Abstract
The forward premium puzzle, or violation of the uncovered interest rate parity (UIP), has been documented by many. Studies, using methodologies more sophisticated than regression analysis, tend to suggest that the rejection of UIP may be false, due to cointegration, biased parameter estimates, time-varying risk premium, etc. In this study, from data for the two decades (1979–1998) and nine major currencies of the developed world, it is shown that the forward premium puzzle is not always present for any arbitrary time period, or for any pair of countries, even using the regression analysis that established the puzzle in earlier works.
Acknowledgements
Some of these findings were initially made when I was completing my doctoral dissertation at Stanford University. I thank all members of my dissertation committee at Stanford—Peter Hammond, Peter Glynn, Darrell Duffie, and Michael Saunders, for their support in the dissertation process.