How big is the premium for currency risk?1We thank Geert Bekaert, Tim Bollerslev, Peter Bossaerts, Mark Carhart, John Cochrane, Magnus Dahlquist, Wayne Ferson, Linda Goldberg, Campbell Harvey, Pierre Hillion, Robert Hodrick (the referee), Olivier Ledoit, John Matsusaka, Hans Mikkelsen, Angel Serrat and Ivo Welch, as well as workshop participants at INSEAD, University of California at Los Angeles, University of Southern California, University of Rochester, Southern Methodist University, University of California – Irvine, Koc University and participants at the 1996 UBC Global Investment Conference (Whistler, BC), 1996 Western Finance Association meetings (Sunriver, OR), 1996 European Finance Association meetings (Oslo, Norway), 1996 NBER Asset Pricing meeting (Evanston, IL), and the 1997 Econometric Society winter meetings (New Orleans, LA) for their comments. The paper was written while the second author was visiting the Anderson School at UCLA. Both authors acknowledge the financial support of a CIBEAR grant.1
We estimate and test the conditional version of an International Capital Asset Pricing Model using a parsimonious multivariate GARCH process. Since our approach is fully parametric, we can recover any quantity that is a function of the first two conditional moments. Our findings strongly support a model which includes both market and foreign exchange risk. However, both sources of risk are only detected when their prices are allowed to change over time. The evidence also indicates that, with the exception of the U.S. equity market, the premium for bearing currency risk often represents a significant fraction of the total premium.