Fiscal Decentralization and the Quality of Public Services in Russian Regions
Freinkman, Lev, Public Finance and Management
The paper provides empirical analysis of the relationship between fiscal decentralization and the quality of public services in the Russian regions. The analysis suggests that fiscal decentralization has no significant effect on the key inputs into secondary education, such as schools, computers, or availability of preschooling, but has a significant positive effect on average examination results, controlling for the key observable inputs and regional government spending on education. Decentralization also has a positive impact on the quality of municipal utilities provision. Both effects can be attributed to strengthened fiscal incentives rather than to superior productive efficiency of municipal governments.
Empirical evidence in support of the theoretical advantages of decentralization has been generally in-conclusive, as highlighted by Treisman (2007). This may be partly explained by the difficulties of measuring the extent of decentralization or its relevant aspects such as the degree of voter accountability or decision making autonomy (Treisman, 2002; Stegarescu, 2005). In fact, correlation between different measures of decentralization proves to be astonishingly low (Voigt and Blume, 2008).
Inconclusive empirical evidence may also reflect theoretical ambiguities of the impact of decentraliza-tion. There has been growing consensus in the literature that the results of decentralization depend critically on the local conditions, in particular on comparative inten-sity of distortions in the incentive structure at different tiers of government (Bardhan and Mookherjee, 2005; Blanchard and Shleifer, 2001). The extent of capture at different levels of government, and hence the potential benefits of decentralization, may accordingly depend on the quality of institutions and a number of other factors, including the extent of political centralization, i.e. pres-ence of strong nation-wide parties (Enikolopov and Zhuravskaya, 2007)
Finally, measuring performance under decentra-lization is itself a significant challenge, since the quality of public services tends to be benchmarked through ob-servable input indicators or process-oriented measures, such as pupil enrolment ratios, student-teacher ratios, or teacher absenteeism in education, rather than output in-dicators (such as quality of instruction), which is often much more difficult to assess.
In the light of these complications, more empiri-cal analysis is needed to strengthen the case for further decentralization. Of particular interest would be a direct comparative study of the quality of public service deli-very in more and less decentralized regions, to comple-ment cross-country studies of decentralization and stu-dies of the impact of country-wide reforms. Cross-country studies, for instance those based on perfor-mance in standardized PISA examinations conducted by the OECD (Programme for International Student As-sessment), face the difficulty of meaningfully compar-ing outcome or input indicators across countries given vast differences in objectives, priorities, curricula, and sampling. In addition, production functions that trans-form various inputs in public service provision into out-puts may differ substantially across countries complicat-ing estimation and interpretation (see Nabeshima (2003) on education in South-East Asia). In the case of studies of system-wide reform initiatives, it is often difficult to credibly establish a counterfactual, i.e. the quality of public services in the absence of reforms.
Cross-regional studies, by contrast, combine reasonably comparable data on inputs and outcomes with substantial heterogeneity in some institutional ar-rangements, including the extent of fiscal decentraliza-tion-the role of municipal governments in public ser-vice provision relative to that of central regional gov-ernments. The latter approach was adopted by Galiani and Schargrodsky (2002) and Barankay and Lockwood (2007), and is followed below. …