The Feldstein‐Horioka puzzle in an ARIMA framework

To test the Feldstein‐Horioka hypothesis that the investment‐to‐output ratio moves one‐for‐one with the saving‐to‐output ratio, suggesting international capital mobility.

The paper uses the econometric framework developed by Fisher and Seater, interpreting the Feldstein‐Horioka hypothesis as a long‐run phenomenon, and paying particular attention to the integration properties of the data, since meaningful tests critically depend on these properties. The paper also investigates the power of the long‐horizon regression tests, using the inverse power function of Andrews.

The paper tests the Feldstein‐Horioka hypothesis for 15 European countries, as well as for the USA and Japan, using annual data for the period from 1960 to 2002. Evidence is found against the Feldstein and Horioka hypothesis of low international capital mobility.

Although the findings are in contrast to those of Feldstein and Horioka, they are consistent with neoclassical growth theory according to which there is no reason to expect a relation between saving and investment if there are no barriers to capital movements.