la Pobreza Extrema: Es el porcentaje 6 proporci6n de la poblaci6n total que se encuentra por debajo de la linea de pobreza extrema. La linea de pobreza extrema en 1998 es C$2,246 per cApita por afno 6 alrededor de US$212. Una persona es clasificada en pobreza extrema si su consumo per capita por afno es menor a la linea de pobreza extrema y por tanto no puede satisfacer su, Las f6rmulas de los indicadores de pobreza se encuentra en el Apendice 1. Annex 19, Page 8 requerimientos minimos cal6ricos diarios auin cuando dedicase la totalidad de su consumo a alimentos. (2) Nuimero Estimado de Pobres Extremos: Es el numero de personas extremadamente pobres cuyo consumo per capita anual se encuentra por debajo de la linea de pobreza extrema. (3) Porcentaje de Pobres Extremos Rurales: Es el porcentaje 6 proporci6n del total de pobres extremos que vive en areas rurales. (4) Brecha de la Pobreza Extrema: Representa la cantidad necesaria para Ilevar el consumo de todos los pobres extremos hasta la linea de pobreza extrema, expresada como un porcentaje de la linea de pobreza extrema y tomando en cuenta la proporci6n de la poblaci6n pobre extrema en el total de la poblaci6n nacional. (5) Brecha de la Pobreza Extrema en miles de dMlares de 1998: Es la cantidad necesaria para llevar el consumo de todos los pobres extremos hasta la linea de pobreza extrema, expresada en dMlares. (6) Proporci6n de la Brecha de la Pobreza Extrema Nacional: Es el porcentaje 6 proporci6n de la Brecha de la Pobreza Extrema Total Nacional que corresponde a cada area geografica. Se calcula una vez que se han sumado las Brechas de la Pobreza Extrema en d6lares (columna 5). A esta sumatoria se le define como Brecha de la Pobreza Extrema Total Nacional, y cada area geogrAfica representa una proporci6n de este total. 18. El Cuadro 3 agrupa a los 151 municipios de Nicaragua en cuatro estratos de pobreza. Los Estratos de Pobreza del Mapa de Pobreza Extrema se generan al jerarquizar a los 151 municipios que integran Nicaragua en cuatro diferentes niveles (pobreza severa, pobreza alta, pobreza media, y pobreza menor), de acuerdo a los valores de la brecha de la pobreza extrema (columna 4) en cada municipio como se detall6 anteriormente en el parrafo 14. 19. A manera de ejemplo, la primera etapa del Mapa de Pobreza Extrema para Waspan genera los siguientes resultados: la extensi6n de la pobreza extrema estimada con base en el analisis conjunto del Censo95 y la EMNV98 es 64.1 por ciento (columna 1), que representa un nuimero estimado de pobres extremos de 18,561 (columna 2). El 95.0 por ciento de los pobres extremos vive en areas rurales (columna 3), lo que implica que tan s6lo el 5.0 por ciento de los pobres extremos vive en zonas urbanas. La brecha de la pobreza extrema 6 profundidad de la pobreza extrema es 40.0 por ciento (columna 4). Esta ulitima columna indica que Waspan se encuentra en la segunda posici6n en orden de jerarquia para priorizar esfuerzos para la reducci6n de la pobreza extrema. Asi mismo, Annex 19, Page 9 Waspan se encuentra en el estrato de pobreza extrema mas aguda, que es el de Pobreza Severa. 20. Usando nuevamente Waspan a manera de ejemplo, la segunda etapa del Mapa de Pobreza Extrema genera los siguientes resultados: la brecha de la pobreza extrema en miles de US$ de 1998 es de 2,462 (columna 5) que representan el monto requerido para que todos los pobres extremos alcancen un consumo de al menos la linea de la pobreza extrema; y, la proporci6n de la brecha de la pobreza extrema total nacional es 4.18 por ciento (columna 6), que corresponde a la proporci6n que la brecha de la pobreza extrema municipal de Waspan aporta a la brecha de la pobreza extrema total nacional. Esta ultima columna indica que del total de recursos nacionales asignables a la reducci6n de la pobreza, Waspan debe recibir proporcionalmente un 4.18 por ciento. 21. El Cuadro 4 presenta un Resumen de Indicadores del Mapa de Pobreza Extrema de Nicaragua por Estrato. En los resultados de la primera etapa del Mapa de Pobreza Extrema es importante hacer notar que en todos los 151 municipios de Nicaragua hay pobres generales y pobres extremos. Sin embargo, y de acuerdo a los niveles de la brec-a de pobreza extrema, 31 municipios se encuentran en Pobreza Severa, 34 en Pobreza Alta, 34 en Pobreza Media, y 52 en Pobreza Menor. Esto se debe a que la profundidad 6 brecha de la pobreza extrema varia desde 48.0 por ciento en Prinzapolka (ver Cuadro 3 columna 4) hasta 0.61 por ciento en Managua. Los resultados de la segunda etapa del Mapa de Pobreza Extrema, corresponden a la proporci6n que la brecha de la pobreza extrema en cada estrato aporta a la brecha de la extrema pobreza total nacional. Asi, al estrato de Pobreza Severa le corresponde el 30.8 por ciento de los recursos asignables a la reducci6n de la pobreza, al de Pobreza Alta el 26.8 por ciento, al de Pobreza Media el 18.8 por ciento y al de Pobreza Menor el 23.6 por ciento. Annex 19, Page 10 Cuadro A19.1 Indicadores del Mapa de Pobreza Extrema de Nicaragua por Region 1 2 3 4 5 6 Brecha de la Proporcion de Extension Numero % Pobres Brecha de la Pobreza la Brecha de Region de za Estimado de Extremos Pobreza Extrema la Pobreza Pobreza Pobres Rurales Extrema (%) US$1998 Extrema Extrema (%) Extremos ~~~~(miles) Nacional (%) Atlantico 36.0 187,652 84.4 13.7 15,203 25.8 Central 32.2 433,907 81.2 10.0 28,537 48.5 Pacifico 18.2 248,819 67.5 4.6 13,427 22.8 Managua 3.6 39,194 15.8 .7 1,713 2.9 Cuadro A19.2 Indicadores del Mapa de Pobreza Extrema de Nicaragua por Departamento 1 2 3 4 5 6 Extension Numero Brecha de la Proporcion de dela Estimado de % Pobres Brecha de la Pobreza la Brecha de Departamento Pobreza Pobres Extremos Pobreza Extrema la Pobreza Extrema Pobres Rurales Extrema (%) US$1998 Extrema (%) Extremos (miles) Nacional (%) RAAN 43.7 7,173 92.0 20.3 7,817 13.3 Jinotega 37.0 89,533 92.2 12.6 6,485 11.0 Rio San Juan 36.3 25,369 86.7 12.1 1,793 3.0 Madriz 37.1 39,695 85.7 11.3 2,572 4.4 Nueva Segovia 34.1 55,151 68.2 10.3 3,534 6.0 Boaco 32.8 44,787 86.9 10.0 2,891 4.9 RAAS 30.8 83,109 76.5 9.7 5,592 9.5 Matagalpa 31.9 121,852 83.4 9.5 7,736 13.1 Chontales 29.4 42,253 68.9 9.2 2,819 4.8 Esteli 23.4 40,636 70.0 6.8 2,500 4.2 Chinandega 20.9 72,924 64.0 5.5 4,041 6.9 Rivas 20.3 28,427 83.9 5.2 1,529 2.6 Leon 19.0 63,736 71.4 4.9 3,484 5.9 Granada 17.0 26,203 62.6 4.3 1,414 2.4 Carazo 15.5 23,047 64.4 3.8 1,203 2.0 Masaya 14.3 34,482 60.1 3.4 1,756 3.0 Managua 3.6 39,194 15.7 .7 1,713 2.9 Annex 19, Page 11 Cuadro A19.3 Indicadores del Mapa de Pobreza Extrema de Nicaragua por Municiplo 1 2 3 4 5 6 Extension Brecha de Proporcion de Estrato de la Estimado % Pobres laPNBrecha de la Pobreza la Brecha de la de Departamento Municipio Pobreza Estimado Extremosa Pobreza Extrema Pobreza Pobreza Extrema de Pobres Rraes Extrema US$1998 Extrema (%/) Extremos (% (miles) Nacional (%) RMN Prinzapolka 76.3 3,291 Y6./ 48.0 440 .75 RAAN Waspan 64.1 18,561 95.0 40.0 2,462 4.18 RAAS Desemb. c.r.g. 43.3 1,233 55.7 20.8 126 .21 RAAN Bonanza 37.9 4,179 94.7 20.4 477 .81 RAAN Pto. Cabezas 33.7 12,509 84.0 18.5 1,458 2.48 Jinotega Cua-bocay 43.8 24,682 97.7 18.3 2,188 3.72 Jinotega Wiwili 49.2 18,679 96.0 17.5 1,408 2.39 Madriz Totogalpa 49.3 4,341 92.1 16.1 301 .51 Nueva Segovia Santa Maria 48.9 1,808 94.8 16.0 126 .21 RAAS Ayote 46.0 3,912 71.1 15.4 278 .47 Esteli San Nicolas 45.6 2,824 95.8 15.1 199 .34 RAAN Waslala 43.0 14,098 94.5 14.7 1,021 1.73 r< Madriz S.J. de Cusmapa 46.6 2,653 87.1 14.6 176 .30 > RAAS C.R. Grande 43.0 5,838 94.9 14.5 419 .71 w Chontales Comalapa 44.5 4,668 97.3 14.5 323 .55 Rio San Juan El Castillo 41.1 3,985 98.8 14.2 293 .50 u Nueva Segovia Macuelizo 45.2 2,143 97.2 14.0 141 .24 m RAAN Siuna 41.2 21,831 93.6 14.0 1,571 2.67 o RAAS Paiwas 42.2 13,876 91.7 13.9 971 1.65 - Rio San Juan San Miguelito 39.7 5,368 89.4 13.6 390 .66 Madriz San Lucas 43.5 4,555 96.4 13.4 298 .51 RAAS El Tortuguero 39.9 3,463 88.2 13.1 240 41 Nueva Segovia C. Antigua 42.7 1,454 73.8 13.0 94 .16 Nueva Segovia Mozonte 41.9 2,171 78.0 12.9 142 .24 RAAN Rosita 33.2 4,704 81.4 12.9 389 .66 Madriz Las Sabanas 41.2 1,664 85.4 12.8 110 19 Nueva Segovia Wiwli de Abajo 41.4 5,746 87.5 12.7 375 .64 Madriz Telpaneca 41.4 6,259 93.2 12.6 404 .69 Nueva Segovia Quilali 40.2 8.277 65.3 12.6 550 .93 Matagalpa Rancho Grande 41.2 7,029 97.7 12.5 454 .77 Chontales Santo Domingo 37.5 4,686 79.2 12.2 325 .55 Matagalpa Matiguas 38.8 14,907 91.4 12.0 979 1.66 Jinotega S.M. dePantasma 39.8 11,801 94.8 12.0 754 1.28 Esteli San Juan de Limay 38.1 4,761 83.0 11.9 316 .54 Nueva Segovia Murra 39.7 4,370 98.0 11.9 278 .47 Matagalpa Muy Muy 37.4 4,871 89.9 11.6 321 .55 Rio San Juan San Carlos 35.1 9,991 80.3 11.6 701 1.19 Matagalpa Tuma-la Dalia 38.7 16,948 94.5 11.6 1,080 1.83 Chontales La Libertad 36.6 3,574 92.3 11.6 241 .41 Matagalpa San Ramon 38.1 8,787 96.8 11.6 566 .96 Boaco San Lorenzo 37.4 8,338 84.7 11.5 543 .92 Boaco Teustepe 38.0 8,082 94.0 11.3 512 .87 Matagalpa San Dionisio 37.6 6,016 90.1 11.1 376 .64 Matagalpa Rio Blanco 35.9 9,357 77.5 11.0 612 1.04 Rio San Juan S. Juan del Norte 32.2 87 23.4 11.0 6 .01 < Rio San Juan Morrito 34.2 2,076 87.9 10.9 141 .24 < Chontales Acoyapa 33.7 5,634 73.7 10.8 384 .65 Chontales El Coral 35.0 2,504 63.0 10.8 164 .28 N Chontales S.F. de Cuapa 35.0 1,744 78.5 10.7 114 .19 CnMadriz Yalaguina 36.3 2,723 84.7 10.7 171 .29 0 Jinotega S.S. deYali 35.4 7,162 92.1 10.5 452 .77 EL Boaco Camoapa 32.9 10,576 83.3 10.5 717 1.22 Rio San Juan El Almendro 32.8 3,862 87.9 10.5 262 .45 Madnz S.J. Rio Coco 34.8 5,256 95.3 10.4 335 .57 Boaco S.J. Remates 35.0 2,676 92.5 10.4 168 .29 Matagalpa Terrabona 36.0 3,812 94.1 10.4 233 .40 Nueva Segovia El Jicaro 34.9 7,674 77.9 10.4 483 .82 RAAS N. Guinea 33.5 26,479 63.8 10.3 1,727 2.93 Nueva Segovia Dipilto 35.0 1,357 95.2 10.3 85 .14 RAAS El Rama 31.7 14,471 84.6 10 1 977 1.66 Matagalpa Esquipulas 33.4 4,923 82.5 9.9 308 .52 Chontales Villa Sandino 31.2 4,136 69.2 9.8 276 .47 Jinotega S.R. del Norte 32.9 4,611 90.9 9.6 285 .48 Nueva Segovia Jalapa 31.9 13,274 56.0 9.5 840 1.43 Esteli Pueblo Nuevo 32.4 6,327 93.2 9.5 393 .67 Annex 19, Page 12 Cuadro 3 Indicadores del Mapa de Pobreza Extrema de Nicaragua por Municipio (cont.) 1 2 3 4 5 6 Extension umero Brecha de Brecha de Proporcion de Estrato de la Nutmero %Pbelaobreha dela Pobreza la Brecha de la de Departamento Municipio Pobreza de Pobres Extremos Extrema Extrema Pobreza Pobreza Extrema Ruraemo US$1998 Extrema Extremos R (miles) Nacional (%) Madriz Somoto 30.5 8,712 69.3 9.18 557 .95 Rivas Cardenas 32.2 1,484 94.6 9.09 89 .15 Chinandega S.Pedro del N. 34.8 1,406 94.1 9.03 78 .13 Matagalpa C. Dario 30.6 10,964 85.1 8.90 676 1.15 Madriz Palacaguina 29.9 3,531 76.4 8.77 220 .37 Nueva Segovia San Fernando 30.3 1,801 73.0 8.76 111 .19 Matagalpa San Isidro 29.4 4,501 85.8 8.74 285 .48 esteli Condega 30.0 7,477 81.0 8.66 458 .78 Chontales San P. de Lovago 27.9 1,990 82.7 8.61 130 .22 RAAS Muelle de B. 28.3 6,575 93.6 8.59 424 .72 Leon Sta. Rosa del P 32.3 2,952 89.4 8.55 166 .28 Boaco Boaco 28.5 12,852 84.1 8.52 816 1.39 Chinandega Villanueva 31.8 7,100 94.1 8.50 402 .68 < Chinandega Sto. Tomas del N. 32.7 2,220 96.4 8.41 121 .21 : Jinotega La Concordia 29.2 2,054 88.6 8.23 123 .21 uI RAAS K. Hill 27.4 2,039 71.8 8.12 128 .22 Chinandega Posoitega 29.5 4,524 81.7 8.02 261 .44 Chinandega S.Fco. del N. 31.1 1,858 96.6 7.96 101 .17 r Rivas Altagracia 30.1 5,299 94.8 7.94 296 .50 O Jinotega Jinotega 26.8 20,543 81.4 7.82 1,275 2.17 ML Matagalpa Sebaco 26.3 6,524 46.8 7.81 411 .70 Leon Achuapa 30.5 4,018 91.9 7.79 218 .37 Boaco Santa Lucia 27.7 2,263 96.3 7.77 135 .23 Chontales Sto. Tomas 25.2 4,015 60.9 7.73 262 .44 Rivas Tola 29.2 5,800 94.5 7.72 325 .55 Chinandega Pto. Morazan 28.2 3,214 72.9 7.69 186 .32 RAAS L. Perlas 23.5 1,447 58.1 7.56 99 .17 Leon El Sauce 28.6 7,405 87.0 7.36 405 .69 Leon El Jicaral 29.2 2,915 97.1 7.26 154 .26 Chinandega Somotillo 27.8 6,858 70.1 7.20 377 .64 Leon Telica 26.9 6,124 82.8 7.17 347 .59 Chinandega ElViejo 26.0 17,899 72.2 7.12 1,042 1.77 Chinandega Cinco Pinos 28.3 1,758 95.2 7.03 93 .16 Esteli La Trinidad 25.1 4,641 80.5 7.00 275 .47 Annex 19, Page 13 Cuadro A19.3 Indicadores del Mapa de Pobreza Extrema de Nicaragua por Municipio (cont.) 1 2 3 4 5 6 Extension Numero Brecha de Brecha de Proporcion cle Estrato de la Numado % Pobres Brezad la Pobreza la Brecha de la de Departamento Municipio Pobreza de Pobres Extremos Extrema Extrema Pobreza Pobreza Extrema Rurales (%) US$1998 Extrema (%) Extremos (miles) Nacional (%,) Leon Larreynaga 25.0 7,443 86.7 6.60 417 .71 Matagalpa Matagalpa 22.4 23,212 67.7 6.52 1,435 2.44 Leon Quezalguaque 25.1 1,944 92.8 6.28 103 .18 Carazo La Conquista 25.1 960 95.0 6.27 51 09 Chontales Juigalpa 20.5 9,302 37.9 6.22 600 1.02 Rivas Belen 24.2 3,846 86.4 6.11 206 .35 Granada Diriomo 22.9 4,606 75.4 5.82 248 .42 Leon La Paz Centro 21.8 5,972 55.4 5.79 336 .57 Carazo La Paz de Carazo 22.4 909 70.0 5.75 49 AB Nueva Segovia Ocotal 19.7 5,077 5.8 5.62 308 .32 Granada Nandaime 21.0 6,823 77.8 5.32 367 .62 Carazo Sta. Teresa 20.1 3,489 91.8 5.04 186 .32 Rivas S.J. del Sur 19.7 2,566 84.7 4.87 135 .23 Masaya Tisma 20.2 2,017 74.9 4.86 103 .17 Granada Diria 19.3 1,170 70.1 4.82 62 .11 Masaya Niquinohomo 18.6 2,502 79.0 4.52 129 .:22 Chinandega El Realejo 17.9 1,452 72.4 4.51 78 .13 esteli Esteli 15.9 14,604 41.8 4.40 859 1.46 Rivas Buenos Aires 18.8 905 81.8 4.40 45 A8 Masaya La Concepcion 17.9 4,955 76.3 4.26 250 .42 Carazo El Rosario 17.1 671 63.1 4.24 35 .)6 Rivas Potosi 17.5 1,873 77.4 4.23 96 .16 r Rivas Moyogalpa 17.2 1,493 76.5 4.20 77 .13 o Carazo San Marcos 16.2 4,182 47.4 4.07 223 .38 u Masaya Nandasmo 16.3 1,283 47.3 3.90 65 .11 2 chinandega Chinandega 15.0 17,394 27.6 3.77 928 1.58 < Chinandega Chichigalpa 14.8 6,169 55.1 3.64 323 .55 u Masaya Masatepe 15.4 3,933 61.8 3.64 197 .33 o Granada Granada 14.2 13,604 50.1 3.62 737 1 25 0 o Masaya S.J. de Oriente 15.0 465 62.6 3.60 24 34 Carazo Jinotepe 13.8 5,128 65.9 3.41 270 46 Masaya Nindiri 14.2 4,102 81.6 3.36 206 35 Leon Leon 13.0 20,892 53.9 3.32 1,137 1 33 CArazo Diriamba 13.9 7,080 59.2 3.29 356 30 Leon Nagarote 14.0 4,071 48.7 3.28 202 34 Masaya Catarina 13.9 989 75.8 3.26 49 08 Masaya Masaya 12.1 14,236 42.4 2.94 734 1 25 Carazo Dolores 11.6 627 18.5 2.81 32 05 Rivas San Jorge 11.5 809 34.0 2.76 41 07 Rivas Rivas 11.6 4,353 65.0 2.73 217 37 RAAS Bluefields 9.2 3,425 34.0 2.38 187 32 Managua S.F. Libre* 7.9 691 83.5 1.66 31 .05 Chinandega Corinto 6.3 1,073 8.3 1.45 53 .09 RAAS C. Island 6.6 350 0.0 1.45 16 .03 Managua El Crucero* 6.2 781 4.0 1.36 36 .06 Managua S.R. del S.* 6.2 2,250 49.5 1.31 101 17 Managua Villa C.Fonseca* 6.3 1,530 90.6 1.30 67 11 Managua Mateare' 6.1 1,076 36.2 1.27 48 08 Managua Tipitapa* 5.9 4,883 18.0 1.26 221 .38 Managua Ticuantepe' 5.1 1,022 61.4 1.09 46 08 Managua C. Sandino' 3.8 2,038 4.1 .74 85 14 Managua Managua* 3.0 24,924 4.3 .61 1,078 1 83 Annex 19 14 Cuadro A19.4 Resumen de Indicadores del Mapa de Pobreza Extrema de Nicaragua por Estrato de Pobreza Estrato de Pobreza Indicador Pobreza Severa Pobreza Alta Pobreza Media Pobreza Menor Total Numero de Municipios 31 34 34 52 151 Numero Estimado de Pobres Extremos 220,487 243,168 184,765 261,149 909,571 Pobres Extremos Nacional (%) 24.3% 26.7% 20.3% 28.7% 100% Numero Estimado de Pobres Extremos Rurales 201,209 202,568 150,118 131,027 684,922 Pobres Extremos Rurales Nacional (%) 29.4% 29.6% 21.9% 19.1% 100% Brecha de Pobreza Extrema US$1998 (mill) 18.15 15.80 11.04 13.89 58.88 Proporcion de ia Brecha de la Pobreza Extrema Nacional (%) 30.8% 26.8% 18.8% 23.6% 100% 14 Annex 19, Page 15 VI. EL MAPA DE POBREZA EXTREMA COMO UNA HERRAMIENTA PARA FOCALIZAR RECURSOS 22. La utilizaci6n del Mapa de Pobreza Extrema como herramienta de focalizaci6n de la Estrategia Reforzada de Reducci6n de Pobreza busca el mayor impacto posible con los recursos disponibles, tanto en la pobreza como en la pobreza extrema. En el disefio y focalizaci6n de politicas y programas es indispensable diferenciar las caracteristicas de la poblaci6n objetivo (entre otras cosas, si la poblaci6n objetivo es de pobres extremos o de pobres en general), asi como el tipo de problematica que se busca resolver. El enfasis en cerrar y prevenir el agravamiento de la brecha de pobreza extrema es parte de la preocupaci6n.de protecci6n social mientras que, paralelamente, se promueve el desarrollo econ6mico y la inversi6n en capital humano de todos los pobres en general. Las decisiones acerca del gasto publico en la lucha contra la pobreza deberan ser guiadas por la buisqueda de un punto 6ptimo entre la equidad para atender a los pobres extremos, y la eficiencia en la reducci6n de los niveles de pobreza en general y de pobreza extrema en particular. 23. El Mapa de Pobreza Extrema de Nicaragua se convierte en una herramienta para focalizar recursos una vez que las entidades del Sector Puiblico lo utilizan para priorizar sus acciones por area geografica, en particular a nivel municipal. El Mapa de Pobreza Extrema se debe combinar con otros indicadores para asi priorizar aquellos municipios donde hay una coincidencia entre la pobreza mas aguda y los mas altos vacios de cobertura de servicios publicos. 24. Para el Sector Educaci6n, la Asistencia Escolar a nivel primario es un indicador utilizado para priorizar inversiones. El Grafico 2 contiene informaci6n para cuatro municipios del "Mapa de Pobreza Extrema" y del "Mapa de Hogares con al menos un nifio de 7-12 afnos que no asiste a la Escuela." Observese que el municipio de San Nicolds tiene un nivel severo de pobreza extrema que es mas agudo que el de San Sebastian de Yali (nivel alto de pobreza extrema). No obstante esta diferencia, ambos municipios comparten el mismo estrato (alto) de hogares con al menos un ninio de 7-12 afios que no asiste a la escuela. La yuxtaposici6n de estos dos mapas muestra claramente que el municipio de mayor prioridad entre los cuatro, para incidir en la Asistencia Escolar, es San Nicolas puesto que en este coinciden un estrato de pobreza severa con el estrato mas alto de inasistencia escolar. 15 Annex 19, Page 16 GRAFICO A19.2 - MAPA PARA PRIORIZAR INVERSIONES EN EDUCACION (NIVELES DE POBREZA EXTREMA v ASISTENCIA ESCOLAR A NIVEL PRIMARIO) Nicrargua: Mapa de Pobreza Hogxes con al menos un nin 7-12 afios que Rangos de la Brecha de la Pobreza Extma no asiste a la Fcuela (%) Condega ~ _ Conde . H:::::::::::::::::::::::.S^il. de Ya ..(17 .3) ........ ajltj 'tllEs '2"te' 0BIgSiS >J .. ................ .. ; s _SanN~an icolas Sniolas{37.1) _ ~~~~~Pobreza Severa _ Bajo (10.8-19.6, 2 obreza AtIda _Medio (1 9.7-28.5' El Pobreza Menor - Alto (28.6-37.1) 25. Asi mismo, para los Sectores Salud y Agua, el Acceso a Agua Segura es un indicador utilizado para priorizar inversiones. El Grafico 3 contiene informacion para cuatro municipios del "Mapa de Pobreza Extrema" y del "Mapa de Hogares con Acceso a Agua Segura." San Nicolas tiene un nivel severo de pobreza extrema que es mas agudo que el de San Sebastian de Yali (nivel alto de pobreza extrema), y simultaneamente comparte con San Sebastian de Yali el estrato mas bajo de hogares con acceso a agua segura. Por tanto, la yuxtaposici6n de los mapas muestra claramente que, para incidir en el Acceso a Agua Segura, el municipio de mayor prioridad entre los cuatro es San Nicolas, ya que es aqui donde coincide un estrato de pobreza severa con el estrato mas bajo de acceso a agua segura. 16 Annex 19, Page 17 GRAFICO A19.3 - MAPA PARA PRIORIZAR INVERSIONES EN SALUD Y AGUA (NIVELES DE POBREZA EXTREMA Y ACCESO A AGUA SEGURA) Nicaragua: Mapa de Pobreza Hogare: con Accesso a Agua Segura (%) Raangos de la Brecha de la Pobreza Extrema Conde r Condie (0.1,7.0, E Pobreza Meno Bj (50 .2-60)12 26 Eso do ejmpo 1lutra la gr, utlia qul Map dePbeaEtead uno de los secto.es relevantes en la ta .d disio . liE a fseguir -s que los sectores5se.apropien5 de esta herramienta s. de Yaque [ .:-- l . . ~~~~~~~~~~~~~~~~~~(64.3) | - \ ~~~~~~~~~~~~~~~~~~~. . . . . . . . . . . .::::::::::::::::: ntermedios ave San cNsensuads enlSEtatgaReozaade idcolan de0 la Pobreza (ERR). Con esta rerramelta sepeen opirzrefElzo Alt asignar.3 objEstosmet los recurslos pausran la greacin utldadqe ela paadobreza,sntmin disrema sisema deimniaragapedey teneraluaciombdenarsmeta yan deinmids deindlaaER rP. piriaisecd uno de los sectore relevantes en la tomadedecisionesdeocalizaci6nElsiguientepas a teure u o scoe earpe de e ta herramienta y elaboren mapas que::: cobnnifrainsbel itiuind la pobreza:::a::nivel:::municipal:::con inomc6 sor lo iniaoe setrae m.' ipratsylsndiaoe in ermdo clae y cosenuads e laEstateia efozad deRedccin d l Porz ER) o sahrrrinas ud osl prorza esfuerzos y asignar:::: obEtiaetel lo reuro par lareucc6nde a obrza sio trnindief:r istma de mntrey vlaind las mtayadfnasenla ERRPr r r xro sx r::: :::::::::;~~~~~~~~~~~~~~1 Annex 19, Page 18 VII. RECOMENDACIONES 27. El Mapa de Pobreza Extrema de Nicaragua es una herramienta para la focalizaci6n a traves de sus dos etapas. La primera jerarquiza a los 151 municipios que integran Nicaragua en cuatro estratos de pobreza extrema (severa, alta, media y menor), utilizando la brecha de la pobreza extrema. La segunda etapa calcula la Proporci6n de la Brecha de la Pobreza Extrema Nacional, indicador que representa el porcentaje o proporci6n de recursos asignables a la reducci6n de la pobreza que debe corresponder a cada region, departamento y municipio del total de recursos disponibles a nivel nacional. Recomendamos que los tomadores de decisi6n utilicen la informaci6n de los indicadores estimados en ambas etapas del Mapa de Pobreza Extrema para focalizar las acciones y recursos dedicados a la reducci6n de la pobreza y asi maximizar su impacto. De tal manera, no s6lo se debe utilizar el Estrato de Pobreza Extrema en el que se ubica cada municipio para priorizar esfuerzos, sino tambien utilizar la Proporci6n de la Brecha de la Pobreza Extrema Nacional para asignar los recursos destinados a la reducci6n de la pobreza. 28. Recomendamos que los tomadores de decisi6n encargados de planear politicas y programas utilicen el Mapa de Pobreza Extrema para la focalizaci6n, incluyendo la priorizaci6n de regiones, departamentos y/o municipios, y la asignaci6n objetiva de recursos para la reducci6n de la pobreza enmarcados en la ERRP. Este mapa es claro y facil de interpretar. No hay grandes suposiciones implicitas en la metodologia, ni hay ninguna duda en cuanto a la calidad de los datos utilizados. 29. La pobreza es un fen6meno multidimensional que no puede, ni debe medirse uinica y exclusivamente con base en el consumo per capita de los hogares y su relaci6n con las lineas de pobreza. Por este motivo recomendamos que la informaci6n del mapa sea complementada con otros indicadores disponibles a nivel municipal. La combinaci6n de estos indicadores con el mapa hara posible una mejor asignaci6n de recursos buscando maximizar los impactos de los diversos programas sobre la reducci6n de la pobreza. 30. El mapa describe la pobreza en los municipios antes del huracan Mitch. Por tanto, tambien advertimos que las decisiones se hagan tomando en cuenta que aun cuando el perfil de pobreza por regiones no cambi6 significativamente9 cuando el huracan Mitch azot6 el pais, pueden existir bolsones de pobreza en comarcas y barrios especificos que no son reflejados en los indicadores estimados para el nivel municipal. 9 El Apendice 5 incluye una tabla comparativa de la extensi6n de la pobreza general y extrema pre- y post- Mitch por regiones. 18 Annex 19, Page 19 VIII. BIBILOGRAFiA Alderman, Harold, Miriam Babita, Jean Lanjouw, Peter Lanjouw, Nthabiseng Makhatha, Amina Mohamed, Berk Ozier, and Olivia Qaba (2000), Is Census Income an Adequate Measure of Welfare? Combining Census and Survey Data to Construct a Poverty Map of South Africa. Statistics South Africa Working Paper. En proceso de publicaci6n. Elbers, Chris, Jean Lanjouw, Peter Lanjouw (2000), Welfare in Towns and Villages. Micro-Level Estimation of Poverty and Inequality. Tinbergen Institute Working Paper. En proceso de publicaci6n. Gobiemo de Nicaragua (2000), Estrategia Reforzada de Reducci6n de la Pobreza. Hentschel J.. Lanjouw J., Lanjouw P & Poggi J (Enero 2000), Combining Census and Survey Data to Trace the Spatial Dimension of Poverty: A Case Study of Ecuador. The World Bank Economic Review, 14 (1). Instituto Nacional de Estadisticas y Censos, Censo Nacional de Poblacion y Vivienda de Nicaragua - 1995 Instituto Nacional de Estadisticas y Censos, Encuesta de Nacional de Hogares sobre Medici6n de Nivel de Vida - 1998 Instituto Nacional de Estadistica y Censos (Marzo, 2000), Encuesta Nacional de Hogares sobre Medici6n de Nivel de Vida - Informe General. Proyecto MECOVI. World Bank (2000), Nicaragua Poverty Assessment: Challenges and Opportunities for Poverty Reduction. Report No. 20488-NI 19 Annex 19, Page 20 IX. APENDICES APENDICE A19.1 MEDIDAS DE POBREZA La extensi6n y la brecha de la pobreza son dos medidas que nos ayudan a entender las caracteristicas de la pobreza en Nicaragua. Este documento utiliza las siguientes f6rmulas: EXTENSI6N DE LA POBREZA Es el nuimero de pobres como proporci6n de la poblaci6n total n Donde: q = numero de pobres n = tamafio de la poblaci6n BRECHA DE LA POBREZA B = N z )Y n j=1 (Z Donde: Z = linea de pobreza Y1 = Consumo per capita. Ya que se suman solamente hasta el hogar "q", se incluyen uinicamente los hogares pobres. VALOR TOTAL DE LA BRECHA Valor minimo necesario para hacer llevar el consumo de los pobres a la linea de pobreza v = (Z- = B*n*Z i=2 20 Annex 19, Page 21 PROFUNDIDAD PROMEDIO DE LA POBREZA COMO PROPORCION DE LA LiNEA DE POBREZA Valor medio de la brecha de los pobres expresado como proporci6n de la linea de pobreza. P= I ( Z-Y' = B* E-' = B z ym q ,1Z) q Z Donde: yM = Consumo promedio de los pobres PROPORCION DE LA BRECHA DE LA POBREZA EXTREMA NACIONAL El valor total de la brecha puede utilizarse para desarrollar un indice compuesto de las tasas de pobreza y la poblaci6n basado en la proporci6n de la brecha de la pobreza extrema. Si consideramos el problema de asignar una cantidad especifica de recursos disponibles, RD,r)iAL, para reducir la pobreza. Las estimaciones de la Brecha de Pobreza, BD, en cada municipioj se pueden usar. Dado este valor, podemos calcular facilmente el total de recursos necesarios en principio para llevar a cada individuo extremadamente pobre a la linea de pobreza extrema (en terminos monetarios). Entonces, para cada municipioj, tenemos que el total de recursos necesarios es, RN,: RN; =B. n z, done n es la poblaci6n total del municipioj, y z es la linea de pobreza extrema. Entonces, obviamente, el total de recursos necesarios a nivel nacional es la suma de los recursos necesarios de cada municipio (6 la Brecha de la Pobreza Extrema Total Nacional es la suma de las Brechas de la Pobreza Extrema de cada municipio): k RNTOTAL = RN, En la mayoria de los casos, los recursos disponibles, RD, no seran los mismos que los recursos necesarios, RN. Por tanto, para asignar los recursos disponibles se toma como base la contribuci6n de cada municipio a los fondos necesarios para cerrar la Brecha de la Pobreza Extrema. Por tanto, los recursos asignados a cada municipioj serian la Proporci6n de la Brecha de la Pobreza Extrema de cada municipio en la Brecha de la Pobreza Extrema Total a nivel Nacional: RN. RNI RDJ = RN * RDTOTAL donde ' = Proporci6n de la brecha NTOTAL RNTOTAL Siempre y cuando los fondos disponibles se asignen al interior de cada municipio de la misma manera (proporcionalmente a la brecha del consumo de cada individuo con respecto a la linea de pobreza extrema, 6 brecha de la pobreza extrema municipal), estas asignaciones pueden contribuir a reducir la brecha de pobreza extrema de cada individuo en la misma proporci6n. 21 Annex 19, Page 22 APENDICE A19.2 DESCRIPCION DE VARIABLES UTILIZADAS COMUNES AL CENSO95 Y LA EMNV98 1. BAGUA (Buena agua). Se considera como buena agua, tanto en el area urbana como rural, en caso que los hogares obtengan el liquido de una tuberia dentro de la vivienda. 2. BLUZ (Buena luz). Si los hogares tienen energia electrica. 3. BPARED (Buena pared). Es buena pared si esta construida de bloque de cemento o concreto, piedra cantera, y Humina plycem o nicalit. 4. BPISO (Buen piso). Se considera buen piso si es de ladrillo de barro, de ladrillo de cemento, mosaico o terrazo. 5. BTECHO (Buen techo). Solamente si es de zinc y de Iaminas plycem o nicalit. 6. BVIVIEN (Buena vivienda). Si el tipo de vivienda es una quinta, es buena. 7. COCINA (Cocina en cuarto exclusivo). Si el hogar cocina en un cuarto dedicado para cocinar. 8. CONAGNE (Inodoro con aguas negras). 9. HACIN (Indice de hacinamiento). Nuirmero de personas por cuarto de dormir en la vivienda. 10. JEFCAS (Jefes casados). Se tom6 en consideraci6n el estado civil o conyugal actual del jefe. En este caso, s6lo si esta casado. 11. JEFIND (Jefes que hablan lengua indigena), si el jefe habla miskito y/o sumo. 12. JEFUNI (Jefes unidos), Se tom6 en consideraci6n el estado civil o conyugal actual del jefe. En este caso, s6lo si esta unido o juntado. 13. JMUJER (Jefa de hogar mujer), incluyendo s6lo al jefe de la vivienda que sea mujer. 14. LETRINA (Existencia de letrina). En caso de que el hogar cuente con el servicio higienico de letrina. 15. M1865 (Promedio de educaci6n de personas entre 18 y 65 afios). 16. MAXJECY (Grado maximo de educaci6n del jefe o c6nyuge). 17. MPARED (Mala pared). Es mala pared si esta construida de bambui, caha o palma; ripio o desechos, y otro. 18. MPISO (Mal piso). Se considera mal piso si es de tierra u otro. 19. MRAGUA (Mala agua rural). En el area rural, es mala agua si la obtienen de un rio, manantial o quebrada. 20. MTECHO (Mal techo). Solamente si es de paja o similares y de ripio o desechos. 21. MUAGUA (Mala agua urbano). En el area urbana, es mala agua si la obtienen de un puesto publico o de un rio, manantial o quebrada. 22. MVIVIEN (Mala vivienda). Si es rancho o choza u vivienda improvisada, es mala. 23. PALFAB (Porcentaje de personas alfabetizadas), tomando en cuenta a las personas mayores e iguales a doce afios y que sepan leer y escribir. 24. PASIPRI (Porcentaje de personas que asisten a primaria), tomando en cuenta a las personas entre 6 y 12 afios que se matricularon en algun grado en ese anio y que tuvieran alguin nivel de estudio. 25. PASISEC (Porcentaje de personas que asisten a secundaria), tomando en cuenta a las personas entre 13 y 18 afios que se matricularon en algun grado en ese afio y que tenga alguin nivel de estudio. 22 Annex 19, Page 23 26. PINDIG (Porcentaje de personas que hablan lengua indigena), si la lengua materna que hablan desde la nifiez es miskito y/o sumo, se consideran personas que hablan lengua indigena. 27. PMAI65 (Porcentaje de personas mayores e igual a 65 afios). 28. PMEI12 (Porcentaje de personas menores e igual a 12 afios). 29. PRICOMP (Porcentaje de viviendas con personas de primaria completa). Se consider6 a las personas que estuvieran en el rango de 18 a 65 afios y que tuvieran aprobados entre 6 y 10 grados. Se asume que las personas con 6 grados aprobados tienen la primaria completa y si las personas tienen mas de 6 y menos de 10 grados, incluye los que estan cursando la secundaria, pero no han egresado aun. 30. PTRSP (Porcentaje de personas que trabaj6 la semana pasada). Las personas mayores e iguales a 10 afios que trabajaron la semana pasada. 31. PTRSP8 (Porcentaje de personas que trabaj6 la semana pasada y ocho horas o mas). Las personas mayores e iguales a 10 afios que trabajaron la semana pasada y 8 horas a mas en la misma semana. 32. PVIVAGR (Porcentaje de viviendas con alguien de ocupaci6n principal = 6 segan CIUO). 33. PVIVNCA (Porcentaje de viviendas con alguien de ocupaci6n principal = 9 segun CIUO). 34. PVIVOCP (Porcentaje de viviendas que tienen a alguien con alguna ocupaci6n principal). Se tom6 a personas mayores e igual a 10 afios y con ocupaciorLes principales que le dedicaron mas tiempo en el trabajo durante la semana pasada. Esias ocupaciones segun la CIUO son: Miembros de la Administraci6n Publica y Empresas, Profesionales Cientificos e Intelectuales, y a Tecnicos y Profesionales del nivel medio. 35. SECCOMP (Porcentaje de viviendas con personas de secundaria completa). Se consider6 a las personas que estuvieran en el rango de 18 a 65 ahos y que tuvieran aprobados entre 11 y 16 grados. Se asume que las personas con 11 grados aprobados tienen la secundaria completa y si las personas tienen mas de 11 y menos de 16 grados, incluye los que estan cursando la universidad, pero no han egresado auin. 36. SINAGNE (Inodoro sin aguas negras). 37. THIJNVI (Hijos nacidos vivos). 38. TNEGOC (Existencia de negocios en la vivienda). 39. TPERV (Total de personas en la vivienda). 40. UNICOMP (Porcentaje de viviendas con personas de universidad completa). Se consider6 a las personas que estuvieran en el rango de 18 a 65 a-nos y que tuvieran aprobados s6lo 16 grados o mas. 41. VALQUIL (Vivienda alquilada). Solamente si es alquilada. 42. VPROPIA (Vivienda propia). Propia con escritura. 23 Annex 19, Page 24 APENDICE A19.3 METODOLOGIA Y RESULTADOS DE PRIMERA ETAPA 1. La metodologia utilizada se puede diferenciar claramente en dos partes: la estimaci6n de las ecuaciones de la primera etapa utilizando los datos de las encuestas de hogares EMNV98'0 y la estimaci6n de las medidas de pobreza en todo el pais utilizando el Censo95. PRIMERA ETAPA 2. La primera etapa utiliza exclusivamente informaci6n de la EMNV98 y consiste en una regresi6n multivariada entre el logaritmo natural del consumo per capita por vivienda (In yi) y las diferentes caracteristicas del hogar cuyas preguntas se encontraban tanto en la EMNV98 como en el Censo95 (vector Xi): In y, = Xi, + e, (Ecuaci6n 1) 3. Esta estimaci6n se hizo individualmente para cada una de las siete regiones significativas de la EMNV9811 y fueron evaluadas para detectar problemas de las violaciones de los supuestos de normalidad, homoscedasticidad, y efectos fijos'2. 4. Para evaluar el supuesto de normalidad se utiliz6 la prueba de Kilmogorov-Smirnov (con la correcci6n de significancia Lilliefors)'3. Para determinar problemas de heteroscedasticidad se realiz6 una prueba F de la regresi6n multivariada entre el cuadrado los residuos de la regresi6n (variable dependiente: (£j )2 estimado ) y el mismo vector de variables utilizado en la regresi6n (variables independientes: X ;). Finalmente para detectar la presencia de problemas por efectos fijos se realiz6 una prueba F de la regresi6n multivariada entre el cuadrado los residuos de la regresi6n (variable dependiente: (s1 )2 estimado ) y una variable dic6toma para cada uno de los municipios. 5. Para eliminar el problema de falta de normalidad fue necesario eliminar algunos de los hogares de la muestra. Debido de que el nuimero de hogares eliminados fue muy bajo, no esperamos que este proceso altere los resultados. 6. Se identific6 una variable responsable del problema de heteroscedasticidad en dos de las siete regiones la cual no se incluy6 para las regiones afectadas; las otras cinco regiones no presentaron problemas de heteroscedasticidad. 10 Debido a que el censo recoge la informacion en base a viviendas, la EMNV 98 tuvo que ser transformada a nivel de viviendas. Despues de esta transformaci6n, quedaron 3,874 observaciones. "Ver pie de pagina 1. 12 Estos supuestos se encuentran incorporados en el programa en SAS que nos calcula los estimados de pobreza. Futuras versiones de este programa se estan desarrollando donde al menos dos de las restricciones impuestas no seran necesarias. 13 Con la excepci6n de la regi6n Central rural donde se utilizaron las pruebas de Kurtosis, Skewness, Chi- cuadrado y el test de Shapiro-Wilk-Francia. 24 Annex 19, Page 25 7. Problemas de efectos fijos fueron detectados en dos regiones, donde fue necesario hacer pequefias modificaciones a la especificaci6n de los modelos 14. Los modelos pasaron las tres pruebas a los que se les someti6 y los resultados se presentan en el Cuadro A3.1. Cuadro A19.3.1 Resultados de la evaluacion de las ecuaciones de la primera etapa REGIONES N R' ajustado Prueba de Heterosced Efectos eliminados Normalidad asticidad fijos 1. Managua 5 0.629 > 0.20 0.119 0.388 2. Pacifico urbano 8 0.586 > 0.20 0.060 0.274 3. Pacifico rural 2 0.486 > 0.20 0.106 0.057 4.Central urbano 3 0.610 > 0.20 0.957 0.074 5. Central rural 15 0.513 > 0.05 0.465 0.662 6. Atlantico urbano 2 0.620 > 0.20 0.810 0.406 7. Atlantico rural 3 0.449 > 0.20 0.585 0.132 Nota: Probabilidades mayores de 5% (0.05) se consideran satisfactorias 8. Los resultados de las regresiones de la primera etapa, los datos especificos de las pruebas F, R cuadrado ajustado, valores , estimados y su correspondiente error estandar se encuentran en el Cuadro A3.2. Es importante hacer notar que estos modelos no pretenden ser explicativos y los valores individuales estimados no deben de ser interpretados como una medida que relacione las caracteristicas del hogar con el nivel cle consumo. SEGUNDA ETAPA 9. La segunda etapa utiliza las caracteristicas de los hogares en el Censo95 y estima la probabilidad de ser pobre de cada vivienda (Pi* ) a traves de la siguiente ecuaci6n: pi f [P IXi,,,x6] = L l C 8 j (Ecuaci6n 2) Donde: cD = Distribuci6n normal acumulada In Z = logaritmo natural del valor de la linea de pobreza (general o extrema). A = valores estimados en la regresi6n de la primera etapa. a," = desviaci6n estandar estimada en la regresi6n de la primera etapa. C, = valores del censo de las caracteristicas de cada hogar "i". 10. Los otros parametros que se calculan son la brecha y profundidad de la pobreza. El calculo de la proporci6n (y = 0), la brecha (y = 1) y la profundidad (y = 2) de la pobreza se puede expresar con la siguiente ecuaci6n: 14 En la regi6n del Central Rural fue la unica excepci6n donde fue necesario una exploraci6n mas a fondo para obtener un modelo que satisficiera todas las pruebas hechas. 25 Annex 19, Page 26 p = n=[Z ] (Ecuaci6n 3) Donde: Hq = el hogar pobre con mayor consumo. Z = la linea de pobreza. yi = consumo de cada hogar. n = el nuimero de observaciones totales (pobre y no pobres) 1 1. Para formamos una idea de cuanto habian cambiado las condiciones entre 1995 y 1998, se tabularon los promedios de las variables utilizadas en la primera regresi6n comparandose los valores provenientes del Censo95 y de la EMNV98; sus valores y el correspondiente error estandar se presentan en el Apendice 4 Cuadro A4. 1. 12. Los resultados del Mapa de Pobreza Extrema incluyen la extensi6n de la pobreza general y extrema a nivel de regiones, departamentos y de municipios. 13. Por otra parte, y respondiendo a las necesidades del Gobierno de Nicaragua, tambien se calcularon las brechas de pobreza general y extrema. Estos calculos se hicieron a todos los niveles de agregaci6n pero sin distinguir entre areas urbanas o rurales.15 15 La desviaci6n estandar de la nueva medida es: V(Var X) + (Var Y) + 2COVx, donde VAR X es el cuadrado de la desviaci6n estAndar de la brecha urbana y VAR Y es el cuadrado de la desviaci6n estandar de la brecha rural que se promediaron. Ya que las estimaciones son generadas por diferentes modelos, la covarianza entre ambas medidas es cero. 26 Annex 19, Page 27 Cuadro Al 9.3.2 Resultados de las Regresiones de Primera Etapa - Parametros Beta de regresiones iniciales con la EMNV981 Regie n Sienificativa Variables independientes 1 2 3 4 5 6 7 _ M BPARED 0.1823 0.0397 0.2117 0.0878 -0.0790 0.0347 0.2438 A Buena Pared (0.0483 (0.0386) (0 0464) (0.0475) (0.0731) (0.0709) (0. 271 ) T MPARED . 0.0406 0.0517 . 0.2246 -0.0730 E Mal pared (0.0907) (0 0674) (0.2064) (0.1164) R BPISO 0.0880 0.0523 0.0623 0.0856 0.1148 0.1473 0.4362 V ] Buen piso (0.0551 (0.0533) (0.0602) (0.0613) (0.07361 (0.0968) (0.29711 I A MPISO . -0.1310. -0.1160 0.0000 -0.2500 -0.0481 V L Mal piso (0.0557) (0.0578) (0.0000) (0.0822) (0.07473 I E BTECHO 0.1337 0.0823 0.0090 0.1081 0.0943 0.5054 0.0621 E S Buen techo (0.0897 (0.0383) (0.0393 (0.0501) (0.0336) (0.1960) (0.12741 N MTECHO . 0.1672 -0.1090 . . 0.2694 -0.0421 D Mal techo (0.1295) (0.0899) . (0.3358) (0.1314 A BVIVIEN -0.7130 2.6830 0.5074 . 0.9178 0.6704 0.1566 Buenavivienda (0.5025) (0.4476) (0.3138 . (0.1919) (0.4737) (0.5570C T MVIVIEN . -0.0520 -0.0860. . -0.0830 -0.1940 I Mala vivienda (0.1215) (0.0917) (0.2865) (0. 1175_ p COCINA 0.1364. 0.0221 -0.0040 0.1870 -0.0750 0.1695 0 Cocina dormitorio y lefna (0.0484) . (0.0406) (0.0515) (0.0400 (0.0690) (0.0695A VPROPIA -0.0140 0.0288 0.0963 -0.0410 0.1157 0.1814 0.260" Vivienda propia (0.0482) (0.0384) (0.0396 (0.0499) (0.0327) (0.0618) (0.0586) VALQUIL 0.4499 0.0346 -0.1520 0.2293 0.9311 0.0687 -0.395/ Vivienda alquilada (0.1116 (0.0637) (0.2313 (0.0896) (0.1410) (0.1191) (0.3308) BAGUA 0.0563 0.1374 0.1219 0.1374 0.2324 0.0354 0.2201 Buena agua (0.0526) (0.0411) (0.1064) (0.0532) (0R0876) (0.2254 S MRAGUA . . . . . . 0.028 E A Mala agua rural (0.060C' R G MUAGUA . -0.0850. . . -0.0900 . V U Mala agua urbana (0.0647) . (0.0763) I A LETRINA -0.0100 0.0039 -0.0200 0.1092 0.0080 0.0205 0.0828 C S Tiene letrina (0.1210 (0.0869) (0.0465) (0.0863) (0.0363) (0.0909) (0.065 1_ CONAGNE -0.0030 0.1523 1.5020 0.2802 0.4147 -0.3010. O Con aguas negras (0.1297) (0.0997) (0.4189) (0.1000) (0.4216) (0.3651)1 _ s SINAGNE -0.1940 0.0921 0.4668 0.2871 0.4395 0.4402. Sin aguas negras (0.1483 (0.1030) (0.1732) (0.1163) (0. 2501 (0.1566) LUZ BLUZ -0.0250 0.1970 0.1254 0.2306 0.1169 0.2223 0.0924 _ u Z Buena luz (0.1114 (0.0631) (0.0413) (0.0648) (0.0416 (0.0797) (0.1208) TP PERV . -0.0610 -0.0460 -0.0660 -0.0690 -0.0550 -0.0601 E Total personas (0.0096) (0.0104) (0.0129) (0.0091) (0.0120) (0.0163 R PME112 -0.5520 -0.4070 -0.3980 -0.2400 -0.4320 -0.3830 -0.4100 S % menores de 12 aflos (0.1340 (0.1022) (0.1132 (0.1271) (0.0949) (0.2105) (0. 1661' O PMA165 -0.0680 0.0386 0.0877 -0.1130 -0.1920 0.2805 -0.7640l D N % mayores de 65 affos (0.1475 (0.1122) (0.1150) (0.1546) (0.1204) (0.2574) (0.228&) E A M1865 . . 0.0174. . 0.0403. M _ Personas entre 18 v 65 (0.0175) . (0.0215) O JMUJER 0.025 -0.0300 -0.0310 -0.0720 -0.0350 0.1898 -0.2300 G J lefe muier (_O0712) (0.0497) (0.0642) (0.0698) (0.0608) (0.0860) (0. 1210O R E JEFUNI -0.0030 -0.1080 -0.0030 -0.0940 -0.0110 0.1792 -0.1901 A F Jefe Unido (0.0788 (0.0572) (0.0657) (0.0762) (0.0605) (0 0960) (0.1179) F E JEFIND -0.530 0.0634 . 0.0330 -6.450.) I Jefe indigena (0.6341) (0.5418). (0.2984) (1.774(J) A JEFCAS -0.0190 -0.0230 -0.0030 0.0196 0.0297 0.1584 -0.2080 Jefe casado (0.0775) (0.0569) (0.0690) (0.0764) (0.0621) (0.0997) (0.1152) O PINDIG 0.3954 -1.3800 0.0000 0.0000 -1.3100 0.1633 6.318f T % de indigenas en hogar (0.8201 (2.3470) (0.0000) (0.0000) (2.3120 (0.3061) (1.7880) R THIJNVI -0.0250 -0.0020 -0.0230 -0.0060 -0.0080 0.0085 0.0000 O Hiios nacidos vivos (0.0099) (0.0077) (0.0063) (0.0087) (0.0059) (0.0119) (0 008t S HACIN -0.0820 -0.0180 -0.0150 -0.0610 -0.027 -0.0460 -0.024) _ Hacinamiento (0.0119 (0.0102) (0.0106) (0.0143) (0.00961 (0.0169) (0.016_1 27 Annex 19, Page 28 Cuadro Al 9.3.2 Resultados de las Regresiones de Primera Etapa (cont.) Regi n Significativa Variables independientes T 2 3 4 5 6 7 A PALFAB 0.0804 0.1546 -0.0800 -0.0150 0.1727 0.1086 0.0115 S % de alfabetos (0.1206) (0.0818) (0.0772) (0.1064) (0.0649) (0.1304) (0.1118) E S PASIPRI -0.0230 -0.0010 0.1314 0.0162 -0.0060 0.0064 0.2957 D T % en primaria (0.0552) (0.0417) (0.0483) (0.0539) (0.0457) (0.0792) (0.0982) l E PASISEC 0.0462 0.0072 0.1319 0.0587 0.0943 -0.0880 -0.1200 C % en secundaria (0.0541) (0.0428) (0.0534) (0.0551) (0.0527) (0.0809) (0.1011) A- C C PRICOMP 0.1252 0.0727 . 0.1152 0.3272 . 0.0505 I M % primaria completa (0.0830) (0.0610) . (0.0753) (0.0691). (0.1394) 0 P SECCOMP 0.0838 0.1497. 0.3095 0.3395. -0.8800 N L % secundaria completa (0.1250) (0.0843) . (0.1104) (0.1440). (0.3400) E UNICOMP 0.8670 0.7952 . 0.5952 2.9170 T % universidad completa (0.1697) (0.1426) . (0.2010) (0.7520) . MAXJECY 0.0236 0.0084 0.0091 0.0155 -0.0020 0.0129 0.0348 Max. educ. Jefe o comp. (0.0087) (0.0066) (0.0087) (0.0085) (0.0091) (0.0115) (0.0167) PVIVAGR 0.4339 0.0440. . 0.1575 -0.2540 -0.2270 L T % en agricola y similar (0.3169) (0.1877) . . (0.1083) (0.2415) (0.1908) A P PVIVNCA -0.3040 -0.2120 . . 0.0945 -0.1330 -0.5500 B O % trabajos no calificados (0.1174) (0.0842) . . (0.0918) (0.1546) (0.1648) O PVIVOCP 0.4123 0.1858 0.2426 0.1872 0.21 03 0.0298 0.0774 R % profes. tecnico, ejecutivo (0.1684) (0.1178) (0.1707) (0.1534) (0.1790) (0.1975) (0.3204) A CAN PTRSP 0.0078 0.8461 -0.0790 -0.5940 0.5521 -1.5300 0.9169 L TI > 8 horas ultima semana (0.5783) (0.2804) (0.3129) (0.4469) (0.6001) (1.9390) (0.8543) DAD PTRSP8 0.1343 -0.6470 0.3319 0.6364 -0.4670 1.8610 -0.6160 > 8 horas y le pagaron (0.5812) (0.2780) (0.3104) (0.4525) (0.5976) (1.9380) (0.8453) O TNEGOC 0.0000 0.1200 0.0217 0.0974 0.1654 0.2044 -0.0370 TRO Negocio en la vivienda (0.0561) (0.0404) (0.0574) (0.0487) (0.0591) (0.0792) (0.1179) INTERCEP 8.8680 8.5790 8.4420 8.6490 8.21 30 7.9200 8.5460 (0.1976) (0.1316) (0.0966) (0.1660) (0.0935) (0.2709) (0.2134) Significancia prueba F 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Pruebas y R' ajustado 62.9 58.6 48.6 61.0 51.3 62.0 44.9 Supuestos Casos Eliminados 5 8 2 3 15 2 3 Regresi6n Prueba de Normalidad > 0.20 > 0.20 > 0.20 > 0.20 > 0.05 > 0.20 > 0.02 Heteroscedasticidad 0.119 0.060 0.106 0.957 0.465 0.810 0.585 _Efectos regionales (fijos) 0.388 0.274 0.057 0.074 0.662 0.406 0.132 Errores est6ndar en parentesis. El c6lculo correcto del error estandar es una parte fundamental de cualquier Mapa de Pobreza Extrema ya que sin esta medida es imposible determinar si las diferencias observadas en las estimaciones son significativas o no. Celdas sin informaci6n corresponden a variables que no fueron utilizadas para la correspondiente regi6n significativa. 28 Annex 19, Page 29 APENDICE A19.4 VALIDACION DEL MAPA DE POBREZA 1. Este Mapa de Pobreza tambien nos permite realizar pruebas estadisticas para evaluar la relevancia y validez de las estimaciones espaciales de pobreza. La primera prueba examina el valor promedio de todas las variables utilizadas comunes al Censo95 y la EMNV98 para averiguar si ocurri6 un cambio significativo entre 1995 y 1998 (Cuadro A4. 1) Se encontr6 que en las siete regiones algunas variables han cambiado y otras no. pero las diferencias no son importantes. 29 Anne~ age30 Cuadro A19.4.1 Valores promedios de variables en el CENS095 y EMNV981,2 usadas en regresi6n de primera etapa Nacional Managua Pacifico Central Atlantico Variable Urbano Rural Urbano Rural Urbano Rural 95 98 95 1 98 95 1 98 95 1 98 95 98 95 99 1 98 95 98 N 751632 3,874 202,358| 491 134,2131 785 103,6351 591 84,042 561 141,758 774 28,519 308 57,107 364 BAGUA .322 .229 .524 .490 536 .413 .096 .039 .471 .353 .046 .046 .245 .189 .023 .019 (.0005) (.0067) (.0011) (.0226) (.0014) (.0176) (.00009) (.0080) (.0017) (.0202) (.0006) (.0075) (.0025) (.0224) (.0006) (.0071) BLUZ .617 .626 .927 .957 .881 .902 .422 .478 .794 .834 .178 .319 .657 .691 .058 .106 (.0006) (.0078) (.0006) (.0092) (.0009) (.0106) (.0015) (.0206) (.0014) (.0157) (.0010) (.0168) (.0028) (.0264) (.0010) (.0162) CONAGNE .172 .120 .428 .399 .183 .187 .007 .002 .189 .213 .004 .002 .035 .006 .004 .000 (.0004) (.0052) (.0011) (.0221) (.0011) (.0139) (.0003) (.0019) (.0014) (.0173) (.0002) (.0016) (.0011) (.0045) (.0003) (.0000) HACIN 4.059 4.290 3.470 3.608 3.717 4.227 4.424 5.083 3.688 3.668 4.838 4.503 3.875 3.821 4.984 4.959 (.0029) (.0429) (.0049) (.0994) (.0064) (.1022) (.0079) (.1296) (.0078) (.0910) (.0072) (.0895) (.0149) (.1363) (.0114) (.1430) JEFCAS .393 .384 .367 .348 .383 .348 .338 .311 .412 .405 .437 .432 .405 .401 .460 .477 (.0006) (.0078) (.0011) (.0215) (.0013) (.0170) (.0015) (.0191) (.0017) (.0207) (.0013) (.0178) (.0029) (.0280) (.0021) (.0262) JEFIND .017 .033 .001 .002 .000 .001 .000 .000 .000 .000 .006 .000 .129 .130 .137 .238 (.0001) (.0029) (.0001) (.0020) (.00000) (.0013) (.0001) (.0000) (.0000) (.0000) (.0002) (.0000) (.0020) (.0192) (.0014) (.0223) JEFUNI .352 .327 .342 .304 .318 .310 .448 .439 .284 .245 .366 .332 .328 .311 .366 .340 (.0006) (.0075) (.0011) (.0208) (.0013) (.0165) (.0015) (.0204) (.0016) (.0182) (.0013) (.0169) (.0028) (.0264) (.0020) (.0249) JMUJER .269 .272 .334 .339 .336 .359 .193 .187 .347 .363 .167 .189 .305 .289 .143 .155 (.0005) (.0072) (.0010) (.0214) (.0013) (.0171) (.0012) (.0160) (.0016) (.0203) (.0010) (.0141) (.0027) (.0259) (.0015) (.0190) LETRINA .565 .635 .471 .505 .693 .643 .714 .734 .657 .633 .458 .651 .776 .804 .352 .464 (.0006) (.0077) (.0011) (.0226) (.0013) (.0171) (.0014) (.0182) (.0016) (.0204) (.0013) (.0171) (.0025) (.0227) (.0020) (.0262) M1865 2.530 2.528 3.851 3.986 3.349 3.413 1.632 1.730 2.939 3.386 1.016 1.349 2.378 2.632 .783 1.048 (.0030) (.0404) (.0064) (.1376) (.0072) (.0948) (.0053) (.0702) (.0086) (.1117) (.0034) (.0558) (.0135) (.1407) (.0047) (.0727) MAXJECY 4.945 4.791 7.083 6.802 6.126 6.066 3.526 3.500 5.610 6.136 2.487 2.962 5.123 5.499 2.197 2.637 (.0050) (.0687) (.0103) (.2097) (.0120) (.1548) (.0100) (.1368) (.0148) (.1875) (.0074) (.1193) (.0240) (.2551) (.0113) (.1637) MRAGUA .1S2 .144 .021 .025 .000 .000 .187 .211 .000 .000 .401 .312 .000 .000 .585 .499 (.0004) (.0056) (.0003) (.0070) (.0000) (.0000) (.0012) (.0168) (.0000) (.0000) (.0013) (.0167) (.0000) (.0000) (.0021) (.0262) MUAGUA .053 .060 .046 .043 .084 .091 .000 .000 .166 .146 .000 .000 .184 .190 .000 .000 (.0003) (.0038) (.0005) (.0092) (.0008) (.0103) (.0000) (.0000) (.0013) (.0149) (.0000) (.0000) (.0023) (.0224) (.0000) (.0000) PALFAB .746 .755 .890 .876 .867 .858 .687 .721 .827 .854 .532 .604 .775 .793 .455 .561 (.0004) (.0051) (.0005) (.0110) (.0007) (.0089) (.0010) (.0128) (.0009) (.0104) (.0010) (.0126) (.0018) (.0170) (.0015) (.0199) PASIPRI .389 .373 .453 .431 .464 .434 .386 .372 .456 .427 .255 .300 .428 .432 .209 .187 (.0005) (.0072) (.0011) (.0212) (.0013) (.0166) (.0014) (.0178) (.0016) (.0198) (.0010) (.0150) (.0027) (.0251) (.0015) (.0174) PASISEC .269 .283 .341 .380 .338 .349 .213 .241 .350 .346 .138 .180 .322 .343 .130 .155 (.0005) (.0069) (.0010) (.0210) (.0012) (.0163) (.0012) (.0161) (.0016) (.0193) (.0008) (.0130) (.0026) (.0255) (.0013) (.0175) PINDIG .014 .034 .001 .003 .000 .000 .000 .000 .000 .001 .005 .000 .108 .136 .111 .237 (.0001) (.0029) (.0000) (.0015) (.0000) (.0003) (.0000) (.0000) (.0000) (.0004) (.0002) (.0002) (.0016) (.0191) (.0012) (.0222) 30 Anne age 31 PNAI65 .052 .059 .049 .065 .065 .059 .062 .076 .050 .058 .046 .052 .047 .044 .036 .049 (.0002) (.0026) (.0003) (.0083) (.0005) (.0058) (.0006) (.0077) (.OOOS) (.0064) (.0004) (.0055) (.0009) (.0073) (.0005) (.0079) PMEI12 .353 .338 .312 .282 .319 .310 .374 .338 .336 .311 .398 .364 .379 .376 .437 .429 (.0003) (.0036) (.0005) (.0101) (.0006) (.0077) (.0007) (.0094) (.0007) (.0093) (.0006) (.0079) (.0013) (.0123) (.0009) (.0115) PRICONP .306 .295 .426 .425 .378 .359 .236 .235 .371 .395 .141 .156 .348 .363 .126 .158 (.0004) (.0056) (.0008) (.0166) (.0010) (.0126) (.0010) (.0128) (.0013) (.0156) (.0007) (.0098) (.0022) (.0202) (.0011) (.0153) PTRSP .422 .469 .426 .382 .403 .466 .404 .462 .411 .491 .452 .505 .376 .450 .453 .508 (.0003) (.0044) (.0006) (.0129) (.0008) (.0103) (.0008) (.0112) (.0010) (.0114) (.0006) (.0091) (.0017) (.0165) (.0011) (.0135) PTRSP8 .419 .463 .423 .376 .398 .456 .400 .455 .408 .486 .451 .502 .373 .449 .450 .505 (.0003) (.0044) (.0006) (.0128) (.0008) (.0103) (.0008) (.0113) (.0010) (.0112) (.0006) (.0090) (.0017) (.0165) (.0011) (.0134) PVIVAGR .131 .081 .022 .014 .034 .021 .200 .096 .051 .038 .294 .151 .086 .058 .355 .212 (.0003) (.0027) (.0002) (.0032) (.0003) (.0033) (.000 7) (.0076) (.0005) (.0047) (.0007) (.0074) (.0010) (.0076) (.0010) (.0107) PVIVNCA .120 .214 .130 .147 .117 .184 .136 .280 .113 .187 .125 .284 .089 .129 .079 .228 (.0002) (.0042) (.0005) (.0101) (.0006) (.0089) (.0007) (.0111) (.0007) (.0112) (.0006) (.0098) (.0011) (.0135) (.0007) (.0135) PVIVOCP .053 .057 .087 .083 .069 .076 .020 .031 .066 .082 .015 .031 .054 .071 .023 .033 (.0002) (.0025) (.0004) (.0079) (.0005) (.0062) (.0003) (.0048) (.0006) (.0076) (.0002) (.0040) (.0009) (.0104) (.0004) (.0071) SECCOMP .097 .109 .157 .180 .160 .170 .035 .047 .131 .165 .013 .041 .097 .131 .007 .024 (.0003) (.0039) (.0006) (.0126) (.0008) (.0101) (.0004) (.0062) (.0009) (.0119) (.0002) (.0060) (.0014) (.0145) (.0003) (.0069) SINAGNE .039 .054 .038 .062 .084 .122 .017 .016 .057 .088 .007 .005 .073 .068 .009 .000 (.0002) (.0036) (.0004) (.0109) (.0008) (.0117) (.0004) (.0052) (.0008) (.0119) (.0002) (.0025) (.0015) (.0144) (.0004) (.0000) THIJNVI 4.624 2.887 3.900 2.408 4.414 2.351 5.330 3.170 4.466 2.465 5.287 3.242 4.553 3.204 5.021 3.857 (.0052) (.0505) (.0086) (.1167) (.0132) (.0932) (.0154) (.1427) (.0142) (.1137) (.0120) (.1237) (.0275) (.1792) (.0225) (.2000) TNEGOC .138 .186 .141 .209. .190 .270 .085 .134 .202 .293 .088 .099 .158 .195 .118 .077 (.0004) (.0063) (.0008) (.0184) (.0011) (.0159) (.0009) (.0140) (.0014) (.0192) (.0008) (.0107) (.0022) (.0226) (.0014) (.0140) TPERV 5.782 5.736 5.394 5.269 5.624 5.583 5.936 6.114 5.575 5.299 6.220 5.846 5.923 6.083 6.396 6.228 (.0035) (.0474) (.0063) (.1203) (.0083) (.1084) (.0095) (.1375) (.0099) (.1018) (.0083) (.1044) (.0178) (.1781) (.0126) (.1540) UNICOMP .029 .025 .065 .059 .041 .043 .005 .008 .027 .041 .002 .001 .014 .015 .001 .000 (.0002) (.0020) (.0004) (.0088) (.0004) (.0057) (.0002) (.0029) (.0004) (.0066) (.0001) (.0009) (.0006) (.0053) (.0001) (.0000) Error estandar de los promedios en par6ntesis 2 No se hicieron los programas automAticos para determinar el valor del error estandar combinado. La f6rmula es EE(c) = Raiz cuadrada[(EE(1)*EE(1)) + (EE(2)*EE(2)) ] + 2 COV [ EE(1), EE(2)], debido a que los estimados que se est4n combinando vienen de dos diferentes ecuaciones la covarianza es 0 dejdndonos con la siguiente f6rmula EE(c) = Raiz cuadrada[(EE(1)*EE(1)) + (EE(2)*EE(2)) ]. Donde EE(c) = Error est4ndar combinado; EE(1) = Error estandar del primer estimado; y, EE(2) = Error est4ndar del segundo estimado. 31 Annex 19, Page 32 2. La segunda prueba compara los estimados de pobreza en cada una de las siete regiones de Nicaragua de los resultados obtenidos con la EMNV98 y la EMNV93 (Cuadro A4.2). Se encontr6 que los estimados del Mapa de Pobreza Extrema son muy similares a la EMNV98.16,17 CUADRO A19.4.2 ESTIMADOS REGIONALES DE POBREZA EN 1993, 1995 Y 1998 Mapa EMNV98 EMNV93 Reg iones Extensi6n desviaci6n Extensi6n e.e.' Extensi6n e.e. de la estandar de la de la Pobreza Pobreza Pobreza (estimado) (media) (media) Managua 20.1 0.33i 18.5 1.71 29.9 1.61 Pacifico Urbano 39.5 0.221 39.6 1.69 28.1 2.05 Pacifico Rural 70.8 0.35 67.1 1.89 70.7 2.62 Central Urbano 45.9 0.23 39.4 2.03 49.2 2.28 Central Rural 79.5 0.32 74.0 1.55 84.7 1.32 ~Atlantico Urbano 48.2 0.30 44.4 2.84 35.5 3.311 Atlantico Rural 78.7 0.49 79.3 2.11 83.6 2.55 Error estandar de la Media 16 Solamente en dos regiones (Central Urbano y Central Rural), se acepta la hip6tesis nula de que nuestros estimados son iguales a los de la EMNV98 con una probabilidad del 10%, pero la hip6tesis se rechaza con un criterio mas estricto (probabilidad del 5%). Esta diferencia minima es esperada debido al cambio de pobreza demostrado en comparaciones independientes entre la EMNV93 y la EMNV98 y al cambio de los valores promedio en algunas de las variables explicativas usadas en el Censo de 1995. 17 Dado que los mapas son estimados de pobreza en un momento dado (puntuales), el periodo en el cual el censo y la encuesta de hogares fueron recolectados es una consideraci6n importante. La relevancia de un Mapa de Extrema Pobreza disminuye cuando el periodo de uso se aleja del periodo de recolecci6n de los datos. Es importante notar que en Nicaragua el Censo95 y la EMNV98 s6lo se encuentran separados por tres afnos. 32 Annex 19, Page 33 APENDICE A19.5 POBREZA PRE Y POSMITCH Cuadro A19.5.1 Nicaragua-Pobreza Pre y Post-Mitch por regi6n, 1998 & 1999 Incidencia de y Cambio en Incidencia de y Cambio en la Pobreza Extrema la Pobreza General Extensi6n (%) Extensi6n (%) Regi6n 1998 1999 Cambio 1998 1999 Cambio Nacional 17.3 17.3 -0.1 ns 47.9 47.9 0.1 ns Urbano 7.6 7.5 -0.1 ns 30.5 30.3 -0.2 ns Rural 28.9 28.9 0.0 ns 68.5 69.0 0.4 ns Managua 3.1 3.1 0.0 X 18.5 18.5 0.0 X Pacifico Urbano 9.8 9.6 -0.3 ns 39.6 39.0 -0.6 ns Rural 24.1 20.6 -3.6 ** 67.1 63.1 -4.0 ** Central Urbano 12.2 12.1 -0.1 ns 39.4 39.4 0.0 ns Rural 32.7 35.7 2.9 ** 74.0 77.6 3.6 ** Atlantico Urban 17.0 17.0 0.0 X 44.4 44.4 0.0 X Rural 41.4 40.6 -0.8 ns 79.3 80.6 1.3 ns Fuente: Nicaragua EMNV 1998 y EMNV 1999. Simbologia: ns = no-significantivo a p<=10%; ** = significantivo a p<=1%; y, X Managua y Atiantico Rural no tuvieron ningun hogar entrevistado por la EMNV 1999. 33 Annex 20, Page 1 Annex 20 - Nicaragua: Qualitative Poverty and Exclusion Study (QPES)1.2 by Man, Lisbeth Gonzalez EXECUTIVE SUMMARY i. This paper discusses the main findings of a study of urban and rural poverty carried out in Nicaragua from October 1999 to Februarv 2000. The conclusions are based on a qualitative methodology for which the unit of analysis is the housenold. The goal is to broaden the analytical understanding of the data captured through living standard measurement surveys (LSMS). ii. This paper has two objectives: To contribute t design a poverty reduction strategy for Nicaragua by learning the perceptions of poverty's causes and perpetuation from the poor themselves. This document also serves as the social assessment of the Agricultural Technology and Technical Education Project. iii. This study's hypothesis is that poverty (and peoples' perception of poverty) is not solely the result of economic measures but of social, political and economic processes acting in concert. At the national level, perceptions of poverty have been influenced by radical changes in regimes over the last 50 years. At regional level, the incidence and perception of poverty varies. Differentiated participation of regions on the formation of Nicaraguan nation-state has had an incidence on nature, evolution and people's perception of poverty. Central and Pacific regions (west) have been fully integrated to national development, while Atlantic region (east) has been more isolated. Differences between west and east are wide, encompassing topographic, ethnic, and cultural factors as well as economic and political development. iv. During the 1990s, Nicaragua had two great challenges: to stabilize the economy, promote economic growth, and to pacify, democratize, and reawaken civil society. This paper explores people's perceptions of the measures taken to achieve such objectives. The study concludes show that with the new millennium, the national development agenda should include strengthening the rule of law and institutional arrangements on which economic growth and social and political development depend. Specifically, * Make governmental institutions work for the poor. Reduce corruption at all levels. e Develop safety nets for the poor, focusing on human capital, mainly health and education. * Expand job opportunities for the poor by reduce financial vulnerabilities for private sector in order to increase private sector investment. This report is the product of a team effort including the National Institute of Statistics (INEC), the MECOVI project, the Institute for Nicaraguan Studies (IEN) and the World Bank. The author is indebted to several people and institutions for making this research possible. Florencia Castro-Leal. Task Manager of the study from the World Bank, provided great overall guidance and was involved at every stage. Gonzalo Cunqueiro, Intemational Coordinator of the MECOVI project, made staff available, including Melva Bemales who contributed to the sample survey design and supervised INEC staff during the fieldwork. INEC contributed with fifteen staff members who prepared semi-structured interviews. INE was in charge of fieldwork logistics and the organization of focus groups and case studies. Dagoberto Rivera coordinated the Central region. Ana Julia Moreno coordinated the Pacific region and Miguel Castellon coordinated the Atlantic region and prepared the annex on indigenous peoples. Miguel also provided invaluable assistance in preparing the final report. Many people inside the World Bank provided valuable contributions and comments. Maria Valeria Juno is peer reviewer. Particular thanks are due to Nancy Gillespie, Ana Maria Arraigada, Norman Hicks, Helena Ribe, Katherine Bain, and Laura Rawlings, who had the patience to review and comment earlier drafts. Special thanks are due to the Nicaraguan Resident Mission and to Norman Piccioni, task manager of the Agricultural Technology and Technical Education Project, for their support. 2 The report has eight annexes in Spanish. available upon request: (I) Indigenous Communities Study; (2) Methodology; (3) Glosary of Concepts; (4) Team Members; (5) Sample of Communities. (6) Characteristics of Visited Communities; and, (7) Qualitative and quantitative Analysis: An Integral Approach. Annex 20. Page 2 Main Conclusions and Findings a. Poverty as Defined by the Poor v. People have a broad concept of poverty that goes beyond the lack of employment or income. Poverty is a state of being that encompasses not only economics but also social, cultural and moral spheres. Poverty means limited access to assets and basic services. To be poor means to be affected bv disparity of opportunities. It means to be excluded from political and economic decisions concerning their lives, lack of legal protection, and vulnerability from a variety of sources. . Moreover, it is to feel excluded and unprotected due to absence of safety nets. vi. The majority of the population continues to live in poverty. Hyperinflation has been controlled, and the opportunities to advance are higher. The opportunities to be less poor are higher, but poverty still widespread. Poverty decreased by 2.4 percent and extreme poverty by 2.1 percent. However. in 1998, 5 out of each 10 Nicaraguans were poor and two out of 10 are ex- tremely so. The extent of poverty varies across regions, but rural communities are most severely affected, particularly those in the Atlantic region, where extreme poverty increased by 9 percent in urban areas and 11.1 percent in rural areas. Sources of Political Instability vii. People feel trapped within a circle of poverty, leading them to believe that they are helpless and that poverty is inherited: "We have been poor for generations." It is a tragic sense of life. Demoralization and lack of trust in institutions are potent engines of social conflict and political instability, which jeopardize macroeconomic gains. Vulnerabilities and Risk Management viii. Poor households have an impressive ability of rebusque and to design coping strategies, but face problems to be effective because the social risk management is weak and counts with few assets. The poor development and outreach of State institutions coupled with households limited asset-base increases the impact that market fluctuations, natural disasters, legal and civil insecurity have on households. Structural adjustment policies have changed the rules of game in terms of credit, subsidies and commercial policies, which has undermined household's capacity to respond efficiently in the short-term. ix. Households coping and risk management strategies could resolve short-term problems but might increase vulnerability over the long-term. Child work is the best example. Sending children to work may increase household consumption in the immediate term, but also might increase vulnerabilitv over the long-term. This decision-making deprives younger generations of the benefits of being educated. In average, each additional year in education increases. Annex 20. Page,3 Empowerment State Institutions. Make State institutions work for the poor x. State Institutions and Government Programs. For the population, State institutions do nor: work for the poor. People question financial resources management. They also question decisio n- making process to (a) selection beneficiaries of basic services, (b) allocation of physical and human assets to miake effective basic services. Exclusion xi. As with the concept of poverty, people also have a broad and multidimensional view of exclusion. To be poor and to live under high levels of ill-being means also to be excluded. Where to draw the line between exclusion and poverty? These two concepts are ciosely related. When people talks about their socioeconomic condition both concepts - along with vulnerability -- arc used to describe their situation. Being excluded means to be affected by unemployment and economic crisis. It means to be affected by inefficient safety net and by precarious social services. People distinguish the following types of exclusion. (a) Political. (b) Legal. (c) Economical. (d) Financial. (e) Gender and age group. (f) Ethnic. (g) Basic services and (h) Insecurity. xii. Corruption. Both urban and rural populations state that there are high degrees of corruption at all levels. Private Sector. Increase opportunities for the poor have to go through the creation of opportunities for the private sector. xiii. Urban and rural populations complain about lack of employment opportunities. In the rural areas, people not only lack employment but also land.3 Employment opportunities for the poor could only be generated if private sector is willing to invest, specialize and diversify production. To do so, a framework of incentives and a relatively safe and accountable investing environment (economic and political) have to be promoted. Manpower and labor force is the most abundant asset available in Nicaragua. The problem is that it is unskilled. These two factors restrict wage raise. The form of agricultural production is extensive more than intensive and/or technified. It requeires seasonal employement more than permanent skilled workers. Children and women participation in agricultural contributes to keep wages low. Local Organizations Social capital. xiv. Existence of a varietv of organizations does not mean that there is a substanctial and effective social capital. There is awareness among community people that membership in social networks secure benefits. They perceive the concept as "it is not what you know, it is whom you know." However, the population is not fully aware of the extent to which building social capital and working in an organized way could substantially reduce transaction costs, and increase gains and productive rates. People are afraid of indebtnesss. However, farmers are willing to work together if a third party organizes the group and ensures credit and technical assistance. In urban areas is unlikely to organize people on working groups because the poor are linked to informal sector. Opportunities Inequality and Disparities in Opportunities. 3 Until tlhe or-going titling process that began in 1990 finishes the land tenure issue would continue to limit strategies to target the rural poor. Annex 20. Page 4 xv. There is awareness on the incidence between low education and low access to health services with higher levels of poverty and with poverty perpetuation. DIsParities in opportunities and quality between non-poor and poor are reflected in: (a) location of services, (b) physical assets (low equipment). (c) human assets (limited and inappropriate personal), (d) high costs (subsidies have been removed) and lack of confidence and trust in the quality of the services for the poor. Security xvi. Insecurity and violence affect all social sectors, but the poor are affected more because there is disparity in opportunities to protect the most vulnerable groups. Domestic, street, and countryside violence have been identified by consulted people. xvii. People correlate poverty, alcohol abuse, overcrowding, extended families, and limited income with domestic violence, with women and children its main victims. Social, economic, physical, sexual and psychological abuse of children and women abound. Coupled with a lack of opportunities for youth, these forms of abuse create an environment in which the young grow up dysfunctional and are easy prey for gangs and drug dealers. Technical Assistance xviii. According to consulted farmers, three types of services are needed. (a) Traditional training consists of seminars and workshops to provide information on entrepreneurial, managerial, and organizational skills; animal husbandry production; veterinary care; agricultural production. (b) Technical assistance, when most effective, combines traditional training with applied, hands-on assistance. (c) Specialized Non-formal Education: To achieve cultural change in animal husbandry and agriculture two specialized non-formal education programs are recommended: Economies of Scale: Groups of farmers acting together allow economies of scale. Technical assistance can provide them the training and strengthen social capital. The chief goal, however, is to develop entrepreneurial activities. Small private activities avoid clientism on the part of the donor and dependence on governmental institu.tions or donors on the part of the beneficiary. Patio Economy: In all rural communities, the patio economy is considered important in al- leviating deprivation. Our findings lead us to the hypothesis that the patio economy allows woman to be close to the family and to allow rural households to eat better than do urban households. Traditional training and technical assistance can make the patio economy more productive while promoting cultural change. These methods should target diversified production, animal nutrition, hygiene, water conservation, and household nutrition. Annex 20, Page I Example of Physical and Social Maps Comarcaf Ierc Cruz - :E jlrnendro - ND. 6aOaa5 l Sur AcIpa 1151C0 cS l I is- Wrde5IIr O i'derce CmmXJn,Ja o d ye. c t ogares i brts - 4xpoes de, ?obrezj g- t1Hocar2m No iRbbesv _ CUlitU05 , - onodo Voce7nO v- 6o5queK . 7F_ a4e.j5 4t k / t -p - / 4\1,LO E ~~ -._.e-w.-.-io rt Annex 20. Page 2 yil, e S t! E f/I ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I g~~~~~ ~ tv\k"FKSC'c, 1tY ~ ~ ,-! - -z - -' '' 1 a . - _ } - ''U fE,W'--, - -E's EL~~L Y Li rP L] * > v½\ (k.~'l 4~ ~ ~ ' -4 A'1'.E tS )9' Q Annex 20, Page I INTRODUCTION 1. This report consists of six chapters and eight annexes. Chapter I provides a brief historical overview of trends and perceptions of poverty across regions. 2. Chapter II provides people's concepts of poverty, well-being and ill-being. Field data are organized in a conceptual framework of assets, which correlates economics with social and political processes. This chapter also analyzes how Consulted people understand and classify socioeconomic groups. 3. Chapter III examines how the poor and extremely poor face risk and vulnerabilities The poor use a variety of coping strategies, both legal and illegal. The study's data confirmed the working hypothesis that people have both the ability and motivation to design coping strategies, and that communities maintain an impressive capacity to rebusque and agenciarselas. People must face their problems to be effective, but lack of social capital prevents them from doing this. 4. Chapter IV deals with the role of governmental and local organizations in alleviating povertv. The study's working hypothesis is that corruption undermines development by weakening both rule of law and the institutional arrangements on which economic growth and development depend. This chapter discusses the correlation between poverty and exclusion, different tvpes of exclusion and the ways in which exclusion is made manifest. This chapter also discusses the correlation between local organizations and social capital. The hypothesis is that when organizations and entrepreneurial practices are strong, farmers form alliances that reduce transaction costs and lessen risk and vulnerability. 5. Chapter V assesses how disparity of opportunity and gaps in basic services cause and perpetuate poverty. The Nicaraguan poor know that it is not enough to build infrastructure; rather, effective infrastructure depends on qualified personal and reliable equipment. The fieldwork corroborated the hypothesis that gaps in access to education and health services result from disparity of opportunity, which include distance, schedules, and physical and human assets. 6. Chapter VI evaluates the impact of technical assistance, an important instrument of the Nicaraguan government in addressing the needs of small and medium farmers. Findings verified the working hypothesis that technical assistance is often inappropriate because it is not permanent and does not always seek to change productivity patterns or increase beneficiaries' social capital. It also verified that social aspects of organization and culture are often not taken into account when providing technical assistance. Finally, a model of informal education for entrepreneurial development is presented. Annex 20, Page 2 PART I: COUNTRY OVERVIEW A. GEOGRAPHICAL CHARACTERISTICS AND DEMOGRAPHIC TRENDS 7. Nicaragua is the largest country in Central America, with a total area of 130,728 square kilometers (or 121,428 square kilometers excluding Lakes Cocibolca and Xolotlan). Administratively, Nicaragua is divided into three regions: the Pacific lowlands, the Central highlands, and the Atlantic lowlands. In 1995. the estimated population was 4.4 million. Population has tripled in the past 25 years and is expected to double in the next 25 years. 8. Nicaragua's population historically has been unevenly distributed. The Pacific and Central regions, making up 44.8 percent of the territory, are the home of 87.7 percent of the population. The Atlantic region, with 55.2 percent of the territory, is sparsely populated, with only 12.3 percent of the population. 9. The Pacific lowlands extend about 75 kilometers inland from the Pacific coast. Most of the area is flat, except for a line of young volcanoes, many of which are still active. The total area is 18.429 square kilometers, or 15.2 percent of national territory. The smallest region, it has the highest percentage of total population and highest population density. It is the most developed and most politically important region, comprising Managua and the other large cities. Its main economic activities are animal husbandry at industrial and commercial levels. It has the most developed road infrastructure and basic services. 4 Table A20.1 - National Territory, Total Population and Population Density By Regions Region i % of National % of Total Population l Territory Population Density Pacitic 15.2 56.6 133.9 Central 29.6 31.1 37.7 Pacific and Central 44.8 87.7 -171.6 Atlantic 55.2 12.3 I Z.- Source: Bureau of Census. 1995. National Population Census of Nicaragua. 10. The Central highlands lie northeast and east of the Pacific lowlands. These mountains are forned of ridges 900 to 1,800 meters high with a mixed forest of oak and pine. In size, percentage of population, and density, it falls between the Pacific and Atlantic. The total area is 35,960 square kilometers. The economy is based on coffee, grains, tobacco, milk, and sugar cane. 11. The Atlantic region is cultural and ethnically diverse. It embodies the autonomous regions of northern Atlantic (RAAN), southern Atlantic (RAAS) and Departments of Rio San Juan, which are becoming autonomous regions.5 The Eastern Atlantic lowlands are hot and humid. Tropical rain forests are characteristic from Pearl Lagoon to the San Juan River, while in the interior lie the savannas. Fertile soils are mainly found along the natural levees and narrow floodplains of numerous rivers. The region occupies the 55.2 percent of total territory but only 12.3 percent of the population. It is the least densely populated. Its economy is based on gold extraction, industrial fishing, and forestry. It has the greatest potential for eco-tourism but road infrastructure and social services are deficient. 4 From 1950 to 1975, increasing exports of coffee, cotton, and beef in the Pacific lowlands and Central highlands generated rapid GDP growth and a notable concentration of land and wealth. As a result, small farmers were forced to sell their landholdings and came to depend on the agricultural exporters for employment and wage labor. These exporters diversified their investment portfolios into industry and services. Economic growth and diversification encouraged migration to the cities (especially Managua), fostering urban expansion. 5 This means that once the Law Manual is approved these territories will have autonomous governance. Annex 20. Page 3 12. In the ethnic hierarchy, indigenous groups - Miskito, Sumu, and Rama - occupy the bottom ranks. These groups are the poorest and least educated and are generally relegated to the least desirable jobs. Above them, at successively higher ranks, are recently arrived poor and middle-class mestizos and Creoles. Today the Atlantic region, mainly the Atlantic lowlands. faces an expanding mestizo population, many of whom live in mining areas. Since the 1950s. expansion of export agriculture in the western half of the country has forced many dispossessed peasants to seek new land on agricultural frontiers. On the Atlantic side of the central highlands, this migration has produced bitter clashes between mestizo pioneers and Miskito and Sumu farmers over what indigenous people regard as communal lands. Box I - The Atlantic Region Occupying half of the territory of Nicaragua, the Atlantic Region has the highest extreme poverty rates - 17 percent urban and 41.4 percent rural - but only 10 percent of the country's roads. In many other ways, the Atlantic region stands apart from the rest of Nicaragua. The Atlantic region's political and economic development has differed from the Pacific and Central regions. As part of the United Kingdom, the region was headed by a Miskito King named by the British Crown. With the independence of the Atlantic region and the arrival of fruit exporters from the United States, the Atlantic coast was transformed into an 'enclave" economy. in which Creole migration was encouraged and Miskito peoples relegated to lower status under a slave labor system. The Atlantic coast has undergone changes in political and economic regimes that have fostered social stratification. Both the British and enclave economy isolated the region as a political and economic strategy/. The British Monarchy protected a political regime, a territory, and natural resources. The enclave economy protected the economic interests of a private company (closely related to U.S. financial interests), a contract labor force, and political beliefs. After the overthrow of the Miskito-British regime, the Atlantic region was incorporated into the Republic of Nicaragua. However, the mestizo regime of the Pacific and Central regions continued to isolate the Atlantic coast, an isolation that the Atlantic population itself accepted. Development in the Atlantic region rests on many institutional reforms now in process: * The land titling process * Indigenous peoples land demarcation process * Operational execution manual of the Law of Autonomy for the Autonomous Regions of the Atlantic Coast (otherwise, the law will not be effective) * Definition of natural resources and protected areas policy in the region B. INCIDENCE OF POVERTY 13. Since 1990 Nicaragua has promoted economic growth as the first step in reducing poverty. Restoration of economic and political stability after hyperinflation and eight years of war was a great concem over the last decade. From 1993 to 1998, the proportion of people living in poverty decreased by 2.4 percent; those in extreme poverty by 2.1 percent. However, povertv is still widespread. Five out of 10 Nicaraguans are poor, 2 of them extremely so. Of a total population of 4.8 million, 2.3 million live in poverty while 831,000 live in extreme poverty. Widespread malnutrition, inadequate water, sanitation and sewage systems depict a nation plagued by poverty. Annex 20, Page 4 14. The incidence of povertv varies by region and by urban and rural concentration. Poverty rates in 1993 to 1998 have been higher in rural than in urban areas. The highest rates have been recorded in eastern autonomous territories of the Atlantic region, the lowest in Managua. Between 1993 and 1998, poverty decreased most in Managua and in the Central region. Managua recorded a reduction of 11.4 percent while the Central region registered one of 9.8 percent in urban and 10.7 percent in rural areas. Rural areas of the Pacific and Atlantic regions registered small reductions of 3.6 percent and 4.3 percent each, while urban areas of these regions recorded an increase of poverty rates of 11.5 percent and 8.9 percent. respectively. Though the indigenous population is not statisticallv significant in Nicaragua. there is a correlation between poverty and ethnicity; the Eastern Atlantic registered the highest urban and rural poverty rates. C. PEOPLE'S PERCEPTION OF THE EVOLUTION OF POVERTY 15. Nicaragua has suffered dramatic political and economic transformation during the last 50 years, not to mention natural disasters and bloody civil wars that have devastated the economy. Box 2 Economic and Political Changes Natural Disasters Somoza regime 1930-72 Earthquakes: 1931 & 1972, Managua Pre-revolution period 1972-79 Hurricane Joan: 1988 Sandinista Revolution 1979 Volcanic eruption: 1992, Cerro Negro Junta de Gobierno 1979-83 Tsunami: 1992 US Embargo 1983-90 El Nifio: 1996-1998 Electoral Victory of Sandinistas 1986 Hurricane Mitch: 1998 Civil War 1988-90 Electoral Victory for 1990 Democracy 16. Nicaraguans explain the -evolution of poverty using regimes as reference. During the Somoza era poor families depended upon wealthy patrons. Some poor peasants believe that larger farmers protected them while others believe that larger farmers treated them as slaves. Box 3 - Regime Change Forced People to Take Higher Risks During the Somoza period, rural people in Nicaragua worked as peons. They were used to obeying a boss, a foreman. They were not used to making decisions, holding their own land or managing a farm. With the Sandinistas' land reform and cooperative movement, most of these people got land and began to make decisions. With strong governmental support, they obtained subsidized credit and technical assistance. The government provided facilities to forgive debts. After 1990, the rules changed. Now, people have to play by market rules. They have to pay their debts and there is no subsidized credit or technical assistance. Source: Communal leader, Reparto Shick, Managua. 1 7. During the Sandinista Administration, which ruled the country from 1979 to mid-1990, education and health were paramount. To make up for long-standing gaps in the educational system, particularly in rural areas, the Sandinistas invested a significant proportion of GDP in education. The 1980 literacy campaign reduced illiteracy from 50 to 23 percent. Secondary and superior enrollment and completion rates increased. Investments in the health sector also increased along with an institutional reorganization and massive campaigns. The health sector increased its coverage and impact in preventive medicine. In later years the Sandinistas' increased defense budget forced it to reduce spending in health and education sectors. Annex 20, Page 5 18. The Sandinista Administration attempted to transform the country by expropriating land and wealth from the agro-export sector. The Agrarian Reform Law transferred about one-third of the land to small farmers. Expropriated lands became cooperatives and government-run farms later were transformed into cooperatives, as the Sandinistas sought broad popular and political support in Pacific and Central regions. The Agrarian Law was drafted to guarantee property rights and land titles. However, land titling did not begin until the early 1990s and is still in process. 19. Interviewees throughout the country perceive that during the Sandinista era, people had more access to basic services. However, perceptions vary among regions according to the degree of military presence, armed conflict, and land reform. Military organization and armed conflict was greater in the northern areas, mainly the Central and Atlantic regions. There, consulted people perceive the Sandinistas' community-based work as more political and military than economic. In contrast, community-based work, where people were organized into productive cooperatives, was more economically driven in the Pacific region. Therefore, in this region, consulted people perceive that the Sandinistas created basic social services and provided primary support to those lacking land, production inputs, and technical assistance. Meanwhile in the Atlantic region, Sandinistas were highly unpopular because the revolution was perceived as a mestizo event, culturally and politically distant from Atlantic society. 20. Hyperinflation during the mid-1980s reached 33,000 percent. Exports were one-third of imports. Exacerbated by the U.S. embargo, the economy declined. In 1988, the GDP started to fall, eventuallv reaching productive levels of 1960s. The civil war increased public expenditures for defense instead of education and health. In 1998, GDP was US$410 per capita, considerablv lower than the average of US$3,940 for Latin America and the Caribbean. 21. The greatest impact of the 1980s' economic crisis took effect in the early 1990s.The Chamorro Administration's economic reform program stabilized the economy, reduced hyperinflation, and began to liberate trade to stimulate the commercial and service sectors. External aid began to flow in support of social sectors and reform of public institutions. However, instead of associating the 1990s with economic stabilization, some Nicaraguans associate it with public sector downsizing and privatizing governmental enterprises, which reduced poor families' employment and household revenues. Annex 20, Page 6 PART II: THE NATURE OF POVERTY A. POVERTY DEFINED BY THE POOR 22. Poor people's definition of poverty goes beyond low income, low food intake, and low levels of education. Poverty for them is a social phenomenon resulting from economic, social, political, and cultural spheres acting together. 23. To be poor means to face disparity of opportunities. It means to be excluded from society's benefits. Moreover. it is to feel excluded and to live in high levels of ill-being. Consulted people define well-being as a situation in which people satisfy their socioeconomic needs and have access to basic services. They define household status by material, financial, human, and natural assets. The determining factors of low levels of well-being are high food deficit, limited diet, bad housing, low income, limited land tenure, and exclusion from the formal financial system. Consulted people perceive that if social capital is not well developed, a household's vulnerabil- ity increases. 24. In both urban and rural areas, poor people strongly correlate low development of human capital and high social vulnerability. They are aware of significant gaps in health services such as preventive medicine, which directly affect mortalitv and birth rates. Interviewees correlate low education and higher levels of poverty but also perceive that educational quality is lower for the poor. Malnutrition, limited access to land holdings and material capital, and inequality of income reflect disparity of opportunity. Table A20.2 - Households in Well-being as Defined by Consulted People Capital Result Natural Access to good land for agriculture and animal husbandry Large land holdings Human No food deficit Access to good and private clinics and hospitals (urban) Good education centers Children attend school and have no need to work Access to sewage, sanitation, and garbage collection Access to paid technical assistance Favorable dependency ratio Financial Capital and financial solvency Access to financial markets Access to well-paid jobs Access to social security and pension plans I Access to paid labor force Material Access to equipment, machinery and other means of production Access to production inputs Drinking and irrigation water-supply systems Diversified and specialized farm Strong and well-built houses Adequate furniture and appliances Access to means of transportation Social Domestic and international network of financial support Political and social connections Annex 20, Page 7 25. Both rural and urban poor perceive that to live in a state of well-being is to be safe, to be protected by the system, and to have access to material and social assets that make life comfort- able. Safety means having access to social services, or at least to be able to afford the most basic ones such as health. Safety also means having enough income to feed a family and afford a comfortable house. i. Material Capital 26. Considerable weight was given to housing. A well-built house with basic furniture is an indicator of well-being. In rural areas, a backyard or plot for patio production is also considered an indicator of well-being. The poor seldom own the land on which their house stands. 27. With the exception of the Atlantic region and Managua. most poor households have dirt floors. In Managua, urban houses have cement floors. In urban areas walls are built with cement while in the rural areas adobe, mud, and wood predominate. Urban houses have zinc roofs while rural houses have tile or straw roofs. The urban poor build their houses with cement and wood, n a style known colloquially as minifalda (miniskirt). 28. One of the social consequences of high fertility rates is the large extended family, which results in overcrowded households and high dependency rates. LSMS confirms that on average there are six persons, and often more than one family, per household. Sixteen people living in a two-room house was the densest example of a crowded household. In rural areas, houses average 60 square meters, usually of one or two rooms. Promiscuity, physical abuse, and lack of privacx are some of the results of these crowded households. Table A20.3 - Average Household Characteristics. l j ~~~~~~ ~~~Urban Rural ' |Paciic Central Atlantic R Pacific I Central Atlanti] Average age of respondent 4 44 W45 43 47 42 l Average number ot people 5 '4 l 6 6 Average number ot rooms I l 2 3 2 1 2 Range ot persons per household 3-9 3-6 4-lIU 1 -1 I 2-14 2-14 Average # ot people who work 2 l1 F1 2 2 '2' Iype ot construction House House |house tiouse/ house/ house/ l ype o construcnrancho rancho l ranchoue Floor materials OirU dirt cement/ dirt drt/ wood/J Cement cement wood cement cement dirt Wall materials Wood! T Cement/ cement adobe/ mud/ F wood cement mud / wood wood/ wood/ adobe cement adobe l Root materials Zinc Zinc zinc tiles/ zinc I iles/ zinc/ ___ _ zinc thatch Source: Semi-structured Interviews. 29. Most rural houses have little furniture. Most houses do not have one tijera (a rustic, foldable bed) per family member. In rural areas, some people use hammocks. ii. Natural Capital 30. For rural small farmers and extremely poor who eat only what they produce (a subsistence economy), land is the most important natural asset. The poor have small land holdings while ex- tremely poor have no land. A bare majority of households (52 percent) have private titles, while 13 percent have agrarian reform titles, and 22 percent of land-owning households do not have any title to their land. Only 64 percent formally registered their land (Benjamin Davis: p. 13). Annex 20, Page 8 31. There is a difference between owning land and having access to land. Both the poor and extremely poor manage to have access to land by providing services or labor for access to plots. Particularly in the Central and Pacific Regions people take care of a number of hectares without pay. In exchange, landowners allow them work a few plots of land or to keep part of the harvest. Alternatively, people hired to oversee a farm can use their earnings to rent plots of land. The average rental cost per nmanzana (0.7 hectares) is 200 C6rdobas (US$16). In the best cases, access to these lands can just satisfy annual consumption needs, while in the worst cases such access provides food only seasonally, for six months at most. 32. In the Atlantic region, Indigenous peoples'access to land and land tenure is governed by a complex system of legal, contractual, and customary arrangements. Agriculture, forestry, and grazing take place on communal properties while housing and the patio economy take place on private plots. Since the 1960s, non-indigenous human settlements along the littoral have increased. From indigenous peoples' perspective, their land is under permanent threat from immigrants, miners, oil companies, forestry concessionaires, and hunters. 33. Not to own land means not only to lack a key asset but also to be excluded from the formal financial system and to have limited access to technical assistance. Qualitative fieldwork corroborated the results of the LSMS survey. Of all rural households, only 17 percent received some kind of credit, and only 12 percent received loans for productive purposes. Those who did not seek credit said they lacked collateral to secure it. iii. Human Capital 34. Consulted people emphasized the importance of food intake, education, and health. Most perceive education as a route out of poverty' even though they also believe that the educational system for the poor is inefficient, inappropriate, and expensive. 35. Although good health is considered an asset, interviewees do not always correlate nutrition and sanitation with health. They think of health care as curative, not preventive. 36. Infant mortality has decreased, but most people still consider it a problem. Most families have lost more than two children and do not know why. They describe the symptoms but do not know the sickness. The most common symptoms are high temperature, infant diarrhea, and acute respiratory infections. 37. High fertility rates prevail across the country. However, consulted people do not often correlate the impact of multiple pregnancies with health and poverty. Some women have been pregnant 8 to 10 times, and there is a high incidence of adolescence pregnancies in both urban and rural areas. Young females are mothers at 13-15 years old, when they still are children themselves. There is little knowledge of family planning. 38. Men are always considered the head of the household. Only widows or single mothers living alone are considered household heads. If there is a man in the house - an adult son or brother - both men and women identify him as the household head, especially if he works. The head of the extended family is the father. iv. Financial Capital 39. The poor and extremely poor have occasional jobs which pay poorly. Most urban poor work in the informal sector, while the rural poor work as peons. 40. In Managua, monthly income per person averages 800 Cordobas (US $62.00) while in rural areas the jornal (daily wage for six-hour day) ranges from 15 to 20 Cordobas (US$1.16 to US$1.55) and averages 17 C6rdobas (US$1.3 1).-Assuming that a person works the whole month, rural earnings average 510 C6rdobas (US$39.53). Annex 20, Page 9 41. Unless household heads have a stable job, the family does not have a stable income. The dependency ratio is very high. In general. there is only one source of income, and rarely there are two. Access to job opportunities is limited, and most people consulted have not looked for jobs. claiming that the few jobs available are highly specialized and consequently far beyond their skills. 42. The rural poor consume most of what they produce. They have little skill in selling their produce, and know little about fluctuation of prices and dealing with intermediaries. Box 4 - Mitigating the Impact of Poverty The urban poor face unemployment and underemployment. Transaction costs such as transportation are more expensive in urban than in rural zones. In contrast, the patio economy and credit from local stores or among families are widespread and seem to mitigate the impact of poverty on rural households. That it is why a backyard or a plot near the house is considered a key asset by the rural population. Unfortunately, there is no statistical information on the impact of the patio economy on rural poverty. v. Social Capital 43. Social capital is the "ability of individuals and households to secure benefits by virtue of membership in social networks or other structures" (WDR 2000/1: S.] 2). * Existence of a variety of organizations does not mean that there is a notable development of' social capital. There is positive and negative social capital. The positive social capital implies benefits to the participants. This form of social capital could be positive to participants while it could negative to other community members who are excluded. Gangs are forms of negative social capital. These groups could be well organized that carried out negative impacts to the community. * Social capital as any other type of capital involves stocks. In this case. stocks are trust, norm s and networks. If communities have a wide dissemination of "formally" built organizations does not mean that communities have an important development of social capital. * Communities have an important level of social capital if local forms of organization generate positive externalities or benefits' These externalities could be economic such as reduction o-l transaction cost or social externalities such as the result from coping and survival strategies during natural disasters, for example. * Externalities are feasible if community people trust each other and are able to organize a sel of norms to achieve such benefits. 44. There is a relationship between social capital and entrepreneurial tradition. In Nicaraguan history, people have not been exposed to conditions that develop entrepreneurial tradition and social capital. With exception of indigenous communities, local forms of organizations are more discussion forums where they debate over community problems instead of being productive entities. However, these forms of customary organizations have great potential if their managerial capabilities are strengthened. Annex 20, Page 10 45. With exception of indigenous communities and some mestizo communities (such as Nindiri) who have had a positive experience with an NGO or other donors, most urban and rural areas avoid organizing to jointly raise animals or crops. They make this decision even though they know it might benefit them economically. Distrust among neighbors and fear of indebtedness cause people to continue working independently. Rural communities are willing to join groups organized by donors because these activities include credit and some technical assistance, which they cannot obtain otherwise. In addition, projects that provide rural credit require communities to identify people's preferences for agriculture or animal husbandry. The degree of involvement of communitv members during the project identification phase determines people's willingness to join these rural credits. Greatest beneficiary participation increases trust and ownership. 46. During the crisis caused by hurricane Mitch, communities organized rapidly to evacuate people and store water and food. However, most of these organizations weakened or disappeared once donors began to channel aid to Nicaragua. 47. This studv found that communities trust their own organizations and communitv-based projects more than government institutions, which consulted people vieu as remote. Rural communities feel closer to local governments than to the central government. Urban communities, particularlv Managuans, perceive both local and State Institutions part of their reality, but do not perceive that such institutions work for the poor. 48. According to the LSMS 98, "only four percent of rural households participate in producer organizations, with another two percent in a productive project. Far more prevalent are non- producer organizations, involving over a third of rural households. The bulk of these organizations are religious in nature" (Benjamin Davis). 49. Design and execution of community development projects must involve local leadership to legitimate the process within the community. This would avoid creating parallelforms of organization and would empower the existing ones. There are three types of "positive" local leadership: (a) traditional communal leaders normally linked to political parties, who control, for instance, Community Assemblies (Asambleas Comarcales); (b) indigenous leaders in charge of community justice, medicine, religion and cultural affairs; and (c) modem leaders who have been trained by NGOs and have executed communal development projects. Some of these have had formal education such is the leader of El Cacao, who completed secondary education and was by the NGO Tierra y Vida and Reparto Shick lawyer linked to Movimiento Communal. This type of leadership is associated with newly formed organizations for and could foster cultural change in production, nutritional habits and environmental conservation. 50. There is also a "negative" form of leadership exercised by gangs, arms sellers and drug dealers. This leadership normally does not have call for meeting power at community level, but they do have such type of power with certain groups of youth. Annex 20, Page II Local Organizations 51. Consulted people identified three types of social networks: Basic Networks, formed by families and neighbors, help each other in death, birth, sickness, natural disasters, childcare, and farm labor. Not formally organized, such networks respond to specific circumstances. Community Organizations are local forms of governance headed by a community leader. These organizatioils make decisions for the community. In indigenous groups, decisions are made in General Assemblies in which the whole communitv participates. Leaders are often tribal elders. These as- semblies have the political capacity to call for meetings, but they cannot effectively organize to increase production or foster strategies to reduce transaction costs. Instead, they can organize the community to cope with natural disasters. 52. There are four types of community-based development organizations: (a) Infrastructure Development Organizations: committees created by donors such as water committees. (b) Productive Development Organizations: micro-credit lending programs that organize groups of peasants into projects to grow crops or raise livestock. (c) Religious Organizations: church-supported groups that seek to combine religious train- ing with technical assistance to build infrastructure or to support communal projects. In some rural areas religious organizations form groups of peasants to whom they provide in- kind credit, technical assistance, and religious instruction. In some urban areas these organizations provide technical assistance and religious instruction to fight drugs and violence, (d) Cooperatives: Mostly the legacy of Sandinismo, these cooperatives' functions vary byv region. In the Pacific region, cooperatives address economic and productive goals, whereas in the Central region they work more as political entities. Regional and National Organizations 53. In the Pacific region, indigenous peoples have the Federacion de Comunidades Indigenas and Union of Cooperatives for the Sutiaba Indigenous Peoples, while in the Atlantic, indigenous communities have the Asociaci6n de Sindicos del Atlantico. All these organizations function as spokesmen for the community. Regional mestizo trade unions also include indigenous peoples. 54. At the national level, trade unions such as the Uni6n Nacional de Agricultores y Ganaderos (National Union of Farmers) includes individual farmers, local associations, and cooperatives. They also include communities. Indigenous Organizations 55. The basic indigenous form of organization is the community. Within the community, traditional organizations and social systems deal with religion, education, justice, and political and economic power. Indigenous communal organizations discuss critical community issues, formulate coping strategies, and approve donor's development initiatives. They also function as pressure groups in front of govemmental institutions, mainly municipalities. 56. These organizations have a notable power to convene that makes them potential agents o'f change. All indigenous communities have spiritual guides, healers, moral or ethical judges, coun- cils of elders, sports groups, local and communal committees, and family associations. Most communities have three types of organizations: local, traditional authorities, and civic authorities. Table A20.4 - Community Organizations Annex 20. Page 12 Level i Type Social Function Local Cotriadas, spiritual guides. |'o preserve and promote the Education. culture T raditional Legal: Sukias, council of jIao keep order following Authorities elders and tribes councils. indigenous peoples customary law. Medical. healers, midwives, To maintain mental, spiritual, and doctors and physical health. CSivic | Indigenous mayor, "sindicos" I o connect indigenous peoples Authorities and council members. culture, traditional authorities and I customary law with institutional arrangements and legal system of govemment. 57. Local organizations preserve and transmit cultural heritage. Local authorities preserve social order and ensure that the customary legal system works. They resolve community problems and to maintain citizens' well-being. Traditional authorities resolve land conflicts, using customary law. Civic authorities are political actors that link communities with local governments. B. SOCIOECONOMIC GROUPS 58. Men and women similarly define poverty and classify socioeconomic groups. Consulted people identified three socioeconomic groups: the extremely poor, the poor, and the non-poor. It was easier for the poor to identify the extremely poor and poor, and in some places consulted people, especially the extremely poor, had difficulty identifying non-poor and rich groups, because wealthy living standards are beyond the socioeconomic reality of the rural extremely poor. Only in municipalities or sub-regions with significant private investment do the poor have no difficulty defining the non-poor and the rich. Annex 20, Page 13 Table A20.5 - Key Characteristics of Poverty by Gender, Ethnicity, and Urban and Rural Areas Area Both Genders Gender Differences Indigenous Indigenous and (levels of emphasis) Peoples non-indigenous (both genders) I (Similarities) ] (levels of l_______________ l__________________ em phasis) I Men Women Urban Similarly detine: Employment* Health t* Culture - Poverty | Income * Education Rural | * Socioeconomic C Credit * Domestic * Traditional groups * Means of violence medicine * Vulnerabilities production * Family values * Education * Priority needs . Family * Natural [ * Social capital planning resources !ain Rural: Poor and extremely poor have no problem detining poor and extremely -- dif- poor. but often have problems defining rich and non-poor. Wealthy ference living standards are far beyond their socioeconomic reality. The poor did not have problems to identify rich and non-poor in municipalities or areas with important private investments. Urban: Easily define all socioeconomic sectors. Source: Social mapping, social capital inventory, and wealth ranking exercises. 59. Rich and non-poor are easily identified in agricultural frontier areas, where there is significant private sector investment. For example, on the southern Atlantic coast (El Almendro. El Coral. and Nueva Guinea, in the animal husbandry triangle, rich farmers are growing up a large animal husbandry and milk industry. They are buying large tracts of land, some of which used to belong to the desmovilizados, (former Army soldiers, revolutionaries, and contras) In El Almendro, consulted people associated rich people with large farms and associate non-poor with medium-sized farms. 60. In urban and semi-urban areas and in zones with a notable presence of private capital, consulted people also easily identify the rich and non-poor. In urban Esteli, Managua, and Nindiri, and rural Sabana Verde and Chontales, wealth is associated with medium and large land holdings and with commerce. 61. Only indigenous peoples distinguish between "extreme poverty" and "precarious extreme poverty." For them, the latter are the destitute poor,6 a small group of extremely poor, physically or mentally handicapped that impedes them from working or taking care of themselves. They are often very old and have lost their ability to manage risk or to cope with crises. Box 5 -The Destitute Poor The very poor people here are those who have health problems, or who are handicapped, crazy or blind. They are those who lack jobs, food, or housing, or are orphaned. No one cares for them. Source: Creole women. Pearl Lagoon. Atlantic region. 6 Kozel, Valerie and Barbara Parker. Poverty in Rural India: The Contribution of Qualitative Research in Poverty Analysis. Page 18. Annex 20, Page 14 The Rich 62. Consulted people generally perceived that the rich-both urban and rural-have inherited land, farms, or family businesses. The poor perceive the rich as "other people" as "different," as if they do not belong to the municipality. The rich build their political and economic network, along with their social capital, elsewhere. Their children attend private schools away from these communities or even overseas. Their social capital and financial network exist at national and international levels. Locally, thev obtain basic services such as electricity, drinking water, and technical assistance by purchasing them privately. While they might have property and businesses there, they do not necessarily live in rural communities. The Non-Poor 63. Rural non-poor live in the community where they own small and medium sized plots of land. Some have access to credit and paid technical assistance. Some can hire peons and farm workers. They use health and educational services available in municipalities, in the Department, or in Managua. Some diversify their eamings through small businesses such as groceries or agriculture supply stores. The dividing line between non-poor and poor is access to land, the quantitv and quality of land, the quality of food intake, and housing. The non-poor produce enough food to feed their family and have una casita bonita (a nice little house), interviewees said. The Mobile Poor 64. The mobile poor have little income but are free of debt and have assets or employment that improves their quality of life. Their numbers are few. Box 6 -Two Cases of Economic Mobility El Guanacaste, in the Pacific region, is semi-urban, with a well developed service sector. Two families there have managed to move upward economically. Both have established their own small business. One, a carpenter, was been trained by NGOs, ecological brigades and by the Ministry of Construction and Infrastructure to establish a cooperative to maintain roads and paths. The other man, originally from La Conception, migrated to Managua where he was trained as a brick maker. He saved money and moved to El Guanacaste. While working as a peon, he saved enough money to begin his own brick factory. Both men saved money while their wives and children supported the family business. The brick maker was supported by his parents while the carpenter was supported by NGOs and the government institutions. The brick maker makes a clear correlation between education and breaking the circle of poverty. Therefore, he has encouraged his children to study. The Poor 65. The rural poor may own small tracts of land (from 1.4 to 3.5 hectares), not always of good quality. They may rent land and work as peons for non-poor or rich people. Their houses are in bad shape, but they consider themselves better off than the extremely poor. 66. Indigenous people consider themselves poor, but not as poor as the non-indigenous poor in the rest of the country. They perceive that having communal land minimizes their vulnerability. Annex 20, Page 15 Box 7 - Poverty As Viewed by Atlantic Indigenous Peoples "We believe that the poor are in verv bad condition. The poor are those living in the Pacific. Their houses are made of cartons, and they eat out of garbage cans. Here, we have houses and food." Source: Miskito leader. Puerto Cabezas. Northern Atlantic. 67. For indigenous peoples, vulnerability increases as their customary land arrangements and cultural heritage is threatened by land demarcation, migration within Nicaragua, and private investments. Private enterprises such as oil and mining jeopardize their ancestral lands and culture rather than creating jobs for indigenous peoples. 68. Households headed by women do not always perceive themselves as vulnerable. Thev are aware of domestic violence and of the triple burden of labor, childcare, and housekeeping, but they consider themselves to have assets equal to those of other households. Box 8 - A Female Head of Household Orfa Maria Cruz, 55, is separated. She lives in Veracruz, a small community in El Almendro Municipality. She used to lived in San Miguel, but was forced to migrate because of the war, as were most of the other residents of San Miguel. "More than being a single mother, having little education is my biggest problem. Living alone is manageable because I make my own decisions, but being uneducated limits my economic mobility," she says. However, her father taught her how to read and write. She has nine children. The first one was born when she was 19 years old. The four eldest attended school in San Miguel, but in Veracruz the school is too far away and she is afraid of sending her children there. Four of her sons migrated to Costa Rica. Today, only two of her children live with her. Her economic activity varies. She grows corn, beans, fruits and raises pigs and chickens. She sells firewood and milk. She saves monev to be prepared for crises. Orfa has never requested credit or loans. She would like to request credit from the local cooperative, but the Municipality of El Almendro excludes the community of Veracruz. There is no heath post in Veracruz. The community is highly dispersed and is moribund because most of the population was war veterans who sold their lands. There is no community-based organization. There is no trust or solidarity among people and most are illiterate. The main vulnerabilities are: social exclusion, low access to health services, and lack of latrines, which increases sickness. Source: Case Study, El Almendro, Atlantic Region. The Extreme Poor 69. The rural extreme poor lack land or have small plots of less than 4 hectares. They do not own a house. Most live in makeshift structures made of materials that have been trashed (ripio). They lack access to water or sanitation, and their only source of income is working asjornaleros (agricultural day laborers). Annex 20. Page 16 Box 9 - Extremely Poor Family Maria Crecencia Reyes, 45, is a native of Nueva Guinea (southern Atlantic region). She migrated with her children, parents, brothers and sisters to Los Mollejones in Santo Tomas because of the war. Seven people form the household. She has two sons. Her older son is married to a 24-year-old woman who has been pregnant four times. One child died of asthma; the surviving children are 8 months, 4, and 5 years old. Al] family members are illiterate, and none went to school. With the exception of the oldest son who has been trained by INTA, none of the family has received any training. Her sons work as peons. The family has one cow and one horse. They have a patio economy. They grow animal husbandry and grains. Monthly income is 400 C6rdobas (US$32.20) including the older son's wages as a peon. The Reyes family does not own a house or land. They live on the farm of a rich family. They do not have water or electricity. The water used at home is brought from the river. Social networks are linked to family and relatives. Her brother lends money to the Reyes family at 10 percent interest. This family avoids participating in community organizations because they believe that most are highly politicized. Source: Case study. Central Region. 70. The poor and extremely poor face a dilemma. They cannot afford to send their children to school, and they know of no scholarships or educational aid programs. Most consulted people rank illiteracy and low school attendance among the main causes of the perpetuation of poverty, but do not believe that the educational system provides good quality education for their children. Sending their children to school represents a loss of income since children who work contribute to household income. Thus, for them the opportunity cost of sending children to school is not justified. 71. The incidence of child fabor is higher in rural than in urban areas. Boys work more than girls. According to LSMS, 16 percent of the rural labor force are children from 10 to 14 years old. In urban areas, 6 percent of the total labor force are children. 72. Poverty is higher in relative terms in the countryside where about 63 percent of poor and 78 percent of the extremely poor live. However, rural interviewees do not always consider them- selves poorer or more vulnerable than the urban poor. The patio economy-owning a small plot of land or a backyard to grow food for the family- tends to mitigate people's perception of poverty. There are no economic estimates of the patio economy's effect on poverty. 73. Consulted people stress that most of the urban non-poor have formal jobs, a perception confirmed by the LSMS. In Managua 51 percent of the non-poor were employed in formal jobs in 1998. Except for Managua, the Atlantic region registered the highest urban employment rate, 47 percent, while the Central region registered the lowest, 39 percent. In urban areas, 98 percent of the employees that have formal jobs work for gas, electricity, and water companies; others are drivers or work for rich families. They can afford a more diversified diet, and most own a house, which gives them economic collateral and access to formal financial sources. Annex 20. Page 17 Figure A20.1 - Proportion of Urban Households Employed in Formal Jobs 1993-98 50. - i 40. 000-_ 20 ,t1 10,; Managua Pacifico Central Atlantico Total 74. During 1993-98, the proportion of formal jobs in total households increased slightly, from 3 1 percent to 36 percent. The most important changes occurred in Central and Atlantic regions, where the proportion of urban households emploved in formal jobs increased by 6 and 5 percem respectively. In these regions the proportion of underemployed also decreased. In the Atlantic region, urban underemployment dropped from 40 percent to 17 percent, in the Central region from 44 percent to 21 percent. Table A20.6 - Underemployed, by Urban Areas (1993-1998) Region Underemployed Managua 1993 1998 Diiflerence | Managua _ 43.0 42.0 l l Pacitic 44.0 32.0 1 -12 Central i 40.0 21.0 1 -19 Atlantic i 40.0 l 17.0 -23 | otal 41.0 l 32.0 -9 75. People's perception of permanent employment coincides with LSMS findings. Only a few urban poor have permanent jobs. Most urban poor are underemployed or depend on the informal sector, which consists of small-scale enterprises that use simple technologies and as a whole operate outside the legal regime of labor protection. The informal sector encompasses activities such as mattress makers, bakers, shoemakers, shoe cleaners, and carpenters. It also includes people who take in laundry and ironing or prepare food for sale in the streets (fritangas and desserts), people who sell small plastic bags with water and a wide variety of merchandise. 76. The urban extremely poor work in similar small enterprises or are unemployed. Thev also prepare food to sell in the streets, sell plastic water bags, and sell inexpensive merchandise at stoplights. Informal sector earnings are generally extremelv low. Typically, a man works in an) of these activities, while his wife makes tortillas or other food to sell in the streets. Their children beg, sell cigarettes, or clean car windshields at stoplights. Table A20.7 - Characteristics of Non-Poor, Poor and Extremely Poor as Defined by Interviewees Rural Non-Poor Poor Extremely poor Own or rent tarms from 3.5 to Own or rent land holdings Lack land 35 hectares from 1.4 to 3.5 hectares Have livestock (up to 50 cows) Have livestock (I to 5 cows) Lack livestock Have patio economies Some have patio production Some have patio production Have transportation (pigs and poultry) (pigs and poultry) Annex 20, Page 18 Urban and Rural Non- oor Poor Extremely poor Do not migrate or migrate to Emigrate during harvest Migrate according to harvest the United States seasons or permanently to seasons or permanentlv within Managua, Costa Rica, other Central America, mainly to Central American countries, Costa Rica or the United States Produce employment Are selt-subsistent Are sell-subsistent Have diversified economies Some can sell production Some get family remittances - Some get familv remittances Some get tamily remittances Have a diversitied diet and eat Eat twice and sometimes three Eat once and sometimes twice three times daily times daily Diet has two or a day Diet has one or two ! ~~~~~~~~~three staples (rural) stapies (rural)i i Have collateral (own land or Have collateral (own land Lack collateral and are thus house) They have access to a or house) excluded from the formal wide variety of credit and can financial system afford technical assistance Send most otftheir children to Send some children to school Make most children work school I instead of sending to school, Houses have more than three FHouses have up to two rooms i Houses have one room rooms and are in good shape in bad shape I generally made of salvaged l ! ~~~~~~~ ~ ~ ~~~~~~~materialsl Up to two tamilies per house I wo to three tamilies EI nree or more tamilies l _____________________________ per house per house Sleep in own tieras (rustic Sleep in ttyeras, often shared Sleep on bare Iloor on plastics i type of foldable bed) l and/or cardboard I Have appliances (radio, T V Have some appliances, mainly Have no appliances and domestic appliances) radio _ I Most can attord private medi- Some can attord some Cannot afford medicines cal system and medicines medicines Source: Focus groups, semi-structured interviews, and case studies C. PEOPLE'S PERCEPTION OF ILL-BEING 77. III-being is the opposite of well-being. The poor perceive degrees of ill-being. For all, the priority need is to eat daily. What are we going to eat today? How many times can we eat today? Must someone in the family eat little or nothing to have enough food for the children? A family that cannot eat three times per day has a significant degree of ill-being. "Ellos no estan bien. " 78. In all areas, people perceive the extremely poor as those who lack access to food. They also lack the means to afford minimum food intake, and therefore have to: * Go through hungry periods. * Reduce their food intake by up to two meals per day. Limit their meals to one or two staples (rice and beans, or beans and tortillas, or rice and tortillas).7 * Exchange food for food. For instance, if they have extra sugar they exchange it for beans, rice, or corn tortillas. This trade pattern is widely found in rural areas. * Borrow money. To buy food, some borrow from private money lenders. Interest rates vary but can be high. * In rural areas, ask for credit at stores such as the pulperia (convenience store) and tienda campesina (a cooperative that functions as a hardware store, seed distributor, credit lender, and department store). 7 In the Nicaraguan diet, the bastimento is a staple such as tortillas, plantains, or corn. Depending on the region, rice and beans or other ingredients are added. When describing their daily diet most peasants start with bastimento as a given and then list other ingredients as additions to daily food intake. Annex 20, Page 19 * Live in houses built of salvaged materials * Live in overcrowded houses, a situation that promotes promiscuity 79. In urban and rural areas, food is cited as the first priority. Most rural people cite land as a second priority, and most urban people cite employment second. In all areas, housing is the third priority, followed by health, education, and water and sanitation. Table A20.8 - Priority Needs As Defined by the Poor Priority Rural Urban I | Food 0Food 0 11 I Land and Housing Employment and Housing III | Health Health IV Lducation Education V Water and Sanitation iWater and Sanitation 80. It is worth noting that none of the consulted people cite water and sanitation as a priority. Moreover, none cited clothing as a priority, although some cite having to purchase school uniforms as a financial burden. However, none mentioned lack of shoes or clothing as a manifestation of ill-being 81. Both sexes say health is a priority, but women strongly correlate it with ill-being. For urban women, civil insecurity is a manifestation of ill-being. Men strongly correlate lack of income with ill-being. However, rural and urban men perceive income differently. In rural area-,. men want financing to work the land. They prefer to be self-employed farmers because that is w.hat they know how to do best. A permanent job is not their first priority, as it is for urban men. Annex 20, Page 20 Table A20.8 - Priority Needs. As Defined by the Poor E Area I Both Men Women Urban Food T 1- Financial Capital: 1- Human Capital: Employment/ Well paid jobs. Health Housing 2- Human Capital: Nutrition. I Health Social security and pension plans. Family planning. Civil security (street violence). L3 Violence control (street and Garbage collection. Sewage and sanitation services. [ domestic). 3- Social Capital: Pollution and sanitation services. Organizational skills.Education l Organizational skills. Children: Appropriate education for poor and indigenous peoples. Trade and skills. Adolescents. Trade & skills. Adults. Trade & skills. 2- Social Capital: Organizational Skills Rural Food 1- Financial Capital: Land / Housing Credit Health 2- Human Capital: Appropriate technical assistance. Information and knowledge (trade, fluctuation in market prices, risk management). 3- Material Capital: Equipment, machinery, and means of production. Roads and paths. l Irrigation systems 4- Social Capital: Organizational skills. Reduction of transaction costs. Nees Common to All l- Hiuman Capital: Nutrition Health. Education. Appropriated education for urban or rural poor. Trade and skills. Culturally appropriated education for indigenous peoples (indigenous heritage and natural resources management). C Domestic violence. C Trade off: Large families- high dependency ratio- children work C Trade off: Children labor versus education. 2- Financial Capital: E Job Opportunities. :1 Income. 3- Material Capital: E Roads, paths, and transportation. Source: Social mapping, social capital inventory, and wealth ranking exercises. Nicaragua. October 1999- February 2000. Annex 20, Page 21 PART III: VULNERABILITY AND COPING STRATEGIES 82. The objective of this section is to show that the poor are aware of the multiple sources vulnerabilities and wide variety of risks they face. The population has a variety of self-protection mechanisms, some of them within the legal framework, others not. Our fieldwork corroborated the hypothesis that people are not losing their capability and motivation to cope. Both rural ani urban vulnerable interviewees have an impressive capacity of rebusque (adaptability), but they have problems being effective because social risk management is weak and counts with few assets. 83. Social risk management involves public and private assets. Public institutional infrastructure and policies can help poor households respond to shocks, and households can use their own assets as well. The poor development and outreach of public institutions, coupled with households' limited asset-base, increase the impact of market fluctuations, natural disasters, legal and civil insecurity on households. Structural adjustment policies have changed the rules of game in for credit, subsidies and commercial policies. These changes have undermined households' capacity to respond efficiently in the short-term. 84. Households' coping strategies can resolve short-term problems, but can also increase vulnerability over the long term. The share of household members employed and the number of hours worked have increased household incomes, most likely with high participation of child work. In 1998, the incidence of child labor (10-14 years old) was 16 percent of the labor force in rural areas and 6 percent of the labor force in urban areas. Sending children to work may increase household consumption in the immediate term, but also might increase vulnerability over the long-term because this strategy deprives younger generations of the benefits of being educated. On average, each additional year of education increases the social rates of return on income by 8 percent for males and 7 percent for females. 85. The poor are aware-of the multiple sources of vulnerability and risks they face. Annex 20, Page 22 Table A20.9 - Types of Vulnerability as Described bv the Consulted Population Nutritional Malnutrition Hungry periods Low food intake Poor gastronomic culture Lack of knowledge to link nutrition and sanitation to health Basic Services -E-ducation Lack of physical access Lack of money Lack of confidence in services Unqualified teachers Poor infrastructure and educational quality Deteriorating school infrastructure Lack of materials High repetition and dropout rates Low levels of learning Health Ignorance of services available at municipal and department level Lack of confidence in services Little equipment, human resources, or medicine Agricultural A high-risk activity Non-diversified production Limited access to technical assistance Cultural resistance to production changes Lack of irrigation and drinking water Low Productivity Low soil fertility and productivity Economic Limited sources ot income Limited access to credit Indebtedness Market prices fluctuations Lack of trade skills Lack of transportation High dependency on intermediaries Environmental Hurricanes River floods Dry season Plagues Civil Street violence Insecurity Drug dealers Gangs Assaults E Cattle theft Stealing Domestic Male alcoholism Violence Promiscuity Incest Disproportionate impact on women and children ________ _ [Rape A. VULNERABILITY i. Nutritional 86. In all sites, the poor and extremely poor have serious difficulty feeding their family and maintaining a balanced diet. The national diet is based on gallo pinto, a dish of rice and beans. Annex 20. Page 23 Box 10 - Meals Made of One Staple "Sometimes we eat beans three times a dav. If the chicken lays an egg, we are in better shape. One of the most difficult parts is that if we have beans, we do not have rice or tortillas. We can handle it, but not the children..." Source: Woman. El Cacao, Santa Teresa. 87. Nutritionally, there are two differences between poor and extremely poor households: the amount of staples used per meal, and the amount of meals households eat per day. Some poor households can afford to eat three meals per day, but these meals are not nutritionally balanced. Their dailv diet consists of rice, beans, tortillas and a beverage of tropical fruits (fresco). Each meal can have up to three staples, but often it has only two. Sometimes families add fresh cheese (cuajada) but can rarely afford meat. The extremely poor can afford to prepare a maximum of two meals per day. comprising one to two staples. 88. There are cultural and informational gaps about nutrition between the poor and non-poor. Neither poor nor extremelv poor families know what a well-balanced meal is. They rarely eat vegetables and fruits, which are available at reasonable prices. When asked what a well-balanced diet means, they say adding milk and meat. ii. Agricultural and Environmental 89. There are no major regional differences in perception of agricultural and environmental risks. 90. The rural poor live daily with the low profitability and high risk of agriculture. Limited control over weather and natural disasters increases people's feeling of vulnerability. Indeed, Hurricane Mitch and El Niiio caused ecological changes that increased agricultural vulnerability. Lack of material, financial, and social capital make the poor more vulnerable to such environmental risks. 91. Inequities in garbage collection, water and sanitation infrastructure increase people's health vulnerability. Although environmental problems are widespread and none of the rural poor communities have a garbage collection system, the rural poor do not seem to directly correlate pollution with health as the urban poor do. Rural communities that have worked with donors on environmental issues are aware of the environmental impact of deforestation, burning soils, and erosion (Nindiri , El Cacao in the Pacific region; Santa Lucia in the Central region and Laguna d- Perlas, Wasakin and Tasbapain in the Atlantic region). 92. The urban poor in Managua are highly critical of pollution and lack of sanitation, which they link to poor health. They correlate deficient garbage collection, absence of sewage systems, and poor treatment of wastewater and solid residues with health and the environment. Most wastewater and solid residues in the area are dumped into Lake Managua. Burning garbage in streets and backyards is common, and the poor, mainly the urban poor, link frequent respiratory diseases with the resulting dust, pollution, and smoke. The urban poor live amid filth - unpaved streets, sewage, garbage-filled corners and alleys full of waste. Many poor live along the banks of polluted rivers. In cities such as Bluefields where seawater is visibly polluted, consulted people do not cite water pollution as a social problem and do not link it to poor health. Annex 20. Page 24 93. Qualitative findings concur with the LSMS 1998, which finds that 31.1 percent of households in the whole country have garbage collection service by truck. The most unsanitary conditions are in the Pacific and Atlantic rural areas, where 0.8 percent and 0.7 percent of the households have garbage collected by trucks, respectivelv. In urban areas the coverage ranges from 50.1 percent in Managua to 57 percent in the urban Central region. Burning is the first option to get ride of garbage in both urban and rural areas; 46.8 percent of households burn garbage. Percentages are higher in rural areas; 80.1 percent of rural areas in the Pacific region follow this practice. The second option is to dump garbage into rivers. Nationwide, 15 percent of households pollute rivers with garbage. Percentages are higher in rural communities of the Atlantic (51.4 percent) and Central (35.1 percent) regions. On average, 0.6 percent in both urban and rural areas use garbage as fertilizer. iii. Economic 94. In rural areas, economic and agricultural vulnerability are closely related. Consulted people classify economic vulnerability as poor earnings. limited access to credit, fluctuations in market prices, lack of trade skills, lack of transportation, and high dependency on intermediaries. In contrast, the urban poor are vulnerable because they suffer from unemployment and have limited resources and collateral to access credit. iv. Information and Knowledge Gaps 95. Lack of knowledge and information are critical barriers to advancement. If poor people don't know how NGOs, the financial system, and government programs work, they cannot be said to have true access to these resources. More specifically, they need to know rules. procedures, regulations, options, opportunities, sources of technical assistance, and sources of financing. v. Post-War 96. People identified three sources of vulnerability caused by the civil war. Reinsertion of Army Soldiers. Young poor and extremely poor army conscripts had diffi- culty returning to civil society, as did former revolutionaries. The economy was unable to provide new job opportunities, and the war veterans did not have the skills required by existing jobs. There were no social or psychological programs to help them move from soldiers to civilians. The Veterans (Desmovilizados) Plan sought to resettle war veterans in the Atlantic region, but success was limited. First, most veterans came from other regions of Nicaragua. Lack of family, social, and financial networks impeded rapid development of their social capital. Second, the resettlement plan included land, but neither technical assistance nor financial support to encourage production. As a result, most beneficiaries sold their farms at low prices for quick cash. This has contributed to the land concentration that has been taking place in the southern areas of the Atlantic region, a process known as "la Chontalizaci6n."8 Landmines. Landmines threaten the safety and lives of peasants, and much potentially pro- ductive land cannot be cultivated because it is mined. 8 Chontales is one the richest Departments known by larger farms and land concentration among few farmers. Annex 20. Page 25 Box 11 -Post -War Vulnerability In the Pacific and Central regions, fields studded with mines threaten the safety of peasants and children (northern part of Chontales, Boaco, Matagalpa, and Jinotega in Central region and from Wiwili to the Pacific coast). In the Central regon, people say that abolishing military service set them back economically and socially. Thousands of young men and women were returned to civil society without job opportunities. Trained to be soldiers, they did not have skills needed for civilian jobs. In the Atlantic region, people are concerned about the resettlement plan designed for war veterans. (contras, FSLN, and desmovilizados). This program was unable to settle them in the southern agricultural frontier lands because, among other factors, it did not include technical assistance or credit. vi. Civil Insecurity and Violence 97. Consulted people perceive violence as a serious national problem that affects all social classes. However, they also place violence at the core of poverty, because the poor lack the means to protect themselves from violence. 98. People distinguish three types of violence: (a) domestic violence, which encompasses mistreatment toward women, children, and in some cases men; (b) street violence, such as delinquency, means civil insecurity; and (c) rural countryside violence, which means farm insecurity, such as cattle theft. (Caroline Moser and Elizabeth Shrader 1999, page 3). 99. Domestic violence prevails in all areas, and women repeatedly correlate men's abuse of alcohol with domestic violence and waste of limited household income. There are two main types of domestic violence: (a) Sexual Partners. In the majority of cases, men are the aggressors and women the victims. However, there are cases in which women are the aggressors. (b) Adults against Children. Biological parents, one parent's sexual partner, or any other adult may physically or emotionally abuse children. 1 00. Because neighbors often abuse children most women, especially in Managua. avoid leaving their children alone at home. Instead they bring them along to the streets, making them beg and clean cars at stop lights. They use child labor as a safeguard against child abuse and as a. way of keeping the family together for security. Except in Managua (Reparto Shick), people say child abuse occurs more frequently than violence against women. Box 12 -Wife Beating vs. Parking Fines The fine for beating a woman is 25 C6rdobas (US$2.00), whereas the lowest parking fine is 600 C6rdobas (US$48.38). Source: Commissioner for Woman Protection. Bluefields. 101. The family is the basic social institution of every society. Family is an asset. One's values, ideas, beliefs and forms of relationship are learned within families. Violence erodes household relationships. damages children, and increases the risk that they will grow up socially dysfunctional, join gangs, and become drug dealers. Annex 20, Page 26 Box 13 - No Opportunities for Young People "Half of our youth believe that there are no opportunities in Nicaragua. Unemployment is the major problem." Source: Police of Nicaragua. Plan de Prevencion de las Pandillas. Internal Document. Page 3 . 102. Street violence prevails in urban and semi-urban areas, which are threatened by drug dealers, gangs, and common delinquency. Children and youth face a permanent threat, houses are subject to frequent robberies, while public service infrastructure (such as electricity) is vandalized. Recently, street violence is also occurring in rural communities such in Somotillo, where in February gangs destroyed most of the public streetlights. 103. Although most gangs are in Managua. they are a growing presence throughout Nicaragua. Gangs are the prototypical example of negative social capital. Their structure varies, ranging from 10 to 100 members, from 10 to roughly 25 years old, and including girls and boys. Relations with communities also vary. Some gangs prey on the community, such as Los Come Muerlos (the Dead Eaters) in Reparto Shick, Managua. Others co-exist with community residents, while still others act as informal vigilantes, as in El Calvario, Esteli. According to the Nicaraguan Police Department, easy access to firearms is making gangs more violent and sophisticated. 104. Well organized gangs have norms, recruitment procedures, ideas and beliefs, and punishment, and compensations actions. Trust, loyalty, and solidarity are fundamental to group cohesion. Membership benefits and externalities include cash, drugs, and social protection, but these benefits do not extend to communities and households. On the contrarv, gangs cause family breakdown and civil insecurity. Annex 20, Page 27 Table A20.10 Total Gangs in Nicaragua by Department Department Number of Gangs 'I'otal Members Managua 108 1627 Chinandega 15 12 Matagalpa 8 202 (iranada 4 120 Esteli 1 8 284 Jinotega 8 84 ! Total 161 2329 Source: Police of Nicaragua. Interview 12/01/99. 105. Some gangs are more violent than others. Less dangerous groups are normally unarmed, make street noise, and primarily enjoy disturbing the peace. According to the police, there are 9 such groups, 5 in Masaya and 4 in Nueva Segovia, with a total membership of 161. Box 14 - Adolescent's Problems As Defined by Adolescents Poverty is destroying young people. Their main problems are: * Unemployment * Drugs * Violence * Family disintegration * Undesired pregnancies * Suicides Source: Police of Nicaragua. Plan de Prevenci6n de las Pandillas. Intemal Document. 106. Gangs attract children and adolescents because Nicaraguan families are under social, economic and psychological stress, and the country lacks adequate sport, cultural and recreational centers for the youth and job opportunities for youth.. 107. Countryside violence (primarily cattle theft and burglary) exists in rural areas such as Santa Teresa, Achuapa and Santo Tomas (Pacific region), El Almendro, Nueva Guinea, Siuna, and Tasba Pain (Atlantic region). Only the Atlantic region has armed groups who burglarize houses and public transportation. These groups also kidnap tourists and natives for ransom. vii. Legal Insecurity 108. Legal insecurity also besets poor people. In the southern Pacific region where there is a growing animal husbandry industry, colonos (settlers) who do not work for these farms lack legal rights, titles, or certificates and could be evicted at any moment. B. COPING STRATEGIES 109. The poor use a wide range of strategies, both legal and illegal, to cope with vulnerability. There are no great differences in strategies across the country. i. Within the Legal Framework 110. Hungry periods. Households in extreme poverty go through hungry periods to feed their children. Annex 20, Page 28 111. Community support. Indigenous and mestizo peoples rely on family and neighbors for support. This includes taking care of children, the elderly and the sick; taking care of animals; tending agricultural plots; borrowing small amounts of money: exchanging food for food; obtaining credit at local stores; and bartering food for cash or labor. 112. Selling assets. Rural people sell livestock, poultry and grain during crises. Selling land and houses is a strategy of last resort for both rural and urban households. 113. Labor. The poor use family labor-especially women and children-to earn money. Women increasingly work in agriculture or sell food and handicrafts. Children work in agriculture and in the infornal sector. Girls sometimes work as maids while boys work as peons. Both genders sell lottery tickets or food. or work local mercados (markets). Miskitos have survived by alternating subsistence activities with wage labor, often in foreign or enclave enterprises. 114. Child work versus child education. There are plenty of references in all sites on high costs of education. Tuition and school supplies are expensive. These expenses along with uniforms, transportation, and meals make education unaffordable. There is an opportunity cost in terms of lost labor. For the poor, the economic returns of education are lowv. What is the difference between completed or uncompleted primary? What is the difference between completed primary and uncompleted secondary? Are they going to be better off? The generalized perception is that most educated children would still face problems finding a job. 115. Migration and Remittances. Migration is an important survival strategy. Both the poor and extremely poor (rural and urban) migrate often within Nicaragua - to Managua, other cities, or to regional harvest areas - and internationally to Costa Rica, Honduras, and El Salvador to follow the harvest. The majoritv migrate to Costa Rica, but Atlantic region residents prefer to migrate to the Caribbean isles and the United States. 116. Men and women, families and single people, indigenous and non-indigenous peoples, urban and rural people migrate. Poor people perceive migration as both positive and negative. On one hand, it increases family earnings, because family remittances are key income sources for Nicaraguans. On the other hand, migration separates and destabilizes families. Children and wives are often left behind, increasing emotional problems. 117. Damaging the Environment. Most rural and semi-urban communities cut trees for their own use and also sell them as firewood. ii. Outside the Legal Framework 118. Drugs. Selling drugs is an easy way to rapidly increase earnings. This strategy is evident in urban areas such as Managua and semi-urban areas such as Esteli. Team members also observed drug problems in most cities of the Atlantic region and in cities on the border of other countries. 119. Gangs. Lack of jobs drives young people to join gangs as a way to obtain cash. 120. Illegal use of basic services. In poor urban and in some rural areas, people illicitly connect to electricity and water (El Calvario, Bluefields, and Managua) and water systems (El Guanacaste). 121. Prostitution. Mostly in urban cities, female prostitution is cited as a survival strategy. Male, homosexual, and child prostitution is not cited in any area. PART IV: EMPOWERMENT OF STATE INSTITUTION AND LOCAL ORGANIZATIONS Annex 20. Page 29 A. GOVERNMENTAL INSTITUTIONS AND CORRUPTION "Many organizations have visited us. All claim that they, will do something to help us. Like you, they all ask questions, take notes and photograph us to show their bosses what good work they have done. But we never see the outcome. Nothing changes for us! All remains the same.'" Source: Woman. Nindiri. 122. This study's hypothesis is that corruption undermines development by weakening rule of law and institutional arrangements on which economic growth and development depend. The poor are the most affected by falling labor demand, increased unemployment, migration, and price fluctuation. Corruption is particularly damaging to the poor, who depend more on basic services and anti-povertv programs. A recent survey by the Institute for Nicaraguan Studies (IEN) reflects that 86.5 percent of the consulted people believe that corruption affects their well- being. 123. During the site visits, people's distrust of and apathy toward governmental institutions, authorities, political leaders, and the private sector were evident. Perception of corruption reduces public support in civil society, donor countries, and agencies for development assistance, and it discourages private investors. Poor people perceive that society's benefits are generally not available to them. They perceive that they lack access to justice. Most felt the legal system is distant from them, that they are powerless and voiceless. They are confronted by corrupt and unaccountable institutions that lack any transparency. 124. People perceived corruption at both local and national levels. Primarily they trust their own organizations, community-based organizations, and demand-driven commnunity-based projects that work closely' with them. 125. According to surveys by the Institute for Nicaraguan Studies, the Nicaraguan people believe that corruption is widespread. Ministers, political leaders and legislators have the worse reputation. The Nicaraguan Govemment has four branches: Executive, Legislative, Supreme Court of Justice, and Supreme Electoral Council. The Executive Branch has the worst image, with 88 percent of consulted people believing the Presidency is corrupt; 85.8 percent believe that the Congress is corrupt and 81.8 percent believe the Supreme Court of Justice is corrupt. The Supreme Electoral Council has a slightly better image, but over 80 percent of consulted people consider that it is also corrupt. Annex 20, Page 30 Table A20.11 - People's Perception of Corruption Institution Yes % No %/o Don't Know Presidency 88.4 3.4 8.2 l Congress 85.8 4.2 10.( Supreme Electorai Council 1.S 7. 11.2 | Supreme Court ot Justice 80. 6.5 | 12.7 Municipalities 84.3 7.0 8.7 Police 7.1 9.4 Cfourts : 82.5 T7.71 9. i Accounting ottice 80.7 | 8.1 11.2 Autonomous institutions 72.9 |1321 13.9 Schools 69.9 i 22V.1 8.0 Health centers 69.4 | 0.8 9.8 Army | 69.4_| 16.3 14.3 Source: Institute for Nicaraguan Studies (IEN). Survey on governance, transparence. accountability and integrity. Figure A20.2 - Do you believe that there is corruption? 1 JJJI1JJJJ'No DlNo EYes 126. The main manifestations of corruption are: (a) graft; (b) influence trafficking (defined as nepotism and the use of one's position to obtain commissions, privileges, and exemptions); and (c) formation of mafias. 127. High functionaries betray their corruption by their: (a) a sumptuous life style. (b) rapid acquisition of houses and goods, (c) high salaries in dollars, and (d) influence trafficking. Only 7.5 percent of the surveyed people believe that no action has been taken to resolve corruption, while 90.1 percent consider that some measures have been taken, but not enough to resolve the problem. Only 7.1 percent consider that corrupt public functionaries are tried transparently, while 34.3 percent believe most use their connections when they are tried, and the majority, 54.5 percent, say that public functionaries are never brought to trial. Box 15- People's Recommendations to Combat Corruption * Increasing control and audits * Appointing honest people * Increasing public denunciations of corruption. Increase Participation in Social Control and Accountability An overwhelming majority (92.5 percent) of respondents believe that citizens should participate in good governance. Of each 100 Nicaraguans, 85 are willing to participate in social control to ensure accountability and integrity. Annex 20, Page 31 128. Interviewees believe that corruption takes place in all public institutions. Municipalities have the worst image, with 84.3 percent of consulted people considering these institutions unaccountable and corrupt. The IEN survey does not examine why people believe this, but during fieldwork for the qualitative study, the population expressed doubts about government financial management and quality of auditing systems. Governmental institutions - mainly central institutions -- are viewed as remote and unconcerned. People do not fully believe that local governments work for the poor. However, local governments are viewed as being closer to their reality than is the central government. 129. The armv is considered less corrupt (69.4 percent) than the police, who are perceived as less accountable (83.5 percent). About 70 percent of the sampled population believe that there is corruption in schools and health centers. Unfortunately, the IEN survey does not explore why people believe that corruption is widespread in these centers, nor whether rural and urban consulted people share the same opinion. 130. Nationwide, 68.5 percent believe that private sector is also corrupt and 84 percent perceive widespread corruption throughout the country. 131. The lEN survev confirms information gathered in our study's fieldwork and explains whv people perceive that wealth and poverty are inherited. "Tle power of the rich has been transmitted tlhrough generations, and so has deprivation: We have been poorfor generations! There is a general perception that economic opportunities have been unevenly distributed. that wealth has always been concentrated among few families down through the country's history. and that it will be ever so. 132. When visiting poor and extremelv poor areas, it becomes evident that these groups feel trapped in the circle of poverty. Widespread distrust in institutions explains why people cannot visualize a way out of poverty. When people live under deplorable socioeconomic conditions, distrust the system, and believe that governmental institutions do not work for them, people live not under the poverty line but in deprivation. B. EXCLUSION 133. Consulted people say they are excluded from the benefits of society. As with the concept of poverty, people also have a broad and multidimensional view of exclusion. To be poor and to live under high levels of ill-being means to be excluded. The two concepts are closely related. When people talk about their socioeconomic condition, they use both concepts - along with vulnerabilitv - to describe their situation. Being excluded means to be affected by unemployment and economic crisis. It means to have an inefficient safety net and precarious social services. Legal and political exclusion coupled with the perception of widespread corruption cause people to question the existence of good govemance and the rule of law. 134. People distinguish many types of exclusion: Political. Complaints that the poor do not participate in decisions affecting their lives are widespread. In some communities such as El Cacao people said they are excluded by their political affiliation. If the Mayor and communities belong to different political parties, the likelihood of excluding communities increases. Legal. The poor cannot afford lawyers and do not trust that the legal system works for them Economic. People lack access to employment and lack the assets to own a small business. The perception that wealth has historically been concentrated in a few families is widespread, so that people claim that poverty as wealth is inherited. Financial. Lack of assets excludes the poor from the financial system. Annex 20, Page 32 Gender-Based. Women are aware that gender discrimination is widespread. Women are at risk of prostitution and sexual abuse. The structures to protect women do not reach poor and extremely poor communities. Age-Based. The elderly are excluded from pension and medical plans. There are no institutions to take care of elders in poor and extremely poor communities. Youth have no opportunity to grow into healthy human beings. The risk of dropping out of school and joining gangs is very high. The majority of youth believes that the future is hopeless. Child labor abounds. There are no structures to take care of mentally or physically handicapped children in poor and extremely poor areas. Ethnic. Indigenous peoples feel excluded from the benefits of the mestizo society and see threats to their ancestral lands and cultural heritage. Indigenous peoples from the Atlantic region and from Central (Sabana Grande) and Pacific regions (Le6n) feel excluded from political and economic decisions. Basic services. Limited access to basic services - education, health, and technical assis- tance - abounds in poor and extremely poor neighborhoods. Insecurity. Neglect by police and civil security mechanisms is widely perceived. 135. There are five mechanisms of exclusion: 1. Information and knowledge. Limited capacity to make decisions and to know where to go or what to do in cases of emergency. 2. Assets. Exclusion from economic and social benefits of society. 3. Decision-making. Limited participation in political and economic decisions affecting their lives 4. Segregation. Exclusion from the legal system. 5. Basic Services. Limited access or lack of access to basic services. C. STRENGTHENING LOCAL ORGANIZATIONS FOLLOWING AGE AND GENDER ROLES 136. Evidence captured through fieldwork shows that local consulted people trust their own organizations and community-based organizations more than they trust governmental institutions. Two main reasons account for this. Local people are more closely aligned with local than with national government, and they perceive that governmental institutions are corrupt and unaccountable. These elements increase distrust. 137. Communities' customary organizational practices, age and gender division of labor determine the ability to build local social capital. Women do not normally have decision-making positions in community-based organizations. However, during the Hurricane Mitch crisis, women were very active in organizing food and water storage and evacuation strategies. Specific research is needed to determine if these organizations still exist or disappeared after the crisis. 138. Empowerment, social capital development, and strengthening of local human capital are related. Strengthening organizational capabilities and entrepreneurial practices promotes strategic alliances among farmers to reduce transaction costs and cope with vulnerability. 139. To achieve these objectives, the Consulted people make the following recommendations: Annex 20, Page 33 * Sociopolitical framework and good governance: * Strengthen communities to be active and accountable members of civil society (political participation). * Participate in decentralization and fiscal control. * Enforce law and order (civil security). * Target tenure rights. * Target access to common resources. * Target access to financial resources and assets (irrigation systems, seed and food storage, and commercialization). * Sector investment (public and private): * Reinforce local networks to ensure effectiveness of multi-grade educational svstem by making parents accountable and responsible partners. * Increase access to traditional or official social safety nets. * Increase and strengthen participation of water and electricity users. * Increase household's participation in improved access to health services, nutritional and familv planning programs. * Increase household's participation in building paved roads and latrines. * Entrepreneurial development * Increase household's productivity of patio economy. * Strengthen women's productive and administrative capabilities. * Strengthen farmers organizational capabilities. 140. Consulted people want to empower govemmental institutions, local organizations, and th-e private sector to work on their behalf Poverty is widespread in rural areas. The main asset available in these areas is the labor force, but there are no job opportunities. In the urban areas, people perceive that unemployment is likely despite job vacancies because these jobs require highly specialized workers. 141. In rural areas, people recommend the following measures: * Governmental institutions: * Enforce rule of law and civil order. * Enforce public trust and confidence in the judicial system. * Punish corruption at all levels. * Private sector * Create employment opportunities by favoring private investment in agro-industry and services. * Enforce participation of grassroots organizations and NGOs to restructure safety net pro- grams and provide technical assistance, food, and nutritional and water programs. * Population: * Increase people's participation in decision-making. * Increase people's participation in building local infrastructure such as roads and water and sanitation systems. Annex 20. Page 34 PART V: ACCESS TO BASIC SERVICES 142. The objective of this chapter is to capture and examine people's view of access to basic services. To consulted people, population disparities in opportunities and gaps in basic services cause and perpetuate poverty. The Nicaraguan poor are aware that access to basic services is fundamental, but also are aware that building infrastructure is not enough if buildings lack qualified staff and equipment. A. HEALTH 143. Gaps between poor and non-poor result from disparity of opportunities. These include location of health posts and centers, schedules, and quality of personnel. In most regions, distance limits people's access to health services. According to LSMS, the average distance to a health post or center is 2.4 kilometers. The survey corroborates findings of the qualitative study. For instance, in the rural areas of the Central region, the average distance is 4.9 kilometers. Even worse, the rural Atlantic region centers and posts are an average of 6.6 kilometers away from beneficiaries. 144. Taking into account road conditions and limited public transportation, such distances are crucial in an emergency. Other people say that distance is not as important as inconvenient schedules and low quality of physical and human assets. "I am willing to walk if health centers are open and have doctors and medicine." Source: Poor woman. Central Region. 145. Schedules limit people's use of services. Most health centers in rural communities are closed evenings and weekends. Poor physical assets directly affect the quality of services. Most people complain that rural centers lack equipment and medicines. Basic equipment is useful for curative, not preventive, medicine. Some centers even charge for syringes and bandages. Although prices might be lower than drug stores, such expenses are a burden to the poor and extremely poor. Centers provide prescriptions which people take to drug stores, but poor people's incomes are normally not high enough to buy medicines. Box 16 - Limited Human Assets Restrict Effective use of Physical Assets Wasakin, an indigenous community located in the Atlantic region, has only one health post built by the Emergency Social Investment Fund (FISE). The building is in good shape but lacks qualified staff. Tasba Pain, also an indigenous community in the Atlantic region, has a health post built by an American organization. The post has been closed for two years. It lacks qualified staff and medicines and so community members use it as a jail. 146. Inappropriate human assets create health gaps between poor and non-poor. Not all physicians like to work in remote rural communities eaming low salaries. Most people complain that the health post's staff are rude and that their professional skills are inappropriate for rural areas. Indigenous peoples say that most of the staff from western regions, culturally insensitive and disdainful of traditional medicine. Indigenous peoples questioned the way in which physicians are assigned to health posts. They wonder why indigenous peoples who study medicine are sent to other regions, instead of sending them back to work with their people. Annex 20, Page 35 Box 17 - A Prejudiced View of Reproductive Health Opinion is prejudiced against family planning. Men do not want women to expose themselves to a physician, especiallv a male physician. Some men are too shy to purchase condoms. They are not particularly sympathetic to vasectomy and do not know that it can be reversed. 147. Gaps also result from household conditions. Disparities in knowledge and information preclude the poor from adopting better behavior such as sanitation habits that improve househoid health. Most communities, and particularly in rural areas, know little about reproductive health care. 148. Lack of confidence in services often prevents people from using health centers. Box 18 - Lack of Confidence in Health Services "As sooIn as we find out that we are once again pregnant we must start saving to pay for private health services. It is a matter of taking care of your own life and that of the child." Source: Poor urban woman. El Calvario, Esteli. 149. Access to health care and services contribute to well-being, labor force performance, and productivity rates. Across the country, health services were appraised as unavailable, inappropriate, or inefficient. Disparities in the distribution of services reflect regional inequalities. 150. Of all interviewed communities, 41 percent appraise health services as bad and 27 percent do not have services at all. 9 In the Atlantic region, 57 percent of the interviewed communities did not have such services, while in the Pacific all the visited sites have access to health care. However, 75 percent of people appraised them as bad. In the Central region 29 percent of the people considered that health care is good. Table A20.12 - Access and Quality of Health Services Pacific Central Atlantic Not available - 29 7 Bad 75 28 Medium 13 14 (ood 12 29 14 I otal 10010 100 Source: Commr unity and Regional Reports. 9 To classify the services three variables were used; distance, physical assets, and human assets. The "bad" category means that services are located more than 2 km away from the consulted community, "medium,' that they are less than 2 km away, and "good" that they are located in the community. Physical assets means the services are well equipped - medical equipment in the case of health,and educational materials in the case of education. Human assets means that according to the people schools and health centers have staff trained to provide services for the poor. Annex 20, Page 36 151. The scenario is even worse for rural communities, where 38 percent do not have health services. Only 13 percent of all communities had health services. The Atlantic region faces the worse conditions since 80 percent of the visited sites lacked health centers. In the Pacific region, 83 percent had bad health service, and in the Central region, 40 percent had bad health services. B. EDUCATION 152. Poor people are aware of the relationship between low quality of education and poverty. People believe that there are disparities in access to education and in quality of education. 153. Disparities in opportunities such as availability, location of schools, and human assets increase gaps between the poor and non-poor. Not all rural communities have schools. Not all the schools teach the entire primary curriculum, although most extend through fourth grade. Most of the rural teachers are empiricos. The majority follows the multi-grade (multigrado) system, meaning that one teacher simultaneously instructs all grades in one single classroom. Box 19 - Disparities in Education Of the 16 consulted rural communities, only two offer secondary education, and only 1. in Wasakin (Atlantic region) has full secondary education. Most urban sites have primary and secondary schools. 154. In most regions, distance limits people's access to education and affects dropout rates. According to the LSMS. on average elementary schools are 0.9 kilometers from sites, and the principal access road is paved in only 22.3 percent of the communities. Poor physical assets directly affect educational quality. With the exception of schools built by the Emergency Social Investment Fund (FISE) most schools are in bad shape and lack educational materials and appropriate furniture. 10 155. Inappropriate human assets increase educational gaps between poor and non-poor. Low salaries discourage teachers from working in rural areas. With the exception of indigenous peoples, most people feel that teachers are polite. Indigenous peoples feel that teachers from western regions disdain the indigenous culture and are unwilling to learn about customary educational methods. Bilingual education does not exist, particularly in areas where indigenous peoples live. Finally, not all teachers have appropriate skills to teach in rural areas. 156. Parents are aware of teacher's overwhelming workload. They also perceive that the multi- grade system does not encourage learning capabilities. 157. Disparities in financial opportunities and gaps in education, knowledge, and information between poor and non-poor parents limit poor households from participating in educational decisions. Box 20 - Voluntary Fee is Compulsory and is a Financial Burden for the Poor Schools operating autonomously request a monthly contribution of 5 to 10 C6rdobas (US$0.40 - 0.80) per student. Some parents complain that if they do not pay, their children are expelled. 10 People were asked: Is the school in bad or good shape? Who built the school? Annex 20, Page 37 158. For school autonomy, the parent's financial contribution is indispensable, but it is also overwhelming for parents with three or more children. School materials, uniforms, shoes, and food make it extremely difficult for poor rural household to afford education. Lack of scholarships and financial aid increase educational disparities in opportunities between poor and non-poor households. 159. Parents' poverty and educational level limit their abilitv to be reliable partners in autonomous schools, limiting the model to be participatorv and accountable. Parents are unaware of their role and responsibilities as partners in decision-making and claim that teachers do not know their own role and parents' role in such a process. 160. Nine percent of interviewed communities did not have any schools. Only 45 percent of all interviewed areas have primary education up to fourth grade and 5 percent have completed primary schools. Only 14 percent have full primary and 18 percent full secondary education. In terms of water services, 5 percent of interviewed communities did not have water, 36 percen: appraised services as bad and 41 percent as good. 161. All the visited sites in the Pacific region have schools, but 63 percent of the sites have primary education only up to fourth grade. In the Atlantic region 29 percent of sites have no schools, 57 percent have primary education only up to the fourth grade, and 14 percent have full primary. 162. In all rural communities, 13 percent do not have schools and 63 percent have multi-gradc primary only up to the fourth grade. In the Atlantic region, 40 percent of the visited sites do not have schools and only 20 percent have multi-grade primary up to fourth grade. All visited sites in the other two regions had schools at least up to the fourth grade. C. WATER, SANITATION AND ELECTRICITY 163. Limited water and sanitation services increase the poor household's health vulnerability. Only urban communities have piped water systems. According to the LSMS, 9.1 percent of households in Nicaragua-mostly rural sites-use water from rivers. This practice is more common in the rural areas of Atlantic coast (48.3 percent) and Central region (23.6 percent). Only 27 percent of all households have indoor pipes. In the rural areas of the Atlantic region, only 1.8 percent of the households have pipes inside. Public or private wells predominate in the rural areas of Pacific (40.6 percent), Central (28.9 percent), and Atlantic (48.3 percent) regions and in urban areas of the Atlantic region (44.4 percent). 164. People in rural areas claim that their water is unsafe. LSMS confirms these findings. In all of Nicaragua, according to the survey, only 16.4 percent of the households have unsafe water, but in rural communities 40.5 percent of the extremely poor and 38.1 percent of the poor live without safe water. When asked what they do to purify water, people do not know basic treatments such as boiling water. 165. Most of urban poor have latrines, whereas the majority of rural sites lack them, particularly extremely poor. None of the rural communities has sewage systems. Annex 20, Page 38 166. Only 9 percent of the total consulted sites did not have any water, sanitation, or electricity. According to the survey, 15.9 percent of the households lack latrines and 31.3 percent lack elec- tricity. Only 32 percent of all sites have water and sanitation and 55 percent have electricity. In the Atlantic region 29 percent of the visited areas did not have any of the services. LSMS findings coincide with qualitative work: 89.8 percent of Atlantic rural sites lack electricity, 53.2 percent lack latrines, and about 50 percent lack safe water. Qualitative findings state that in this region 14 percent of the interviewed sites have only water. In contrast, in the Pacific, 75 percent of the interviewed areas have all three services. The intermediate scenario was found in the Central region, where 43 percent of the sites have all the services and 57 percent have water and sanitation. 167. Twenty percent of the studied rural communities in the Atlantic region do not have water. In the other regions, all the visited sites have some type of water service. On average, 64 percent of the rural sites in the Central and Pacific regions have good water and 26 percent have bad water service. 168. Only 38 percent of all studied rural communities have water, sanitation, and electricity while 13 percent have none of these services. The Atlantic region again fares the worst since 40 percent of visited communities have none of these services and only 20 percent have all three services. 169. In general, people perceive that water service is expensive and irregular. In communities such as El Guanacaste, the service was installed, but the community service was suspended for more than three months and had to bring water from a public well located in the nearest community. People claim that although water does not run, they are billed for the service every month and INAA(the water company) does not acknowledge receipt of their complaints. Concerning electrical service, electrical outages occur frequently in Nicaragua, particularly during the summer, even though they are billed for the service. D. ROADS AND MEANS OF PUBLIC TRANSPORTATION 170. Disparities in physical infrastructure and means of public transportation increase poor households' vulnerability by limiting trade, commerce, and access to education and health services. 171. Urban roads are in reasonably good shape. The urban poor rely on public transportation, mainly buses. In Nicaragua, services of public transportation such as buses are called rutas (routes), whose service can be erratic. People say that most are in bad condition. Pickpockets rob passengers. Independent operators use trucks with no seats as means of public transportation, which are obviously unsafe. 172. In most rural communities, roads and paths are generally in bad shape and get even worse during the rainy season. This increases the length of trips. Travel between the Atlantic and the other regions is done mostly by airplanes; the few available roads are dangerous either because they are in bad shape or because armed groups assault automobiles. 173. Thirteen of the 16 interviewed rural communities have public transportation, including rutas. In rural communities, these services are sporadic. Some work twice a day (in mornings and/or evenings), whereas others operate every two days or twice a week, such as in indigenous communities. Annex 20, Page 39 174. Only 18 percent of the interviewed communities considered that rutas provide good service, while 50 percent considered the services bad. 1 The Atlantic region has the worst conditions since 86 percent of the communities considered services bad. In the Pacific region, 63 percent think service is good, while in the Central region 57 percent think so. 175. Roads are indispensable to economic and social development. Data shows that there are great differences among regions. The Atlantic region has the fewest roads. Because the Pacific region has the highest concentration of economic and political activity and the highest population density, it has a high concentration of roads. This facilitates development and access to basic services. 11 Good service means that at least there are two rutas (round trip) for day. Annex 20, Page 40 PART VI: TECHNICAL ASSISTANCE 176. From 1993 to 1998, technical assistance was the principal instrument by which the Nicaraguan government supported small and medium farmers. Findings show that technical assistance is often inappropriate because it is not permanent and does not seek to change productive patterns and increase beneficiaries' social capital. According to LSMS, 24 percent of farm households have access to technical assistance and 15 percent actually use technical assistance. Of those who use this assistance, 54 percent received it from the government, whereas 41 received it from NGOs. Only 3 percent of the surveyed households say they paid for technical assistance (Benjamin Davis). 177. Only three of the 22 visited areas had farmers who say they received technical assistance. The assistance is not fully reaching small and medium farmers, who cannot afford to pay for such services. The main way in which farmers can get free technical assistance is by forming groups that demand it. This study revealed that the poor do not know how to organize for this purpose. Moreover, as noted in social capital), small and medium size farmers show little inclination to form groups. 1 78. The Nicaraguan Institute for Technical Assistance (INTA) has three technical assistance programs for farmers, but two (ATPI and ATP2) are intended mainly for medium and large farmers. The ATPM is designed for small farmers, but it does not directly assist them. Rather, this program trains community leaders who in turn train farmers. Leaders and community farmers must request training from an INTA agricultural extension officer, who must commit himself to work with the group. Box 21 - Technical Assistance Programs (INTA) The Participatory Technical Assistance I (ATP1) program works with producers who have land, inputs, machinery, and labor force. Co-financing ranges from 20 to 30 percent, depending on the farmer's financial capabilities. Co-financing also increases according to the length of time the farmer wants to receive technical assistance. The Participatory Technical Assistance 2 (ATP2) program is an integrated program that establishes consulting firms provide technical assistance-to farmers. The government subsidizes up to 70 percent of the consulting firms' costs. Participatory Massive Technical Assistance (ATPM) is a free program targeted to small farmers. It favors marginal zones, plots near rivers, hills and agricultural frontiers. A demand-driven activity, it must be requested by community leaders. INTA provides one agricultural extension officer to train ten leaders. Each leader then trains 50 small farmers. Source: Alfredo Betanco. INTA. Interviewed by Mary Lisbeth Gonzalez. 12/01/99. INTA 179. The ATPM program requires beneficiaries to be familiar with the program and its administrative procedures. It also requires significant community organization. According to our study, the poor normally do not know how these organizations work. 180. The study confirms that technical assistance changes traditional patterns of crops and increases productivity. Community-based projects by NGOs with a permanent community presence foster organizational efficiency, higher productivity, and change in household nutritional habits. Nindiri, Santa Lucia and Mojellones are good examples of such projects. Annex 20, Page 41 Box 22 - Nindiri Case Study Poor farmers have been helped by an integrated project of the Nicaraguan Institute of Human Promotion (INPRHU). The project aimed to change productive patterns of the farmers, who own one hectare of land, by moving them from single-crop production to a mix of fruits and vegetables. Another objective is to change nutritional patterns through this diversification and to promote a variety of products throughout the year. The project's main components are (a) Soil conservation and appropriate land use, (b) Production diversification, (c) Financing and repayment in kind (seeds and plants), which reduces the stress of indebtedness, (d) Commercialization and market identification, and (e) Organic production. This project changed production patterns, improved land use, introduced commercialization into the production cycle, and changed nutritional habits. Among the lessons learned: * Long-term technical assistance (two years) helps sustainability. * Loans in kind do not cause debt crisis or stress. * Surplus is an economic objective. * Diversified production changes nutritional habits and improves land use and productivity. * The permanent presence of agricultural extension officer helps sustainability. * Support is needed to encourage subsistence farmers to grow and sell surpluses. 181. If technical assistance follows an integrated development strategy, it can strengthen organization building and social capital. These two elements can eventually strengthen human capital (as shown in Nindiri) by changing production patterns, increasing food production, and modifying nutritional habits. 182. Findings of the qualitative work corroborate the LSMS's conclusions that technical assistance, credit, and organization are closely related. Participation in producer organizations seems to be a channel for credit and technical assistance. However, such organizations exclude farmers who are less organized and have less social capital. This is important to take into account because the LSMS shows that households that receive services from any agrarian institution have higher well-being than those who do not. Therefore, development of social capital leads to greater well-being, which develops greater human capital. 183. Empowerment of Local and Community Organizations: Limited access to informaticon and knowledge restricts people from doing things differently. Production is based on grains and starchy root crops. Most poor and extremely poor households produce only enough for their own consumption and have no surplus to sell. Through permanent technical assistance and participa- tory methods, beneficiaries develop of ownership. There is a direct correlation between owner- ship and project sustainability. Through good technical assistance, farmers also become socially and financially accountable and thus strengthen local social capital. Annex 20, Page 42 184. Any program or project aimed at promoting technical assistance, strengthening farmer's organizations, or building their capacity must take into account gender customary division of labor. Men and boys are generally in charge of agriculture although women work in both the patio economy and in the production field. In indigenous communities, women prepare the soil, harvest, and sell the produce. In both indigenous and non-indigenous communities, the women play an important role during the post-harvest season, when they take charge of cleaning produce (corn, rice, or beans) to prepare it for consumption or sale. 185. Cultural sensitivity is needed when providing technical assistance in ethnically and culturally diverse regions. 186. Rural Micro-Credit. Rural micro-credit systems, if accompanied by sustained technical assistance, tend to foster product diversification and strengthen social capital. Membership qualifications, terms of repayment, and quality of technical assistance are among the main problems pointed out by rural communities. In most rural credit systems the debt is collective, not individual. The risk of indebtedness is high because agricultural activities are by nature riskyv and because success of the system depends on the honesty of participating members. Payment arrangements often do not take agricultural calendars into consideration. Technical assistance is often inappropriate because it is not permanent and does not always seek to change productive patterns and increase beneficiaries social capital. 187. Technical assistance is often expensive in most rural micro-credit projects because beneficiaries have to pay for administrative financial services. One of the greatest weaknesses of these programs is that farmers have no control over financial companies or funds administrators. which often use funds more as financial leverage than to benefit farmers. Consequently, some rural credit systems, rather than capitalizing beneficiaries, drag them into a cycle of debt that irremediably perpetuates poverty. 188. According to the study, the government's technical assistance does not extend to the entire country. In the Central region, a variety of organizations provide technical assistance, but only one of the visited communities has had assistance from government programs (INTA and MAGFOR). In the Pacific region, INTA has been present sporadically and most sites visited by this study have been assisted by NGOs. In the Atlantic region, none of the visited sites has received any technical assistance. 189. Technical Assistance Model: Informal Education to Prepare Farmers for Entrepreneurial Development, as Defined by Farmers. Technical assistance and training, key components of informal education, are adequate mechanisms to answer modernization needs of agricultural sector. Farmers are the target population, butyouth should also be included in the training, as should members of the household other than household heads. 190. Farmers say they need three different types of services: (a) Traditional training,which involves seminars and workshops to provide general information on: (i) entrepreneurial, managerial and organizational skills; (ii) animal husbandry production cycle; (iii) veterinary care; (iv) agricultural production cycle. The objective of traditional training is to contribute to eliminate information and knowledge gaps. (b) Technical assistance: To ensure effectiveness technical assistance services need to combine short seminars and workshops (traditional training) and applied and demonstrative assistance to increase multiplying factors. (c) Specialized Informal Education: To achieve cultural change in growing animal husbandry and agriculture, three specialized non-formal educational programs are recommended in the table below: Annex 20, Page 43 Table A20.13 - Technical Assistance Model. Informal Education to Prepare Farmers for Development T vpe of Program Beneficiaries Content Entrepreneurial and I Agricultural and animal Administration Management husbandry growing farmers Farm technical planing Farms technical and financial analysis. o Animal Husbandry Animal husbandry growing Soil management o farmers. Irrigation systems. Animal nutrition Sickness Storage systems Commercialization and market analysis Milk and meat quality control analysis. Agriculture Agricultural tarmers Soil management Irrigation systems Plagues and risk control management Storage systems Production quality control analysis Commercialization and market analysis. 191. Scale Economies. Groups of farmers acting together create economies of scale. Technical assistance can provide them the training and strengthen social capital. The chief goal, how- ever, is to develop entrepreneurial activities. Small private activities avoid donors' clientism and beneficiary's dependence on governmental institutions or donors. 192. Patio Economy. In all rural communities, the patio economy is considered important in alleviating depravation. Traditional training and technical assistance can make patio economy more productive while promoting cultural change. These informal education methods must aim at teaching diversified production, animal nutrition, hygiene, water care and family nutrition. Our findings lead us to the hypothesis that the patio economy allows households in rural areas lo have better nutrition than urban ones. Female participation in the patio economy allows her to be close to her family. In contrast, women in urban areas working outside home in informal jobs have to leave their children alone or take them along to the streets. Urban households have no patio economy (as an asset), so the cost of women joining the informal labor sector (transportation and food) are higher than their rural counterparts. All these factors reduce urban households' food budget and reduces the time available for cooking and nutrition. Annex 21, Page I Annex 21 - Public Social Sector Spending and Analysis of Programs by Ema Budinich B. PART A. TRENDS IN CENTRAL GOVERNMENT SPENDING FROM 1991 TO 2000 1. This chapter contains an analysis of the Central Government Spending over the last I 0 years. A review is made of the tendency for total spending, its relation to the GDP, the per capita spending levels, and the share of the total for the different sectors, as well as the weight debt service has had during this period. 2. Following the official classification, a review is made of the evolution of spending on the social sectors and especially the tendency of this in recent years. This background serves a complement to the evaluation of social spending by program presented in Chapter 2. 3. In the analysis, use is made of the information about public spending prepared by the World Bank in Nicaragua, with the background about budgetary execution prepared by the Ministry of the Treasury. For the last two years of the series, data from the Budget are used because it is still being executed and is not available. i. Trends in total Central Government spending 4. The Total Spending of the Central Government, expressed as a percentage of the GDP, has shown a tendency to increase during the 1990s. In 1991, total spending amounted to 28.7% of the GDP, rising to 34.5% of the GDP in 1994. This is explained by the major increase in foreign debt service, tripling in three years from 2.9% of the GDP in 1991 to 9.6% of the GDP in 1994. 5. As of 1994, two clear periods were observed. From 1994 to 1997, debt service remained between 7% and 9% of the GDP, and starting that year, the tendency changed and it was reduced to 5% of the GDP in 2000, but without returning yet to the level it was at the start of the decade. Figure A21.1 - Trends in Central Government Spending (as % of the GDP) 40% 35%7 . 30% ! / *-*~-iii Total spending 30% O__ _lth -- - -41 -41- 36 39 39 39- --37--- 37 34 -36 Ho&ing - 0 2 3 0 0 Wkatar Sar itafiriVc 2 2 1 2 2 3 2 13 Scal Proectiond- -0 ° 12 10 9 8 6 4 T 4 Local Dew1jiru/e I 2 2 4r 6 6 4 5 8 8 Oitf~ 6 5 2 I 1 1 1 II 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 as %oftheGDP)1 SOCIAL SERVICES 10_3% l0%_ 109% 11_9%_11-5% 11.0 108% 100% 16.3% _1_54_ Educ lini 5.1% 5. 48 5. 4 55% 4.9% 4.8% 5% 5.--.1-% -6.3% 6.5° Hlaith 4.2 42% 3.9% 4.7% 4.5% 4.3% 4.0%/o 37-/o 5.6% 5.50 HocEirE 0.0% Off,o 0.3% 0.0%o 0.1W0% 0.3% 0. 4 lff/or 1.0% 0.9% 1.8%/o 1.7% 2.1% 2.0% 1.6% 1.2% 4.1% 3.0° 1991 1992 1993 1994 1995 1996 1997 1998 1999 200(i (in nill. US_) SOCLAL SERVICES(a 170.4 1848 196t3 217.7 218.3 217.4 219.1 212.6 376.46 388.2 Ionl ib 84.4 91.9 86.3 100.2 92.8 93.9 106.5 108.0 144.47 163.8 Health 69.8. 76.6 71.3' 85.7, 85.6 84.7 81.1 78.3 129.56 138.7 H)usi 0.3 0.3 6.0 0.1 0.1 6.78 10.i Water and Snirtatiorvc 3.4 3.9 1.9 3.6 4.4 6.2 5.0 48.52 9. Social Prcectiid 0.1 0.2 23.4 21.4 19.47 18.3 13.6 8.3 13.61 29.i0 LDcal Dveloprert/e 20 2.9 3.0 8.7 13.8 14.1- 9.7 11.0 30.63 31.4 rs/f 10.5 8.9 4.5 1.8 T. 1 19 2.-- Z0 2.1 2.89 5.5 1991 1992 1993 1994 1995 1996 1997 1998 1999 2001 (n UtSS per capita) SOCALSERVICESta 44.2 46.5 47.9 51.5 51).1 48.4 47.3 44.5 76.47 76.5 ltic~non/b 21.9 23.1 21.11 23.7 21.3. 20.9 23.0 22.6 29.35 323 1-alth 18.1 19.3 17.4 20.3 19.7 18.9 17.5 16.4 26.32 27.I 1HomiTn 0.1 0.1 1.5 0.0 0.0 1.38 ateamlSarnitaticFVc 0.9 1.0 0.5 0.8 1.0 1.3 1.1 9.86 1. Social Protectior/d 0.0 0.1 5.7 5.1 4.4 4.1 Z9 1.7 2.76 5. 7 -LlDevelorn/e 0.5 0.7 0.7 21 3.2 3.1 21 2.3 6.22 6.2 O-t1xs'f 27 221 1.1 0.4 0.7 0.4 0.4 0.4 0.59 1.I Sore: Dita Base of Wbrld B* Nicaragia, lbaed ni official inoniin fromn te Mrasirv of tde Treastrv Exqedis exoauted 1991-1998. E sidirtt Brdgefedor 1999-2000. NUMS' a/ inciluies the FSS simn 1998onm Edratim ad 1th and in 2)00, also c the MFAM and thr FISE. b/ iJcudas transfe-s to the uwiinsiti INATEC,INIEA, INAP, 1N; the R& nDleino Tlr and te huitsue ir Youth md Spxt cl irr lus catnrihtns from the Treasay to ENACAL for imvesnt pojcs. OrI nEas a fraticm of puiic uTwstnmt in the area d/ inlud SAS, MFAN and th Emxwnzy Social nmtn Fund (FISE) e! inrlues FISE, INIFCM and ccrtribiis to th Lcai Ckvaiuzts f? Oire Jcuds: MTRAB Particular Expx1ihr, ISS. and cenral gvnit trasfers to private irsatinurs. For 2000 these are for tIe Red Crcs, the Ni'ncau Couiacil to Figli Dug, and the Officl of the Hmn Rigits Onixudmn Annex 21, Page 46 b. Social Service spending by type of expenditure (in mill. US$) 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 SOCIAL SERVICES 170.4 184.8 196.3 217.7 218.3 217.4 219.1 212.6 376.5 388.2 Current expenditures 159.7 1688 165.4 155.3 145.3 141.2 146.7 166.6 231.0 223.6 Capital expenditures 10.7 15.9 30.9 62 4 73.1 76.2 72 4 46.0 145.5 164.6 Education /a 84.4 91.9 86.3 100.2 92 8 93.9 106.5 108.0 144.5 163 8 Current expenditures 79 7 85.4 82.3 80.6 75 3 73.3 75.7 86.7 109.1 110 6 Capital expenditures 4 7 6.6 4 0 19.6 17.5 20.6 30.9 21.3 35.4 53 2 Health 69.8 76.6 71.3 85.7 85 6 84 7 81.1 78.3 129.6 138.7 Current expenditures 68.5 73 0 67 8 67.8 65 5 63.5 66.0 74.2 107.5 97.5 Capital expenditures 1.3 3.6 3.5 17.9 20 1 21.2 15.0 4.1 22.0 41.1 Housing 0.3 03 6.0 0.1 0 1 6.8 10.0 Current expenditures 0.3 0.3 6 0 Capital expendtures 0.1 0.1 6 8 100 Remainder/b 15.9 16.0 32 8 31.8 39.9 387 31 5 26.4 95.7 757 Current expenditures 113 10.2 9 3 6 9 4.5 4.4 50 5.8 14.3 15.5 Capital expenditures 4.7 5 8 23 4 24 9 35 4 34.4 26.5 20.6 81.3 602 (in US$ per capita) 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 SOCtAL SERVICES 44.2 46.5 47.9 51.5 50.1 48.4 47.3 44.5 76.5 76.5 Current expenditures 41 4 42.5 40.4 36.8 33.3 31 4 31 7 34.9 46 9 44 i Capital expenditures 2.8, 4.0 7.5 14.8 16.8 17.0 15.6 9.6 296 32.4 Education /a 21.9 23.1, 21.1 23.7 21.3 20.9 23 0 22.6 29.3 32.3 Current expenditures 20.7. 21.5 20.1 19.1 17.3 16.3 163 18.2 22.2 21.8 Capital expenditures 1.2 1.7, 1.0 4.6 4.0 4.6 6.7 4.5 7.2 10.5 Health 18.1 19.3 17 4 20.3 19.7 18.9 17.5 16(. 4 26.3 27.3 Current expenditures 17.8 18.4 16.5 16.0 15.0 14.1 14.3 15.5' 21.8 19.2 Capital expenditures 0 3 0.9 09 4.2 4 6 4 7 3.2 0.9 4 5 8.1 Housing 0 1 0.1 1.5 0 0 0.0 1 4 2.0 Current expenditures 0.1 0. 1 1.5 Capital expenditures 0 0 0 0 1.4 2.0 Remainder /b 4.1, 4.0, 8.0 7.5 9.2, 8 6 6.8 5.5 19.4 14.9 Current expenditures 2.9. 2.6. 2.3 1 6 I10, 1.0, 11 1.2 2.9 3.1 Capital expenditures 1.2 1.4 5.7 5.9 8.1 7.6 5.7 4.3 16.5 11.9 Source: Data Base of World Bank Nicaragua. based on official infornation from the Ministry of the Treasury Expendiures executed 1991-1998. Expenditures Budgeted for 1999-2000. a' Besides the MECD, includes transfers to the universities, INATEC,INTECNA, INAP,INC,the Ruben Dario Theater and the Institute for Youth and Sport. b Includes: SAS, MITRAB, MIFAN, and transfers to INIM, FONIF, FISE, FSE, INSS, ENACAL, and other public and private entities that carry out social actions. - - -- I I - -- - - MEMORANDUM: 91 92 93 94 95 96 1 97 98 ' 99 2000 Tipode cambiopromedio[C$/1 4.50 5.00 6.12 6.72 7.53 8.44' 9.451 10.581 11,81 12.68 1ndicePreciosConsumidor[l99 62.32 77.07 92.79 100.00 110.94 123.83! 135.241 152.89 174.14 191.73 PtB nominal imill. CS] 7,424.6 9,217.2 11,053.1 12,310.61 14,246.8 16,649.2, 19,115.8 22,483.3 27,281.0 31,879.3 PIB nominal [mill. USS] 1 1,649.91 1,843.4| 1,806.1 1,831.3' 1,892.1' 1,973.7 2,023.2 2,124.7, 2,310.0 2,513.3 Poblacion inuiles personas] 3.853.9. 3,974.0 4,097.8 4, 225 5 4 356.9 4,492.0 4,631.2 4,774.8! 4,922.8 5.075 4 Annex 21, Page 47 c. Gato en Fwadi6n y Salud, cxipoicid6n y tendercia. -- -~ -- 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 EDIXCAWON (In ri11i- US$) _ _ _ _ _ Total spendirig 84.4 91.9 86.3 1(0.2 92.8 93.9 106.5 1080 144.5 163.8 Oirrenxpeiq-itrees 797 85.4 82.3 80.6 75.3 73.3 75.7 86.7 109.1 110.6 Capital iexntuures 4 7 6.6 4.0 19.6 17.5 =20.6 30.9 21.3 35.4 53.2 (7oage compaiticn) 1991 1992 1993 1994 1995 1996 1997 1998 1999 20(0 On tex itresn 94 93 95 80 81 78 71 80 76 6E' Capital exienture 6 7 5 20 19 22 29 20 24 32 Total spaMing 1OD 1(X) 1O( 1OD 100 100 1O0 1(0) 10X 10) I1rex 1991=1 - _ _ oa] spending 100 109 102 119 110 111 126 128 171 19p waTent expendicuresi(X) -107 103 101 95 92 95 109 137 139 capital expelitures 1(X 139 85 415 370 436 654 451 750 1,12- HEALTH- (in nill.US$) Taal speMxibg 69.8 76.6 71.3 85.7 85.6 84.7 81.1 78.3 129.6 138 7 Onrent expenclitzes 68.5 73.0 67.8 67.8 65.5 63.5 66.0 74.2 107.5 97 5 Capital expaenitures 1.3 3.6 3.5 17.9 20.1 21.2 15.0 4.1 22.0 41 1 (7.age ccmpsition) Totalspeling O 1030 1D 100 100X 100 1(o0 10) 1() 10) Chieii eenitnes 98 95 95 79 76 75 81 95 83 70 Capital expieriues 2 5 5 21 24 25 19 S 17 3) J,tex 1991=103 ttal speriing 100 110 102 123 123 121 116 112 -4186 193 current expediures 100 107 99 99 96 93 96 108 157 142 capital expenitues 100 283 273 1,399 1,575 1,66D 1,175 322 1,721 3,216 Annex 21, Page 48 BIBLIOGRAPHY Arcia. Gustavo and Castro, Vanessa, <, Consultancy Report for the IDB, February 1999 Arcia, Gustavo and Castro, Vanessa, oExperiencias en educaci6n rural: que lecciones puede aprender Nicaragua?>>, Consultancy Report presented to the World Bank, Managua, June 30, 1999 World Bank, <>, Nadeem llahi, LCSPR, draft, nov. 15, 99, chapter for the Poverty Assessment of Nicaragua, 1999. World Bank, >, Benjamin Davis, Rinku Murgai, draft, Oct. 18, 1999 World Bank, >, Julia M. Davton, First draft, Nov. 99. World Bank, draft chapters for Education, Health, Population, Nutrition for the Poverty Assessment of Nicaragua, 2000. World Bank, Nicaragua, Central Government Spending, Data Base 1991-2000 based on official information from the Ministry of the Treasury. 1991-1998: Expenditures executed. 1999- 2000: Expenditures budgeted. IDB Nicaragua, < (NI-0075) Informe de Proyecto (borrador) Budinich, Ema; Quinteros, German and Ubilla, Rodrigo, (>, Notas Tecnicas de la Unidad de Educaci6n, Departamento de Desarrollo Sostenible, BID, presentado en XII Seminario Regional de Politica Fiscal, Taller "Evaluaci6n de la Gesti6n del Gasto Publico en Educaci6n", Santiago de Chile, January 27, 2000. Norwegian and Dutch Cooperation, <>, Informe final de consultoria de evaluaci6n, sept. 98. FISE - World Bank, Berk Ozler, Development Economy Study Group, <>, (Slides, undated). FISE, <>, Gustavo Bermuidez (coordinator), Ligia Maria Castro (principal consultant) and Ana Victoria Fiallos (support consultant) and support group of engineers and architects, January 2000. FISE, oAn4lisis institucional del Fise>> , Gustavo Bermudez (coordinator), Ligia Maria Castro and Luz Marina Garcia Fonseca, November 1999. Annex 21, Page 49 FISE, World Bank, Berk Ozler, Development Economy Study Group, <> (charts, undated). FISE - Suecia (;, Pablo Schneider June 30,. United Nations Population Fund - Nicaragua <;, Managua, March 1997. Fondo Social Suplementario (FSS), Reglamento operativo, Nov. 4,. 98. Fondo Social Suplementario (FSS), Informe de Avance ler. semestre 99. Fondo Social Suplementario (FSS). Resumen Ejecutivo, disponibilidad de recursos y ejecucion e i enero-agosto 1999. Government of Nicaragua, "Preliminary Document on the Initiative for Heavily Indebted Poor Countries (HIPC)", August 24, 1999 (document prepared jiontly by the IMF and IDA staffs) Government of Nicaragua, Technical Secretariat of the Presidency, General Directorate for Public Investments ((Resumen Ejecutivo Programa de Inversiones Publicas 2000»>, Managua Nicaragua. Government of Nicaragua, Technical Secretariat of the Presidency, General Directorate for PubliC Investments. <>. Government of Nicaragua, <, Managua, January 21, 2000. Government of Nicaragua, Ministry of the Treasury and Public Credit, General Directorate for the Budget, <;. Government of Nicaragua - UNICEF - Government of Sweden, <;, Consultants: E. Lewin, K Boman, M Medina, September 1999. INIFOM , <«Programa de Desarrollo Local en 5 capitales departamentales Esteli, Somoto, Ocotal. Le6n y Chinandega>>. INIFOM - Suecia, <. MAG, <, Managua, October 1999. MAG, Planning Office for the Agriculture Sector, < , Managua, Nicaragua, July 1999. MECD - World Bank, (, Basic Education Project, Progress Report 1997, January 1998. MECD, Agency for International Development (USAID), Academy for Education Development (AED), <>, Contract N° 524-C- 00020-00 Base 11, Operative Plan 1999, May 99. MECD/UNESCO/NETHERLANDS, <> Final report of the External Evaluation Mission, Managua Nicaragua, April 98. MIFAM, MECD, WFP: <>, baseline study prepared by the Centro de Investigacion y Asesoria Socioeconomica (CINASE). Summary, Managua, Nicaragua December 1998 MINSA, <>, Chontales Sept. 95. MI7NSA - Royal Embassy of the Netherlands, <>, Dr. Ena Liz Torrez, Direcci6n Primer Nivel de Atenci6n, MINSA, Dr. Marta Medina, Clarity ApS, Royal Embassy of the Netherlands, Dr. Peter Gondrie, KNCV, Royal Embassy of the Netherlands, March 2000. MINSA - Netherlands <, Oct-Nov. 98. MINSA - Sweden, Project for Support and Consolidation of the Sistemas Locales de Atencion Integral de Salud - Prosilais, <>, M. Medina, U. Farnsveden, R Belmar, Nov. 1999. Mirette, Seireg, <>, Managua, Nicaragua, March 2000. Movimiento Comunal Nicaragiiense/United Nations Population Fund, <>, Evaluation Report, April-June 1997 United Nations, CEPAL, Panorama Social de America Latina 1998, United Nations, April 1999. Annex 21, Page 51 UNDP Nicaragua, <. Final Report of the Evaluation Mission, December 9, 1999. WFP, <) abril. 99. WFP/NIC 4571-01 Proyecto "Apovo al Mejoramiento de la seguridad alimentaria de familias rurales pobres en areas afectadas por sequias e inundaciones", Project Summary /(no year or author) WFP, < WFP, Project NIC 4515 Amplification 1) <» (Design of second stage) WF P (WFP/EB. 1/97/6/Add. 1) Junta Directiva, January 97. WFP- NIC 2593(2), «Informe de Misi6n de Revisi6n Tecnica del Proyect'o Desarrollo Lechero PMA-NIC 2593" (2), Wolfram Herfurth, Joaquin Secco and Denis Pommier, May 12, 1997. PRODES, Report of the First Mission for accompaniment of PRODES, Nicaragua NEI Netherlands Economic Institute, Feb. 1999 PROSUR - Holland, <>, Annual Report 1998 (Jan..99), Annual Plan 1999 (Nov. 1998), Quarterly report October-December 1999 (Jan. 2000). PROTIERRA/INIFOM, <(Proyecto Municipios Rurales. Evaluaci6n de efectos y sostenibilidad potencial> World Bank, final draft, December 1999. PROTIERRA/INIFOM, <>, Nicaragua, December 20,1999. PROTIERRA/INIFOM, <(Evaluacion de efectos y sostenibilidad potencial del Proyecto de Municipios Rurales. Analisis Costo-unitario Protierra»>, Dagoberto Rivera, November 1999 Annex 21, Page 52 PROTIERRA/INIFOM, (> Aidan Gulliver, FAO Investment Center. PROTIERRA/INIFOM, <>, Mary Lisbeth Gonzalez, December 1999. Protierra/ Inifom, <, August 1999, (copy, no author or date) Republic of Nicaragua, <>, document prepared for meeting with donor community, Managua, Nicaragua, June 26, 1998. UNFPA . "Evaluaci6n final de NIC/96/PO3 Servicios de Salud Reproductiva y planificaci6n familiar en Matagalpa Jinotega y la RAAN". Sissel Hodne Steen and Mabel Munist, Sept. 99. UNFPA, Onforme Misi6n de Revisi6n de Medio Termino, Programa Pais 1994-1996»>, Nicaragua, April 29, 1996. UNFPA, <>, Caspar Peek, April- June 1997. lJNFPA, <>, Managua, March 1997. UJNOPS/PNUD, oInforme de evaluaci6n externa del Proyecto "Apoyo al desarrollo humano sostenible de las comunidades indigenas y campesinas de la zona norte de la Region Aut6noma del Atlantico Sur" (PRORAAS -bridge stage I y II/ UNDP-NIC 96/010 and DGIS-N1004302 y Nl-4304)>», Managua, April 1998. USAID, Programas de AID vigentes en el pais. ejecuci6n y desembolsos 97-98 (Report remitted to the World Bank), March 2000. Annex 21, Page 53 Government of Nicaragua, "Estrategia de Reducci6n de la Pobreza, Primera Parte: Diagn6stico v Lineamientos," Managua, January 21, 2000 " Government of Nicaragua, Ministry of the Treasury and Public Credit, General Directorate for the Budget, "General Budget of the Republic 2000." Although this can be explained in part by the Supplementary Social Fund (FSS), the observation about long-term sustainability of the still persists. Only 11% of public spending on secondary education is received by students in the poorest 40% of households, while 55% is received by students in richest 40%. (See World Bank "Incidencia del Gasto Publico en Educaci6n y Salud en Nicaragua en 1998", Julia M. Dayton, first draft, November 1999. The study of influence (Note 4) does not analyze public spending on higher education. MIFAM, MECD, PMA: "Situaci6n alimentaria y nutricional de niios en edad preescolar y de primer ciclo de ensenianza primaria , beneficiarios del proyecto PMA/NIC 4515," baseline study by the Center for Socio/economic Research and Advice (CINASE), Managua, Nicaragua, December 1998. vii CEPAL-IDB-UNESCO, "El costo-efectividad de las politicas de educaci6n primaria en America Latin a: Estudio basado en la opini6n de expertos," Technical Notes from the Education Unit, Department ror Sustainable Development, IDB, pp. 12-14. ' In the Budget Program it mentions that 250.000 students are benefited and the annual cost is US$3.9 million, for a per-child cost of US$15.60 per year. ix CEPAL-IDB-UNESCO, "El costo-efectividad de las politicas de educaci6n primaria en America Latirna: Estudio basado en la opini6n de expertos," Technical Notes of the Education Unit, Department {or Sustainable Development, IDB, presented at the XII Regional Seminary on Fiscal Policv, Workshop."Evaluaci6n de la Gesti6n del Gasto Puiblico en Educaci6n," Santiago, Chile, January 27, 2000. There is evidence that they are not used properly, especially because the poorest have no money to buv lime. xi Although the SILAIS have made progress in defining the essential elements of health care for their populations, it seems the task is not complete. MINSA, in its "Analisis del Sector Salud de Nicaragua" formulates the national policy for 1997-2000 land mentions the needs to "define and delivery a basic package of health services to the whole population.... that includes sanitary education, the provision of preventative clinical services." Xii Government of Nicaragua - UNICEF - Government of Sweden, Monitoring Mission August 3C0 - September 17, 1999. Report to SIDA, Consultants: E. Lewin, K Boman, M Medina, Managua, Nicaragua, September 1999. xiii Project to Strengthen Primary Health Care FORSAP (Evaluation mission, Chontales, September 1995). xiv Project for Support and Consolidation of the Local Systems for Integral Health Care - PROSILA IS, executed in six SILAIS. In "Apoyo de ASDI al Sector Salud en Nicaragua PROSILAIS 1992-1998." M. Medina, U. Famsveden, R. Belmar. November 1999. xv Mirette Seireg, "Estudio de fortalezas y obstaculos que enfrentan los Programas de Nutrici6n y Alimentaci6n en Nicaragua." xvi INIFOM, "Programa de Desarrollo Local en 5 capitales departamentales Esteli, Somoto, Ocotal. Le6n y Chinandega." xvii Government of Nicaragua - UNICEF - Government of Sweden, Monitoring Mission, August 31 - September 17, 1999 Report to SIDA, Consultants: E. Lewin, K Boman, M Medina, Managua, Nicaragua, September 1999. XViiI Evaluation of the Program for Integral Attention to Nicaraguan Children (PAININ), consultancy report for the IDB. G. Arcia, V. Castro. February 1999. xix USAID, AID programs in effect in the country (information remitted to the World Bank). One of these has a component for giving supplementary food to women and malnourished children under three years of age xx Mirette Seireg, "Estudio de fortalezas y obstaculos que enfrentan los Programas de Nutrici6n y Alimentaci6n en Nicaragua." xxi Cooperation from Norway and Holland, "Evaluaci6n del Proyecto Ampliaci6n y fortalecimiento de las Comisarias de la mujer y la nifiez en Nicaragua", final report from consultancy for evaluatoin, September 1998. Nicaragua, Rural Municipalities Project (PROTIERRA), "Evaluaci6n de efectos y sostenibilidad potencial," World Bank, final draft, December 1999. XXII MAG: "Gasto del Sector Publico Agropecuario en Nicaragua," Managua October 1999. Annex 22, Page I Annex 22 - The Incidence of Public Spending on Health and Education by Julia M Dayton INTRODUCTION I. This paper analyzes the distribution of public spending on health and education in Nicaragua in 1993 and 1998. It focuses on assessing changes over time in the distribution of spending across income groups, between genders and across the seven geographic regions of the country. BENEFIT INCIDENCE METHODOLOGY 2. The benefit incidence method seeks to measure how government subsidies on services such as health and education are distributed across groups in society. It became an established approach after the ground-breaking work on Malaysia by Meerman (1979) and on Colombia by Selowsky (1979), and there has been a recent resurgence of interest in the approach, as reviewed in Van de Walle and Nead (1995). The distribution of the subsidies is determined by two broad factors. First, it depends on government spending itself, and how it is allocated within the sector. The lower the spending, and the greater the effective cost recovery, the lower will be the subsidy embodied in the service provided. Second, the distribution will depend on individual or household behavior -on who uses the service that the government provides. It is only by using the service (by sending a child to a primary school, or visiting the outpatient department at a hospital) that individuals and households can lay claim to the in-kind transfer that is implicit in the subsidy. Benefit incidence analysis, therefore, brings together two sources of information: data on the government subsidy (estimated as the unit cost of providing the service, less any cost recovery back to the government) allocated to the different categories of service (primary schooling, in-patient hospital care, etc.) and information on the use of these services by individuals and households, which is usually obtained from household surveys. 3. In general, government expenditures will be more equally distributed when the spending is concentrated on services that are widely used by the population, and especially by poorer groups. If public expenditures are concentrated on primary education or on primary-healthcare facilities such as clinics or health centers -which are widely used services benefiting poor and non-poor- public expenditures will tend to be more equally distributed. However, if governments spend more on high-cost services that are not generally used by poorer groups (such as university education or in-patient hospital care), the incidence of spending is likely to be more unequal. In sum, the benefit incidence of public spending depends on both the allocation of public expenditures within the sector and on the behavior of households. 4. An important limitation of the approach is that it does not necessarily measure the effect of government spending on the welfare of households. The real benefit to a household in using a public service lies not so much in the monetary value of that service (usually approximated by unit cost), as in the direct benefits it gives -better health, literacy and better income-earning potential in the future. This should be kept in mind when interpreting the results of this paper. Annex 22, Page 2 PUBLIC PROVISION OF HEALTH SERVICES A. III health and household response to ill health 5. What is the response of a household to an incidence of illness or injury? To what extent do households seek care for a sick or injured member? Information on the patterns of healthcare use was provided in two national surveys: the Living Standards Measurement Survey, or LSMS. for 1993 and 1998. Individuals were asked about whether they had been ill during the past four weeks and if they had sought medical care. The incidence of public expenditures on health is determined fundamentally by this pattern of illness and treatment response. 6. In 1993, on average 23% of the population reported having an illness in the past four weeks, and this share rose to 36% in 1998 (Figures 1 and 2). In 1993, the non-poor reported slightly higher rates of illness than the poor (25 versus 21 %). This is a typical finding of surveys that rely on the self-reporting of illness, as it is generally observed that the poor (and uneducatecl) are less inclined to observe and therefore report an illness in the household. In 1998, people in a I quintiles reported roughly the same proportion of illness.' In both years, women were somewhar more likely than men to report an illness (24 versus 22% in 1993 and 36 versus 42% in 1998). Figure A22.1 - Pattern of illness reporting in Nicaragua, 1993 c 100% 80% rc -60% Well *Sick 40% 20% X 0% Quintile Household members were assigned to quintiles according to the per capita total expenditures of the household to which they belong. Quintile I represents the poorest 20% of the population and quintile 5 the richest. Annex 22, Page 3 Figure A22.2 - Pattern of illness reporting in Nicaragua. 1998 o 100% 60% 1W1 _ll o 40% _ _ SLIII t _ < X Sick c 20% 0% - ~ ~ ~ ~ ~ ~~0- Quintile 7. Among those who reported an illness, the propensity to seek care varied directly with household expenditure level, with the poor less likely to seek health care than the non-poor. In 1993, 34% of the poor who were ill sought medical care for illness, whereas 52% of the non-poor sought care (Figure 3). In 1998, this share increased to 37% among the poor and decreased to 47% among the non-poor (Figure 4). The type of health facility consulted also varied by income level. In 1993, the poor mainly sought care at public primary-level health facilities (health posts or health centers), while the non-poor were more likely to seek care at a public hospital or at a private facility. In 1998, while the poor relied increasingly on public primary-level care, the non- poor increasingly sought care in the private sector. This pattern of increasing reliance on the private sector by the richer groups (who can usually afford to pay more for health care) was desirable, as it allowed the pubic system to focus its scarce resources on the poor. However, it remained somewhat problematic that non-poor were more likely than the poor to use public hospital-based services in both years, since these services are more costly to the public health system, as will be discussed below. There were few gender differences in the pattern of health care use reported in the two surveys, although in 1998 women were more likely to seek care at a public primary health facility than were men. 8. There were substantial differences across geographic regions in the propensity to seek care in 1993, with people living in Managua and the Pacific region much more likely to seek care than those living in the Central and Atlantic regions. These differences diminished substantially in 1998. In both years, residents of Managua and the Pacific region were more likely to seek private care, whereas residents of the Central and Atlantic regions were more likely to seek care at a public facility. Annex 22. Page 4 Figure A22.3 - Response to illness, among those reporting illness, 1993 00% =E 0 Sick-no care 0 Sick-pnvate care U Sick-public hospital care 0 Sick-public primar_ care 20%/|_ILtWL LLI- o , C ° 0)C CD (Z = o C Quintile Figure A22.4 - Response to illness, among those reporting illness, 1998 1 00%/ T 800/ 0 Sick-no care 600% F 0 Sick-private care 4 Sick-public hospital care 40:" U Sick-public prinmarv care 0% (N -t ~~~ C ~ ~ 00 Quintile - Annex 22, Page 5 B. Public spending on health 9. Table I shows the pattern of public health expenditures by level of health care and region (for 1998 only)2 and public recurrent spending per health care consultation.3 From 1993 to 1998, total public expenditures on health services doubled in nominal terms. After taking into account inflation and contributions to the budget by donors (included in the budget for 1998 but not 1993), there was little real increase in real government spending for public health services. 10. The annual number of consultations for each type of service is shown from two sources: (a) the Ministry of Health, based on client service data. and (b) the 1993 and 1998 surveys, which are based on self-reports of the use of services.4 Although it is not possible to know which source is more accurate, the survey estimates are used to ensure consistency within this benefit-incidence analysis, which must rely on the survey data for the distribution across the population of the health care consultations. Fewer outpatient visits to public health centers were reported in the survey as compared with Ministry of Health records, with the opposite pattern observed at the hospital level. I. Spending is distributed fairly evenly between the two levels of health care, although the number of consultations at the primary level is about four times greater (depending on whether estimates of numbers of consultations from the Ministry of Health or the LSMS survey are used). The per-unit cost of a consultation at a primary health care center was C$60 in 1998, compared to $32 in 1993, and the cost of a consultation at a hospital was C$480 in 1998, up from $174 in 1993. The cost to the government of an outpatient visit was between 5 and 8 times more expensive at the hospital level than at the primary health care level, probably reflecting higher overhead costs.5 C. The incidence of public expenditures on health 12. What do the patterns of health care response by household members (as reported in the surveys) imply for the incidence of public spending? To translate this into public expenditure terms, incidence analysis simply allocates the subsidy (or unit cost) embbdied in the care they received to those households that used publicly-provided health services. Use of subsidized health services is therefore considered as an 'in-kind transfer' to those households that use the system. In effect, the analysis poses the question: What additional income would households need if they had to pay for services? 2 Public expenditures were only available for the four geographic regions of the country (Managua, Pacific, Central and Atlantic), with no disaggregation between urban and rural areas within these regions. Thus the incidence analysis for the health sector will be limited to these 4 regions. 3Only outpatient visits are considered in the per unit subsidy. At the hospital level in particular, this may not be an accurate reflection of the true unit cost, as a substantial subsidy is probably allocated to inpatient consultations. In order to incorporate hospital inpatient costs into this analysis, it will be necessary to estimate the relative cost of an inpatient stay/case as compared to a hospital outpatient consultation. 4 The annual number of consultations from the survey data is obtained by multiplying the number of consultations reported during the 4-week recall period by 12. 5 Again, inpatient consultations have not yet been taken into account and this may alter these ratios. Annex 22, Page 6 Table A22.1 - Public Health Expenditures by Region and Level of Care, 1993 & 1998 Primarv-level healthi care Hospital 71993 1998-g 1993 J9 Recurrent expenditures Managua - 79,100,298 - 181,350,604 Pacific - 95.046,300 - 101,352,658 Central - 98,835,193 - 65,302,5 10 Atlantic - 33,237,050 - 15,142,41 ' Total 139,475,646 306,218,842 171,345,012 363,148,188 Consultations (Ministry of Health) Managua 1,379,214 1,201,103 387,531 307,65' Pacific 1,559,362 1,760,052 300,430 214,151; Central 1,170,359 1,683,938 239,839 182,6301 Atlantic 301,663 472,765 57,867 52,266) Total 4,410,598 5,117,858 985,667 756,71.' Consultations (Survey) Managua 702,670.68 570,206 273,734.88 266,50W Pacific 856,767.57 1,570,328 429,378.56 387,936 Central 1,038,551.61 1,897,977 186,228.92 351,984 Atlantic 179,962.14 548,209 131,674.04 184.591 Total Survey consultations 2,777,952 4,586,720 1,021,016 1,191,019' Per Visit Subsidy (MOH consultations) Managua 66 581) Pacific 54 473 Central 59 3583 Atlantic 70 290 National Average 32 60 174 480 Per Visit Subsidy (Survey consultations) Managua 139 681) Pacific 61 261 Central 52 185 Atlantic 61 8 2 National Average 50 67 168 305 Annex 22. Page 7 13. The results of this exercise for 1993 and 1998 are shown in Figures 5, with additional details provided in Appendix Tables 1-6. The incidence of government spending on health is reported for each quintile, by poor/non-poor and for females and males. The regional allocation is then considered separately. 14. Tite primary-level health subsidy became increasingly targeted to thIe poor between 1993 and 1998. The top two panels in Figure 5 show the distribution of the public health subsidy for primary health care in 1993 and 1998. respectively. In 1993, the poorest two quintiles received 13 and 22% of the primary health subsidy, whereas by 1998, these two quintiles received 23 and 26%. This shift in spending is most evident when assessing the relative share of the subsidy received by the poor and non-poor groups: the share of the subsidy received by the poor increased from 47% in 1993 to 61% in 1998. This increase is largely due to an increase in consultations by the poor to primary-level health centers, perhaps as a result of explicit efforts to target this group by decentralizing and improving delivery at this level. 15. Tlhe hospital-level subsidy became more equitably distributed, but stillfavored tlhe richest groups. The poorest two quintiles received 8 and 18% of the hospital subsidy in 1993, and this share increased to 13 and 21%, respectively, in 1998. The total share accrued to the poor increased from 34% in 1993 to 40% in 1998, but the majority of the hospital subsidy continued to be allocated to the non-poor in 1998. Of particular concern at this level of care is how little the poorest quintile of the population shares in the public subsidy. 16. The total healtlh subsidy became fairly evenly distributed across income groups. The bottom two panels in Figure 5 show the distribution of the total health subsidy (combining primary and hospital levels of care). The distribution between the poor and non-poor shifted from 40/60 in 1993 to 50/50 in 1998, mainly as a result of changes at the primary level of care. 17. The allocation across genders remains relatively equal. The per capita health subsidy for females is slightly higher than that for males, but this is likely the result of the greater health needs of women, especially those of reproductive health age. 18. Regional distribution favored Managua. The regional allocation of health expenditures was only available for 1998, so comparisons over time were not possible. However, in 1998 the total per capita health subsidy overwhelmingly favored Managua (Figure 6). There was a concentration of hospitals in Managua, with greater use of hospital outpatient services there, resulting in a per capita hospital subsidy that was more than twice as much in Managua than in other regions. At the primary level, the per capita subsidy was distributed fairly equally across all regions. 19. Primary-level health subsidy became mildly pro-poor between 1993 and 1998. Figures 7 and 8 show the targeting of the public health subsidy compared with the distribution of income (measured by per capita total household expenditures) -also called the Lorenz curve- for 1993 and 1998, respectively. The cumulative shares of individuals in the population, ranked by per capita expenditure, are measured on the horizontal axis. The vertical axis measures the cumulative shares of expenditures and public health subsidy. In 1993, all levels of the health care subsidy were progressively distributed; their concentration curves were situated closer than the Lorenz curve to the 45-degree diagonal. The health subsidy could not be described as pro-poor in 1993. as none of the curves were above the 45-degree diagonal. 20. There was a distinct shift in the curve for the subsidy for health centers from below the 45-degree diagonal in 1993 to above the diagonal in 1998, in a pattern that was mildly pro-poor. Annex 22. Page 8 The hospital subsidy continued to have a progressive distribution, in that this curve was situated above the Lorenz curve, but still below the 45-degree diagonal. Taken together, total health expenditures (primary and hospital together) were distributed equitably in 1998, with the concentration curve falling almost exactly on the 45-degree diagonal. Annex 22, Page 9 Figure A22.5 - Distribution of Public Health Subsidy Across Poverty Status and Gender 60 60 40 f n 40 202 2 34 Poo 2 34 Po M Poore Riche Non- M'ema Poore Riche PNon Maema Primary-level, Primary-level, 60 60 - 40 40 ~n 20 20 i il20 2 34 2 34 Poore Riche PN°oo- MalemP Poore Riche Poo M+era Por HeahNon- Merna No n Hospital, Hospital, 60 60 40 40 20 20 0 0 2or 3 iche Poo 234Poo Poore Riche Non- M?terna Poore Riche Non- Mema All Health, All Health, Annex 22 Page [0 Figure A22.6 - Per Capita Government Health Subsidy by Facility and Quintile, 1998 ,200 150 143 74 150 1Hospital 100 64 4330 7 Pia 30 D Prim_ ~~~~~~~) ~~ ~ ~ A ~50 c0 Managua Pacific Central Atlantic Average Region Figure A22.7 - Concentration Curves for the Public Health Subsidv. 1993 100 80 - health '/tz / C centers / / 60 U +-hospitals //- - ~all health /,/tww / - 40 40 ____ / oPC Tot Exp 20c - Diagonal 0 0 20 40 60 80 100 Share of subsidv (%) Annex 22, Pag-e 11 Figure A22.8 - Concentration Curves for the Public Health Subsidy, 1998 I i M *100 80 Sh health : / /// / are centers / / / 60 of * hospitals pul - - all health 40 ati on PC Tot Exp 20 Diagonal 0 20 40 60 80 100 Share of subsidy (%) THE PUBLIC PROVISION OF EDUCATION A. Educational outcomes: enrollments. 21. Table 2 shows gross and net enrollment rates for the primary level and gross enrollment rates at the secondary level. Gross enrollment rates are commonly over 100% in settings where some of the children enrolled in school are older than the official school age.6 22. Primary enrollment rates. At both the primary and secondary levels, there was a direct relationship between income group and enrollment rates, and children from poorer families were less likely to be enrolled in school than children from non-poor families. In 1993, net enrollment rates were 55% for the poor as compared with 80% for the non-poor, with the poorest quintile substantially lower at 44%. Nevertheless, there were substantial improvements between 1993 and 1998 at the primary level for all income groups, with the greatest gains for the poor, for whom the average net enrollment rate increased to 78 in 1998. Boys had slightly lower enrollment rates than girls. Regional disparities remained acute, with the Rural Atlantic Region falling the furthest behind with a net primary school enrollment rate of 60%, followed by the Rural Central Region with an average net enrollment rate of 75% in 1998. Enrollment rates for those two regions did not improve much between 1993 and 1998, although other rural regions such as the Rural Pacific and Rural Central Regions made substantial gains during this period. 23. Secondary enrollment rates. Secondary enrollment rates grew tremendously between 1993 and 1998, from a national average of 23% in 1993 to 57% in 1998. The gains were most substantial for the non-poor, almost tripling the rate from 32 to 86. The rate for the poor doubled from 13% to 23%. As with the primary level, there are also stark regional differences, with rural 6 In Nicaragua, primary school age is 7-12 years old and secondary school age is 13-17 years of age. Annex 22. Page 12 areas having gross enrollment rates one-third those in Managua. Gross enrollment rates remained below 25% in 1998 in the Rural Atlantic and Rural Central Regions. 24. Public vs. private enrollment. At the primary level, the lower four quintiles dominated enrollment in public schools, with less than 10% enrolled in private schools. Among the richest quintile. over 40% of the students were enrolled in private schools. At the secondary level, however, the pattern is reversed, with the middle three quintiles having the largest enrollment in public schools. In the richest quintile, fully half of the students in 1998 were enrolled in private school. Table A22.2 - Gross and net enrollment rates for primary and secondary school, 1993 and 1998 Primary Secondary Neta Grossb Grossb 1993 1998 1993 1998 1993 ] 99 Poorest 44 70 59 92 4 13 2 6] 83 86 108 14 32 3 72 88 106 109 27 62 4 76 91 113 120 29 80 Richest 88 98 121 109 36 103 Male 66 83 99 106 24 52 Female 67 87 101 107 22 62 Poor 55 78 80 102 13 26 Non-poor 80 93 128 113 32 86 Managua 83 85 129 107 38 90 Pacific Urban 87 91 109 115 36 69 Pacific Rural 73 88 100 116 21 33 Central Urban 86 88 127 109 19 79 Central Rural 59 75 81 92 3 24 Atlantic Urban 87 88 119 116 27 56 Atlantic Rural 60 60 89 81 6 11 Average 66 85 108 107 23 57 a Share of children aged 7-12 in enrolled in primary school. b Total enrollments in school as share of school aged children. Annex 22. Page 13 B. Public spending on education 25. Table 3 shows the public recurrent spending for education at the national level for 1993 and 1998. Using enrollments reported by the Ministry of Education and by households in the 1993 and 1998 national surveys, the table also shows two estimates of the per-student subsidy implied in education expenditures at each level. The total enrollment reported by the Ministry of Education is generally lower than that reported by households in the survey, resulting in larger per-student public subsidies when using the Ministry of Education enrollment figures. As with the health sector, the survey estimates will be used in the incidence analysis to provide consistency across the analysis.7 26. In nominal terms, the average per-student subsidy at the primary level doubled between 1993 and 1998, from C$221 per-student to C$430 in 1998. This was largely due to expanded public spending on primary public education (mainly caused by inflation), whereas enrollment increased by only 10%. At the secondary level, there was a two-fold expansion in enrollments accompanied by a modest increase in public spending, which resulted in a decline of the average per-student subsidy from C$590 to C$266. At the technical and university levels, the enrollment figures given by the line Ministries more accurately reflect the number of full-time students, making them more comparable with enrollment at the lower levels. Per-student subsidies for technical education increased by about one-third, going from C$2,582 in 1993 to C$3,394 in 1998. 8 times more than the primary school per-student subsidy. At the university level, the per- student subsidy more than doubled, and conservative estimates (of enrollment) suggest it was 40 times more than the primary-level subsidy. 27. The regional variation in the per-student subsidy implied in education spending is shown in Table 4 for primary and secondary education, the only two levels of education for which regional expenditure data was available. (Annex Table 7 provides more details.) Regional data on public expenditures were not broken down into rural and urban. At the primary level, it was assumed that expenditures were allocated equally between rural and urban areas within each region. This assumption was not made at the secondary level, due the concentration of secondary schools in urban areas. At the primary level, the per-student subsidy in Managua was about one- third more than the national average, indicating a concentration of educational resources in the capital. Per-student subsidies are lowest in the Atlantic region. At the secondary level, per-student subsidies were lower in Managua than other regions in both 1993 and 1998. perhaps indicating the concentration of pupils here. ' This includes survey estimates at the vocational and university level, which may be overestimates of full- time enrollment as they include both part-time and full-time students. Nevertheless, these estimates are used because we also know the distribution of these enrollments across the population. Annex 22, Page I4 Table A22.3 - Public Spending for primary and secondary education Primary Education 1993 Cs 1998 Cs Total recurrent expenditures* 151,239,710 331.692,296 Total enrollment (Ministry of Education) 613,044 664,491 Total enrollment (Survey) 685,448 770,836 Per-student subsidy (Ministry of Education) 247 499 Per-student subsidy (Survey) 221 430 Secondary Education Total recurrent expenditures* 50,732,306 67,804,866 Total enrollment (Ministry of Education) 92,196 192,884 Total enrollment (Survey) 85,998 260,837 Per-student subsidi' (Ministry of Educatiow,) 550 352 Per-student subsidy (Survey) 590 260 Vocational Education Total recurrent expenditures 39,354,000 45,464,650 Total enrollments -- full-time equivalent (SETEC) 15,243 13,397 Total enrollments (Survey) 44,379 4,481 Per-student subsidy (SETEC) 2,582 3,394 Per-student subsidy (Survey) 887 10,147 University Education Total recurrent expenditures 152,112,500 348,855,000 Enrollments (full-time equivalent) under Scenario A 15,000 15,000 Enrollments (full-time equivalent) under Scenario B 20,000 20,000 Enrollments (Survey) 32,021 38,162 Per-student subsidy- Scenario A 10,141 23,257 Per-student subsidy- Scenario B 7,606 17,443 Per-student subsidy (Survey) 4,750 9,142 * This total does not include recurtent expenditures tor the subsidized line item. Annex 22. Page 15 Table A22.4 - Per-student subsidy by region for primary and secondary education, 1993 and 1998 (using Survey enrollment figures) Primari' Secondary 1993 1998 1993 1998 Per Student Subsidy with Survey enrollments Alanagua 180 601 441 250 Pacific Urban 225 420 - - Pacific Rural 318 372 - - All Pacific 264 395 611 312 Central Urban 237 650 - - Central Rural 218 287 - - All Central 227 398 935 225 Atlantic Urban 211 285 - - Atlantic Rural 148 323 - - All Atlantic 174 303 529 237 NationalAverage 221 430 590 260 Source: Annex table 7. C. The incidence of public spending on education 28. By bringing together information on per-enrollment expenditures provided with household data on enrollment, it is possible to trace the incidence of public spending on education across the population.8 The results of this exercise for 1993 and 1998 are reported in Figure 8 (with more information provided in Annex Tables 8-x). Households were allocated the appropriate subsidy for each child enrolled in public primary and secondary schools. 29. Thie primary school subsidy was better targeted to the poor in 1998 relative to 1993. The top two graphs in Figure 8 show the distribution of the public primary subsidy in 1993 and 1998. In 1993, the primary school subsidy was allocated equally among the poor and non-poor, but the allocation to the poor increased to 57% in 1998, with the non-poor receiving only 43%. This shift reflects increase enrollment among the poorest groups, but because the poor have more school- age children than the non-poor. the share of the subsidy for the poor would need to continue to increase to reflect full enrollment among the poor. 30. Secondary school public subsidy continued tofavor non-poor. The second row of graphs in Figure 8 presents the distribution of the secondary school public subsidy across quintiles in 1993 and 1998. In 1993, the poorest two quintiles received 4 and 15% of the secondary school public subsidy, and these shares only modestly increased to 6 and 15 in 1998. The total share received by the poor actually decreased from 30% in 1993 to 27% of the total in 1998, indicating that significant gains remain to be made at this level. 31. Share of public technical school subsidy increasedfor the poor. In 1993, 70% of the total public subsidy for technical education was accrued to the non-poor, with the poor receiving only 30%. In 1998, the share for the poor increased to 40%, with a clear shift toward women. The result of this shift is that the top four quintiles benefited evenly from the technical education subsidy in 1998, and the poorest quintile received none. 8 Public expenditures on subsidized education are not included in the incidence analysis. Annex 22, Page 16 32. University-level subsidy increasingly targeted to the richest quintile. The public university system. however, still overwhelmingly favored the non-poor, and the share they received increased from 87 in 1993 to 93% in 1998. These benefits were increasingly captured by the richest 20% of the population. 33. Total education subsidy continued tofavor thze non-poor. When considering all levels of education together, the poor received only 32% of the education subsidy in 1998, a slight increase from 30% in 1993. This inequality was driven largely by the university subsidy, which is as large as the total primary-level subsidy and goes almost entirely to the richest quintile. 34. Boys and girls continued to share equally. This analysis reveals very little gender disparity. At both the primary and secondary levels of education, girls and boys shared equally in the educational subsidy. There was a slight increase in the share of the public subsidy that goes to girls at the secondary level, from 49 to 54%, which reflects a slightly higher per-person subsidy. but also the demographics of this age group, in which there are more girls than boys. 35. Clhanges in regional allocation of primary-level subsidyfavored Managua, and tlhe Atlantic and Central regions stiifell behiind. Figures 9 and 10 show the per capita subsidies fcr the primary and secondary level educational subsidies (data for the regional analysis were not available at the technical and university levels). In 1993 the primary-level per capita subsidy varied across regions with the Pacific and Central Regions receiving the largest per capita subsi Ay and the Rural Atlantic Region receiving the smallest subsidy. In 1998, the subsidy continued to be distributed in a similar pattern, although the subsidy for Managua was relatively larger than in 1993. 36. At the se.condary level, the regional distribution of the subsidy remained constant between 1993 and 1998, with the Urban Pacific and Managua Regions benefiting from the highest per capita subsidies and the Rural Central and Atlantic Regions having the smallest per capita subsidies. This was the result of smaller per-student subsidies in the Central and Atlantic Regions, coupled with lower enrollment rates in these areas. 37. Between 1993 and 1998 the primary school subsidy became pro-poor in its distribution, but other levels of education were also progressively distributed. Another way to examine the targeting of the public education subsidy is to compare the distribution of the education subsidy to the distribution of income in the country (measured by per capita total household expenditures), also called the Lorenz curve. This is shown in Figure 11 for 1993 and in Figure ] 2 for 1998. The cumulative shares of individuals in the population, ranked by per capita expenditure, are measured on the horizontal axis. The vertical axis measures the cumulative shares of expenditures and public education subsidy. In 1993, all education subsidies except the university subsidy were distributed in a progressive pattern, as these curves lie between the diagonal and the Lorenz curve. Spending on primary education was distributed equitably: its concentration curve was along the 45-degree diagonal. 38. By 1998, however, the primary-level subsidy was distributed in a pro-poor way, as it curved well above the 45-degree diagonal. (This is probably the result of two factors: efforts to bring more poor children into school have been effective and the richer groups have opted for private education.) The distribution at other levels of education did not change but continued to be progressively distributed in that they were distributed more equitably than per capita total expenditures, but none of the other subsidies were any more equitably distributed than they were in 1993. As mentioned before, the university-level subsidy became even more regressive, favoring the rich even more than in 1993. Annex 22, Paue 17 Figure A22.9 - Distribution of Public Education Subsidy Across Population Poverty Status and Gender, 1993 and 1998 60%/o 60% 4o0/ %l ff 40% 20% iffll n n H ~H 209O%fll Rn Hn 11 0% 0% 2 34 Pr23 4P Poorest Richest Mattemale goon-poor Poorest Richest MaIeemale tQorn-poor Primary, 1993 Primary, 1998 60% . _ .f17 60% H. 40% nf b 40% f/oI 20%flfll HiI RLIJ 20% 1 n fl HH 1IL 0% 0 Poorest Richest Makemale lRorn-poor Poorest Richest Maemale Plon-poor Secondary, 1993 Secondary, 1998 60% I 60% - 40% * 40% - * 20% l II 1 20% l _________________________ 0% Poorest Richest Maemale Plorn-poor Poorest 34Richest Maremale P°on-poor Technical, 1993 Technical. 1998 80% - 80% - 60% 60%- 40%-** 40% - Ii 20% - * 20% * * .l 0% 0% Poorest Richest Maleemale fon-poor Poorest Richest Maemale lNon-poor University, 1993 Ulniversity, 1998 60% - 51i 4950 70 60% - 5 i - M 40% - 25% 35% n 30% 40% 35% 32% 90/ n° n n 0 ni i 20%- 2O 14%20% 21% _ 0% Ilrnf 111 0% rin0nF I Poorest Richest Maleemale 1norn-poor Poorest Richest MaWemale lWorn-poor All Education, 1993 All Education, 1998 Annex 22. Page 18 Figure A22.10 - Per capita public education subsidy by level of education and quintile, 1993 1 00 __ __ __ __ _ __ __'__ __' _ ___: _ __ _ i_ 80 141 60 4 12 0* Secondary 40 3 OPrimary 20 ~ 3 4~~~~ Figure A22.11 - Per capita public education subsidy by level of education and region, 1998 60 ¢s Secondary 40 * Primary 20 FPX <)v° + 4? Annex 22, Page 19 Figure A22.12 - Concentration Curves for Public Education Expenditures, 1993 100 _ _ _ _ _ _ _ _ _ _ _ _80 -*-- + Primary /-° -- + Secondary 7 __- 60 , -a-- Technical /~~~~~~~ r ..jl..--.- University 40 - All education .< ---- 0 9PC Tot Exp .- ~~~~20 0 20 40 60 80 100 Share ofsubsidy(%) Figure A22.13 - Concentration curves for public education expenditures, 1998 100 80 ; -4--, / + Primary ----/_ +Secondarv 60 = + Technical 8//fit~~~~~~~C ...... University 40 ° - AUEducation // jt / - ' +~~~~~~PC TotExp / 20 45 Diagonal 0 20 40 60 80 100 Share of subsidy(%) Annex 22. Page 20 COMPARISONS WITH OTHER COUNTRIES 39. How does the incidence of public sector social spending in Nicaragua compare with other countries? Table 5 shows the share of total government subsidies in health and education reaching the poorest and richest quintiles in Nicaragua as compared with other countries and elsewhere for which data are available. Comparisons must be handled with care, given differences in methodologies across countries, but these results provide a general order of magnitude. In both health and education, the distribution of the public subsidy in Nicaragua is less equitable than in the other Latin American countries for which data are available. Uruguay, in particular, has beer more successful in targeting public health and education expenditures to the poor. In public health, Colombia has also been quite successful. Table A22.5 - The Incidence of Public Health and Education Expenditures in Selected Countries Percentage share of subsidy Healtlh Education Year Poorest Richest Poorest Richest Quintile Quintile Quintile Qu/intile Nicaragua 1993 10 26 9 35 1998 18 18 11 35 Latin America Brazil 1985 17 42 14 1 9 Colombia* 1992 28 12 20 21 Uruguay 1989 37 11 33 15 Other countries Indonesia 1989 12 29 15 29 Vietnam 1992 12 29 11 45 Kenya 1993 14 24 17 21 Ivory Coast 1995 8 35 13 35 Source: Demery, Sen and Vishwanath (1995). Poverty, Inequality and Growth. ESP Discussion Paper Series No. 70. World Bank, Washington, DC. * Quintiles defined over households. CONCLUSION 40. This analysis suggests that the public subsidies implied in public health and education expenditures have increasingly benefited the poor between 1993 and 1998. Improvements are more evident in the health sector than in the education sector. In health, 61% of the primary-level public subsidy was accrued to the poor in 1998, an increase from 47% in 1993. The hospital-level subsidy, while progressively distributed as compared with the distribution of income in Nicaragua, still largely benefited non-poor groups and residents of urban areas. In education, the share of the primary level subsidy received by the poor increased from 49 in 1993 to 57% in 1998, but at the secondary level it declined from 30 to 27%. The public subsidy for technical and university education continued to favor the wealthiest groups. In summary, public social expenditures are distributed more equitably than income in Nicaragua, but they could be targeted Annex 22, Page 21 more effectively, to poor Nicaraguans. Comparisons with other Latin American countries indicate that other countries have targeted their public spending more effectively to the poor. Experience in many countries teaches that investing in the human capital of the poor is one of the most effective ways to reduce poverty. Annex 22, Page 22 Annex Table 1. Incidence of Public Spending on Health, 1993 Quintile Government Subsidv Total Per capita Row share Primary-level care 50 subsidy per visit Poorest 18,701,760 23 13% 2 30,228,420 37 22% 3 33,850,140 41 24% 4 30,492,360 37 22% Richest 25,624,920 29 18% Poor 65,808,000 32 47% Non-poor 73,089.300 35 53% Male 63,747,600 31 46% Female 75,150,000 35 54% Total 138,897,600 33 100% Hospital 168 subsidy per visit Poorest 13,228,387 16 8% 2 30,175,488 37 18% 3 28,911,456 35 17% 4 45,130,176 54 26% Richest 54,085,248 62 32% Poor 58,498,272 28 34% Non-poor 113,031,072 54 66% Male 73,084,032 36 43% Female 98.443,296 46 57% Total 171,530,755 41 100% Total health Poorest 31,930,147 39 10% 2 60,403,908 73 19% 3 62,761,596 76 20% 4 75,622,536 91 24% Richest 79,710,168 91 26% Poor 124,306,272 60 40% Non-poor 186,120,372 88 60% Male 136,831,632 67 44% Female 173,593,296 82 56% Total 310,428,355 74 100% Annex 22, Page 24 Annex Table 3. Incidence of Public Spending on Health, Managua Region, 1998 Quintile Government Subsidy Total Per capita How share Primary-level care 139 subsidy per visit Poorest 5,995,547 124 8% 2 11,021,518 97 14% 3 20,669,108 87 26% 4 22,163,059 59 28% Richest 19,251,549 39 24% Poor 24,016,514 103 30% Non-poor 55,083,784 53 70% Male 32,303,128 53 41% Female 46,797,170 71 59% Total 79,100,781 62 100% Hospital 680 subsidy per visit Poorest - 0 0% 2 11,685,025 103 6% 3 59.000,801 249 33% 4 30,044,640 80 17% Richest 80,620,138 162 44% Poor 22,769,059 98 13% Non-poor 158,584,811 153 87% Male 73,460,509 121 41% Female 107,892,545 163 59% Total 181,350,604 143 100% Total Poorest 5,995,547 124 2% 2 22,706,542 199 9% 3 79,669,909 336 31% 4 52,207,699 140 20% Richest 99,871,687 201 38% Poor 46,785,573 201 18% Non-poor 213,668,595 206 82% Male 105,763,637 174 41% Female 154,689,715 234 59% Total 260,451,385 205 100% Annex 22, Page 25 Annex Table 4. Incidence of Public Spending on Health, Pacific region, 1998 Quintile Government Subsidy lotal Per capita Row share Primary-level care 61 subsidy per visit Poorest 18,409,763 60 19% 2 28,625,321 77 30% 3 22,674,139 61 24% 4 18,219,919 63 19% Richest 7,860,275 32 8% Poor 57,908,286 70 60% Non-poor 37,881,344 50 40% Male 35,718,277 46 37% Female 49,297,016 61 51% Total 95,789,418 60 100% Hospital 261 subsidy per visit Poorest 8,900,831 29 9% 2 26.884,148 72 27% 3 20,542,162 55 20% 4 22,334,292 78 22% Richest 22,592,682 92 22% Poor 39,659,890 48 39% Non-poor 61,593,912 82 61% Male 57,701,462 74 57% Female 57,659,494 71 57% Total 101,254,115 64 100% Total Poorest 27,310,594 89 14% 2 55,509,470 149 28% 3 43,216,301 117 22% 4 40,554,211 141 21% Richest 30,452,957 124 15% Poor 97,568,175 117 50% Non-poor 99,475,256 132 50% Male 93,419,739 120 47% Female 106,956,509 132 54% Total 197,043,532 124 100% Annex 22, Page 26 Annex Table 5. Incidence of Public Spending on Health, Central Region, 1998 Quintile Government Subsidy l'otal Per capita Row share Primary-level care 52 subsidy per visit Poorest 26,706,950 60 27% 2 26,322,048 71 27% 3 28,548,911 95 29% 4 10,936,786 47 110% Richest 6.180,383 33 6% Poor 67,453,788 71 68% Non-poor 31,240.810 54 32% Male 38,508,363 51 39% Female 60,186,984 77 61% Total 98,695,079 64 100% Hospital 186 subsidy4 per visit Poorest 16,914,096 38 26% 2 14,810,436 40 23% 3 6,894,871 23 11% 4 15,476,465 67 24% Richest 11,374,049 60 17% Poor 35,453,088 37 54% Non-poor 23,321,498 40 36% Male 31,271.882 41 48% Female 34,198,034 44 52% Total 65,469,917 43 100% Total Poorest 43,621,046 98 27% 2 41,132,484 112 25% 3 35,443,782 118 22% 4 26,413,250 114 16% Richest 17,554,432 92 11% Poor 102,906,876 108 63% Non-poor 54,562,308 94 33% Male 69,780,245 92 43% Female 94,385,018 121 57% Total 164,164,995 107 100% Annex 22. Page 27 Annex Table 6. Incidence of Public Spending on Health, Atlantic Region, 1998 Quintile Government Subsidy lotal Per capita Row share Primary-level care 61 subsidy per visit Poorest 11,049,379 65 33% 2 7,442,354 66 22% 3 6,612.646 84 20% 4 4,792.082 56 14% Richest 3,544,444 62 11% Poor 21,863,669 70 65% Non-poor 11,577,110 60 35% Male 16.120.195 65 48% Female 17.320,591 68 52% Total 33,440,905 67 100% Hospital 82 subsidy per visit Poorest 2,712,593 16 18% 2 3,733,001 33 25% 3 2,716,440 34 18% 4 3,366,707 40 22% Richest 2,607,531 46 17% Poor 4,842,392 16 32% Non-poor 10,293,919 54 68% Male 6,682,442 27 44% Female 8,453,642 33 56% Total 15,136,272 30 100% Total Poorest 13,761,972 81 28% 2 11,175,355 100 23% 3 9,329,087 118 19% 4 8,158,789 96 17% Richest 6,151,975 108 13% Poor 26,706,061 86 55% Non-poor 21,871,029 114 45% Male 22,802,637 92 47% Female 25,774,234 102 53% Total 48,577,177 97 100% Annex 22, Page 28 Annex Table 7. Primary and Secondary Public Education Expenditures by Region, 1993 & 1998 Primary Secondary 1993 1998 1993 1998 Public expenditures* Managua 30,134,484 99,973,400 15,246,512 26,004,271 Pacific Urban 28,090,300 54,439,890 Pacific Rural 28,090,300 54,439,890 All Pacific 56,180,599 108,879,781 18,527,127 22,858.240 Central Urban 25,410.828 48,924,771 Central Rural 25,410.828 48,924,771 All Central 50,821,655 97,849,542 13,397,244 16,019,735 Atlantic Urban 7,051,486 12,494,787 Atlantic Rural 7,051,486 12,494.787 All Atlantic 14,102,972 24,989,573 3,561,422 2,922.620 Subventionado 6,470,299 1,614,986 Total (excluding subv.) 151,239,710 331,692,296 50,732,305 67,804,866 Enrollment (Ministry of Education) Managua 142,278 141,069 31,522 71,639 Pacific Urban 108,112 111,558 26,301 56,198 Pacific Rural 116,099 124,881 7,854 9,777 All Pacific 224,211 236,439 34,155 65,975 Central Urban 74,534 79,895 20,357 42,942 Central Rural 120,620 128,621 1,834 1,802 All Central 195,154 208,516 22,191 44,744 Atlantic Urban 16,060 19,540 3,832 7,375 Atlantic Rural 35,341 58,927 496 3,151 All Atlantic 51,401 78,467 4,328 10,526 Total Enrollments 613,044 664,491 92,196 192,884 Survey Enrollments Managua 167,390 166,368 34,591 104,092 Pacific Urban 124,637 129,731 21,376 46,667 Pacific Rural 88,414 146,236 8,965 26.683 All Pacific 213,051 275,967 30,341 73,350 Central Urban 107,336 75,319 11,047 39,930 Central Rural 116,733 170,655 3,288 31,127 All Central 224,069 245,974 14,335 71,057 Atlantic Urban 33,422 43,890 4,588 9,532 Atlantic Rural 47,516 38,637 2,143 2,806 All Atlantic 80,938 82,527 6,730 12,338 Total Survey enrollments 685,448 770,836 85,998 260,837 Annex 22, Page 29 Annex Table 7, continued Primary Secondary 1993 1998 1993 1998 Per Student Subsidy with MOE enrollments Managua 212 709 484 363 Pacific Urban 260 488 - - Pacific Rural 242 436 - - All Pacific 251 460 542 346 Central Urban 341 612 - - Central Rural 211 380 - - All Central 260 469 604 358 Atlantic Urban 439 639 - - Atlantic Rural 200 212 - - All Atlantic 274 318 823 278 NationalAverage 247 499 550 352 Per Student Subsidy with Survey enrollments Managua 180 601 441 250 Pacific Urban 225 420 - - Pacific Rural 318 372 - - All Pacific 264 395 611 312 Central Urban 237 650 - - Central Rural 218 287 - - All Central 227 398 935 225 Atlantic Urban 211 285 - - Atlantic Rural 148 323 - - All Atlantic 174 303 529 237 National Average 221 430 590 260 *Public expenditures at the regional level were not disaggregated by rural/urban location. At the primary level, expenditures are assumed to be equally distributed betwoen rural. This assumption is not, however, made at the secondary level. Annex 22. Page 30 Annex Table 8. All Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1993. Per capita Quintile Total subsidy subsidy Row share Primary 221 per student subsidy Poorest 24,960,646 30 16% 2 31,578,845 38 21% 3 33,428,814 40 22% 4 33,033,953 40 22% Richest 28,481,771 33 19% Male 77,462,047 37 51% Female 74,021,961 35 49% Poor 74,014,447 36 49% Non-poor 77,470,025 36 51%/o Total / Average 151,484,028 36 100% Secondary 590 per student subsidy Poorest 2,195,762 3 4% 2 7,490,062 9 15% 3 13,527,679 16 27% 4 11,434,489 14 23% Richest 16,090,928 18 32% Male 26,003,188 13 51% Female 24,735,774 12 49% Poor 15,290,558 7 30% Non-poor 35,448,203 17 70% Total 50,738,920 12 100% Total Primary and Secondary Poorest 27,156,408 33 13% 2 39,068,907 47 19% 3 46,956,493 57 23% 4 44,468,442 54 22% Richest 44,572,699 51 22% Male 103,465,235 50 51% Female 98,757,735 47 49% Poor 89,305,005 44 44% Non-poor 112,918,228 53 56% Total 202,222,948 48 100% Annex 22, Page 32 Annex Table 10. Managua, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1993. Per capita Quintile Total subsidy subsidy Row share Primary 180 per student subsidy Poorest 1,318,608 29 4% 2 4,768,920 40 16% 3 7,451,820 35 25% 4 8.420,760 32 28% Richest 8,170,020 17 27% Male 16,326,360 30 54% Female 13,803,660 24 46% Poor 9,484,560 76 31% Non-poor 20,645,640 46 69% Total 30,130,128 27 100% Secondary 441 per student subsidy Poorest 226,317 5 1% 2 1,784,771 15 12% 3 3,892,090 19 26% 4 4,014,644 15 26% Richest 5,336,982 11 35% Male 9,139,725 17 60% Female 6,115,347 11 40% Poor 3,113,857 25 20% Non-poor 12,141,171 27 80% Total 15,254,803 14 100% Total Primary and Secondary Poorest 1,544,925 34 3% 2 6,553,691 55 14% 3 11,343,910 54 25% 4 12,435,404 48 27% Richest 13,507,002 28 30% Male 25,466,085 46 56% Female 19,919,007 35 44% Poor 12,598,417 101 28% Non-poor 32,786,811 72 72% Total 45,384,931 41 100% Annex 22. Page 33 Annex Table 11. Pacific Urban, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1993. Per capita Quintile Total subsidy subsidy Row share Primanr 225 per student subsidy Poorest 1,484,685 45 5% 2 3,570,300 47 13% 3 6,405,525 44 23% 4 9,242,775 42 33% Richest 7,339,950 25 26% Male 14,127,750 39 50% Female 13,915,350 35 50% Poor 7,930,350 46 28% Non-poor 20,112,750 34 72% Total 28,043,235 37 100% Secondary 611 per student subsidy Poorest 203,744 6 2% 2 2,004,874 27 15% 3 3,584,920 25 27% 4 3,210,622 15 25% Richest 4,056,857 14 31% Male 6,475,989 18 50% Female 6,584,747 17 50% Poor 3,605,083 21 28% Non-poor 9,455,836 16 72% Total 13,061,017 17 100% Total Primary and Secondary Poorest 1,688,429 51 4% 2 5,575,174 74 14% 3 9,990,445 69 24% 4 12,453,397 57 30% Richest 11,396,807 39 28% Male 20,603,739 56 50% Female 20,500,097 52 50% Poor 11,535,433 66 28% Non-poor 29,568,586 50 72% Total 41,104,252 54 100% Annex 22. Page 34 Annex Table 12. Pacific Rural, Nicaragua: Incidence of public spending on pirmary and secondary education by quintile and gender, 1993. Per capita Quintile Total subsidy subsidy Row share Primary 318 per student subsidy Poorest 8,153,838 59 29% 2 5,653,086 46 20% 3 5,792,052 60 21% 4 3,842.076 50 14% Richest 4,674,600 57 17% Male 13,702,938 52 49% Female 14,412,714 58 51% Poor 17,316,372 55 62% Non-poor 10,799,598 54 38% Total 28.115,652 54 100% Secondary 611 per student subsidy Poorest 699,290 5 13% 2 1,214,179 10 22% 3 1,552,368 16 28% 4 474,753 6 9% Richest 1,537,032 19 28% Male 2,782,983 10 51% Female 2,694,632 11 49% Poor 3,133,514 10 57% Non-poor 2,344,102 12 43% Total 5,477,621 11 100% Total Primary and Secondary Poorest 8,853,128 64 26% 2 6,867,265 56 20% 3 7,344,420 77 22% 4 4,316,829 56 13% Richest 6,211,632 76 18% Male 16,485,921 62 49% Female 17,107,346 68 51% Poor 20,449,886 64 61% Non-poor 13,143,700 66 39% Total 33,593,273 65 100% Annex 22. Page 35 Annex Table 13. Central Urban Province, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1993. Per capita Quintile Total subsidy subsidy Row share Primarp 237 per student subsidy Poorest 2,590,173 51 10% 2 5,217,318 58 21% 3 6,609,219 55 26% 4 5,423,745 43 21% Richest 5,598,177 33 22% Male 12,678,078 50 50% Female 12,760,791 42 50% Poor 11,757.807 57 46% Non-poor 13,681,062 39 54% Total 25,438,632 46 100% Secondary 935 per student subsidy Poorest 781,080 15 8% 2 1,442,986 16 14% 3 2,858,295 24 28% 4 1,447,193 12 14% Richest 3,799,747 22 37% Male 3,788,714 15 37% Female 6,540,606 22 63% Poor 3.603,116 17 35% Non-poor 6,726,203 19 65% Total 10,329,300 19 100% Total Primary and Secondary Poorest 3,371,253 67 9% 2 6,660,304 74 19% 3 9,467,514 79 26% 4 6,870,938 55 19% Richest 9,397,924 55 26% Male 16,466,792 65 46% Female 19,301,397 64 54% Poor 15,360,923 74 43% Non-poor 20,407,265 58 57% Total 35,767,932 64 100% Annex 22. Pane 36 Annex Table 14. Central Rural Province, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1993. Per capita Quintile Total subsidy subsidy Row share Primary 218 per student subsidy Poorest 9,336,940 28 37% 2 8,688,826 32 34% 3 4,064,392 27 16% 4 2,320,392 28 9% Richest 1,037,331 16 4% Male 12,798,562 28 50% Female 12,649,232 28 50%zO Poor 20.176,118 29 79% Non-poor 5.271,894 25 21% Total 25,447,881 28 100% Secondary 935 per student subsidy Poorest 695,154 2 23% 2 601,504 2 20% 3 914,280 6 30% 4 535,362 6 17% Richest 327,774 5 11% Male 1.212,602 3 39% Female 1,861,492 4 61% Poor 1,937,788 3 63% Non-poor 1,136,306 5 37% Total 3,074,074 3 100% Total Primary and Secondary Poorest 10,032,094 30 35% 2 9,290,330 34 33% 3 4,978,672 33 17% 4 2,855,754 35 10% Richest 1,365,105 22 5% Male 14,011,164 30 49% Female 14,510,724 33 51% Poor 22,113,906 32 78% Non-poor 6,408,200 31 22% Total 28,521,956 31 100% Annex 22, Page 37 Annex Table 15. Atlantic Urban Province, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1993. Per capita Quintile Total subsidy subsidy Row share Primary 211 per student subsidy Poorest 448,776 45 6% 2 1,156,850 58 16% 3 1,753,326 54 25% 4 2,141,650 41 30% Richest 1,551,525 28 22% Male 3.599,871 43 51% Female 3,452,382 40 49% Poor 2,641,720 56 37% Non-poor 4.410,322 36 63% Total 7,052,126 41 100% Secondary 529 per student subsidy Poorest - 0 0% 2 330,218 17 14% 3 660,351 20 27% 4 589,359 11 24% Richest 846,929 15 35% Male 896,972 11 37% Female 1,529,815 18 63% Poor 624,432 13 26% Non-poor 1,802,356 15 74% Total 2,426,856 14 100% Total Primary and Secondary Poorest 448,776 45 5% 2 1,487,067 75 16% 3 2,413,676 74 25% 4 2,731,009 52 29% Richest 2,398,454 43 25% Male 4,496,843 53 47% Female 4,982,197 58 53% Poor 3,266,152 69 34% Non-poor 6,212,678 51 66% Total 9,478,983 56 100% Annex 22. Pa2-e 38 Annex Table 16. Atlantic Rural Province, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1993. Per capita Quintile Total subsidy subsidy Row share Primary 148 per student subsidy Poorest 2,588,964 29 37% 2 2,278,904 22 32%/O 3 1,234,172 18 18% 4 866,229 19 12% Richest 64,112 7 1% Male 3,649,680 22 52% Female 3,382,688 24 48% Poor 5,599,284 23 80% Non-poor 1,433,099 21 20% Total 7,032,381 22 100% Secondary 529 per student subsidy Poorest 80,207 1 7% 2 300,774 3 27% 3 217,699 3 19% 4 534,713 12 47% Richest - 0 0% Male 608,244 4 54% Female 525,159 4 46% Poor 380,880 2 34% Non-poor 752,238 11 66% Total 1,133,393 4 100% Total Primary and Secondary Poorest 2,669,171 30 33% 2 2,579,678 25 32% 3 1,451,871 21 18% 4 1,400,942 31 17% Richest 64,112 7 1% Male 4,257,924 25 52% Female 3,907,847 27 48% Poor 5,980,164 25 73% Non-poor 2,185,337 31 27% Total 8,165,774 26 100% Annex 22. Page 39 Annex Table 17. All Nicaragua: Incidence of public pending on primary and secondary education by quintile and gender, 1998. Per capita Quintile Total subsidy subsidy Row share Primary 430 per student subsidv Poorest 76,970,774 79 23% 2 79,588,270 82 24% 3 77,718,673 79 23% 4 63,674,099 65 19% Richest 33,507,608 34 10% Male 168,242,660 70 51% Female 163,216,820 65 49% Poor 190,244,470 82 57% Non-poor 141,214,666 55 43% Total / Average 331,459,424 68 100% Secondary 260 per student subsidy Poorest 4,066,800 4 6% 2 10,081,711 10 15% 3 16,475,677 17 24% 4 20,195,781 21 30% Richest 16,997,630 17 25% Male 31,089,526 13 46% Female 36,728,510 15 54% Poor 18,591,950 8 27% Non-poor 49,225,852 19 73% Total 67,817,599 14 100% Total Primary and Secondary Poorest 81,037,574 84 20% 2 89,669,981 93 22% 3 94,194,350 95 24% 4 83,869,880 86 21% Richest 50,505,238 51 13% Male 199,332,186 83 50% Female 199,945,330 80 50% Poor 208,836,420 90 52% Non-poor 190,440,518 74 48% Total 399,277,023 82 100% Annex 22, Page 40 Annex Table 18. Nicaragua: Incidence of public spending on technical and university education by quintile and gender, 1998 (using Survey enrollments) Per capita Quintile Total subsidy subsidy Row share Technical School 10,147 per student subsidy Poorest - 0 0% 2 11,449.875 12 25% 3 12,462.545 13 27% 4 9.977,444 10 22% Richest 11,573,668 12 25% Male 11,975,388 5 26% Female 33,488,651 13 74% Poor 18.151,359 8 40% Non-poor 27,312,578 11 60% Total 45.463,532 9 100% University 9,142 per student subsidy Poorest 4,783,861 5 1 % 2 6,033,236 6 2% 3 49,912,826 51 14% 4 73,426,703 75 21% Richest 214,698,283 217 62% Male 180,272,146 75 52% Female 168,582,854 67 48% Poor 25,016,930 11 7% Non-poor 323,838,253 126 93% Total 348,854,909 71 100% Total Technical and University Poorest 4,783,861 5 1% 2 17,483,110 18 4% 3 62,375,371 63 16% 4 83,404,147 85 21% Richest 226,271,951 228 57% Male 192,247,534 80 49% Female 202,071,506 81 51% Poor 43,168,289 19 11% Non-poor 351,150,831 137 89% Total 394,318,441 81 100% Annex 22, Page 41 Annex Table 19. Managua, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1998. Per capita Quintile Total subsidy subsidy Row share Primary 601 per student subsidy Poorest 5,853,620 121 6% 2 12,735,791 112 13% 3 27,611,142 116 28% 4 34,958.367 94 35% Richest 18,828,128 38 19% Male 50,894,483 84 51% Female 49,092,685 74 49% Poor 26.679,592 115 27% Non-poor 73,306,975 71 73% Total 99,987,048 79 100% Secondary 250 per student subsidy Poorest 282,650 6 1% 2 2,013,800 18 8% 3 4.744,000 20 18% 4 10,273,000 28 39% Richest 8,709,500 17 33% Male 11,704,000 19 45% Female 14,319,000 22 55% Poor 2,857,750 12 11% Non-poor 23,165,250 22 89% Total 26,022,950 20 100% Total Primary and Secondary Poorest 6,136,270 127 5% 2 14,749,591 129 12% 3 32,355,142 136 26% 4 45.231,367 121 36% Richest 27,537,628 55 22% Male 62,598,483 103 50% Female 63,411,685 96 50% Poor 29,537,342 127 23% Non-poor 96,472,225 93 77% Total 126,009,998 99 100% Annex 22. Page 42 Annex Table 20. Pacific Urban, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1998. Per capita Quintile Total subsidy subsidy Row share Primary 420 per student subsidy Poorest 8,633,520 88 16% 2 14,413,140 99 26% 3 14,587,860 73 27% 4 11,146,800 62 20% Richest 5,705,700 29 10% Male 28.698,180 74 53% Female 25,789,260 60 47% Poor 29,702,820 92 55% Non-poor 24,784,200 50 45% Total 54,487,020 66 100% Secondary 312 per student subsidy Poorest 1,370,959 14 9% 2 1,822,735 13 13% 3 5,053,152 25 35% 4 3,798,288 21 26% Richest 2,515,063 13 17% Male 6,609,408 17 45% Female 7,950,696 18 55% Poor 4,923,672 15 34% Non-poor 9,636,432 19 66% Total 14,560,198 18 100% Total Primary and Secondary Poorest 10,004,479 102 14% 2 16,235,875 112 24% 3 19,641,012 98 28% 4 14,945,088 83 22% Richest 8,220,763 42 12% Male 35,307,588 91 51% Female 33,739,956 78 49% Poor 34,626,492 108 50% Non-poor 34,420,632 69 50% Total 69,047,218 84 100% Annex 22. Page 43 Annex Table 21. Pacific Rural, Nicaragua: Incidence of public spending on primarv and secondary education by quintile and gender, 1998. Per capita Quintile Total subsidy subsidy Row share Primary 372 perstudentsubsidy Poorest 16.610,544 79 31% 2 17,526,408 77 32% 3 11,332,608 67 21% 4 6,671,448 62 12% Richest 2,258,710 45 4% Male 26,317,884 68 48% Female 28,081,536 74 52% Poor 39,243,024 77 72% Non-poor 15,156,768 59 28% Total 54,399,718 71 100% Secondary 312 per student subsidy Poorest 1,087,538 5 13% 2 2,250,113 10 27% 3 2,466,610 15 30% 4 1,843,889 17 22% Richest 676,978 13 8% Male 3,609,840 9 43% Female 4,715,256 12 57% Poor 4,354,584 9 52% Non-poor 3,970,512 16 48% Total 8,325,127 11 100% Total Primary and Secondary Poorest 17,698,082 84 28% 2 19,776,521 87 32% 3 13,799,218 81 22% 4 8,515,337 79 14% Richest 2,935,687 58 5% Male 29,927,724 77 48% Female 32,796,792 87 52% Poor 43,597,608 86 70% Non-poor 19,127,280 75 30% Total 62,724,845 82 100% Annex 22. Page 44 Annex Table 22. Central Urban Province, Nicaragua: Incidence of public spending on primary and secondarv education by quintile and gender, 1998. Per capita Quintile Total subsidy subsidy Row share Primary 650 per student subsidy Poorest 7,684,950 104 16% 2 11,391,250 129 23% 3 9,608,300 107 20% 4 11,074,700 95 23% Richest 9,198,150 63 19% Male 23,875,800 103 49% Female 25,081,550 89 51% Poor 23,705,500 117 48% Non-poor 25,251.200 81 52% Total 48,957,350 95 100% Secondary 225 per student subsidy Poorest 314,438 4 3% 2 1,188,698 13 13% 3 1.962,270 22 22% 4 2,513,025 22 28% Richest 3,005,775 21 33% Male 3,412,350 15 38% Female 5,572,125 20 62% Poor 2,456,100 12 27% Non-poor 6,528,150 21 73% Total 8,984,205 17 100% Total Primary and Secondary Poorest 7,999.388 109 14% 2 12,579,948 142 22% 3 11,570,570 128 20% 4 13,587,725 117 23% Richest 12,203,925 84 21% Male 27,288,150 117 47% Female 30,653,675 109 53% Poor 26,161,600 129 45% Non-poor 31,779,350 102 55% Total 57,941,555 113 100% Annex 22. Page 45 Annex Table 23. Central Rural Province, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1998. Per capita Quintile Total subsidy subsidy Row share Primary 287 per student subsidy Poorest 18,735,073 50 38% 2 12,924,471 46 26% 3 11,085,088 53 23% 4 4,108,979 36 8% Richest 2,124,489 48 4% Male 25,643,163 49 52% Female 23,334,822 47 48% Poor 37,859,031 50 77% Non-poor 11,119,241 41 23% Total 48,978,100 48 100% Secondary 225 per student subsidy Poorest 722,903 2 10% 2 2,393,100 9 34% 3 1,928,520 9 28% 4 837,068 7 12% Richest 1,122,030 25 16% Male 4,196,475 8 60% Female 2,807,325 6 40% Poor 3,359,250 4 48% Non-poor 3,644,550 13 52% Total 7,003,620 7 100% Total Primary and Secondary Poorest 19,457,976 52 35% 2 15,317,571 55 27% 3 13,013,608 62 23% 4 4,946,047 43 9% Richest 3,246,519 73 6% Male 29,839,638 57 53% Female 26,142,147 53 47% Poor 41 ,218,281 55 74% Non-poor 14,763,791 55 26% Total 55,981,720 55 100% Annex 22. Page 46 Annex Table 24. Atlantic Urban Province, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1998. Per capita Quintile Total subsidy subsidy Row share Primary 285 per student subsidy Poorest 3,613,800 70 29% 2 2,306,220 53 18% 3 3.075,720 63 25% 4 2,179,224 41 17% Richest 1,333,544 27 11% Male 6,177,375 52 49% Female 6,331,275 50 51% Poor 6,795,825 62 54%/o Non-poor 5,712,825 42 46% Total 12,508,508 51 100% Secondary 237 per student subsidy Poorest 272,621 5 12% 2 282,954 6 13% 3 590,557 12 26% 4 648,195 12 29% Richest 464,710 10 21% Male 1,177,274 10 52% Female 1,081,739 9 48% Poor 671,066 6 30% Non-poor 1,587,971 12 70% Total 2,259,037 9 100% Total Primary and Secondary Poorest 3,886,421 75 26% 2 2,589,174 59 18% 3 3,666,277 75 25% 4 2,827,419 54 19% Richest 1,798,253 37 12% Male 7,354,649 62 50% Female 7,413,014 58 50% Poor 7,466,891 68 51% Non-poor 7,300,796 54 49% Total 14,767,544 60 100% Annex 22. Page 47 Annex Table 25. Atlantic Rural Province, Nicaragua: Incidence of public spending on primary and secondary education by quintile and gender, 1998. Per capita Quintile Total subsidy subsidy Row share Primary 323 per student subsidy Poorest 4,609,856 39 37% 2 3,816,891 56 31% 3 1,745,524 58 14% 4 2,079,119 64 17% Richest 228,481 27 2% Male 6,378,281 49 51% Female 6,101,470 48 49% Poor 9,559,508 47 77%/o Non-poor 2,920,308 53 23% Total 12,479,871 49 100% Secondary 237 per student subsidy Poorest 206,294 2 3 1% 2 131,206 2 20% 3 119,920 4 18% 4 207,584 6 31% Richest - 0 0% Male 289,306 2 44% Female 375,692 3 56% Poor 393,657 2 59% Non-poor 271,341 5 41% Total 665,003 3 100% Total Primary and Secondary Poorest 4,816,150 41 37% 2 3,948,097 58 30% 3 1,865,444 62 14% 4 2.286,702 70 17% Richest 228,481 27 2% Male 6,667,587 51 51% Female 6,477,162 51 49% Poor 9,953,165 49 76% Non-poor 3,191,649 58 24% Total 13,144,874 51 100% Annex 23, Page I Annex 23 - Between prosperity and poverty: rural households in Nicaragua' By Benjanin Davis and Rinku Murgai EXECUTIVE SUMMARY i. Over the past decade Nicaragua has consistently had among the lowest per capita GDP in Latin America. Much of this poverty is concentrated in the rural sector; the 1993 LSMS showed 75 percent of rural households living in poverty, with a third in extreme poverty (World Bank, 1995). The new round of the LSMS provides evidence, however, that the incidence of extreme and moderate rural poverty was significantly lower in 1998. Perhaps not coincidentally, since 1994 agriculture and livestock has been the most dynamic sector of the Nicaraguan economy. ii. In this background paper we analyze rural households in Nicaragua using data from the 1998 LSMS. The paper has six main sections. We begin in Section 11 with brief background information on the rural sector in Nicaragua. We then present the two typologies, rural and farm, used in this report. These typologies are the primary analytical tool in Section IV, which generally describes the demographic, welfare, and economic characteristics and strategies of rural households. This sets the stage for an econometric examination of the determinants of rural prosperity in Section V. In Section VI, we analyze the change in consumption from 1993 to 1998. We then focus in Section VII on households with access to agricultural and pasture land and examine the impact of technical assistance programs, a key policy instrument, on a variety of outcomes. a. Rural poverty should be confronted from a rural rather than a solely agricultural perspective. Agricultural policy may or may not reduce poverty, since the poorest Nicaraguans may be wage workers, staple crop producers, or both. iii. The rural sector in Nicaragua is highly heterogeneous in its assets, spatial attributes, and economic strategies. Sectoral issues must be dealt with from the perspective of rural development, as the majority of households pursue both on- and off-farm, agricultural and nonagricultural strategies. iv. This heterogeneity suggests a variety of solutions, a perspective adopted by the government. Sectoral initiatives must have clearly defined objectives-for example, alleviating poverty or increasing supply response-and policymakers must take into account how objectives complement or work against each other. For example, technical assistance programs may increase the supply response of small and medium producers but may not help the poorest households at all. v. Similarly, incentives for exportable production may create agricultural wage employment but reduce poverty only slightly, since the benefits will mostly accrue to growers instead of agricultural wage laborers, the poorest of the poor in rural Nicaragua. b. Overall, off-farm activities bring in more income than on-farm activities. ' We would like to thank Florencia Castro-Leal. Carlos Arce, Alain de Janvry, Mario Arana, Mario de Franco, Roberto Godoy, Tom Weins, Isabelle Tsakok, and participants at a seminar in Nicaragua for comments and INTA staff for a description of their programs. Only we as authors are responsible for content and errors. Annex 23, Page2 vi. We find that in general rural farm households compare poorly to rural nonfarm households. Those who work in nonagricultural wage employment and family businesses tend to have higher adult educational levels and be nearer to towns and infrastructure. c. Increasing women's access to education may greatly reduce poverty. vii. Econometric analysis shows that providing all levels of education to women has a significant and positive impact on per capita consumption and income. These results are particularly strong for poorer households. d. Not all off-farm activities have a positive impact on welfare; there is evidence of differential returns to wage labor. viii. A large distinction can be found between better remunerated activities (nonagricultural wage labor and self-employment), which require education and access to infrastructure, and worse-remunerated activities (agricultural wage labor), which require little education. The latter do not seem to lead out of poverty, but rather are part of a subsistence-level survival strategy, particularly when used in combination with farm activities. Therefore, a strategy to increase agricultural exports, which relies on increased demand for agricultural wage labor, may not reduce poverty in the long term. e. Nevertheless, on-farm activities play a fundamental role in assuring prosperity-or survival-for a large segment of the rural population. ix. Agriculture is a key economic activity for most rural households but can play a variety of roles. For medium sized producers it is the principal source of livelihood in terms of both cash liquidity and consumption. For small producers it serves as an insurance policy, a source of consumption, and a complement to off-farm activities. Though returns may be higher off farm, farming provides food security and insurance, and small landowners may be better able to absorh shocks than landless wage laborers. For rural families with lower levels of education, agriculture is also an important source of wage labor demand. x. Livestock production is the secondary farm economic activity for most rural households. For small farmers, livestock holdings can serve as insurance or investment. For large farmers, livestock production is a principal source of cash income. f. Access to land is the key determinant of on-farm economic success. xi. Econometric results suggest that access to land is an important determinant of household welfare, as measured by consumption. This is particularly true for poorer households. Increased access to land-or alternatively, increased returns to land-could thus be one source of the drop in extreme and moderate poverty from 1993 to 1998. xii. The results on rental land, the most prevalent form of access to land for poor farmers, are ambiguous. While the returns to renting and owning in combination appear to be high, renting alone does not increase welfare levels. We were unable to control for selection bias in rental market participation, however, and thus we are limited in the conclusions we can draw regarding this strategy. g. Education increases returns to both on- and off-farm activities. Annex 23, Page3 xiii. Higher levels of male and female education increase total per capita consumption and income, as well as off-farm income and participation. In addition, male primary education brings higher agricultural income. h. Despite government efforts, access to the services provided by agrarian institutions, an important factor in microeconomic success, is still minimal. xiv. The survey results show that, despite high sectoral growth and govemment efforts, microeconomic problems hinder productivity, particularly for small and medium farmers. Agriculture remains primarily extensive rather than intensive. This is due to low levels of modern input use, credit constraints, segmented markets, and the scarcity of agrarian institutions. xv. While technical assistance and use of credit appear to have increased somewhat since 1996, large gaps in access to these services still exist These services may benefit farms, but coverage remains minimal. Participation in producer organizations is a common mode of accessing those services that do exist. xvi. Despite the government's emphasis on commercialization, staple crop markets remain highly segmented. We see no improvement since 1996 in market participation in beans, and the share of households participating in corn markets has fallen. Farm households cite significant transaction costs as a major impediment. i. Technical assistance programs lead farmers to begin using fertilizers and pesticides. xvii. Data limitations prevented us from determining whether technical assistance programs have a direct impact on household welfare. We were also unable to separate effects by government and nongovernmental programs. We were able to show, however, that s technical assistance programs lead farmers to adopt modern technologies that in turn increase supply response and agricultural income. j. Those who lack access to sufficient land or education in rural Nicaragua are destitute. xviii. The worst-off households in rural Nicaragua are minifundistas and agricultural wage workers; that is, households poor in both education and land. It is unclear to what extent these households have benefited from agricultural sector growth, but it is evident that neither agricultural sector growth alone, nor access to agrarian institutions, will bring them out of extreme poverty in the long term. In addition to education and land poor people need economic incentives, markets, and institutional support. k. Increases in per capita consumption from 1993 to 1998 stem principally from changes in family size and regional variations. xix. Decomposing the change in rural per capita household consumption levels from 1993 to 1998, we find that the decrease in family sizes played the most important role in increasing per capita consumption. While increases in the levels of adult education also helped increase consumption over all rural households, the retums to both male and female education actually declined over the period. Finally, higher returns to living outside Managua, particularly to the rural Central region, accounted for an important share of increased consumption. This result could derive from the leading role of agriculture in driving economic growth over the period, as well as increased employment opportunities outside of agriculture. Annex 23, Page4 INTRODUCTION 1. Over the past decade Nicaragua has consistently ranked among the poorest countries in Latin America in per capita GDP. Much of this poverty is concentrated in the rural sector; the 1993 LSMS showed 75 percent of rural households living in poverty, with a third in extreme poverty (World Bank, 1995). The new round of the LSMS provides evidence, however, that the incidence of extreme and moderate rural poverty was significantly lower in 1998. Perhaps not coincidentally, since 1994 agriculture and livestock has been the most dynamic sector of the Nicaraguan economy. 2. In this background paper we analyze rural households in Nicaragua using data from the 1998 LSMS. The paper has five main sections. We begin in Section II with brief background information on the rural sector in Nicaragua. We then present the two typologies, rural and farm, used in this report. These typologies are the primary analytical tool in Section IV, which generally describes the demographic, welfare, and economic characteristics and strategies of rural households. This sets the stage for an econometric examination of the determinants of rural prosperity in Section V. In Section VI, we analyze the change in consumption from 1993 to 1998. We then focus in Section VII on households with access to agricultural and pasture land and examine the impact of technical assistance programs, a key policy instrument, on a variety of outcomes. We finish with the policy implications of our analysis. 3. Our basic unit of analysis is the household. Household level analysis of the rural sector is justified first by the extent of household level heterogeneity. Households vary in their control over productive assets, access to agrarian institutions and public infrastructure, and price bands and transaction costs in both product and factor markets. Second, given the structure of families in Nicaragua and indeed most of Latin America, we consider the household to be the basic decisionmaking unit.2 4. Household level data in the rural or agricultural sector in Nicaragua are sparse. A few small household surveys were conducted during the 1980s. In 1993, the World Bank (1995) sponsored the first LSMS in Nicaragua, but the agricultural module was very limited. Nitlapan-UCA (1995) carried out a regional household survey, focusing on land titling. The FAO, in conjunction with the European Union and the MAG, sponsored a national household survey in late 1996 (Davis, Carletto, and Sil, 1997). A follow-up to this survey by the World Bank and the Nicaragua government, focusing on land market issues, is planned for the year 2000. In 1995, FIDEG carried out a nationwide survey of 6,000 households, focusing on gender issues (Renzi and Agurto, 1996). Finally, Ruben, Rodriquez, and Cortez (1999) refer to a 1998 nationwide, 500- observation survey focusing on farm incomes, with which we are not familiar. 5. In terms of design, the FAO survey is closest to the present LSMS, but covers only those households with land. The number of observations and level of representativity in each survey is similar, though the sample designs are very different.3 Comparisons must be taken with some caution, however, and only as an indication of change. We do not expect to find much change over the two surveys, given that only two years separate them, with no major exogenous shocks. We hope to gauge the extent and impact of certain agricultural policy reforms initiated by the 2 Though obviously individual choice exists, and bargaining takes place in the interior of households, the household as a decisionmaking unit remains the most logical and tractable conceptualization of family decisionmaking for our purposes. See Steiner, 1995. for a description of the FAO sample design. Annex 23, Page5 government since the FAO survey. We refer to the FIDEG study, to which we had access only to tabulated data. BACKGROUND 6. While a decade of reform has stabilized Nicaragua's macroeconomy, austerity measures and adjustment policies of the government of President Violeta Chamorro (1990-1996) worsened rural povertv (World Bank, 1995). These programs reduced public spending and the fiscal defic:t, restricted credit, privatized more than 350 state enterprises, liberalized the financial sector, liberalized domestic and foreign trade, and drastically downsized the state's role in agriculture (World Bank, 1995). These policies have been continued, in large measure, by the administration of President Arnoldo Aleman, which signed agreements with the IMF in 1997 and 1998. 7. The highi level of rural poverty has led the current government to target the agricultural an J livestock sectors as keys to Nicaragua's recovery and economic growth (MAG, 1998). This strategic focus appears sound, as Nicaragua is among the countries in Latin America with the highest share of the primary sector in GDP (29 percent in 1999), with approximately 35 percenl of the economically active population working in this sector during the 1990s. Agriculture and livestock constitute the most dynamic sector of the economy, with an average annual growth rale of more than 7 percent, and more than 4 percent in per capita terms, from 1994 to 1998 (BCN, 1999). 8. Despite high growth rates, the experience within the sector itself is mixed. Two background studies (Valdes and Bastos, 1999; Rose and Neira, 1999) show that in real terms exportable crops have outperformed importable crops, primarily corn and beans, even though beans have higher protection rates. While the real price of exportable crops has increased 25 percent since 1994/94, the real price of importables has fallen 20 percent. Rose and Neira argue that while overall growth in yields was evident, growth in exportables was largely dues to expansion of cropped area. Valdes and Bastos argue that stagnating labor productivity limits agriculture's role in reducing poverty. They also argue that reducing the protection of corn and beans will increase export production, thus increasing demand for agricultural wage labor. However, this increased demand will reduce poverty only if wages increase concomitantly. 9. The Nicaraguan government has shifted policy on the agricultural and livestock sector dramatically over the last two decades. During the Sandinista regime (1979-1990) the government intervened heavily in the agricultural sector. The Chamorro government (1990-96) sought to reverse this policy by drastically reducing credit, liberalizing input prices, curtailing the government's technical assistance services, and liberalizing foreign and domestic output markets (Spoor, 1995, and Banco Central de Nicaragua). As in other countries undergoing similar reforms, neither the government nor the private sector promoted institutions to bring about competition in input and output markets, nor did credit or technical assistance increase. As such, markets became highly segmented, and few households had access to services (Davis, Carletto. and Sil, 1997). The current government has acknowledged these microeconomic problems, and at least in its public discourse and programmatic documents has identified them as key bottlenecks to rural development in Nicaragua. Thus, as in other Latin American countries following stabilization and adjustment, Nicaragua's rural development requires more attention to microeconomic problems that inhibit the productivity and response capacity of producers. Among these problems are failures in labor, land, and product markets; the absence of agrarian institutions; and the lack of public investment (de Janvry and Sadoulet, 1997). Annex 23. Page6 AN ASSET-BASED TYPOLOGY OF RURAL HOUSEHOLDS 10. A typology of rural households is useful as a tool in analyzing the LSMS data and has been used before in Nicaragua. The most important date from the early years of the Sandinista period. Competing typologies differed in their interpretation of the importance of control over land, the hiring in of labor, and participation in off-farm activities. These typologies formed part of a debate on the agrarian structure in Nicaragua, which ultimately had important implications for the formation of policy during the Sandinista administration.4 11. A new generation of agricultural and livestock producer typologies have been constructed for Nicaragua over the last few years. The key tradeoff inherent in building typologies of use in policymaking is between detail/disaggregation and statistical representativity. Davis, Carletto, and Sil (1997) construct two basic typologies, one based on land use, with five categories, and the second on cattle ownership, with four categories. These typologies were chosen for two reasons. First, land and livestock constitute the most important productive assets at the disposal of producers. Second, this categorization is statistically representative; that is, inferences may be drawn from these categories about similar producers nationwide. Secondary typologies were developed to analyze specific issues, such as corn and bean market participation and participation in off-farm activities. 12. Maldidier and Marchetti (1996) take a different approach, building upon the typologies of the early 1980s, and construct a disaggregated typology with 21 categories, based primarily on land and cattle ownership, agroecological conditions, and labor supply and demand. Such a typology, using data from case studies, provides a closer and more detailed approximation of producer types, and is more practical when developing policy interventions targeted to specific groups of producers.' 13. However, such a detailed typology lacks statistical representativity. Inferences cannot be made with statistical certainty from case study data to the nation as a whole. Nitlapan-UCA (1995) attempts to apply the Maldidier and Marchetti typology to a large sample of producers, but most of the typology's categories have too few observations to make reliable inferences. 14. For this purpose we construct two statistically representative typologies. The first broadens the scope of agricultural producer or livestock typologies described above to include all rural households. The second is limited to those farm households who owned or used land for agriculture or livestock production during the survey period. 15. The rural household typology takes into account the principal productive assets to which households have access. In rural Nicaragua these are land, cattle, and human, the latter divided into labor experience and education. Heads of cattle tend to be associated with access to land, so we do not need them both. We have no characterization of labor experience beyond current activity, so we cannot use this variable, but education is a good proxy for labor market participation. We are thus left with two variables, land and education. We consider that these two exogenous assets determine in large part the choices made by Nicaraguan rural households, and for this reason we use them to divide households into categories, expecting to find divergence in key choice variables. ' See Maldidier and Marchetti (1996) for a discussion of this debate. s In fact IFAD uses this typology in the targeting of beneficiaries for the Agricultural Technology and Technical Education Project (IFAD. 1999). Annex 23, Page7 16. Land assets are determined by what land households controlled during the survev period, regardless of whether the land is owned or rented. Rather, we categorize households b' farrm size. Nonfarm households are categorized by the average educational level of adults in the household. We also add a category of urban farm households for comparison. 1 7. We use the farm typology to analyze issues specifically related to agricultural and livestock production. In this typology we distinguish between owners and renters of land. If land markets were perfect, then there would be no useful distinction between owned and rented land, as operated land would not be determined by land ownership. Obviously land markets are not perfect, but are imperfect to varying degrees, which may be governed by parcel size. The rental market is very active in Nicaragua, particularly for small parcels. thus blurring the distinction owner and renter. But agricultural households are often constrained in credit and insurance markets, which may imply further differences between owners, renters, and sharecroppers, again by parcel size. We believe that these constraints are sufficiently binding to merit separation of rental and owned households. Finally we mix urban farm households with rural farm household in the farm typology, as our interest is in the agricultural and livestock sector, and not rural households in general. The distribution of households by the two typologies can be seen below, in Tables I and 2. Table A23.1 - Rural household typology % of % of total units number households land reported Total 2093 100 100 Farm households Minifundia 0-2 mzs 438 21 2 Small 2-5 mzs 229 11 3 Medium 5-20 mzs 229 11 1 1 Large >20 mzs 164 8 59 Non farm households Low education <4 years 408 19 0 High education >4 years 435 21 0 Urban farm households 189 9 24 Annex 23, Page8 Table A23.2 - Farm household typology % of % of total units number households Land reported Total 1254 100 100 Owner households Minifundia e-2 mzs 180 14 1 Small 2-5 mzs 186 15 3 Medium 5-20 mzs 203 16 10 Large >20 mzs 177 14 74 Renter households Minifundia e-2 mzs 346 28 2 Other >2 mzs 162 13 10 18. All farm sizes are in adjusted manzanas.6 Note that the largest landholding size, greater than 20 manzanas, is a much smaller definition of "large" than is usually employed in Nicaragua' The spatial distribution of households by the rural and farm typologies can be found in Tables 3 and 4. Farm households are located predominantly in the Atlantic and Central regions; nonfarm households, in the Pacific region and Managua. 19. For each topic we use a series of typologies to contrast households that have, for example, chosen different strategies, or have differential access to services. CHARACTERISTICS OF RURAL HOUSEHOLDS Economic activities 20. Rural households in Nicaragua are heterogeneous and use a variety of economic strategies, as found in Table 5. The key economic activities include wage-earning activities, self- employment, agricultural production, and livestock production. Off-farm activities prevail. In more than 75 percent of rural households, and in more than 60 percent of farm households, at least one household member works off the farm. This represents an increase from the FAO survey, where in 50 percent of farm households at least one household member worked off the farm. This increase appears to derive principally from increased off-farm wage activities. " In aggregating land types to construct a total farm size, we adjusted for differences in quality of land. The following weights were given-perennial (1), annual (.75), pasture (.5), and forest (.25). The relative importance of each type was determined by regressing total household income on land owned. by type, using the FAO data. The results can be found in Davis, Carletto, and Piccioni (1999). We do not use exact relationships, as nonlinear effects are difficult to sort out. 7 We have some questions as to how the monte and bosque categories, which are very low, compare to the tacotal and bosque categories of previous surveys. Annex 23. Page9 21. Agricultural wage labor is the most frequent off-farm activity, with 40 percent of households with at least one member participating, followed by nonagricultural wage labor with 39 percent, and self employed, 23 percent.! A small part of these activities involved migration, only I I percent of all households, with only 4 percent of households having at least one member who migrated to urban areas, and barely 2 percent migrating to another country. 22. Analysis of household income in Table 6 for rural households, and Table 7 for farm households, shows a similar trend.9 Among rural households, off-farm income dominates on-farm income in share of total household income. Among rural households, almost 60 percent of income comes from off-farm labor, while for farm households, the corresponding number is 44 percent. Though participation rates are similar, the share of nonagricultural wage labor is more than four times as high as agricultural wage labor for both rural and farm households. This is due primarily to differences in wage rates (liahi, 1999). Income from self-employment accounts for 18 percent of average total rural income, while other off-farm sources, principally imputed rent, constitute 21 percent. 23. In agricultural and livestock production (Table 5), 57 percent of rural Nicaraguan households planted crops during the survey period. The bulk of agricultural production involved corn, beans, and fruit. Another 17 percent of households witlhout access to agricultural lands had backyard production, almost exclusively in fruit. More than 20 percent of rural households owned cattle during the survey period, more than a third kept pigs, and more than 60 percent some kind of fowl. For farm households, these percentages increase to 33 percent for cattle, 49 percent for pigs, and 76 percent for fowl. 24. Agricultural and livestock production, however, account for only 22 percent of average ru-al income, and 36 percent among farm households, much smaller than the 75 percent share found by the FAO survey. Part of this difference is undoubtedly due to methodological differences in collecting information and calculating agricultural and livestock figures; further, the FAO numbers did not include imputed rents."' Given the increased participation in off-farm activities since 1996, combined with some evidence of lower returns to agriculture, it is not surprising that the share of agricultural and livestock production in total income has diminished. 25. Participation and share of income is conditioned by access to productive assets, particulariv land and education, the two elements upon which our typology of rural households is built. We would expect higher education to be associated with increasingly nonagricultural forms of off- farm labor. We expect lower off-farm participation and share of income with increased farm size, given increasing demands on family labor. In Table 5, we presented the distribution of households in economic activities by the rural household typology. Nonagricultural wage labo- is most prevalent among households without access to land and is associated with higher levels cf education. While 40 percent of households in the low education category participate in nonagricultural wage labor, more than 68 percent in the high education category do so. Similarly, a greater share of higher education households derived their income from self-employment. Conversely, while 55 percent of households in the low education category participate in agricultural wage labor, the percentage drops to 38 percent for higher educated households. ' Self-employment is defined as receiving cash income from an own business activity, not including agricultural or livestock production. ' See Sobrado. 1999, for details on the construction of the income aggregate, as well as for in-depth analysis of the income data. Corral and Reardon, 1999. using their own version of the income aggregate, find similar trends. "' Indeed, Corral and Reardon find a higher share of agricultural and livestock income in total income among rural households, around 34 percent. Annex 23, PagelO 26. For farm households, Table 5 shows that only 32 percent of the smallest landholders work in nonagricultural wage labor. For the largest landholders that number drops to 15 percent. Agricultural wage labor also falls with increasing land size, though maintaining at a higher level. Self-employment is greatest amongst medium size farmers, and households with higher levels of education. In contrast, though not surprisingly, household members who consider themselves patrones are most prevalent among larger landholders and among urban landholders. 27. Agricultural activities are most prevalent among the middle sized landholders, both overall as well as by crops. More than 80 percent planted corn during the survey period, and almost 60 percent grew beans. We analyze agricultural production in more detail below. Planting percentages are lower for small holders, who tend to be single crop producers, and for the largest landholders, who are primarily cattle producers. This is confirmed by Morduch's (1992) plot diversification index, which shows that the largest and smallest landholders are the least diversified in plot numbers and size." The share of households with cattle increases with land size. reaching 65 percent of the largest category. Similar increases are seen with pig production. Access to land appears to be correlated with welfare. In Figures I (over rural households) and 2 (over landholding households), farm size increases with consumption levels. Demographic characteristics of rural households 28. As shown in Table 8, rural households in Nicaragua are primarily headed by males, have low educational levels, and have large families. Only 17 percent of households are headed by females, and the heads of households, on average, have less than three years of education. Households have on average 5.8 members, with 3.4 adults. The household head is, on average, 45 years old. Differences in the demographic characteristics among farm households are minimal. When considered by land size, no clear patterns emerge in gender, age, or education of the head of the household, nor in family size, with the exception of the youthfulness of minifundia households. A similar observation can be made for nonfarm households, with the exception again that households with higher levels of education have heads of household who are significantly younger than the households with less education. 29. Contrasts emerge, however, when comparing farm with nonfarm and urban farm households. More than a quarter of nonfarm households are headed by females and have on average almost double the level of education. These households are significantly smaller, both overall as well as in terms of adults. Urban farm households, on the other hand, have far fewer female headed households, but similar levels of education as rural nonfarm households. In family size, they are almost identical to rural farm households. 30. When viewed by deciles of consumption, as in Figure 3, the share of female headed households shows a U- shaped trend, indicating that female-headed households are among the poorest as well as wealthiest. The poorest are primarily farm households and poorly educated, while the wealthier female heads of household are mostly nonfarm with higher levels of education (not shown). Level of education, in Figure 4, both for household heads and average adult, increases slowly, but consistently with higher deciles of consumption, confirming trends found in the tables shown so far. A similar trend is evident when we restrict the sample to farm households (not shown). Seen another way, classifying male and female education attainment by poverty grouping, in Table 9, the link between poverty and the lack of education is clear. " We were unable to do a crop diversification index since the data do not provide area dedicated to each crop. Annex 23, Page 1I Welfare indicators 31. In Table 10, we present a variety of welfare indicators by rural household typology. All three groups of indicators-ownership of consumer durables, dwelling characteristics, and per capita consumption-illustrate that increasing the two basic productive assets used in the typology, land and education, increases levels of welfare. For landed and renting households. the differences are most evident between the largest landholding group and the smaller categories. A higher share of the largest landholding households have refrigerators, color televisions, and stereos, as well as most other durables, not shown in the table. For dwelling characteristics, differences are not that evident. This is primarily an issue of spatial considerations; as will be discussed below, this group is made up primarily of prosperous large-scale farmers living in relative isolation. The clearest differences are found in per capita household consumption, whicl increases linearly by land size, and levels of extreme and moderate poverty, which decrease linearly by land size. 32. The comparative welfare status ofthe largest landholding group is eclipsed, however, by 1re nonfarm and urban farm households. Rural nonfarm households have significantly higher levels of welfare than farm households, by each of the three groups of indicators. Among the rural nonfarm households, more education is associated with much higher levels of welfare, again by al I three groups of indicators. A quarter of these households have color televisions, and more than 60 percent have running water. Only 4 percent live in extreme poverty, and 14 percent in moderate poverty. The general correlation between improved dwelling characteristics and ownership of consumer durables, and deciles of consumption, can be found in Figures 5 and 6. 33. Despite these trends, dwelling and consumer durables levels do not tell all, and some interesting counterintuitive results are evident. The largest landholding group clearly consumes more than those groups with less land. This group also consumes more than the low education, nonfarm households. These nonfarm households tend to have higher levels of some durables and dwelling characteristics. This has to do in part with spatial considerations as mentioned above. Many of the households in the large landholder category reside in relatively isolated communities. 34. High variability in welfare is noticeable in off-farm activities. Similar to the results found by the FAO survey, agricultural wage laborers appear to live in poverty and have a lower per capita level of consumption. Nonagricultural wage laborers and the self-employed have higher levels of household consumption, as shown in Figure 7. The share of households involved in agricultural wage labor falls linearly with increasing consumption deciles. The opposite is true of nonagricultural wage labor and owning a business. Identifying groups of poor 35. Tables I I and 12 compare the characteristics of rural (excluding urban landed) households sorted into four land-labor groups: households involved in neither nonagricultural nor agricultural wage labor; those involved only in nonagricultural wage labor; households involved only in agricultural wage labor; those involved in both. Each category is subdivided by access to land.12 With regard to landed agricultural and nonagricultural wage households (columns 5 and 7), households involved in nonagricultural wage labor have significantly more land and higher levels of education than agricultural wage households. Higher levels of these assets translate into 2 These groupings represent decisions made by households under different constraints. They do not represent dynamic strategies, and we cannot determine causality. Annex 23, Page 12 significantly higher levels of per capita consumption, accompanied by lower extreme and moderate poverty rates. These households also have more access to credit, as well as better dwelling characteristics and more consumer durables. 36. With regard to households with agricultural and nonagricultural labor and without access to agricultural land (columns 6 and 8), similar differences can be observed. Agricultural wage households have lower levels of education, fewer self employed members, lower levels of credit use, significantly lower levels of consumption, higher incidence of poverty, and smaller quantities of durables. 37. Among nonagricultural wage households. those without land have significantly higher levels of per capita consumption than those with land. Further, these households also have higher levels of education, as well as lower levels of extreme poverty. Among agricultural wage households, however, despite differences in education levels, there is no significant difference in per capita consumption levels. While agricultural wage households with land have higher levels of extreme poverty than landless agricultural wage households, the reverse is true for moderate poverty levels. 38. It should be noted that in both cases described above, a strong regional factor is present, as nonagricultural wage households are concentrated in (rural) Managua and the Pacific, while agricultural wage households are located primarily in the Central and Atlantic regions. 39. The 1993 LSMS identified two groups of rural poor: landless agricultural wage households and small farmer households (World Bank. 1995). We find similar groups to be poor in 1998.'3 In fact, the only segments of the rural population that could not be considered poor as a group are nonagricultural wage households, and the nonfarm households who do not participate in wage activities (column 4). The 1993 LSMS found that the small farmer households were significantly poorer, in a variety of dimensions, than the landless agricultural wage households (columns 2 and 8). We find similar results, located in the same table, but none of the differences are significant.'4 40. We did identify another group, however, to be the most destitute in rural Nicaragua: households that combined both agricultural wage and farm activities (column 7). Almost 75 percent of these households are considered poor, and 39 percent live in extreme poverty, the highest of all categories. The household head has little more than one year of education on average, the lowest of all categories. These households include both landowners (45 percent) and renters (61 percent) and are located predominantly in the Central region. 41. Finally, one group of prosperous households that emerges are those that neither have land nor participate in on- or off-farm wage activities (Column 4). These households are dedicated primarily to family businesses, and almost 40 percent receive transfers. A large share of these households are headed by older women. Land ownerskip, renting, and market transactions 42. For Nicaraguan farmers both land ownership and renting play crucial roles in production strategies. Sixty percent of farm households own land, while 47 percent either rent or borrow, as shown in Table 13. Average amount of land owned, over all households, was almost 15 (adjusted) manzanas, while average rental size was just under three manzanas. The bulk of owned land is in '3We define a group as poor if it has a higher total poverty rate (50 percent) than the overall rural sample. " Our definition ofthe categories could be slightly different, resulting in the nonsignificant results. Annex 23, Page 13 annuals and pastures. The vast majority of rental plots are under five manzanas, while almost half of owned plots are under five manzanas. Only 6 percent of households rented out land, for an average of .28 manzanas, suggesting that the sample or the survey instrument is not adequately catching this aspect of the rental/loan markets. We have included urban farm households, who make up approximately 15 percent of the total landed sample, in this and following tables. 43. When viewed by land size, it is evident that the smallest landholders hold primarily annual land; the share and amount of permanent and pasture land increases with land size. The share of owner households that also rent in land ranges from 10-16 percent among the smaller categories and falls to 5percent in the largest category. For the two smallest landowning groups, rental land accounts for between 16 and 12 percent of total land controlled. Merging owners and renters together, rental land accounts for 20 and 25 percent of total land in these categories. Conversely, and not surprisingly, the percentage of households that rent out increases with farm size. Similarly, while the share of households who sold land in the last five years is relatively constant around 5 percent, the larger land groups show signs of recent land accumulation, with between 20 and 25 percent purchasing land in the last five years. Both buying and selling of land also appeal positively correlated with per capita consumption, as illustrated in Figure 8. A similar trend for farm households is seen in Figure 9. Deininger, Lavadenz, and Zegarra (1999) provide tentative nonparametric evidence that land purchases are dominated by farmers with more than 50 manzanas of arable land. While this requires more detailed study, the accumulation of land by larger landowning and wealthier households may threaten the gains in equity of more than two decades of land reform in Nicaragua.15 44. Gini coefficients were calculated, in Table 14, to gauge the level of inequality in the major household assets. Surprisingly, Nicaragua, even after many years of agrarian reform, has a very unequal distribution of land (.86), confirming what is evident in Table 116 Looking at the overall figures, perennial and pasture are the most unequally distributed; this is due in part to a large number of zeros, particularly in the case of perennials. Table 13 illustrates that the bulk of perennial and pasture land is found in the greater than 20 manzana category!'' 45. Returning to Table 14, inclusion of rental land reduces the overall land Gini coefficient to .80. This is not surprising, as we are replacing zeros with a large quantity of minifundia renters. A similar result was found by the FAO survey. Cattle is also extremely unequally distributed, with an overall Gini of .80. We have added consumption and education Ginis for comparison. Land titling 46. Looking at the titling status of landowners, in Table 15, we find that 22 percent of landowning households do not have any title to their land. A slight majority (52 percent) have escrituras, or private titles, while 13 percent have agrarian reforn titles. Only 64 percent have formally registered their land. These percentages vary by total land size; while the largest landowners tend to have more titled land (principally with escrituras), almost three times as many small landholders as large landholders do not have any title. Similarly, more than 70 percent of the landholders have formally registered their land, compared to 51 percent of the smallest landholders. is While this may represent the most efficient outcome under current conditions, constraints in credit and liquidity may unduly for-e small and medium size farmers to sell out to larger farmers. However, one would need to look at the pretransaction land position of purchasers to see if these are small or medium farmers who are accumulating. 16 We used unadjusted land for the Ginis, since we look at land by type, while Table I is based on adjusted manzanas. '' Comparing the Gini coefficients from 1996 and 1998 (taken only over farm households), which must be interpreted with caution, the Gini coefficient has risen from .72 to .80 in two years. Annex 23, Page 14 47. Land titling status also corresponds to welfare levels, as seen in Figure 10. The share of landowning households with private title increases with quintiles on consumption. Conversely, the share of those households with no title decreases with increased consumption. The share of households with agrarian reform titles is spread evenly across quintiles. Finally, the share of households who have formerly registered their properties increases with consumption. Causality cannot be inferred from this graph, however; wealth can lead people to formalize land titles, as well as vice versa. Manv economists maintain that land titling increases agricultural efficiency and thus welfare (Deininger and Binswanger, 1999). 48. As mentioned earlier, 47 percent of farm households in Nicaragua have access to land without owning it, and 38 percent of the landless access land through rental markets. We use the catch all phrase of rentals, but a variety of contracts are used. As seen in Table 16, almost 40 percent of these transactions are fixed rent, while 47 percent are borrowed. Less than 10 percent are sharecropped. 49. The source of these lands is primarily individually owned plots (65 percent). as seen in Table 17. Twenty-two percent derive from family lands, and less than 10 percent from cooperative lands. Classifying by contract arrangement and land source (Table 18). we find that a higher share of borrowed contracts involve family lands, while the same is true for individual plots and rented land. Classifying by contract arrangement and length of contract. in Table 19, rental transactions tend to be shorter in duration, while borrowed transactions go back further in time, particularly to before 1990. Overall, as seen in Table 20, almost a quarter of nonownership mechanisms date back to pre-1990, and another quarter from 1991-1995. As seen in Table 21, as the original date of the contract becomes more distant, the share of individual lands decreases, while that of cooperative and family lands increases. Comparing owners and renters 50. The presence of many options for households to access land reduces the theoretical differences between landowners and renters. In practice the households differ widely in their characteristics and their access to productive assets. Landowners control more than five times as much land as renters, as can be seen in Table 22. The rental market is vibrant: about 40 percent of agricultural and livestock households depend exclusively on its functioning. But this market is confined in large part to parcels under two adjusted manzanas, and almost exclusively to parcels under five manzanas. With smaller landholdings comes lower participation in livestock activities, while 45 percent of landowners have cattle, only 14 percent of renters do. Similarly, given the lower access to land, renters are involved more in off-farm activities, with the exception of family businesses. Renters and owners have similar shares of female-headed households, but renters tend to be younger and have lower education levels than landowners. This is true even when comparing minifundia owners and renters. 5 1. Few differences in agricultural practices emerge between owners and renters, who have similar shares of households planting crops, including corn, beans, and vegetables. Surprisingly, a high share of renters grow fruit, but very few grow coffee. Owners hire in labor with more frequency and participate more in product markets. Input use is similar, with the exception of fertilizers. Again these similarities, as well as the few differences, hold even when comparing only minifundia owners and renters. Both plant primarily corn and beans, while a significantly lower share of renters cultivate coffee. Neither use modern agricultural inputs, with the exception of fertilizer. Annex 23, Page 15 52. Very surprisingly, similar shares of owners and renters use credit. As we discuss later when looking specifically at credit, collateral is not the important limitation in credit access. Renters consider themselves more constrained by lack of collateral. Owners overall somewhat higher use of technical assistance, particularly from government sources, as well as higher levels of participation in producer organizations. These differences disappear when controlling for land size. 53. Ultimately, land ownership is associated with higher levels of welfare, though this is attributable more to large average landholdings than to ownership per se. Owners have significantly higher levels of consumption, though few differences are found in ownership of durables. This can be seen from a different perspective in Figure 11, where the share of farm households renting in land decreases with increasing consumption decile. Among landowners, however, in Figure 12, the share of households renting in is even across consumption quintiles, illustrating that poorer and richer landowning households alike rent in land. Further, controlling for land size, there are no significant differences in consumption levels. These results are confirmed in the econometrics section. The role of agriculture in rural Nicaragua 54. Agricultural production is a key economic strategy for rural Nicaraguan households. As mentioned earlier, more than half of rural households planted crops during the survey period. Corn (72 percent), fruits (62 percent), and beans (52 percent) are far and away the most frequently planted crops, with sorghum (18 percent) a distant fourth. Only 9 percent of households grew coffee. As shown in Table 23, medium- size farms (2-5 and 5-20 manzanas) show the highest frequencies; minifundia farms are too small to plant a variety of crops, while she presence of livestock farms in the largest grouping reduces the share of households growing a variety of crops. The results coincide with the conclusion of Davis, Carletto, and Sil, who find that medium-size households produce the bulk of Nicaragua's staple food. Little trend is found in planting by consumption deciles, with the exception of a slight drop in staple crop production in the highest deciles, as seen in Figures 13 and 14. Rather, the majority of producers, rich or poor, small or large, produce staple crops. 55. Agriculture contributes only a small share of total household income, as discussed earlier. Such a monetary measure, however, underestimates the importance of on-farm activities to rural households. For minifundia and small producers, on-farm activities complement off-farm activities, allowing these families to spread the risk and to have another source of food. For medium- sized producers, on-farm activities are the principal source of cash income. Livestock complements savings and investment for all types of producers. Finally, for large producers agriculture complements livestock production, as well as family businesses. 56. Only a third of agricultural farms hire nonfamily labor, as seen in Table 22. These farms almost always mix family and hired labor; only I percent rely completely on hired labor (not shown in table). Use of hired labor increases with farm size, though the share of households in the two largest categories are essentially the same, at just under 50 percent. Landowners are more likely to hire in labor than renters, even among farms of similar sizes. Not surprisingly, given the association with land size, the share of farm households hiring in also increases with deciles o f consumption, as observed in Figure 15. 57. The level of use of modem agricultural technology is very low, and remains essentially unchanged from the 1996 FAO survey. As seen in Table 22, only 5 percent of farm households Annex 23, Pagel 6 use HYV seeds, a little more than a third use fertilizer, and less than half use chemicals. Very few differences are found by land size, with the exception of lower fertilizer use by the largest landowning class. Renting households also show lower use of fertilizer. Access to agrarian institutions 58. Agrarian institutions generally foster economic development, and agricultural production in the rural sector. In this section we focus on the three principal services provided by agrarian institutions in Nicaragua-technical assistance, credit, and producer organizations. The descriptive statistics confirm the findings from the FAO survey. regarding the nexus of the receipt of technical assistance, credit, and producer organizations. While overall we find some increase in access since the 1996 FAO survey, the vast majority of agricultural producers do not have access to the services provided by agrarian institutions. Technical assistance 59. Technical assistance has constituted. during the last few years, the principal policy instrument of the Nicaraguan government in support of small and medium sized agricultural producers. We describe these technical assistance programs in more detail in Section Vl. Though access to technical assistance has improved somewhat since 1996 (using FAO and FIDEG results'8), overall use remains quite low. Only 24 percent of farm households have access to technical assistance in their communities, and only 15 percent actually used technical assistance, as seen in Table 24. This technical assistance was supplied in almost equal shares by government and NGO or project sources. A small minority (3 percent of all households, and 13 percent of households receiving technical assistance) reported paying for these services. 60. Overall levels of technical assistance use are low, and some bias is evident in terms of who uses technical assistance. Minifundia renters have disproportionately lower access and use of technical assistance, while medium sized farms have slightly higher usage. Governmental programs appear to be more successful than NGOs in reaching the smallest farms. Landowners use technical assistance, especially government-sponsored assistance, more than renter households do. In Figure 16, technical assistance shows a bimodal trend, with peaks between deciles three and five and eight through ten. Technical assistance programs are also biased towards urban areas. In Table 25 we see that technical assistance exists disproportionately in urban areas-about 50 percent of the Managua households, and 34 percent of the urban Central households report the existence of technical assistance.'9 Technical assistance is also used by a disproportionately higher share of urban households than rural households, and this is particularly true for government-supplied services. Paid services are almost exclusively urban. Both governmental and NGO projects are concentrated in INTA zone B3,which encompasses a disproportionately high share of minifundista and small owner farms, as seen in Table 26. 61. For those who did receive technical assistance, in 86 percent of the cases it was provided by technicians (Table 27). Most of this assistance was geared towards agricultural production. More than a quarter of all services focused on the use of fertilizer and other chemicals, 14 percent on crop diversification, and 12 percent focused on improvement in seed selection (Table 28). Services concerned with livestock production (between improvement in cattle and veterinary advice) accounted for about 13 percent of the programs' content. 18 In the FAO survey, 9 percent of household used technical assistance, while in the FIDEG. 8 percent of adults used technical assistance. '9 The total number of urban households is small, so caution is due in interpreting these results. Annex 23, Pagel 7 62. Thirteen percent of all farm households received a visit from INTA or MAG (Table 29); these visits were somewhat more concentrated among larger landholding groups. Only 7 percen'- of farm households participated in an agricultural technology show or event (Table 30). For tho-,e who did not receive technical assistance,65 percent said that services weren't available to them . while another 12 percent claimed that services were not available close to home (Table 3 1). This suggests that INTA and nongovernment providers have a long way to go to make technical assistance services available to a small and medium-sized farmers. Only 10 percent said services were too expensive. Then again, if services had been available, a higher percentage ma' have found them too expensive. Few differences are evident in this report among categories in the farmer typology. Credit 63. State-sponsored financial services to the agricultural sector have steadily decreased since the heady days of the Sandinista government, while private sources have increased dramatically (BCN, 1999). The percentage of households receiving credit appears to be roughly the same compared to the FAO survey, excluding credit for purchases, which was not collected in that survey. As noted in Table 32, 17 percent of all rural households received some kind of credit.20 Only 12 percent received a loan for productive purposes; another 7 percent obtained purchase credit. Loans for productive purposes were primarily for agriculture and derived principally fromn producer organizations, followed by banks and informal loans from friends or relatives. 64. In sources and uses of credit, there are few differences among land size categories, with the exception of purchase credit for the largest landholder group. Overall, a similar share of farm and nonfarm households receive credit, particularly productive credit. Among nonfarm households however, almost twice as many households with high education receive credit as did those with low education. This contrast derives principally from credit for productive purposes from organizations. Nonfarm households are more credit-constrained,21 principally from lack of collateral. The poor use production credit significantly less and have a significantly higher share of credit-constrained households. 65. Few differences are found between landowners and renters, as seen in Table 33, with the exception of credit constraints. Owners are significantly less credit-constrained, again due to ownership of land. When viewed by deciles of consumption, in Figure 17, access to credit is clearly associated with higher levels of welfare, particularly productive credit or loans. Access ':o credit, more than either of the other two agrarian institutions discussed here, is associated with greater wealth. Finally, 4 percent of rural households loaned money, but only 2 percent had any kind of bank deposit. These low levels of loans and deposits reinforce the perception that the formal banking system has little presence in the Nicaraguan countryside. 66. Households that did not solicit credit during the survey period were asked why they had not done so. Among rural households the most frequently cited reason was lack of collateral (28 percent), as seen in Table 34, in large part by the smallest landholders and nonfarm households. Another 20 percent considered credit too risky and 13 percent too expensive. Only 10 percent said credit was not available in their community. Large landholders cite lack of availability as the biggest constraint, while small holders, primarily renters, cite lack of collateral. Medium-sized farmers are the least constrained. Among nonfarm households, low education households are more constrained than high education households, with the difference divided equally among lack 2" FIDEG reports that only 7 percent of rural adults received credit. 2" The definition of credit-constrained households can be found in Table 21 a, plus those hiouseholds that solicited credit and were denied. as well as credit recipients who were rationed on the amount they received. Annex 23. Page 8 of availability and collateral. Overall. nonfarm households are more constrained, due to lack of collateral. Similar trends are found comparing owners and renters among farm households, in Table 35. Renters are more credit-constrained, again primarily due to lack of collateral. Producer Organizations 67. The last of the three principal agrarian institutions is producer organizations. Only 4 percent of rural households participate in these organizations, with another 2 percent participating in a productive project (Table 36). Far more prevalent are nonproducer organizations, which involve more than a third of rural households. The bulk of these organizations are religious. These figures are confirmed by the FIDEG survey, which collected information on participation in nonproducer organizations. Both producer and nonproducer organizations are more prevalent among farm households than nonfarm households. Between the two nonfarm household categories, however, highly educated households tend to participate more in nonproducer organizations, a trend due primarily to the high frequency of school organizations in this group. Urban households have similarly high levels of participation in nonproducer organizations. 68. Looking at farm households, in Table 37, the larger landowning participates somewhat more in producer organizations. Landowners in general participate much more than renters. No differences stand out with regard to membership in nonproducer organizations. Analyzed bv consumption deciles, in Figure 18, no type of organization appears associated with welfare level. Households in extreme poverty, however, have a significantly lower share of family members participating in either producer or nonproducer organizations. 69. Digging into the detail of household participation in producing organizations reveals evidence of the nexus between the three agrarian institutions. Most of the primary benefits cited from participation in producer organizations involve technical assistance, as shown in Table 38. However, 20 percent of the participants say they derived no benefit from participating in organizations. More than half of the primary benefits from projects also involved technical assistance. This is not surprising since at least for some governmental programs, households had to join together to receive technical assistance, or belong to previously existing producer organizations. Forty-four percent of the projects involved NGOs or other civil organizations, followed by MAG with 23 percent and other government offices with 16 percent. Only 10 percent involved other producer organizations (Table 39). Household characteristics of the users of agrarian institutions 70. In Tables 40-42, we present the household characteristics of the users of agrarian institutions. In Table 40, the data clearly illustrate the nexus of technical assistance, credit, and organization. Among those households that received technical assistance, more than twice as many received credit as compared to those who did not, and more than 10 times as many (31 percent) participated in a producer organization. Similarly, comparing those who received credit with those who did not, more than twice as many households used technical assistance and participated in producer organizations. Finally, among those belonging to producer organizations, more than two-thirds used technical assistance, compared to only 11 percent of those not in producer organizations. This was true for technical assistance supplied by both government and NGOs. More than twice as many received some kind of credit, and almost three times as many received productive credit. Thus, participation in producer organizations appears to be an important avenue to technical assistance and credit. From higher levels of education. These households have a significantly higher frequency of participation in the product market, Annex 23, Page 19 particularly credit-receiving households. These households use more modem agricultural technology and hire in more labor. 71. Households that receive services from any of the three agrarian institutions have higher levels of welfare compared to those who do not, particularly those participating in producer organizations. As shown in Table 42, these households tend to have better infrastructure and dwellings. own more durable consumer goods, and have higher per capita consumption. A significantly lower percentage of these households live in extreme or overall poverty. Participation in product markets 72. Corn and beans are the two primary crops in Nicaragua, both in terms of agricultural output as well as consumption by rural households. Since corn and beans are both cash crops and consumed on farm, analyzing the consumption and production of these crops sheds light on hovw their markets function. 73. The internal market for these two staple crops has been at the mercy of large swings in agricultural policy through the last three administrations. Under the Sandinista government, traditional market intermediaries were displaced by state control over marketing, and prices had a distinctly urban bias. The government of President Chamorro severely curtailed the role of the state in internal grain markets, but neglected to promote the institutions needed to foster competition and transparent markets. Thus, under the government of President Aleman, commercialization has been viewed-by both state and civil actors-as a major bottleneck for staple producers (Davis, Carletto, and Sil. 1997). 74. To analyze staple product markets from the perspective of the household, we characterized producer households by participation in each market in the following fashion: both purchased and sold (the grain or bean, not byproducts such as tortillas); purchased only; self-sufficient; and sold only. Several factors, both structural and temporary, determine the category into which a household falls: * Internal household demand (family size and livestock holdings); * Internal supply (land available and technology utilized); and * Household level transaction costs (determined by the price band, or the difference between the market price and the effective selling and/or purchase price in a recent transaction). 75. Transaction costs may come about through lack of information, difficult access to markets, and monopsony or monopoly market structure. Also, at any given moment bad weather or family sickness may influence market participation. 76. Characterizing households by market participation, we find that the corn market has become more segmented since 1996. During the survey period only 31 percent of corn producing households sold part or all of their output, as shown into Table 43. Six percent also purchased corn. The bulk of producers (45 percent) were self-sufficient; they neither purchased nor sold the grain. The remaining 20 percent purchased grain to complement their output. These numbers represent a reversal from the results of the FAO survey, in which 48 percent of corn-producing households sold production, while only 30 percent were self-sufficient. It is unclear whether this change is due to structural or temporary factors. Annex 23, Page2O 77. While corn and bean output has grown since 1994, increasing supply, real producer prices have dropped 20 percent over the same period, lowering the incentive to sell. Not surprisingly, medium landholders show the highest level of market participation, while in the largest landholders, with a high level of diversification into livestock activities, more than 50 percent are self-sufficient. The smallest land size groups, both owner and rental, had the highest share of households purchasing grain. 78. The story is similar with beans, even though beans are considered more of a cash crop than corn. In this case, 40 percent of households sell beans, and 33 percent are self-sufficient (Table 44). The remaining households, 28 percent, only buy and produce. These numbers are more comparable to the 1996 FAO results, where a combined 42 percent (27 percent sellers only) of households sold part or all of their production. By land size, once again the medium landholders show the greatest level of market participation. Considering buyers and sellers as well, the smallest landowner group also shows considerable market participation (44 percent). Small renters, on the other hand, are more apt to be self-sufficient. 79. In 1996 we found that those households who sold in basic grains markets had higher levels of assets and income than those who did not. We do not find this differentiation when comparing seller with nonsellers in this data. In fact, the share of households in extreme poverty who sell corn or beans is approximately the same as the nonpoor (not shown). 80. These results are disappointing, given recent emphasis on improving commercialization of basic grains. Market access is an important element of rural development and of assuring strong supply response to sector policy initiatives. We find that, at least during the survey period, basic grains production did not provide this access. Market segmentation can result from high transaction costs and low prices, which combine to induce households to consume most of their production. While we do not conduct a formal study of price bands and transaction costs, Table 45 presents farm household identification of the principal problems of commercialization during the survey period. Almost 30 percent of households identify a variety of high transaction costs as the principal problem. These include distance, poor roads, lack of transportation, and high intermediary commissions. The share of households identifying transaction costs increases with land ownership and land size, reflecting that many large farms are relatively isolated, but also that increasing land size usually indicates more surplus production. 81. Focusing on basic grains is somewhat deceptive, however. Considering all crops, almost 63 percent of farm households participated in agricultural markets. This percentage reaches 75 percent for medium-sized farms. A higher share of owners sell their crops, but more than half of renters do too. After com and beans, as seen in Table 46, fruit is next in importance, followed by sorghum, coffee, and vegetables. Of households that produce a specific crop (Table 47), 74 percent of coffee growers sell their crop, followed by 53 percent of rice growers, and 49 percent of fruit growers. Contrasted with basic grains described above, a significantly higher percentage of sellers of fruits, and to a lesser extent vegetables, are found among the nonpoor than the poor. Livestock production 82. Similar to agriculture, livestock production plays a variety of roles in rural economies in Latin America. Livestock is a source of on-farm income and consumption (including derivatives). It is also a mechanism for insurance and savings, given the relative absence of formal insurance and savings institutions in rural Nicaragua. Livestock production is the second major on-farm economic activity. More than 20 percent of all rural households, and a third of farm households, Annex 23, Page2l maintain cattle, the activity requiring the most investment and offering the highest returns (Table 48). Less than 5 percent of non-farrm households have cattle stocks. As expected, the share of households with cattle stocks increases with larger farm size, land being a key input into livestock production, particularly of the extensive kind that predominates in Nicaragua. Sixty-five percent of households in the largest category owned cattle, holding on average more than 17 heads of cattle at the time of the survey. A lower share of renters hold cattle, as shown in Table 22, even among comparable groups of landowner. Uncertainty over control of land for pasture most likely plays a role in the lower propensity of renter households to keep cattle. 83. More than a third of rural households, and almost half of farm households, kept pigs during the survey period. Pigs, which require less space and inputs, are also kept by just under 20 percent of nonfarm households. Again, the share of households with pigs increases with land size, thougn less so than do cattle, as more than 40 percent of the smallest landholders owned pigs. Although fewer renters than owners keep pigs. this cannot attributed to their condition as renters, as a similar share of households by land size keep pigs (Table 49).Finally, fowl is the most commonly kept animal, in 62 percent of all rural households, more than 70 percent of farm households. and just under 50 percent of nonfarm households. Little correlation is found with land size, but farm households have more than twice as many fowl as nonfarm households and significantly higher numbers than renters. 84. In general, rural households accumulated livestock over the survey period, if we compare stock at the beginning of the reference period to that reported at the time of the household interview. For cattle, this change is relatively constant across landholding categories, providing hopeful evidence that small farmers are staying afloat, particularly those who own land (Table 49). These are the farmers typically forced to sell first during times of economic distress. While ownership of pigs overall also increases by I I percent, by land size the changes show high variation among categories. Surprisingly, while overall fowl increased 13 percent, all farm households, including urban, experienced a decrease in fowl stocks. 85. Livestock production is primarily extensive and not modemized, as described in detail in the FAO survey. Similar results are found as well in Nitlapan-UCA (1995). The LSMS data, while sketchy, seem to confirm these conclusions. Only 37 percent of all farm households with livestock vaccinated their animals, although 64 percent of the largest landholders did22 Owners show a higher use of inputs than renters, but viewed by land size, these differences decrease. 86. Cattle ownership is clearly associated with a higher level of welfare, confirming another conclusion of the FAO survey, that the wealthiest farm households in Nicaragua are cattle ranches. This is evident in Figure 19, showing the share of fanm households owning cattle, as well as their stocks, by consumption decile. It is important to note, however, that a significant share of the poorest landowning households keep cattle, presumably to hedge risk and increase savings. Pigs and fowl are owned more or less equally by rich and poor alike, as seen in Figure 20. 87. In Table 50 we classify households by ownership of cattle resources and land use: non-farn households with no cattle; farm households with no cattle; households with 1-5 heads of cattle; and households with more than 5 heads of cattle. The households in the latter group are a category apart, but the small livestock producers have more in common with the farm households that do not own livestock. Households in this category have far more land, particularly annual and pasture land, than the other households. These households are much less likely to be renters as well. Large cattle owning families also have significantly higher levels of education than the 22 This figure may underestimate the use of inputs in specific types of livestock production, as input use was reported over all production. Annex 23, Page22 other categories. with the exception of the nonfarm households. The heads of households tend to be male and older than the heads of other households. A smaller share of these households participate in off-farm wage labor, and not surprisingly, a larger percentage considers themselves patrones. A higher percentage of these households plant corn, an important input to livestock production. More than half hire in labor, more than twice as frequently as the other farm households. These households have much more access to technical assistance and in organization, but have similar credit usage compared to the other categories. Finally, large cattle owners have significantly higher welfare, particularly in terms of per capita consumption and poverty levels. Interestingly, this category also has the highest share of urban households (20 percent). This suggests that these are a combination of wealthy urban cattle ranchers, with all the urban infrastructure benefits this entails, with wealthy but relatively isolated livestock producers who live on or near their land. For this reason the picture on dwelling characteristics and access to public goods is mixed, despite the high per capita consumption and low poverty levels. Slhocks 88. Problems of physical security, theft, and land insecurity generate lots of publicity, but these are not among the principal problems identified by Nicaraguan farmers. Instead, they cite the classic problems of drought (53 percent) and crop infestation (33 percent), as shown in Table 51. These responses are essentially invariant across farm size and between owners and renters.3 89. Household response to these shocks, however, is governed by access to land, as seen in Table 52. Although the largest group of households claim to do nothing (38 percent), this response is biased towards larger landowners, who can absorb these shocks better. Almost a quarter of households reduce household consumption, but such a drastic response is concentrated among smaller landowners and renters, who obviously have less capacity to absorb these shocks by other means. Another 1 I percent worked extra hours, and 8 percent either sold assets or spent savings. These latter practices particularly prevalent among the largest landowners (19 percent), who presumably have more to sell. 90. The shortage of liquidity among farm households is evident from the data in Table 53. More than 40 percent cite lack of resources for investment (21 percent lack of capital, 20 percent financing) as their principal problem. Again, this is not surprising given the virtual absence of formal and informal financing mechanisms. Other problems affecting producers are high variable costs of production (inputs and labor, together 25 percent), and lack of land (8 percent). Only I percent say that government regulation is a problem. V. Determinants of rural prosperity 91. To better understand the determinants of rural prosperity, we regressed a variety of "successful" outcomes on household characteristics, assets, and regional variables. We begin with two overall indicators of household welfare, per capita consumption and income, and follow with outcomes reflecting channels through which assets increase welfare. In particular, household income by source and household participation in off-farm activities have clear associations with prosperity in the descriptive statistics.24 23 Similar responses are found in Davis, Carletbo, and Sil, 1997 and Nitlapan-UCA, 1995. 24 We do not analyze other successful agricultural outcomes, such as the adoption of modem technology, due to unavailability of data, low incidence in some cases, and estimation problems in others. Annex 23, Page23 92. We use reduced form specifications to find the determinants of welfare. Considering consumption and welfare as alternative proxies for household welfare, we model them as a function of a variety of household assets and characteristics, as well as access to infrastructure. One of the earliest instances of this approach is Glewwe, 1991, which has been used time and again in poverty and welfare analysis. The general specification we estimate is W = f (, I, R) 93. Where W is the measure of welfare, Z is a series of household characteristics and assets, I i. access to infrastructure, and R is geographical location. The type of estimation procedure varies with the characteristics of the dependent variable. We try to avoid using explanatory variables endogenous to the survey period consumption decision, this is especially important for the income equations and economic activity outcome variables. For this reason we exclude labor activity choices as explanatory variables, choosing instead to view these as outcome variables. We also exclude livestock holdings, which are highly liquid in the face of shocks. Independent variables include the following: * Land assets, by type of land (prior to survey) * Jointly owning and renting land (prior to survey, dummy) * Female head of household (dummy) - Age, head of household - Number of male adults with different levels of education - Number of female adults with different levels of education * Log of family size * Age and gender of household members (number in each category) * Value of own (nonagricultural) business stock * Credit-constrained (dummy) * Had running water in 1993 (dummy) * Did not have toilet in 1993 (dummy) * Has electricity (dummy) * Community has paved road (dummy) * Regional dummies 94. A number of interactions between assets and characteristics were also tested; those that had a significant effect on outcomes are discussed accordingly. Per capita consumption 95. Per capita consumption is modeled using a simultaneous quantile regression in order to correct for heteroskedasticity in the cross sectional data. Three Least Absolute Deviation (LAD) regressions (at the 25 h, 50'h, and 75th income quantiles) were estimated simultaneously, producing an estimate, by bootstrapping, of the entire variance-covariance estimator.25 Comparison of the coefficients permits us to see changes in the impact of the explanatory variables at different points in the income distribution. The results can be found in Table 54. 25 See Deaton. 1997, for an introduction to the benefits, in the presence of heteroscedasticity, of the LAD estimator, and StataCorp. 1999, for a description of the simultaneous quantile estimator. Annex 23, Page24 96. Land plays a critical role in determining welfare among rural Nicaraguan households. Across all three equations, land assets26 are associated with increasing levels of welfare, though the type of land, and impact, vary across equations. Surprisingly, perennial land is important for households at the lower end of the consumption distribution and has no significant impact on wealthier households. Although perennial crops are grown by households across the consumption distribution, as we saw in Section IV, the cultivation of crops such as coffee and fruit does not account for differences in income among wealthy households. 97. Annual land, on the other hand, is significant for the middle and upper level regressions, further, the coefficients across equations are significantly different. The larger coefficient on the 75thl quantile most likely stems from the complementary role annual land plays in livestock production, a frequent activity among wealthier households. Pasture land is also significant across all three equations.2 98. Although the land coefficients tend to be small, we find that renting in conjunction with land ownership brings a significant increase in welfare, for all three quantile regressions, on the order of 20 percent of consumption. This suggests a positive role for rental markets in reducing poverty, but only for those households who already own land. We were unable to find any effect for renting in the absence of land ownership.28 99. Education is the second key determinant of household consumption levels. We model the impact of education by separating by gender and level of schooling achieved. We find differential impacts among male and female education, and coefficients reflect access and quantity; higher levels of education equate to higher levels of consumption. Increasing access to all levels of education, particularly for women, greatly increases household welfare. The coefficients are large; for example, an additional female obtaining a primary education increases per capita household consumption, in the median regression, 1 I percent; secondary education 22 percent; and higher education 27 percent. In wealthy households, only for secondary education and above is female education associated with higher levels of consumption. This suggests that basic levels (primary) of education for women especially are still lacking and, once provided, could have a big payoff among poorer rural households. This result squares with the descriptive statistics showing the very low level of schooling among the poorest households. 100. Once controlled for male and female education, household demographic composition becomes less important. Larger family size means lower levels of consumption, which is not surprising with a per capita measure. The presence of small children means lower consumption among poor households. Of all the adult categories, only the number of females aged 35-59 has a significant (positive) impact, and only among wealthier households. This result is consistent with the descriptive statistics, which show the important role older women play in off-farm family businesses. Female headed households, once controlled for female education levels, have significantly lower levels of consumption, ranging from 8 to 16 percent less than male headed households. 101. A third key determinant of household consumption level is the value of assets (measured through stock) in own business enterprises. This coincides with the results of the descriptive analysis. 26 Measured as land owned and rented prior to 1998. This does insert a selectivity issue, as other households not currently renting may have left the rental market because of either success or failure. Thus it is not immediately clear whether we are underestimating or overestimating the impact. Similar comments may be made about land assets as well. 27 Some of the land results, particularly for the 25"' quantile, are sensitive to specification of the education and rental variables, suggesting collinearity. 28 We were unable to control for selection bias in rental market participation, thus this result must be treated with caution. Annex 23. Page25 102. A fourth determinant of household welfare is access to financial resources. Across all three equations, the credit constraint is associated with significantly lower household welfare, on the range of 10 to 15 percent. This includes households both without access to credit as well as those whose access to credit was limited. From this perspective the increase in the volume of credit to the agricultural sector over the past five years could have helped reduce poverty; then again, bias towards wealthier households in terms of access and amounts of credit make this unlikely. Similarly, it suggests that targeting of additional financial resources towards the sector would be useful.29 Surprisingly, coefficients are not significantly different across equations, since we would expect wealthier households to be less bound by credit and liquidity constraints. 103. Finally, spatial considerations play a critical role in determining household welfare, for which we employ three types of proxies. First, higher levels of infrastructure (such as electricity and piped water) are associated with higher levels of welfare. Presumably this stems from their correlation with increased off-farm labor activities, as well as access to agricultural output and input markets. Second, access to paved roads, a measure of isolation, is most important for poorer households. Third. we included dummy variables for regional location. In all three equations, living in (rural) Managua, brings significantly higher levels of welfare. The reasoning is the san-me as above. Per capita income 104. Total per capita income is modeled using both robust and median regressions, again to minimize the bothersome presence of heteroskedasticity.30 In general the results are similar to the consumption equation (Table 55). Again, land is an important determinant of household welfare. As in the consumption equation, no quality differences emerge between land types, as coefficients are not significantly different. All levels of female education bring higher incomes, but only primary and higher education for men have a similar significant effect. Coefficients again reflect the quality of higher levels of education. Infrastructure, self-employment assets, and credit constraints show similar results. 105. Contrasting with the consumption equation, however, are the impact of household demographic characteristics. Adult males of any age, with exception of over 60, bring higher household incomes, but only females between the years of 35 and 59 have a similar effect. Several reasons could account for this difference. First, education is more necessary for women than men in income generation. Second, more of male income could be spent on items which are undercollected on the consumption side-namely savings or vices such as alcohol and tobacco-- thus reducing male importance in consumption equations. Third, males tend to participate more in activities with easily quantifiable income. 106. In equations 56 and 57, we focus on the determinants of income by source of economic activity. For agricultural and livestock income, we find, not surprisingly, that land assets are key determinants of agricultural income. Here, coefficients reflect differences in the quality of land, and perennial land brings significantly higher returns than other land types. Moreover, increasing levels of education increase the returns to annual land. Similarly, male primary education brings higher agricultural income. Typically such results are found in a context of a modernized agricultural production system, which is not the situation in Nicaragua. In any case, this result suggests that increased rural education levels increase returns not only to off-farm activities, but "' This variable could be considered endogenous; households could be excluded based on perceived welfare (and thus repayment potential) level, though we are controlling for the key signals of repayment potential and collateral. 30 We abandoned the simultaneous quantile equation after finding few differences among the quantiles. Annex 23, Page26 also to agricultural and livestock production as well. Further, under certain conditions (specifically, higher levels of household education), annual land may have higher returns than perennial land. 107. Not surprisingly, credit-constrained households had significantly lower levels of agricultural and livestock income. Infrastructure and distance variables, however, have opposite signs than expected. Access to infrastructure and paved roads lowers agricultural incomebecause households become increasingly involved in off-farm activities, at the expense of agricultural and livestock production. 108. In Table 57, we look at the determinants of nonagricultural self-employment, agricultural wage labor, and nonagricultural wage labor income Greater access to land lowers wage income, but perennial and pasture land increase nonagricultural self-employment income. Presumably, agricultural and livestock production are used as an input in off-farm activities. An alternative explanation would be that households controlling large amounts of land tend to have greater access to the capital necessary for profitable self-employment. 109. Differences are also found with education and demographic variables. Education plays no role in nonagricultural self-employment, and again the presence of older female adults is a key determinant of this source of income. Education also has no role in agricultural wage labor, and instead young males (15-34) and females (15-19) are associated with this source of income. Higher levels of education, on the other hand, are associated with nonagricultural wage labor income. Again the presence of adult males is key; returns for adult males are more than double those of agricultural wage labor. Females aged I -19 also have higher levels of nonagricultural wage income. 110. Finally, better infrastructure brings high levels of off-farm, nonagricultural income. These variables are more important than geographical location, except in the Atlantic region, which notably lacks wage opportunities. Regional location is more important for agricultural wage labor income; households in the Pacific and Managua regions have a higher level of this type of income, compared to the other two regions. Participation in off-farm activities I ll . Using a series of probits, we modeled the probability of a household having at least one member who participates in one of the three off-farm activities. The results, in Table 58, while confirming the sectoral income equations, show that nonagricultural wage labor and self- employment are the high quality off-farm activities, as we posited. Higher levels of education increase the probability that a household participates in both of these activities. Access to electricity also increases the probability of these activities, which is not surprising, since it implies more markets and job opportunities. Further, female headed households are more likely to participate in both of these activities. Few gender differences can be found in the participation in nonagricultural wage labor, but the dominant role of older women in nonagricultural self- employment is confirmed. Finally, although nonagricultural wage labor is associated with living in Managua, and to a lesser extent in the Pacific region, no geographic differentiation is found with self-employment. 112. Agricultural wage labor, on the other hand, is the lower quality off-farm source of employment. Both higher levels of education and older age of head of household lower the probability of a household participating in agricultural wage labor. Males of all ages are the chief Annex 23, Page27 agricultural wage laborers. This kind of labor is also associated with landlessness: higher levels of owned annual and pasture land reduce the probability of agricultural wage labor. Similarly, credit-constrained households, often so because they lack land as collateral, have a higher probability of participation. DECOMPOSITION OF THE CHANGE IN HOUSEHOLD CONSUMPTION, 1993-1998 113. Up to this point we have focused exclusively on the data provided in the 1998 LSMS. since it does not constitute a panel to the 1993 survey. The two surveys are not strictly comparable, particularly regarding rural households, because of differences in sampling methodology. Furthermore, the 1993 survey lacked an agricultural and livestock module. Adjustments have been made, however, to make the consumption aggregates comparable over the two periods (see Sobrado, 1999 for a description of this adjustment). Preliminary comparison of the two data sets has shown that the levels of extreme and moderate rural poverty have fallen significantly between 1993 and 1998. Overall, per capita household consumption in rural areas increased more than 20 percent between 1993 and 1998. 114. In this section, we attempt to determine econometrically what may have led to this decrease in rural poverty. Because we do not have panel data, we are unable to follow directly those households that left poverty over this period; instead we are limited to looking at changes in average characteristics. Given these limitations, we must refrain from implying causality in analyzing these changes. Instead, we use this analysis to formulate hypotheses about the determinants of the change in consumption and poverty rates over this period. 115. First, we look at changes in the mean value of key household characteristics over the period to ascertain how the group of poor households may have changed over the period. Next wve run poverty probits by year to examine how the role of different characteristics may have changed over the survey period. Finally, we decompose the change in consumption into what is attributable to changes in the returns to different household characteristics and assets, and what is due to changes in holdings and composition of these characteristics and assets. Because two key' groups of variables, agricultural assets and labor activities, are not comparable, we excluded them from the analysis. Chtanges in mean values 116. We begin by looking at the general changes in household characteristics of the two survey periods, overall and for urban and rural populations. As can be seen in Table 59, the most salient tendencies are demographic; families are older and more educated. The average number of uneducated adults drops in both urban and rural areas. Overall, however, educational gains are more evident among rural households, as the average number of household members with all levels of education increases. This is true for both men and women. Also, in calculating the poverty rate, average family size decreased. This change appears due primarily to fewer numbers of children and young adolescents. In productive assets, cattle stocks increased among rural households, and decreased for urban. 117. Table 60 classifies households by poverty status and year. Some of the results are as expected: poor households have lower levels of all types of assets, particularly education. however, extreme and moderately poor households appear to have benefited along with nonpoor households by increasing levels of education. On the other hand, the extreme poor have larger Annex 23. Page28 households than the moderate and nonpoor. Again, this difference appears to stem from larger numbers of small and old children, and young adolescents. Larger family sizes are also reflected in greater housing density (household members per room) for extremely poor households. Poverty probits 118. In Table 61, we present standard poverty probits for urban and rural households for 1993 and 1998. Comparing the results for 1993 and 1998, we can interpret differences in the coefficients of specific variables as the changing influence that a variable has on the probability of a family living in poverty (extreme and moderate combined). Again, most action is found among the education, demographic composition, and regional variables. In general, in 1998, both male and female education, particularly higher education, is less likely to move a family out of poverty than in 1993. However, primary education for men, and secondary education for women, appear to be more likely to reduce poverty. 119. Demographically, larger family size, and particularly the presence of large numbers of young children, increases the probability of living in poverty more in 1998 than in 1993. In terms of regions, the default is Managua. In 1993, living in the Atlantic, Central, or Pacific regions implied a much higher probability of being poor, on the order of 22, 28, and 12 percent, respectively. By 1998, however, the differences between the regions had declined. While living in the Atlantic and Central regions still meant a higher probability of living in poverty, the coefficients had decreased to 15 and 20 percent, and living in the Pacific region was now no different from living in Managua. Decomposition analysis 120. In this section we use decomposition techniques first used by Oaxaca (1973) to determine which household characteristics were associated with the increase in welfare that occurred from 1993 to 1998. Following a recent application by Davis, Handa, and Soto (1999), for urban and rural households overall and by region, in we regress total log per capita household consumption in both survey years on a set of household characteristics and assets divided into nine categories: log Con, = C, +/,, * Xi +i, where C and the vector B are the terms to be estimated, X is the vector of the eight categories, and pt is a random error term. The nine groups are: * Participation in agriculture and the number of heads of cattle * Age and gender of the head of household * Log of family size * Number of male adults with different levels of education (none, less than primary, primary, secondary, and higher) * Number of female adults with different levels of education * Demographic composition (number of females and males by age category) * Ownership of dwelling * Access to infrastructure (running water, toilet, and electricity) * Region Annex 23, Page29 121. Using the estimates, we decompose the change in consumption into the share corresponding to changes in the mean level of household characteristics (the vector X in each equation), and the share corresponding to differences in returns to these characteristics, or the Betas. More specifically, ACon = (C93 - C98 ) + X 98 * (93 b98 ) + 93 * (X 93 -A 98) where Con is (log of) per capita consumption, C93 and C98 are the constant terms in the regression for corresponding time period, X93 and X98 are the mean characteristics of households, and b9, and 698 are the coefficient vectors. Thus the change in consumption between 1993 and 1998 can be decomposed into three components. The first is the difference in the estimated constant term. The second is the difference in coefficients, which reflect the environments in the two periods, leading to different returns to household characteristics. The third is the difference in characteristics of households, or the change in level or composition of endowments.' 122. The results for the decomposition analysis. found in Tables 62 and 63. confirm the conclusions drawn from the comparison of the means and probits. Since per capita household consumption among rural households increased from 1993 to 1998, a negative change reported in the table means that the category helped increase consumption, and a positive change means the category decreased it. The first column is the change in Betas, or return to categories, and the second column is the change in levels of characteristics. As can be seen in Table 62, approximately 83 percent of the total change in consumption among rural households comprises changes in endowments. Foremost are smaller family sizes, which account for 64 percent of the change. Increases in both male and female education have smaller, though positive, shares in the change in consumption over the period. 12'3. Although only 17 percent of total change derives from changes in returns, this percentage masks considerable positive and negative movement among categories. Again, the change in returns to family size plays the dominant role. This is followed by the increasing returns to regions, which reflects the differential impact of economic growth between Managua and the rest of the country between 1993 and 1998. Surprisingly, the interior regions, particularly the Central, appear to have benefited most. Also surprisingly, the returns to education and agricultural and livestock activities, as well as demographic composition, worsened over the period. 124. Table 63 presents the decomposition results for the Central and Pacific regions. Some interesting divergences with the overall rural estimation emerge. In the rural Central region, while changes in family size and education play similar roles, both the returns to agriculture and the constant term are negative, reflecting growth in agriculture and generally better conditions in the Central region not attributable to any of our explanatory variables. Here, increased returns to demographic characteristics, principally adult males and females aged 20-59, play an important role in the increase in consumption over the period. Since real agricultural wages appear to have fallen in the Central region over this period, these increased returns are likely due to better employment opportunities for adults. 125. Similarly, in the rural Pacific region, smaller family sizes constitute the largest share of increased consumption levels. Again, however, changes in returns to demographic characteristics play an additional role. In this case, these increased returns are attributable to two specific cases, 3' One caveat to the analysis is that the behavior of the independent variables in the time between the two surveys is not completeli exogenous; for example. households may join or split for welfare reasons. Thus we must be cautious in terms of claiming causality for changes in consumption levels. Of the variables chosen, ownership of cattle is probably the least exogenous. Annex 23, Page3O males ages 20-34 and females 35-59. Agriculture saw a large fall in retums, as well as unattributable environmental factors reflected in the constant term. These results can be contrasted with those of the urban Pacific, where poverty increased and consumption decreased from 1993-1998. A large share of this decrease derives from a drop in the returns to both male and female labor, in combination with a large constant term. These large decreases overshadowed increased returns to education and smaller family sizes. THE IMPACT OF TECHNICAL ASSISTANCE PROGRAMS Brief description of major technical assistance programs 126. Technical assistance has been the principal policy instrument of the Nicaraguan government in support of small and medium sized producers during the last few years. In 1993, extension services were restructured to combine agricultural and livestock research under the Instituto Nicaraguense de Tecnologia Agropecuaria (INTA). Driven in part by budgetary difficulties, soon after INTA was formed its technology assistance programs were redesigned into a demand-driven extension system. Currently, three classes of service are provided: free mass media and demonstration (ATPm), publicly co-financed (Asistencia Tecnica Publica Cofinanciada - ATPI), and private extension (Asistencia Tecnica Privada - ATP2). INTA's staff provides the first two, and private extension firms provide the third (Dinar and Keynan, 1998). These programs differ in their objectives, as is reflected in each program's process of selecting communities and producers, type of technology transferred, and extension method. 127. The ATPm program is targeted to small producers with scarce resources who reside in marginal areas in each of the five INTA zones, with the objective of supporting basic grain production for home consumption. The principal methods of extension used include establishment of reference farms to demonstrate new technologies during farm days, monthly group meetings for training of group leaders, and radio programs. No payment is required of the farmers, but the selection of farmers depends on their participation in community producer organizations and on the community leaders. 128. ATP I and ATP2 were designed to improve the effectiveness of extension services by making extensionists accountable for results. As a result, recipients of ATP I and ATP2 are required to pay for service provided by the technicians. Both programs target small and medium producers who farm in favorable agro-ecological zones. ATPI seeks to improve production of basic grains, selected other crops, and small livestock among producers who have at least some resources. ATP2 covers a wide range of farm and livestock production and marketing aspects and is targeted to producers who have sufficient resources to engage in agriculture for commercial purposes. 129. The two programs also use different extension methods. Producers who demand ATPI services, and are willing to pay a percentage of the costs, organize into groups of a dozen. These groups can be previously formed, formed by the technician, or formed by the users themselves expressly to take advantage of ATPI. An INTA technician visits the producer group every 15 days, or according to the demands of the clients, demonstrates new technologies, and holds training workshops. ATP2 participants are individually trained. The firms that provide ATP2's Annex 23, Page31 services recruit producers selectively. They look for professional skills indicating that their recommendations will be implemented and their services paid for.3- 130. Extension is also provided by other organizations such as UNICAFE, the coffee growers association; UNAG, an association of small farmers; and numerous nongovernmental organizations. These programs, however, are much smaller than the government's programs (Keynan, Olin, and Dinar, 1997). 131. Despite the massive reorganization of the national agricultural technology assistance programs, overall use of these programs remains quite low (Table 24). Only 24 percent of the farm households had access to technical assistance in their communities and a mere 15 percent actually used technical assistance. Government and NGOs were the main providers 33Fifty-four percent of the recipients received technical assistance from the government, and 41 percent received assistance from NGOs. Almost 10 percent of the recipients had more than one source o-f technical assistance. Thirteen percent of the recipients reported paying for the services, but we cannot assume that services were received from ATP I or ATP2 since some of the private organizations also impose a fee for service. 132. Table 64 reports the percentage distribution of technical recipients and nonrecipients classified by degree of poverty, farm household typology, and consumption quartiles. When viewed by farm household typology, 39 percent of technical assistance recipients are small and medium sized landowners even though these households constitute only 31 percent of the sampl e population. Together, these percentages show some evidence of targeting towards small and medium producers.34 Minifundia renters are most under-represented among technical assistance recipients. While these households are 28 percent of the farm household sample, only 17 percent of the recipients come from this group. By contrast, minifundia owners are not under-represented suggesting that land ownership rather than access to land explains why renters are less likely to participate in technical assistance programs. 133. When viewed by welfare indicators, such as the degree of poverty or consumption quartiles, it is clear that the extremely poor households are under-represented among technical assistance recipients. For example, while 20 percent of farm households are categorized as 'extremely poor,' only 13 percent of recipients are from this group. This result is not surprising since 38 percent of extremely poor households are minifundia renters. 134. As seen in Table 28, most technical assistance is geared towards agricultural production. Use of fertilizer and chemicals accounts for 28 percent, and crop diversification, seed selection, and soil conservation each account for 12-15 percent of the technical assistance received. In the remainder of this section, we assess the impact of technical assistance on the use of fertilizer and pesticides.35 12 This description was provided by INTA. "The LSMS data do not allow us to distinguish whether govemment services were received from ATPm. ATPI. or ATP2. 3The numbers cannot be taken to be an accurate representation of INTA's targeting since the distributions of technical assistance recipients include households that received extension from any source - govemment. NGO, or other private sources. Because of the relatively small number of technical assistance recipients in the sample, we did not split the sample into govemment provided and other sources of technical assistance for the purpose of analysis in this section. 15 We do not have data to evaluate the impact on crop diversification (since land area planted to different crops is not available), secd selection, or soil conservation. Nor do we have information on crop yields. Ultimately, all technology transfer methods should translate into enhanced farm profits, whether through reduction in costs, improved crop management, or improved yields and expanded area. Therefore, it is better to use farm profits - inclusive of both costs and benefits - as an indicator of farm-level success when evaluating the impact of a technical assistance program. If data become available, we will perform the analysis using farm profits. Annex 23, Page32 Impact of technical assistance on fertilizer and pesticide use 135. The impact of technical assistance on any outcome, say farm profits, is the difference between farm profits with the program and without it. The "with'" data are observed in a household survey that records outcomes for technical assistance recipients. But the "without" data are fundamentally unobserved, since a household cannot both participate and not participate in the same program. To determine the effects of technical assistance we need to identify what the outcomes at the household level would have been if they had not participated. 136. A control group of households chosen from the nonparticipants in the survey should resemble the participant group, with the exception that the control group did not receive technical assistance. The outcomes of nonparticipants may differ systematically from what the outcomes of participants would have been without the program, producing selection bias in estimated impacts. There are three sources of bias that we need to be concerned about. 137. First, the criteria for choosing communities that receive INTA's technical assistance, such as degree of marginal lands for ATPm, or favorable agro-ecological zones for ATP I and ATP2, are likely to be related to household outcomes of interest. Therefore, a comparison of communities with and without the technical assistance program would yield a biased measure of impact (program placement bias). For example, since marginal areas are targeted for ATPI, comparison of farm profits of ATP I households to control group households in other communities would downwardly bias the estimated impact. 138. Second, households that receive technical assistance are likely to be different from the ones that do not. For example, ATP2 expressly targets households that can better apply the recommendations and can qualify for cofinancing. Comparing profits of these households to others is likely to provide an upward bias of the programs' impact since the pre-intervention profits of ATP2 households are also likely to have been higher than those of nonrecipients (self- selection bias). 139. A third source of bias arises when the control group differs from the-participants in their distribution and supports of explanatory variables. The criteria used for selecting participants can be expected to entail considerable rationing of participation according to observable characteristics, making this bias likely. In their comparison of non-experimental methods of evaluating a training program with a benchmark experimental design, Heckman et. al. (1997, 1998) find that failure to compare participants and controls at common values of matching variables is the single most important source of bias - more important than the problem of selection bias due to differences in unobservables. 140. Drawing upon the literature on impact evaluation, several methods correct for selection biases, depending on what data are available. Econometrically, program placement bias can be eliminated by using community-level fixed effects to control for community unobservables that influence the placement of programs and are also correlated with household outcomes. The standard approach to control for household-level unobservables that lead to the self-selection bias is instrumental variables: an instrumental variable is a factor that affects a household's probability of participation in a technical assistance program but does not affect its outcome. Unfortunately, it is difficult to find any household-specific characteristics to serve as valid instruments. For example, the criteria used to select farmers for the governmental programs - participation in a producer organization, cultivation of favorable or marginal lands, etc. - are also likely to influence farm profits. Typically the availability of technical assistance in the community is used as an instrumental variable, as long as the sample includes households from communities with the Annex 23, Page3D program as well from communities without it (Moffitt, 1991). However, since government technical assistance programs identify farm-level impact by comparing households in non- program communities with households in program communities will yield biased estimates of program impact. Therefore, LSMS data do not yield instruments that allow us to control for selection on unobservables. and the estimates reported below may be biased because the distributions of unobserved characteristics of recipients and non-recipients are different. 141. The third source of bias, differences in distribution and supports of observed variables, can be reduced by using matching estimators. In matching methods, the counterfactual group is constructed by matching program participants to nonparticipants on the basis of similarities in observed characteristics (Jalan and Ravallion. 1998; Heckman et al., 1997, 1998). The predicted probability of participation - the "propensity score" - is calculated for each participant and non- participant in the sample using a logit regression. Using the propensity score, each participant is matched to the five closest nonparticipants in the sample36 Average characteristics of these five matched nonparticipants' provide an estimate of the pre-intervention outcome for each participant. 142. Table 64 reports the results of the logit regression used to estimate the propensity score on the basis of which matching is subsequently done. Variables are included in the model on the basis of two criteria: (a) minimization of classification error and (b) statistical significance of the included regressors (Heckman, et. al., 1998). The participation regression suggests that households that receive technical assistance have more adults and older household heads. Recipients have more farm assets (particularly work animals, fumigation pumps, irrigation equipment, and water pumps) and are less likely to be credit-constrained. Number of cattle is also significant but was not retained in the final specification because it is highly correlated with the number of work animals.-Surprisingly, land ownership is not a significant determinant of the probability of participation, once ownership of other farm assets is controlled for. The model also predicts that controlling for farm assets, indicators of standards of living such as the number of rooms or availability of a toilet in the house have a positive significant impact on the probability of participation. 143. Table 65 reports our estimates of the impact of technical assistance on the probability of using fertilizer and pesticides.37 The share of technical assistance recipients who use fertilizer and pesticides in the overall sample is 48 percent and 62 percent respectively. By contrast, only 30 percent and 41 percent of the matched non-recipients use these chemical inputs, suggesting that technical assistance significantly increases the probability of using these inputs. When viewed by the degree of poverty, the increase in probability is greatest for the non-poor, and significant for the moderately poor households. However, the technical assistance does not result in an increase in fertilizer use among the extremely poor. This suggests that other factors limit the usefulness of applying new technologies by the poorest households. 144. Table 66 reports the impact on the probability of using fertilizer and pesticides, by the farm household typology. The estimates are consistent with those reported in the previous table. While technical assistance significantly increases the probability of chemical input use for most groups of households, it does not significantly increase use by the minifundia renters who are al;o among the poorest households in the sample. Neither do large farmers alter their fertilizer use 36 To ensure that matching is done only over common values of the propensity score, nonparticipants for whom the estimated density of participation is zero are excluded from the sample. We also exclude 2 percent of the sample from the top and bottom of the non- participant distribution (see Jalan and Ravallion (1998) for a good application of matching methods). 3 We also plan to estimate the impact on fertilizer and pesticide expenditures. Annex 23, Page34 significantly. The estimates indicate that the impact of technical assistance is greatest among the small and medium producers, in accordance with the design and objectives of INTA's programs. CONCLUSIONS AND POLICY IMPLICATIONS Rural poverty should be confronted from a rural rather than a solely agricultural perspective. 146. Agricultural policy may or may not reduce poverty, since the poorest Nicaraguans may be wage workers, staple crop producers, or both. 147. The rural sector in Nicaragua is highly heterogeneous in its assets, spatial attributes, and economic strategies. Sectoral issues must be dealt with from the perspective of rural development, as the majority of households pursue both on- and off-farm, agricultural and nonagricultural strategies. 148. This heterogeneity suggests a variety of solutions, a perspective adopted by the government. Sectoral initiatives must have clearly defined objectives-for example, alleviating poverty or increasing supply response-and policymakers must take into account how objectives complement or work against each other. For example, technical assistance programs may increase the supply response of small and medium producers but may not help the poorest households at all. 149. Similarly, incentives for exportable production may create agricultural wage employment but reduce poverty only slightly, since the benefits will mostly accrue to growers instead of agricultural wage laborers, the poorest of the poor in rural Nicaragua. Overall, off-farm activities bring in more income than on-farm activities. 150. We find that in general rural farm households compare poorly to rural nonfarm households. Those who work in nonagricultural wage emplovment and family businesses tend to have higher adult educational levels and be nearer to towns and infrastructure. Increasing women's access to education may greatly reduce poverty. 151. Econometric analysis shows that providing all levels of education to women has a significant and positive impact on per capita consumption and income. These results are particularly strong for poorer households. Not all off-farm activities have a positive impact on welfare; there is evidence of differential returns to wage labor. 152. A large distinction can be found between better remunerated activities (nonagricultural wage labor and self-employment), which require education and access to infrastructure, and worse remunerated activities (agricultural wage labor), which require little education. The latter do not seem to lead out of poverty, but rather are part of a subsistence survival strategy, particularly when used in combination with farm activities. Therefore, a strategy to increase agricultural exports, which relies on increased demand for agricultural wage labor, may not reduce poverty in the long term. Annex 23, Page35 Nevertheless, on-farm activities play a fundamental role in assuring prosperitv-or survival-for a large segment of the rural population. 153. Agriculture is a key economic activity for most rural households but can play a variety of roles. For medium sized producers it is the principal source of livelihood in terms of both cash liquidity and consumption. For small producers it serves as an insurance policy, a source of consumption. and a complement to off-farm activities. Though returns may be higher off farm, farming provides food security and insurance, and small landowners may be better able to absorb shocks than landless wage laborers. For rural families with lower levels of education, agriculturc is also an important source of wage labor demand. 154. Livestock production is the secondary farm economic activity for most rural households. For small farmers, livestock holdings can serve as insurance or investment. For large farmers, livestock production is a principal source of cash income. Access to land is the key determinant of on-farm economic success. 155. Econometric results suggest that access to land is an important determinant of household welfare, as measured by consumption. This is particularly true for poorer households. Increased access to land-or alternatively, increased returns to land-could thus be one source of the drop in extreme and moderate poverty from 1993 to 1998. 156. The results on rental land, the most prevalent form of access to land for poor farmers, are ambiguous. While the returns to renting and owning in combination appear to be high, renting alone does not increase welfare levels. We were unable to control for selection bias in rental market participation, however, and thus we are limited in the conclusions we can draw regarding this strategy. Education increases returns to both on- and off-farm activities. 157. Higher levels of male and female education increase total per capita consumption and income, as well as off-farm income and participation. In addition, male primary education brings higher agricultural income. Despite government efforts, access to the services provided by agrarian institutions, an important factor in microeconomic success, is still minimal. 158. The survey results show that, despite high sectoral growth and government efforts, microeconomic problems hinder productivity, particularly for small and medium farmers. Agriculture remains primarily extensive rather than intensive. This is due to low levels of modem input use, credit constraints, segmented markets, and the scarcity of agrarian institutions. 159. While technical assistance and use of credit appear to have increased somewhat since 1996, large gaps in access to these services still exist. These services may benefit farms, but coverage remains minimal. Participation in producer organizations is a common mode of accessing those services that do exist. 160. Despite the government's emphasis on commercialization, staple crop markets remain highly segmented. We see no improvement since 1996 in market participation in beans, and the share of households participating in corn markets has fallen. Farm households cite significant transaction costs as a major impediment. Annex 23, Page36 Technical assistance programs lead farmers to begin using fertilizers and pesticides 161. Data limitations prevented us from determining whether technical assistance programs have a direct impact on household welfare. We were also unable to separate effects by government and nongovernmental programs. We were able to show, however, that s technical assistance programs lead farmers to adopt modem technologies that in turn increase supply response and agricultural income. Those who lack access to sufficient land or education in rural Nicaragua are destitute 162. The worst-off households in rural Nicaragua are minifundistas and agricultural wage workers; that is, households poor in both education and land. It is unclear to what extent these households have benefited from agricultural sector growth, but it is evident that neither agricultural sector growth alone, nor access to agrarian institutions, will bring them out of extreme poverty in the long term. In addition to education and land poor people need economic incentives, markets, and institutional support. Increases in per capita consumption from 1993 to 1998 stem principally from changes in family size and regional variations. 163. Decomposing the change in rural per capita household consumption levels from 1993 to 1998, we find that the decrease in family sizes played the most important role in increasing per capita consumption. While increases in the levels of adult education also helped increase consumption over all rural households, the returns to both male and female education actually declined over the period. Finally, higher returns to living outside Managua, particularly to the rural Central region, accounted for an important share of increased consumption. This result could derive from the leading role of agriculture in driving economic growth over the period, as well as increased employment opportunities outside of agriculture. Annex 23, Page37 Table A23.3 - Share of Household in Regions by Rural Household Typology land ownership and rental education urban overall farm non farm adjusted mzs years units total e-2 2-5 5-20 ->20 4 >4 # of observations 2093 438 229 229 164 408 435 189 1063 8:39 Regions Atlantic. urban share .02 .00 .00 .00 .00 .00 .00 .21 .00 .10 Atlantic. rural share .10 .07 .14 .16 .39 .06 .04 .00 .15 .1)5 Central. urban share .04 .00 .00 .00 .00 .00 .00 .40 .00 .o0 Central, rural share .42 .54 .46 .53 .41 .45 .38 .00 .50 .42 Pacific. urban share .02 .00 .00 .00 .00 .00 .00 .27 .00 .() Pacific, rural share .30 .31 .37 .22 .16 .41 .38 .00 .28 .A0 Managua share I 10 .08 .03 .10 .03 .07 .20 12 .07 . 4 Atlantic: RAAN, RAAS. Rio San Juan Central: Nueva Segovia. Madriz. Esteli. Jinotega- Matagalpa. Boaco. Chontales Pacific: Chinandega. Leon. Masaya. Carazo. Granada. Rivas Table A23.4 - Share of Household in Regions, by Farm land renter overal owne rente adjusted unit total e-2 2-5 5- e-2 >2 # of 125 180 186 203 17 346 16 746 538 Region Atlantic, shar .0 .02 .00 .03 .1( .02 .0 .04 .2 Atlantic. shar .1 .06 .10 .16 .3 .05 .1 .16 .3 Central, shar .0( .08 .07 .07 .0, .06 .0 .07 .3 Central, shar .4 .45 .39 .47 .3( .47 .4( .41 .45 Pacific, shar .OZ .02 .06 .02 .0 .06 .0 .03 .0 Pacific, shar .2 .27 .32 .19 .1: .26 .2( .23 .26 Managu shar .0 .09 .05 .07 .04 .08 .0 .07 .09 Annex 23, Page38 Table A23.5 - Household economic activities, by rural household typology farm non farm rural overall urban non test land ownership and rental education farm farm farm farm vs. adjusted m:s years nonfarm units total e-2 2-5 5-20 >20 <4 >4 # of observations 2093 438 229 229 164 408 435 189 1060 843 Labor activities hh has at least one member share 75 71 .57 56 47 89 97 75 .61 .93 *** nonagwageworker share 39 .32 20 .21 15 40 68 49 .24 .55 *** agwageworker share 40 47 .35 .31 32 55 38 19 .39 .46 *** selfemployed share 23 15 .21 .20 12 .24 35 32 17 30 * businessowner share 10 .08 .11 19 20 .01 03 28 .13 02 * withoutpay share .33 42 .54 .56 52 .11 .13 35 49 12 2 ** Migration hh member migrated share 11 .15 .09 10 03 1 3 10 12 .11 .12 ** tocity share 04 .05 .01 03 01 .06 .04 .07 .03 05 * toothercountry share 02 .03 .01 01 00 02 02 02 02 02 Receives transfers share .17 .15 .18 13 11 23 .16 14 15 20 * Agricultural activities planted crops share .57 .95 98 98 93 00 .00 .90 96 00 corn share 43 .68 .81 .80 .76 00 00 .59 .74 00 beans share .31 .50 .57 .59 44 00 00 45 53 00 veggies share 08 .11 18 .15 05 .00 00 .12 13 00 fruits share .38 .59 .65 .67 .63 .00 00 .57 63 .00 coffee share 05 .06 .10 .16 .09 .00 .01 06 10 00 planted backyard share 17 .00 .00 .00 00 .39 .47 00 00 43 fruits share .17 .00 .00 .00 00 39 47 .00 00 43 Livestock activities cattle share .21 .15 .32 45 65 05 .03 .32 .33 .04 * pigs share .36 .42 .46 .55 .65 .22 .16 .33 .49 19 **9 fowl share .62 .70 .81 .80 .81 .51 .45 .47 .76 48 * ***significant at I percent **significant at 5 percent *significant at 10 percent Annex 23, Page39 Table A23.6 - Household income, by rural household typology farm non farm rural overall land ownership and rental education urban farm non farm farmrr adjusted mzs years annual, per capita total e-2 2-5 5-20 >20 <4 >4 # of observations 2093 438 229 229 164 408 435 189 1060 843 Total income 4994 3075 3684 3851 6243 4489 6359 9288 3865 5450 .4s share of total income Agriculture and cattle 22 26 41 50 62 6 2 19 43 4 Off farm, labor activities 58 50 34 28 18 69 80 67 34 76 Wage labor. agricultural 7 10 4 6 1 14 9 3 6 11 Wage labor, non agricultural 32 32 18 12 9 41 46 30 19 44 Selfemploved 18 8 11 10 8 13 25 34 9 20 Offfarm, other activities 21 24 25 22 2] 26 17 14 23 21 Imputed rent 10 13 10 12 9 9 1 1 8 11 10 Gifts 2 3 3 3 2 4 1 1 3 3 Remittances 5 5 7 4 7 7 3 3 6 4 Capital 0 0 2 1 1 0 0 1 1 0 Pension 0 1 1 0 0 4 0 0 0 1 Other 3 2 3 3 1 2 2 1 2 2 Annex 23, Page4O Table A23.7 - Household income, by farm household typology, land owners renters overall owner renter adjusted mzs annual, per capita total e-2 2-5 5-20 >20 e-2 >2 # of observations 1254 180 186 203 177 346 162 746 508 Total income 4688 3677 3963 4769 10277 3156 3705 5612 3331 .4s share of total income Agriculture and cattle 36 27 39 46 42 22 30 40 25 Off farm, labor activities 44 48 37 32 44 56 46 41 52 Wage labor, agricultural 5 8 5 3 1 11 5 3 9 Wage labor, non agricultural 22 24 21 16 17 33 28 19 31 Self employed 17 15 12 12 25 12 13 18 12 Off farm, other activities 20 25 24 23 14 22 25 19 23 Imputed rent 10 13 10 13 7 12 11 10 12 Gifts 2 2 1 2 1 3 6 1 4 Remittances 5 7 6 4 4 4 6 5 5 Capital 1 0 2 0 1 0 0 1 0 Pension 0 1 1 0 0 0 0 1 0 Other 2 2 3 3 1 2 2 2 2 Annex 23, Page4l Table A23.8 - Demographic characteristics, by rural household typology land ownership and rental education urban rural overall I , farm non firm adjusted mrs years farm onits total e-2 2-5 5-20 >20 <4 >4 # ofobservations 2093 438 229 229 164 408 435 189 1060 843 Head of household temale head of hh share IS :12 13 12 0,7 .29 .26 .08 12 .7 T age, head ofhh years 45.31 43.43 48.35 46.56 49.13 47.35 39.82 49.21 46.05 43.48 ** education, head ofhh years 2.75 202 1.54 2.16 1.97 1.29 5.74 376 1.94 3.56 ** Adult education levels education, adults years 2.58 2.29 2.79 2.56 1760 6.59 4.99 2.56 4.16 t # adults no education f 1.08 1.21 1.47 1.33 1.59 1.37 .23 89 1.35 .79 **. # adults less primarv education # 1.01 1.09 1.27 1.31 1.15 .91 75 .87 1.19 .83 '+' ft adults primary education # .53 45 .38 .44 .40 .21 .99 .75 42 .61 . ft adults secundary education .23 .09 .10 20 .16 02 59 .50 .13 .31 ** f adults high education # .09 .03 .02 .04 .08 00 21 .32 04 .11 Famils' composition tamily size # *76 5 73 642 6 33 6 59 23 5.17 6.15 6.14 5720 .7 # children e-4 vears f 88 97 .98 94 .91 90 .71 .78 95 .80 ** # children 5-10 # 1.06 1.10 1.11 1.09 1.26 .96 .97 1 09 1 12 97 ** f males 11-14 .33 30 45 .39 43 .32 .25 .34 .37 .28 + # females 11-14 .31 .25 .33 .36 .33 .29 .30 .38 30 .29 # males 15-19 # .36 .36 47 A41 .57 .30 .23 .38 42 .26 * # females 15-19 f .35 .36 .40 .38 .42 .29 .33 .34 .38 .31 1 # males 20-34 f .60 67 .65 .68 .53 43 69 .53 .65 .57 # females 20-34 # .59 .56 .50 .62 .55 .50 .70 .70 .56 .60 f males 35-59 f .51 .48 .54 .55 .67 .44 .42 .68 .54 .43 *' # females 35-59 f .47 .41 .57 48 .54 41 .46 .54 .48 .44 #males60+ # .17 .15 .26 .26 .22 .18 .04 .21 .21 .11 I #females60+ # .15 .13 .15 .17 16 .21 .07 18 .15 .14 Marriage status tree union share 37 .45 38 .26 .35 .36 .35 39 35 married share .40 .36 .42 .44 .62 .30 .38 .52 43 .34 * Annex 23, Page42 Table A23.9 - Adult education levels, by poverty classification Poor Non extreme moderate poor units total Adult education levels # men no education .56 .85 .61 .39 # men less primary education ft .54 .66 .58 .44 ft men primary education .27 .16 .26 .34 ft men secundary education ft .10 .03 .08 .15 ft men high education ft .05 .00 .02 .10 # women no education ft .52 .78 .59 .32 ft women less primary education ft .48 .59 .51 .40 ft women primary education ft .26 .19 .28 .28 ft women secundary education ft .13 .03 .09 .22 ft women hig,h education ft .04 .01 .02 .09 I able A23.10 - Household welfare, by rural household typology land ownership and rental education urban rural overall IL.s adjusted mr yvears farm non farmni , f0arm units total e-2 2-5 5-20) >20) <4 >4 of observations 2093 438 229 229 164 408 435 189 1060 843 Ownership of durables refrigerator share .09 .02 .03 .05 .08 .04 .18 .23 .04 .12 B/W TV share .25 .23 .18 .20 .10 .25 .35 .33 .19 .31 Color TV share .11 .03 .01 .04 .07 .08 .25 .26 .03 .17 Fan share .14 .06 .06 .04 08 .12 .29 3 1 .06 .21 * Dwelling characteristics has dirt tloor share .66 .76 81 .68 .64 .72 .50 39 .74 .61 has running water share .33 .19 19 .22 .08 .30 .62 6 1 As8 47 has no toilet share .28 .35 .38 34 .37 .33 .12 09 .36 .22 has electricity share 43 .31 .22 .24 .15 .46 .71 73 .25 .59 Welfare indicators hhconsumption, per capita - 13 36 432 472 53-5861 -66 431 52 extreme poverty share .21 .32 .24 .23 .20 .25 .05 .17 .26 .15 ** moderate poverty share .37 40 .46 .30 .32 .42 .33 .27 .38 .3 8 Annex 23, Page43 I able A23.1I - Household characteristics by land--labor strategy, rural households no wage non agricultural agricultural both wage wage wage (1) (2) (3) (4) (5) (6) C) ( 8) (9) 717)) min total <3 >3 no land land no land land no land land no land 8 ofobservations 1902 226 275 146 151 304 302 235 109 R;4 Agriculture and livestock Iotal land mz 1.68 30.39 00 1434 00 1 -17T(. owns land share .33 .50 92 .00 .51 00 .45 .00 42 110 rents land share .27 .54 .18 .00 .55 .00 61 .00 .62 .00 Has cattle share .20 .25 .58 .04 .26 02 .25 .05 .17 (7 Heads of cattle # 2.44 1.26 10.89 .10 2.56 09 1 66 16 3.17 .27 Demographic Female headed ho usehold share .14 .12 09 .44 17 30 I() 1 .16 2' Age, head of household years 44.92 44.54 47.70 50.58 47.12 40.80 45.82 41.52 44.21 45.06 Education, head ofhousehold years 2.65 1.89 1.97 2.93 2.90 5.17 1.36 2.45 2.25 2.70 Education, adults years 3.27 2.23 2.35 3.41 4 43 5.53 1 74 2.82 3.47 4,13 Selfemployed share .23 .17 .16 .59 28 .32 .10 .14 .21 .,1 Receives transfers share 17 .14 .12 .38 .17 .19 13 12 .23 .15 Welfare has dirt floor share 68 .83 .68 .58 .63 51 7 69 80 electricity share 40 .21 .13 .58 .57 71 16 47 .44 .5 PC household consumption C 4869 4075 5113 6340 4939 6455 3859 4531 3543 4435 Extreme poverty share .21 .24 .21 .13 17 .06 .39 .25 .25 .19 Moderate poverty share .38 .44 .38 .27 .28 .33 .35 .47 .45 .42 Regions _T 393 Atlantic share .13 .28 09 07 .05 1 03 3 0 Center share 46 .53 .48 .44 .32 .33 .59 57 48 .:4 Pacific share .33 .28 .22 .39 .41 .38 .24 33 .37 53 Managua share 10 .05 02 .09 .20 .24 .03 07 12 1 3 Table A23.12 - Significance tests on selected variables, by participation in labor activities and land use Education, Education, per capita extreme moderate columns head of hh adults consumption povertv poverty land: non agricultural wage vs. 5 vs. 7 agricultural wage no land: non agricultural wage vs. 6 vs. 8 *** ** * agricultural wage non agricultural wage: land vs. 5 vs. 6 * no land agricultural wage: land vs. 7 vs. 8 ** ** *** ** no land small land, no wage vs. 2 vs. 8 no land, agricultural wage small land, no wage vs. 2 vs. 7 ** *** ** land, agricultural wage Annex 23. Page44 Table A23.13 - Land ownership and transactions, by farm household typology _ land owners renters overall adjusted m:s owners renters units total e-2 2-5 5-20 >20 e-2 >2 # of observations 1254 180 186 203 177 346 162 746 508 Owned land total land, adjusted mzs 14.73 1.17 3.28 10.43 87.69 .00 .00 24.77 .00 total land mzs 20.93 1.53 4.46 14.43 125.35 .00 .00 35.18 .00 permanent share .08 .07 .06 .05 .08 na na .08 na annual share .66 .92 .85 .79 .63 na na .66 na pasture share .26 .01 .07 .16 .28 na na .26 na owns land share .59 1.00 1.00 1.00 1.00 .00 .00 1.00 .00 Total land mzs 16.91 1.39 3.71 10.63 88.13 1.06 13.10 25.08 4.90 Rent in rents in land share .47 .11 .16 .10 .05 1.00 1.00 .11 1.00 total land rented in. adjusted mzs 2.18 .22 .43 .20 .44 1.06 13.10 .32 4.90 share oftotal land rented in share .13 .16 .12 .02 .00 1.00 1.00 Rent out rents out lant share .06 .04 .10 .15 .14 .00 .00 .11 .00 total land rented out, adjusted mzs .22 .03 .20 .37 .91 .00 .00 .37 .00 share of owned land rented out share .02 .02 .06 .04 .01 na na .02 Purchases and sales, last S years sold land share .05 .05 .04 .09 .06 .03 .03 .06 .03 bought land share .11 .12 .14 .19 .23 .02 .05 .17 .03 Plot diversification index .04 .04 .08 .06 .04 .01 .03 .06 .01 Rural households share .85 .86 .85 .87 .78 .85 .88 .84 .86 Annex 23, Page45 Table A23.14 - Asset Gini coefficients, by region region overal Atlanti Pacifi Centra # of 171 202 631 88 Over farm Owned .8( .73 .88 .86 perenni .9S .99 .98 .99 annua .8v .76 .89 .85 pastur .9& .96 .97 .98 Rented .7 .93 .87 .84 Owned and rented .8C .69 .80 .77 Cattl .9C .85 .88 .90 Consumpti .4t .36 .37 .38 Educatio .4E .54 .39 .49 Over non farm Consumpti .5C .36 .36 .37 Educatio .3 .54 .39 .48 Table A23.15 - Land documentation, by farm household typology land owners adjusted manzanas units total e-2 2-5 5-20 >20 # of observations 746 180 186 203 177 Land documentation Title share .52 .46 .48 .55 .61 Agrarian Refonn share .13 .10 .16 .15 .14 Other share .14 .12 .12 .11 .14 No title share .22 .32 .26 .20 .12 Formnally registered share .64 .51 .62 .66 .72 Annex 23, Page46 Table A23.16 - Type of rental arrangement, by rent parcel size parcel size adjusted mzs units total e I # of observations 647 440 207 Type of arrangement rental % 39 39 39 borrowed % 47 46 47 taken in possesion % 3 5 3 sharecropped % 9 9 10 other % I I Table A23.17 - Source of rental lands, by rent parcel size parcel size adjusted mzs units total e 1 # of observations 648 439 209 Source of rental lands state lands % 2 2 1 comunal land % 1 1 I cooperative % 9 6 11 family % 22 20 22 individual % 65 67 63 other % 1 3 1 Annex 23, Page47 Table A23.18 - Source of rental lands, by contract arrangement contract arrangement share units total rented borrowed possesion cropping other # of observations 646 253 305 22 59 7 Source of rental lands cooperative % 9 9 7 -- 13 family % 22 7 34 -- 24 individual % 65 82 54 -- 61 -- other % 4 2 5 -- 2 -- Table A23.19 - Contract arrangement, by length of contract length of contract units total before 1990 1991-1995 1996 1997 1998 # of observations 647 147 165 74 198 63 Arrangement rented % 39 23 34 39 56 37 borrowed % 47 56 50 51 36 46 possesion %3 -- -- -- -- -- sharecropped % 9 10 9 11 7 15 other %I -- -- -- -- -- Table A23.20 - Length of contract, by rent parcel size parcel size adjusted mzs units total e I # of observations 647 440 207 Year of initial contract before 1990 % 23 21 23 1991-1995 % 25 22 27 1996 % 12 12 11 1997 % 31 36 28 1998 % 10 9 10 Annex 23, Page48 Table A23.21 - Source of rental lands, by length of contract length of contract inits total before 1990 1991-1995 1996 1997 1998 # of observations 647 147 165 74 198 63 Source of rental lands cooperative % 9 16 7 10 9 5 family % 22 28 23 16 20 14 individual % 65 51 65 67 68 79 other % 4 5 5 7 3 3 Annex 23, Page49 Table A23.22 - Household characteristics, by farm household typology. land owners renters overall /C.,/ miwnundo.3 adjusted m:s owners renters --et, , units total e-2 2-5 5-20 >20 e-2 >2 nmner, reLnierv # ofobservations 1254 180 186 203 177 346 162 746 508 Agricultural and livestock Total land mz 16,91 1.39 3 71 10.63 88.13 1.06 13.10 25.08 4.90 * ** Has cattle share .33 .21 .40 48 72 .11 .21 45 .14 ** Heads ofeattle # 5.05 1 04 2.45 5.79 23.48 .47 1.13 8.03 .68 . Head of household Female headed household share I I 13 .14 .15 .07 11 07 .12 .10 Age, head of household years 46.53 46.92 50.67 47.88 50.24 42.50 44.21 48.91 43.05 * Education, head ofhousehold years 2.22 2.68 2.02 2.24 2.75 1.90 1.99 2.42 1.93 * Education and labor activities Education, adults ears 2.93 3 17 2.77 3.13 3.53 2 64 2.57 3.14 2.61 *** Non agricultural wage labor # .28 .33 .25 .19 .20 .35 .31 24 .34 Agricultural wage labor # 36 33 .31 20 .23 .50 45 .27 .49 * Migrates # 11 .08 .10 .10 05 18 .08 .08 .I5 *** Migrates to city s 04 03 01 04 .03 .07 01 .03 .05 * Migrates to other countrv # 02 01 .01 .01 .01 .03 01 .01 03 * Agricultural practices Planted crops share .95 95 97 .97 .88 95 .98 .95 .96 corn share .72 .64 .77 .77 .71 .69 .79 .72 72 beans share .52 .51 .57 .60 .46 .51 .49 .53 .50 fruits share .62 .68 .66 .72 .60 .53 .60 .67 .55 . coffee share .09 .10 .14 .19 .09 .03 .03 .13 .03 *** HYV seeds share 05 .06 .04 .08 .08 .03 06 .06 .04 Fertilizer share .36 .41 .46 .45 .28 .27 .35 40 .30 G Pesticides share 47 .43 .52 .54 .46 44 .45 .49 .44 Hired in labor share .33 .31 .35 .47 .48 .18 .28 .40 .21 *** Sold crops share .60 .63 70 72 .50 .50 .59 .64 .52 *** Agrarian institutions Technical assistance exists share .24 .22 .27 .29 .23 .20 .25 .25 .22 used share 15 .13 .19 .20 .16 .11 .14 .17 .12 ** government share .07 .08 .11 .09 .06 .05 .06 .09 .05 * Used any kind ofcredit share .16 .18 .18 .18 .14 .14 .15 .17 15 Credit constrained share .38 .37 .33 .30 .34 .46 .43 .33 45 * * Producer organization or project share .08 .07 .09 .11 .14 .05 .07 .10 .05 ** Welfare indicators Flas dirt floor share .68 .70 .74 .64 .53 73 74 .65 .73 ** Piped water share .25 .29 .24 .24 .20 .24 .27 .24 25 No toilet share .32 .29 .31 .34 .32 .34 .28 .32 .32 Electricity share .32 .38 .30 .27 .29 .36 .30 .31 .34 PC household consumption C 4854 4201 4655 5540 7508 3699 451 3 5464 3958 *** Extreme poverty C .25 .24 .24 .25 .18 .34 .16 .23 .28 Moderate poverty C .36 .46 .40 .29 .30 .35 .40 .36 .36 ** Annex 23. Page5O Table A23.23 - Agricultural activities, bv farm household typology land owners renters adjusted mzs units iotal e-2 2-5 5-20 >20 e-2 >2 # of observations 1254 180 186 203 177 346 162 Agricultural activities planted crops share .95 .95 .97 .98 .88 .95 .98 com share .72 .63 .78 .77 .71 .68 .79 beans share .52 .50 .56 .59 .45 .51 .48 rice share .09 .07 .10 .09 .16 .04 .11 sorghum share .18 .14 .23 .15 .11 .19 .28 roots share .14 .14 .10 .17 .22 .09 .18 veggies share .13 .13 .17 .17 .04 .11 .15 fruits share .62 .68 .66 .72 .59 .53 .60 coffee share .09 .10 .14 .18 .09 .03 .03 Table A23.24 - Share of households having access to or using technical assistance, bv farm household typology land owners renters overall ,ei adjusted mzs owner renter units total e-2 2-5 5-20 >20 e-2 >2 # ofobset-vations 1254 i80 186 203 177 346 162 746 508 Technical assistance exists in community share .24 .22 .27 29 .23 .20 .25 .25 .22 used .15 .13 .19 .20 .16 .11 .14 .17 12 ** provided by govt .07 .08 11 .09 .06 .05 .06 .09 .05 ** provided by NGO/project 05 04 .07 .07 06 .04 .06 .06 .05 NGO .04 .02 .03 .05 .05 .03 .04 .04 04 * project .02 .01 .04 02 .01 .01 .01 02 .01 provided by other 02 02 .01 03 .03 .02 .02 .02 02 professional 00 00 00 .00 .01 .01 .00 00 .00 cooperative .01 .01 .00 .02 .01 .01 .00 .01 .01 private business .01 .01 .01 .00 .01 .00 .01 .01 .00 paid for services (over all households) .02 .04 .02 .03 .02 .00 00 .03 .00 paid for services (over households using servic s) .13 .18 .02 Annex 23, Page51 Table A23.25 - Share of households having access to or using technical assistance, by region and INTA zone Regions INTA zone Mana- Pacific Central Atlantic no un,r, otall gua urban nrurl urhan rural urhan rural zone Al A2 B3 B5 C6 t of observations 1254 92 52 300 76 532 39 162 261 170 182 212 290 13 Technical assistance exists in community share .24 .48 .21 .20 .34 .25 .24 .10 .24 .19 .21 .39 .18 I used .15 .22 .09 .12 .24 .18 .07 .04 .10 .13 I i .30 .11 .15 provided bygovt .07 .13 .07 .04 .15 .09 .01 .01 .05 .03 .07 .15 .05 .13 provided by NGO/project .05 .05 .00 .04 .08 .07 .05 .03 .04 .05 .02 .14 .04 3( NGO .04 .05 .00 .03 05 .05 .04 02 .03 .03 .02 .09 .03 .( 2 project .02 .00 .00 .01 .03 .02 .01 .01 .01 .01 .01 05 .01 ( I provided by other .02 .06 .02 .01 .03 .02 .03 .00 .02 .01 .01 .03 .02 2 professional .00 .03 .00 .00 .01 .00 .02 .00 .01 .00 .00 .00 .00 () cooperative .01 .03 .00 .01 .02 .01 .01 00 .01 .01 .00 .02 .00 ( 2 private business .01 .00 .02 .00 .00 .01 .00 .00 .00 .01 .00 .01 .01 0.) paid for services (over all househol s) 02 .00 .08 .01 .07 .02 .00 .00 .00 .03 .02 .04 .02 .02 Al: Leon ) Cthinandega_ A2: Managua. Masaya. Granada. Carazo and Rivas B3: Esteli, Madriz and NuLeva Segovia B5: Matagalpa. Jinotega and RAAN C6: Boaco, Chontales. Rio San Juan and RAAS Table A23.26 - Share of households in technical assistance zones, by farm household typology land owners renters overall adjusted m-s owner renter units total e-2 2-5 5-20 >20 e-2 >2 # of observations 1254 180 186 203 177 346 162 746 508: Technical assistance zones zona_Al share .14 .09 .23 .10 .13 .12 .15 .14 .13 zona_A2 share .15 .20 .15 .12 .02 .20 .15 .12 .18 zona_B3 share .17 .31 .20 .18 .05 .17 .08 .19 .14 zona_B5 share .23 .14 .22 .27 .20 .26 .27 .21 .26 zona_C6 share .11 .08 .05 .11 .18 .12 .12 .11 .12 Al: Leon y Chinandega A2: Managua, Masaya, Granada, Carazo and Rivas B3: Esteli, Madriz and Nueva Segovia B5: Matagalpa, Jinotega and RAAN C6: Boaco, Chontales, Rio San Juan and RAAS Annex 23, Page52 Table A23.27 - Primary source of technical assistance, by farm household typology overall owner renter units overall # of observations 207 140 67 % 100 68 32 Primary source neighbor 7 7 6 technician 86 85 89 radio 2 2 0 other 6 6 5 Table A23.28 - Content of technical assistance, by farm household typology. overall owner renter units overall of observations 351 230 121 Content new crops, diversification % 14 13 17 seed improvement 12 11 14 use of fertilzers and chemicals 28 27 29 irrigation 1 2 1 cattle improvement 6 8 2 veterinary advice 7 7 6 reforestation 7 8 7 post harvest 8 7 9 soil conservation 15 15 15 other 2 2 1 Annex 23, Page53 Table A23.29 - Received visit from INTA or IAG. by farm household typologv _ land owners renters overall owner renter adjusted mzs units overall e-2 2-5 5-20 >20 e-2 >2 # of observations 1254 180 186 203 177 346 162 746 SCS Received visit from INTA or MAG yes % 12 13 13 16 16 9 11 15 19 no % 88 87 87 84 84 91 89 85 9C T able A23.30 - Participated in any agricultural technology show/events, be farm household typology land owners renters overall owner renter adjusted mzs units overall e-2 2-5 5-20 >20 e-2 >2 # of observations 1254 180 186 203 177 346 162 746 508 Participated in event yes % 7 8 9 9 9 3 7 9 7 no % 93 92 91 91 91 97 93 91 93 Table A23.31 - Main reasons why technical assistance not used, by farm household typology land owners renters overall adjusted mzs owner renter units overall e-2 2-5 5-20 >20 e-2 >2 # of observations 1044 150 148 157 147 303 139 603 443 Main reason lack of access % 76 76 78 77 85 79 58 79 ,2 not available % 64 67 68 65 69 63 53 67 60 not available close to home % 12 9 9 12 16 16 5 11 12 too expensive % 4 9 2 2 3 2 6 4 3 not interested % 2 3 3 1 2 1 4 2 2 does not have time % 5 5 3 6 3 6 4 4 5 not needed % 10 5 15 8 3 9 26 8 14 other % 3 2 0 6 4 4 2 3 3 Annex 23. Page54 Table A23.32 - Share of households receiving credit, by rural household typology farm non farm and ownership and rental education urban overall ,.., farm non farm v.i adjusted mzs years farm farm n,,n farm units total e-2 2-5 5-20 >20 <4 >4 # of observations 2093 438 229 229 164 408 435 189 1060 843 Share of households receiving credit any kind of credit share .17 .16 .18 .17 .11 .13 .24 .19 .16 .19 loan .12 .12 .12 .11 .11 .08 .16 .11 .11 .12 loan from bank .02 .01 .02 .03 .04 .02 .02 .04 .02 .02 loan from organization .05 .04 .05 .05 .04 .01 .07 .06 .05 .04 loan from relative/friends .03 .03 .03 .02 .02 .04 .04 .01 .02 .04 loan for agriculture .05 .08 .09 .09 .08 .01 .01 .08 .09 .01 *** loan for non agriculture .03 .01 .00 .01 .02 .02 .07 .03 .01 .05 * loan for other .04 .03 .02 .01 .01 .05 .09 .00 .02 .07 * credit on purchases .07 .06 .08 .07 .01 .07 .10 .09 .05 .08 * Share of credit constrained households .35 .38 .29 .25 .35 .47 .34 .27 .33 .41 * Share of households providing loans .05 .03 .02 .09 .09 .06 .06 .03 .05 .06 Share of households with bank deposits .02 .00 .02 .01 .03 .02 .03 .06 .01 .03 Table A23.33 - Share of households receiving credit, by farm household typology land owners renters overall ,e. ownier v% adjusted mzs owner renter renler units total e-2 2-5 5-20 >20 e-2 >2 # of observations 1254 180 186 203 177 346 162 746 508 Share of households receiving credit any kind of credit share .16 .17 .18 .19 .14 .15 .15 .17 .15 loan .11 .12 .12 .11 .12 .10 .11 .12 .11 loan from bank .02 .01 .02 .03 .05 .00 .02 .03 .01 *** loan from organization .05 .05 .05 .05 .05 .04 .05 .05 .04 loan from relative/friends .02 .04 .01 .01 .02 .02 .03 .02 .02 loan for agriculture .09 .07 .10 .09 .09 .08 .08 .09 .08 loan for non agriculture .01 .01 .01 .01 .03 .01 .01 .02 .01 loan for other .02 .04 .01 .01 .01 .02 .02 .02 .02 credit on purchases .06 .06 .07 .08 .02 .06 .06 .06 .06 Share of credit constrained households .32 .29 .26 .22 .27 .42 .38 .26 .40 | Share of households providing loans .04 .02 .01 .04 .04 .02 .16 .03 .07 Share of households with bank deposits .02 .00 .03 .03 .05 .00 .00 .03 .00 * Annex 23, PageSS Table A23.34 - Why did not solicite credit, by rural household typology farm non farm and ownership and rental education urban overall farm non adjusted m:s years farm farn! units overall e-2 2-5 5-20 >20 <4 >4 # of observations 1795 372 191 190 143 378 357 165 Reason Credit constrained % 42 45 35 31 40 52 42 32 38 4i no credit available in community 10 8 1] 15 20 9 5 13 12 7 no collateral 28 32 20 12 15 38 35 16 22 37 other 4 5 4 4 5 5 2 3 4 4 Nof credit constrained 56 53 64 66 56 45 56 66 59 59 too risks 20 19 19 18 17 18 22 26 18 29 too expensive 13 14 17 20 18 8 10 1 4 17 7 does not need 23 20 28 28 21 19 24 26 24 21 other 2 2 0 3 4 3 3 2 2 3 Table A23.35 - Why did not solicite credit, by farm household typology land owners renters overall adjusted mzs owner rentei units overall e-2 2-5 5-20 >20 e-2 >2 # of observations 1234 172 178 195 176 350 162 Reason Credit constrained % 38 36 32 28 32 48 44 32 45 no credit available in community 12 11 13 15 22 7 11 15 8 no collateral 22 17 16 8 6 37 30 12 35 other 4 8 3 5 4 4 3 5 3 Not credit constrained 62 64 69 73 68 52 56 6- J 3 too risky 20 23 22 22 18 19 13 21 ,17 too expensive 16 14 16 14 23 13 23 16 16 does not need 24 27 30 33 23 18 20 28 18 other 2 0 1 4 4 2 0 2 2 Annex 23. Page56 'I'able A23.36 - Share of households participating in organizations, by rural household typology land ownership and rental education urban overall test Jarm non larm vs. adjusted mzs years farm non farm units total e-2 2-5 5-20 >20 <4 >4 # of observations 2097 438 229 229 164 408 435 194 1060 843 Share of households participating producer organization share .04 .04 .04 .11 .11 .00 .00 .09 .06 .00 * project share .02 .03 .03 .04 .04 .00 .00 .04 .04 .00 non producer any organization share .34 .32 .39 .42 .37 .24 .34 .39 .36 .29 ** religious share .26 .24 .32 .32 .26 .21 .25 .29 .28 .23 school share .04 .05 .05 .03 .04 .01 .06 .08 .04 .03 Table A23.37 - Share of households participating in organizations, by farm household tvpologv land owners renters overall Ownler *iw adjusted mzs owner renter renrer units total e-2 2-5 5-20 >20 e-2 >2 # of observations 1254 180 186 203 177 346 162 746 508 Share of households participating producer organization share .07 .07 .07 .11 .13 .03 .05 .09 .03 project .04 .03 .05 .04 .06 .03 .03 .04 .03 non producer any organization .37 .37 .46 .38 .36 .29 .42 .39 .33 * religious .28 .32 .37 .31 .24 .20 .33 .31 .24 school .05 .05 .06 .01 .05 .05 .07 .04 .06 'I'able A23.38 - Primary benefit received from participation in organizations and projects organization project units # of observations 98 57 Primary benefit investment % 9 4 technical assistance % 48 55 inputs % 12 8 credit % 3 9 other % 9 13 nothing % 20 12 Annex 23, Pa,e57 Table A23.39 - Project institutions units overal7 # of observations 57 Project institutions MAG % 23 other government % 16 NGO % 44 producer % 10 other % 7 Table A23.40 - Use of agrarian institutions, by type of agrarian institutions technical organization/ assistance credit project total no yes no yes no yes #of obs 1254 1070 184 111] 143 1154 100 Technical assistance exists in community .24 .11 1.00 na .22 .39 ** .19 .71 * used .15 .00 1.00 na .13 .27 ** .10 .63 * provided by govt .07 .00 .49 na .07 .11 * .05 .32 *' provided byNGO/project .05 .00 .36 na .05 .10 ** .03 .26 NGO .04 .00 .26 na .03 .07 * .02 .22 * project .02 .00 .11 na .01 .03 .01 .05 paid .02 .00 .13 na .02 .05 * .01 .08 Financial services provided loans .04 .04 .05 .05 .04 .04 .08 bank deposit .02 .01 .07 ** .01 .06 ** .01 .10 * any kind of credit .16 .14 .29 ** .05 1.00 ** .14 .39 * loan .11 .10 .21 ** .00 1.00 na .10 .29 *" loan from bank .02 .02 .05 * .00 .19 na .02 .04 loan from organization .05 .03 .12 ** .00 .41 na .03 .18 * purchase credit .06 .05 .10 ** .05 .09 na .05 .12 credit constraint .32 .33 .24 .36 .00 na .33 .20 Producer organization .08 .04 .35 ** .07 .21 ** .00 1.00 na organization .07 .03 .31 ** .05 .17 ** .00 .81 na project .04 .01 .17 ** .03 .10 ** .00 .44 na Non producer organization .37 .35 .49 ** .36 .42 .36 .51 religious .28 .27 .35 * .28 .32 .28 .36 Annex 23, Page58 Table A23.41 - Household characteristics of farm households, by type of agrarian institutions technical organization/ assistance credit project total no yes no yes no yes #of obs 1254 1070 184 1111 143 1154 100 Assets Total land, adjusted 14.73 13.69 20.78 13.93 21.02 12.41 40.32 perennial 1.63 .62 7.47 1.60 1.80 .59 13.03 annual 13.84 13.94 13.25 13.70 14.97 13.40 18.65 pasture 5.45 5.22 6.74 4.09 15.99 3.53 26.60 owns land .59 .58 .68 ** .59 .63 .58 .74 rents land .47 .48 .41 .47 .49 .48 .35 no documentation .13 .13 .15 .14 .13 .13 .14 land registered .37 .36 .46 ** .37 .40 .37 .47 has cattle .33 .31 .42 ** .32 .34 .31 .48 heads of cattle 5.05 4.41 8.75 * 4.47 9.56 4.10 15.51 education, adults 3.01 2.84 4.01 ** 2.91 3.74 ** 2.87 4.51 Labor activities non agricultural wage labor .32 .31 .37 .30 .42 ** .31 .35 agricultural wage labor .33 .33 .34 .34 .28 .33 .35 self employed .19 .19 .22 .19 .23 .19 .24 Agricultural practices grew coffee .09 .08 .13 * .09 .11 .09 .12 sold any crop .60 .59 .65 .57 .79 ** .59 .71 HYV .05 .04 .12 ** .05 .10 ** .04 .18 organic fertilizer .05 .05 .07 .05 .05 .04 .12 fertilizer .36 .33 .50 ** .33 .61 ** .34 .58 pesticides .47 .44 .61 ** .45 .63 ** .45 .63 hired in labor .33 .30 .50 ** .30 .51 ** .30 .59 Annex 23, Page59 Table A23.42 - Welfare indicators, by type of agrarian institutions technical organization/ assistance credit project total no yes no yes no yes #of obs 1254 1070 184 1111 143 1154 100 has dirt floor .68 .70 .57 ** .69 .62 .69 .56 ** has piped water .25 .23 .35 ** .24 .32 .23 .38 ** no toilet .32 .34 .19 ** .33 .21 ** .33 .16 ** electricity .32 .30 .45 ** .30 .48 ** .31 .49 ** refrigerator .07 .06 .13 ** .07 .09 .06 .18 ** TV, black and white .21 .21 .24 .19 .36 .21 .24 TV, color .07 .06 .11 .06 .10 .05 .21 Stereo .04 .03 .07 .04 .05 .03 .13 ** VCR .02 .01 .04 .02 .02 .01 .09 ** Sewing machine .09 .08 .09 .08 .11 .08 .10 Vehicle .03 .03 .05 .03 .05 .03 .12 ** percapitaconsumption 4854 4700 5746 * 4731 5815 ** 4616 7477 ** extreme poverty .18 .19 .12 * .19 .10 ** .19 .08 ** moderate povertv .44 .44 .43 .44 .40 .44 .40 Table A23.43 - Participation in corn market, by land categories owner renter adjusted units total e<2 2<5 5<2 >20 e<2 2< #of 1068 134 170 184 145 279 151 Buyand % 6 8 3 4 1 9 8 Net % 2 29 20 13 16 27 12 Self % 4 38 43 41 53 46 49 Net % 21 25 33 42 21 19 Annex 23, Page6O Table A23.44 - Participation in bean market, by land categories owner renter adjusted units total e<2 2<5 5<2 >20 e<2 2< # of 82 114 132 152 10 221 98 Tota % 10 100 100 100 10 100 10 Buy and % 13 19 16 7 8 12 13 Net % 2 30 25 25 2 29 33 Self % 33 24 24 34 4 41 24 Net % 27 25 34 34 26 18 29 Table A23.45 - Principal problems for comercialization of agricultural products, by farm household typology land owners renters overall owner renter Percentage of responses adjusted mzs units total e-2 2-5 5-20 >20 e-2 >2 # of observations - 1982 264 289 335 316 519 258 1204 777 Principal problems lo", sale price 21 19 20 22 23 21 21 high intermediary commisions 3 3 2 2 4 2 4 3 3 high transaction costs 27 26 28 32 35 19 23 31 20 high transport costs 6 9 9 7 5 4 5 7 5 lack of transportation 3 3 4 4 4 1 3 4 2 point ofsale very far 8 8 7 9 12 8 5 9 7 poorroad 4 4 5 5 5 3 3 5 3 lack of road 4 1 2 5 8 2 5 4 3 dangerous I 0 1 3 2 1 1 1 1 does not sell 16 19 14 10 8 26 16 13 22 low demand 2 2 0 2 3 2 2 2 2 belongs to cooperative I I 0 1 0 0 2 0 1 other 2 2 2 1 4 1 2 2 2 nothing 28 29 33 30 22 28 30 28 28 Annex 23, Page6l Table A23.46. Share of all farm households participating in agricultural markets, by farm household typology owner renter total adjusted units e<2 2<5 5<2 >20 e<2 2< #of 125( 179 185 202 17 345 16 Cro shar .63 .67 .72 .75 .56 .52 .59 corn .26 .22 .29 .37 .25 .20 .31 bean .21 .24 .29 .25 .17 .16 .21 whea .01 .02 .02 .02 .0 .01 .00 rice .05 .04 .06 .05 .09 .01 .07 sorghu .0 .04 .11 .06 .06 .10 .13 roots .05 .08 .04 .06 .06 .03 .07 veggie .07 .07 .08 .10 .01 .05 .06 fruit .1 .23 .22 .30 .17 .12 .15 citru .0 .11 .12 .09 .05 .04 .03 coffe .0 .08 .12 .15 .1C .01 .00 other .04 .02 .06 .06 .06 .01 .07 Table A23.47 - Share of crop specific farm households participating in agricultural by farm household typology owner renter unit total S Crop (over all producers of each shar .6: .61 .55 corn .3 .3 .31 bean .3 .4 .34 whea .3 .44 .24 rice .53 .5 ~ .51 sorghu .4 .4( .46 roots .36 .3, .35 veggie .21 .2 .12 fruit .4 .5 .40 citru .3 .3 .22 coffe .7 .8 .12 other .8' .8: .91 number of observations varnes by crop and Annex 23, Page62 Table A23.48 - Livestock ownership and technology, bv rural household typology farm non farm urban farm adjusted mcs units overall e-2 2-5 5-20 >20 0-4 >4 # of observations 2097 438 229 229 164 408 435 194 Share of households with cattle share .21 .15 .32 .45 .65 .05 .03 .32 pigs share .36 .42 .46 .55 .65 .22 .16 .33 fowl share .62 .70 .81 .80 .81 .51 .45 .47 horses share .06 .09 .09 .14 .17 .01 .00 .06 Cattle beginning survey period # 4.28 .55 1.57 3.55 14.38 - - 8.13 end survey period 5.11 .63 1.89 4.03 17.57 .16 .13 9.67 change % 20 16 20 13 22 - - 19 Pigs beginning survey period # 2.06 1.15 3.08 3.36 2.97 - - .63 end survey period # 1.93 1.21 1.41 3.38 3.67 .46 .58 .97 change % -6 6 -54 1 24 - - 54 Fowl beginningsurveyperiod # 13.61 9.94 14.35 19.40 17.29 - - 10.96 end survey period 4 12.02 9.57 13.27 16.66 15.55 5.24 5.03 7.44 change % -12 -4 -8 -14 -10 - - -32 Share of households in livestock production using technology vacunas share .26 .24 .32 .42 .55 .01 .02 .51 corrals . share .07 .02 .05 .14 .26 .00 .00 .18 veterinary shre 02 .00 .02 .03 .07 .00 .00 .04 Annex 23. Page63 Table A23.49 - Livestock ownership and technology, by farm household typology land renter adjusted units total e-2 2-5 5- >20 e-2 >2 # of 125 180 186 203 17 346 16 Share of households cattl shar .3 .21 .39 .49 .7 .11 .2 pigs shar .4 .36 .43 .56 .6 .43 .4 fow shar .7 .69 .78 .78 .7 .65 .7 horse shar .1 .05 .12 .15 .1 .09 .0 Cattl beginning survey # 4.2 .85 1.79 5.10 19.7 .45 .9 end survey 5.1 1.06 2.45 5.81 23.7 .47 1.1 chang % 2 24 37 14 20 5 1 Pigs beginning survey 2.0 .66 3.42 1.11 2.9 1.23 4.1 end survey # 1.9 .83 1.37 1.89 3.4 1.26 3.6 chang %- 26 -60 69 1 3 -1 Fowl beginning survey # 13.6 8.87 13.3 19.8 18.0 9.77 14.8 end survey # 12.0 9.29 13.0 17.4 15.7 8.83 9.8 chang % -1 5 -2 -13 -1 -10 -3 Share of households in livestock using vacuna shar .3 .29 .38 .45 .64 .23 .2 corral shar .11 .01 .09 .17 .31 .03 .0 veterinar shar .0 .01 .01 .02 .09 .01 .0 Annex 23, Page64 Table A23.50 - Household characteristics. bv cattle nroduction categories cattle categories none none units total no farm farm 1-5 >5 of obs 2091 808 840 234 210 total land, adjusted mz 8.83 .00 7.06 11.59 47.01 perennial mz .97 .00 1.62 .17 3.06 annual mz 8.29 .00 6.80 14.74 39.11 pasture mz 3.26 .00 .68 .73 29.25 has land share .36 .00 .48 .64 .88 rents land share .28 .00 .57 .32 .16 heads of cattle # 3.08 .00 .00 2.59 28.02 4 pigs 1.36 .42 1.45 2.57 3.32 vaccinated share .19 .01 .18 .39 .73 corral share .05 .00 .02 .09 .34 produces milk share .17 .01 .03 .61 .93 female head of household share .18 .27 .12 .14 .07 age, head of household years 45.31 43.16 45.66 46.14 51.17 education, adults years 3.42 4.19 2.81 2.61 3.83 non agricultural wage share .43 .60 .36 .25 .26 agricultural wage share .36 .39 .38 .33 .18 self employed share .23 .30 .19 .18 .23 patron share .10 .02 .11 .12 .33 planted com share .43 .00 .70 .68 .76 planted bean share .31 .00 .51 .50 .47 planted coffee share .06 .00 .08 .10 .08 hired in labor share .20 .01 .26 .28 .57 technical assistance exists share .14 .00 .23 .20 .29 used share .09 .00 .13 .15 .20 govemment share .04 .00 .06 .07 .11 NGO share .03 .00 .05 .06 .06 used credit share .17 .18 .16 .18 .18 loan share .12 .11 .11 .12 .13 bank share .02 .02 .02 .02 .05 organziation share .05 .04 .05 .03 .06 purchase credit share .07 .08 .06 .06 .05 credit constrained share .41 .46 .43 .38 .21 producer organization share .05 .00 .06 .07 .15 has dirt floor share .66 .60 .71 .75 .54 has piped water share .33 .47 .26 .20 .24 electricity share .43 .59 .33 .27 .39 Vehicle share .03 .03 .03 .03 .06 per capita consumption C 5123 5461 4379 4473 7564 extreme poverty share .21 .15 .28 .22 .12 moderate poverty share .37 .38 .37 .37 .27 Atlantic. urban share .02 .00 .03 .02 .05 Atlantic. rural share .10 .05 .11 .15 .18 Central, urban share .04 .00 .06 .04 .08 Central, rural share .42 .43 .44 .38 .36 Pacific. urban share .02 .00 .04 .02 .05 Pacific. rural share .30 .39 .24 .31 .21 Managua share .10 .13 .08 .09 .06 Annex 23, Page6S Table A23.51 - Types of shocks experienced by households, by farm household typology land renter overal Percentage of adjusted owne rente units total e-2 2-5 5- >20 e-2 >2 # of 243' 321 372 398 36 668 31I 1452 98 Shoc drough % 53 53 53 52 51 52 55 52 53 infestatio 33 33 36 31 2 36 32 32 3 floo 3 3 2 5 3 3 3 3 3 theftof 1 I 0 1 5 1 0 2 0 theft 4 2 5 6 7 3 5 5 4 extorsio 0 0 0 0 0 0 0 0 0 physical 0 0 0 0 0 0 0 0 0 land 0 0 0 0 0 0 0 0 0 kidnappin 0 0 0 0 1 0 0 0 0 othe 1 I 1 1 3 1 1 2 1 No 4 7 2 3 2 4 2 4 4 Table A23.52 - Response to household shocks, by farm household typology land renter overal Percentage of adjusted owne renme responses units total e-2 2-5 niziO >20 e-2 >2 r r # of 1782 223 260 290 27( 474 26( 1049 73 Respons loan without % 1 1 1 0 0 0 1 1 1 Itith 1 0 0 1 2 0 1 1 1 dMO paying 1 1 3 1 1 1 3 1 2 2 0 0 2 7 2 2 3 2 6 3 3 5 12 3 7 6 5 bgm 2 2 0 3 1 2 1 2 2 h4tIi;om 1 0 1 0 0 1 1 0 1 Np¢d extra 11 8 8 12 1s 9 1 11 l h%Ied household 24 30 20 21 20 30 2 22 21 planted other 3 2 3 5 3 2 5 3 3 * organic 8 11 13 8 4 5 6 9 5 trnent 3 2 6 2 4 3 2 3 3 nothin 31 40 41 39 29 41 3 37 :3 Annex 23, Page67 Table A23.54 - Determinants of household consumption 25th quantile 50th quantile 75th quantile Kegresslon-analysis-Using simultaneo-u--q-0-iaTiTe-- regression with bootstrapped errors # obs 1843 Dependent variable = log per capita household Psuedo R2 .25 .27 .28 consumption H-=251h= Coef P>Itl Coef P>lt Coef P>tI 50h=-5th=0 Land assets perennial .0022 .00 * .0024 .00 *** .0022 .31 annual .0011 .15 .0016 .19 .0030 .01 * ** pasture .0015 .04 ** .0027 .00 * .0043 .00 * owns and rents in land .2215 .02 * .1835 .07 * .2223 .02 ** Human capital assets female headed household -.0813 .10 * -.1156 .02 ** -.1626 .04 * age, head of household -.0012 .57 -.0013 .36 -.0006 .78 # men less then primary education .0319 .22 -.0028 .88 .0099 .70 # men primary education .0759 .09 * .0549 .13 .0827 .05 ** # men secundary education .0383 .61 .1187 .06 * .1325 .17 # men high education .3092 .00 *** .2665 .01 *** .2143 .03 ** * # women less then primary education .0693 .06 * .0808 .00 t .0257 .17 *** # women primary education .1142 .00 ** .0991 .00 *** .0246 .48 # women secundarv education .2246 .00 .1891 .00*** .1723 .01 *** # women high education .2775 01 *** .2952 .00 *** .1634 .10 * log familv size -.5478 .00*** -.5966 .00 -.6232 .00** # children e-4 years -.0657 .01 *** -.0561 .00 -.0195 .53 #children 5-10 -.0163 .43 -.0107 .57 -.0136 .64 # males 11-14 .0039 .90 .0016 .95 .0155 .59 # females 11-14 -.0094 .67 .0049 .84 .0108 .81 # males 15-19 .0086 .81 .0116 .67 -.0191 .43 4 females 15-19 .0138 .74 -.0156 .66 -.0008 .98 # males 20-34 .0099 .70 .0143 .60 -.0120 .74 4 females 20-34 .0425 .22 .0034 .92 .0198 .60 # males 35-59 .0540 .25 .0380 .23 -.0094 .84 # females 35-59 -.0052 .90 .0448 .21 .1261 .03 ** # males 60t -.0038 .94 -.0065 .89 -.0623 .40 # females 60+ -.0783 .21 -.0298 .31 .0673 .29 Other assets value total stock, own business .0000 .08 * .0000 .00 * .0000 .00 * pension -.1367 .42 -.0576 .68 .0207 .89 Agrarian institutions credit constrained -.1053 .01 O -.1232 .00 M -.1582 .00 ** Infrastructure had running water in 1993 .0861 .07 * .1239 .00 * .0953 .07 * did not have toilet in 1993 -.0361 .48 -.0608 .16 -.0796 .10 * has electricity .2276 .00 *** .2239 .00 * .1904 .00 *'* Distance community has paved road .0943 .07 ' -.0400 .36 -.0110 .91 Regions Atlantic -.1978 .08* -.1402 .13 -.3703 .03* Center -.2297 .08 * -.2154 .01 *** -.4295 .01 * Pacific -.1172 .30 -.1523 .06* -.3145 .05* Constant 8.8002 .00 9.2130 .00*** 9.7412 .00 *"signiicant at I percent **significant at 5 percent *significant at 10 percent Annex 23, Page68 Table A23.55 - Determinants of household income robust median Regression analysis # obs 1843 4 obs 1843 Dependent variable = log per capita household Prob>F 0 Psuedo R2 .19 income Coef P>l Coef P>lt. Land assets perennial .0028 .00 ** .0024 .01 annual .0017 .00 *** .0019 .01 pasture .0026 .00 * .0022 .31 owns and rents in land .0(007 1.00 -.0242 .77 Human capital assets temale headed household -.0178 .76 .0330 .61 ace. head of household -.0007 .72 -.0003 .89 # men less then primary education .0235 .43 .0213 .36 # men primary education .1248 .00 * .1294 .00 # men secundary education .0576 .38 .1252 .11 # men high education .3160 .02 * .3434 .00 # women less then primary education .0823 .01 * .0983 .05 # women primary education .0919 .03 ** .0907 .06 # women secundary education .1934 .01 * .1956 .00 # women high education .2819 .01 *** .3089 .01 log family size -.7379 .00 *** -.833 7 .00 4 children e-4 years -.0533 .03 ** -.0317 .21 # children 5-10 -.0061 .80 .0102 .67 # males 11-14 -.0037 .91 -.0186 .66 # females 11-14 .0230 .53 .0567 .23 # males 15-19 .0705 .05 ** .1122 .00 4 females 15-19 .0267 .49 .0600 .14 # males 20-34 .0932 .01 * .1131 .00 # females 20-34 .0608 .17 .0578 .31 # males 35-59 .1029 .03 * .1496 .00 4 females 35-59 .1129 .02 * .0534 .35 # males 60+ -.0052 .94 .0395 .48 4 females 60+ .0166 .79 .0351 .68 Other assets value total stock. own business .0000 .00 * .0000 .00 Agrarian institutions credit constrained -.1187 .00 *** -.1330 .02 Infrastructure had running water in 1993 .1256 .02 ** .1136 .14 did not have toilet in 1993 -.0967 .02 * -.1145 .00 has electricity .2466 .00 * .2371 .00 Distance community has paved road .0787 .26 .0689 .41 Regions Atlantic .0637 .56 .0179 .88 Center .0716 .49 .0103 .92 Pacific .0815 .42 .0481 .59 Constant 8.7178 .00 *** 8.8202 .00 significant at I percent **significant at 5 percent *significant at 10 percent Annex 23, Page69 I able A23.56 - Determinants of agricultural income I obit analysis #obs 18TM Dependent variable = per capita Prob>y 0 agricultural income Coef P>ltj Land assets perennial 39 .00 S*+ annual 0 .95 pasture 24 .00 owns and rents in land 1124 .06 * Human capital assets female headed household -721 .05 ** age, head of household 4 .76 # men less then primary education 54 .77 # men primary education 497 .07 * # men secundary education -275 .51 4 men high education -507 .56 # women less then primary education 116 .57 # women primarv education 381 .16 # women secundary education -395 .38 # women high education 93 .89 log family size -827 .14 perennial*average education -2 .07 * annual*average education 8 .00 ** pasture*average education 0 .86 4 children e-4 years -173 .27 4 children 5-10 3 .98 4 males 11-14 -29 .89 # females 11-14 -101 .66 # males 15-19 -74 .74 4 females 15-19 30 .90 4 males 20-34 -62 .78 # females 20-34 58 .83 # males 35-59 634 .04 ** # females 35-59 214 .49 # males 60+ 410 .31 # females 60+ 36 .93 Other assets value total stock, own business 0 .58 Agrarian institutions credit constrained -780 :00 * Infrastructure had running water in 1993 -202 .56 did not have toilet in 1993 445 .08 * has electricity -1175 .00 Distance community has paved road -1257 .01 '* Regions Atlantic 555 .43 Center 634 .34 Pacific -353 .59 Constant 1097 .23 490 left censored observations 1353 uncensored observations ***significant at I percent **significant at 5 percent *significant at 10 percent Annex 23. Page7O Table A23.57 - Determinants of off farm income non agricultural agricultural non agricultural Tobit analvsis self employment wage labor wage labor 4 obs 1843 1843 1843 Dependent variable = per capita income Prob>X 0 0 0 (by off farm category) Coef P>11n Coef P>lt Coef P>t Land assets perennial 14 .02 ** -430 .09 * -10 .18 annual -I .71 -6 .10 * -13 .00 pasture 11 01 *** -38 .02 ** -20 .02 owns and rents in land 993 .18 888 .19 -2330 .01 Human capital assets female headed household 698 .11 520 .21 255 .56 age. head of household -14 .33 7 .61 -36 .01 4 men less then primary education 319 .18 206 .32 -272 .22 4 men primarv education 227 .50 475 .11 165 .60 #men secundarv education -613 .21 589 .17 152 .75 #men high education 559 .53 -1303 .20 2123 .02 4 women less then primary education -74 .78 256 .27 265 .29 # women primary education 47 .89 -143 .63 -1 1.00 # women secundarv education 760 .12 -18 .97 314 .53 # women high education 648 .38 64 .94 1682 .02 log family size -320 .65 -578 .35 -2743 .00 4 children e-4 years 4 .98 126 .47 284 .13 # children 5-10 97 .61 61 .71 270 .13 # males 11-14 -60 .82 -47 .84 112 .65 # females 11-14 -127 .66 -99 .70 460 .09 # males 15-19 -571 .06 * 441 .08 * 1021 .00 4 females 15-19 -142 .65 454 .09 * 776 .01 4 males 20-34 -145 .61 533 .03 ** 1096 .00 # females 20-34 302 .38 373 .23 464 .16 # males 35-59 422 .26 110 .75 1332 .00 # females35-59 982 .01 *' -162 .66 583 .13 # males 60+ 456 .35 -699 .15 777 .12 # females 60+ 106 .83 -728 .14 -277 .57 Other assets value total stock, owrn business .294 .00 *'* 0 .14 -.049 .08 Agrarian institutions credit constrained 827 .00 ** 394 .13 -72 .79 Infrastructure had running water in 1993 597 .12 -99 .79 656 .09 did not have toilet in 1993 -605 .07 -100 .73 -727 .02 has electricity 1628 .00 *** 418 .20 1535 .00 Distance community has paved road 31 .95 153 .75 1227 .01 *** Regions Atlantic -417 .61 -2029 .01 * -1413 .07 Center 99 .89 -1803 .01 *** -399 .58 Pacific 1016 .15 -691 .28 192 .79 Constant -4826 .00 * -2429 .01 *** 1504 .14 left censorted observations 1401 1349 983 uncensored observations 442 494 860 ***significant at I percent "*significant at 5 percent *significant at 10 percent Annex 23, Page7l I aDie AZJ.-a - Llelerminanis o1 paruiciparmon in onj 1arm activities non agricultural non agricultural agricultural self employment wage labor wage labor #obs 1843 # obs 1843 # obs 1843 Probit analysis Prob>x 0 Prob>7 0 Prob>x 0 Dependent variable = participation in Psuedo R2 .10 Psuedo R2 .24 Psuedo R2 .10 off farm activities dF/dx P>IzI dF/dx P>Izl dF/dx P>Izi Land assets perennial -.0110 .30 -.0810 .02 M* .0002 .68 annual -.0003 .32 -.0002 .37 -.0028 .00 pasture -.0009 .16 -.0030 .01 *** -.0015 .04 owns and rents in land .1871 .00 *** -.0351 .51 -.0503 .44 Human capital assets female headed household .0514 .09 * .1286 .00 *** -.0282 .47 age. head of household -.0004 .64 -.0018 .06 * -.0006 .64 # men less then primary education .0345 .03 * .0245 .12 -.0330 .1 ( # men primarv education .0337 .12 .0535 .02 -.0231 .43 # men secundary education .0256 .40 .1096 .00 ** -.1691 .00 # men high education .0925 .12 .1639 .04 ** -.1656 .08 # women less then primary education .0206 .22 .0379 .03 ** -.0151 .50 # women primary education .0353 .09 * .1005 .00 *** -.0902 .00 #womensecundaryeducation .0150 .64 .1126 .00 *** -.2193 .00 # women high education .0191 .70 .2811 .00 *** -.1185 .18 log family size .0576 .21 .0663 .17 -.0165 .78 # children e-4 years .0021 .87 -.0256 .06 * .0297 .08 # children 5-10 -.0041 .74 -.0136 .29 -.0085 .60 # males 11-14 -.0065 .70 -.0153 .38 .0249 .26 # females 11-14 -.0177 .33 .0152 .43 .0349 .16 # males 15-19 -.0461 .02 ** .0073 .70 .1104 .00 # females 15-19 -.0418 .04 ** .0111 .58 .0766 .00 # males 20-34 -.0485 .01 *** .0135 .47 .1264 .00 # females 20-34 .0081 .72 .0084 .72 .0489 .10 # males 35-59 .0046 .85 .0101 .69 .0926 .00 # females 35-59 .0535 .03 ** -.0289 .28 .0376 .28 # males 60+ .0167 .61 -.0265 .44 .0568 .20 # females 60+ -.0143 .66 -.0480 .17 -.0708 .11 Other assets pension -.1292 .03 * -.0431 .56 -.0360 .76 Agrarian institutions credit constrained .0247 .18 .0300 .12 .0581 .02 Infrastructure had running water in 1993 .0208 .42 .0225 .41 -.0596 .10 did not have toilet in 1993 -.0221 .29 -.0284 .18 -.0252 .35 has electricity .0978 .00 *** .0938 .00 *** .0138 .66 Distance time required to get to school -.0370 .04 ** -.0648 .00 *** .0012 .94 community has paved road .0358 .29 .0854 .03 ** .0635 .19 Regions Atlantic -.0240 :64 -.1917 .00 * -.0186 .81 Center -.0166 .73 -.2032 .00 * .0908 .21 Pacific .0640 .19 -.0958 .06 * .0883 .22 ***significant at I percent **significanlt at 5 percent *significant at 10 percent Annex 23, Page72 Table A23.59 - Household Characteristics by year and poverty category extreme moderate non poor poverty poverty Variable units 1993 1998 1993 1998 1993 1998 in agriculture share .65 .70 .63 .59 .53 51 heads of cattle 4 1.21 .94 1.99 1.92 4.03 5 05 female headed household share .18 .22 .15 .15 .23 17 age. head of household vears 45.61 46.48 43.67 44.79 45.47 45 79 aver years of education. adults years 1.42 2.02 2.28 2.97 3.89 4 61 aver vears of education. head years .85 1.31 1.68 2.28 3.08 3 70 4 men no education 4 1.04 .97 .82 .67 .48 4I # men less than primary education # .57 .74 .48 .66 .52 54 # men primary education .14 .17 .25 .28 .36 41 4 men secundary education # .02 .04 .07 .09 .14 16 # men high education .00 .00 .00 .02 .06 10 4 women no education 1.05 .86 .76 .68 .52 36 4 women less than primary education 4 .51 .66 .59 .59 .56 49 4 women primary education .12 .24 .20 .32 .32 33 # women secundarv education # .02 .04 .07 .09 .19 26 # women high education # .00 .00 .01 .02 .02 J9 # adults # 3.48 3.83 3.27 3.51 3.17 3 23 family size 4 8.36 8.95 7.21 7.12 6.12 5 56 dependency ratio share .57 .56 .52 .49 .43 38 # children e-4 years # 1.56 1.67 1.30 1.13 .94 70 # children 5-10 # 1.92 1.96 1.53 1.40 1.14 89 # males 11-14 4 .62 .70 .45 .44 .36 29 # females 11-14 # .53 .52 .47 .43 .32 28 # males 15-19 .53 .66 .42 .46 .40 35 # females 15-19 # .44 .49 .37 .49 .38 .37 # males 20-34 # .62 .70 .66 .70 .61 66 4 females 20-34 # .67 .70 .72 .68 .68 66 # males 35-59 # .62 .57 .52 .57 .49 54 # females 35-59 # .59 .64 .53 .55 .49 51 # males60+ # .12 .20 .14 .14 .16 .19 4 females 60+ .14 .14 .11 .15 .17 .13 housing density 6.59 7.17 5.51 5.38 4.09 3.86 per capita consumption (1993 C) 1993 C 820 900 1778 1771 5257 4204 Atlantic share .09 .16 .15 .11 .08 .08 Central share .58 .52 .42 .48 .29 .38 Pacific share .29 .29 .33 .37 .41 .36 Managua share .03 .03 .10 .05 .22 .19 Annex 23. Page73 Table A23.60 - Determinants of p verty, 1993 and 1998, r ral households. 1993 1998 187# 207 Probit Prob X 0 Prob X0 Dependent variable living in Psuedo .2 Psuedo 4 dF/d P>Izl dF/d P>IzI Land in .014 .58 .40 heads of - .00*** .00 Human capital female headed - .24 .051 .23 age, head of .001 .34 .003 .01 # men less then primary - .22 .004 .86 # men primary - .24 - .05 # men secundary - .07 * - .07 # men high - .00*** - .47 # women less then primary - .07 * - .34 # women primary - .01** - .46 # women secundary - .00*** -.00 # women high - .02** -.03 log family .319 .00*** .387 .00*** # children e-4 .003 .85 .044 .03 * # children 5- .000 1.00 .029 .13 #males I 1- .58 .013 .62 # females I I- .49 .91 # males 15- .29 .015 .58 # females 15- .001 .94 .025 .40 # males 20- .009 .72 .025 .38 # females 20- .001 .95 - .69 # males 35- .004 .89 - .59 # females 35- .053 .13 - .70 # males .14 - .37 # females .55 .051 .25 Housin owns .001 .96 -.00 rents .78 .006 .85 Infrastructu had dirt .093 .00*** .130 .00*** had running .003 .91 -.01 did not have .022 .42 .082 .01 had .00*** -.00 housing .00*** -.00 Region Atlanti .218 .00*** .146 .05 Cente .277 .00*** .200 .01 Pacifi 122 .00*** .085 .25 ***significant at 1 **significant at 5 *significant at 10 Annex 23, Page74 Table A23.61 - Decomposition of change in consumption, 1993-1998 rural urban 1993-1998 1993-1998 AB DX AB DX Agricultural/livest 8 0 2 2 Age head of household -40 0 -37 3 Gender head of household 2 0 14 0 Family -36 -64 -447 -30 Male 14 -7 -91 -1 Female 23 -8 -64 -1 Demographic composition 16 0 368 2 Housing -3 1 -32 2 Infrastructure -7 -8 -20 2 Atlantic region 1 -I -4 0 Central region -30 3 -15 3 Pacific region -6 0 60 - Constant 41 407 Iota -17 -81 141 Annex 23, Page75 Table A23.62 - Decomposition of change in consumption 1993-1998, urban and rural, by region rural urban Central Pacific Central Pacific 1993-1998 1993-1998 1993-1998 1993-1998 AB DX AB DX AB DX AB DX Agriculturalflivestock -5 0 40 -2 5 0 11 -8 Age head of household -7 -1 -44 0 18 1 -35 - I Gender head of household 2 0 7 0 -19 2 0 0 Family size 39 -35 -59 -94 -138 -109 -74 -208 Male education 17 -7 34 3 35 2 -169 -12 Female education 15 -6 67 3 4 -3 -40 31 Demographic composition -58 -7 -58 -24 -18 -26 289 36 Housing 4 2 -51 -13 22 1 -30 14 Infrastructure -8 -9 -20 3 -40 9 70 31 Constant -34 109 155 195 Total -36 -64 25 -125 23 -123 217 -118 Demographic composition breakdown # children e-4 years 4 -3 25 -13 -10 -14 -35 -30 # children 5-10 2 2 15 -5 6 -4 7 3 #males 11-14 -2 -1 19 3 15 6 2 -1 #females 11-14 -2 0 0 -1 4 8 -2 0 # males 15-19 3 1 -18 -10 -9 -7 51 3 # females 15-19 -6 0 -2 0 -4 -4 6 1 # males 20-34 -14 0 -31 -9 -13 -10 66 23 # females 20-34 -16 -1 -2 13 20 0 97 9 # males 35-59 -18 -I -4 -2 -32 -2 54 6 # females 35-59 -15 -4 -56 -2 13 -1 23 9 # males 60+ 4 0 4 0 -11 0 29 13 # females 60+ 1 0 -8 2 4 1 -9 -I Total -58 -7 -58 -24 -18 -26 289 36 Annex 23, Page76 Table A23.63 - Percentage distribution of technical assistance recipients, by degree of poverty, farm household typology, and consumption quartiles. Technical Assistance Non-Recipients overall share Recipients in sample Degree of Poverty Non-Poor 42.7 33.1 34.48 Moderately Poor 44.2 45.3 45.12 Extremely Poor 13.1 21.6 20.4 Farm Household Typology Owner households Minifundia 15.1 13.8 14 Small 18.5 13.9 15 Medium 20.9 15.8 16 Large 17.0 17.7 14 Renter households Minifundia 17.0 26.6 28 Other 11.7 12.1 13 Consumption Quartiles* Poorest 10% 17.0 26.3 Quartile 2 25.7 24.9 Quartile 3 27.2 24.6 Quartile 4 30.1 24.2 Note: households who do not own or operate land are excluded. *quartiles formed from the distribution of household consumption per capita for all households. Annex 23, Page78 Table A23.65 - Impact of technical assistance on probability of using chemical inputs. Matching over entire sample Degree of Share of recipients Share of Share of matched Test of Poverty who use input non-recipients non-recipients diff. who use input who use input Fertilizer Non-Poor 0.61 0.41 0.36 Moderately Poor 0.42 0.29 0.25 ** Extremely Poor 0.30 0.18 0.30 -- Overall sample 0.48 0.33 0.30 Pesticide Non-Poor 0.68 0.51 0.46 Moderatel) Poor 0.57 0.47 0.37 Extremely Poor 0.59 0.34 0.37 * Overall sample 0.62 0.45 0.41 Note The 'matched non-recipient' was classified as a fertilizer or pesticide user if the majority of the closest 5 matches used fertilizer or pesticides respecitively. **. **:*. significant difference between recipients and matched non-recipients at 1%. 5%, and 10%. Table A23.66 - Impact of technical assistance on probability of using chemical inputs. Matching by farm household typology Farm household Share of recipients Share of Share of matched Test of typology who use input non-recipients non-recipients diff who use input who use input Fertilizer Owner households Minifundia and small 0.53 0.38 0.39 Medium 0.69 0.33 0.49 ** Large 0.38 0.17 0.29 -- Renter households Minifundia and other 0.34 0.27 0.38 -- Pesticide Owner households Minifundia and small 0.65 0.45 0.49 * Medium 0.64 0.51 0.50 -- Large 0.56 0.42 0.32 * Renter households Minifundia and other 0.61 0.45 0.39 ** Note: The 'matched non-recipient' was classified as a fertilizer or pesticide user if the majority of the closest 5 matches used fertilizer or pesticides respecitively. ***; **:*: significant difference between recipients and matched non-recipients at 1%, 5%, and 10%. how to read the cells: e.g., among the recipients who are large land owners, 48% use fertilizers. for fertilizer, matched non-recipient was classified as a user based on average of 5 matches. gives the same results. Annex 23, Page79 Figure A23.1 - Annual, periennial, and pasture land, by consumption deciles, rural households 25 total land. adjusted 20 j mz s 10 5. pasture - 0 2 - - - - - - eren nial 1 2 3 4 5 6 7 8 9 10 consumption deciles Figure A23.2 - Annual, periennial, and pasture land, by consumption quintiles, landowning households 40 3 5 30- 25- mz 20 s~~~~~~~~~~~~~~~~~ pasture 15 10 5 - .- perennial 0 .: 1 2 3 4 5 consumption quintiles Annex 23. PafRe8 Figure A23.3 - Share of households with female head of Figure A23.4 - Years of education, adults and head of household household 2S 7 20 . .' sha I 5 yea4 aut re - rs adults 2- head 00 _ 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 consumption deciles consumption deciles Figure A23.5 - Dwelling Figure A23.6 - Ownership of Characteristics Consumer Durables 90 45 45 color TV 80 difloo, ' _ electricity 40 60 . > , -3 BIV TV / \ 7 30 ~ ~ ~ .3 sha 0- sha .25 40 * ,Ix20 30- o .10 -,- running water no toilet 05 ... -1 -,~ ehicle 00 00 1 2 3 4 5 6 7 8 9 lE 1 2 3 4 5 6 7 t 9 tO consumption deciles consumption deciles Figure A23.7 - Participation in off farm activities .60 non agricultural wag 50_ ~~~0gricultural w wg .50 e .40\ sha. re 30 .'s .20 .10 self employed .00 1 2 3 4 5 6 7 8 9 10 consumption deciles Annex 23, Page8l Figure A23.8 - Share of Figure A23.9 - Share of farm households landowning households buying or selling land buying and selling land .20 .25 .18 .16 .20 .20 bought 5h I2.14 bought mz .5sha 12 mz.15 / ~~~~~~~~~~~~re .10 ------- ------ .10 r .08 .05 . - . 8old - ~~~~~~~~~~~~~.04 .00 . -. .02 1 2 3 4 5 .00 consumption quintiles 1 2 3 4 5 6 7 i 9 10 consumption deciles Figure A23.10 - Land title status of landowning households .80 .70 registered .60 .50 . . - ' has escritura sha ........... re 40 no document .30 - - - - . .20 .10 at .00 1 2 3 4 5 consumption quintiles Annex 23. Page82 Figure A23.11 - Share or farm households owning ___ and renthin Figure A23.12 - Share of landowning households renting in and out land .80 9 70 landed rent out .60 -.15 sha 0 mz re 40 1 rents in .30 .20 .05 .10 .00 - .00 - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 2 3 4 5 6 7 8 9 10 1 2 3 4 5 consumption quintiles consumption deciles Fi-ure A23. 13 - Share of FEarm Figure A23.14 - Share of Farm Households F, Planting Fruits or Vegetable 1.00 anv CTOP .80- 900- re ~ 70 - *fruits , * 800 - - > - . , ~~~corn .60 ,._ 70 .- sha .60 '.50 re .50 re 40- .40 .30 .30 i beans .20 vegetables .20 I .00 0 ,.- 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 consumption deciles consumption deciles Annex 23, Page83 Figure A23.15 - Share of farm households using family labh4irgficulture 1.00 .90 famil .80 y .70 sh .60 ar .50 _ .40 .30 ___ .20 - - -- - hire .10 d .00 1 2 3 4 5 6 7 8 9 10 consumption deciles __ _ Figure A23.16 - Share of farm households using or with access to technical assistance .25 - .20- use sh15~~ ... sh .1 5 ar .10- * * use of government .05 supplied \/ - _--- use ofNGO supplied .00 1 2 3 4 5 6 7 8 9 10 consumption deciles Annex 23, Page84 Figure A23.17 - Share of households with credit .30 .25 anv credit .20 loan sha re .15 .10 .05 credit .00 1 2 3 4 5 6 7 8 9 10 consumption deciles Figure A23.18 - Share of households with members participating in organizations .40 non producer organization .35 .30 .25 . -. - . religious sha religious re .20 .15 .10 producer organization .05 . * - * - .00 1 2 3 4 5 6 7 8 9 10 consumption deciles Annex 23, Page85 Figure A23.19 - Cattle ownership of farm households .60 25 .50 20 .40 15 sha he re .30 share ads 10 .20 5 .10 .00 0 1 2 3 4 5 6 7 8 9 10 consumption deciles Figure A23.20 - Pig and fowl stocks of farm households 18 16 - 14 - 12 fowl / 1 0- 8 6 4. pigs 2 .; .. .. . .... -.-.-........... . 1 2 3 4 5 6 7 8 9 10 consumption deciles Annex 23, Page86 Bibliography Banco Central de Nicaragua, 1999. Indicadores Econon2icas, Webpage, October. Corral, L. and Reardon, T. 1999. "Rural nonfarm and farm incomes in Nicaragua: evidence from the 1998 Living Standards Measurement Survey," preliminary version of a background paper prepared for the World Bank Nicaragua Agricultural Sector Study, October. Davis, B., Carletto, C. and Sil, J. 1997. Los Hogares Agropecuarios en Nicaragua: Un Ancilisis de Tipologia. Food and Agriculture Organization of the United Nations and the University of California, Berkeley, November. Davis, B., Carletto, C. and Piccioni, N. 2000. "Income Generation Strategies among Nicaraguan Agricultural Producers," forthcoming in Zoomer, A. and G. Van Der Haar, eds., Land and sustainable livelihood in Latin America, (Amsterdam: CEDLA/KIT). Davis, B., Handa, S., and Soto. H. 1999. "Crisis, Poverty, and Long-term Development: Examining the Mexican Case," IFPRI and Progresa, December. Deaton, A. 1997. The Analysis of Household Surveys. A Microeconomic Approach to Development Policy, Baltimore: The John Hopkins University Press. Deininger, K. and Binswanger, H. 1999. "The evolution of the World Bank's land policy." in de Janvry, A., Gordillo, G., Platteau, J. and Sadoulet, E. (eds). Access to land, ruralpoverty, and public action. Oxford University Press. Deininger, K. Lavadenz, I., and Zegarra, E. 1999. "Rural land markets in Nicaragua," preliminary version of a background paper prepared for the World Bank Nicaragua Agricultural Sector Study, October. Dinar, A. and Keynan, G. 1998. "The Cost and Performance of Paid Agrtcultural Extension Services: The Case of Agricultural Technology Transfer in Nicaragua." Policy Research Working Paper, No. 1931. de Janvry, A. and Sadoulet, E. 1997. "Agrarian Heterogeneity and Precision Policies," Paper for presentation in the Latin American Seminar on Agrarian Heterogeneity and Differentiated Policies, Mexico, November 27-29. Heckman, J., Ichimura, H. and Todd, P. 1997. "Matching as an Econometric Evaluation Estimator: Evidence from Evaluating a Job Training Programme." Review of Economic Studies, 64: 605-654. Heckman, J., Ichimura, H. and Todd, P. 1998. "Matching as an Econometric Evaluation Estimator." Review of Economic Studies, 65: 261-294. Heckman, J., Ichimura, H., Smith, J, and Todd, P. 1998. "Characterizing Selection Bias using Experimental Data." Econometrica, 66(5): 1017-98. IFAD. 1999. "Programa nacional de tecnologia y formaci6n agropecuaria: Fondo de Asistencia Tecnica (FAT). Misi6n De Evaluaci6n Ex-Ante. Volumen 1: Texto Principal." Divisi6n de America Latina y el Caribe, Departamento de Administraci6n de Proyectos, November. Annex 23. Page87 llahi, N. 1999. "An Analysis of Labor Markets and Time use in Nicaragua," preliminary versioi of a background paper prepared for the World Bank Nicaragua Poverty Assessment, November. Keynan, G., Olin, M. and Dinar, A. 1997. "Cofinanced Public Extension in Nicaragua." World Bank Research Observer, 12(2): 225-47. MAG. 1998. Marco de Politicas y Acciones para el Ciclo Agricola 1998-1999. Managua. Maldidier. C. and Marchetti, P. 1996. El Campesino-Finquero y el potencial economico del campesinado nicaraguense. (Managua: Nitlapan-UCA). Moffitt, R. 1991. "Program Evaluation with Non-Experimental Data." Evaluation Review, 15(.3): 291-314. Morduch, J. 1992. "Risk, production, and saving: theory and evidence from Indian households.' Working Paper. Harvard University. N itlapan-UCA, 1995. Diagn6stico de la producci6n agropecuario. Analisis de encuesta rural 1995. Informe preparado para el Proyecto de Tecnologia Agraria y Ordenamiento de la Propiedad Agraria, Managua. Oaxaca, R. 1973. "Male-female wage differentials in urban labor markets." International Economic Review, 14:693-709. Renzi, M. and Agurto, S. 1996. La Mujery los Hogares Rurales Nicaraguenses (Managua: FIDEG). Rose, H. and Neira, 0. 1999. "Desempenjo del sector agropecuario y politica de incentivos: elementos para una politica de incentivos sectorial," preliminary version of a background paper prepared for the World Bank Nicaragua Agricultural Sector Study, October. Ruben, R., Rodriquez, L. and Cortez, 0. 1999. "Land reform, rural organization, and agrarian incomes in Nicaragua" paper for presentation at the Workshop Land in Latin America: New Context, New Claims, and New Concepts, Amsterdam, Holland, May 26-27. Sobrado, C. 1999. "Indicaciones para el uso del agregado del consumo de la Encuesta de Mediciones de Vida Nicaragua 98," mimeo, The World Bank. Spoor, M. 1995. The State and Domestic Agricultural Markets in Nicaragua From Interventionism to Neo-Liberalism (New York: St. Martin's Press). StataCorp, 1999. Stata Statistical Software: Release 6.0 (College Station, TX: Stata Corporation). Steiner, M. 1995. "Area Frame Sampling Methodology for the 1994/95 Nicaraguan Agricultural Survey," mimeo, agosto. Valdes, A. and Bastos, G. 1999. "Reflexiones sobre el sector agricola de Nicaragua con enfasis en la estructura de incentivos," preliminary version of a background paper prepared for the World Bank Nicaragua Agricultural Sector Study, October. Annex 23, Page88 World Bank, 1995. Republic of Nicaragua Poverty Assessment. Volumes I and II. Washington, D.C. Annex 24. Page I Annex 24 - Distributional Effects of Agricultural Incentives Policies in Nicaragua By Diana Kruger 1. INTRODUCTION 1. The role of agriculture in the Nicaraguan economy has been growing in the last decade, contrary to the trend in other countries of the Central American region (Valdes, 1999a). The primary sector' represented 29 percent of the GDP in Nicaragua in 1998, compared to 25 percent in 1993 2 The total participation of agriculture in the GDP, which includes productive activities in forestry, agro-industry, and complementary commercial and financial services, has been estimated at 40 percent (MAGFOR 2000, p.6). Thus, it is not surprising that in a recent study (Valdes, l999a), the author stipulates that Nicaragua's medium-term growth potential and poverty alleviation policies are based on the revitalization of agriculture. 2. Pricing policies for agricultural products have a sizeable impact on Nicaragua's population because people's lives are related to agricultural activities either directly or indirectly, both as producers and consumers of agricultural goods. Specifically, agricultural pricing policies will have a direct impact on the poor: 70 percent of the rural population was poor in Nicaragua in 1998, and they are probably engaged in agriculture as both producers and consumers.' Furthermore, a greater share of total expenditures by the poor relative to the non-poor is for food: almost 63 percent of spending by the poor is for food, compared to 45 percent for the non-poor. 3. The objective of this study will be to measure the impact of agricultural tariff reforms on the incomes of consumers and agricultural producers in Nicaragua. Six major agricultural products will be covered: coffee, sugar, meat, rice, red beans, and corn. On average, 62 percent of family food expenditures are for these six goods. The analysis will quantify the impact of a free trade reform scenario on household income across poverty groups.4 Answers will looked for to the following questions: What effects do pricing policies have on domestic production and consumption of agricultural goods? What are the welfare implications of the current incentive framework on the poor-that is, who benefits and who loses from these policies? 4. The rest of the paper is organized as follows: Section 2 contains a brief background of the evolution of the agriculture sector during the 1990s; Section 3 discusses agriculture's current incentive framework: prices, commercial policy (tariffs), fiscal incentives, and the current exchange rate policy; Section 4 summarizes nominal and effective protection rates in Nicaragua. The core of the paper is found in Section 5, which reports the results of the exercise of measuring welfare implications of incentive framework reform. The final section contains conclusions and lessons. ii. AGRICULTURE IN NICARAGUA, 1990-1998 5. Since 1990, Nicaragua has been undergoing a transition from a highly planned economy, to one that responds to market forces. The Stabilization Plan of March 1991 succeeded in stabilizing hyperinflation and stopping massive devaluations, through the exercise of discipline in monetary ' The primary sector includes agriculture, livestock and fishing activities. 2 Agriculture alone represented 19 percent of GDP in 1998, compared to 14 percent in 1993. 3 The 1998 LSMS data indicates that 70% of rural households live in poverty. Furthermore, Davis and Murgai's (1999) analysis of the corn market indicates that 6, 20, and 29 percent of rural, landed households buy and sell, are net buyers, and are net sellers of corn, respectively. Similar data for the bean market is 13, 28 and 27 percent of homes buying and selling, being net buyers, and being net sellers, respectively. 4 Results by region were not possible because the sample did not contain sufficient observations. Annex 24. Page 2 and fiscal policies. The Government proceeded to undertake major structural transformations that included privatization of former state enterprises, deregulation in domestic and international trade, and creating institutional avenues to resolve property rights conflicts, among others. 6. The agriculture sector was not exempt from reform. During the 1980s, the sector was characterized by high State participation in the different stages of agricultural production, including: planned production, state monopolies in domestic and foreign commerce, strict regulation of the use of foreign exchange earned from exports, direct intervention in wholesale and retail distribution activities, and direct price interventions. All of these created negative incentives to production, as well as severe distortions and inefficiency that led to the growth of the informal sector and a decline in production.5 7. In the early 1990s, reforms in the agricultural sector followed the country's policies of de- regulation and price liberalization with the objective of achieving efficient production. Speciiic reforms in the agriculture sector have included the privatization of domestic and foreign commerce, as well as the liberalization of agricultural prices6 These have led to increased participation of the private sector in agricultural production and to the introduction of Nicaragua's products in international markets. The credit market was gradually liberalized. Private banks began to operate in 1991 and provided loans to the agriculture sector on a competitive basis. Through the early and mid- 1990s, the largest State-run bank, BANADES. remained the most important provider of agricultural credit, but its operations were phased out gradually until it was shut down in 19977 The private credit market has slowly been replacing the role that BANADES used to play. 8. The remainder of this section will briefly describe the major issues and developments of the agricultural sector through the past decade, including evolution, export performance, and technology and productivity constraints. This section should provide a broad context in which the welfare exercise can be assessed. 2.1 Evolution of tIhe Primary Sector 9. The comprehensive economic reforms undertaken in the last decade resulted in a turn-around from stagnation to positive GDP growth rates in 1994. After negative or almost no growth in 199 1 - 93, Nicaragua's GDP maintained an average annual growth rate of 4.5 percent for the period 1994- 98. In a similar pattern, agricultural GDP (ADP) exhibited negative growth rates through 1993; then, between 1994 and 1998, ADP grew at an average annual growth rate of 8.8 percent.! 10. The main source of total economic growth between 1990-98 was primary sector growlth (farming and ranching). The sources of growth vary within that period: between 1990-1994, while farming remained stagnant, the ranching sector grew by a total of 17 percent (at an average annual growth rate of 4.0 percent). The factors that facilitated this expansion were the privatization of cati le processing plants and the liberalization of the meat market, as well as high protection rates on meat imports.9 The high growth of the ranching sector can be interpreted as a "recovery" from the MEDE (1996). 6 A price-band mechanism was in place (for certain imported agricultural products) between April 1992 and June 1997, where the price floors and ceilings were based on international prices. The objective was to give temporary protection to domestic producers and consumers, to stimulate efficient and competitive domestic production (MAGFOR (1999a), p.6). 7 For a full discussion, see Sanchez 1999. 8 Growth rates are estimated from: BCN, "Indicadores Econ6micos", November 1999, web page. They refer to the geometric average rates of growth of GDP in constant 1980 c6rdobas. 9 Valdes (1999a), p. 3. See Table 6 in this paper. Annex 24. Page 3 stagnant, previous decade-that is, the high rates of growth are the result of very low initial levels of production. The high rate of growth is explained by the level of stagnation of the sector during the years 1985-89, a result of the armed conflict and the international trade embargo against Nicaragua (MAGFOR and World Bank 1999). 11. The reverse phenomenon occurred between 1994 and 1998: while ranching remained stagnant, farming experienced high rates of growth. During these years, the sector grew by almost 40 percent at an average annual rate of 8.8 percent. This was due mainly to high growth rates for agricultural export products-a resulted of comprehensive export promotion measures -and the increase in production for domestic consumption. The share of agriculture in the total GDP increased from 16 to 19 percent between 1994 and 1998. 12. The period 1993-98 saw a large expansion in the area under basic grains, with very small increases in yield, while the expansion in the area for export crops was modest (sometimes decreasing), with high growth in production yields. The total harvested area grew 33 percent between 1990 and 1998. Table A24.1 - Agricultural Production Growth 1993-98 (Output, Area, and Yield, % change) Area Area Output Harvested Yield Output Harvested Yield Traditional Exports Basic Grains & Others Sesame -74.0 -54.8 -42.5 Rice 21.6 25.2 -2.8 Cotton -100.0 -100.0 -100.0 Beans 49.0 56.2 -4.6 Coffee 52.6 19.8 27.4 Corn 7.0 20.9 -11.5 Sugar Cane 99.7 40.1 42.6 Sorghum -48.1 -22.0 -33.5 Tobacco -2.3 89.9 -48.5 Peanuts 86.7 124.2 -16.7 Banana 61.1 15.5 39.5 Sovbean 199.2 187.9 3.9 Source: BCN, "Indicadores Econ6micos` web-page: www.bcn.gob.ni 2.2 Agricultural Exports 13. Nicaragua's economy is very open: total trade represented 93 percent of total GDP in 1998. The following table indicates that the agriculture sector is also highly tradable, but to a lesser degree than other countries in the region. Both imports and exports of goods and services measure openness to trade. Table A24.2 - Degree of Openness of the Nicaraguan Economy, 1990 and 1996 Volume of total trade over GDP Volume of agricultural trade over (%) agricultural GDP (%) Country 1990 1996 1990 1996 El Salvador 28 55 45 71 Costa Rica 57 81 82 130 Nicaragua 46 76 81 84 Honduras 65 80 121 99 Mexico 21 42 52 79 Source: extracted from Valdes (1 999a). 14. In Nicaragua, despite high growth rates in agriculture, agricultural exports perforned modestly during the decade. As the following table illustrates, in the period 1990-1999, total agricultural exports grew at an average nominal annual rate of 4.8%, led by the high growth rate of non-traditional agricu!tural products. Traditional agricultural exports, which represented an average of 85% of total agricultural exports, grew at a modest nominal annual rate of 2.5%. Annex 24. Page 4 Table A24.3 - Agricultural Exports, 1990-1999 (Millions of U.S. dollars) Year Traditional Non-Traditional TOTAL 1990 247.6 19.2 266.8 1991 201.5 15.0 216.5 1992 171.5 18.1 189.6 1993 151.1 30.4 181.5 1994 214.0 56.8 270.8 1995 321.6 44.0 365.6 1996 322.9 44.2 367.1 1997 324.1 91.2 415.3 1998 350.0 65.8 415.8 1999 309.6 95.5 405.1 Annual Growth Rate (%) 2.5 19.5 4.8 total Growth Rate 1990-99 (%) 25.0 397.6 51.8 |Source: BCN 15. The modest performance of agricultural exports, despite favorable international prices and different policies for promoting exports, may signal an anti-export bias in Nicaraguan agriculture. Export promotion strategies, especially for the agriculture sector, have vielded important poverty reduction results in other parts of the world and should be considered a viable priority for Nicaragua's future development. 2.3 Agricultural Productivity and Poverty 16. Nicaraguan agriculture has been characterized by slow (or negative) growth in productivity in the recent past. While in almost all countries the use of technology in agriculture has increased significantly over the last 15 years, the trend in Nicaragua has been one of decreasing technological levels in agriculture.10 Three useful indicators that shed light on the evolution of productivity in the agricultural sector in several countries are summarized below (Table 4) for the periods 1979/81 and 1995/97. Table A24.4 - Productivity in Nicaraguan Agriculture (1) Agric. Labor Country Productivity Growth (%) (2) Fertilizer Consumption (3) Agricultural Machinery 1979-81 1995-97 1979/1995 1979-81 1995-97 1979-81 1995-97 Argentina 12.195 13,833 13 46 254 73 112 Australia 20.880 29,044 39 269 376 75 65 Brazil 2,047 3,931 92 915 898 139 142 Chile 2.612 5,211 100 321 1.131 86 119 Colombia 1,926 2.890 50 812 2,853 77 118 CostaRica 3.159 4,627 47 2.650 3.636 210 246 Mexico 1,482 1,690 14 570 538 54 71 New Zealand n.a. n.a. n.a. 1,965 4,247 367 488 Nicaragua 1,334 1,407 6 392 147 19 11 ource: Extracted from Valdes, Alberto (1999b). (l) Agricultural value added per agricultural worker. in 1995 US$. (2) Hundreds of grams per hectare of arable land. (3) Tractors per thousand hectares of arable land. 17. Low levels of education, high levels of poverty, and inadequate basic services (sanitation, safe water, housing structure, etc.) characterize farming families, In rural farm households, adults complete an average of 2.6 years of education compared to 4.2 years in rural non-farm homes (Davis and Murgai 1999). '° Furthernore, the rate of growth of labor productivity in Nicaragua is substantially lower than for other countries. Although it is expected that countries with greater levels of GDP will have higher levels of agricultural value added, the rate of growth of value added is much lower in Nicaragua. Annex 24, Page 5 18. In fact, Valdes (1999b) concludes that stagnant labor productivity in Nicaragua guarantees stagnant incomes for poor households involved in agricultural production. It should not be surprising that the profit margins enjoyed by producers with higher levels of technology are larger, so low levels of technology and education have direct negative implications on the income of poor rural households involved in agricultural production. III. THE INCENTIVE FRAMEWORK 19. In 1997, the Government of Nicaragua designed a comprehensive set of measures (MAGFOR 1997) aimed at stimulating sustainable growth of agricultural production, which comprises the central pivot of the economic reactivation strategy for Nicaragua. The country's official agricultural policy is to increase production by improving sector competitiveness through channels that increase profitability, reduce risks, and increase funds directed to agricultural and livestock production. 20. This section will summarize the existing incentive framework under which agricultural production takes place in Nicaragua." It consists of four broad areas: prices, tax incentives, tariff structures, and exchange rate policy. The goal with current policies is to have them be "pro- agriculture" and remove any biases against the sector. 3.1 Agricultural Prices 21. Market forces determine prices in Nicaragua.'2 Prices for the country's principal agricultural products had a tendency to increase during the 1990's. Table 5 below summarizes the evolution of the prices for the six products being studied. Prices for all of these (except meat) increased between 1992 and 1998. Table A24.5 - Producer Prices and Price Index, 1992-1998 Prices Price Index (1992=100) Year Rice Bean Corn Coffee Sugar Meat Rice Bean Corn Coffee Sugar Meat 1992 62.1 71.5 38.9 56.0 10.1 1.0 100.0 100.0 100.0 100.0 100.0 100.0 1993 68.6 244.6 44.2 54.6 13.0 1.1 110.6 342.3 113.7 97.5 128.7 110.0 1994 73.6 179.9 66.0 91.4 13.0 1.1 118.6 251.7 169.9 163.2 128.7 110.0 1995 80.9 151.5 60.4 149.3 14.1 1.0 130.4 212.0 155.5 266.6 139.6 98.0 1996 98.0 360.9 91.3 109.5 14.6 0.9 157.9 505.0 234.9 195.5 144.6 86.0 1997 106.2 387.8 93.1 141.0 11.2 0.9 171.1 542.7 239.4 251.8 110.9 88.0 1998 113.6 411.7 90.9 146.7 13.4 0.9 183.1 576.1 233.9 262.0 132.7 92.0 Source: Export prices from BCN, Indicadores Economicos. Basic grains: MAGFOR. Prices are farn gate prices, annual average. ** Prices for rice, beans, and com are in C6rdobas/quintal. Coffee and sugar prices are in dollars/quintal; meat prices in dollars/lb. 22. Producers responded to these incentives by increasing output and the area harvested, as indicated in Table 1.'3 3.2 Tax Structure 14 23. Agricultural production is officially subject to income, value-added, fuel, property, and municipal taxes, and if production is for export, fees are paid to an export transaction center (Centro " For a more detailed description of the incentive framework, see MAGFOR (2000). 12 With the exception of the price-band in effect from 1992-1997 for imported products (mentioned in footnote 6). These products are: rice, sorghum and yellow corn. 3 As mentioned earlier, productivity (yields) increased only for exportable goods. 4 Sections 3.2, 3.3, and 3.4 draw heavily from MAGFOR (2000), Part 11. Annex 24. Page 6 de Tramite de Exportaciones, CETREX). Some of these taxes, such as the property tax, seldom have any effect due to weak mechanisms for revenue collection and property registry systems and to tax incentive schemes provided to the sector to promote agricultural production. 24. One of the main objectives of recent economic policy in Nicaragua has been export promotion, which has been supported by fiscal (and other) incentives. An Export Promotion Lavh'5 was passed in 1991. which contemplated the following fiscal benefits for exports: * Profits from exports were exonerated from income tax for a determined time period. * Locally-purchased inputs used in export production were exonerated from the national value- added tax. * Exporters had immediate access to the foreign exchange derived from the sale of their products. * Non-traditional exports were awarded a tax benefit certificate (Certificado de Beneficio Tributario, CBT), equivalent to export value on a decreasing scale of 15% to 5% through 1997. The program was gradually phased out until ending in December 1997.16 25. In 1997. new tax legislation" was passed which replaced the Export Promotion Law and continued with the objective of supporting exports. The new tax law (effective since 1998) stipulates that exports are not subject to any trade tariffs, provides all agricultural export producer . a tax credit of 1.5 percent of export value (f.o.b.). and exempts agricultural production services from the 15% value-added tax. 3.3 Trade Policy 26. Consistent with the liberalization program implemented in the early 1990s, Nicaragua's tracle policy has fallen in line with global trends towards free trade, and since the mid-1990s the country has drastically reduced trade tariffs for all products. Furthermore, to promote international commerce, trade policy has included: institutional reforms to remove trade barriers; simplification of the trade process; reinsertion of Nicaragua's products in traditional export markets; and promotion of new markets through regional and bilateral preferential (and free) trade agreements. 27. Tariffs in Nicaragua have two components: the Central American common external tariff'8 and temporary protection tariffs. Imported products from other Central American countries are traded freely with the following exceptions: sugar, coffee, wheat flour, ethyl alcohol, rurn, petroleum ether, and gasoline, among others. The average tariffs to the agriculture sector were 32, 13, 21, and 17 percent during the years 1991-1994, respectively. 28. Agricultural trade policy was reformed in 1997, which consisted in substituting a price-band mechanism for imported rice, sorghum and yellow corn with a fixed tariff for these products as well as revising import tariffs for certain products (reflected in Table 6, below). The 1997 reforms also included reducing high import tariffs to certain "exceptional" products: the tariff on chicken parts was reduced from 250% to 180%, the tariff on beef was reduced from 480% to 40%, and sugar imports, which were formerly restricted by the requirement for an import permit, were subject tc a fixed 55% tariff in 1997. Finally, restrictions to curtail the dumping of certain agricultural products were also contemplated. 15 Official Decree No.37-91. 16 MAGFOR (2000). " Law 257 "Ley de Justicia Tributaria y Comercial", June 4, 1997. 18 Nicaragua is a member of the Central American Common Market. Annex 24. Page 7 Table A24.6 - Nicaragua: Import tariffs for selected products (%) Product 1996 1997 1998 1999 Beef Cattle n.a. n.a. 20 15 Meat (beef) 480 40 40 40 Sugar Import pernit 55 55 55 Soybeans 20 20 0 0 Peanuts 15 15 15 10 Sesame 10 n.a. 5 0 Cotton 10 n.a. 5 0 Corn 5 18 25 15 Sorghum (industrial) 5 15 20 15 Beans 25 25 25 15 Rice 12 24 35 20 Cantaloupe 25 n.a. 20 15 Onion (for export) 30 n.a. 20 15 Coffee 25 n.a. 20 15 Banana 35 n.a. 20 15 Poultry (parts) 250 ISO 180 180 Poultry (whole) n.a. 60 50 40 Source: MAGFOR, Valdes (1999a) 29. The Export Promotion Law effective from 1991-1997 included an exemption from trade tariffs for imported inputs and capital goods used for export production. During that period, temporary trade decrees were authorized which exonerated import tariffs on agro-chemical inputs used in agricultural production. The 1997 reforms expanded these tariff exemptions to all agricultural production, including intermediary and capital goods. 30. Non-tariff trade barriers in Nicaragua included exclusive representation rights, the requirement of Import Licenses (at a cost of US$250.00 each), as well as quantitative restrictions on imports from certain Central American countries.t9 In 1998 new measures were taken to continue the reduction of trade barriers such as: the prohibition of non-tariff restrictions to exports and imports,20 the elimination of all charges on export licenses or permits and taxes on export operations (except service fees which were effective in 1998), and the elimination of exclusive Representative Agent laws, effective July 1998. 3.4 Exchange Rate Policy 31. The government faced two major challenges in the early 90s: to control inflation, and to stimulate growth via exports. The first objective calls for a stable nominal exchange rate to adjust the population's inflationary expectations, while the latter requires sustaining a competitive real exchange rate (relative to other countries) to stimulate the production of export items. 32. To control inflation, in 1991 a stabilization program was implemented that aimed to control inflation through fiscal discipline and that set the nominal exchange rate as an anchor. The nominal exchange rate was fixed at 5 c6rdobas to I U.S. dollar. By 1993, the goal of controlling inflation had been achieved, and the government aimed to achieve its second goal of export stimulation, and '9 Such as: dairy products from Costa Rica and Honduras; eggs from El Salvador and Costa Rica; all basic grains from Costa Rica; coffee from Honduras, Costa Rica, and Guatemala; and instant coffee from El Salvador. 20 Except for sanitary, public health, consumer or enviromnental safety, or national emergency reasons. Annex 24. Page 8 adjusted its exchange rate policy accordingly. The nominal exchange rate was devalued 20% to 6 c6rdobas to I U.S. dollar, and a "crawling-peg" system of daily mini-devaluations was set in place (still in effect today) in an attempt to make the real exchange rate more competitive. 33. The sustained stabilization and devaluation efforts have had a positive effect on economic growth, and specifically on exports. Furthermore, the rate of devaluation has consistently been greater than the rate of inflation, which means that the real exchange rate has been devalued since 199322 The behavior of the real exchange rate is summarized in Table 7, below. Table A24.7 - Real Exchange Rate Index (1994=100) Year Bilateral with U.S.A. Multilateral 1990 106.36 104.81 1991 91.54 90.57 1992 90.95 90.53 1993 95.95 95.95 1994 100.00 100.00 1995 103.81 108.28 1996 107.24 109.1] 1997 112.53 115.64 1998 113.30 117.55 1999 115.98 120.34 Source: BCN An increase Tepresents a devaluation in the real exchange rate. 34. Despite the success of the exchange rate policy in devaluing the real exchange rate, Valdes (1 999b) points out that a devaluation of the exchange rate does not guarantee there is 110 misalignment of the exchange rate. In fact, there are signs of a possible imbalance in the country's external balance, including a trade deficit equivalent to an average 20% of GDP since 1993, a current account deficit of an average 29% of GDP, periods when international reserves cover less than two months of imports, and a high level of foreign debt. If a country's exchange rate is misaligned (over-valued), then the foreign exchange obtained from exports buys less domestic currency, making the export sector less profitable, thereby exhibiting an anti-export bias. 35. An alternative interpretation is given by the BCN in a quarterly publication: the high trade deficit is not necessarily linked to an overvalued exchange rate-it is explained by high private and official capital flows (BCN, 1999). Furthermore, the article points out that to obtain future gains in competitiveness, it is not enough to rely on exchange rate policy, which affects only a nominal variable: the nominal exchange rate. To increase the export sector's competitiveness, additional measures affecting real variables are required, such as: preventing wage adjustments (increases) that are unrelated to productivity gains; increasing public investment in ports, transportation, and communications infrastructure; eliminating the energy sector monopoly; reducing distortions in cost structures; strengthening the judicial framework to promote export production and foreign investment; and providing management and marketing training to those who work in the export sector (BCN 1999, p.8). 21 The real exchange rate is defined as: (Price of Tradables x Nominal Exchange Rate)/Price of Non- Tradables 22 The change in the real exchange rate = foreign inflation (z0) + change in nominal exchange rate - domestic inflation. If this sum is positive, the real exchange rate has been devalued. This occurs when the change in nominal exchange rate is greater than domestic inflation. Annex 24. Page 9 36. In essence, the BCN argument is that the exchange rate is not the "culprit" of the existing anti-export bias in Nicaragua, rather, it lies in structural problems in the sector that must be addressed. IV. NOMINAL AND EFFECTIVE PROTECTION RATES23 37. Despite government efforts to stimulate exports and move toward a free market economy- specifically to more open trade policies-there are both market price interventions and government subsidies for agricultural goods in Nicaragua.:4 Nominal rates of protection (NRP) and effective rates of protection (ERP) can measure the effect that these interventions have on the market for agricultural products. 38. The NRP measures the relative difference between observed domestic prices and the world market price at the same trading point, while the ERP measures the ratio of value added to a product at domestic prices to the value added had it received world prices (Valdes and Kray, 1999). A positive NRP means that producers receive higher prices than without interventions, while a negative NRP means there is price discrimination against producers. A similar interpretation applies to the effective rate of protection: a positive ERP means that value added with intervention is greater than at free market prices, and negative ERP indicates a lower value added with intervention than without (Valdes and Kray, 1999). 39. A MAGFOR study estimated these protection rates for several agricultural products, using both the official and parallel exchange rates. This study will analyze the redistribution effects using only the protection rates calculated with the official exchange rates. The main results for Nicaragua are summarized in Table 8 below. Table A24.8 - Nominal and Effective Rates of Protection and Prices, 1998 Official Exchange Rate Parallel Exchange Rate Product NRP ERP NRP ERP Sector Total 9 13 -4 1 Export items -1 -1 -13 -12 Coffee -2 -1 -13 -12 Sugar Cane -3 1 -15 -13 Meat 1 0 -10 -I] Import items 24 35 12 23 Corn 38 56 22 37 Rice 29 36 14 19 Beans 13 17 -1 4 Source: Estimated from MAGFOR (2000), which reported Nominal and Effective Protection Coefficients. 40. The resulting NRPs and ERPs for Nicaragua indicate that the incentive framework for the agriculture sector has an anti-export bias: import-competing goods are strongly protected, while export goods are slightly taxed.25 23 A through discussion of the methodology for estimating nominal and effective exchange rates is found in MAGFOR (2000). 24 Valdes and Kray ( 1999). Market price interventions refer to price-related supports such as import tariffs, export subsidies, and quantitative restrictions, while government subsidies include subsidies for capital goods, credit, direct transfers, and others that do not directly affect prices. 25 The anti-export bias is stronger if parallel exchange rates are considered. Annex 24, Page 10 V. THE DISTRIBUTIONAL EFFECTS OF AGRICULTURAL INCENTIVE POLICIES 41. According to international trade theory, the elimination of trade tariffs will generate net gains for the economy as a whole due to efficiency gains associated with the elimination of distortions in incentives to production and consumption.26 Partial equilibrium analysis predicts that the elimination of tariffs will have negative effects on some groups of the population and positive effects on others. In the case of eradicating import tariffs, consumers stand to gain while producers will lose.27 However, "free trade is the optimal policy for a competitive, small economy" (Bhagwati, Panagariya, and Srinivasan, 1998). 42. As described earlier, in 1997 tariffs were raised on basic grains in an effort to stimulate their production, done mostly by small farmers. Thus, this increase in tariffs was an effort to target sector policies and programs on assisting the most vulnerable rural groups-small farmers, who are usually poor. This paper will examine the effects of eliminating all import and export barriers, i.., setting net and effective rates of protection equal to zero, or free trade. The estimations are detailed in Appendix 2; the tables in the sections below contain a summary of the main findings. 43. If theory proves correct, in Nicaragua consumers stand to gain from eliminating import tariifs on those goods that are highly protected-basic grains-because prices would be lower with free trade. On the other hand, producers of basic grains stand to lose once the distortion is removed because of the lower prices received and thus, lower income. Since the magnitude of discrimination against exports is relatively small, the effect of eliminating this distortion should not have a large impact on consumers (the share of spending by consumers on export goods is half what they spend on basic grains). Producers of export items, on the other hand, stand to gain from the elimination of "'negative" protection rates. 44. As Table 9 reveals, the share of expenditure on the six items we are analyzing varies across poverty groups: on average, expenditures on these six products represent 85% of extremely poor families' food expenditures, compared to 52% for non-poor families. Table A24.9 - Food expenditures on the agricultural items being analyzed (as a proportion of total food expenditures) Poverty Group Urban Rural Total Extreme Poor Sub-total: grains 0.59 0.62 0.61 Sub-total: exports 0.23 0.24 0.24 Total 0.82 0.86 0.85 4ll-poor Sub-total: grains 0.45 0.55 0.52 Sub-total: exports 0.22 0.22 0.22 Total 0.68 0.77 0.74 Non-Poor Sub-total: grains 0.27 0.39 0.30 Sub-total: exports 0.22 0.21 0.21 Total 0.48 0.60 0.52 4ll Sub-total: grains 0.32 0.49 0.40 Sub-total: exports 0.22 0.22 0.22 Total 0.54 0.71 0.62 Source: LSMS 1998. Grains: rice, corn and beans. Exports: coffee, sugar and meat. 26 Yrarrazaval, Lindert and Wiens, 1998. 27 The opposite welfare effects occur from eliminating export tariffs. See, for instance, Mankiw (1997), p. 180. Annex 24. Page 11 45. Given these shares and the high degree of protection of basic grains. poor consumers-both extreme poor and all poor-stand to gain the most from free trade relative to non-poor. Likewise, poor producers will be the larger losers relatively. The net welfare gains will of course depend on which of these two effects is larger. 5.1 Metlhodology 46. The methodoiogy for estimating welfare effects from trade reform (from Schiff and Valdes,1992) is described in detail in Appendix 1; in this section, we have summarized the distribution effects from reform. 47. The effects of changes in agricultural incentives will be estimated for two separate populations, producers and consumers, and across poverty groups. The first step is to classify households by poverty level. This specific study will follow the poverty classification used in the Nicaragua Poverty Assessment. Households are also classified as either net consumers or net producers of agricultural items. In general. urban households likely will be net consumers while rural households involved in agriculture likely will be net producers. This will be defined using the 1998 LSMS data. 48. The impact of tariff reforms on household income will be measured differently for net producers and consumers. For producers, two indicators of net income are measured based on the quantity28 of a product sold in the market at the two alternative prices: current domestic price and world market price (without intervention). (a) Actual real income: y = Y/PI where Y is nominal income (defined in Appendix) and PI is a price index of the farmer's CPI at prevailing prices (b) Real income in the absence of direct price intervention: y' = Y/PI' where Y' is nominal income (defined in Appendix) and PI' is a price index of the farmer's CPI at prices that would prevail without intervention The direct effect of protection on the real income of producers of agricultural goods is: Ay = [y - y'] / y' 49. For consumers, the income effect of agricultural tariff changes will be measured by the implied changes in food prices in a reform scenario, and thus by the changes in food expenditures. If tariff changes increase food prices, then other things being equal, consumers will spend more on food and their income will decrease. Since the poor spend a greater proportion of their income on food, rising food prices will have a greater impact on their income. 50. After classifying households into a poverty category, a consumer price index will be estimated for each group to measure the effect of changing food prices on the different households. The calculation of the CPIs is detailed in the appendix; the CPI for each poverty group is denoted by the subscript "i". CPIs will also be constructed under alternative food prices after the tariff reform. 51. To estimate the income effect on consumers from the trade reform, the exercise will consist in comparing real incomes under the different tariff scenarios. Letting NY = nominal income, we estimate real incomes: (a) Actual real income in 1998: Y = NY / CPI (b) Real income under tariff scenario: Y '= NY / CPI' 28 As mentioned in the Appendix, the study assumes that supply elasticity is equal to zero, so despite the change in price, producers continue to produce the same amount. This is a realistic response in the short run. Annex 24, Page 12 52. If the effect of tariff reform was to increase food prices, then (b) will be less than (a) and consumers are worse off. If tariff reform decreases food prices, the opposite is true. 5.2 Tlhe Data 53. The data for the analysis is from the 1998 LSMS. Consumption estimates were calculated from the expenditure section of the survey, which provided details of the patterns of consumption by poverty group. From this information, it was possible to construct the shares of total food expenditure that were spent on the different agricultural products analyzed in this study by poverty group and by location (rural/urban). These shares served as weights in the estimation of the different consumer price index constructed for each poverty group, which allowed the comparison of household income under the alternative trade reform scenarios. 54. The 1998 LSMS survey contained a section on agriculture that provides information on tile items and quantities produced by (mainly rural) households. The data allowed for an estimation of total quantities produced by farmers, as well as the quantity that was actually sold in the market (total production less self-consumption). The marketed output is what we needed for the estimations in order to see what effect a change in prices/value added received by producers would have on their income. 55. The sampling framework for the LSMS was designed to be representative of Nicaragua's population at the national and area (rural/urban) levels, as well as for seven geographical regior-s. Sampling was not designed to be representative of agricultural producers or of agricultural production, which means that the results of agricultural production from this study are strictly sample results. It is not possible to draw inferences from these sample results and generalize them to production at the national level, nor to redistribution effects on the entire population. 5.3 Effects on Producers 56. As Table 8 revealed, agricultural producers in Nicaragua enjoy high rates of protection for import-competing goods (basic grains). Although not reported, MAGFOR (2000) demonstrates that the average rates of protection for all products increased between 1997 and 1998, probably as a result of the trade policy measures implemented in 1997, which sought to stimulate agricultural production by creating incentives through high import tariffs. The motivation behind the 1997 reform was to support basic grain producers, who are mostly small farmers and whose income 29 depends on the sale of these products. 57. As mentioned above, economic theory predicts that producers of import-competing goods that are highly protected lose from free trade. Appendix 2 contains tables of the effects of free trade on the income of farmers, with details by product. Theory is verified: producers of rice, beans and corn, who enjoy positive protection rates in Nicaragua, lose from free trade (under both measures, NRP and ERP). Producers of export crops with negative rates of protection (coffee and sugar cane) 30 gained from free trade. It is noteworthy that the largest free trade gains were obtained by non-poor producers, the group usually involved in export agriculture. Table 10 below summarizes the total effects by poverty group for estimations of both nominal and effective rates of protection. 29 Most farmers, regardless of size or poverty group, grow basic grains and a sizeable portion their production of beans and corn is used for household consumption (see Appendix 3). 30 Although meat is an export good, meat producers lose from free trade because they receive positive protection under the existing incentive framework. Annex 24. Page 13 Table A24.10 - Producer's Income Effect by Poverty Group Extreme Poor All Poor Non-Poor All Nominal Rate of Protection Percent -2.3% -2.3% 0.2% -0.8% C6rdobas -48.382 -179.757 23.948 -155,809 Effective Rate of Protection Percent -2.4% -2.4% -0.6% -1.4% C6rdobas -50,471 -187,938 -75,896 -263.834 Aggregate Income 2.062.414 7.672.572 11,683,928 19.356,500 58. In absolute terms, poor producers lose and non-poor producers gain from free trade using the NRP measure. The last line of the table includes the aggregate income of farmers involved in the sale of the six agricultural products we chose (total income is greater in rural areas because that is where most agricultural producers are located.) In relative terms, free trade causes losses in income of approximately 2.4 percent to poor farmers and a gain of 0.2% to non-poor producers using the NRP measure. The interpretation for ERP results is straightforward. Poor producers stand to lose more from free trade than non-poor producers do. 59. An alternative interpretation is that the current policy of protecting poor farmers has been effective-i.e., poor farmers would lose if tariffs were eliminated. Non-poor farmers are also protected with the current policy, and to a greater extent than poor producers are3' 5.4 Effects on Consumers 60. The elimination of high rates of protection is bound to reduce food prices, which implies a benefit to consumers. Consumers who purchase a high share of import-competing goods will benefit most: in Nicaragua, 61 percent of total food expenditures by the extreme poor is for basic grains (rice, beans and corn), while for the non-poor it is 30 percent. Therefore, the relatively larger "winners" will be the extreme poor consumers, as Table 11 verifies. Table A24.11 - Consumers' Income Effect by Poverty Group Extreme Poor All Poor Non-Poor Urban Rural All Percent 3.9% 3.1% 1.3% 1.3% 2.7% 1.9% C6rdobas 301.903 943,374 1.245.850 1,159.339 1.087.505 2.399,705 Aggregate Income 7,821.788 30.918,062 94,963.616 85.890.656 39.983.108 125,873.760 Note: These results wvere estimated using Nominal Rates of Protection. because ERPs applv onlv to producers. 61. All consumers in Nicaragua gain from free trade. The last line of Table 11 reports the aggregate income of consumers in each of the poverty groups and by the area where they live. The proportionate gains are more than twice as large for extreme poor consumers relative to non-poor (3.9% versus 1.3%); likewise, rural consumers gain more relative to urban ones. 5.5 Net Welfare Effects 62. Total welfare gains from free trade in the six products analyzed in this paper are positive, for all poverty groups. The total gains are slightly higher when the NRP is used to measure producer effects. ' Experts on the topic have stated that the effective rate of protection is more appropriate. Annex 24, Page 14 Table A24.12 - Total Welfare Effects, by Poverty Group Extreme Poor All Poor Non-Poor All -Nominal Cordobas-- Nominal Rate ofProtection Producers -48.382 -179.757 23.948 -155.809 Consumers 301.903 943.374 1.245,850 2.189.224 Net Effect 253.521 763.617 1.269.798 2,033,415 Effective Rate of Protection Producers -50,471 -187.938 -75,896 -263.834 Consumers (NRP) 301.903 943,374 1.245.850 2,189.224 Net Effect 251,432 755.436 1.169,954 1,925,390 Note. Consumer effects were estimated withz nominal rate of protection in both cases. 63. Due to the different aggregate incomes of both groups-producers and consumers-it is no t possible to obtain a relative measure (percentage) of these gains. 64. Once again, it has to be kept in mind that these results must be interpreted with caution: t3e welfare gain is likely smaller than the magnitude presented here since the loss to producers is probably under-estimated. The LSMS was not designed to be representative of agricultural producers. Table A24.13 - Number of Observations by Poverty Group and Area Urban Rural Total Farm Households: Ext. Poor 24 138 162 Non-Ext. Poor 34 238 272 Non Poor 71 221 292 Total 129 597 726 All Households: Ext. Poor 154 488 642 Non-Ext. Poor 459 711 1.170 Non Poor 1.574 654 2.228 Total 2,187 1,853 4,040 65. As Table 13 illustrates, the sample used to estimate the welfare effects contained a total of 726 homes that were involved in the sale of agricultural products and over 4,000 homes with consumers of the same goods. Despite the limitations because of the data, this study provides a first exercise in estimating whether the current agricultural incentives benefit or harn the population at large. VI. CONCLUSIONS 32 66. The results in this study are consistent with findings in other rural/agriculture sector studies by Davis and Murgai (2000), Valdes (1 999a), and Valdes (1 999b) and they are indicative of the general situation of the agriculture sector in Nicaragua. 32 It is important to emphasize once again that results from this study are strictly sample findings, so that the conclusions drawn here should be interpreted with this limitation in mind. Annex 24. Page 15 (1) Free trade in rice, beans, corn, coffee, sugar, and meat is indeed an optimal policy for Nicaragua. Total welfare is increased with the removal of trade tariffs; some producers lose, yet all consumers gain substantially from the policy change. (2) The policy goal of protecting the small, poor farmer in Nicaragua has been effective i.e., free trade would harm them-but the total cost of this policy is potentially very high. By protecting poor farmers, non-poor ones inevitably benefit-more so than the target group. This study finds that net welfare benefits accrue to the extreme poor from free trade. (3) Free trade is a progressive policy. Consumers gain from removing tariffs, and as a proportion of their income, extremely poor families benefit relatively twice as much as non-poor ones. Although limited, results indicate that gains to the poor from reduced food prices in consumption outweigh their losses in income as producers. (4) The need for high protection rates for basic grains signals that production by that sector is inefficient relative to other countries. This is likely due to low levels of technology and education in Nicaragua's agriculture sector. Gains in competitiveness could be achieved through policies that emphasize increasing both the human and physical capital of small farmers. In fact, MAGFOR stated succinctly, "Innovative efforts in rural education are a necessary component of a rural development strategy, because only then can the capabilities of rural residents of gathering and assimilating information about technology, production, and marketing alternatives be increased, thus helping them find their own, decentralized way to development." (MAGFOR and World Bank 1999). (5) The results demonstrate an anti-export bias in the agriculture sector. The results in MAGFOR (2000) revealed that the existing incentive framework in Nicaragua taxes the export sector. Some studies suggest that this is due to an over-valued exchange rate; an alternative view is that in order to increase the competitiveness of exportable goods, policies that improve export-related infrastructure (roads, transportation, ports, etc.) and strengthen export institutions should be implemented. Free trade brings gains to the producers of export goods, who are usually non-poor. 67. As pointed out earlier, strategies that promote agricultural exports have yielded important poverty reduction results in other parts of the world. Agriculture sector exports should be an important priority for Nicaragua's future development. Annex 24, Page 16 REFERENCES Bhagwati, Jagdish. Arvind Panagariya, and T.N. Srinivasan (1998): Lectures on International Trade, Second Edition. Cambridge, Massachusetts: The M.I.T. Press. BCN (2000): "Indicadores Econ6micos". Web page: www.bcn.gob.ni BCN (1999): "Reducci6n del Deslizamiento del Tipo de Cambio Oficial."Boletin Econ6mico. June 1999. Managua, Nicaragua. Davis, Benjamin and Rinku Murgai (2000): "Between prosperity and poverty: rural households in Nicaragua," Background Paper for the Nicaragua Poverty Assessment, World Bank, Washington, D.C. Garcia, Magdalena (1999): "Nicaragua, Precios Agricolas: Politicas y Resultados," consultant report to the World Bank Nicaragua Rural Sector Study, World Bank, Washington, D.C. Ilahi, Nadeem (1999): "Labor Markets in Nicaragua," Background Paper for the Nicaragua PovertvAssessment, World Bank, Washington, D.C. MAGFOR (Ministry of Agriculture and Forestry) (1997): "Marco de Politicas y Acciones para el Ciclo Agricola 1997-1998," MAGFOR Publication, Managua, Nicaragua. MAGFOR (1999): "Nicaragua: Politica Comercial hacia Agro," draft. MAGFOR and World Bank (1999): "Desempefio del Sector Agropecuario y Politica de Incentivos: Elementos para una Politica de Incentivos Sectorial," Managua, Nicaragua. MAGFOR (2000): "Tasas de Protecci6n en el Sector Agricola de Nicaragua." Managua, Nicaragua. Mankiw, N. Gregory (1997): Principles of Microeconomics. Fort Worth, Texas: The Dryden Press. MEDE (Ministry of Economy and Development) (1996): Informe sobre la Gestion del MEDE. 1990-1996. Nicaragua. (Note: the Ministry has since been renamed Ministry for Promotion of Industry and Commerce, MIFIC). Sanchez, Susana (1999): "Nicaragua: Financial Markets." Background Paper for the Nicaragua Poverty Assessment, World Bank, Washington, D.C. Schiff,, M. and Alberto Valdes (1992).' The Political Economy of Agricultural Pricing Policy in Developing Countries, Volume IV: A Synthesis of the Economics in Developing Countries, John Hopkins University Press, Baltimore, Maryland. Valdes, Alberto (1999a): "Rural Sector Diagnosis and Performance," Working Paper for the World Bank's Nicaragua Rural Sector Study, World Bank, Washington, D.C. Valdes, Alberto (1999b): "Reflexiones Sobre el Sector Agricola de Nicaragua con Enfasis en la Estructura de Incentivos," Consultant Report, World Bank, Washington, D.C. Valdes, Alberto and Holger A. Kray (1999): "Lithuania: Adjustments of Agricultural and Trade Policies" ECSSD Environmentally and Sociable Development, Working Paper No. 16, World Bank, Washington, D.C. Yrarrazaval, Rafael, Kathy Lindert, and Tom Wiens (May 1998): "Redistributive Impact of Agricultural Trade and Pricing Reforms." Panama Poverty Assessment Working Paper (Draft). World Bank, Washington, D.C. Annex 24. Page 17 APPENDIX A24.1 Methodology of Welfare Implications of Incentive Policies 33 The first step is to identify which households are producers of the agricultural items being analyzed. The impact of tariff reforms on household income will be measured differently for net producers and consumers, as detailed below: PRODUCERS: Assume that farmers produce only one product (the formulas can be adjusted to include more products, but for ease of exposition, the case of one product will be presented here). We assume that the elasticity of supply is equal to zero, thus, output is fixed so that the impact of interventions on nominal income is measured entirely by price changes. I . Define Nominal Income Y = YF + YNF (farm + non-farm income). There are two ways to estimate farm income: (1) final price and (2) value added in production: (l) YF = Pi * Qi s where = Pi is the actual market price of good i Qi s = marketed surplus of the agricultural good i (2) YF = (Pi - Xj a,j*Pj) * Qis where P, = price of input j, and aij is the share of input j in the production of good i 2. There will be two definitions of Real Income when analyzing the income effect of tariff reform: (a) Actual real income: y = Y/PI where PI is a price index of the farmer's CPI at prevailing prices. Specifically, (1.1) PI = -ja, * Pi + (I - aij) * PNA where as is the share of good i produced, Pi is the market price of good i, and PNA is a price index of non-agricultural goods. (b) Real income in the absence of direct price intervention: y' = Y'/PI' where PI' is a price index of the farmer's CP1 at prices that would prevail without intervention (in (1.1) above, Pi would be substituted by P,' 3. Finally, the direct effect of protection on the real income of producers of agricultural goods is: Ay =[y-y'] /y' where y' = [YF' + YNF] / PI' and ' refers to the price that would prevail without intervention. CONSUMERS: Assuming that consumers do not change their demand for food items (a short-run scenario), the income effect of agricultural tariff changes will be measured by the implied changes in food prices, and thus changes in food expenditures. If tariff changes increase food prices, then other things being equal, consumers will spend more on food and their income will decrease. Since the poor spend a 33 From: Schiff and Valdes (1992). Annex 24, Page 18 greater proportion of their income on food, rising food prices will have a greater impact on their income. In the case of Nicaragua, there is detailed data about the share of expenditures of tile individual food items under study in total food expenditures by poverty group. After classifying households into a poverty category, a consumer price index will be estimated for each group to measure the effect of changing food prices on the different households. The calculation of the CPls is detailed below. 1. Consumer Price Index: First, it will be necessarv to construct a CPli , where i= Extreme poor, overall poor, non-poor. Specifically, the CPI will be constructed as follows: CPI = D * PA + (l]-)*PNA where D = share of the agricultural product in total consumption expenditure, measured at retail prices in 1998, PA= I:jPj*PAj = index of the various agricultural products under study that enter the CPI. To estimate the income effect. CPIs will also be constructed under the alternative food prices af;er the tariff reform. Let these be P': 2. Income Effect: Nominal income is available from the LSMS data. The exercise will consist in comparing real incomes under the different tariff scenarios. Letting NY = nominal income, we estimate real incomes: (a) Actual real income in 1998: Y =NY / CPI (b) Real income under tariff scenario: Y = NY / CPI If the effect of tariff reform were to increase food prices, then (b) would be less than (a) and consumers are worse off; if tariff reform decreases food prices, the opposite is true. Annex 24, Page 19 APPENDIX A24.2 Redistribution Effects - Producers Extreme Poor Producers Sugar Net Index Unit Rice Beans Corn Coffee Cane Meat (lbs.) Gain/Loss Border equivalent price (a) C$ / metric ton 2,420.8 6,785.6 1,249.8 506.4 184.3 134.9 NRP (b) Percent 29.4 13.0 37.6 -1.8 -3.1 1.0 Domestic price (c) CS /metric ton 3,132.4 7,654.9 1,719.6 497.4 178.6 135.7 Quantity Sold Hundredweight 119.0 440.0 1,348.0 189.0 0.0 10,204.0 Revenue Cordobas 16,912.7 152,821.0 105,175.5 94,015.8 0.0 25,176.1 Revenue-Border eq. Price Cordobas 13,070.5 135,466.3 76,442.3 95,712.3 0.0 25,027.6 Gain/loss-free trade Cordobas -3,842.2 -17,354.8 -28,733.2 1,696.5 0.0 -148.4 -48,382.2 Value Added -Domestic price C<:rdobas 14,636.6 123.394.6 84,690.2 84,605.0 0.0 22,219.4 ERP (b) Percent 35.7 16.9 55.0 -1.3 0.5 -0.4 Value added-border eq. price Cordobas 10.787.4 105,587.1 54,635.4 85,746.7 0.0 22,318.3 Gain/loss-free trade Cdrdobas -3,849.2 -17,807.5 -30,054.8 1,141.7 0.0 98.8 -50,471.0 (a) CIF price at C'orinto port. at the official exchange rate. Price unnisfor coffee and sugar are c6rdobas per quintal Meat=c6rdobas per box (55 lbs.). (b) NRP and ERP at official exchange rate (c ): Producer price atfarm gate. (d): Question of the LSMS asks about producton in the previous 12 months. The question was asked in April 1998. Non-Extreme Poor Producers Sugar Net Index Unit Rice Beans Corn Coffee Cane Meat (lbs.) Gain/Loss Border equivalent price (a) C$/metric ton 2,420.8 6,785.6 1,249.8 506.4 184.3 134.9 NRP (b) Percent 29.4 13.0 37.6 -1.8 -3.1 1.0 Domestic price (c) C$/metric ton 3,132.4 7,654.9 1,719.6 497.4 178.6 135.7 Hundredweigh Quantity Sold t 14.0 1,566.0 3,417.0 416.0 0.0 3,748.0 Revenue Cordobas 1,989.7 543,904.0 266,605.9 206,934.3 0.0 9,247.3 Revenue at Border equiv. Price Cordobas 1,537.7 482,136.7 193,771.0 210.668.3 0.0 9,192.8 Gain/loss if free trade Cdrdobas -452.0 -61,767.3 -72,834.9 3,734.0 0.0 -54.5 -131,374.7 Value Added Cordobas 1,722.0 439,172.7 214,678.5 186,220.5 0.0 8,161.3 ERP (b) Percent 35.7 16.9 55.0 -1.3 0.5 -0.4 Value added at border equiv. Price Cordobas 1,269.1 375,794.2 138,493.6 188,733.4 0.0 8,197.6 Gain/loss if free trade Cordobas 452.9 -63,378.6 -76,184.9 2,512.9 0.0 36.3 -137,467.1 (a): CF price at Corinto port. at the official exchange rate. Price unitsfor coffee and sugar are cordobas per quintal. Meat=cordobas per box (55 lbs.). (b): NRP and ERP at official exchange rate (c ): Producer price atfarm gate. (d): Question pf the LSMS asks about production in the previous 12 months. The question was asked in April/ 998. Annex 24. Page 20 Non Poor Producers Sugar Net index Unit Rice Beans Corn Coffee Cane Meat (lbs.) Gain/Loss Border equivalent price (a) C$ /metric ton 2.420.8 6.785.6 1.249.8 506.4 184.3 134.S NRP (b) Percent 29.4 13.0 37.6 -1.8 -3.1 .(' Domestic price ( c) CS /metric ton 3.132.4 7.654.9 1,719.6 497.4 178.6 135.7, Hundredweigh Quantity Sold t 32.0 1.444.0 2.894.0 9.091.0 10.868.0 15.287.(. Revenue Cordobas 4,548.0 501.530.9 225,799.6 4,522.211.0 1,940,704.8 37.717.2 Revenue at Border equiv. Price Cordobas 3,514.8 444,575.6 164,112.7 4.603.812.3 2.002.948.8 37.494.8 Gain/loss if free trade C6rdobas -1,033.2 -56,955.3 -61,686.9 81.601.3 62,244.0 -222.4 23,947.6 Value Added Cordobas 3,935.9 404,958.8 181,820.2 4.069.545.0 1.356.292.2 33.287.7 ERP (b) Percent 35.7 16.9 55.0 -1.3 0.5 -0.4 Value added at border equiv. Price Cdrdobas 2,900.8 346,517.7 117,296.0 4,124.460.8 1.349,332.4 33.435.8 Gain/loss if free trade Coirdobas -1,035.1 -58.441.0 -64,524.2 54,915.8 -6.959.8 148.1 -75,896.2 (a): CIF price at Corinto port. at the official exchange rate. Price units for coffee and sugar are cordobas per quintal. Meat =c6rdobas per box :55 lbs.). (bf) NRP and ERP at official exchange rate (c)- Producer price atfarm gate. (d) Question of the LSAIS asks about production in the previous 12 months. The question was asked in April 1998. All Producers Meat Net dex Unit Rice Beans Corn Coffee Sugar Cane (lbs.) Gain/Loss Border equivalent price (a) C$ /metric ton 2,420.8 6,785.6 1,249.8 506.4 184.3 134.9 NRP(b) Percent 29.4 13.0 37.6 -1.8 -3.1 1.0 Domestic price ( c) CS /metric ton 3,132.4 7,654.9 1,719.6 497.4 178.6 135.7 Hundredweigh Quantity Sold t 165.0 3,451.0 7,659.0 9,696.0 10.868.0 29.240.0 Revenue Cordobas 23,450.4 1.198,603.3 597,581.0 4,823.161.1 1,940,704.8 72,143.1 RevenueatBorderequiv. Price C6rdobas 18,122.9 1,062,486.5 434.326.0 4,910.192.9 2.002,948.8 71.717.7 Gain/loss-free trade Cordobas -5,327.5 -136,116.8 -163,255.0 87,031.8 62,244.0 -425.3 -155,848.8 Value Added Ciirdobas 20,294.5 967.806.6 481,188.9 4.340.370.5 1.356.292.2 63,670.7 ERP (b) Percent 35.7 16.9 55.0 -1.3 0.5 -0.4 Value added at border equiv. Price C6rdobas 14.957.3 828,139.0 310,425.0 4,398,941.0 1,349.332.4 63.953.9 Gain/loss-free trade Cdrdobas -5,337.2 -139,667.6 -170,763.9 58,570.4 -6,959.8 283.3 -263,874.8 (a): C/F price at Corinto port. at the official exchange rate. Price unitsfor coffee and sugar are cordobas per quintal. Meat-cordobas per box 55 lbs.). (b): NRP and ERP at official exchange rate (c): Producer price atfarm gate. (d): Question of the LSMS asks about production in the previous 12 months. The question was asked in April 1998. Annex 24, Page 21 APPENDIX A24.3 FORM OF ACQUISITION OF PRODUCTS, BY POVERTY GROUP Extreme Non-Extreme Extreme Non-Extreme Product Form of Acquisition Poor Poor Non-Poor Total Poor Poor Non-Poor Total ----Cdrdobas Percent of Total-- Rice Purchased 705.202 1.782,037 3.584.954 6.072.193 93.1% 91.8% 90.9% 91.4% Own production 17,153 61.557 77.781 156,491 2.3%/o 3.2% 2.0% 2.4% In payment 1.772 8.516 21,195 31.483 0.2% 0.4% 0.5% 0.5% From own store 7,135 62.162 236,344 305.641 0.9% 3.2% 6.0% 4.6% Donation 26.148 25,008 21,767 72.923 3.5% 1.3% 0.6% 1.1% Other 1.440 1.620 3.060 0.0% 0.10% 0.0% 0.0% Total 757,410 1,940,719 3,943.661 6,641,790 100% 100% 100% 100% Bean Purchased 396,457 1.101.608 2.259.829 3.757,893 69.4% 69.1% 78.3% 74.4% Own production 132,686 394.785 419.877 947,348 23.2% 24.8% 14.6% 18.8% In payment 3,432 17.804 40.850 62.086 0.6% 1.1% 1.4% 1.2% From own store 29,832 79.032 108.863 0.0%/6 1.9% 2.7% 2.2% Donation 38,698 47,952 83,574 170.224 6.8% 3.0% 2.9% 3.4% Other 2,160 2.340 4.500 0.0% 0.1% 0.1% 0.1% Total 571,273 1,594,141 2,885,502 5,050,915 100% 100% 100% 100% Corn Purchased 355,532 582,659 511,657 1,449,848 65.6% 61.8% 61.2% 62.5% Own production 170,346 313,857 299,644 783,847 31.5% 33.3% 35.8% 33.8% In payment 3,694 14,668 3.866 22,228 0.7% 1.6% 0.5% 1.0% From own store 10,465 8.347 18,812 0.0% 1. 1% 1.0% 0.8% Donation 12,011 20,455 12.678 45,143 2.2% 2.2% 1.5% 1.9% Other 520 520 0.0% 0.0% 0.1 % 0.0% Total 541,583 942,103 836.711 2,320,398 100% 100% 100% 100% Coffee Purchased 190,477 433,341 860,290 1,484,108 88.9% 87.4% 88.0% 88.0% Own production 9.640 36,742 32,418 78,800 4.5% 7.4% 3.3% 4.7% In payment 732 2,350 8,704 11,786 0.3% 0.5% 0.9% 0.7% From own store 1,095 7,747 49,741 58.583 0.5% 1.6% 5.1% 3.5% Donation 12,246 13,553 21,821 47,620 5.7% 2.7% 2.2% 2.8% Other 162 1,926 4.330 6,418 0.1% 0.4% 0.4% 0.4% Total 214,351 495,659 977,305 1,687,315 100% 100% 100% 100% Sugar Purchased 357,529 878,047 1,803.829 3.039,405 96.8% 96.1% 92.8% 94.2% Own production 0 0 0 0 0.0% 0.0% 0.0% 0.0% In payment 372 4.019 17,412 21,803 0.1% 0.4% 0.9% 0.7% From own store 5,807 27.111 110,233 143,151 1.6% 3.0% 5.7% 4.4% Donation 5.681 3,644 10.689 20,014 1.5% 0.4% 0.5% 0.6% Other 1,035 1,362 2.397 0.3% 0.1% 0.1% 0.0% Total 369,389 913,856 1,943,524 3,226,769 100% 100% 100% 100% Meat Purchased 43.470 374,956 2.809.640 3,228,066 96.2% 96.84% 96.81% 96.8% Own production 0 0 600 600 0.0% 0.0% 0.02% 0.02% In payment 0 0 1,860 1.860 0.0% 0.0% 0.1% 0.1% From own store 728 7,343 76,896 84,967 1.6% 1.9% 2.6% 2.5% Donation 981 4,888 13.362 19,231 2.2% 1.3% 0.5% 0.6% Total 45,179 387,187 2,902,358 3,334,724 100% 100% 100% 100% Annex 25, Page I Annex 25 - Labor Markets and Poverty Reduction By Nadeem Ilahi EXECUTIVE SUMMARY i The Nicaraguan labor market has gone through substantial changes due to the economic transition, but in some ways, the full gains of economic restructuring have not yet been realized. Economic reform and restructuring in the mid-I 990s have altered the structure of employment in Nicaragua. High-paying sectors such as construction. transport and financial services now have a higher share of employment and wages in these sectors have increased in real terms. However, the restructuring within manufacturing does not seem complete. The share of, and wages in, this sector have declined substantially, suggesting that manufacturing activity has not picked up. Low productivity growth in the agricultural sector has meant that this sector remains a "'parking lot" for redundant labor rather than an engine for growth and labor-related poverty reduction. ii. There are wide regional disparities in the evolution of wages between 1993 and 1998. The average hourly wage in Nicaragua in 1998 is C$7.7. In nominal terms, wages in Nicaragua have risen by 38 percent since 1993. However, there is a stark regional disparity in this increase. In the Urban Pacific region, where poverty rates increased significantly, the increase in nominal wages was only 4 percent; which translates to a significant drop in real terms. On the other hand, nominal wages in Managua jumped by 71 percent-an increase in real terms-that might be a contributing factor in the large decline in poverty. iii. Thte changes in labor markets do explain some of the changes in the regional pattern of poverty between 1993 and 1998. The general trend in the labor market between 1993 and 1998 has been one of increased participation in the labor force and employment. Labor force participation has increased (though it still remains low for women), unemployment is slightly lower, and underemployment has declined. At the household level there appears to have been an increase in hours of work as well as an added worker effect whereby secondary workers have increased their labor supply to compensate for sluggish wage growth. This increase in labor supply has more than compensated for the wage declines and we see a general trend toward lower poverty in Nicaragua. iv. Government policies do not interfere in the functioning of the labor market; tltis has allowed the formal sector to become increasingly important in the Nicaraguan labor market. Labor markets in Nicaragua are not made rigid by government policy distortions. First, minimum wages in Nicaragua do not have distortionary effects on employment. This is either because they are not set at too high a level or because they are not enforced. The minimum wage seems to have been targeted well-only 29 percent of the poor have wages below the minimum wage, compared to 22 percent of the non-poor. Second, there are few government regulations governing the formal sector with the result that there is much less segmentation in the labor market in Nicaragua than in similar countries in the region. Formal jobs do not pay an excessively high premium-only 12 percent. The share of formal sector jobs in the economy has risen and a larger share of household wage income now comes from this source. The findings suggest that employers do not encounter policy-related disincentives to create jobs in the formal sector. v. The sectoral composition of growth has constrained the demandfor labor, and as a result, poverty reduction. The overall growth that was expected during the transition has not been Annex 25, Page 2 strong enough to translate to high demand for labor. There is a clear need to explore why growth has been slow. One area of concern is the composition of growth. The manufacturing sector, which has played a significant role in export-led growth in many countries in the region, has in fact declined in its share of employment and real wages in Nicaragua between 1993 and 1998. The role of agriculture in reducing poverty is also not strong. The sector has not seen productivity improvements and real wages in this sector have largely fallen. Those deriving incomes from this sector seem to do so because other better paying options are missing. vi. Thte lack of education is a constraint on adequate labor supply. High rates of educational attainment are necessary for economic growth. In Nicaragua, the lack of education remains a major constraint to the availability of an adequately skilled labor force. Female participation in the labor force is particularly constrained by low education. Educational attainment has been increasing over time but it still remains low. Returns to education in Nicaragua are not low (8 percent) so there are incentives for individuals to invest in education. Poor supply of schools and poverty seem to be contributing to low educational attainment among the poorer sections of society. Thus targeted cash transfer programs that require children attend school, should improxe educational attainment among the poor. vii. Tltere are significant gender disparities in laborforce participation and unemployment. Education reduces these disparities and poverty. First labor force participation is much higher among men (72 percent) than women (35 percent). Second, consistent with eVidence from other countries in the region, unemployment is higher for women (15 percent) than for men (11 percent). Education significantly increases women's labor participation, and women who are in the labor force are more likely to be non-poor, which implies that increases in educational attainment will increase women's labor force participation rates and may indirectly reduce poverty by potentially increasing household income. viii. Women earn significantly less than men; some of this may be due to gender discrimination. Work interruptions hurt women more than men. Regression analysis reveals that women earn 22 percent less than men, after controlling for experience, education and other factors. The possibility that this may be due to gender discrimination canniot be ruled out. Tenure in the current job increases wages much more strongly for women than for men-an additional year of tenure raises female wages by about 3.5 percent but male wages by only 1.5 percent. Thus work interruptions, due to child bearing and rearing, exact a much larger wage penalty on women than on men. In addition, the formal sector premium is much greater for women (18 percent) than for men (8 percent). This suggests that while labor markets in Nicaragua are not segmented along formal-informal lines, there are large gender differences. Women who are unable to access the formal labor market pay a higher penalty in terms of wages than men do. ix. Child labor exists not only in poor rural households in Nicaragua, but also in non-poor ones. The incidence of child labor in Nicaragua is about I I percent. Many more boys work ( 17 percent) than girls (5 percent): As could be expected, the incidence of child labor is higher in rural than urban areas. Almost 27 percent of all rural boys work and child labor tends to be quite high in farm households. The intensity of child labor is quite high-working children spend 31 hours per week in labor. Child labor of boys is particularly high in mid-size farms of five to 20 manzanas (46 percent) than in the smallest farms of less than two manzanas (33 percent). There is also a stark difference in child labor between farm and non-farm households in rural areas, the latter have an incidence of child labor for boys of only about 9 percent. x. Work activities seriously hinder children's educational attainment. Working children pay heavy penalties in terms of their educational attainment. Compared to being in school-only, Annex 25, Page 3 combining school-and-work retards educational attainment. The retardation is even higher for those who only work and do not attend school at all. Grade-for-age of those children who are in school-only is 65 percent versus 21 percent for those who only work. xi. As child labor is particularly high in mid-size farms, poverty targeted cash grant programs for education cannot both increase enrollment and reduce child labor. The opportunity cost of a large number of working children is likely to be high as they are located in mid-size farms. Unlike their counterparts on small farms, they are not working because they are poor. This suggests that poverty targeted cash transfers that require investments in education (on the lines of Mexico's Progresa and Brazil's Bolsa Escola) may be effective instruments in increasing education but not eliminating child labor. This is because while education in Nicaragua is strongly correlated with poverty, a large amount of child labor is not. xii. Time poverty is not severe. The time use module in the LSMS allows us to look closer at one of the main non-monetarv aspects of poverty-namely time poverty. A way of telling if time poverty is severe is to assess whether the total time spent working (in income-generating activities as well as in household chores) is excessively high. The results indicate that work burdens are not excessively high in Nicaragua-the total time spent in all work activities is about 8.2 hours per day on average. This total work does not vary by poverty, i.e. those who are poor (i.e. lower than the poverty line) are no more likely to work more than the non-poor. Even more surprisingly, women do not necessarily work more than men. However there are sharp differences between the types of work men and women engage in-men are in income- generating activities while women are heavily in chores. Constraints in the supply of basic services (water, energy) do not appear to be adding a time burden of women. Thus it is unlikely that these constraints hinder their supply of labor to income-generating activities. xiii. Government policy actions regarding labor policy shouldfocus on both micro and macro aspects. The three main micro-economic challenges of the labor market in Nicaragua are the low qualifications of the labor force (refer to education indicators), gender disparities (in wages and in low labor force participation) and child labor. Real wages in agriculture have declined during the period 1993-1998, which is largely explained by lack of productivity increases, lack of an increased level of technological improvements in production and higher labor force participation-i.e. labor supply outstripping labor demand. Labor market challenges related to the macro-economy are thus associated with the need for economic growth, particularly in high paying sectors, technological change in order to have an increase in real wages and a decline in unemployment rates. Annex 25, Page 4 LABOR MARKETS AND POVERTY REDUCTION Labor Force Participation I . The overall labor force participation in Nicaragua is 54 percent.' Tables 1.1 through 1.7 summarize labor force participation by other variables. There are big differences between men and women-men are almost twice as likely to be in the labor force (73 percent) as are women (36 percent). There is a small difference in national labor force participation rate along rural- urban lines. However when the rural-urban distinction is broken down further by gender, sharp differences in labor force participation appear. 2. Regression analysis for the deterninants of labor force participation is presented in Table 1.8. Since labor force participation is a discrete variable, probit regressions are used. The first set pools men and women in the same sample and includes a female dummy (to test whether after controlling for other factors women indeed participate less in the labor force). The other two sets are separate regressions for men and women,(recognizing that different factors influence their decision to enter the labor force. After controlling for other factors, women are significantly less likely to participate in the labor force than men, by 39 percent.2 Separate regressions for men and women also show that the determinants of labor force participation are different by gender. 3. Education has opposite effects on the labor force participation of men and women3 Educated women are more likely to be in the labor force than uneducated women, but educated men are somewhat less likely to be than uneducated men. This indicates that female educational attainment is a critical determinant of increasing women's participation in income-generating activities. 4. Poverty also affects male and female labor force participation differently. After controlling for other factors that influence labor force participation, men in quintiles 2 through 5 are significantly less likely to be in the labor force compared to their counterparts in the first quintile.4 There is little association between female labor force participation and poverty. These results suggest that as we go from non-poor to poor households, male labor force participation increases but there is little change in female labor force participation. Later in this chapter we explore in greater detail the determinants of female employment behavior to look closely at this question. 5. Sociodemographic characteristics are critical in explaining different labor force participation rates between men and women. A high share of dependents in household lowers participation of men, but not of women. Women in female-headed households (on a self-reported basis) are 7 percent more likely to be in the labor force than their counterparts in male-headed households. No such relationship exists for men. Household heads are 24 percent more likely to be in the labor force than other family members. A labor force participant in Nicaragua is someone of economically active age and is either employed or looking for work. Economically active age in Nicaragua begins at age 12 and ends at age 60. 2 This result is statistically significant and robust to inclusion or exclusion of other explanatory variables. 3 No education is the excluded dummy. 4 Quintile I is the excluded dummy. Annex 25. Page 5 Unemployment 6. Unemployment rate here follows the Nicaraguan definition.' The overall unemploymeni rate in Nicaragua is 12.5 percent (Table 2.1). It is 15 percent in urban and 9 percent in rural areas. Rural women have a higher unemployment rate (14 percent) than rural men (8 percent). Tables 2.2 through 2.7 summarize the unemployment rate by age, region, education, and income. 7. Regression analysis for the determinants of unemployment is presented in Table 2.86 The first two columns test, after controlling for other factors, whether women exhibit a higher unemployment rate than men. The results show that women are significantly more likely to be unemployed. Their unemployment rate is about 2 percent higher than that of men. Most of the male-female difference in unemployment rate occurs in rural areas. There are significant differences in the unemployment rate according to educational attainment. There is higher unemployment among the educated in Nicaragua than among the uneducated. For instance, those with 10 to 12 years of education are 8 percent more likely to be unemployed than those with no education. 8. Poverty is positively associated with unemployment- the poor are more likely to be unemployed than the non-poor.' 9. Regions other than Managua exhibit significantly lower unemployment. Managua is taken as the reference dummy and therefore is excluded. The summary statistics for unemployment by regions confirm that the unemployment rate in Managua is as high as 18 percent, much higher than the national average of 12.5 percent. Underemployment 10. Underemployment is defined as employment in which the individual works less than 40 hours per week. About 32 percent of the employed are underemployed, with little difference between rural and urban areas (Tables 3.1 through 3.7). Underemployment rates are highest in Managua and the Pacific Rural region. Females in poor households are more likely to be underemployed than their non-poor counterparts. 11. Regression analysis for the determinants of underemployment is presented in Table 3.8. The first column shows that, after controlling for other factors, women are 12 percent more like ly to be underemployed than men. Underemployment falls with age, implying that the older the person the less likely they are to be underemployed. After controlling for other factors, it appears Specifically, an unemployed person is one who did not work in the 7 days prior to the survey but was actively seeking a job, or was a discouraged worker, or was disheartened by labor market prospects. 6 It is difficult to completely account for the determinants of unemployment in a regression based solely on a household survey. This is because unemployment is determined by labor demand factors that are impossible to capture in a survey of households. 7 One needs to exercise caution when trying to explain unemployment as a function of poverty. This is because the latter could in turn be influenced by the former, that is, the poor could be poor because they are unemployed. 8 Ideally, it should be for those who work less than full time and are seeking more work. The latter information, however, was not collected in the LSMS. Annex 25, Page 6 that underemployment is significantly higher in Managua (the reference dummy) compared to other regions. In which activities do the employed work? 12. The picture of employment by activity is provided in Tables 4.1 through 4.3. The pattern that emerges is not significantly different from those of other developing countries (see for example World Bank. 2000). In rural Nicaragua. almost two-thirds of the employed work in agriculture. About 11 percent work in community services and another 10 percent in commerce. However there are vast differences according to poverty status. The extremely poor in rural areas are over represented in agriculture (80 percent) compared to the non-poor (51 percent; Table 4.2). Commerce (18 percent) and community services (14 percent) account for a significant share of the non-poor's employment. 13. This picture is not substantially different for the urban employed. While II percent of the urban employed say they are engaged in agricultural work (Table 4.1), as many as 37 percent of the urban poor are in this sector (Table 4.2). Commerce and community services account for the rest (23 percent and 20 percent each respectively). For the urban non-poor, commerce (38 percent) and community services (25 percent) are the main employers with manufacturing also significant at 15 percent. 14. It is instructive to assess how the composition of employment by activity changed in Nicaragua between 1993 and 1998 and whether this structural change explains changes in poverty. Table 4.3 compares the distribution of employment by activity and region between 1993 and 1998. Here we discuss what happened to the four main activities (which together account for 88 percent of employment in 1998) over the period. The trends in these four activities indicate that the economic transition is underway but not complete. Jobs in community services-which were predominantly public sector jobs-dropped in share drastically. At the same time manufacturing jobs should have picked up as a result of overall economic reform and growth of the private sector, but they have not occurred. The share of manufacturing dropped from 11.5 percent to 9.5 percent in 1998. Not surprisingly the largest growth in jobs was in commerce-a category comprising activities that typify the informal sector. 15. The regional picture of changes in poverty can partly be explained bv structural changes in employment between 1993 and 1998. First, the significant drop in poverty in Managua can be partly explained by the substantial increase of jobs in high paying activities (manufacturing and financial services) in the city. The share of commercial sector jobs also increased in Managua. On the other hand, in the urban Pacific-where there has been a sharp rise in poverty--manufacturing jobs dropped dramatically, from 26 percent of employment in 1993 to 14 percent in 1998. This, it seems, led to a crowding of the employed in low paying sectors such as agriculture; the share of agriculture doubled from 5.5 percent to 11 percent. Commercial jobs in the region also increased, from 27 percent in 1993 to 34 percent in 1998. Are the poor more likely to work in some occupations than in others? 16. The results with respect to occupation are given in Tables 5.1 and 5.2. Two occupational categories are analyzed here: wage-employment, and self-employment and unpaid family work. A much larger proportion of the employed in urban areas are wage-employed (61 percent) than in rural areas (45 percent). Interestingly, the split between wage-employment and self-employment and unpaid family work does not vary by poverty status in urban areas. This is an indication that Annex 25, Page 7 the formal-informal sector segmentation in urban Nicaragua is not severe by developing country standards. In rural areas, the non-poor employed are more likely to be in wage- and self- employment and less in unpaid family work compared to the extremely poor (see below). The formal-informal divide: Is there segmentation? 17. By distinguishing between the formal and informal sectors we can judge the segmentation in the labor market. Such segmentation may partly be due to excessive regulations that increase the non-wage costs of labor and hence discourage formal firms from hiring. Wages tend to be higher in formal jobs than in informal ones, after accounting for education, experience, and compensating differentials.9 Typically, formal jobs are filled by the more educated, the urban, and the non-poor (Mazumdar, 1975). Is Nicaragua any different? Tables 6.1 through 6.7 summarize degree of formality of the employed in Nicaragua by various characteristics. Overall, 36 percent: of the employed are in formal employment, with 43 percent of the urban and 29 percent of the rural in such jobs. The degree of formality is lowest in agriculture and commerce, the two activities in which the majority of the Nicaraguan poor engage. Not surprisingly, the degree of formality is the highest in Managua and generally high in other urban centers (Table 6.4). The most striking pattern is that of education-the higher one's level of education, the more likely one is to hold a formal sector job. 18. The regression results in Table 6.8 confirm our findings on formal employment in Nicaragua but show that once we control for other factors the relationships are more nuanced. First, women are 13 percent less likely than men to work in formal jobs. Second, men in urban areas are only 5 percent more likely to work in formal jobs than men in rural areas. This is a small difference compared to other countries in the region. Third, there is a clear positive association between level of education and the probability of having a formal job, but the magnitudes are not substantial. Those with 13 or more years of education are 5 percent more likely than the uneducated to have formal jobs. Fourth, sociodemographic characteristics are also important determinants. As the share of dependents in the household increases, the probability of holding a formal job falls. This may be that because the inflexible hours that characterize formal jobs make it more difficult to accommodate childcare. How many hours do people work? 19. Those that work spend an average of 48 hours per week on the job. There is surprisingly little variation between the poor and non-poor and between rural and urban areas (Table 7. 1). Hours of work do vary significantly, however, by activity. Compared to agriculture, hours in transport and financial services are much higher (Table 7.2). There also vast differences by regions, with the lowest hours in Managua-consistent with the finding above that Managua also has high underemployment. What determines wages? 20. How do hourly earnings or wages vary? We count hourly wage (both cash and in-kind) received in employment as remuneration. We use wages that have been adjusted for regional 9 Compensating differentials are wage differences that arise from differences in quality of work. Thus because of a compensating differential, a worker in a nuclear power plant would earn more than one in a less vulnerable activity, all other factors being equal. Annex 25, Page 8 price differences. The only case where such wages have not been used are in the discussion of minimum wage, where using nominal wages allows direct comparison with the officially set wage, which is not adjusted regionally. 21. Tables 8.1 through 8.8 profile wages by various characteristics. Wages vary by location and gender. Summary data indicate that women earn less than men. Workers in rural areas make significantly less than their urban counterparts. Wages appear to differ widely by poverty; and the broad wage disparity between the extremely poor and the non-poor appears much greater in urban than in rural areas. Wages in agriculture average C$4.8 per hour, far below the national average of C$7.7 per hour. Workers in the financial sector and transport earn the most. Again, the within- sector wage disparity is much greater in urban than in rural areas. Education is positively associated with returns to labor. For both men and women, wages peak in the 40-50 age group. 22. Before confirming these findings by regression analysis, we examine the evolution of wages between 1993 and 1998, presented by activity in Table 8.8. The 1993 nominal wages have been adjusted at 1998 prices to make them comparable to current nominal wages in Nicaragua. Overall real wages declined from C$9.3 in 1993 to C$7.7 in 1998. Thus the overall reduction in poverty experienced in Nicaragua over the period has to be due to factors other than real wages- reduction in unemployment/underemployment or an increase in hours worked. It is nonetheless interesting to note from Table 8.8 that changes in wages over the period were not equal by activity. Manufacturing wages fell by twice the national average, suggesting that economic transition has hit this sector hard. Note from Table 4.3 that the share of this sector in employment also fell quite a bit. The retrenchments in manufacturing and in community services (largely in the public sector) increased employment in the marginal commercial sector, as noted above (see Table 4.3), where wages fell 25 percent. However, restructuring also increased the share of employment in high-paying sectors such as construction, transport, and financial services. Real wage improvements in these sectors explain some of the improvements in poverty in urban areas. 23. Regional changes in poverty between 1993 and 1998 can also be explained by changes in real wages. In the urban Pacific, the drastic retrenchment in manufacturing (much greater than the national average) and the subsequent increase in marginal sectors (commerce and agriculture) went hand-in-hand with drastic drops in real wages in all these sectors. Overall wages in this region dropped 37 percent between 1993 and 1998. Manufacturing wages dropped more than 50 percent (the result of a lack of growth in this sector) and consequently crowding in agriculture and commerce caused a 33 percent and 38 percent drop in wages in these activities respectively. On the other hand in Managua, where there was a substantial reduction in poverty, wages and employment in high-paying sectors (construction, financial services, and transport) compensated for the drop in employment and wages in manufacturing. 24. Summary data can cloud the true relationships between individual and job characteristics and wages. Therefore it is important to assess the effects of these characteristics on wages by a regression analysis. However, those who work are a self-selected (higher ability) sub sample and therefore regression results based on only those who work are likely to be biased. A Heckman wage correction was used to correct for self-selection (Greene, 1997).1' Three regressions are reported in Table 8.9. The first column reports the results of a regression that pools men and ° It is important to use exclusion restrictions to identify this procedure. Specifically we need to identify variables that influence the "decision" to enter the labor force (and report wages) but not the level of wages. We use number of children, marital status, and education of head of household as exclusion restrictions. Annex 25, Page 9 women and tests whether after controlling for factors such as education and experience women do indeed make less than men. The other two regressions are conducted separately. 25. Do women eam less than men, after controlling for education, experience, location, job characteristics and self-selection into the labor market? Our results, shown in Table 8.9, indicate that women earn 22 percent less than men. This result is statistically significant and quite robusi. To explore the determinants of wages further, we therefore split the sample by gender. The strongest result in the two regressions is that education has a strong effect on earnings. The return to a year of education in Nicaragua is 7.6 percent. It is higher for men (8 percent) than for women (7 percent). Tenure in the current job increases wages much more strongly for women than for men. On the margin, an additional year of tenure raises female wages by about 3 percent.'1 This suggests that women who continue in the same job without interruption are likely to experience faster growth in wages compared to those that interrupt their employment. Urban wages are significantly higher than rural ones. Job characteristics also influence wages. After controlling for all other characteristics the formal sector wage premium-i.e. the difference in pay between formal and informal jobs-is only 12 percent. This is quite low compared to other countries in the region (see, for example, World Bank, 2000). The formal sector premium is much greater for women (18 percent) than for men (8 percent), suggesting that dualism in the labor market affects women a lot more than it does men. Another way to look at this is that women who are unable to enter the formal labor market pay a higher penalty in terms of wages than men do. Does the minimum wage benefit the poor? Does it distort employment? 26. The main reason to enact a minimum wage is to protect workers from being paid wages considered too low for sustenance. Since the level of subsistence is subjectively determined, there usually is no set formula for setting the minimum wage. A minimum wage set too high is likely to be binding for a large number of workers and therefore will be poorly targeted-i.e. it would leak to the non-poor. It can also distort the labor market by increasing unemployment, for example. If set too low, it does not allow those with the lowest wages to earn enough to live on and is therefore ineffective.'2 27. In Nicaragua minimum wages were established in 1997. They vary by sector of activity. Table 9.1 compares the existing average nominal wage (not regionally adjusted) and minimum wages for 1997. It does not appear that the minimum wage was set at an excessively high level. Table 9.2 shows that only 25 percent of the workers reported wages lower than the minimum wage.'3 Further, for those who make more than the minimum wage, average wages are many times higher than the minimum wage (Table 9.3). For instance, the average wage of those workers in financial services who make more than the minimum wage is more than 6 times the minimum wage; in agriculture, this ratio is 4 to 1. In other words, there was no "clustering" of wages at the minimum wage level. This indicates that the observed wages are not bound by the lower limit of minimum wage. The return to a year of tenure was calculated at the sample mean--6.6 years for women and 10.7 years for men. The return to tenure for men is 0.02 percent only. 12It would be poorly targeted for another reason-because it would not cover all the poor. 13 This is not very different from other countries in the region. For instance the recently concluded povery assessment for Panama also found about 26 percent of the workers were receiving below the minimum wage (see World Bank, 2000). Annex 25, Page 10 28. One test of the effectiveness of the minimum wage is how well it targets the poor.4 Table 9.4 indicates that minimum wage may have benefited the poor. First, only 29 percent of the workers in the bottom quintile (extremely poor) received less than the minimum wage. This compares quite favorably with the non-extremely poor group, where as many as 15 percent received less than the minimum wage.'5 Second, Table 9.2 indicates that minimum wage has not had distorting effects on unemployment, especially among the poor. A minimum wage can fail by causing higher unemployment, which particularly hurts the poor. However, the results in Table 2.5 show that unemployment is not significantly higher among the poor than among the non- poor.'6 Table 9.5 further breaks down the information in Table 9.3 by activity. The figures in this table show that in some sectors (community services, financial services, gas, electricity and water, construction, and agriculture) the poor have much lower coverage of the minimum wage than the non-poor. In others (transport and mining) the coverage is more even. One should exercise caution in interpreting these results since the degree of intrasectoral disparity in wages influences the extent to which poor and non-poor are covered by the minimum wage. Who are the child workers? 29. Children aged 10-14 are perhaps the most vulnerable group in developing countries. The tendency of their families to send them to school is lower than when they are younger because the need for them to contribute economically to the family income is high (see Ilahi, Sedlacek and Sasaru, 2000). We define child labor here as all work done in economic activities (on the family farm, in the family business, or outside work for wages). As shown in Table 10. 1, in Nicaragua, almost 75,000 children work, the vast majority of them boys (about 77 percent). The incidence of child labor, defined as the proportion of those in the 10-14 age group who work, is 11 percent. This is quite low compared to other countries in the region. As many as 16 percent of rural children work, compared to only 6 percent of urban children (Table 10.2). Rural boys, 27 percent of who participate in some form of economic activity, are the most vulnerable. Working children spend about 31 hours per week in labor activities (Table 10.3). 30. There is evidence that some child labor in rural Nicaragua is due to failures in the labor markets, rather than from poverty. Table 10.4 shows that while child labor tends to be higher in the farmn households, it is much higher in the mid-size farms (46 percent among boys) than in the smallest farm households (33 percent among boys). This suggests that poorly functioning labor markets force mid-size farm households to rely on the labor of young children. There is a stark difference between farm and non-farn households, where the incidence of male child labor is below 9 percent. There is little to suggest that child labor varies according to the head of household's gender. 31. Where do children work? Table 10.12 shows that in rural areas, most working boys engage in agriculture (94 percent). Rural girls also work in agriculture (53 percent, but manufacturing (13 4This is certainly not the only test of the effectiveness of the minimum wage. However, since the primary objective of having a minimum wage is to prevent wages from falling below sustenance levels, this may be a good criterion. 15 These numbers are quite different from Panama where while 26 percent of all workers earned wages below the minimum wage, as many as 68 percent of the extreme poor and 17 percent of the non- extremely poor had wages below minimum. 6 This is a crude test of the unemployment creating effects of minimum wage because unemployment tends to be higher among the poor even without a minimum wage. Annex 25. Page 11 percent) and commerce (30 percent) are also significant employers. In urban centers, most of the children who work have jobs in commerce (50 percent for boys and 75 percent for girls). More than one quarter of the working boys in urban areas work in agriculture. 32. The picture of child labor in the preceding paragraph is interesting but inaccurate because the relationships do not control for other determinants of child labor. Table 10.14 lists the results of probit regressions on determinants of child labor. The first column lists the results of a pooled regression. After controlling for age, household characteristics, quintiles and regions, the incidence of child labor among girls is on average 10 percent less than for boys. The second and third columns of Table 10.14 list results of separate regressions for boys and girls. The probability of being a child worker increases with age. Within the 10-14 age group, an additional year of age increases the incidence of boys' labor by 4 percent and of girls' labor by I percent. No consistent picture of child labor by region emerges from the analysis. 33. The association between child labor and poverty in Nicaragua is complex. Monetary indicators of poverty, as captured by consumption quintiles, do not show a clear association between poverty (i.e. belonging to the bottom two quintiles) and child labor. However, non- monetary (and more long-term) measures of poverty-such as lack of access to modern fuels arld tap water-are significantly affect child labor. For instance, boys who have in-house water taps are 7 percent less likely to work than are those that do not. For girls the effect is 5 percent. Those with access to modern fuels in the household are 8 percent less likely to work (2 percent for girls). The evidence suggests that some of the child labor in Nicaragua may be associated with the long-term poor. Does child labor affect educational attainment? 34. Table 1 1. I shows the distribution of school and work by poverty status in rural and urban areas. In rural areas, current enrollment rate in the 10-14 age group is 78 percent. Of these, 88 percent are in school only. However this distribution is very different when disaggregated by poverty status. Only 64 percent of the poor children are in school, compared to 92 percent of the non-poor. Table 11.3 shows the average educational attainment of children in school/work combination by poverty status. Educational attainment is highest among children who only attend school, followed by those who both attend school and work, followed by those who only work. The difference in attainment between those who both attend school and work and those who only attend school is much greater for boys than for girls. 35. Poverty is in general associated with lower educational attainment. Children in extremely poor households in Nicaragua have half the educational attainment of those in non-poor ones. Table 11.4 shows the distribution of educational attainment by school/work combinations and age. Instead of summarizing raw educational attainment, the results are presented by grade-for- age.'7 Using grade-for-age better captures progression through the school system (Patrinos and Psacharopoulos, 1997). Boys who combine school and work have an average grade-for-age of 58 percent while those in "school only" have a grade-for-age of 65 percent. The difference in attainment between those who are in school and work and those in "work only" is even greater. The grade-for-age of those who only work is 21 percent. These results resoundingly suggest that 17 Grade-for-age is defined as [{educational attainment/(age-7)}* 100] where 7 is the typical age of entry in school. A grade-for-age measure of 100 indicates complete educational attainment (i.e. no falling behind), and one of zero indicates no educational attainment (or complete falling behind). See Patrinos and Psacharopoulos (1997). Annex 25, Page 12 when children work their educational attainment is seriously hindered. As would be expected, combining school and work hinders attainment, but not as much as for those who only work. TIME USE AND TIME POVERTY 36. An analysis that focuses only on labor market behavior of individuals does not adequately assess poverty. Labor market analysis tells us the factors that affect labor force participation, employment, wages etc., and hence the income-earning capacity of individuals. However a large portion of the people's labor in developing countries occurs outside the labor market, in chores inside and outside the household. It is important to study time use when assessing poverty because of the intimate relationship between time-use and poverty (see llahi, 1999 for details). 37. Time is a resource, but if we concentrate on labor behavior alone we are likely to miss a large portion of the use of this resource.'8 For instance, the nature of time spent in household chores can give us a good indication of time constraints and their relation to poverty. As any other resource in the household, time is not equally distributed across members. There are significant differences not just along gender lines but also by age, social status, wealth etc.19 Third, there is a potential connection between the study of time use and the study of poverty. As already mentioned, a significant number of poor households in developing countries survive by home production, attained primarily by the time of household members. Similarly, leisure, the flip side of "work, can be counted as a good from which people obtain welfare. An important question then is whether we should worry about the shortage of time as an indicator of poverty as much as we think of a shortage of money as an indicator of poverty. Last, development policy interventions, be they poverty alleviation, safety nets, basic services projects, or agricultural extension programs, can be improved by the information that comes from time use. Constraints on the time of some or all household members can critically affect the success of projects. For instance, if household members spend a lot of their time travelling to work on foot, there is a high likelihood that the provision of rural roads would have high returns. In this section we examine the unique time-use module that is a part of the Nicaragua LSMS. We split the analysis of time for a sample of adults (ages 15 through 60) and children (ages 10 through 14). How do adults use their time? 38. Tables 12.1 through 12.13 summarize statistics for time spent by adults in various activities. Time allocated to all work activities (income generation plus household chores) is not excessively burdensome (Table 12.1). Women spend 8.2 hours per day in all work, and men spend slightly less (7.7 hours per day). These results are largely consistent with those found in other developing countries (see Ilahi, 1999, for a review). In rural areas men spend more time in work than women do. The poor and extremely poor work more hours than the non-poor, but the difference is not large. Essentially, time allocated to all work activities does not vary by poverty. 39. Work activities can be broken down further into housework and income-generating activities. Table 12.4 provides the summary data for time spent in household chores. The results 18 Household chores are usually hidden to those who look for market interactions and are in fact not recorded in national income accounts. However they form an important part of the household's basic needs. 19 There is today an extensive literature on the intra-household allocation of resources in developing countries (see for instance Haddad et al, 1997). Annex 25, Page 13 show that time spent in household chores is not excessively high, but that there is a sharp division of labor by gender. Women do most of housework (5.5 hours per day) compared to men (1.5 hours per day). Are household chores excessively burdensome, especially for the poor? The summary results for total time spent working and the time spent in chores suggest that the scarcity of time in the lives of the poor may not be a problem in Nicaragua. Chores account for 54 perce nt of total adult time. They occupy 75 percent of the time of adult women and only 28 percent of that of men (Table 12.5). Disaggregating chores into fuel and water collection shows that time spent in such activities is not high either (Tables 12.6 through 12.11). Men spend a lot more time than women in fuel collection (Table 12.11), women more than men in water collection (Table 12.8).20 The time women spend in childcare is quite high-about 3 hours per day-but is not necessarily higher for poor than for non-poor women. Men allocate little time to this activity. What affects the use of adult time? 40. In the time-use regressions in this section, the explanatory variables include: a) demographic information about the individual (age, years of education), b) demographic characteristics of the household (number of children in different age groups, gender of head, share of adult females in household population, c) household quintile rank as well as "affluence" (access to modern fuels, access to tap water, materials used in walls, floors and ceiling), d) cluster level wages as indicators of opportunity cost of time, e) regional dummies, f) an urban-rural indicator, and g) a dummy for whether the respondent was interviewed about his or her time use during the weekend.2' Because the allocation of time exhibits large differences by gender in most studies, separate regressions for men and women are estimated. All Work 41. Table 13.1 provides results of regressions for men and women for time allocated to all work. Note that time spent in all work is the opposite of time spent in leisure activities. Thus this regression explains the determinants of total work burden. The first column pools the sample of men and women and tests whether after controlling for other factors, women work more than men. The results show that while the difference in the time men and women spend working is statistically significant, it is not large: on average women work only 36 minutes more than men in a day. Time allocated to all work activities rises with age but at a decreasing rate. Those with more education tend to work less, that is, they take more leisure time, with the impact being greater for men than for women. The presence of very young children (under 5) increases work burdens significantly for women and for men, but the impact is much greater on women. On the other hand, the presence of older children (10-14) allows both adult men and women to lower their total work, again with the effect being much greater in magnitude for women. 42. The direction of the association between work and poverty depends upon which metric of poverty is used. If non-monetary measures are used, then there is a positive association between poverty and work burdens. Household members in poor quality dwellings seem to work more than do those in better dwellings. Those in dwellings with dirt floors work significantly more than their counterparts in better quality dwellings. However the poor as defined by consumption 20 Fafchamps and Quisumbing (1998) find similar results in Pakistan. 21 It is crucial to include "day of the interview" dummies in our regressions because individuals tend to work significantly less during the two days of the weekend. Since the question was asked about timle use during the previous day, we include dummies for Sunday and Monday to control for weekends. Annex 25, Page 14 aggregate do not appear to work any more than the non-poor. Access to fuel and water inside the household also affects the total workload. Access to modem fuels (such as gas or kerosene) lowers the work time of men but not of women (see details below). Compared to those with access to modern fuels, those without access spend 79 more minutes per day working. In-house access to tap water significantly lowers the work time of adult women. These two results are consistent with our tabulations showing that fuel collection is a largely male activity and water collection a female one. Finally, most regional dummies and one urban dummy show significant effects. Compared to Managua, almost all other regions exhibit lower work burdens. Housework 43. Table 13.2 provides results of the regression for time spent doing housework. Note that this regression further disaggregates total work into time spent in income-generating activities and time spent on housework. We know a lot about time spent in income-generating activities; here we use the standard variables of labor supply analysis to peek into the determinants of time men and women spend in household chores. Our objective is to identify constraints that force individuals to spend time in household chores at the expense of participating in income- generating activities.22 The results indicate that men spend significantly less time than do women in such activities, and that housework varies significantly over the life cycle. Even though men spend much less than women in such activities, the time men spend does respond to constraints. Access to a water tap significantly lowers the chore time of women only, and access to modem fuels affects the time of men only. The presence of very young children (less than age 5) increases women's time in chores but has little effect on men's time. For women, the presence of older children (6-9) also increases time in chores. The presence of adolescent children (10- 1 4) lower the time that both men and women spend on housework, but the results are not statistically significant. The number of prime-age females in the household also tends to lower the time spent on chores. Those in poor-quality dwellings tend to spend significantly more time than do those in better dwellings. However, after controlling for all other factors, poverty as defined by consumption aggregates does not affect time spent on chores. Nor are there clear regional patterns of time spent on chores. Water and Fuel Collection 44. Tables 13.3 and 13.4 show the determinants of time spent collecting water and fuel. Since the total time spent is such activities is not excessively high (see Tables 12.6 and 12.9) we restrict the discussion of the results to some critical variables. In the regression for water collection time (Table 13.3) we also included dummies for whether the household collects water from a well or a river (as opposed to from an in-house source). The objective here is to see whether the lack of water supply infrastructure adds to the chores of individuals in households that do not have this access (Ilahi and Grimard, 2000). The results indicate that access to tap water has no effect on time use, either of men or of women. 45. Contrary to popular wisdom, men in Nicaragua spend significantly more time collecting firewood than do women. Time allocated to fuel collection is closely related to distance traveled to collect wood. Each kilometer of distance increases collection time by about 10 minutes. After controlling for access to modem fuels, there is no association between poverty and fuel 22 The determinants of time spent on work are discussed in a previous section of this chapter. Annex 25, Page 15 collection. Men in poorer households do spend significantly more time in this activity than their counterparts in non-poor households. Childcare 46. Table 13.5 lists the regression results for time allocated to childcare. 3 Childcare can constrain women's participation into income-generating activities. Because few men provide childcare. only the regression for females is reported here. Not surprisingly, the number of very young children in the household (under age 5) is perhaps the most important determinant of adult hours allocated to the care of children. The presence of other adult females in the household lowers the time a woman spends on childcare. There is a weak relationship between childcare and poverty. How do children use time? 47. Tables 14.1 through 14.5 present summary statistics for children's time use. First, the time spent by children in all work is not high at all-2.5 hours per day for boys and 1.5 for girls (Table 14.1). There is a wide variation between urban and rural areas-children in rural areas work twice as long as children in urban areas. It also varies by poverty-children in poor households work significantly more than those in non-poor households. Child work is quite burdensome. Among those children who work, time spent in income-generating activities is as high as 5.4 hours per day (Table 14.2). There is little difference in this between boys and girls but a large difference between poor and non-poor. On the other hand the time children spend in chores is quite small (Table 14.3) and on average does not exceed I hour per day. Children in extremely poor households attend school-significantly less than do their counterparts in non-poor households. What affects their use of time? 48. Table 15.1 provides results for the determinants of time children allocate to all work. There is a sharp distinction between the nature of work done by boys and girls. toys mostly work outside the home or work in the family farm or non-farm enterprise, whereas girls mostly do housework. To obtain a complete picture of the work of both boys and girls, we presents regressions for both housework and total work, in addition to time allocated to income-generating activities. Girls work significantly longer than boys. 49. Boy child labor is affected by access to school. Distance to school was used as a proxy for school supply. The results indicate that distance to school is important in the work of boys but not girls. Because boys are more likely to do "outside" child labor, households farther away from schools are likely to have higher work burdens for boys than those that lie closer. The presence of other children in the 10-14 age group and of prime-age females in the household lowers the work time of boys significantly but not of girls. Children's time in all work does not appear to be directly associated with poverty status. However, indirect measures of poverty do appear to correlate with child labor. For example those in households with dirt floors and adobe walls spend significantly more time on work. 23 It is comparatively more difficult to measure time allocated to childcare than time allocated to other activities. This is because childcare is often a "joint" or simultaneous activity. To completely account for childcare time I have defined childcare time here as the time devoted solely to childcare and the time devoted to it as a simultaneous activity. Annex 25. Page 16 Housework 50. Table 15.2 provides the results of regression for time spent in all household chores by children. Girls spend a lot more time on this activity than do boys. The presence of prime-age females significantly lowers the time spent on housework for boys but not for girls. On the other hand, access to water at the household level affects the time of girls but not boys. Annex 25, Page 17 APPENDIX A25.1 - LABOR FORCE PARTICIPATION Table A25.1.1 - Labor force participation rates by gender and location Rural Urban Total Male 0.81 0.67 0.73 Female 0.27 0.42 0.36 Total 0.55 0.54 0.54 Note: For individuals older than 12. Table A25.1.2 - Labor force participation rates by age group Rural Urban Total Age (;'rs) .4Aale Female Male Female Male Female 12-19 0.62 0.16 0.30 0.16 0.45 0.16 20-24 0.90 0.26 0.75 0.39 0.82 0.34 25-29 0.93 0.36 0.86 0.58 0.89 0.49 30-34 0.93 0.45 0.92 0.61 0.92 0.55 35-39 0.89 0.32 0.90 0.64 0.90 0.51 40-44 0.89 0.35 0.90 0.69 0.90 0.56 45-49 0.92 0.33 0.88 0.50 0.90 0.44 50-54 0.88 0.35 0.89 0.50 0.89 0.43 55-59 0.87 0.19 0.87 0.48 0.87 0.36 60-64 0.92 0.19 0.67 0.32 0.79 0.27 Over 65 0.91 0.18 0.78 0.27 0.87 0.21 Note: For individuals older than 12. Table A25.1.3 - Labor force participation rates by region Region Male Female Managua 0.64 0.39 Pacific Urban 0.69 0.47 Pacific Rural 0.77 0.32 Central Urban 0.75 0.43 Central Rural 0.84 0.22 Atlantic Urban 0.68 0.34 Atlantic Rural 0.85 0.19 Note: For individuals older than 12. Annex 25, Page 18 Table A25.1.4 - Labor force participation rates by education Rural Urban Total Years of education Male Female Male Female Male Female 0 0.89 0.22 0.78 0.42 0.86 0.29 1 - 6 0.77 0.27 0.63 0.38 0.71 0.33 7 - 9 0.77 0.25 0.67 0.38 0.69 0.35 10-12 0.87 0.54 0.73 0.52 0.75 0.53 13-17 0.93 0.68 0.80 0.65 0.81 0.65 Note: For individuals older than 12. Table A25.1.5 - Labor force participation rates by poverty status Rural Urban Total Poverty Status Male Female Male Female I Male Female Extremely poor 0.80 0.24 0.73 0.41 0.79 0.28 Poor 0.81 0.25 0.68 0.42 1 0.77 0.31 Non-poor 0.79 0.31 0.67 0.42 0.70 0.39 Note: For individuals older than 12. Table A251.6 - Labor force participation rates by consumption quintiles Rural Urban Total Quintiles Male Female Male Female 1 Male Female I 0.81 0.24 0.73 0.38 0.79 0.28 2 0.82 0.26 0.65 0.43 0.76 0.32 3 0.81 0.26 0.66 0.44 0.73 0.37 4 0.81 0.31 0.70 0.40 0.74 0.38 5 0.72 0.40 0.64 0.43 0.65 0.42 Note: For individuals older than 12. Table A25.1.7 - Labor force participation rates by rural household typology Typology Male Female < 2 manzanas 0.83 0.26 2-5 manzanas 0.85 0.25 5-20 manzanas 0.88 0.24 > 20 manzanas 0.85 0.12 Non-ag; low education 0.74 0.30 Non-ag; high education 0.73 0.34 Note: For individuals older than 12. Annex 25, Page 19 Table A25.1.8 - Probit regression for labor force participation Variable All Male Female Education Grade 1-6 0.072** 0.067** 0.028* 0.018 0.106** 0.107** Education Grade 7-9 0.031* 0.015 -0.044** -0.071** 0.108** 0.107** Education Grade 10-12 0.1 16** 0.094** -0.009 -0.050** 0.207** 0.204** Education Grade 13-17 0.175** 0.145** 0.036 -0.015 0.291** 0.283** Age20-24 0.231** 0.238** 0.205** 0.212** 0.130** 0.131** Age 25-29 0.368** 0.373** 0.268** 0.272** 0.282** 0.284** Age 20-34 0.421** 0.425** 0.259** 0.263** 0.438** 0.440** Age 35-39 0.434** 0.435** 0.254** 0.256** 0.477** 0.477** Age 40-44 0.413** 0.415** 0.247** 0.249** 0.461** 0.461** Age 45-49 0.408** 0.409** 0.232** 0.234** 0.480** 0.480** Age 50-54 0.365** 0.367** 0.225** 0.227** 0.377** 0.378** Age 55-59 0.351** 0.352** 0.215** 0.218** 0.377** 0.378** Age 60-64 0.325** 0.326** 0.207** 0.209** 0.308** 0.308** Single 0.049** 0.049** -0.011 -0.011 0.061** 0.062** Hlousehold Male 0.246** 0.244** 0.166* 0.160** 0.169** 0.169** Urban -0.012 -0.026** -0.111** -0.132** 0.093** 0.091** Share of dependents -0.001 0.006 0.016 0.023 0.027 0.030 Female-headed 0.041** 0.042** 0.012 0.017 0.074** 0.073** household Children under6 0.013** 0.018** 0.013* 0.023** 0.010 0.010 Children 6-9 -0.001 0.004 -0.005 0.006 0.000 0.000 Children 10-14 -0.003 0.002 -0.012** -0.006 0.001 0.002 Quintile 2 -0.013 -0.062** 0.042** Quintile 3 -0.026 -0.077** 0.032 Quintile 4 -0.039** -0.069** 0.003 Quintile 5 -0.087** -0.174** 0.009 Female -0.393** -0.394** Log likelihood: No. Observations: 14283 14283 6989 6989 7294 7294 LR Chi-Sq 4874.01 4849.61 1998.01 1943.90 1333.24 1325.76 Pseudo R-Sq: 0.247 0.246 0.247 0.240 0.140 0.139 Note: Coefficients are marginal effects. * significant at 5 percent. * * significant at 10 percent. Annex 25, Page 20 APPENDIX A25.2 - UNEMPLOYMENT Table A25.2.1 - Unemployment rate by gender and rural/urban location Rural Urban Total Male 0.080 0.148 0.113 Female 0.136 0.151 0.147 Total 0.094 0.149 0.125 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.2.2 - Unemployment rate by age group I Rural i Urban Total Age (yrs) Male 1 Female Male Female Male | Female 12-19 0.158 0.226 0.181 0.270 0.166 0.250 20-24 0.063 0.242 0.202 0.243 0.129 0.243 25-29 0.052 0.097 0.107 0.159 0.081 0.140 30-34 0.051 0.048 0.096 0.133 0.076 0.108 35-39 0.081 0.068 0.141 0.121 0.116 0.107 40-44 0.059 0.100 0.135 0.104 0.103 0.103 45-49 0.013 0.104 0.150 0.103 0.091 0.103 50-54 0.009 0.136 0.147 0.075 0.085 0.097 55-59 0.024 0.127 0.122 0.081 0.068 0.090 60-64 0.110 0.041 0.192 0.012 0.148 0.020 Over 65 0.000 0.081 | 0.000 0.000 0.000 0.038 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.2.3 - Unemployment rate by region Region Male Female Managua 0.178 0.177 Pacific Urban 0.158 0.145 Pacific Rural 0.112 0.173 Central Urban 0.097 0.105 Central Rural 0.050 0.121 Atlantic Urban 0.114 0.103 Atlantic Rural 0.029 0.068 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.2.4 - Unemployment rate by education Rural Urban Total Years of Education Male Female Male Female Male Female 0 0.057 0.148 0.116 0.104 0.071 0.126 1 - 6 0.083 0.119 0.151 0.121 0.112 0.120 7-9 0.155 0.233 0.151 0.185 0.152 0.192 10-12 0.114 0.136 0.172 0.193 0.163 0.182 13-17 0.032 0.023 0.112 0.168 0.102 0.152 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Annex 25. Page 21 Table A25.2.5 - Unemployment rate by poverty status Rural Urban Total Poverty, Status Male Female Male Female Male Female Extremely poor 0.084 0.248 0.145 0.154 0.097 0.213 Poor 0.085 0.185 0.168 0.173 0.109 0.179 Non-poor 0.072 0.064 0.140 0.144 0.117 0.128 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.2.6 - Unemployment by consumption quintiles Rural Urban Total Quintile Male Female Male Female Male Fentale 0.077 0.227 0.157 0.166 0.095 0.203 2' 0.105 0.152 0.178 0.188 0.127 0.171 3 0.073 0.097 0.144 0.150 0.110 0.136 4 0.069 0.055 0.178 0.127 0.140 0.111 5 0.037 0.039 0.104 0.152 0.091 0.139 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.2.7 - Unemployment by rural household typology Typology Male Female < 2 manzanas 0.08 0.14 2-5 manzanas 0.08 0.21 5-20 manzanas 0.03 0.10 > 20 manzanas 0.02 0.11 non-ag; low education 0.12 0.14 non-ag; high education 0.13 0.12 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Annex 25. Page 22 Table A25.2.8 - Probit regression for unemployment Variable All Urban Rural Education Grade 1-6 0.004 -0.005 0.020 0.008 -0.009 -0.015 Education Grade 7-9 0.054** 0.029** 0.052** 0.028 0.072** 0.050** Education Grade 10-12 0.080** 0.042** 0.097** 0.059** 0.042* 0.010 Education Grade 13-17 0.059** 0.008 0.072** 0.023 -0.023 -0.056* Age -0.012** -0.012** -0.011** -0.011** -0.011** -0.012** Age squared 0.000** 0.000** 0.000° 0.000°°* 0.000° * 0.000** Quintile 2 -0.001 0.003 0.000 Quintile 3 -0.045** -0.043** -0.034** Quintile 4 -0.049** -0.038* -0.050** Quintile 5 -0.075** -0.074** -0.070** Region: Pacific Urban -0.029** -0.017* -0.035** -0.023* Region: Pacific Rural -0.058** -0.038** -0.056** -0.032** Region: Central Urban -0.067** -0.060** -0.084** -0.076** Region: Central Rural -0.1 12** -0.096** -0.120** -0.091** Region: Atlantic Urban -0.057** -0.049** -0.069** -0.061** Region: Atlantic Rural -0.106** -0.101** -0.082** -0.078** Female 0.021** 0.020** 0.005 0.003 0.046** 0.048** No. Observations: 7875 7875 4190 4190 3685 3685 LR Chi-Sq (17) (13) (14) (10) (14) (10) LR Chi-Sq 381.71 321.24 113.65 91.05 264.36 212.87 Pseudo R-Sq: 0.064 0.054 0.032 0.026 0.115 0.093 Note: Coefficients are marginal effects. The unemployment rate in dependent variable is measured in proportions, so a rate of 0.125 means 12.5%. * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 23 APPENDIX A25.3 - UNDEREMPLOYMENT Table A25.3.1 - Underemployment rate by gender and rural/urban location Rural Urban Total Male 0.283 0.280 0.281 Female 0.441 0.372 0.394 Total 0.318 0.319 0.319 Note: Rates are in proportions. so a rate of 0.125 means 12.5%. Table A25.3.2 - Underemployment rate by age group Rural Urban Total Age (yrs) 1 Male Female Male Female | Male Female 12-19 0.409 0.578 1 0.405 0.599 0.408 0.589 20-24 0.268 0.302 0.263 0.446 i 0.266 0.400 25-29 0.223 0.396 0.275 0.339 0.249 0.357 30-34 0.281 0.465 0.238 0.317 0.258 0.363 35-39 0.221 0.439 0.298 0.303 0.265 0.340 40-44 0.212 0.417 0.213 0.375 0.213 0.385 45-49 0.228 0.388 0.242 0.274 0.235 0.304 50-54 0.211 0.374 0.263 0.351 0.238 0.359 55-59 0.193 0.590 0.224 0.453 0.206 0.480 60-64 0.336 0.497 0.342 0.305 0.339 0.352 Over 65 0.276 0.577 0.156 0.269 0.242 0.405 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.3.3 - Underemployment rate by region Region Male Female Managua 0.395 0.453 Pacific Urban 0.265 0.393 Pacific Rural 0.373 0.488 Central Urban 0.176 0.273 Central Rural 0.202 0.322 Atlantic Urban 0.141 0.242 Atlantic Rural 0.284 0.549 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.3.4 - Underemployment rate by education I Rural 1 Urban Total Years of Education Male Female Male Female Male Female 0 0.250 0.412 0.217 0.339 0.243 0.375 1 - 6 0.288 0.434 0.295 0.399 0.291 0.412 7 - 9 0.386 0.447 0.280 0.392 0.307 0.400 10-12 0.407 0.530 0.283 0.343 0.308 0.381 13-17 0.033 0.472 0.285 0.320 0.251 0.339 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Annex 25, Page 24 Table A25.3.5 - Underemployment rate by poverty status Rural Urban Total Poverty Status Male Female Male Female Male Female Extremely poor 0.290 0.347 0.209 0.516 0.274 0.415 Poor 0.280 0.414 0.245 0.429 0.271 0.421 Non-poor 0.286 0.476 0.293 0.353 0.291 0.379 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.3.6 - Underemployment by consumption quintiles I Rural I Urban I Total Quintiles I Male Female I Male Female Male Female 0.300 0.361 0.234 0.448 0.286 0.396 2 0.269 0.477 0.262 0.415 0.267 0.446 3 0.254 0.422 0.273 0.379 0.263 0.391 4 0.297 0.465 0.286 0.348 0.290 0.376 5 0.307 0.538 0.303 0.350 0.304 0.375 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Table A25.3.7 - Underemployment by rural household typology Typology Male Female <2 manzanas 0.28 0.45 2-5 manzanas 0.31 0.62 5-20 manzanas 0.30 0.60 > 20 manzanas 0.28 0.41 non-ag; low education 0.26 0.25 non-ag; high education 0.26 0.44 Note: Rates are in proportions, so a rate of 0.125 means 12.5%. Annex 25, Page 25 Table A25.3.8 - Probit regression underemployment Variable All Female Male Education Grade 1-6 0.012561 0.0123 0.0128 Education Grade 7-9 0.006973 -0.0043 0.0118 Education Grade 10-12 0.022106 0.0180 0.0262 Education Grade 13-17 -0.021389 -0.0193 -0.0215 Age -0.026034** -0.0285** -0.0253** Age squared 0.000309** 0.0003** 0.0003** Children under 6 -0.002643 -0.0069 0.0000 Children 6-9 -0.013012* -0.0028 -0.0176** Children 10-14 0.009551 0.0208** 0.0046 Region: Pacific Urban -0.093315** -0.0593** -0.1 123** Region: Pacific Rural -0.017292 0.0068 -0.0344* Region: Central Urban -0.188603** -0.1817** -0.1899** Region: Central Rural -0.186153 * -0.154 1 ** -0.1979* * Region: Atlantic Urban -0.207113** -0.2047** -0.2053** Region: Atlantic Rural -0.082151** 0.0876 -0.1 155** Female 0.121554** No. Observations: 6976 2213 4763 LR Chi-Sq 511.78 117.24 327.6 Pseudo R-Sq: 0.059 0.040 0.058 Note: Coefficients are marginal effects. The underemployment rate in dependent variable is measured in proportions, so a rate of 0.125 means 12.5%. * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 26 APPENDIX A25.4 - DISTRIBUTION OF EMPLOYMENT BY SECTOR OF AQ7IVITY Table A25.4.1 - Distribution of the employed by activity and poverty status Rural Activity Extremely poor Poor Non-poor Total Agriculture 84.44 75.22 50.19 66.17 Mining 0.05 0.24 1.31 0.62 Manufacturing 3.66 4.64 7.58 5.71 Gas, Elec, Water 0.26 0.20 0.73 0.39 Construction 1.92 2.65 3.73 3.04 Commerce 2.81 6.47 17.26 10.37 Transport 0.53 1.42 2.50 1.81 Financial services 0.14 0.40 1.88 0.94 Community services 6.20 8.75 14.83 10.95 Total 100.00 100.00 100.00 100.00 Urban Activity Extremely poor Poor Non-poor Total Agriculture 40.52 23.51 6.52 11.03 Mining 0.54 0.69 0.43 0.50 Manufacturing 6.68 11.10 13.30 12.72 Gas, Elec, Water 0.21 0.19 1.21 0.94 Construction 5.41 6.86 5.80 6.08 Commerce 23.79 27.86 36.73 34.38 Transport 2.41 3.49 6.19 5.47 Financial services 0.35 2.06 4.45 3.81 Community services 20.10 24.24 25.35 25.06 Total 100.00 100.00 100.00 100.00 Annex 25, Page 27 Table A25.4.2 - Distribution of the employed by activity and consumption quintiles Rural Quintiles of Consumption Activity 1 2 3 4 5 Total Agriculture 80.74 71.58 54.84 51.48 43.41 66.17 Mining 0.08 0.32 1.48 0.95 1.16 0.62 Manufacturing 4.45 4.64 8.46 5.44 8.25 5.71 Gas. Elec, Water 0.34 0.08 0.27 0.76 1.34 0.39 Construction 2.50 2.49 3.58 4.54 2.88 3.04 Commerce 4.01 8.59 14.30 18.10 18.23 10.37 Transport 0.66 2.37 3.43 0.98 1.96 1.81 Financial services 0.12 0.40 2.22 1.44 2.01 0.94 Community services 7.10 9.53 11.41 16.30 20.77 10.95 Total 100 100 100 100 100 100 Urban Quintiles of Consumption Activity 1 2 3 4 5 Total Agriculture 36.78 19.99 9.73 5.60 5.33 11.03 Mining 0.41 1.18 0.72 0.53 0.07 0.50 Manufacturing 9.74 9.84 14.76 14.85 11.45 12.72 Gas, Elec, Water 0.16 0.00 1.37 1.40 0.86 0.94 Construction 5.08 7.18 5.92 7.21 5.10 6.08 Commerce 23.14 30.29 35.37 37.80 35.73 34.38 Transport 2.29 4.86 3.26 5.92 7.93 5.47 Financial services 2.16 0.80 4.32 2.14 6.57 3.81 Community services 20.22 25.86 24.56 24.55 26.96 25.06 Total 100 100 100 100 100 1o0 Annex 25 28 Table A25.4.3 - Distribution of the employed by activity and region (1993 and 1998) 1993 Activity Managua Pacific Pafirc Central Central Atlarttic Atlantic Total Urban Rural Urban Rural Urban Rural Agriculture 5.27 5.45 45.87 22.51 82.07 22.44 78.81 35.98 Mining 0.17 0.00 0.00 0.00 0.18 3.88 0.00 0.30 Manufacturing 12.46 25.76 10.56 13.03 2.86 17.17 1.90 11.53 Gas, Elec, Water 1.04 2.73 0.66 1.12 0.06 1.66 0.00 0.99 Construction 5.45 3.33 2.15 4.07 1.49 1.94 1.43 3.22 Commerce 29.72 26.57 20.63 24.34 6.02 21.88 7.38 19.97 Transport 5.97 4.85 2.97 3.87 0.36 3.88 1.90 3.47 Financial services 3.07 2.32 0.66 1.22 0.00 0.28 0.(0 1.37 Community services 36.85 28.99 16.50 29.84 6.97 26.87 8.57 23.15 Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 1998 Activity Managua Pacific Paciflc Central Central .1tliantic Atlontic Total Urban Rural Urban Rural Urban Rural Agriculture 8.74 10.96 53.73 19.29 76.49 27 52 83.71 36.20 Mining 0.15 0.00 0.56 0.79 0.15 4.12 2.68 0.56 Manufacturing 14.11 14.45 6.96 9.59 4.77 4.60 1.42 9.52 Gas, Elec, Water 0.86 0.70 0.87 1.13 0.22 0.95 0.00 0.69 Construction 6.41 4.99 3.82 6.09 3.21 4.82 0.66 4.69 Commerce 33.75 33.87 16.69 32.78 5.75 27.45 5.74 23.42 Transport 5.78 5.56 2.89 3.42 1.41 4.41 1.17 3.80 Financial services 5.92 2.09 0.85 2.73 0.74 1.06 0.11 2.50 Community services 24.28 27.4 13.63 24.18 7.26 25.07 4.52 18.62 Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Annex 25, Page 29 APPENDIX A25.5 - DISTRIBUTION OF EMPLOYMENT BY OCCUPATION Table A25.5.1 - Distribution of the employed by occupation and poverty status Rural Occupation Extremely poor Poor Non-poor Total Wage-employed 41.87 44.40 47.33 45.46 Self-employed or unpaid family 58.13 55.60 52.67 54.54 Total 100.00 100.00 100.00 100.00 Urban Occupation Extremely poor Poor Non-poor Total Wage-employed 62.52 62.71 60.26 60.91 Self-employed or unpaid family 37.48 37.29 39.74 39.09 Total 100.00 100.00 100.00 100.00 Table A25.5.2 - Distribution of the employed by occupation and consumption quintiles (%) Rural Quintiles of Consumption Occupation 1 2 3 4 5 Total Wage-employed 42.82 45.22 51.26 42.46 48.36 45.46 Self-employed or unpaid family 57.18 54.78 48.74 57.54 51.64 54.54 Total 100.00 100.00 100.00 100.00 100.00 100.00 Urban Quintiles of Consumption Occupation 1 2 3 4 5 Total Wage-employed 64.86 59.64 62.28 61.84 58.45 60.91 Self-employedorunpaidfamily 35.14 40.36 37.72 38.16 41.55 39.09 Total 100.00 100.00 100.00 100.00 100.00 100.00 Annex 25, Page 30 APPENDIX A25.6 - FORMAL/INFORMAL SECTOR CHOICE Table A25.6.1 - Proportion of the employed who are in formal jobs Rural Urban Total Male 0.30 0.49 0.39 Female 0.27 0.34 0.32 Total 0.29 0.43 0.36 Table A25.6.2 - Distribution of formal jobs by activity Rural Urban Total Activity Male Female Total Male Female Total Male Female Total Agriculture 0.25 0.45 0.27 0.39 0.63 0.42 0.27 0.49 0.29 Mining 0.62 0.62 0.37 0.00 0.34 0.50 0.00 0.48 Manufacturing 0.66 0.21 0.47 0.66 0.53 0.61 0.66 0.43 0.57 Gas, Elec, Water 0.90 1.00 0.91 0.98 1.00 0.98 1 0.96 1.00 0.96 Construction 0.30 0.30 0.41 0.00 0.40 0.37 0.00 0.37 Commerce 0.23 0.05 0.13 0.27 0.12 0.19 0.26 0.11 0.18 Transport 0.38 0.00 0.37 0.44 0.72 0.47 0.43 0.66 0.45 Financial services 0.43 0.00 0.29 0.67 0.71 0.68 0.63 0.54 0.61 Community services 0.67 0.34 0.43 0.81 0.48 0.59 0.78 0.44 0.55 Total 0.30 0.27 0.29 0.49 0.34 0.43 0.39 0.32 0.36 Table A25.6.3 - Distribution of formal jobs by age group Rural Urban Total Age (yrs) Male Female Male Female T Male Female 12-19 0.25 0.32 0.34 0.18 0.28 0.24 20-24 0.35 0.19 0.50 0.35 0.42 0.30 25-29 0.31 0.35 0.51 0.47 0.42 0.43 30-34 0.40 0.21 0.57 0.43 0.49 0.37 35-39 0.21 0.39 0.51 0.40 0.38 0.39 40-44 0.35 0.21 0.52 0.32 0.44 0.29 45-49 0.34 0.23 0.54 0.31 0.44 0.29 50-54 0.27 0.27 0.48 0.24 0.38 0.25 55-59 0.25 0.10 0.46 0.19 0.34 0.17 60-64 0.16 0.08 1 0.45 0.04 0.28 0.05 65 and over 0.24 0.28 | 0.30 0.14 0.26 0.20 Total 0.30 0.27 0.49 0.34 0.39 0.32 Annex 25, Page 31 Table A25.6.4 - Distribution of Formal Jobs by Region Rural Urban Total Activity Male Female Male Female Male Female Managua 0.42 0.31 0.59 0.37 0.56 0.37 Pacific Urban 0.47 0.30 0.47 0.30 Pacific Rural 0.35 0.21 0.35 0.21 Central Urban 0.39 0.33 0.39 0.33 Central Rural 0.28 0.34 0.28 0.34 Atlantic Urban 0.37 0.42 0.37 0.42 Atlantic Rural I 0.13 0.17 0.13 0.17 Total 0.30 0.27 0.49 0.34 0.39 0.32 Table A25.6.5 - Distribution of formal jobs by education Rural Urban Total Years of Education [ Male Female Male Female Male Female 0 0.21 0.20 1 0.36 0.12 0.25 0.16 1 - 6 0.31 0.22 0.40 0.17 0.34 0.19 7 - 9 0.44 0.31 0.49 0.37 1 0.48 0.36 10-12 0.59 0.55 0.66 0.58 1 0.65 0.57 13-17 0.46 0.67 0.76 0.82 0.72 0.80 Total 0.30 0.27 0.49 0.34 0.39 0.32 Table A25.6.6 - Distribution of formal jobs by poverty status Rural Urban Total Poverty Status Male Female Male Female Male Female Extremely poor 0.26 0.34 0.37 0.27 0.28 0.31 Poor 0.26 0.26 0.41 0.24 0.30 0.25 Non-poor 0.36 0.30 0.52 0.38 0.46 0.36 Total 0.30 0.27 0.49 0.34 0.39 0.32 Table A25.6.7 - Distribution of formal jobs by consumption quintiles Rural Urban Total Quintiles 1 Male Female I Male Female | Male Female 1 0.26 0.34 0.41 0.25 0.29 0.30 2 0.27 0.20 0.36 0.23 0.30 0.22 3 0.35 0.22 0.48 0.27 0.41 0.26 4 0.29 0.26 0.49 0.38 0.42 0.35 5 0.42 0.41 0.57 0.43 0.54 0.43 Total 0.30 0.27 0.49 0.34 0.39 0.32 Annex 25, Page 32 Table A25.6.8 - Probit regression for characteristics of formal sector jobs Variable All Male Female Education Grade 1-6 0.071907** 0.068375** 0.082804** 0.025686 Education Grade 7-9 0.197977** 0.191389** 0.171037** 0.22018** Education Grade 10-12 0.379302** 0.373358** 0.326666** 0.414896** Education Grade 13-17 0.488251** 0.483407** 0.38855** 0.603752** Age 20-24 0.107395** 0.110387** 0.113169** 0.093196 Age 25-29 0.129568** 0.132174** 0.158291** 0.046374 Age 20-34 0.153948** 0.154317** 0.137062** 0.164246** Age 35-39 0.172633** 0.172988** 0.205755** 0.105294* Age40-44 0.138981** 0.140231** 0.105649** 0.200228** Age 45-49 0.129749** 0.133064** 0.172403** 0.070114 Age 50-54 0.155922** 0.158157** 0.19476** 0.079835 Age 55-59 0.141623** 0.144178** 0.152496** 0.132287 Age 60-64 0.095985** 0.095191** 0.140081** -0.009999 Single -0.035289** -0.035147** -0.081007** 0.024914 Household Male -0.018562 -0.020666 -0.06561 * * -0.031205 Urban 0.052519** 0.049496** 0.096995 -0.040064* Share of dependents -0.105201** -0.095927** -0.094260** -0.120085 Female-headed household 0.019716 0.021464 -0.004177 0.031921 Quintile 2 -0.0511 ** Quintile 3 -0.023625 Quintile 4 -0.038943* Quintile 5 -0.033038 Female -0.132076** -0.133724** 6976 6976 4763 2213 LR Chi-Sq 846.98 839.34 479.99 417.93 Pseudo R-Sq: 0.093 0.092 0.076 0.1505 Note: Coefficients are marginal effects. * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 33 APPENDIX A25.7 - HOURS WORKED Table A25.7.1 - Hours worked per week by gender and rural/ urban location Extremely poor Poor I Non-poor I Total Rural Urban Total I Rural Urban Total j Rural Urban Total I Rural Urban Total Male 46.0 50.7 46.9 i 47.8 50.8 48.6 50.2 49.7 49.9 48.6 50.0 49.3 Female 46.6 41.4 44.5 45.6 44.0 44.8 42.8 47.7 46.7 44.3 46.8 46.0 Total 46.1 47.2 46.3 47.3 48.0 47.5 48.2 48.9 48.7 47.6 48.6 48.2 Table A25.7.2 - Hours worked per week by activity, 1993 and 1998 Activity Rural Urban Total 1993 1998 1993 1998 1993 1998 Agriculture 44.4 46.1 49.9 48.1 45.2 46.4 Mining 59.0 50.9 45.5 53.5 47.5 52.1 Manufacturing 40.8 43.8 44.8 47.3 44.1 46.4 Gas, Elec, Water 52.8 59.2 51.2 55.5 51.4 56.5 Construction 41.8 53.8 49.7 45.9 47.8 48.2 Commerce 52.5 50.6 56.4 50.6 55.5 50.6 Transport 52.4 60.8 53.5 52.3 53.3 54.1 Financial services 63.8 58.3 50.6 53.3 51.4 54.2 Community services 44.2 50.2 43.9 46.2 43.9 47.3 Total 45.2 47.6 49.1 48.8 47.4 48.2 Table A25.7.3 - Hours worked per week by region and activity Activity Region Agriculture Mining Manufacturing Gas, Elec, Construction Commerce Transport Financial Communini' Total Water Services Services Managua 42.7 67.0 46.6 56.5 38.2 48.1 52.6 55.0 42.7 46.3 Pacific Urban 44.9 48.0 58.2 51.8 51.6 49.3 55.2 47.4 49.2 Pacific Rural 44.1 49.1 42.7 64.7 51.7 48.1 55.5 75.6 52.6 46.9 Central Urban 49.4 59.1 44.9 50.4 50.1 51.3 54.7 51.3 49.3 49.9 Central Rural 47.8 48.0 48.2 44.4 58.6 60.9 72.3 31.2 55.1 49.6 Atlantic Urban 50.4 50.6 43.4 58.3 56.1 55.4 56.4 60.2 47.1 51.3 Atlantic Rural 45.5 48.0 35.6 66.4 50.0 55.0 40.0 40.5 45.7 Annex 25. Page 34 APPENDIX 8- WAGES Table A25.8.1 - Wages by gender, rural/ urban location and poverty status (CS per hour) Rural Urban Total Poverty Status Male Female Total Male Female Total Male Female Total Extremely poor 2.68 2.27 2.59 3.52 3.26 3.42 2.91 2.72 2.86 Poor 3.56 3.70 3.59 5.37 4.32 4.94 4.19 4.04 4.14 Non-poor 7.60 6.73 7.34 12.88 8.35 10.99 11.31 8.03 10.05 Total 5.10 5.09 5.10 10.90 7.32 9.41 8.30 6.69 7.73 Table A25.8.2 - Wages per Quintiles (C$ per hour) Rural Urban l Total Quintiles Male Female Male Female Male Female 1 2.85 2.56 3.43 3.22 3.01 2.85 2 4.27 4.33 5.66 4.56 4.75 4.46 3 5.21 5.63 7.01 4.83 6.23 5.04 4 6.59 5.71 8.10 5.27 7.61 5.37 5 13.56 10.19 20.65 13.14 19.48 12.76 Table A25.8.3 - Wages by formality of jobs Rural Urban Total Male Female Total Male Female Total Male Female Total Informal 5.10 4.84 5.03 9.66 5.93 7.86 7.34 5.59 6.65 Formal 5.11 5.67 5.22 12.00 9.60 11.19 9.39 8.65 9.17 Table A25.8.4 - Wages by activity (C$ per hour) Activity Rural Urban Total Agriculture 4.59 5.76 4.83 Mining 6.43 9.30 7.88 Manufacturing 5.87 7.37 6.98 Gas, Elec, Water 4.97 7.59 6.91 Construction 4.92 10.52 8.90 Commerce 5.67 8.38 7.85 Transport 8.05 17.27 15.26 Financial services 7.60 20.90 18.55 Community services 5.39 9.21 8.18 Annex 25, Page 35 Table A25.8.5 - Wages by education (C$ per hour) Years of Education Male Female Total 0 4.52 3.67 4.27 1 - 6 6.08 5.24 5.80 7 - 9 8.07 5.63 7.23 10-12 11.32 7.70 9.60 13-17 32.18 20.09 27.23 Table A25.8.6 - Wages by age group (CS per hour) Age (yrs) Male Female 12-19 4.03 3.52 20-24 4.58 5.06 25-29 5.74 6.27 30-34 9.16 7.82 35-39 8.51 6.39 40-44 12.41 8.39 45-49 10.33 9.80 50-54 11.88 5.84 55-59 16.43 6.57 60-64 11.24 4.83 Over 65 5.78 6.40 Table A25.8.7 - Wages by region (CS per hour) Region Male Female (i) otal Managua 13.51 9.63 12.00 Pacific Urban 7.84 5.87 6.99 Pacific Rural 5.17 5.52 5.27 Central Urban 9.09 6.02 7.78 Central Rural 3.75 3.38 3.67 Atlantic Urban 11.90 5.27 9.36 Atlantic Rural 5.69 3.89 5.41 Annex 25. 36 Table A25.8.8 - Wages by region and activity, 1993 and 1998 1993 (in 1998 priees,) Region giclue Mining Manufactutring' Gas, EIec, Construction iCommerce Transport Financial Commtunity Total Water Services Services Man'a g"ua 9.75 11811.56 16191.41 37 22.14 10.74 11.56 Pcfia Urbac 7.76 12.56 9.42 9.09 10.08 10.74 19.00 I1.1 07 .1.0 Pac i fi c Ru"r-a . 8.596.77 7.27 5.78 1 6.94 4.96 5.12 5 57.27 . 10.74 ~~~~~~~~~~~~~~~.613.22 I 15.20 13.22 6.77 9.58 Ce n t ral - -R u r al 3.63 2.159 9.09 7.10 4.46 0.00 5.62 4.46 Atlanti'c -Urb,-an .... H.8..... ji 4.30 12.06 109 .48.76 13.05 5.1 7.43 94 Ailan.tic Ru-r-a I" il.56 6.4 8.76 4.13 0.006.159 An 5~~~~~~~~~~~~'.78i61 69 .6 10.41 12.06 19.33 9.09 92 1998 Regi'on Agriculture Mining I anufacturing, Gas, Elec Construction Comre Tasot Financial Community Total I ~~~~~~~~~~~Water Services Services Ma'nagua ju.7 7.69 8.73 5.213.36 9.37 24.63 24.57 11i.03120 P"a"c"ifi'c Urban { 5',2'2 160847.5 6.21 8.96 9.35 1 8.06 6.99 N66c'k~a-f 4.899 J 41 6.62 4.96 11 4.28 I 6.07 10.14 3.01 4.23 5.27 C'entral Urban I .4 2.7" 5.899 10.43 1 5.86 8.906 11.32 1.1383977 - I I . I . I~~~~~~~~~~~~~~~~~~~~~~~~~... .. . ... ... Cent"ra"l R-u-ral- 3.43 6.67 4.42 4.98 5.14 I 2.95 2.69 1I 3.28 367 Atla-nti-cU"rb,a-n, 7.81 i12.0 j 83 8 14.46 106.50 16.02 8.62 7.21 9.36 Aii'n t'1 ' R7.18ai5.91 I 599.613.49 5.95.07 54 All 4~~~~~~.83 { 7.8,8 6.98 6.91 8.90 7.85 152185 1 8.873 Annex 25, Page 37 Table A25.8.9 - Regression for the determinants of wages (Heckman selectivity corrected wage regression) All Female Male Years of education 0.076834** 0.071083** 0.082601** Tenure 0.022483** 0.035198** 0.015988** Tenure squared -0.00045** -0.00071 ** *-0.00031** Urban -0.01838 -0.23443* 0.10582 Formal 0.117983** 0.182182** 0.077056 Formal x agric. -0.08271** -0.29544** -0.00073 Union -0.10507* * -0.20364* -0.05723 Region: Pacific Urban -0.33975** -0.2389** -0.4276** Region: Pacific Rural -0.44381** -0.35678** -0.50386** Region: Central Urban -0.37438 -0.26502** -0.45063** Region: Central Rural -0.78955 -0.85149** -0.7625** Region: Atlantic Urban -0.11468 -0.24763** -0.05268** Region: Atlantic Rural -0.63789** -0.51934** -0.63835** Female -0.21601** Intercept 1.555175** 1.190563** 1.498895 Log likelihood: No. Observations: 7014 3355 3355 Wald Chi-Sq 807.68 356.14 747.98 * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 38 APPENDIX A25.9 - MINIMUM WAGES Table A25.9.1 - Actual and minimum wage by activity Activity Actual Wage Minimum Wage Agriculture 4.70 1.70 Mining 7.25 3.41 Manufacturing 6.84 2.84 Gas, Elec, Water 6.80 3.41 Construction 8.61 2.73 Commerce 7.59 3.13 Transport 14.79 2.56 Financial services 18.13 3.98 Community services 7.95 2.67 Note: Wages in C$ per hour. Table A25.9.2 - Proportion with actual wages below minimum wage, by poverty status Poverty Status % wit/li wages below minimum Unemployment rate Extremely poor 0.286 0.123 Poor 0.292 0.126 Non-poor 0.215 0.120 Total 0.249 0.123 Note: Wages in C$ per hour. Table A25.9.3 - Average wage by activity of those whose wages are above minimum wage Activity Mean wages of those whose Minimum Wage wages are above minimum wage Agriculture 6.46 1.70 Mining 8.82 3.41 Manufacturing 8.34 2.84 Gas, Elec, Water 9.35 3.41 Construction 9.93 2.73 Commerce 11.40 3.13 Transport 17.06 2.56 Financial services 27.49 3.98 Community services 11.13 2.67 Note: Wages in C$ per hour. Annex 25, Page 39 Table A25.9.4 - Proportion with wages below minimum wage, by consumption quintiles Quintiles I 1 0.29 2 0.29 3 0.28 4 0.23 5 0.15 Table A25.9.5 - Proportion with actual wages below minimum wage, by activity Poverty Status Activity Extremely poor Poor Non-poor Agriculture 0.23 0.21 0.15 Mining 0.00 0.40 0.15 Manufacturing 0.30 0.27 0.19 Gas, Elec, Water 1.00 0.78 0.30 Construction 0.28 0.27 0.09 Commerce 0.45 0.39 0.29 Transport 0.07 0.17 0.14 Financial services 1.00 0.61 0.31 Community services 0.60 0.54 0.21 Annex 25, Page 40 APPENDIX A.25. 10- CHILD LABOR Table A25.10.1 - Number of child workers in Nicaragua Male 57,332.9 Female 17,551.7 Total 74,884.6 Note: Child workers are those aged 10-14 Table A25.10.2 - Incidence of child labor by rural/urban location Male Female Total Rural 0.27 0.06 0.16 Urban 0.08 0.05 0.06 Total 0.17 0.05 0.11 Table A25.10.3 - Hours worked per week by working children by rural/urban location Male Female Total Rural 32.37 27.88 31.60 Urban 29.93 25.12 28.14 Total 31.76 26.53 30.53 Table A25.10.4 - Incidence of child labor by rural household typology Typology Male Female Total < 2 manzanas 0.33 0.08 0.21 2-5 manzanas 0.43 0.04 0.24 5-20 manzanas 0.46 0.14 0.31 > 20 manzanas 0.34 0.04 0.20 non-agriculture; low 0.09 0.05 0.07 education non-agriculture; high 0.05 0.02 0.03 education Total 0.27 0.06 0.16 Table A25.10.5 - Hours worked per day by children by rural typology Typology Male Female Total < 2 manzanas 29.9 30.7 30.0 2-5 manzanas 32.0 20.9 31.1 5-20 manzanas 32.9 22.8 30.8 > 20 manzanas 36.4 51.4 37.8 non-ag; low education 37.0 31.3 35.1 non-ag; high education 24.0 19.1 22.2 Total 32.40 27.90 31.6 Annex 25. Page 41 Table A25.10.6 - Incidence of child grade-for-age by rural household typology Typology Male Female Total < 2 manzanas 33.60 47.43 39.97 2-5 manzanas 34.41 45.63 39.80 5-20 manzanas 36.59 39.31 37.84 > 20 manzanas 39.55 39.48 39.52 non-ag; low education 40.51 33.77 37.21 non-ag; high education 67.40 83.94 77.22 Total 41.82 51.61 46.67 Table A25.10.7 - Incidence of child labor among boys by poverty status Poverij Status Rural Urban Total Extremely poor 34.81 36.65 35.02 Poor 34.20 32.54 33.96 Non-poor 25.11 28.22 26.67 Total 32.37 29.93 31.76 Table A25.10.8 - Incidence of child labor by female headship (traditional definition) Headship Male Female Total Male-headed 0.18 0.05 0.11 Female-headed 0.14 0.06 0.10 Note: Traditional definition of headship is based on self-reporting. Table A25.10.9 - Incidence of child labor by female headship (new definition) New measure of headship Male Female Total Male & Female 0.18 0.05 0.11 Fem only 0.10 0.05 0.08 Male only 0.19 0.17 0.18 Other 0.07 0.08 0.07 Note: New definition of headship classifies households by the presence of prime-age males, prime age females, both, or neither. Table A25.10.10 - Incidence of child labor by poverty status Rural Urban | Total Poverty Status Male Female | Male Female Male Female Extremely poor 0.30 0.02 0.13 0.02 i 0.26 0.02 Poor 0.29 0.06 0.09 0.05 0.22 0.06 Non-poor 0.21 0.04 0.08 0.05 0.11 0.04 Annex 25, Page 42 Table A25.10.11 - Incidence of child labor by consumption quintiles Rural Urban i Total Quintiles 1 Male Female Male Female I Male Female 1 1 0.31 0.03 0.10 0.04 0.26 0.04 2 0.26 0.10 I 0.09 0.07 0.19 0.09 3 1 0.24 0.05 0.07 0.03 0.13 0.04 4 0.16 0.02 0.08 0.05 0.10 0.04 5 0.19 0.03 0.07 0.05 0.09 0.05 Table A25.10.12 - Incidence of child labor by activity Rural Urban Male Female Male Female Activity Percent Percent Percent Percent Agriculture 94.2 53.42 27.29 12.02 Mining 0.65 Manufacturing 1.44 13.34 3.83 3.34 Gas, Elec, Water Construction 0.48 3.4 Commerce 3.24 30.73 50.23 74.51 Transport Financial services Community services 2.51 15.24 10.13 Total 100.00 100.00 100.00 100.00 Table A25.10.13 - Incidence of child labor among boys by poverty status Poverty Status Rural Urban Total Extremely poor 0.14 0.19 0.15 Poor 0.12 0.13 0.12 Non-poor 0.27 0.04 0.16 Total 0.15 0.08 0.13 Annex 25, Page 43 Table A25.10.14 - Probit regression for determinants of child labor All Boys Girls Age 0.022303** 0.0358** 0.01 14** Education of Head -0.003495 -0.0041 -0.0030* Share of dependents -0.005499 -0.0670 0.0244 Share of adult females in 0.02064 -0.0100 0.0363 household Cluster wage -0.000052 -0.0003 0.0007 Female-headed household -0.001145 -0.0047 0.0011 Quintile 2 0.042028** 0.0260 0.0629** Quintile 3 0.015185 0.0159 0.0253 Quintile 4 0.014927 0.0048 0.0273 Quintile 5 0.058972** 0.0380 0.0699** Dummy for modern fuel -0.044979** -0.0764** -0.0216 Dummy for water tap -0.061428* * -0.0737** -0.0499** Female -0.09811 * * Region: Pacific Urban -0.012401 -0.0413 0.0127 Region: Pacific Rural -0.006939 0.0077 -0.0154 Region: Central Urban 0.037629* 0.0537 0.0252 Region: Central Rural 0.025477 0.1098** -0.0305** Region: Atlantic Urban -0.045769* -0.0563 -0.0303 Region: Atlantic Rural 0.06384** 0.2062** -0.0292 Log likelihood: No. Observations: 2692 1355 1337 LR Chi-Sq 230.89 151.66 51.76 Pseudo R-Sq: 0.126 0.127 0.091 Note: Coefficients are marginal effects. * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 44 APPENDIX A25.11 - CHILD LABOR AND EDUCATIONAL ATTAINMENT Table A25.11.1 - Distribution of child school attendance and work combination, by poverty status Rural Urban Schooll Work Extremely Poor Non-poor Total Extremely Poor Non-poor Total Combination poor poor School only 26.68 68.32 31.68 100.00 7.36 32.15 67.85 100.00 56.62 64.44 81.89 69.10 68.01 79.78 90.50 86.75 School and work 26.71 70.26 29.74 100.00 2.00 29.71 70.29 100.00 7.36 8.60 9.98 8.97 1.04 4.15 5.27 4.88 Work only 46.50 92.77 7.23 100.00 39.05 66.73 33.27 100.00 10.40 9.22 1.97 7.29 6.61 3.03 0.81 1.59 Neither 56.97 88.75 11.25 100.00 33.69 67.25 32.75 100.00 25.62 17.74 6.16 14.64 24.34 13.04 3.41 6.78 Total 32.57 73.27 26.73 100.00 9.39 34.96 65.04 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Table A25.11.2 - Distribution of child school attendance and work combination, by poverty status Male Female Schooll Work Extremely Poor Non-poor Total Extremely Poor Non-poor Total Combination poor poor School only 15.97 48.08 51.92 100.00 14.86 46.37 53.63 100.00 53.18 65.06 83.68 73.56 66.87 74.78 92.48 83.34 School and work 22.83 54.50 45.50 100.00 4.92 55.81 44.19 100.00 9.73 9.44 9.38 9.41 1.11 4.52 3.82 4.18 Work only 48.12 88.92 11.08 100.00 22.41 78.37 21.63 100.00 16.33 12.26 1.82 7.50 1.25 1.56 0.46 1.03 Neither 48.12 75.49 24.51 100.00 49.76 86.36 13.64 100.00 20.76 13.23 5.12 9.53 30.77 19.14 3.23 11.45 Total 22.09 54.36 45.64 100.00 18.52 51.67 48.33 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Annex 25, Page 45 Table A25.11.3 - Distribution of educational attainment by child school attendance and work combination and poverty status Poverty Status Schooll Work Combination Extremely poor Poor, Non Non-poor Total School only 2.13 2.90 3.72 3.22 School & Work 1.88 2.42 3.84 2.96 Work Only 1.00 1.51 1.76 1.31 Neither 0.93 1.78 1.58 1.32 Total 1.71 2.67 3.62 2.92 Table A25.11.4 - Distribution of grade-for-age by urban and school attendance and work combination Schooll Work Combination Rural Urban Total School only 55.3 72.1 65.1 School & Work 48.6 65.5 55.0 Work Only 18.3 33.7 21.3 Neither 19.3 31.1 23.3 Total 46.7 68.4 58.2 Annex 25, Page 46 APPENDIX A25.12 - ADULT TIME USE Table A25.12.1 - Hours spent per day in all work activities, by poverty status Rural Urban Total Poverit Status Male Female Male Female Male Female Extremely poor 8.39 8.59 8.06 7.96 8.31 8.43 Poor 8.59 7.75 7.81 8.65 8.29 8.15 Non-poor 7.98 7.76 6.97 8.31 7.26 8.18 Total 8.30 7.97 7.19 8.36 7.69 8.20 Table A25.12.2 - Proportion working in income-generating activities, by poverty status Rural Urban Total Poverty Status Male Female Male Female Male Female Extremely poor 0.83 0.21 0.73 0.36 0.81 0.25 Poor 0.83 0.23 0.70 0.39 0.78 0.29 Non-poor 0.74 0.26 0.63 0.41 0.67 0.38 Total 0.79 0.24 0.65 0.41 0.72 0.34 Table A25.12.3 - Hours spent per day in income-generating activities, by poverty status Rural | Urban I Total Poverty Status Male Female Male Female Male Female Extremely poor 8.12 7.13 9.45 6.91 8.38 7.05 Poor 8.30 7.15 9.18 7.33 8.56 7.25 Non-poor 8.54 7.79 9.03 8.34 8.87 8.25 Total 8.39 7.41 9.07 8.1 8.73 7.91 Table A25.12.4 - Hours spent per day in household chores, by poverty status I Rural Urban Total Poverty Status Male Female Male Female Male Female Extremely poor 1.62 7.07 1.14 5.47 1.52 6.65 Poor 1.65 6.46 1.44 5.66 1.58 6.16 Non-poor 1.62 5.70 1.24 4.86 1.35 5.06 Total 1.64 6.18 1.28 5.06 1.45 5.50 Table A25.12.5 - Share of household chores in all work, by poverty status Rural Urban Total Poverty Status Male Female Total Male Female Total Male Female Total Extremely poor 0.23 0.86 0.55 0.26 0.74 0.53 0.24 0.83 0.54 Poor 0.24 0.85 0.55 0.29 0.73 0.54 0.25 0.80 0.54 Non-poor 0.29 0.81 0.55 0.31 0.68 0.52 0.31 0.71 0.53 Total 0.26 0.83 0.55 0.31 0.69 0.53 0.28 0.75 0.54 Note: A share of I implies 100% and 0 implies 0%. Annex 25, Page 47 Table A25.12.6 - Hours spent per day in water collection Rural Urban Total Poverty Status Male Female Total Male Female Total Male Female Total Extremely poor 1.22 1.21 1.21 1.20 0.95 1.02 1.22 1.16 1.18 Poor 1.26 1.20 1.22 1.25 1.14 1.17 1.26 1.18 1.21 Non-poor 1.13 1.02 1.06 1 1.36 1.36 1.36 1.21 1.18 1.19 Total 1.22 1.15 1.17 1.30 1.26 1.27 1.24 1.18 1.20 Table A25.12.7 - Hours spent per day in water collection by distance to water Rural Urban Total Distance to water Male Female Total M Male Female Total Male Female Toia/ <.025 kms 0.97 1.16 1.09 1.22 1.26 1.25 1.02 1.19 1.1. 0.025-0.10 kms 1.07 1.08 1.08 1.02 0.94 0.96 1.06 1.06 1.06 0.10-0.20 kms 1.29 1.24 1.25 2.22 1.55 1.78 1.51 1.29 1.35 0.2-0.4 kms 1.34 1.11 1.19 1.04 1.12 1.09 1.28 1.11 1.17 >0.4 kms 1.63 1.57 1.59 1.24 1.53 1.42 1.58 1.57 1.57 Total 1.22 1.21 1.21 1.39 1.27 1.31 1.25 1.22 1.23 Table A25.12.8 - Share of water collection in all work, by poverty status Rural Urban Total Poverty Status Male Female Total Male Female Total Male Female Total Extremely poor 0.04 0.07 0.05 0.05 0.05 0.05 0.04 0.07 0.05. Poor 0.04 0.08 0.06 0.02 0.03 0.02 0.03 0.06 0.04 Non-poor 0.03 0.04 0.04 0.01 0.01 0.01 I 0.02 0.02 0.02 Total 0.03 0.06 0.05 0.01 0.02 0.01 0.02 0.04 0.0.3 Note: A share of ] implies 100% and 0 implies 0%. Table A25.12.9 - Hours spent per day in fuel collection activities by those who collect Rural Urban Total Poverty Status Male Female Total I Male Female Total Male Female Total Extremely poor 1.60 1.45 1.55 2.02 1.22 1.58 1.64 1.40 1.56 Poor 1.73 1.34 1.62 I 2.31 1.49 1.88 1.79 1.38 1.66 Non-poor 1.68 1.17 1.56 2.34 1.94 2.24 1.82 1.33 1.7 Total 1.72 1.30 1.60 2.32 1.59 2.01 1.80 1.36 1.67 Annex 25, Page 48 Table A25.12.10 - Hours spent per day in fuel collection activities by distance to fuel source Rural Urban i Total Distance Male Female Total Male Female Total Male Female Total <0.1 km 1.47 0.94 1.30 2.55 0.84 1.61 1 1.79 0.89 1.42 0.1 - 0.5kms 1.49 1.29 1.44 1.29 0.49 0.97 1.48 1.22 1.42 0.5- I.Okms 1.59 1.38 1.53) 2.20 1.72 2.01 1.66 1.44 1.6 1.0 - 2.Okms 1.86 1.31 1.69 2.09 1.68 1.92 1.88 1.37 1.71 >2.0 kms 2.10 1.33 1.91 2.70 2.38 2.61 2.23 1.57 2.06 Total 1.72 1.29 1.6 2.36 1.44 2 1.8 1.32 1.66 Table A25.12.11 - Share of fuel collection in all work, by poverty status Rural Urban Total Poverty Status Male Female Total Male Female Total Male Female Total Extremely poor 0.08 0.03 0.05 0.03 0.02 0.03 0.07 0.03 0.05 Poor 0.07 0.02 0.05 0.02 0.01 0.02 0.06 0.02 0.04 Non-poor 0.05 0.01 0.03 0.01 0.00 0.00 0.02 0 0.01 Total 0.07 0.02 0.04 0.01 0.00 0.01 0.04 0.01 0.02 Note: A share of I implies 100% and 0 implies 0%. Table A25.12.12 - Hours spent per day in childcare by quintiles (hours) Female Poverty Status Rural Urban Total Extremely poor 2.70 2.47 2.64 Poor 2.92 2.82 2.88 Non-poor 2.70 3.13 3.03 Total 2.85 3.04 2.96 Table A25.12.13 - Share of childcare in all work, by poverty status Rural Urban Total Poverty Status Male Female Total Male Female Total Male Female Total Extremely poor 0.01 0.14 0.07 0.03 0.14 0.09 0.01 0.14 0.08 Poor 0.02 0.13 0.07 0.05 0.13 0.1 0.03 0.13 0.08 Non-poor 0.03 0.1 0.06 0.03 0.12 0.08 0.03 0.11 0.08 Total 0.02 0.12 0.07 0.04 0.12 0.08 0.03 0.12 0.08 Note: A share of I implies 100% and 0 implies 0%. Annex 25, Page 49 APPENDIX A25.13 - TIME USE OF ADULTS - REGRESSION RESULTS Table A25.13.1 - Determinants of hours spent by adults in all work activities All Male Female Age 0.487825** 0.4550** 0.5065** Age squared -0.005988** -0.0057** -0.0062** Dummy for interview on -0.899155** -1.4804** -0.3702* Sunday/Monday Years of education -0.064827** -0.0808** -0.0463* Dummy for modern fuel -0.446201 ** -0.7905** -0.1508 Dummy for water tap 40.312392* * -0.0468 -0.5513 ** Children under 6 0.533626** 0.2881** 0.7331 ** Children 6-9 0.189686** 0.1722 0.1909** Children 10-14 -0.159872** -0.1314 -0.1882** Single -I.166373** -1.5966** -1.0012** Female-headed household 1.396619** -0.6775 1.2752** (new definition) Prime-age females per -0.350609 -0.5650 -0.3862 capita in household Walls: adobe 0.541735** 0.3821 0.6654** Walls: wood 0.215372 0.2669 0.1209 Floor: wood -0.300291 -0.2977 -0.2740 Floor: dirt 0.35936** 0.6656** 0.0612 Ceiling: clay 0.091921 0.2385 -0.1015 Ceiling: straw 0.166494 0.0223 0.2555 Cluster wage :0.000802 -0.0005 -0.0011 Quintile 2 0.175507 0.2047 0.1167 Quintile 3 0.280886 0.2327 0.2400 Quintile 4 0.328588 0.3721 0.1796 Quintile 5 0.754674** 0.6743* 0.7175** Region: Pacific Urban -0.613338** -0.7077** -0.5258** Region: Pacific Rural -0.952199** -0.5435* -1.3728** Region: Central Urban -0.516373** -0.3057 -0.7099** Region: Central Rural -0.423101** 0.3596 -1.3012** Region: Atlantic Urban -0.460106** 0.3380 -1.1390** Region: Atlantic Rural 0.040421 0.7188 -0.5859 Female 0.36223** Intercept 0.185277 1.0308 0.3121 No. Observations: 5232 2540 2692 * significant at 5 percent. ** significant at 10 percent. Adults means household members between 15 and 60 years old. Annex 25, Page 50 Table A25.13.2 - Determinants of hours spent by adults in housework activities and chores All Male Female Age 0.0219499 -0.0787** 0.0816** Age squared -0.0002303 0.0011** -0.0010* Dummy for interview on -0.0803761 0.2438 -0.2313 Sunday/Monday Years of education -0.0567617** 0.0283 -0.0885** Dummy for modem fuel -0.1684636 -0.3133* -0.0954 Dummy for water tap -0.5281047** -0.0316 -0.7806** Children under 6 0.3339056** 0.0483 0.4696** Children 6-9 0.0976611 -0.1094 0.1731 ** Chiidren 10-14 -0.0894101 -0.0966 -0.1067 Single -0.8454329** -0.2122 -1.1640** Female-headed household -0.2140271 0.1252 -0.0173 (new definition) Prime-age females per -1.53718** -0.1362 -1.5061** capita in household Walls: adobe 0.5726406** 0.6928** 0.4805 Walls: wood 0.255851* 0.6335** 0.1202 Floor: wood 0.1197256 -0.1113 0.1947 Floor: dirt -0.0594592 0.1071 -0.0940 Ceiling: clay -0.1287038 -0.2006 -0.1251 Ceiling: straw 0.2868535 0.0256 0.4754 Cluster wage -0.0003211 0.0008 -0.0017 Quintile 2 -0.0171725 0.3129 -0.1573 Quintile 3 0.1969994 -0.0389 0.3861 Quintile 4 -0.0966027 -0.1634 -0.0298 Quintile 5 0.1781208 -0.3341 0.4760 Region: Pacific Urban -0.2408547 -0.0958 -0.3011 Region: Pacific Rural -0.5060413** -0.3151 -0.5807** Region: Central Urban -0.2501532 -0.3316 -0.1765 Region: Central Rural -0.2110464 -0.1720 -0.0569 Region: Atlantic Urban -0.3781669 0.5711 -0.6998* Region: Atlantic Rural 0.1633467 -0.6606 0.6282 Female 3.290564** Intercept 3.984265** 4.1493** 6.4957** No. Observations. 3666 1250 2416 * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 51 Table A25.13.3 - Determinants of hours spent by adults on water collection All Male Female Age -0.017897 -0.0206 0.0003 Age squared 0.000174 0.0001 -0.0001 Dummy for interview on 0.099473 0.1289 0.1584* Sunday/Monday Years of education -0.02748** -0.0435** -0.0172** Dummy for modern fuel 0.303476** 0.2650 0.3420* Dummy for water tap -0.142513 -0.0274 -0.1987 Children under 6 -0.00761 -0.0617 -0.0090 Children 6-9 0.03407 -0.0544 0.0679* Children 10-14 0.051948* 0.1017* 0.0154 Single 0.027741 -0.1795 0.1457* Female-headed household -0.157604 0.9490 -0.1615 (new definition) Prime-age females per 0.213898 0.4631 -0.0160 capita in household Walls: adobe 0.304772** 0.1148 0.3505** Walls: wood 0.220042** 0.2868** 0.1930** Floor: wood -0.169357 -0.2924 -0.1 171 Floor: dirt -0.2774** -0.2562 -0.2985** Ceiling: clay -0.029628 0.1201 -0.0724 Ceiling: straw -0.025431 -0.0899 -0.0003 Cluster wage 0.00038 -0.0016 0.0023 Quintile 2 0.015956 -0.0593 0.0508 Quintile 3 0.004275 -0.0238 0.0077 Quintile 4 0.033333 -0.0832 0.0187 Quintile 5 0.23632 0.3829 0.0936 Region: Pacific Urban 0.119936 -0.0935 0.2120 Region: Pacific Rural 0.041302 -0.0182 0.0687 Region: Central Urban 0.148785 0.4012 0.0599 Region: Central Rural 0.008025 0.1862 -0.0613 Region: Atlantic Urban 0.021169 0.2817 -0.0976 Region: Atlantic Rural 0.042311 -0.2001 0.1226 Female -0.059103 Water source: well 0.071538 0.1233 0.0398 Water source: river 0.105489 -0.0115 0.1757 Intercept 1.450925** 1.6733** 1.1620** No. Observations 938 287 651 * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 52 Table A25.13.4 - Determinants of hours spent by adults on fuel collection All Age -0.0325 Age squared 0.00032 Dummy for interview on Sunday/Monday -0.0921 Years of education -0.0058 Dummy for water tap 0.30962** Distance to source of wood 0.16861** Children under 6 -0.0903* Children 6-9 -0.0685 Children 10-14 0.04846 Single -0.0565 Female-headed household (new definition) 0.20088 Prime-age females per capita in household -0.2151 Walls: adobe 0.47963** Walls: wood 0.15234 Floor: wood 0.01527 Floor: dirt 0.11387 Ceiling: clay -0.0897 Ceiling: straw 0.05781 Cluster wage 0.00412 Quintile 2 0.07572 Quintile 3 -0.07 Quintile 4 -0.0176 Quintile 5 -0.8978** Region: Pacific Urban -0.0182 Region: Pacific Rural 0.18807 Region: Central Urban 0.35102 Region: Central Rural -0.0033 Region: Atlantic Urban 0.66147 Region: Atlantic Rural 0.23784 Female -0.4282** Intercept 2.18008** No. Observations 534 * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 53 Table A25.13.5 - Determinants of hours spent by adults on childcare Variable Female Age -0.02864 Age squared -3.6E-05 Dummy for interview on Sunday/Monday -0.01715 Years of education -0.03779 Dummy for modern fuel 0.360025 Dummy for water tap -0.04902 Children under 6 0.374956** Children 6-9 -0.12819 Children 10-14 -0.01456 Single -0.35195 Female-headed household (new definition) 1.053926** Prime-age females per capita in household -1.82842** Walls: adobe 0.296104 Walls: wood -0.07899 Floor: wood -0.70155 Floor: dirt -0.26983 Ceiling: clay -0.17408 Ceiling: straw -0.0379 Cluster wage -0.00174 Quintile 2 0.526028* Quintile 3 0.154908 Quintile 4 -0.24063 Quintile 5 0.448536 Region: Pacific Urban -0.40656 Region: Pacific Rural -0.30017 Region: Central Urban -0.73481** Region: Central Rural -0.71808** Region: Atlantic Urban 0.010396 Region: Atlantic Rural 0.439108 Intercept 5.850321 ** No. Observations 1045 * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 54 APPENDIX A25.14 - TIME USE OF CHILDREN (10-14 YEARS OLD) Section A25.14.1 - Hours spent by children per day on all work activities Rural Urban Total Poverty Status Bov Girl Boy Girl Boy Girl Extremely poor 3.37 3.84 3.03 2.38 3.30 3.44 Poor 3.33 3.58 1.76 2.39 2.74 3.17 Non-poor 2.29 2.88 1.52 1.58 1.71 1.93 Total 3.05 3.37 1.61 1.83 2.25 2.53 Rural Urban Total Quintiles Boy Girl Boy Girl - Boy Girl 1 3.28 3.91 2.81 2.35 3.18 3.48 2 3.58 3.32 1.51 2.69 2.60 3.09 3 2.98 2.68 1.56 1.82 1.99 2.12 4 1.55 3.60 1.44 1.79 1.47 2.31 5 0.86 1.91 1.26 1.16 1.22 1.29 Table A25.14.2 - Hours spent by children per day in income-generating activities Boy Poverty Status Rural Urban Total Extremely poor 5.72 7.77 6.06 Poor 5.45 5.63 5.49 Non-poor 5.33 4.86 5.09 Total 5.42 5.19 5.35 Boy Quintiles Rural Urban Total 1 5.45 7.77 5.77 2 5.45 4.66 5.22 3 6.41 4.32 5.53 4 3.61 4.30 4.03 5 2.31 6.49 5.04 Annex 25, Page 55 Table A25.14.3 - Proportion of children who do household chores Rural Urban Total Poverty Status i Boy Girl t oy Girl Boy Girl Extremely poor 1 0.63 0.88 0.38 0.72 0.58 0.84 Poor 0.65 0.81 0.47 0.75 0.58 0.79 Non-poor 0.53 0.75 0.56 0.64 I 0.55 0.67 Total 0.62 0.79 0.53 0.67 0.57 0.73 Rural Urban i Total Quintiles Boy Girl Boy Girl Boy Girl I 1 0.64 0.87 0.42 0.72 0.59 0.83 2 0.69 0.74 0.55 0.81 1 0.63 0.77 3 1 0.51 0.69 0.54 0.65 0.53 0.66 4 0.59 0.80 0.48 0.64 0.52 0.69 5 0.45 0.84 0.60 0.64 0.58 0.67 Table A25.14.4 - Share of household chores in all work by children Rural Urban Total Poverty Status Boy Girl Boy Girl Boy Girl Extremely poor 0.74 0.96 0.66 0.98 0.72 0.97 Poor 0.75 0.92 0.82 0.95 0.77 0.93 Non-poor 0.69 0.97 0.89 0.96 0.84 0.96 Total 0.73 0.94 0.87 0.96 0.8 0.95 Note: A share of I implies 100% and 0 implies 0%. Table A25.14.5 - Proportion of children attending school Rural Urban Total Poverty Status i Boy Girl Boy Girl Boy Girl Extremely poor 0.46 0.43 0.33 0.45 0.43 0.43 Poor 0.48 0.52 0.59 0.59 0.52 0.55 Non-poor 0.37 0.41 0.66 0.70 0.59 0.62 Total 0.45 0.49 0.64 0.67 0.55 0.58 Rural Urban Total Quintiles Boy Girl Boy Girl Boy Girl 1 0.46 0.46 0.39 0.45 0.44 0.45 2 0.50 0.59 0.68 0.67 0.59 0.62 3 0.26 0.38 0.64 0.70 0.52 0.59 4 0.65 0.44 0.63 0.68 0.63 0.61 5 0.29 0.58 0.74 0.71 0.69 0.69 Annex 25, Page 56 APPENDIX A23.15 - TIME USE OF CHILDREN (10-14) - REGRESSION RESULTS Table A25.15.1 - Determinants of hours spent by children in all work activities All Male Female Age 0.18972 0.9257 -0.0004 Age squared 0.005067 -0.0239 0.0130 Dummy for interview on 0.23886 0.2165 0.3052 Sunday/Monday Education of head -0.073272** -0.0375 -0.0703* Dummy for modern fuel -0.116181 -0.8481* 0.3615 Dummy for water tap -0.490605* * -0.3369 -0.6718* * Distance to school 0.157606** 0.2742** 0.0415 Children under 6 0.271013** 0.2737 0.2317* Children 6-9 0.051072 0.0989 0.0001 Children 10-14 -0.34164** -0.5089** -0.1427 Female-headed household 0.77999** 0.9342** 0.7356 (new definition) Prime-age females per -1.18573** -1.5442** -1.0114 capita in household Walls: adobe 0.400466 -0.2144 1.0845** Walls: wood 0.044924 0.0634 0.2310 Floor: wood 0.003045 0.5122 -0.2874 Floor: dirt 0.689044** 0.7800* 0.6218 Ceiling: clay -0.399953 -0.8527** -0.0674 Ceiling: straw 0.042883 -0.5055 0.6334 Cluster wage :0.001062 -0.0009 -0.0042 Quintile 2 0.349531 0.1119 0.5580 Quintile 3 0.512274 0.3562 0.6102 Quintile 4 0.651776 0.3228 0.8878 Quintile 5 -0.141231 -0.2544 -0.1790 Region: Pacific Urban -0.022084 0.2110 -0.2796 Region: Pacific Rural 0.098309 0.0815 -0.0009 Region: Central Urban 0.677316* 1.9309** -0.3604 Region: CentralRural 1.253412** 2.0632** 0.3017 Region: Atlantic Urban 0.742444 1.2595 0.1398 Region: Atlantic Rural 1.906764 2.0920** 1.2214* Female 0.084105 Intercept 0.614628 -3.8139 1.6277 No. Observations 910 428 482 * significant at 5 percent. ** significant at 10 percent. Annex 25, Page 57 Table A25.15.2 - Determinants of hours spent by children in housework activities and chores All Male Female Age 1.7439 1.0417 Age squared -0.033413 -0.0683 -0.0291 Dummy for interview on 0.240097 0.1314 0.2736 Sunday/Monday Education of head -0.058989** -0.0025 -0.0666* Dummy for modem fuel 0.023778 -0.4082 0.1923 Dummy for water tap -0.396089* -0.2260 -0.6093** Distance to school 0.030841 0.0599 0.0528 Childrenunder6 0.165479* 0.1210 0.1855* Children 6-9 -0.034034 -0.0989 0.0298 Children 10-14 -0.174209* -0.2858* -0.0907 Female-headed household 1.01 1988** 1.1670** 0.8506** (new definition) Prime-age females per -0.897366* -1.4226* -0.3662 capita in household Walls: adobe 0.205846 -0.6430 0.7398 Walls: wood 0.047889 -0.0419 0.0374 Floor: wood 0.147579 1.0970* -0.1676 Floor: dirt 0.421997* 0.6662* 0.4270 Ceiling: clay 0.188924 -0.1801 0.4353 Ceiling: straw 0.149583 -0.5900 0.7528 Cluster wage -0.000432 -0.0005 0.0015 Quintile 2 0.219059 -0.3621 0.6401* Quintile 3 0.204702 -0.6268 0.8299 Quintile 4 0.317152 -0.5558 1.0492* Quintile 5 -0.452487 -0.7891 -0.1853 Region: Pacific Urban -0.323742 0.1186 -0.7074 Region: Pacific Rural -0.172925 0.1463 -0.4918 Region: Central Urban 0.227999 1.2038** -0.4512 Region: Central Rural 0.60286* 0.7704 0.2801 Region: Atlantic Urban 0.566011 0.7518 0.3373 Region: Atlantic Rural 1.19792** -0.2118 1.6304** Female 0.813123** Intercept -4.665032 -7.5084 -5.4446 No. Observations 852 376 476 * significant_at percent * significant at 10 percent Annex 25, Page 58 REFERENCES Fafchamps, Marcel and Agnes R. Quisumbing. 1998. "Social Roles, Human Capital and the Intra household Division of Labor: Evidence from Pakistan." Palo Alto: Department of Economics, Stanford University. Mimeograph. Greene, William. 1997. Econometric Analysis, Prentice Hall: New Jersey. 3rd Edition. Haddad, Lawrence, John Hoddinott and Harold Alderman, eds. 1997. Intra-household Resource Allocation in Developing Countries. Baltimore: The Johns Hopkins University Press. Ilahi, Nadeem and Franque Grimard. 2000. "Public Infrastructure and Private Costs: Water Supply and Time Allocation of Women in Rural Pakistan." Economic Development and Cultural Change, forthcoming. Ilahi, Nadeem, 1999, "The Intra-household Allocation of Time and Tasks: What Have We Leamt from the Empirical Literature?" Background Paper for the Engendering Development. Washington, D.C.: The World Bank. Ilahi, Nadeem, Guilherme Sedlacek and Masaru Sasaki. 2000. Child Labor, Poverty, and Education: Empirical Evidence from some Latin American Countries," Washington, D.C.: Human Development Department, Latin America and Caribbean Region, the World Bank. Mazumdar, Dipak. 1975. "The urban informal sector." Staff working paper no. 21 1. Washington. D.C.: The World Bank. Patrinos, Harry A., and George Psacharopoulos, "Family Size, schooling and child labor in Peru-An empirical analysis." Journal of Population Economics, 10: 387-405. The World Bank. 2000. "Panama Poverty Assessment." Washington, D.C.: The World Bank.