33124 INTERNATIONALBANK FOR WORLD BANK R E T C N O E N STRUCTION PM AND DEVELO February 2005 No. 63 A regular series of notes highlighting recent lessons emerging from the operational and analytical program of the World Bank`s Latin America and Caribbean Region MECOVI IMPROVING THE SURVEY AND MEASUREMENT OF LIVING CONDITIONS IN LATIN AMERICA AND THE CARIBBEAN Gilberto Moncada and Haeduck Lee Background Management (PREM), provides the institutional home for the The MECOVI program1 was launched in 1996, as a joint World Bank's contribution to the MECOVI Program2. initiative of the Inter-American Development Bank (IDB), the World Bank, and the United Nations Economic Commis- Household Surveys before the MECOVI Program sion for Latin America and the Caribbean (UN-ECLAC), driven by the need to improve the statistical knowledge base Household surveys are important for public policy They on households poverty and living conditions at the household provide information on poverty and social conditions and, on level in the region. The program has two principal compo- the determinants of these conditions. nents: (i) activities at the country level; and (ii) regional activities (training, workshops and organization of the house- When the MECOVI program was initiated, the program of hold survey database). household surveys in the region was not functioning at a level ap- The MECOVI program has a proaching its potential. The Steering Committee (SC) suveys typically producted low and a Coordination Commit- quality information, had low tee. The SC members are: the utilisation, lacked relevance and IDB's Vice-president, the were carried out on the back of Vice-president for Latin low budgets. As a consequence, America and the Caribbean statistical errors were widespread, from the World Bank, and there was incomplete thematic and ECLAC's Executive Secre- geographic coverage, publication tary. The SC functions are to of results was chronically behind take decisions on and outline schedule. The access to informa- the major policy factors, or- tion for research and public policy ganization and resource allo- decision making was limited. cation. The Coordination Committee members are responsible for the administration of To a great extend, donors were the MECOVI operations in each of the aforementioned part of the problem, and not part of the solution. Many of institutions. them were more interested in pushing their agendas by sponsoring the surveys to satisfy their immediate data needs Funding comes from the sponsoring organizations to comple- than in strengthening capacities of the National Statistical ment national budget allocations. MECOVI has counted on Offices (NSO). financial contributions from bilateral and/or foundations in- cluding: Germany, Canada, Denmark, United States, Japan, Program Objectives and Strategies Norway, UNDP, Sweden, and the Soros Foundation. To respond to these challenges, the MECOVI program set out The Poverty and Gender Sector, Latin American and the to improve the quality, relevance and timeliness of informa- Caribbean Region (LAC), Poverty Reduction and Economic tion from household surveys on living conditions and pov- 1 erty, and also to broaden access and promote poverty close user-producer relations through an inter-institutional research and analysis. committee and adopted an open data dissemination policy (see Box 1). The MECOVI program developed a strategy which aimed to: The Program started with the participation of four countries, (i) establish a multi-year program rather than a project with (Argentina, El Salvador, Paraguay, and Peru) in 1997. As of short term goals; 2004, 11 countries in Latin America have benefited from the (ii) develop or strengthen existing institutional capacities; country-specific MECOVI program, with the addition of (iii) target specific interventions to problem areas; Bolivia, Colombia, Ecuador, Guatemala, Honduras, Nicara- (iv) provide sponsors and donors' financial support accord- gua and the Dominican Republic to the list of beneficiary ing to countries' capacities and comparative advantages; countries3. The MECOVI program has made substantial (v) provide sufficient resources to guarantee an appropriate contributions to the design and implementation of the sur- level of supervision and coordination of program activi- veys, has processed the survey data, and organized the ties; dissemination of the survey data as well as the results. The (vi) ensure close coordination among the sponsors. MECOVI program also contributed to the establishment and consolidation and regularization of household surveys for Program Activities and Results most participating countries4. There are also are two projects under execution to improve the household surveys in the The MECOVI program has provided the National Statistical Caribbean countries5. Offices (NSO) of the beneficiary countries with a package of capacity building resources to finance consulting services for At a regional level, the program has organized 14 workshops technical assistance, upgrade IT capability, organize work- and 12 training courses6, to improve technical and method- shops and training courses, and support research activities ological aspects of household surveys and poverty measure- through a small grant component. The program also achieved ment. The regional data bank component has assembled and organized the regional database of 320 mi- cro-data of household surveys to support Box 1: analysis and evaluation of poverty and socio- Experience with Household Surveys in Peru: the MECOVI Program economic conditions. The National Institute of Statistics and Informatics (INEI) of Peru has been in Lessons Learned charge of the "National Household Survey (ENAHO)" since 1995. There were quarterly surveys until 2002 (there was only one survey in 2001-02), The MECOVI program provides a series of and some of the survey modules on living conditions were permanent while lessons that could guide the design of a others rotated. The ENAHO is monthly implemented since 2003. In 2001 the program for technical cooperation on house- "Permanent Survey on Employment in Metropolitan Lima" was imple- hold surveys and related activities. The most mented. The MECOVI program has introduced the following strategies and important lessons of these are that: activities since 1997: (i) close consultation between sponsors and · Decentralized training of survey personnel (regional offices). beneficiaries, including producers and users groups, during program preparation, creates a · Focus on quality control procedures (questionnaires, data entry, consis- strong sense of country ownership; tency) in a decentralized manner. · Design of questionnaires and manuals in coordination with the users. (ii) it is important to build capacities based on · Introduction of new modules (ethnic group and race; household haz- the current ones; ards; governance, democracy and transparency; citizenship security; etc.). (iii) the focus of efforts should be not just on · Creation of an Internet-based discussion group to standardize solutions improving instruments but also on develop- to fieldwork problems. ing design and execution of statistical opera- · Automatic codification of open questions (economic activities, etc.). tions capacities; · Production of Reports and Documents about imputations. (iv) interaction between statisticians and us- · Formulate the information more representative at a department level. ers make the project more feasible in the long · Generation of reports on poverty and employment results, among oth- term; ers. · Dissemination of primary databases in the internet. (v) wide access to and democratization of information are essential elements to consoli- Note: Excerpts from "INEI Experiences in Household Surveys", presented by INEI at date trust and confidence in the official statis- MECOVI's International Seminar, in Quito, Ecuador, June 25-27 of 2003. tics; 2 (vi) the country's commitment to produce high quality (x) Improve the organization and documentation of sur- statistics improves their credibility and usability; veys databases. (vii) workshops and courses are fundamental instruments (xi) Improve access and dissemination of household sur- to strengthen technical capacities and create collaboration veys information. networks; (xii) Foster the relationship between producers and users. (viii) horizontal cooperation among countries is a practi- cal and affordable mechanism to improve statistical capaci- (xiii)Promote execution of qualitative studies as a comple- ties; ment to the household surveys. (ix) close coordination among the sponsoring organiza- MECOVI and the regional statistical capacity tions helps prioritize the execution tasks for the project. While MECOVI has typically focussed on the capoacity of According to an evaluation report7, the benefits of the NSOs and their activities it is clear that this cannot be viewed MECOVI program reached not only the NSO, but also other in isoluation from National Statistical Systems (NSS) which parts of the government, non-government organizations and need to function ensure that household survey activity the private sector. This has resulted in an increased culture of achieves its potential developmental impact Going forward, statistics in the participating countries. the MECOVI program needs to extend its action framework to strengthen NSSs. However, much still remains to be done to consolidate improvements in household surveys in the region. Demand In this regard, the MECOVI program will increasingly for more information implies on-going challenges for the provide technical assistance for the generation of National participating countries. Strategies for Development of Statistics (NSDS), to support statistical development in 11 countries of the Region (the Challenges for Regional Household Surveys Andean and Central American countries). This is consistent with PARIS 219 and the Marrakech Action Plan for Statistics, endorsed by the Second International Conference on "Man- On-going challenges include the need to: agement for Development Results" in Marrakech, Mo- rocco10. The participation of statisticians and users in the (i) Satisfy new demand for household level statistics. generation of strategies for statistical development would help them realize the importance of statistics in a country's (ii) Achieve broad geographical coverage at lower levels development process, and identify goals and priorities to of desegregation (provinces, districts, municipalities). strengthen their collection and application.. (iii) Adopt flexible strategies for the execution of house- Similarly, in future the MECOVI program could support the hold surveys according to the needs of the country, tak- implementation of country activities established through the ing into consideration its technical and financial ca- NSDS, such as: (i) revision of NSS's legal frameworks; pacities. training on methodology and management; (ii) technological development and roll-out to improve dissemination and the (iv) Obtain resources to finance the recurrent operating production process; (iii) systematization and organization of costs associated with of household surveys. databases for social, economic, demographic, and environ- mental data; and (iv) revision of basic and summary statistics (v) Improve coherence between household surveys and methodologies, among others. other statistical activities: censuses, administrative records, and other surveys. In general terms, official statistical activity in Latin America needs to be strengthened to be able to respond to the (vi) Improve the capacity to measure the evolution of vari- challenges ahead, especially increased demand for informa- ables through the implementation of continuous sur- tion. In this sense, it is important for each country to define veys and panel data observations. policies or medium term strategic plans to identify objectives and prioritize activities. Household surveys need to be (vii) Strengthen the link between surveys and processes for included in these plans. monitoring and evaluation. (viii)Promote methodological research on household sur- About the Authors veys. Haeduck Lee is a Senior Economist, Poverty Sector, Latin (ix) Emphasize training for producers and users at the na- America and Caribbean Region, World Bank. Gilberto tional level. Moncada is a Consultant in the same group. 3 National statistical offices, availability of strategies for statistical development, household surveys, and census in Latin America Population Strategy Planning and Country National Statistical Office for Statistical Household Surveys 2/ Housing Development 1/ Census 3/ Argentina Statistics and Census National Institute (INDEC) No Permanent Household Survey (EPH). 2001 Bolivia 4/ National Institute of Statistics (INE) No Survey on Improvement of Living Conditions. 2001 Brazil Brazilian Institute of Geography and Statistics (IBGE) Yes National Survey by Housing Samples (PNAD). 2000 Chile National Institute of Statistics (INE) No Employment National Survey (ENE). 2002 National Administration Department of Statistics Continuous Household Survey, 2003. Survey of Colombia 4/ No 1993 (DANE) Living Conditions. Costa Rica 4/ Statistics and Census National Institute (INEC) No Multi purpose Household Survey (EHPM). 2000 Integrated System of Household Surveys Ecuador 4/ Statistics and Census National Institute (INEC) Yes 2001 (SIEH).. Statistics and Census General Directorate El Salvador 4/ No Multi purpose Household Survey (EHPM). 1992 (DIGESTYC) Guatemala 4/ National Institute of Statistics (INE) No Survey of Living Conditions (ENCOVI). 2002 Permanent Multi purpose Household Survey 2001 Honduras 4/ National Institute of Statistics (INE) No (EPHPM). National Institute of Statistics, Geography and 2000 Mexico Yes Employment National Survey (ENE). Informatics (INEGI) National Household Survey on Living 1995 Nicaragua 4/ Statistics and Census National Institute (INEC) No Conditions measures. Panama 4/ Statistics and Census Management Office (DEC) No Household Survey. 2000 Peru 4/ National Institute for Statistics and Informatics (INEI) No National Household Survey (ENAHO). 1993 Paraguay Statistics and Census Management Office (DGECC) No Permanent Household Survey. 2002 Dominican National Statistical Office (ONE) No - 2002 Republic Uruguay National Institute of Statistics (INE) No Continuous Household Survey. 1996 Venezuela 4/ National Institute of Statistics (INE) Yes Household Survey by Sample. 2001 / It refers to strategic policies or plans for medium-term statistical development designed and reached by consensus at the NSS level and its users. Some of the countries are at their creation or revision process (Honduras, Ecuador, Peru, Venezuela). It does not refer to NSO's plans for statistical operations. 2/ Most important labor force and living conditions standard surveys. Other periodic surveys implemented by NSOs are: household income and family expenditures, demographics and health. 3/ The year of the last population and housing census. 4/ Countries that will receive support from the World Bank, through MECOVI `s operations platform, in order to create and improve their statistical development strategies between the years 2005-2006. Notes 1The MECOVI Program is short for "Programa para el Mejoramiento de las Encuestas de Hogares y la Medición de Condiciones de Vida", a Spanish acronym for the Regional Program for Improvement of the Surveys and Measurement of the Living Conditions in Latin America and the Caribbean. 2 For further information contact to Gilberto Moncada, Coordinator (gmoncadavigo@worldbank.org); Namsuk Kim, Databank Administrator (nkim@worldbank.org); Marta Cervantes-Miguel, Program Assistant (mcervantes@worldbank.org). 3Colombia is in the process of being incorporated. 4With the exception of the NSO in the Dominican Republic which is developing capacities to do so. 5Projects executed by through CARICOM and the Organization of the East Caribbean States (OECS). 6Including a training course on "Poverty Analysis and Social Development: in St, Lucia (May 2001). 7Jacob Ryten. The evaluation of the MECOVI Program. Dec. 2000. MECOVI/IDB. 8Statistical services in Latin America are provided by National Statistical Systems (NSS), which depend on each country 's public institutions for the collection of statistical information. NSSs are regulated by the National Statistical Offices (NSO). 9The Consortium Partnership in Statistics for the 21st Century (PARIS21) is sponsored by the World Bank, International Monetary Fund, United Nations, and OECD. 4