Handbook of Exchange Rates

14.3 Empirical Evidence

We have reviewed in the previous section that under fairly general conditions, changes in exchange rates and currency risk premia directly depend on the properties of SDFs. This clear link motivates empirical studies of exchange rates using financial tools and reasonings. Along this line, in this section, we first describe a key methodological innovation, that is, the use of currency portfolios, and then turn to recent findings on exchange rates based on this new method.

14.3.1 From UIP Regressions to Currency Portfolios

UIP Regressions

As already noted, it is well known since the work of Hansen and Hodrick (1980) and Fama (1984) that the UIP condition is flatly rejected by the data.³ Previous work on currency risk started from UIP tests, that is, from regressions of changes in exchange rates on a constant and interest rate differentials. Since the constants are not always 0 and the slope coefficients are never 1 and very often negative, UIP tests suggest that these two coefficients are needed to determine expected currency excess returns. They imply that investors obtain excess returns when investing in higher-than-usual interest rate currencies and that interest rate elasticity is key. As a result, in order to study currency excess returns and thus currency risk, researchers had first to estimate UIP-like equations.

Currency Portfolios

Lustig and Verdelhan (2005) are the first to propose an alternative approach to currency risk. They show that building ...