Zusammenfassung:
We investigate the potential presence of time variation in the coefficients of the ''Fama regression'' for Uncovered InterestRate Parity. We implement coefficient constancy tests, rolling regression techniques, and stochastic coefficient modelsbased on state space modelling. Among six major US bilateral exchange rates we find significant evidence for stochastictime variation.Using the statistical equivalence between stochastically varying coefficients and conditional heteroscedasticity we derivea proxy for time-varying 'risk', and investigate whether it explains the well known negative bias or foreign discount biaspuzzle in the foreign exchange rate literature. We contrast our identification scheme to the ARCH-in-mean approach forempirically identifying risk premia.