(World Bank, Washington, DC, 2014-02) Dasgupta, Basab; Skoufias, Emmanuel; Narayan, Ambar; Kaiser, Kai
This paper takes advantage of the exogenous phasing of direct elections in districts and applies the double-difference estimator to measure impacts on (i) human development outcomes and (ii) the pattern of public spending and revenue generation at the district level. The analysis reveals that four years after the switch to direct elections, there have been no significant effects on human development outcomes. However, the estimates of the impact of Pilkada on health expenditures at the district level suggest that directly elected district officials may have become more responsive to local needs at least in the area of health. The composition of district expenditures changes considerably during the year and sometimes the year before the elections, shifting toward expenditure categories that allow incumbent district heads running as candidates in the direct elections to "buy" voter support. Electoral reforms did not lead to higher revenue generation from own sources and had no effect on the budget surplus of districts with directly elected heads.
