a negative relationship between government consumption (as a proxy for government size) and population (as a proxy for country size), most of the countries should lie on or near the diagonal. As can be easily seen, 83 countries (69% of all countries) are situated in the squares of the dia- gonal and in the vicinity of it (directly below or above the diagonal).45 The simple correlation coefficient between the two variables displayed is –0.345 and significant at the 0.1% level.46Nineteen countries exhibiting low government consumption and few inhabitants (southwest of the diagonal) do not show any obvious pattern. Conversely, there may be an apparent rationale for those 18 coun- tries northeast of the diagonal, which display relatively high government consumption expenditure. Ten of those 18 are members of the OECD; namely Sweden, Australia, Canada, Poland, the U.K., Spain, Italy, France, Germany and the U.S.A. Moreover, 86% of the OECD mem- bers can be found on the diagonal or northeast of it. This suggests that OECD members have significantly higher relative levels of government consumption expenditure and therefore larger governments than other countries of similar size. A non-parametric test reveals that government consumption is indeed higher in OECD countries than in the rest (Mann-Whitney-U-Test; two-sided; p = 0.003). The distributions are also signi ficantly different from each other (Kolmogorov-Smirnov-Test; two-sided; p = 0.011). We will control for this difference in the regressi- ons by means of a dummy for OECD membership. It has to be mentioned that we do not intend to address the rela - tionship between trade openness and country size here in greater detail.47 Clearly, countries which are more open to trade can more easily «afford» to be small (or to split up) than countries with little trade and bad rela- tions with adjacent countries. The explanation for this is straightfor- ward. The more open a country is, the more it can exploit economies of scale effects and effects due to specialization in the production of pri vate goods. It simply follows the rule of comparative advantage and pro duces whatever it can sell on world markets.48 46Does 
country size matter for public sector size? 45Countries outside of this area are printed in capital letters. 46Because of the right-skewedness of the distribution of «population», its logarithm is used. 47See, for instance, Rodrik (1996). 48See Section 3.2.4 for a short discussion on openness.