E n v i r o n m e n t S t r a t e g y Rethinking the World Bank's Environmental Portfolio Distribution -- Adjusting for Threats and Country Risk Craig Meisner, Susmita Dasgupta, and David Wheeler T No. he World Bank's Environment Strategy has as one of its objectives im- provement of the quality of life of the poor in developing countries. Of 11 special concern are the health, livelihoods, and vulnerability of the poor. N In line with the Strategy and with the Bank's commitment to the UN Millennium O Development Goals, the institution's projects and programs are being realigned to V incorporate environmental factors. E M How effectively does the current distribution of the environment portfolio address B the Strategy's objectives? Are environmental problems in specific countries' port- E folios being under-represented (or over-represented) in relation to their magni- R tude? The Development Research Group and the Environment Department have recently launched an analysis of portfolio gaps and opportunities. This analysis 2 can help the Bank adjust its portfolio to redress the imbalances. In this Note, we 0 look at portfolio distribution options in light of the severity of environmental prob- 0 4 lems in each country and examine how well current portfolio distribution match the needs suggested by the gap analysis. The quantitative information emerging from the study can serve as an added tool for countries and the Bank in their decision-making process on the distribution of the environment portfolio. METHODOLOGY BOX 1. CALCULATING OPTIMAL The analysis was divided into three stages: PORTFOLIO DISTRIBUTION SHARES 1. Formulation of a portfolio distribution To illustrate the method of estimating optimal portfolio distribution, we look method that incorporates (1) the relative at health damage from water pollution in two countries. In country A 160,000 severity of environmental problems in a people suffer from waterborne diseases and in country B 20,000 are affected. country; (2) country risk, as gauged by the The environmental threat ratio is 8:1 (160,000/20,000). probability of project success; (3) repre- In case 1 the 8:1 ratio is used to determine the countries' shares of the Bank's sentation in the portfolio of all partner portfolio distribution in water and sanitation. Case 2 additionally incorporates countries; and (4) equal valuation of health risk by multiplying each country's environmental threat index by the probabil- improvements across countries. Because ity of project success, as described in the text. In country A 50 percent of projects have been successful; in country B, 80 percent have. For country A detailed poverty information was not avail- the risk-adjusted index is 80,000 (160,000 × 0.50), and for Country B it is able for all client countries, the analysis 16,000 (20,000 × 0.80). After adjustment, the portfolio shares are in the ratio gives equal weight to every human life 5:1 (80,000/16,000). within a country, as well as among coun- Now suppose that $180 million is available for water and sanitation projects tries. in the two countries. In case 1, which does not adjust for risk, the distribution would be $160 million for country A and $20 million for country B. When 2. Construction of indices of environmental adjustment is made for both the environmental threat and country risk (case threat. To keep the study manageable, we 2), the optimal distribution are $150 million for country A and $30 million for focused on three critical quality of life prob- country B. lems: For air pollution, the indicator is total lates, produced by indoor and outdoor combustion, account for a large damage to human health from breath- proportion of health damage from air pollution (Holgate and others 1999). ing fine particulate matter. Fine particu- Damage is expressed in disability-adjusted life-year losses (DALYs) as documented, for instance, in Pandey and others (2003). For water pollution, the indicator is preventable deaths from unsafe wa- Table 1 Actual Portfolio Distribution ter and poor sanitation (see WRI and others 1999, on the importance of Shares by Region (percent) water and sanitation to health). Here too, the impact is measured in DALYs, Fragile Region Total Air Water lands as estimated by Wang and others (2003). Sub-Saharan Africa (AFR) 9 10 10 8 The vulnerability of poor people living on fragile lands--on steep slopes, East Asia and Pacific (EAP) 37 34 39 36 in arid areas, and on land covered by natural forest--has been identified Europe and Central Asia (ECA) 10 6 10 12 as a major determinant of rural poverty and natural resource degrada- Latin America and the Caribbean (LAC) 19 27 20 17 Middle East and North Africa (MNA) 8 1 8 10 tion in developing countries (World Bank 2003). Total rural population is South Asia (SA) 16 22 13 17 used as the indicator because quantitative studies of poverty-environ- Total 100 100 100 100 ment links on fragile lands are sparse.1 Table 2 Case 1: Optimal (Threat Based Only) Portfolio Distribution Shares The Bank's quality of life objective, it should be noted, refers specifically by Region (percent) to the poor. The severity of environmental impacts may differ by popula- tion group, and although in this exercise we did not differentiate among Fragile Region Total Air Water lands subgroups of the population, it would be possible to incorporate poverty Sub-Saharan Africa (AFR) 30 9 43 21 weights favoring benefits that accrue to the poor, assuming the availabil- East Asia and Pacific (EAP) 34 44 30 37 ity of reliable data on the poor population of a country. Europe and Central Asia (ECA) 4 10 3 5 Latin America and the Caribbean(LAC) 6 6 7 5 Middle East and North Africa (MNA) 5 5 2 8 3. Development of a measure of country risk, using the database of outcomes of South Asia (SA) 20 26 15 25 World Bank projects maintained by the Bank's Operations Evaluation De- Total 100 100 100 100 partment (OED).2 For this analysis, the country indicator is the percentage of Table 3 Case 2: Optimal (Threat Based and "successful" projects, defined as those falling into the top four categories in Country Risk Based) Portfolio Distribution Shares the OED's eight-category rating system--highly satisfactory, satisfactory, by Region (percent) moderately satisfactory, and marginally satisfactory. The measures could also be weighted to account for the varying costs Fragile Region Total Air Water lands of intervention. This approach would provide another screen for filtering Sub-Saharan Africa (AFR) 23 6 34 15 feasible portfolio distribution options. In the absence of detailed cost infor- East Asia and Pacific (EAP) 42 50 38 44 mation, it was implicitly assumed that costs across the three environmental Europe and Central Asia (ECA) 4 9 3 4 Latin America and the Caribbean (LAC) 6 6 7 5 themes are the same. Middle East and North Africa (MNA) 5 5 2 7 Portfolio redistribution opportunities were estimated by region and by South Asia (SA) 20 24 15 24 environmental theme, applying two sets of assumptions for each analysis. In Total 100 100 100 100 case 1 optimal country portfolio distribution shares are calculated using the environmental threat measure. Case 2 incorporates, in addition, the OED-based 6), the portfolio distribution structure across estimates of success probability to adjust portfolio distribution shares for coun- environmental themes remains roughly the try risk.3 same, perhaps because the likelihood of Country portfolio distributions are derived by multiplying the calculated project success varies less among themes than portfolio shares by total lending portfolio in environmental and natural resource among regions. The magnitude of the sug- management activities since 1990, using the Bank's new coding system.4 Over- gested redistribution across themes is not as all portfolio distributions are then obtained by summing the optimal lending large as for regions. portfolios for reducing air pollution, water pollution, and threats to fragile lands. INTERPRETATION AND NEXT STEPS Finally, the country results are aggregated for the major regions in which the Bank operates. (See box 1 for an illustration of the methodology). It should be By region. The regional results in table 3 indi- noted that although the total amount of environmental lending appears to be cate that, optimally, the largest share of total fixed (since we sum the total amount of lending over the period 1990­2003), environmental portfolio should go to EAP (42 lending varies substantially from year to year because of the "lumpiness" of percent), followed by AFR (23 percent) and multiyear projects and as countries address each environmental issue. This lumpi- SAR (20 percent). Other regions should receive ness prompted us to take a broad accounting stance and to allow for a more significantly lower shares: LAC, 6 percent; adequate representation of each environmental theme. MNA, 5 percent; and ECA, 4 percent. Com- parison of the optimal shares with actual port- COMPARISON BY REGION folio distribution shares suggests that, on The top panel of table 1 shows actual portfolio distribution, by region and envi- average, LAC, MNA, and ECA are over-repre- ronmental theme. The largest portfolios are in East Asia and Pacific (EAP)-- sented in the overall environment portfolio dis- largely in China--followed by Latin America and the Caribbean (LAC) and by tribution, while AFR and EAP are under-rep- South Asia (SAR). resented. Case 1 takes into account the environmental threat (Table 2). The calcula- Optimal portfolio distribution patterns for tion treats the air, water, and fragile land budgets as fixed (even though in reality each environmental theme vary significantly they are not), and allocates the shares among the regions. EAP still has the highest shares of the total portfolio for air and fragile lands, but AFR gains over- Table 4 Actual Portfolio Distribution all and with respect to water and fragile lands. Corresponding decreases appear Shares by Environmental Theme (percent) for LAC and, to a lesser extent, Europe and Central Asia (ECA) and the Middle Fragile East and North Africa (MNA). This suggests the desirability of shifting resources Region Sum Air Water lands Sub-Saharan Africa (AFR) 100 12 50 38 from the LAC, ECA, and MNA regions to projects in water and fragile lands in East Asia, Pacific (EAP) 100 15 39 46 AFR and, to a lesser degree, to air, water, and fragile lands in SAR. East Europe, Central Asia (ECA) 100 6 44 50 Case 2 incorporates a measure of country risk by using the OED measure Latin America, Caribbean (LAC) 100 1 44 54 of project success (Table 3). Under this scenario, net regional distribution of the Middle East, North Africa (MNA) 100 14 49 37 South Asia (SAR) 100 9 48 42 lending portfolio to AFR declines 7 percent from the case 1 figure (shown in Across all regions 100 10 46 43 Table 2) and EAP gains 8 percent. Factoring in project success thus dampens the Table 5 Optimal (threat-based only) Portfolio shift of resources from EAP to AFR under the threat-only assumption of case 1. Distribution Shares by Environmental Theme The implication is that when project success is taken into account, the lending (percent) (Case 1) portfolio in EAP would have a greater impact on the overall environmental threat Fragile across regions than would the portfolio in AFR. Region Sum Air Water lands Sub-Saharan Africa (AFR) 100 3 67 30 East Asia, Pacific (EAP) 100 13 34 53 East Europe, Central Asia (ECA) 100 23 33 45 COMPARISON BY ENVIRONMENTAL THEME Latin America, Caribbean (LAC) 100 11 22 67 Middle East, North Africa (MNA) 100 11 51 38 Another reality in the Bank's portfolio distribution is the priority the client coun- South Asia (SAR) 100 13 40 46 try accords to environmental concerns and its capacity to deal with them. The Across all regions 100 10 46 43 pressure of other issues may account for the differences between the actual lending portfolio distribution shares in table 1 and the optimal shares in tables 2 Table 6 Optimal (threat-based and country risk based) Portfolio distribution Shares by Environ- and 3. To the extent that country interest in environmental issues is beyond the mental Theme (percent) (Case 2) control of the Bank, it becomes harder to shift resources among regions in the direc- tions suggested by the analysis. Fragile Region Total Air Water lands We therefore next assume that the portfolio in each region is fixed, and we Sub-Saharan Africa (AFR) 100 3 68 29 estimate the shares for air, water and fragile land. Once again, we assume that East Asia and Pacific (EAP) 100 12 36 52 regional lending envelopes are fixed, when in fact they are not. Table 4 which Europe and Central Asia (ECA) 100 22 35 43 Latin America and the Caribbean presents actual portfolio distributions, shows that the shares of portfolio for (LAC) 100 10 24 66 water and fragile lands dominate in all regions. Table 5, which takes into ac- Middle East and North Africa (MNA) 100 11 52 37 count environmental threats, reveals regional lending portfolio under-represen- South Asia (SAR) 100 12 42 45 tation and over-representation. Finally, when project success is factored in (table All regions 100 10 46 43 by region (table 3). AFR would receive a Holgate, Stephen T., Jonathan M. Samet, Hillel Authors large part of the portfolio for safe water S. Koren, and Robert L. Maynard, eds. 1999. Air Pollution and Health. San Diego, Calif.: David Wheeler is a Lead Economist, (34 percent), second only to EAP (38 per- Academic Press. Susmita Dasgupta is a Senior Econo- cent), while 15 percent would be optimal Pandey, Kiran D., Katharine Bolt, Uwe mist, and Craig Meisner is a consultant for South Asia. AFR's theoretically optimal Deichmann, Kirk Hamilton, Bart Ostro, and in the Development Research Group. lending portfolio for cleaner air is strikingly David Wheeler. Forthcoming. "The Human The note is based on P. Buys, S. lower (6 percent) than actual portfolio dis- Cost of Air Pollution: New Estimates for De- Dasgupta, C. Meisner, D. Wheeler, K. veloping Countries." World Bank Policy Bolt, K. Hamilton, and L. Wang (2003) tribution and is in the same range as for Research Working Paper. Washington, D.C. "Measuring Up: New Directions for the lowest-portfolio regions, while EAP Wang, Limin, Katharine Bolt, and Kirk Environmental Programs at the World (largely China) has a significant share, 50 Hamilton. Forthcoming. "Estimating Poten- Bank," World Bank Policy Research percent, and SAR's is 24 percent. The ac- tial Lives Saved from Improved Environ- Working Paper No. 3097. The authors tual portfolio for fragile lands (natural re- mental Infrastructure." World Bank, Envi- would like to thank Anjali Acharya (En- ronment Department, Washington, D.C. vironmental Specialist) and Michiko source management) is close to the overall World Bank. 2003. World Development Report Shima (consultant) of the Environment optimal portfolio, with EAP receiving 44 2003: Sustainable Development in a Dynamic Strategy Implementation Team for pro- percent, SAR 24 percent, AFR 15 percent, World. New York: Oxford University Press. viding updated data and information on and the other regions much smaller shares. World Resources Institute (WRI), United Na- the World Bank's environment portfo- lio required for this analysis. By environmental theme. The results by tions Development Programme (UNDP), United Nations Environment Programme environmental theme are much less dra- Reviewers (UNEP), and World Bank. 1999. World Re- matic. Briefly, the calculations suggest that sources 1998­1999: A Guide to the Global This Note was reviewed by Julia the Bank should invest more in air in ECA, Environment: Environmental Change and Hu- Bucknall, Sr. Environmental Specialist LAC, and SAR, more in water in AFR and man Health. New York: Oxford University (MNSRE); and Elizabeth White, Sr. Results Management Specialist MNA, and more in fragile lands in EAP, LAC, Press. (OPCRX); both of the World Bank. and SAR. Accounting for project success (risk) changes the results only slightly, but NOTES this finding does not diminish the need to The Environment Strategy Notes series aims 1. To estimate the affected population, geo- to provide a forum for discussion on a range improve the policy and institutional condi- graphic information system (GIS) tech- of issues related to the Environment Strat- tions under which projects are undertaken. niques are applied to spatial overlays of egy, to help the transfer of good practices across countries and regions, and to seek Environmental issues are complex, demographic, topographical, climatic, effective ways of improving the Bank's en- and exploring the trade-offs between is- and natural resource information. vironmental performance. sues is a daunting task. Exercises such as 2. The database covers 3,075 World Bank projects implemented in 146 countries the one reported here are significantly lim- The views herein are those of the author(s) since 1990. and should not be considered official policy ited by the quantity and quality of the data 3. Case 1 assumes that damage abatement of, nor attributed to, the World Bank Group. available. It is limited in that it does not in all countries is given equal weight by consider projects supported by other de- the Bank and that future projects have velopment partners. The scope of the study T the same probability of success in all is currently being broadened through the countries. For case 2 we only assume that Executive Editor incorporation of additional environmental the Bank assigns equal weight to dam- Ede Ijjasz-Vasquez indices and the inclusion of non-lending age abatement in all partner countries. Managing Editor services. Although the portfolio redistribu- 4. Under the Bank's new coding system for Poonam Pillai tion approach cannot capture the full com- its lending and non-lending products, en- Editor vironmental and natural resource man- plexity of environmental decision-making, Nancy Levine agement (ENRM) activities are captured the analysis presented here may lead to under seven themes: biodiversity, climate Designer / better understanding of the extent of data Production Manager change, environmental policies and in- gaps and can highlight areas in which Jim Cantrell stitutions, land management, pollution countries, supported by the World Bank management and environmental health, and other development partners, can best water resources management, and other focus limited resources. ENRM. REFERENCES Buys, Piet, Susmita Dasgupta, Craig Meisner, T H A N K S T H E W O R L D B A N K Kiran Pandey, David Wheeler, Katharine Environment Department A special note of gratitude goes to Bolt, Kirk Hamilton, and Limin Wang. 2003. 1818 H Street, N.W. Ede Ijjasz-Vasquez for serving as "Measuring Up: New Directions for Envi- Executive Editor of the Environment Washington, D.C. 20433 USA ronmental Programs at the World Bank." Strategy Notes series. This is his final Tel: 202 473 1000 World Bank Policy Research Working Pa- Fax: 202 477 0565 issue as he is now Sector Manager of per 3097. Washington, D.C. Available at E-mail: eadvisor@worldbank.org the Bank's Water and Sanitation http://econ.worldbank.org/files/ Web: <>