International Capital Mobility Revisited: Resurrecting the Feldstein-Horioka Hypothesis*

In this paper we resolve the long standing Feldstein-Horioka (FH, 1980) puzzle dynamically and explain it statistically. We first study the dynamic properties of capital mobility, and re-estimate the evolution of the FH coefficient using a time-varying-parameter model for each of 67 countries from 1970 to 2009. We find that the trend of the FH coefficient is highly consistent with increased global financial integration. We then rigorously test this notion by examining the relation between our new dynamic estimate of the FH coefficient, per year and per country, and a measure of global financial integration by Lane and Milesi-Ferretti (2007). We are able to reaffirm that, in this dynamic sense the FH puzzle is "resolved". Next, we re-focus on the FH puzzle in cross-sectional sense. We argue that the FH coefficient contains information not only about capital mobility, but also capital market imperfection. We examine this argument by considering a capital market openness index (Chinn and Ito, 2008). Our empirical findings show that financial market imperfections influence the viability of FH coefficient in measuring capital mobility. Specifically, across countries, the FH coefficient is more likely to reveal information about international capital mobility only in deregulated financial markets.