_ _ __ _ __ W6 POLICY RESEARCH WORKING PAPER 2565 Does Fiscal Decentralization Greater fiscal decentralization is consistently associated with Improve Health Outcomes? lower mortality rates. And its positive effects on infant Evidence from a Cross-Country mortality are greater in institutional environments Analysis that promote political rights. David A. Robalino Oscar F. Picazo Albertus Voetberg The World Bank Africa Technical Families Human Development 1 March 2001 I POLICY RESEARCH WORKING PAPER 2565 Summary findings Decentralization of fiscal responsibilities has emerged as mortality rates. The results suggest that the benefits of a primary objective on the agendas of national fiscal decentralization are particularly important for poor governments and international organizations alike. Yet countries. They suggest also that the positive effects of there is little empirical evidence on the potential benefits fiscal decentralization on infant mortality are greater in of this intervention. Robalino, Picazo, and Voetberg fill institutional environments that promote political rights. in some quantitative evidence. Fiscal decentralization also appears to be a mechanism Using panel data on infant mortality rates, GDP per for improving health outcomes in environments with a capita, and the share of public expenditures managed by high level of corruption. In environments with a high local governments, they find that greater fiscal level of ethnolinguistic fractionalization, however, the decentralization is consistently associated with lower benefits from fiscal decentralization tend to be smaller. This paper-a product of Human Development 1, Africa Technical Families-is part of a larger effort in the region to conduct rigorous analysis of the implications of fiscal decentralization on the financing and delivery of social services. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Oscar Picazo, roomJ10-062, telephone 202-458-7954, fax 202-473-8239, email address opicazo@worldbank.org. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The other authors may be contacted at drobalino@worldbank.org or avoetberg@worldbank.org. March 2001. (14 pages) The Policy Research Working Paper Senes dissemibates the findings of work in progress to encourage the exchange of ideas about devclopmzent issuies. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The fintdintgs, interpretations, and conclusions expressed in this paper are entirely those of the autbors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center Does Fiscal Decentralization Improve Health Outcomes? Evidence from a Cross-Country Analysis David A. Robalino Email drobalino@worldbank.org Oscar F. Picazo opicazo@worldbank.org Albertus Voetberg avoetberg@worldbank.org For criticisms, suggestions and help we thank Demissie Habte, Alberto Agbonyitor, Donata Hoetch, Ejaz Ghani, Peter Warr, and the members of Kenya Country Team at the World Bank. 2 Abstract: While decentralization offiscal responsibilities is one of the main objectives that has emerged in the agenda of national governments and international organizations, there is little empirical evidence of the potential benefits of this intervention. This paper is an attempt to fill, in part, the void of quantitative measurement. Using panel data on infant mortality rates, GDP per capita, and the share of public expenditures managed by local governments, we find that higher fiscal decentralization is consistently associated with lower mortality rates. Our results suggest that benefits offiscal decentralization are particularly important for poor countries. The results also suggest that the positive effects that fiscal decentralization has on infant mortality rates increase in institutional environments that promote political rights. Fiscal decentralization also appears as a mechanism to improve health outcomes in environments with high levels of corruption. Finally, we find that environments with high levels of ethno-linguistic fractionalization tend to reduce the benefits from fiscal decentralization. 1. Introduction Decentralization, in its various forms, has become a highly popular component in policy reform. Within the health sector, decentralization of finances and responsibilities is one of the important topics that has emerged in the agenda of national governments and international organizations. Indeed, devolving some of the centralized responsibilities to local levels is expected to improve both technical efficiency and allocative efficiency (see Peabody et al., 1999 for a review). The rationale is that decentralized programs can be designed with knowledge of the local culture and circumstances. They can also be allocatively efficient because resources can be devoted to the most-needed local services. With decentralization, fiscal responsibilities for services rest with local managers who have incentives to improve efficiency given that they can use the savings for other local purposes. Local managers also have more opportunities to reduce costs. They can tailor staff and procedures to local resources and circumstances, rather than relying on centrally determined procedures. In addition, information can be used without delay, rather than after permission is received from central agencies. Another possible benefit of devolution, at least in a large country, is that local governments can experiment with alternative ways of doing things. Some of these ways may turn out to be superior and can then be adopted by other regions. Despite these compelling arguments in favor of decentralization, there is little evidence that countries which have decentralized management and budgets within their health systems have also improved health outcomes. Indeed, studies evaluating the impacts of decentralization in developing countries (see for example Visschedijk et al., 1995; Green and Collins, 1994; Bossert, 1995; Bossert et al., 1991; Gilson, 1993; Kutzin, 1994; and Holland and Pimphachanh, 1995) provide mixed results. Problems associated with decentralization include, for example, an increase in regional disparities in the absence of a mechanism to transfer resources from rich districts to poor districts. Another major problem is associated with the lack of skilled personnel, 3 the lack of information, or the loss of economies of scale which counteract efficiency gains from devolution. In addition, decentralization may impose constraints to the implementation of national policies and the creation of coordination channels across regions (Guldner, 1995). So while qualitative studies have helped to understand the pros and cons of decentralization, the magnitude of its impacts on health outcomes remains unquantified. This is unfortunate since in the absence of such quantitative measurement, there is little that can be said in terms of the economic costs and benefits from devolution. These are critical to inform the design of sound public policies. This paper is an attempt to fill, in part, the void of quantitative measurements of the impact of decentralization. Our focus is on the impacts offlscal decentralization on an indicator of health outcomes, the infant mortality rate. To measure these effects we develop a simple theoretical model that we then estimate on the basis of panel data for countries. This panel includes an indicator of fiscal decentralization among the social and economic indicators derived from the World Development Indicators (World Bank, 2000). The indicator has been computed as the ratio between total expenditures of local governments and total expenditures of the central government on the basis of the Government Financial Statistics (International Monetary Fund, 2000). While the indicator is only a rough proxy for the fiscal decentralization process, it allows us to derive preliminary estimates of the magnitude of the potential correlation with the infant mortality rate. The remainder of the paper is organized in four sections. Section 2 develops a theoretical framework for estimation purposes. Section 3 describes the data and econometric methods. Section 4 summarizes the results. Finally, Section 5 presents concluding remarks. 2. A theoretical framework As previously discussed, the major channel through which fiscal decentralization is likely to affect health outcomes is an increase in levels of allocative and technical efficiency. A decentralized system is expected to be more successful in allocating scarce resources to alternative interventions in order to maximize health outcomes. To formalize this idea, we consider the problem faced by a benevolent policymaker who attempts to maximize the national average of a health outcome indicator such as the infant mortality rate. The assumption of a 4 benevolent policymaker does not necessarily reflect reality, but it allows us to define a best-case scenario to be used as reference in our empirical analysis. We postulate that within each region g in a country, the outcome indicator M is a function of structural characteristics of the economy (e.g., the level of aggregate output and its regional distribution), call them 0, and the allocation of public expenditures xg, among a set of interventions I (say female education, vaccination campaigns, and so on). We write: Mg =fg(0,xgl,...,Xgj), (1) We assume that f(.) is a continuous function which verifies afg > 0 and ' <0, so that an gi g increase in expenditures in any of the interventions increases the health outcome indicator, but the marginal effect decreases with the level of expenditure. Then the problem solved by the policy maker can be written as: Max9, :M M ng2fg(o xgl ... * x1 (2) where ng is the contribution of region g to the national average and Y is the total budget to be allocated'. Optimality implies % ng = i, Vi E I. The interpretation is that expenditures need axgj to be allocated in such a way that the marginal impact of an additional dollar to an intervention i in region g (adjusted by its weight ng), is the same across all interventions and regions. Hence, in theory, it is possible to compute the optimal level xg; of spending in intervention i in region g. We would have, x;, = xgi (0, Y) or sgj = s(0), where s; is the optimal share of expenditure in intervention i in region g. Unfortunately, in reality, various factors deviate expenditures from their optimal levels. First, policy makers at the central level may have scant information about the functions fg (.). ' We take this budget as given. The problem of allocating the entire public budget or the problem of finding the optimal taxation rate is outside the scope of this paper. 5 Another possibility is simply that policy makers are not benevolent and have objective functions that respond to political incentives rather than social welfare. In these cases, the functions used to allocate expenditures will be different fromfg (.) and therefore observed expenditures Xg,bs will be sub-optimal. By extension, the budget allocations to each region g defined by yohS = X 0bs will be sub-optimal as well. It is straightforward to show that for a given region g, xg; = Xgj (0, Y) are also solutions to the problem: Max;,: Mg = fg(0,Xgi- xg,...,xgi)J-'xg; - i) (3) where yg is the optimal budget of region g. Equation (3) gives the problem that would be solved by a local, benevolent policy maker in control of budget yg. We notice that the allocation of resources by local governments will generate a national optimum only if the budget allocated to each local government was optimal in the first place. Nonetheless, even if the budgets Yg = Xbs are sub-optimal, their management by local government can improve the health outcome as long as local authorities have a better knowledge of fg (.). If the budget going to region g is sub-optimal, the resulting level of expenditures in each intervention will be different from the optimal level; however, the relative level of expenditures will be optimal. We would have: obs= E X' yg jx . When the budget is managed centrally, we can measure the level of inefficiency in the allocation A obs to each intervention by: ug, = Exg' - xgb . Notice that we have E Eugi = 0. Hence, we can ~x;i yg bsgI think of U = E E ug; as a general indicator of inefficiency. Clearly, U will be a function of the g i share S of the total budget ygs that is managed (i.e., allocated) by local authorities. The partial derivative of U with respect to S will depend on the relative levels of efficiency of the local and central governments in the management of public expenditures. Hence, we can postulate: 6 au~~~~~7 @sU = C(c - 1), (4) where c and I are indicators of the level of efficiency in managing public resources of the central and local government respectively. So, if c>l (meaning that institutional capacity at the local level is low relative to the center), an increase in the share of public expenditures managed by local governments will increase inefficiency and reduce health outcomes.2 Under this set of assumptions, our health outcome indicator can be written as: M = f(O, xj(9, Y'))- h(U(j)) = g(9, Y) - h(U(s)), (5) where h is a continuous and monotonic function which gives the loss in the outcome indicator resulting from inefficiencies in the allocation of public expenditures. Equation (5) implies that the loss caused by a marginal deviation from the optimal level of expenditure in a given intervention is the same across regions and interventions. This is a rather restrictive assumption, but it is nonetheless necessary given the lack of data at the regional level. A first order expansion of (5) gives: M=g(O)+(6 -900) ag -h(U(So))-(S-So) F 5g c9~~~~~~h aU 1 ag ah X = g(O)-00 ag -h(U(SO))+SO a e ]+ ag -,Ud 5(6 aoo au0 ajso ao suo ~9C-)S' (6) = ao + a1 + a2S Hence, the sign of a2 provides information on the relative levels of efficiency of central and local planners in allocating resources. Now, it is reasonable to expect that a2 is itself a function of the country level of development and/or other structural factors such as the level of corruption or political rights. A simple formulation of this hypothesis would be: 2 Notice that in our framework we have not assumed that fiscal decentralization implies that each region manages its own resources. The efficient allocation of resources across regions requires redistribution of resources across regions. 7 a2 =ao +Xb', (7) where X is a row vector of structural factors and b is a row vector of parameters to be estimated. The next section describes the data and methods used to estimate models such as (6) to shed light on the question of how fiscal decentralization could affect infant mortality rates. 3. Data and methods We use a panel of low- and high-income countries covering the period 1970-1995. For each country, we observe over time the infant mortality rate, the real gross domestic product per capita (adjusted for purchasing power parity or PPP), an indicator of fiscal decentralization, and structural indicators related to institutional capacity such as civil rights, political rights, and corruption. The number of available observations varies between 519 and 664 depending on the model estimated3. The indicator of fiscal decentralization is defined as the ratio between expenditures managed by local governments and expenditures managed by the central government. The indicator was constructed on the basis of the Government Financial Statistics published by the IMF. This indicator has also been used by Fisman and Gatti (1999) to evaluate the linkages between corruption and fiscal decentralization, and by Davoodi and Zou (1998) to look at the relationship between economic growth and fiscal decentralization. WhileFisman and Gatti find that fiscal df centralization is associated with lower levels of corruption, Davood and Zou find no correlation between fiscal decentralization and economic growth in developed countries, and a negative correlation in the case of developing countries. On the basis of models (6) and (7) we estimate equations of the form4: log IMRi, = a0 + a log GDPj, + aO log S1, + log Si, (Xb') + v, + e;, (8) where IMR is the infant mortality rate measured in deaths per 1,000 live births, GDP is Gross Domestic Product (our proxy for the variable 0), S is the percentage of total expenditures 3 The panel is unbalanced in the sense that most of the countries have missing data for some of the variables in some of the years. The estimation methods are adjusted accordingly. 4 Models of the infant mortality rate usually introduce as predictors indicators of female education (a demand factor) and control for policy interventions such as immunization coverage (supply factors). In our framework, however, these indicators are endogenous interventions, which are related to the level of GDP per capita and the level of fiscal decentralization. Hence, we do not include them in the regression. 8 managed by the local government, X is a vector of structural indicators, v; are country specific shocks, and e is white noise. Table 1 presents summary statistics for each of the variables included in the analysis. The high variance of GDP per capita in the sample used for estimation purposes facilitates the generalization of the results from the analysis to countries with a GDP per capita lower than USD 1,000. Table 1 Mean of Model Variables Variables Observations Mean SD Scale Infant Mortality 766 21.6 26.61 GDP Per Capita 989 8,177 6,515 (PPP) Share of Local 1,145 21.4 20.4 Expenditures (%) Ethno-linguistic 1,056 0.23 0.23 0-1 Fractionalization Corruption 1,028 4.27 1.5 0-6 Political Rights 938 2.52 1.7 1-7 Source: This study Given that the variance of vi is not equal to zero5, model (8) cannot be estimated by Ordinary Least Squares (OLS) method. Fixed effects models and random effects models are usual alternatives. Nonetheless, we prefer the fixed effects model given that the assumption of independence between the exogenous variables and v, required for the random effects model is 6 not met. 4. Results We have estimated six models that differ in the vector X of structural variables. The results are summarized in Tables 2 and 3. The first model includes only GDP per capita and the indicator of fiscal decentralization. It shows that, other things being equal, countries where local governments manage a higher share of public expenditures tend to have lower mortality rates. This does not imply that countries that 5 Test: Varvi )=0; X 2=427.4; Prob> X2 =0.0000 6 Test: Cov(x, Vi )=0; X (3)=460.26; Prob> X' (3)=0.0000 9 decentralize the management of public resources will automatically improve health outcomes. Without institutional capacity at the local level, decentralization is unlikely to be successful. In our sample of countries, it is very likely that the share of public expenditures managed by local governments is correlated with their level of institutional capacity. Hence, from the data, we cannot distinguish between local governments with low and high institutional capacity, and assess how decentralization affects health outcomes in each case. The results, however, support the view that if local governments are strengthened then fiscal decentralization is likely to improve health outcomes. For instance, in a country with a GDP per capita of USD 2,000, a 10% increase in the share of expenditures managed by local governments would reduce the mortality rate by 3.6%. The results also show that the marginal benefits of decentralization diminish as GDP per capita increases. For instance, in the case of a country with a GDP per capita of USD 5,000, the reduction in the infant mortality rate would be marginally lower, at 3.3%. One plausible explanation for this result is that problems associated with centralized systems diminish as economic development takes place. Table 2 Estimates from Fixed Effect Models (within estimator) [ Obs 664 664 519 608 532 616 Groups 67 67 46 54 49 55 R-2-Within 0.80 0.80 0.84 0.82 0.83 0.82 R-2- 0.76 0.72 0.54 0.77 0.57 0.79 Between R2-Overall 0.77 0.73 0. 3 0.79 0.51 0.78 Coeff. SD Coeff. SD Coeff. SD Coeff. SD Coeff. SD Coeff. SD LogGdpPPP -0.810 0.053 -0.890 0.056 -0.921 0.060 -0.948 0.058 -0.870 0.057 -0.879 0.055 LogLocal -0.545 0.138 -2.238 0.449 -1.802 0.542 -2.339 0.445 -2.036 0.528 -2.175 0.447 LogLocal* 0.056 0.017 0.412 0.092 0.415 0.103 0.436 0.091 0.393 0.106 0.420 0.091 LogGdpPPP . LogLocal* -0.018 0.005 -0.017 0.005 -0.018 0.005 -0.017 0.005 -0.019 0.005 LogGdpPPP2 I.I_I_I LogLocal* -0.234 0.077 -0.060 0.056 LogCorruption . LogLocal* -0.197 0.065 -0.106 0.051 LogPolRights . LogLocal* logA 0.046 0.026 vElf Constant 9.935 0.430 10.518 0.450 10.629 0.513 11.077 0.472 10.155 0.490 10.473 0.446 Source: This study. 10 Table 3 Elasticity of the Infant Mortality Rate with Respect to the Share of Local Expenditures GDP Per Reduced Non-Linear Full Corruption Political Ethno- Capita Rights linguistic Fraction- alization 1000 -0.377 -1.164 -0.711 -1.197 -1.011 -1.082 1500 -0.367 -1.111 -0.656 -1.140 -0.960 -1.028 2000 -0.360 -1.074 -0.618 -1.101 -0.925 -0.991 2500 -0.355 -1.046 -0.589 -1.070 -0.898 -0.962 3000 -0.350 -1.023 -0.565 -1.046 -0.876 -0.939 3500 -0.346 -1.004 -0.545 -1.025 -0.858 -0.920 4000 -0.343 -0.988 -0.528 -1.008 -0.842 -0.903 4500 -0.340 -0.973 -0.514 -0.992 -0.828 -0.888 5000 -0.338 -0.960 -0.500 -0.978 -0.816 -0.876 5500 -0.336 -0.949 -0.488 -0.966 -0.805 -0.864 6000 -0.333 -0.938 -0.478 -0.955 -0.795 -0.853 6500 X -0.331 -0.929 -0.468 -0.945 -0.786 -0.844 7000 1 -0.330 -0.920 -0.458 -0.935 -0.778 -0.835 Source: This study. In the second model, we add the square of the logarithm of GDP per capita to account for a non- linear effect of the level of "economic development" on the elasticity of the infant mortality rate with respect to the share of public expenditures managed by local governments. The coefficient is highly significant and suggests that benefits from fiscal decentralization will be higher for high income and low-income countries and lower for middle-income countries. One possible interpretation of this result is that, at low levels of economic development, institutional capacity in local governments increases at a lower rate than in the central government, but increases faster for high levels of economic development. Introducing the new variable has an important impact on the size of the effect of fiscal decentralization on the infant mortality rate. Indeed, in the new model, a 10% increase in the share of public resources managed by local governments in a country with a GDP per capita of USD 2,000 would reduce the infant mortality rate by 10% (as opposed to 3.6%). To evaluate the role of governance, the third model adds Knack and Keefer (1995) indicators of corruption and political rights to the vector X. Again, the coefficients for the new variables are highly significant despite a changing sample size (given that the indicators are not available for all countries). The results suggest that the positive effect of fiscal decentralization on infant mortality rates is higher in institutional environments with strong political rights. This does not 11 come as a surprise. One possible explanation is that when political rights are high, communities can better influence policy-making at the local level, and thus encourage an allocation of resources that better meets their needs. The results also suggest that fiscal decentralization appears as a mechanism to improve health outcomes in environments with high levels of corruption. The negative effects of corruption - distorted government resource allocation decisions, reduced economic efficiency, impaired legitimacy of public institutions, and skewed income distribution in favor of the rich, have been widely reported. We speculate that fiscal decentralization, by redistributing the responsibility of the management of resources among a higher number of officials, reduces the marginal pay-off that each official derives from being corrupt and hence reduces the total level of corruption. This result does not hold, however, when the indicator of political rights is dropped (see model 4). This points to the importance of political rights to enable the anti-corruption capabilities of fiscal decentralization. On the other hand, the indicator of political rights remains significant when the indicator of corruption is dropped (see Model 5). The last model introduces an average indicator of ethno-linguistic fractionalization (see Fedderke and Klitgaard, 1998). This indicator measures the probability that two individuals picked at random in a given country belong to different ethnic groups. The results show that when ethno- linguistic fractionalization is high, fiscal decentralization tends to be less effective in reducing infant mortality rates, probably as a result of coordination failures. 5. Conclusions This paper explores the linkages between fiscal decentralization and infant mortality rates. The study is based on a panel of developed and developing countries using socioeconomic indicators such as infant mortality rate, GDP per capita, and the share of public expenditures managed by local governments which is used as a proxy for the level of fiscal decentralization. Five major results follow from the analysis in the paper: i) higher fiscal decentralization is consistently associated with lower mortality rates; ii) benefits from fiscal decentralization are particularly important for poor countries; iii) the positive effects of fiscal decentralization on infant mortality rates are enhanced in institutional environments with strong political rights; iv) fiscal decentralization appears as a mechanism to improve health outcomes in environments with high levels of corruption; v) environments with high levels of ethno-linguistic fractionalization 12 tend to reduce the benefits from fiscal decentralization. Nonetheless, these results need to be interpreted cautiously: first, given measurement problems associated with aggregated country data; and second, given that the indicator of fiscal decentralization used in the analysis is a very crude proxy for the fiscal decentralization process. It is also important to stress that the results presented in this paper do not imply that fiscal decentralization is a magic recipe to improve health outcomes. Appropriate institutional capacity at the local level should be an important pre-condition. An optimal allocation of public expenditures across regions and program interventions also requires appropriate coordination channels between regions and adequate transfer mechanisms (most likely from rich regions to poor regions). Public expenditure reviews at the local level can constitute an important instrument to guide this process. Other instruments include training in management and financial planning and the endowment of modern information technologies. In short, successful decentralization requires strong leadership from the central government. The central government must be able to influence local policy and implementation without compromising the autonomy of local decision-making from which many of the benefits of a devolved system would be expected to flow. As stated by Guldner "decentralization without direction appears to undermine health system effectiveness". 13 References Bossert, T., 1995. "Decentralization". In K. Janovsky, ed. Health Policy and Systems Development. Geneva: World Health Organization. Bossert, T., R. Soebekti, and R. Rai, 1991. "Bottom-Up Planning in Indonesia: Decentralization in the Ministry of Health". Health Policy and Planning 6(l):55-63. Davoodi, H. and H. Zou, 1998. "Fiscal Decentralization and Economic Growth: A Cross- Country Study". Journal of Urban Economics, 43, 244-257. Fedderke, J. and R. Klitgaard, 1998. "Economic Growth and Social Indicators". Reprinted from Economic Development and Cultural Change, Vol. 46, No. 3, April, pp. 455-489, RAND Corporation, RP-72 1. Santa Monica, CA. Fisman, R. and R. Gatti, 2000. "Decentralization and Corruption: Cross-Country and Cross-State Evidence". World Bank. PREM Advisory. Gilson, L., 1993. "Health-Care Reform in Developing Countries". Lancet, 342(8874): 800 Green, A., and C. Collins, 1994. "Decentralization and Primary Health Care: Some Negative Implications in Developing Countries". International Journal of Health Services, Vol. 24, No. 3, pp. 459-475. Guldner, M., 1995. "Health Care in Transition in Vietnam: Equity and Sustainability". Health Policy and Planning, 10 (Supplement): 49-62. Holland, S. and C. Phimphachanh, 1995. Impact of Economic and Institutional Reforms on the Health Sector in Laos. Institute of Development Studies. International Monetary Fund, 2000. Government Financial Statistics. Knack, S. and P. Keefer, 1995. "Institutions and Economic Performance: Cross-Country Tests Using Alternative Institutional Measures". Economic Policy, 7, pp. 207-27. Kutzin, J., 1994. "Experience with Organizational and Financing Reform of the Health Sector". Current Concerns, SHS Paper No. 8. WHO/SHS/CC/984.3 Geneva: World Health Organization. Mills, A, 1994. "Decentralization and Accountability in the Health Sector from an International Perspective: What are the Choices?" Public Administration and Development, Vol. 14, pp. 281-292. Peabody, J., 1999. Policy and Health: Implications for Development in Asia Cambridge University Press. Visschedijk, J.H., I.M. Liywalii, and J. Van Oosterhout 1995. "Pilot Project for Financial Decentralization in Senegal, Zambia". Geographical Medicine. 47(1): 39-42. 14 World Bank 2000. "World Development Indicators". Policy Research Working Paper Series Contact Title Author Date for paper WPS2542 Checks and Balances, Private Philip Keefer February 2001 P. Sintim-Aboagye Information, and the Credibility of David Stasavage 37644 Monetary Commitments WPS2543 When Do Special Interests Run Philip Keefer February 2001 P. Sintim-Aboagye Rampant? Disentangling the Role in 37644 Banking Crises of Elections, Incomplete Information, and Checks and Balances WPS2544 The Uniqueness of Short-Term Leora Klapper February 2001 A. Yaptenco Collateralization 31823 WPS2545 Financing the Future: Infrastructure Marianne Fay February 2001 A. Francois Needs in Latin America, 2000-05 37841 WPS2546 Gender Dimensions of Pension Paulette Castel February 2001 J. Smith Reform in the Former Soviet Union Louise Fox 87215 WPS2547 The Design of Incentives for Health Jeffrey S. Hammer February 2001 H. Sladovich Care Providers in Developing William G. Jack 37698 Countries: Contracts, Competition, and Cost Control WPS2548 International Provision of Trade Alan V. Deardorff February 2001 L. Tabada Services, Trade, and Fragmentation 36896 WPS2549 Measuring Poverty Dynamics and Erzo F. P. Luttmer February 2001 C. Wodon Inequality in Transition Economies: 32542 Disentangling Real Events from Noisy Data WPS2550 Measuring Equity in Health Care Adam Wagstaff February 2001 A. Maranon Financing: Reflections on (and 38009 Alternatives to) the World Health Organization's Fairness of Financing Index WPS2551 Infrastructure Coverage and the Kristin Komives February 2001 S. Minovi Poor: A Global Perspective Dale Whittington 30012 Xun Wu WPS2552 Inventories in Developing Countries: J. Luis Guasch February 2001 J. Troncoso Levels and Determinants-a Red Flag Joseph Kogan 37826 For Competitiveness and Growth WPS2553 The Value of Relationship Banking Giovanni Ferri February 2001 A. Yaptenco during Financial Crises: Evidence Tae Soo Kang 31823 from the Republic of Korea In-June Kim Policy Research Working Paper Series Contact Title Author Date for paper WPS2554 Administrative Costs and the Estelle James February 2001 A. Yaptenco Organization of Individual James Smalhout 31823 Retirement Account Systems: Dimitri Vittas A Comparative Perspective WPS2555 Implicit Pension Debt, Transition Yan Wang February 2001 A. Datoloum Cost, Options, and Impact of Dianqing Xu 36334 China's Pension Reform: Zhi Wang A Computable General Equilibrium Fan Zhai Analysis WPS2556 Household Strategies for Coping with Michael M. Lokshin February 2001 P. Sader Poverty and Social Exclusion in Ruslan Yemtsov 33902 Post-Crisis Russia WPS2557 Decentralization and Accountability: Stuti Khemani February 2001 H. Sladovich Are Voters More Vigilant in Local than 37698 in National Elections? WPS2558 Growth, Inequality, and Poverty: Martin Ravallion February 2001 P. Sader Looking beyond Averages 33902 WPS2559 Deposit Insurance as Private Club: Thorsten Beck February 2001 P. Sintim-Aboagye Is Germany a Model? 38526 WPS2560 Catastrophe Risk Management: John D. Pollner February 2001 J. Pollner Using Alternative Risk Financing 30079 and Insurance Pooling Mechanisms WPS2561 Democracy and Income Inequality: Mark Gradstein March 2001 P. Sader An Empirical Analysis Branko Milanovic 33902 Yvonne Ying WPS2562 Decomposing World Income Branko Milanovic March 2001 P. Sader Distribution: Does the World Have a Shlomo Yitzhaki 33902 Middle Class? WPS2563 A Race to the Top? A Case Study of Tsunehiro Otsuki March 2001 L. Tabada Food Safety Standards and African John S. Wilson 36896 Exports Mirvat Sewadeh WPS2564 Economic Analysis of World Bank Ayesha Yaqub Vawda March 2001 E. James Education Projects and Project Peter Moock 31756 Outcomes J. Price Gittinger Harry Anthony Patrinos