Decentralization Reforms

Decentralization may refer to a wide range of reforms, but it is possible to identify three principal types: deconcentration, delegation, and devolution. Kai Kaiser outlines three levels of analysis for a PSIA on decentralization depending on the institutional characteristics, the probable mechanisms of impact, and the likely available sources of information. At the first level, the focus is on the amount and distribution of public resources across places. At the second level, the analysis concerns the distribution of public resources across people, such as the poor and non-poor, given the prevailing institutional and governance arrangements within places. At the third level, the analysis investigates impacts on local governance and public service delivery. Assessing the distributional impacts of decentralization will typically require a considerable amount of sub-national data.

Kaiser describes instruments and methodologies that offer valuable entry points for the analysis. Public expenditure reviews have helped in analyses of the aggregate level, composition, and operational efficiency of public expenditures. Through public expenditure tracking surveys, one strives to produce detailed analyses of the extent to which public resources actually reach localities and front-line service delivery points. Poverty mapping, using sub-national fiscal data and other socioeconomic data, may be usefully combined to assess incidence across places. Frontline service delivery surveys may provide important insights into decentralized outcomes. Specialized surveys, including household and facility surveys, may also target dimensions of service delivery not usually found in standard household socioeconomic surveys, which normally supply evidence on basic service use and access, but not service quality. Institutional and governance reviews may serve as venues for describing the political economy of decentralization reforms.

 

Inherent in decentralization is the expectation that the outcomes are diverse across subnational jurisdictions. Hence, studies should attempt to gather evidence from a sufficiently large number of localities to be representative and at least indicative of national patterns. The author suggests that even the short-run evidence on decentralization is quite fragmented. One explanatory factor may be that decentralization efforts are, by design and manner of implementation, diverse and difficult to compare. Another issue is that the methodologies for establishing the direction and magnitude of distributional impacts have been idiosyncratic. 

Back to Reforms