HD WEEK WORKSHOP ON SOCIAL SECTOR DECENTRALIZATION

Led by Christine Allison and Don Winkler

MARCH 6, 2002

NOTES.

The workshop on social sector decentralization was well-attended, and the discussion covered a wide range of topics of concern and relevance to operations. To motivate the discussion, brief presentations were made by Don Winkler, Stuti Khemani, and Randy Hatfield. DW discussed the underlying premise of decentralization—that increasing the voice of beneficiaries and citizens will improve accountability and efficiency—and the challenge of monitoring and testing this premise. Both the concepts of voice and accountability are difficult to directly observe and measure, and disparities between de jure and de facto decentralization further complicate the issue. Nonetheless, several studies are under way within the Bank to examine how decentralization affects social service delivery. These studies can be seen as having the objective of answering two broad questions: [1] How can the devolution of powers to local governments be made to work for education and health? [2] How would we in the social sectors design decentralization to best improve social service delivery?

Nigeria. SK discussed the work she is leading in Nigeria to assess how decentralization is affecting the delivery of primary health care. The resumption of democracy in 1999 led to the call for devolution to local authorities. In health this entailed building community participation through the creation of district and village development committees. Since there is lots of variation in health decentralization across states, the Nigeria study will focus on only 2-3 states and will include collection of data at the level of the primary health care facility and the local government. The study will assess how the performance of local health committees affects service delivery. Issue: The fact local governments are creatures of state governments, as opposed to having a legal basis in the consititution, allows states to create new local governments in response to incentives in the revenue sharing formula.

Pakistan. RH discussed the rapid devolution now underway in Pakistan and the implications for the delivery of schooling. The Federal Government has mandated the devolution of social services from the provincial to local [district] governments. As is true in India and elsewhere, the implementation of devolution has created a conflict between the provincial line ministries, which wish to retain power, and the newly elected governments, which wish to exercise their newly granted authority. The devolution includes the re-creation of school committees, which had by and large failed earlier in the 1990s. The existing organizational structure and personnel at the district level and below have been transferred to the district governments. The Bank study is carrying out an institutional mapping of the educational sector to determine the incentives, motivations, and behaviors of key actors and to monitor how these may change with decentralization. The sample consists of one district [and several local governments and schools] in each of Pakistan’s four provinces. Issue: Resource allocation decisions are made at the district level, which is a general purpose government located several levels above the school, and voters influence those decisions only indirectly. Can indirect election of general purpose governments give beneficiaries/citizens effective voice?

The group discussion which followed these brief presentations was especially rich and benefited from the observations of several field staff attending the workshop. The issues raised in that discussion are briefly summarized below. Several speakers noted the richness of the Bank’s worldwide experience and the lack of even qualitative assessments of lessons learned.

  1. In evaluating decentralization, one must distinguish between several sectoral objectives: [a] quickly raising coverage, [b] reducing absenteeism by service professionasl, [c] raising the quality of services. Community monitoring may be more effective in [b] than [c].
  1. Decentralization can be introduced slowly [e.g., Colombia, India, Uganda] or quickly, using the big-bang approach [Argentina, Indonesia, Pakistan]. What lessons have we learned from past experiences with the big-bang approach?
  1. The constitutional status of local governments makes a difference. Total independent local governments [e.g., Poland] are different from local governments which depend legally or financially on higher levels of government.
  1. Decentralization can lead to elite capture and worsened delivery service to the poor. The result of Tanzania’s decentralization in the 1980s was to decrease the provision of preventive medicine in favor of curative medicine [hospitals]. Do local committees charged with monitoring health and education prevent elite capture? How can they be constructed and selected to effectively represent the community?
  1. The means of financing social services is important. Block grants to municipalities may lead to decisions to provide inadequate amounts of some social services, especially to the poor [e.g., Bulgaria]. Municipalities may need incentives to spend on social services, such as matching grants. Should the social sectors be arguing for earmarked rather than block grants?
  1. Weak local governments may need strong action and guidance from higher levels of government; a Harvard study of the Title I program in the US demonstrated that governments serving the poor often make poor decisions.
  1. Intergovernmental financing arrangements are extremely important. If revenue sources are highly centralized [i.e., local governments have few own-source revenue sources], it’s difficult to have decentralized decision making about resource allocation.
  1. The nature of the social services being delivered is important in evaluating the effects of decentralization. The ECA regional study shows a crucial difference between the delivery of curative and preventive care.
  1. How important is capacity building in implementing decentralization? Low local capacity is a common argument used by central ministries to delay decentralization. How much capacity is enough?
  1. What works in terms of capacity building? It’s common for the Bank and other aid agencies to support training and other types of capacity building at the local level. What lessons have we learned? How should we go about delivering training?
  1. How can the Bank and other donors facilitiate capacity building by learning from doing and learning from each other? Nigeria uses a mentoring model where governments with high capacity guide those with low capacity. Brazil has active associations of local government officials which disseminate innovations and lessons learned.
  1. There are differences in perspectives between sectors within the Bank. The delivery of rural infrastructure is easier than the delivery of social services due to the more private nature of the good. The urban sector is more concerned with building sustained local government capacity and independent revenue sources, while the social sector is more concerned with the quantity and quality of services delivered to the poor.
  1. Decentralization is rarely accompanied by the reengineering of central [or regional] government ministries required for them to carry out their new roles of monitoring, evaluation, providing information, technical assistance, etc. Do we have good models of how to design and implement this reengineering?