The paper estimates currency risk premia for the Czech Republic, Hungary, Poland and Slovakia. Three different approaches are applied: a constant premium approach based on rational expectations, while time-varying premia are estimated with a method using financial market analysts’ surveys and also with a Kalman filter technique. A novelty in this paper is a crosscheck based on the three different approaches applied and also making use of implied and historical volatilities. The results highlight the importance of such a crosscheck: in the case of the Czech and the Slovak koruna and the Polish zloty this exercise reveals severe problems with the results, which otherwise would not have been discovered. On the other hand, the estimation methods produce convincing results for the Hungarian forint. The estimated Hungarian premium series reflect the major events that intuitively may have shaped currency risk in the country. A possible reason for these findings is a high signal-to-noise ratio in the case of Hungary where the risk premium has been large and exhibited substantial shifts through time. Finally, the strong comovement of the premium series obtained with the Kalman-filter and the survey data for the Hungarian forint also indicates that the survey expectations are largely in line with both the riskpremium- extended UIP and the rational expectations hypothesis, which is theoretically important as the UIP relates exchange rate expectations to the interest rate differential.
JEL: C30, C42, F31, G15.
Keywords: risk premium, exchange rate, Kalman filter, survey data.