Can Financial Frictions Account for the Cross-Section Feldstein-Horioka Puzzle?

Joint with: Yan Bai

This paper revisits the Feldstein and Horioka finding, the high correlation between long-period averages of savings rates and investment rates across countries. We first confirm this finding remains using updated data. We then show that a calibrated frictionless complete markets model generates a cross-section savings-investment correlation close to zero, as Feldstein and Horioka predicted. We next propose a bond-enforcement model with two types of financial frictions. One is the limited span friction, under which countries can only trade uncontingent bonds. The other is the limited enforcement friction, under which

debt contracts are supported by reputation and sanction. The interaction between these two frictions helps us account for the Feldstein-Horioka finding, while each friction alone fails remarkably. We also find that the reputation mechanism itself cannot generate observed capital flows and the sanction mechanism is quantitatively important in supporting borrowing and lending across countries.

For more information, contact Stephen Yeaple.