Fölster and Henrekson (European Economic Review 45 (2001), 1501–1520) argue that “...the more the econometric problems that are addressed, the more robust the relationship between government size and economic growth appears”. But in failing to control for simultaneity and in ignoring issues of sample-selection bias, the regressions reported by Fölster/Henrekson are flawed. Using theoretically valid instruments, we find that the estimated partial correlation between size of the public sector and economic growth is statistically insignificant and highly unstable across specifications. Moreover, since instruments are weak, all hypothesis tests are unreliable. We conclude that cross-country growth regressions are unlikely to come up with a reliable answer to the question of the growth effects of government spending and taxation.