Testing the quantity theory using long-run averaged cross-country data

Using data from Barro (1990), Dwyer and Hafer (1988), Duck (1993) and Vogel (1974), we revisit the finding that cross-sectional regressions of long-run average inflation on money growth and real income growth support the quantity theory, and conclude that, as is frequently argued, this depends on the inclusion in the sample of a few countries with very high money growth. The most likely reason for the rejection of the theory when these data points are excluded is simultaneity bias, the importance of which is mitigated when high-inflation countries are included in the sample. Omitted variables bias may also play a role, but measurement errors are unlikely to do so.