Decentralisation or Patronage : What Determines Government's Allocation of Development Spending in a Unitary Country? Evidence from Bangladesh

This paper contributes to the decentralisation and distributive politics literature by empirically investigating the determinants of public expenditure at the sub-national level in Bangladesh. We argue that fragmentation in a unitary developing country may not channel higher resources to local areas. However, in such countries, political motive may play a significant role in the allocation process. Using panel data methods and a novel dataset on government’s district-wise allocation of annual development expenditure in Bangladesh covering the period from 2005 to 2009, the analysis focuses on the impact of local government fragmentation and tests key political distribution models (the core voter hypothesis , the swing voter hypothesis , and the political alignment theory ). The results show that local government fragmentation does not have any significant impact on public spending at the district level. However, the core vote share, local elected representative’s political alignment with the ruling party and the raw number of ministers from a district are all significantly associated with higher expenditure growth rate, supporting the idea that political patronage matters for resource allocation. No evidence was found in support of the swing voter hypothesis. Overall, findings of the study suggest that the local units should be strengthened fiscally to ensure an effective, inclusive, and participatory local government system. In the case of public resource allocation, more emphasis should be put on equity and need-based criteria rather than political motives