Public Expenditure and National Income Causality: Further Evidence on the Role of Omitted Variables

Abstract

Whether changes in public expenditure growth help predict changes in national income growth (and/or vice versa) remains an important issue of sustained interest in the empirical public finance literature. In recent years, attention has mainly been confined to two specific areas, namely, estimation of the impact of the public sector on output growth (by means of regression analysis) and causality testing. Unfortunately, the outcome of both types of analysis has been inconclusive. Focussing on the impact studies, we note that a number of empirical investigations have attempted to measure the effect of the size of the public sector on overall economic growth. For instance, Ram [19], utilizing a two-sector model, found that growth of government size has a positive effect on economic growth. Landau [14; 15], however, presented opposite evidence, indicating that the government sector expansion led to a decline in output growth for many DCs and LDCs since 1960. In a recent study, Barth, Keleher and Russek [4] estimated variants of Ram's models with data for 30 countries. Their empirical results largely support a negative relationship between the scale of government and aggregate economic activity. Furthermore, using the Hausman procedure, they rejected the hypothesis that the independent variables (namely, real government spending or other measures of the scale of the government, depending on the specification) in the regression equations are exogenous. This strongly suggests that the conclusions reached by