I use a panel data set covering 26 OECD countries between 1960 and 2004 to revisit the consumption–real exchange rate anomaly. After using demographic variables (in particular, the fertility rate) as instruments, I document a new empirical regularity that higher real exchange rates are associated with higher relative consumption. This positive relationship is statistically significant and has been shown to be robust. Such a finding may suggest that the theoretical prediction in Backus and Smith (Journal of International Economics, Vol. 35 (1993), pp. 297–316) is no longer at odds with the data.