Essays on Optimal Aid and Fiscal Policy in Developing Economies

TITLE OF DISSERTATION: ESSAYS ON OPTIMAL AID AND FISCAL POLICY IN DEVELOPING ECONOMIES Ryan Niladri Banerjee, Doctor of Philosophy, 2010 Dissertation directed by: Professor Enrique G. Mendoza Department of Economics Essay I: Which countries receive aid as insurance and why? A theory of optimal aid policy Empirical evidence shows that developing countries with opaque institutions receive procyclical Official Development Aid (ODA) while developing countries with transparent institutions receive acyclical or countercyclical ODA. This paper provides a dynamic equilibrium model of optimal aid policy that quantitatively accounts for this fact. In the model, the donor wants to (a) encourage actions by the aid receiving government that increase output and (b) smooth out economic fluctuations. The transparency of institutions in the country affects the donor’s ability to distinguish downturns caused by exogenous shocks, from those caused by government actions. The solution to the donor’s mechanism design problem is dependent on the transparency of government actions. If the donor has good information about government actions, aid policy is countercyclical and aid acts as insurance. However, if the donor is unable to infer perfectly the cause of the downturn, aid policy is procyclical to encourage unobservable good actions. The model predicts a similar pattern for ODA commitments for the following year which is supported by the data. For countries with opaque institutions procyclical aid is the result of optimal policies given the information constraints of donors. Essay II: New Evidence on the Relationship Between Aid Cyclicality and Institutions This paper documents a new fact: the correlation between official development assistance (ODA) and GDP is negatively related to the quality of institutions in the receipient country. Differences in institutional indicators that measure corruption, rule of law, government effectiveness and government transparency are particularly important. The results are robust to several modifications. The results hold for both pooled and within regressions specifications and for different sources of institutional quality measures. This fact also reconciles conflicting empirical results about the correlation between ODA and GDP in the literature. For instance, Pallage and Robe (2001) find a positive correlation in two thirds of African economies and half of non-African developing economies, but Rand and Tarp (2002) find no correlation in a different set of developing countries. First, once institutions are accounted for, African economies are not treated differently by donors. Second, the sample in Rand and Tarp (2002) comprises developing economies which have relatively good institutions, therefore, those countries receive acyclical or countercyclical aid. Essay II: Optimal Procyclical Fiscal Policy Without Procyclical Government Spending Procyclical fiscal policy can be caused by either procyclical government expenditure, countercyclical taxes or both. The majority of models which try to explain procyclical fiscal policy as the result of optimal policy have procyclical government expenditures. This paper develops a model which optimally generates procyclical fiscal policy while keeping government expenditures acyclical. Instead, taxes are optimally countercyclical. The model uses endogenous sovereign default to generate an environment where interest rates are lower in booms than in recessions. If household’s have insufficient access to financial instruments it is optimal for the government to lower taxes and borrow during booms. This enables impatient households to benefit from the lower interest rates in booms by helping the consumer bring consumption forward. ESSAYS ON OPTIMAL AID AND FISCAL POLICY IN DEVELOPING ECONOMIES