This paper empirically investigates the impacts of fiscal policy (represented by changes in government spending relative to GDP) on economic activity at both economic growth and business cycle frequencies for Latin American, Asian, and European countries.
Using a panel data set for 1970-2006, the paper estimates both short-and long-run effects of government spending on growth thorough the pooled mean group (PMG) estimator, introduced by Pesaran, Shin and Smith (1999). The differential impacts that the various components of government spending have on economic performance at different time horizons are tested on the entire sample.
The main results of the paper are the following. First, expansionary fiscal policy in the form of government spending on infrastructure (transport and communications, and energy) is favorable to real GDP per capita growth at business cycle frequencies, but it does not necessarily mean that it spurs the long-run economic performance as well. The long-run impacts of energy spending on growth are also positive, while those of transport and communications spending are strongly negative. This result is in contrast to a series of theoretical and empirical studies which predicted the positive impacts of infrastructure spending on economic growth, but is in line with the finding of Devarajan et al. (1996). Devarajan et al. explained the negative impacts of government capital expenditure on growth might be caused by the excessive use of government capital expenditure. In other words, capital expenditure which is normally considered growth enhancing, could become unproductive if it is used excessively and inefficiently. The negative effect of transport and communications spending might also indicate that the level of government spending is above the optimal value so that government spending no longer contributes to macroeconomic performance but it rather adversely affects the economic activity.
Second, social spending such as education and health appears to have a long-run positive relation with growth. However, the short-run impacts of education spending on growth are negative. The result might be interpreted that the returns on government spending on human capital will become measureable only in the long-run. Moreover, the short-run impacts of health spending are positive as long as an increase in the share of health spending is adjusted by a decrease in other spending shares so that total expenditure remains constant.
Finally, increasing government consumption expenditure is negatively correlated to economic growth both in the long-and short-runs.