An empirical puzzle in international finance is the finding of a high correlation between national savings and domestic investment. This result is widely interpreted as evidence of low international capital mobility. In this paper we examine the long run behaviour of national savings and domestic investment for the United States and Canada in the time series context, by testing for evidence of cointegration between the two series. Overall we find little evidence of a cointegrating relationship between the levels of national savings and domestic investment, except for Canada during the Bretton Woods era.