Macroeconomic determinants of fiscal deficits in East Africa

Abstract

This thesis examines the macroeconomic determinants of the fiscal deficit and its tax revenue and public expenditure components in a panel of five East African countries, namely: Burundi, Kenya, Rwanda, Tanzania, and Uganda. The empirical analysis uses both linear and nonlinear panel autoregressive distributed lag models, panel Granger causality tests, and panel impulse response functions. The analysis is based on annual panel data for the period 1980–2019. The results show that the fiscal deficit is positively associated with real per capita GDP, regardless of whether this relationship is estimated during growth accelerations or growth decelerations. While the study concludes that the fiscal deficit is procyclical, the findings do not provide sufficient evidence to conclude that this relationship is nonlinear. In addition, the fiscal deficit is positively associated with the current account and interest rates, and negatively associated with inflation, grants, and debt service. With regard to the components of the fiscal deficit, the results show that both tax revenue and public expenditure are positively associated with real per capita GDP, both during growth accelerations and decelerations. The study thus concludes that tax revenue is countercyclical while government expenditure is procyclical. However, the study findings do not provide sufficient evidence to conclude that the relationship between real per capita GDP and both tax revenue and total government expenditure is nonlinear. In addition, government consumption expenditure is shown to be acyclical while government investment expenditure is procyclical. Multivariate Granger causality results show long-run feedback causality between the fiscal deficit and current account balance, the GDP deflator, interest rates, and grants. In addition, the study findings show long-run unidirectional causality running from real per capita GDP growth to fiscal deficits, and from debt service to fiscal deficits. Short-run Granger causality results indicate one-way causality running from real GDP growth to fiscal deficits, and from fiscal deficits to the GDP deflator. The results from the impulse response functions and multivariate Granger causality tests are qualitatively similar.