This was a paper that I submitted for my “Applied Analytical Statistics” class for the MSc in Social Data Science in Oxford. Final grade: 78
Summary:
In settings where resources are limited as is the case in developing countries like the Philippines, local governments rely on national revenues to provide for the basic needs of its citizens. For instance, in the poorest segments of the Philippines where sources of local income are scarce, reliance on revenue allotments from the central government is as high as 90% (Marzan, 2011).
In many cases, the distribution of national income to local units is formulaic and mandated by law. However in recent years, there persists a growing body of literature—particularly in low and middle-income economies—suggesting the influence of politics in intergovernment resource allocation. My goal for this paper is to contribute to this understanding of political dynamics in fiscal settings.
I particularly explore one strand in this evolving discourse: that national governments use its resources as a “prize” to politically-aligned localities. Given data from the most recent Philippine presidential election, I observe whether municipalities and cities that supported the president-incumbent in the 2016 election subsequently received a higher budgetary allocation from the national government.
Using a regression discontinuity research design, I identified the localities in which the incumbent barely won and barely lost; then quantified the impact on intergovernment transfers between these two groups. The aggregate results of my study show an indication of discontinuity, that is, more was given to the municipalities where the president won. Further, there is evidence of some regional heterogeneity, as it can be seen that locations where the president did not expect to win were “awarded” with a higher fiscal allocation in the following year; compared to all other municipalities which received their legally-mandated endowments. Though observed, these results are inconclusive as they fail to show a statistically-significant p-value. This may be the case because, ultimately, resource allocation to local governments is mandated by law, and while we observe differences in the resource distributions to several municipalities, on the aggregate, these changes need not be significant.
Full version of the paper: email me on LinkedIn!
