TheWorld Bank N o v e m b e r PREMnotes 2 0 0 3 n u m b e r 8 1 Public Sector Estimating financing needs for local services in Madagascar As more countries decentralize public services, there is a growing need to assess local service needs and available financing. A study in Madagascar offers a methodology for conducting such assessments, which should be complemented by tools that measure the efficiency and quality of public spending. This note presents the methodology and find- financing needs with available resource flows ings of a field study on the financing needs of within a given year. This tool is not designed Madagascar's communes--the country's lowest to measure leakage in resource flows, which Local financing but most institutionally advanced level of sub- is analyzed through public expenditure track- national government. Following a first round ing surveys for health and education. The gap analysis offers of municipal elections in 1995, more than 1,500 approach has three steps: communes are now formally responsible for · Estimating unmet needs for basic local services, a rapid assessment maintaining basic administrative services and defined as the difference between a pre- social and economic infrastructure, including defined benchmark for service delivery of local service local waste disposal and sanitation. In addition, and existing local service capacity. communes are responsible for identifying and · Identifying available financing to support local needs and available coordinating local investments and for sup- service delivery, including from local gov- porting implementation of the national Poverty ernments, centralized and deconcentrated financing Reduction Strategy at the local level. service providers, donors, nongovernmental To finance these activities, communes re- organizations (NGOs), and communities. ceive population-based transfers and small · Estimating the financing gap by deducting conditional transfers, and can collect revenue available financing from estimated needs. from property, market, and consumption In Madagascar the health, education, and water taxes as well as user charges. Yet little is known (including sanitation) sectors were analyzed about how much these fiscal assignments sat- because of their large shares of commune isfy local needs. As part of its policy dialogue financing, low externalities beyond local with the government of Madagascar, the boundaries, and importance as priorities in World Bank is engaged in extensive research the national Poverty Reduction Strategy. that includes geographic mapping of social Because local financing gap analysis requires spending and a review of opportunities and intensive empirical field work, the study was obstacles to fiscal and sectoral decentraliza- applied to a sample of just 15 communes. The tion. This research generated the following communes were evenly distributed across six analysis of local and cross-sectoral service provinces and regions, but differed in size and needs and available financing. socioeconomic characteristics. The approach Estimating local service needs Local financing gap analysis offers a rapid Local needs are estimated based on the dif- assessment of local service needs and available ference between a hypothetical benchmark financing by comparing estimates of local of appropriate service delivery and existing from the development economics vice presidency and povert y reduction and economic management network Table 1 Types and sources of financing for basic services in communes in Madagascar, 2000 (U.S. dollars per capita) Sector Type of financing Line ministries Donor projects District committees Local governments Community groups Education Recurrent 1.31 0.00 0.00 0.10 0.12 (1.16 for salaries) Capital 0.05 0.68 0.13 0.00 0.06 Health Recurrent 0.40 0.02 0.00 0.01 0.21 (0.30 for salaries) Capital 0.03 0.47 0.02 0.00 0.11 Water Recurrent n.a. 0.00 0.00 0.00 0.03 Capital n.a. 4.83 0.00 0.00 0.14 Source: World Bank staff estimates based on commune finance study. service capacity, as measured by available which budget data were available from cen- infrastructure, equipment, and staff levels. tral sources. Financing through central min- For this study the level of appropriate ser- istries and community groups is likely to vary vice delivery was based on a conservative across communes--though much less than service benchmark in order to avoid unre- donor financing, which tends to be con- alistically high estimates of local financing centrated in select communes. gaps. This measure was based on techni- Overall, available resources are modest. In cal standards for infrastructure, equipment, 2000 average capital (investment) spending and staff coverage obtained from line min- for education and health was less than $1 per istries or donor projects. capita (table 1). Recurrent education spend- For education these standards were ing was higher, but most was committed to applied to the number of children enrolled teacher salaries and cannot be reallocated to in primary schools. (Although a calculation other needs. Meanwhile, nonsalary recurrent based on the total school-age population spending for education and health was only would have been desirable, it could not about $0.35 per capita. Resources for water be performed due to unreliable demo- and sanitation in the sampled communes graphic data.) For health and water the were significantly higher, at almost $5 per standards were used to calculate the costs capita. But these data reflect above-average of a functioning health facility at the com- donor investments in three of the sampled mune level and a collective potable water communes and so likely overstate the true system at the village level (excluding indi- resources available across communes. vidual household connections). The importance of different financing sources varies significantly for capital and Analyzing available resources recurrent spending, as well as among cen- Efforts to analyze available resources should tral and deconcentrated and local sources include all resources that reach communes, of financing (see table 1). In all three sec- including from the center, intermediate ser- tors, donor investments were the most impor- vice providers, and external and local sources. tant source of capital financing--ranging In Madagascar four sources of local financ- from $0.47 per capita in health to $4.83 in ing were reviewed: commune budgets, cen- water. The next largest resource category tral government and deconcentrated service was recurrent spending by line ministries, providers, contributions by beneficiaries and which was $1.31 per capita for education community groups, and projects, invest- and $0.40 for health. ments, and other contributions by donors Contributions from community groups and NGOs. were higher than expected in all three sec- Despite the small sample, the results for tors. Health care user fees and cost recovery commune budget data proved robust rel- schemes generated $0.21 per capita, account- ative to a sample of 320 communes for ing for almost one-third of all recurrent PREMnote 81 November 2003 spending. Moreover, communities often con- sectors. Access rates for infrastructure are tributed to salary expenditures--especially particularly low in education. Even a con- in education, where many teachers are servative estimate, based on minimum infra- directly employed by local parent-teacher structure requirements and covering only associations. children already in school, points to signif- Communities also provided the second icant capital investment needs for rehabili- largest share of capital financing (except for tation and construction in this sector ($15.0 education, where special investment funds per capita; table 2). Water needs are also administered by district-level assembly del- considerable ($13.8 per capita). In health, egates provide significant contributions). where the service network is better devel- Although all these community inputs tend oped, investment needs are lower at $4.6 per to be small in absolute terms, they typically capita. Financing needs for recurrent spend- Existing financing represent an important subsidiary source of ing in education and health are compara- financing--especially in the water sector, tively manageable and more evenly falls far short where recurrent spending is financed almost distributed (about $2.5 per capita). entirely through user fees and water sales. Existing financing generally compares of service needs The study confirmed that service delivery poorly with these needs. Some 85­95 per- by local governments is hampered by weak cent of local needs for health and educa- administrative capacity and unproductive tion go unmet--particularly in education, local revenue arrangements. For example, where the financing gap is estimated to be most taxes are collected by deconcentrated $16.2 per capita. The gap for water services agents, not communes. Only in education is much smaller, but only because of large do communes finance a noteworthy share (9 donor investments in a few of the sampled percent) of recurrent spending. Still, com- communes. And absolute financing needs munes provide important administrative and in water are still substantial, at $8.8 per organizational functions--such as monitor- capita (see table 2). ing land and population registries, local devel- Extrapolated to the national level, this opment planning, and management of local snapshot of local finance indicates that Mada- security--that affect public service delivery. gascar's rural communes would have required $320 million in resource transfers to attain Calculating the financing gap appropriate service levels in 2000--equal to The local financing gap is calculated by sub- 40 percent of the country's current budget tracting available financing from estimated resources and 40 times current block trans- service needs. An important distinction must fers to these communes. Moreover, this esti- be made between existing, committed funds mate is conservative because it covers only (such as fixed spending on salaries) and fun- minimum sector norms and does not include gible funds (such as capital and recurrent, nonsalary spending) that might be reallo- Table 2 Needs and gaps in financing for basic services in cated to unmet service needs. This study communes in Madagascar, 2000 assumed that existing funds represent avail- (U.S. dollars per capita) able assets that enter only indirectly into the needs estimate, while fungible funds help Type Existing Sector of financing Needs financing Financing gap close the gap. At the same time, fixed spend- ing on salaries was considered part of exist- Education Total 17.5 1.3 16.2 ing funds because it already contributed to Capital 15.0 0.9 14.1 Recurrent 2.5 0.4 2.1 existing levels of service delivery. Capital and recurrent, nonsalary spending were assumed Health Total 6.9 0.8 6.1 Capital 4.6 0.6 4.0 to be fungible allocations that could be used Recurrent 2.5 0.3 2.1 to address unmet service needs. The analysis confirms existing evidence Water Total 13.8 5.0 8.8 on uneven service performance in key social Source: World Bank staff estimates based on commune finance study. PREMnote 81 November 2003 other important priorities (such as security by two expenditure tracking surveys and and transport) and because education cov- reforms of financial management proce- erage had been limited. dures under a World Bank public sector reform project. Conclusion · Despite decentralization of fiscal and political The story of underfinancing of basic local ser- autonomy, communes lack sufficient financial vices in low-income countries has been told capacity to support local service delivery. many times before. But this study has tried to Although communes perform important add a cross-sectoral and local perspective on functions, they have limited capacity to service needs and financing gaps that can be finance local services provision. Donors, summarized in three main conclusions: central ministries, deconcentrated line Reforms could · Even though the gaps presented here are con- agencies, and even community groups con- servative estimates, they are extremely large. Still, tribute noticeably more to the financing strengthen the from a macroeconomic perspective $320 of local services. Other reforms--includ- million in additional financing would be ing strengthening local revenue collection, capacity of local manageable if Madagascar increased its increasing transfers, and streamlining the revenue efforts from the current low 11 control framework--could strengthen the governments to percent of GDP to about 15 percent--the capacity of local governments to finance average for low-income countries--and if and deliver local services. finance and deliver the investment needs identified here were stretched over several years. Further reading local services · Institutional reforms and decentralization of World Bank. 2003. "Decentralization in decisionmaking in the public sector can directly Madagascar." Report 25793-MAG. Wash- affect local service provision. Public spending ington, D.C. in Madagascar is heavily centralized. Thus reallocating resources within the exist- This note was written by Wolfgang Fengler (Econ- ing budget framework to deconcentrated omist, PREM Unit, Africa Region) and Frank- line agencies and to local facilities could Borge Wietzke (Consultant). significantly enhance the financing avail- If you are interested in similar topics, consider able for local services. Moreover, much joining the Decentralization Thematic Group. Con- could be done to increase the efficiency of tact Dana Weist (x82210) or click on Thematic resource use. This issue will be addressed Groups on PREMnet. This note series is intended to summarize good practices and key policy findings on PREM-related topics. The views expressed in the notes are those of the authors and do not necessarily reflect those of the World Bank. PREMnotes are widely dis- tributed to Bank staff and are also available on the PREM Website (http://prem). If you are interested in writing a PREMnote, email your idea to Madjiguene Seck. For additional copies of this PREMnote please contact the PREM Advisory Service at x87736. PREMnotes are edited by Paul Holtz and laid out by Suzanne Luft. Prepared for World Bank staff