E n v i r o n m e n t S t r a t e g y The Poverty/Environment Nexus -- Findings for Lao PDR Susmita Dasgupta, Uwe Deichmann, Craig Meisner, and David Wheeler E No. nvironmental degradation can inflict serious damage on poor people, 5 because their livelihoods often depend on natural resource use, and their living conditions may offer little protection from pollution of the N air, water and soil. At the same time, poverty-constrained options may induce the O poor to deplete resources and degrade the environment at rates that are incompat- V ible with long-term sustainability. In such cases, degraded resources may precipi- E tate a downward spiral, by further reducing the income and livelihoods of the poor. M This "poverty-environment nexus" has become a major issue in the recent litera- B E ture on sustainable development. In regions where the nexus is significant, jointly R addressing problems of poverty and environmental degradation may be more cost- effective than addressing them separately (World Bank 2002). 2 In this note, we use newly-available spatial and survey data to investigate the 0 0 spatial dimension of the nexus in Lao PDR.1 Our data enable us to quantify several 2 environmental problems at the provincial level.2 In a parallel exercise, we map the provincial distribution of poor households. Merging the geographic information on poverty and the environment, we search for the nexus using geo-referenced indi- cator maps and statistical analysis. While our results are certainly suggestive, we recognize that they represent only the first round of empirical work on this issue. Future research using higher-resolution FRAGILE LANDS 5 data may reveal different, more local di- mensions of the nexus in Lao PDR and else- Steep hillsides under intensive cultivation are particularly vulnerable to ero- where. Nevertheless, the current analysis sion and soil degradation and, in any case, the economic return to farming can suggest a geographic focus for further, steeply-sloped areas is generally lower than the return to cultivating alluvial more detailed work. soils in river valleys. If rural people are relatively free to migrate to areas with higher expected returns, then we would expect steeply-sloped areas to MAPPING POVERTY be more sparsely populated than alluvial plains. A very different picture We begin the analysis by mapping poverty emerges, however, if the poor are isolated in highland areas. In this case, at the province level. From a welfare per- rapid population growth, coupled with insufficient means or incentives to spective, we believe that the size of the intensify production, may induce over-exploitation of fragile lands on steep poverty population in each area is a better slopes. Degraded resources, in turn, may precipitate a `downward spiral' by guide for policy than the incidence of pov- further reducing the income of the poor. erty.3 Accordingly, we index provinces by Figures 1 and 2 provide useful insights about population clustering on the number of inhabitants who fall below marginal lands in Lao PDR. Figure 2 shows that the northern and southeast- the international norm for absolute ern regions have extensive steeply-sloped areas. Comparison with Figure 1 poverty.4 suggests that the concentration of poor people on marginal (steeply-sloped) lands may be a major problem in the north, but not in the south. A recent MAPPINGENVIRONMENTALPROBLEMS Poverty Participatory Assessment traces this northern problem to cultural We consider four critical environmental and ethnic issues (ADB, 2001). If marginalized ethnic groups are isolated in problems, two related to natural resource highland areas by a historical pattern of separation and discrimination, then degradation and two to pollution. The population growth and soil degradation may well create a "poverty trap." "Green" problems are deforestation and Figure 2 suggests the potential for this component of the poverty-environ- fragile land, while the "Brown" problems ment nexus in northern Lao PDR. Analysis of higher-resolution data could are indoor air pollution and contaminated help clarify the links among poverty, ethnicity, topography, land-clearing water. and soil quality in the northern highlands. DEFORESTATION INDOOR AIR POLLUTION The rate of deforestation serves as a proxy Recent research has suggested that indoor air pollution from wood fuels is a for the loss of critical ecosystems and major cause of respiratory disease in developing countries.6 Many house- biodiversity, as well as the increased risk holds use wood or charcoal in Lao PDR, so indoor air pollution may be a of soil erosion in steeply-sloped areas. To significant health issue. Gauging the true magnitude of the problem will re- see whether deforestation is particularly quire household-level pollution monitoring and health assessment. In the severe in areas with large poverty popula- absence of indoor air monitoring data, we attempt a preliminary estimate tions, we map the rate of deforestation in from household survey data on the use of wood fuel and charcoal. Lao PDR during the 1990s). Where signifi- The existence of a true poverty-environment nexus in this context de- cant forest areas remain, we assess the pends on more intensive use of charcoal and wood fuel by poor households. spatial correlation of poverty and defores- In fact, this is strongly suggested by the high correlation (Á = 0.74) between tation using maps, rank correlation coeffi- the province-level poverty population and the population using fuel wood or cients, and regressions. For the country as charcoal (Figure 3). Regression results further confirm that use of wood fuel a whole, our findings suggest that priority and charcoal is far more prevalent among poor households than in the gen- areas for poverty alleviation and forest pro- eral population.7 This, together with the mountainous terrain of Northern tection are weakly related (Figure 3). At the Lao, a colder climate, and more indoor cooking, suggests that indoor air regional level, however, different patterns pollution may also be part of the poverty/environment nexus in Lao PDR. emerge. In the north, with a large poverty population, deforestation is high in a rela- ACCESS TO CLEAN WATER AND SANITATION tively small forested area. The south, by Safe water and adequate sanitation are critical determinants of health sta- contrast, has a large poverty population, tus, particularly for children. Ingestion of coliform bacteria from contami- relatively extensive forests, and relatively low rates of deforestation. This leads one nated drinking water or food is a prime cause of diarrheal disease, which is to believe that there are other driving fac- in turn a major cause of infant mortality in developing countries. Although tors relating to deforestation and the north- data remain limited in Southeast Asia, we use the available information to ern poor, and should be the focus of fur- assess the spatial relationships linking poverty, sanitation and diarrheal dis- ther investigation. ease in Lao PDR. At present, many households do not have access to safe Figure 1 Figure 2 Poverty population, 1997/98 Steeply-sloped land Source: Kakwani and others 2002. Source: Mekong River Commission (MRC) 2001. water or sanitation. A poverty/environment linkage may exist if the affected Ezzati, M., and D. Kammen. 2001. "Indoor households are disproportionately poor. Air Pollution from Biomass Combustion as a Risk Factor for Acute Respiratory Our results indicate a close spatial correlation between poverty and lack of Infections in Kenya: An Exposure-Re- access to clean water (Á = 0.85 in Figure 3). A similarly-positive, but somewhat sponse Study," Lancet, 358 (9281): 619- weaker, relationship exists for lack of access to sanitation (Á = 0.43). Further, a 624. strong relationship exists between poverty and diarrhea (Á = 0.75). Jalan, J., and M. Ravallion. 2002. "Geo- graphic Poverty Traps? A Micro Model CONCLUSIONS / KEY FINDINGS of Consumption Growth in Rural Figure 4 summarizes the evidence on the poverty-environment nexus in Lao China." Journal of Applied Econometrics, 17 (4): 329-346. PDR. It suggests that a significant spatial nexus may exist in all four of our Kakwani, N., G. Datt, B. Sisouphanhthong, environmental dimensions: deforestation, fragile lands, indoor air pollution, and P. Souksavath, and L. Wang. 2002. Pov- contaminated water.8 The potential problems appear to be greatest in the north- erty in Lao PDR During the 1990. Mimeo., ern and northeastern regions of the country. For Lao PDR, our results therefore World Bank: Washington, DC. suggest that the welfare of the poor may be significantly enhanced by close World Bank. 2002. Linking Poverty Reduc- integration of poverty-alleviation and environmental strategies. A geographic tion and Environmental Management: focus on the north would appear to be most beneficial. Policy Challenges and Opportunities. The above discussion raises an important question of what causal reasons World Bank: Washington, DC. may be behind the coincidence of poverty and environmental degradation in NOTES some areas of Lao PDR. If there is a high degree of labor mobility, returns to the 1. The spatial dimension is critical, because use of natural resources must still be high enough for poor people to remain in most environmental problems are inherent- degraded regions rather than migrate to higher potential areas. In that case, one ly geographical in nature. 2. One should note that the conclusions from may presume that the poor contribute to--or even cause--degradation. Struc- this analysis are applicable only at the pro- tural explanations, in contrast, would imply that it is the limited endowment of a vincial level. It is the intention of future re- region, and its lack of geographical capital, that has created a concentration of search to test whether such claims can be held at other levels of aggregation, such as poverty. If labor mobility is limited, then poor people are caught in so-called the district or sub-district level. To be more spatial poverty traps, where the poor resource base contributes to or causes precise, the provincial unit of analysis here high incidence of poverty (Jalan and Ravallion 2002). To shed more light on ranges from 3,500 to 21,000 km2. 3. An extreme example will help clarify the these important issues, and to credibly separate cause from effect, would re- underlying logic: Ten poor households quire detailed time series and household data that is currently not available for might constitute the entire population of Lao PDR. an isolated district, whose poverty inci- dence would therefore be 100 percent. On REFERENCES the other hand, one million poor house- Asian Development Bank. 2001. Participatory Poverty Assessment: Lao People's Demo- holds might represent no more than 40 percent of the population in a large urban cratic Republic. ADB: Manila, Philippines. area. Dasgupta, S., U. Deichmann, C. Meisner, and D. Wheeler. 2002. "The Poverty- 4. The poverty line for Lao PDR reflects a cost- Environment Nexus in Cambodia, Lao PDR and Vietnam." World Bank De- of-basic-needs approach: (a) the food pov- velopment Research Group Working Paper, World Bank: Washington, DC erty line is anchored to an energy thresh- (forthcoming). Figure 3 Authors Indicator correlation with poverty 0.90 Susmita Dasgupta is a senior econo- mist, Uwe Deichmann is a senior envi- 0.80 ytrevop ronmental specialist, Craig Meisner is 0.70 a consultant, and David Wheeler is lead economist. All are members of the hti 0.60 World Bank's Development Research w Group. noitalerroc 0.50 0.40 Reviewers This Note was reviewed by Anil Markandya, 0.30 lead specialist in the Environmentally and knaR 0.20 Socially Sustainable Development Unit of the Europe and Central Asia Region; Sudhir 0.10 Shetty, sector manager of the Poverty Re- duction Group; Jan Bojö, lead environmen- 0.00 tal economist, Environment Department; Deforestation Fragile land Indoor air No access to No sanitation Number of and Glenn S. Morgan, lead environmental rate (Slope>16%) pollution (wood/ clean water diarrhea cases specialist of the Environmentally and So- charcoal use) cially Sustainable Development Unit of the Indicator East Asia and Pacific Region; all of the Bank. Figure 4 Quartile ranking of poverty population and environmental Wood/ The Environment Strategy Notes Steep charcoal Unsafe No Diarrhea series aims to provide a forum for dis- Province Poverty Deforestation slopes use water toilet cases cussion on a range of issues related to Savannakhet 1 the Environment Strategy, to help the Champasack 2 transfer of good practices across coun- Huaphanh 3 tries and regions, and to seek effective Luangphrabang 4 ways of improving the Bank's environmental performance. Oudomxay 5 Saravane 6 The views herein are those of the Khammuane 7 author(s) and should not be considered Phongsaly 8 official policy of, nor attributed to, the Xiengkhuang 9 World Bank Group. Vientiane Municipality 10 Vientiane 11 Luangnamtha 12 T Xayabouri 13 Executive Editor Bokeo 14 Magda Lovei Attapeu 15 Managing Editor Borikhamxay 16 Poonam Pillai Sekong 17 Xaysomboon 18 Editor Nancy Levine 1st quartile 4th quartile 2nd quartile 3rd quartile Designer / (Worst) (Best) Production Manager Jim Cantrell old of 2000 calories per person per day; ence soil vulnerability (for example, vege- and (b) the nonfood poverty line allows for tative cover, rainfall erosivity, soil erodibil- basic nonfood expenses (estimated as the ity, lengths of uninterrupted slopes, and nonfood expenditure of households that conservation measures). T H E W O R L D B A N K can just afford the poverty line). The anal- 6. See, for example, Ezzati and Kammen Environment Department ysis indicates that 36.4 percent of the Lao (2001). 1818 H Street, N.W. population, or about 1.9 million people, 7. For further details see Dasgupta et al. Washington, D.C. 20433 USA were deemed poor in 1997-98 (Kakwani, (2002). Tel: 202 473 1000 Datt, Sisouphanhthong, Souksavath, and 8. Similar research has revealed very differ- Fax: 202 477 0565 Wang, 2002). ent patterns in Cambodia and Vietnam. See E-mail: eadvisor@worldbank.org 5. It should be understood that steeply-sloped Dasgupta and others (2002). Web: www.worldbank.org/eadvisor land is only one of many factors that influ-