In India, approximately 86 per cent of rural households and 24 per cent of urban households rely on solid biomass fuels for their cooking needs. These fuels used in traditional stoves, in households often with little ventilation, emit smoke containing significant quantities of harmful pollutants in the immediate proximity of people leading to serious health consequences. It is estimated that up to 444,000 premature deaths in children under 5 years, 34,000 cases of chronic respiratory disease in women under 45 years and 800 cases of lung cancer are attributable to exposure to Indoor Air Pollution (IAP) due to use of solid fuels by households. While health risks drive policy concerns, these are often difficult and costly to estimate. Information on population exposure to IAP is a useful proxy for health risks, and hence guide and facilitate mitigation actions. Better information on patterns of exposure and its determinants would assist in designing more effective interventions and strategies. As part of World Bank's study on Household Energy, Air Pollution and Health in India, a pilot exercise was conducted in rural Andhra Pradesh to collect quantitative evidence on the levels of exposure to IAP and key factors influencing these levels. This issue of the newsletter presents the results of this study designed by the Center for Occupational and Environmental Health, University of California, Berkeley (USA), and undertaken in partnership with Sri Ramachandra Medical College in Chennai and the Institute of Health Systems in Hyderabad (India). Exposure to Smoke from Cooking Fuels: Evidence from Andhra Pradesh A n exposure assessment study of respirable suspended particulate was piloted in three districts of matter (RSPM) in the micro- Chennai Andhra Pradesh (AP), a state in environments of the same households Southern India with a warm climate and household surveys to collect 24-hour Institute, all year around, which does not ch activity patterns of household members require heating. The study included Resear to estimate exposure (RSPM is defined in three major components: and the study as particles with a median Household-level surveys to aerodynamic diameter of 4 microns). College obtain data on various Exposure to RSPM -- a measure household characteristics and combining the number of people, the Medical behavioral factors, with focus on concentration of RSPM, and the amount secondary indicators of IAP, of time spent breathing it -- is one of the such as the distribution of fuel best indicators of health risks due to IAP. Ramachandra use, stove type and ventilation Sri ce: Finally, a statistical analysis to examine conditions. SourField monitoring of respirable particulate relationships between secondary Field monitoring to collect data matter being conducted in an outdoor household indicators of IAP and indoor kitchen in Andhra Pradesh (India) as part on daily average concentrations of the Exposure Assessment Study RSPM concentrations. Table 1: Health Impact of Exposure to Biomass Fuel Smoke Table 1 shows relative risk estimates for health outcomes associated with exposure to smoke from household fuel use based on a review of IAP studies. For example, children under five years of age exposed to indoor air pollution from solid fuel use have a two to three times greater risk of developing lower respiratory infections compared to unexposed children. There is strong evidence to support an association between solid fuel use and acute lower respiratory infections (ALRI), chronic obstructive pulmonary disease (COPD), and lung cancer (for coal only). Although there is epidemiological evidence to suggest an association with blindness (from cataract), asthma, and tuberculosis, more carefully controlled studies are needed to confirm these associations. Associations with adverse pregnancy outcomes (including low birth weight and still-birth) and ischaemic heart disease need further exploration for exposures from solid fuel use. Table- 1: Health Effects of Exposure to Smoke from Solid Fuel Use: Plausible Ranges of Relative Risk in Households Using Solid Fuels Households Health Outcome Population Affected Relative Risk Evidence Low High Acute Lower Respiratory Infections (ALRI) <5 years 2.0 3.0 Strong Asthma Females >15 years 1.4 2.5 Intermediate/ moderate Blindness (cataracts) Females >15 years 1.3 1.6 Intermediate/ moderate Chronic Obstructive Pulmonary Disease (COPD) Females >15 years 2.0 4.0 Strong Lung Cancer (coal only) Females >15 years 3.0 5.0 Strong Tuberculosis Females >15 years 1.5 3.0 Intermediate/ moderate Source: The above table is adapted from Smith KR and Mehta S. 2000. "The burden of disease from indoor air pollution in developing countries. comparison of estimates." Paper presented at the USAID/ WHO Global Technical Consultation on the Health Impacts of Indoor Air Pollution and Household Energy in Developing Countries. Note: A relative risk of 1 indicates that the risk is the same in the exposed and unexposed groups, i.e. there is no increased risk associated with exposure. Study Design Selection of Study Households The exposure assessment study involved the design of a The households were selected from three districts -- household questionnaire to collect data on household, fuel Nizamabad, Warangal, and Rangareddy of Telangana and other characteristics relevant to IAP exposure. region of Andhra Pradesh. Since the main objective of Reviews of national and state level surveys served as a this exercise was to investigate which household basis for designing the questionnaire. Primary data parameters affect exposure, household selection was collection was undertaken for two categories of done purposively, using a cluster sampling method to information: ensure that a variety of kitchen types and cooking fuel types are present within each cluster of households. The Information from households that parallels the cluster-sampling scheme aimed at obtaining information already collected by demographic surveys, approximately 150 households in each district by including the Census and the National Family Health selection of mandals (group of villages; smallest Survey (NFHS). administrative unit in a district) as first stage sampling Information on household characteristics that are not unit (5 from each district), habitations as second stage well-characterised in existing demographic surveys but sampling unit (1 from each mandal) and households as could be incorporated into future surveys if found to be third stage sampling unit (up to 30 from each predictive of IAP levels, such as kitchen type the presence habitation). Only mandals having a higher percentage of of cooking exhaust, number of open doorways, and clean fuel use (i.e., >2 per cent) were considered and information on the type and quantity of fuel consumed. selected using the probability proportionate to size 2 Figure-1: Sketch of kitchen types in sample households indoor kitchen without partition indoor kitchen with partition [ D ] [ W ] [ D ] [ W ] [ W ] Partition separate kitchen outside house open air kitchen outside house [ D ] [ W ] [ D ] [ D ] [ W ] D = Doorway, W = Window opening, - Stove criteria. (Roughly, 10 per cent of the study households Cooking with clean fuels is negligible in rural areas, so used clean fuels). Within each selected mandal, even with a purposive sampling, the number of clean fuel habitations having populations above 2,000 were users came out much smaller than in other categories. considered and one habitation was randomly selected. Finally, within a habitation, a cluster of households was Measuring IAP Concentrations selected to include a similar number of households in In summary, the data collection framework for the five categories described below: four categories of measuring RSPM concentrations and estimating solid fuel users with four different kitchen types (see exposure consisted of: Figure 1) and one category of clean fuel users. 24-hour average concentrations of RSPM in kitchen and The sampling protocol for selecting a cluster of living areas for all study households; households that satisfied the desired criteria above, involved visiting every fourth household starting from In addition, 10 per cent of households were monitored the center of a habitation. In the end, 420 households using a real-time monitoring instrument (PDRAM) to were covered by both the household survey and determine concentrations during cooking and non- monitoring, including: cooking windows; 24-hour outdoor area concentrations for about 20 per Solid fuel users, i.e. households using wood, dung, and a cent of the households combination of dung and wood: A detailed questionnaire was administered to each 1. Kitchen inside the home without a partition household to collect time-activity information as well (113 households) as information on some other variables likely to 2. Separate kitchen inside the home (108 households) influence concentrations (e.g., smoking habits and the use of insect coils). 3. Separate indoor kitchen outside the home (96 households) Findings of the Study 4. Outdoor kitchen (i.e., open air cooking) (95 households) Impact of Fuel Type on RSPM Levels 5. Clean fuel users, i.e., households using kerosene, Depending on the type of primary cooking fuel, liquefied petroleum gas (LPG) and biogas (45 households were grouped in mixed solid fuel users households). (97 households in the sample using wood and dung, or 3 dung, with small amounts of kerosene to start the fire Figure 3. Kitchen and living area RSPM concentrations but too small to affect overall emissions over a whole for different kitchen configurations in solid fuela burn cycle or day), wood users (270 households), households in rural AP kerosene users (11 households) and gas users (!! ''' '&# (34 households using LPG or biogas). Households using $ &(& mixed fuels have the highest RSPM concentrations, '!! 8A;µ &&) followed by wood, kerosene and gas, illustrating the &!! "energy ladder" sequence observed in other settings %!! $&( (Figure 2). Mean 24-hour kitchen concentrations in solid #)( #*! fuel using households were 2.5 (wood users) to 3.5 $!! #"& (mixed fuel users) times higher than that in kerosene #!! households, while in kerosene using households it is "!! #%2/9:42;0512<=>?2.91.@2 Figure 2: Mean 24-hour RSPM concentration for different ! +,-./0123405 6,7,1823405 fuels and house areas in rural AP B1C9942+,-./012D,-/2E54-,-,91 B1C9942+,-./012D,-/9:-2E54-,-,91 F:-C9942+,-./012D,-/2E54-,-,91 F:-C9942FE0123,42.99G,18 *!! ($# $ (!! a ­Solid fuel households include wood and mixed fuel users 8A;µ '!! -- in rural outdoors is 100 µg/m3, and RSPM was found to &!! &!! be approximately 60 per cent of PM10 in the households %!! $'# monitored). $%& #*) $!! Average living area concentrations were the highest in #!$ #!! households having kitchens without partitions (559 ""% ($ )) (& *( "!! µg/m3), followed by kitchens with partitions (357 µg/m3), #%2/9:42;0512<=>?2.91.@2 separate kitchens outside (280 µg/m3) and outdoor open ! +,-./0123405 6,7,1823405 F:-C994H air kitchens (215 µg/m3). Living area concentrations in ?,K0C I99C +049H010 J5H households using outdoor kitchens were much lower than those with indoor kitchens (215 µg/m3 versus 458 µg/m3) but about similar to households with a separate kitchen also about three times higher than that in LPG using households. Households using kerosene were found to (215 µg/m3 versus 280 µg/m3). These results indicate that dispersion can considerably affect adjacent living area be ambiguous in terms of fuel use because many of concentrations during outdoor cooking. these households frequently switch to wood for cooking, while report kerosene as their primary fuel on In households using gas, concentrations did not vary the basis of larger expenditure on kerosene in the across kitchen types. household budget than on traditional biomass fuels. Fuel Quantity and Ventilation Effects Effect of Kitchen Configuration on RSPM Levels in Solid Fuel A regression analysis showed that both kitchen and Households living area concentrations increased with greater consumption of fuel amongst solid fuel users. Living In kitchen areas, no significant difference in concentrations room concentrations were also influenced more by was observed between households having indoor kitchens number of windows and additional rooms, while kitchen with partitions and those without partitions (666 µg/m3 area concentrations were not, indicating the role of versus 652 µg/m3). Inferring from Figure 3, average dispersion in reducing area concentrations. This kitchen area concentrations are significantly higher in indicates that more specialised improvements in kitchen indoor kitchens compared to outdoor kitchens with or ventilation, such as possibly stove hoods to vent out without partitions (659 µg/m3 versus 436 µg/m3). Still, exhaust, are needed to have an impact. Concentrations even in outdoor cooking, concentrations levels are were also not correlated with the number of people being substantial both in the kitchen and living areas (297 µg/m3 cooked for; nor with total cooking duration. and 215 µg/m3), and well above the health guidelines for Concentrations were not significantly influenced by use outdoor air quality. (24-hour Indian average standard for of kerosene lamps, incense coils or mosquito coils; or by PM10 -- airborne particulate matter less than 10 microns presence of smokers in the house. 4 Exploring Determinants of IAP Concentrations Based on data collected from household surveys and After trying a number of modeling approaches and RSPM monitoring, the study examined the relationships specifications, the following observations were drawn: between kitchen/living area concentrations and household Three variables -- fuel type, kitchen type, and kitchen characteristics based on statistical modeling. Variables ventilation -- were found to be good predictors of significantly associated with kitchen and living area kitchen and living-area concentrations in high concentrations were included in the modeling process to concentration households.. Fuel type is the best explore whether and how certain household characteristics predictor of high concentrations, but not a good can be used to predict household exposure levels. predictor of low concentrations. This is due to the wide The following variables were significantly correlated with range of concentrations within fuel categories. In both kitchen and living area concentrations: general, low concentrations appear to be more difficult to predict than high concentrations Type of cooking fuel (mixed, wood, kerosene, gas) Type of kitchen (4 types, as described earlier) Kitchen type is a good predictor of kitchen Separate kitchen (outside the living area or not) concentrations; indoor kitchens are much more likely Kitchen ventilation (poor, moderate, good) to have high concentrations than outdoor kitchens. For Wall type (pucca, semi-pucca, kachha) living-area concentrations, knowing the specific type Floor type (pucca, semi-pucca, kachha) of kitchen is less important than knowing whether or Housing type (pucca, semi-pucca, kachha) not the kitchen is separate from the living area. Stove type (traditional, improved, kerosene, gas) Households with good kitchen ventilation are less likely Linear regression models that were used to predict to have high concentrations -- in both kitchen and living continuous outcome variables for kitchen and living-area areas -- than households with moderate or poor concentrations did not yield sufficient information for ventilation. Kitchen type and kitchen ventilation are explaining great variability in the kitchen and living-area equally good predictors of household concentrations. concentrations, especially within the wood and mixed fuel Each of these variables is unable to better predict categories. Subsequently, modeling was conducted for household concentrations above and beyond the other binary concentration categories (high and low variable. This suggests that it may not be necessary to concentration households) using logistic regression, and collect information on both kitchen type and kitchen classification and regression trees (CART) techniques. ventilation. Gender Differences in Exposure only group in solid fuel using households for which female and male exposures are similar. The study estimated exposure to IAP by combining concentrations measurement data with time activity patterns. There is no difference in exposure among household As shown in Figure 4, among solid fuel users, mean 24-hour members in gas using households . average exposure concentrations for women cooks in the age Figure 4: Exposure of different household members: group 16-60 years is one and a half times greater than that of women non-cooks (442 µg/ m3 vs. 276 µg/ m3 ), and three solid fuel versus gas households in rural AP time greater than that of men in the same age group (442 µg/ &!! m3 vs. 148 µg/ m3). The exposures of cooks are not $ %%# %&! significantly different across indoor kitchen types but are 8A;µ %!! higher than outdoor kitchens. $&! #(' $!! #'# Among non-cooks in solid fuel using households, women #&! in the age groups of 61-80 years experience the highest #!! "%* exposures followed by women in the age group of 16-60 "&! years, while men in the age groups of 16-60 years () "!! (# () (' experience the lowest exposures. This is presumably &! #%2/9:42570458020KE9H:40P2 because older women are most likely to remain indoors ! =9L,C2M:0L J5H2NH,182 leaving the younger women to be involved in assisting the NH,182O9:H0/9LCH O9:H0/9LCH cooks, while men are most likely to have outdoor jobs that M0;5L02Q99GH2R"'S'!2T4HU M0;5L02V91S.99GH2R"'S'!2T4HU ?5L02R"'S'!2T4HU Q/,LC4012:1C042&2T4H lowers their exposures. Young children (2-5 years) are the 5 The study strengthens the evidence that cooking with clean Chennai fuels reduces exposure substantially and, further, makes Institute, it equally low for all household ch members, including for women- Resear cooks who suffer much more and than others in solid fuel using households. At the same time, College the study finding that indoor Medical concentrations are well correlated with the quantity of solid fuel used indicates that the Ramachandra adoption of clean fuels will lead Sri to a tangible reduction in ce: exposure only if these fuels meet Sour The field monitoring team of the Exposure Assessment Study is explaining the monitoring procedure in a substantial portion of cooking an outdoor kitchen setting to village people in rural Andhra Pradesh (India) needs and biomass consumption is reduced considerably. It is Looking to the Future known that in the reality of rural life, complete or The following key observations and conclusions emerge substantial switching to clean fuels is seldom, and from the AP exposure assessment study: people continue to rely on biomass fuels. This highlights the need for multiple interventions to This study provides, for the first time, quantitative reduce exposure to IAP ranging from alteration in information on 24-hour concentrations and exposures housing design to provision of cleaner fuels and better to respirable particulate matter, for a wide cross- stoves that vent smoke outside the house. section of rural homes using a variety of household fuels and under various exposure conditions. There is While exploratory in nature, the effort at modeling a need, however, to exercise caution in the indoor air pollution concentrations provided valuable extrapolation of the study results as the findings are insight to the key determinants of exposures -- fuel based on a sample from only three districts of a single type, kitchen type, and/or kitchen ventilation. agro-climatic zone of one state in southern India, Remarkably, access to clean fuels and ventilation in wherein socio-cultural, housing and climatic IAP-related health issues seems to be analogous to conditions in this region are likely to differ from other access to clean water and sanitation in water-related parts of the country. Further, monitoring was carried health issues. out only in summer months, which is not reflective of Future studies of this kind could address the high the time-activity pattern of household members and intra-household variability in the behavioral aspects of the nature of biomass fuel used for all seasons. This family cooking, e.g., in ways a stove has been operated points to the need to collect quantitative data on a on a particular day. Multiple-day measurements could regional basis to increase the level of confidence in the be taken in future studies to smooth out this variability. findings emerging from this study. The resulting multi-day mean concentrations levels While women-cooks suffer from far greater exposure in would, presumably, be better suited to be predicted by parameters that do not change daily, such as fuel type solid fuel using households, the exposure level could be and house structure. significant for residents of the house other than the cook. An important finding for policy concern is that The findings of this study, if validated with other data even for households that cook outdoors, the 24-hour and refined models, could also influence the design of concentrations and exposures could be significant both large-scale survey instruments, such as the Census or in the cooking place and indoors, well above levels National Sample Survey, by introducing questions on considered acceptable by air quality guidelines. This the key determinants of exposure, namely kitchen type challenges the conventional wisdom, and a frequent and/or ventilation, in addition to the fuel type already excuse to ignore the problem, that cooking outdoors -- used by these surveys, with a view to facilitating as the majority of poor households do -- prevents the classification of population sub-groups into exposure health risks from fuel smoke. sub-categories. 6 Health Benefits of IAP in Xuanwei, China, Journal of The National Cancer Institute, Vol. 94, No. 11, 826-835, June 5, 2002) Mitigation Strategies: Strengthening the IAP and acute respiratory infections in Kenya ­ an Evidence exposure response study Recent studies on household energy and indoor air A recent field study examined the details of personal pollution issues have significantly contributed to a better exposure and the exposure-response relationship for understanding of the impact of interventions on exposure indoor air pollution and acute respiratory infection and health benefits. A brief summary of selected studies is (ARI) in children and adults in rural Kenya. The study presented below. took place at Mpala Ranch in Laikipia District, Central Kenya, where cattle herding and domestic labour are Improved stoves and lung the main occupations. Stoves used by the households cancer risk reduction: were unvented and used firewood or charcoal (and kerosene in 3-4 houses). evidence from China By monitoring both the health and exposure to indoor Stove improvements reduce exposure to indoor air smoke over a three-year period, the researchers have pollution and thereby decrease the risk of lung cancer characterised the `exposure-response' relationship among farmers in rural China, suggests one recent study between indoor air pollution from biomass smoke and published in the Journal of the National Cancer Institute. the rates of acute upper and lower respiratory Traditionally, household cooking and heating in China infections. This, first of its kind, characterisation of the commonly involved burning of coal in unvented health effects of indoor smoke over a wide range of indoor stoves. Such unvented burning produces very exposure levels shows that: high indoor concentrations of airborne particulate The average daily exposure concentration of children matter, benzopyrene and other organic compounds. under the age of 5 to PM10 is approximately 1,500 These pollutants are risk factors for lung cancer and µg/m3 and that of adult women approximately 5,000 other respiratory diseases. In this retrospective cohort µg/m3. By comparison, the latest National Ambient study of 21,232 farmers followed from 1976 to 1992 in Air Quality Standards of the US Environmental Xuanwei, China, the incidence of lung cancer among Protection Agency required the daily average 17,184 farmers who switched from unvented firepits or concentration of PM10 (particulates below 10 microns stoves to stoves with chimneys was compared with the in diameter) to be below 150 µg/m3 (annual average incidence of lung cancer among 4,048 farmers who below 50 µg/m3). continued to use unvented firepits or stoves. Tobacco smoking and frequency of cooking were similar in Young and adult women who regularly participate in both groups. The authors found that levels of indoor cooking activities have exposure levels that are two to air pollution created by vented burning were less than four times higher than men. They are also in average 35 per cent of levels created by unvented burning. twice as likely as men to be diagnosed with a case of Average concentrations of PM10 decreased from 2080 acute respiratory infection (ARI). Children exposed to µg/m3 before stove improvement to 710 µg/m3 after high levels of indoor air pollution are in average improvement. Further, after adjusting for factors such diagnosed with ARI in approximately 20 per cent of as tobacco smoking and duration of cooking, changing weekly examinations and the more severe Acute Lower from unvented to vented stoves was associated with a Respiratory Infection (ALRI) in 6-8 per cent of weekly 41 per cent reduction in lung cancer risk in men, and a examinations. 46 per cent reduction in lung cancer risk in women. The findings of the study provide strong With the best-estimate of the exposure-response encouragement for stove improvements and other relationship and controlling for a number of measures to reduce indoor air pollution from important confounding factors, the study found household fuel burning. that ARI and ALRI are increasing functions of average daily exposure to PM10, but the increase (Q. Lan, R. Chapman, D. Schreinemachers, L. Tian and X. may flatten for exposures above approximately He. Household Stove Improvement and Risk of Lung Cancer 1,000-2,000 µg/m3. 7 Simple, locally manufactured improved-efficiency A combination of interventions: stoves can reduce exposure to indoor smoke by Kenya experience approximately 40 per cent and cleaner fuels (such as charcoal) by approximately 90 per cent International Technology Development Group (ITDG) compared to traditional wood-burning stove. If this experience over the past ten years in Kenya has shown reduction in exposure is maintained, the incidence that an effective program to reduce IAP and ensure of lower respiratory infections among better health and cleaner homes, requires a combination children, the most important health effect from of improvements in cooking and housing depending on exposure to indoor smoke will be reduced by an the needs and preferences of the household members, estimated 20-45 per cent. especially women-cooks. More efficient stoves alone While these first results, which are subject to cannot reduce IAP to levels that could be considered uncertainty due to both statistical factors and safe. Poor Kenyan households in rural areas use difficulties in measurement of exposure and firewood as the primary source of fuel in inefficient health, should be confirmed in future studies, they cooking devices ranging from three stone fires to metal provide an important bridge to understanding the charcoal stoves that consume large amounts of fuel and exposure and health effects associated with this pre-dispose users to harmful emissions. To reduce important risk factor. The results demonstrate the indoor air pollution levels in these houses, participatory potential extensive health benefits of efforts to approaches were used to develop a package of develop and promote improved stoves and cleaner interventions together with local communities that were fuels among the poor households in the durable, affordable and accessible. These included developing world still relying on traditional changes to kitchens, improved cooking stoves, smoke biomass. hoods, and the provision of windows and eaves. Monitoring of the reductions in particulate matter and (Ezzati, M. and D. M. Kammen (2002) "Evaluating the Health carbon monoxide (CO) to assess the impact of the Benefits of Transitions in Household Energy Technology in interventions revealed that smoke hoods were most Kenya" Energy Policy, 30(10), 815-826 effective in reducing concentrations of both pollutants, by 75 per cent and 78 per cent, respectively. Windows Ezzati, M. and D. M. Kammen (2001) "Indoor Air Pollution did not result in any significant changes in kitchen from Biomass Combustion as a Risk Factor for Acute pollutant concentrations. While eaves space reduced Respiratory Infections in Kenya: An Exposure-Response Study" particulate levels by 60 per cent, it cut CO levels only by Lancet, 358(9282), 619-624 (erratum 358(9287)) 28 per cent. Improved stoves brought down particulate matter and CO levels by 54 per cent and 42 per cent, Ezzati, M., H. Saleh, and D. M. Kammen (2000) "The respectively. Contributions of Emissions and Spatial Microenvironments to Exposure to Indoor Air Pollution from Biomass (J. Nyaga, Lessons from Kenya's Household Energy Activities, Combustion in Kenya" Environmental Health Perspectives, Presentation at the Regional Workshop on Household Energy, 108(9), 833-839) Air Pollution and Health, New Delhi, May 9-10, 2002) Feedback We would appreciate if you send us comments and suggestions. Sameer Akbar Kseniya Lvovsky South Asia Environment and Social Development Unit South Asia Environment and Social Development Unit The World Bank The World Bank 70, Lodi Estate 1818 H St., NW New Delhi 110003, India Washington DC 20433, USA Tel: 011-91-4617241 Fax: 011-91-4619393 Tel. 202-473-6120, Fax 202-522-1664 Email: sakbar@worldbank.org Email: klvovsky@worldbank.org If you would like to be included in the mailing list, please email to pritikumar@mantraonline.com Visit us at http://www.worldbank.org/sar/sa.nsf This newsletter is part of a World Bank study in India, Household Energy Air Pollution and Health, and is supported under the joint UNDP/World Bank Energy Sector Management Assistance Programme (ESMAP). The material used in this newsletter does not necessarily represent the views of the World Bank. Produced by macro graphics.comm pvt. ltd.