Most growth models incorporate the deep-seated belief that fiscal policy can promote growth miracles and lead to enduring stagnation, as for example when a rise in income tax reduces net returns to private investment. In Discussion Paper No. 885, William Easterly and Research Fellow Sérgio Rebelo use cross-section data on approximately 100 countries for 1970-88 and historical data for 28 countries over 1870-1988 to examine the association between fiscal policy, the level of development and growth. They construct per capita income growth rates and regress them on initial per capita income, school enrolment ratios and proxies for political instability. They then incorporate various measures of fiscal policy which prove insignificant and often render the initial income level insignificant as well. Controlling for the initial income level causes a strong, negative correlation between growth and the ratio of tax revenues to GDP to disappear.

Easterly and Rebelo also find a strong, positive correlation between investment in transport and communications and growth. The government budget surplus is also consistently correlated with growth and private investment for the cross-section sample. Results for the other fiscal variables are statistically fragile and heavily dependent on the choice of control variables, however, since they are highly correlated with initial income and among themselves. For both data sets, the ratio of government revenue to GDP and the share of taxes on income relative to international trade both rise with per capita income; a rise in the population has the same effect, even after controlling for income and the trade share.

Easterly and Rebelo conclude that the observed empirical regularities concerning population and income growth suggest that the scale of the economy may influence fiscal structure independently of its well-documented dependence on per capita income. The empirical fragility of their results on tax rates contrasts sharply with the robustness of theoretical predictions in most models, while their results on investment in transport and communications support Aschauer's contention that public spending on infrastructure yields supernormal returns.

Fiscal Policy and Economic Growth: An Empirical Investigation
William Easterly and Sérgio Rebelo

Discussion Paper No. 885, January 1994 (IM)