Alesina and Wacziarg (1998, p. 312–313) offer two possible explanations for the counterintuitive result from Table 3.2 that the explanatory power as well as the size of the effect has increased over time: –Newly de-colonized countries in the 60ies may have had a lot to catch up on on their way to establishing their public sectors, which could have covered up the effect of country size on government size. –The figures for the 60ies and partly for the 70ies are estimations for some countries and could, therefore, lack precision. Both explanations are not entirely convincing, especially when data for the 90ies are considered. A further intuitive explanation would be that the size effect has actually become more important over time, but one has to distinguish between the public and the private sector. Suppose the following development within, say, the last two or three decades: The provision prices of public goods have increased due to higher demands 57
Empirical evidence Table 3.2: OLS regressions for government consumption and log popu - lation with control variables (development) Dependent variable: governmentPeriod/yearLogpopAdj. R2Number of consumption in percent of GDPobservations 1960–64–0.3110.16118 (–0.86) 1965–69–0.1580.17119 (–0.44) 1970–74–0.4070.22124 (–1.02) 1975–79–0.875*0.26125 (-1.90) 1980–84–1.235**0.35130 (–3.46) 1985–89–1.121**0.34134 (–3.39) 1993–97–2.847**0.39103 (–2.870) Other control variables (which are not shown) are the same as in Table A.6, column (7). ** significant at the 1% level; * significant at the 5% level; t statistics based on White heteroscedasti - city-consistent standard errors in
parentheses
t statistics in parentheses Results for 1960–1989 from Alesina and Wacziarg (1998, p. 313).