Does the interest parity puzzle hold for Central and Eastern European economies?

This paper examines the uncovered interest parity (or forward premium) puzzle in four Central and Eastern European countries -- Czechia, Hungary, Poland, and Romania -- as well as their aggregates from 1999 to 2019. Because the interest parity is a foundation of open-macroeconomy analyses, with important implications for policymaking, especially central banking, more systematic evidence on interest parities in the CEE economies is needed. In this study, we not only address this need but also add to a broader discussion on the UIP puzzle after the global financial crisis. The UIP is verified vis-à-vis three major currencies: the euro, the U.S. dollar, and the Swiss franc. We start by providing a full set of baseline forward premium regressions for which we examine possible structural breaks and perform a decomposition of deviations from the UIP. Next, we explore augmented UIP models and introduce various factors which potentially account for the UIP puzzle, such as the realized volatility of the exchange rate, a volatility model of the excess returns, and international risk and business cycle measures. The study shows that the choice of the reference currency matters for the outcome of the interest parity tests in the CEE economies. The puzzle prevails for the EUR and the CHF but not for the USD, a regularity that has not been documented in previous studies. Second, we find that structural breaks in the time series used to test the UIP are not an essential reason for the general failure of the parity in the region. Third, we demonstrate that even though the risk-based measures largely improve the baseline testing regression, both from statistical and economic points of view, they do not alter the overall outcomes of our empirical models. Additionally, we show that the exchange rate peg of the Czech koruna to the euro from 2013 to 2017 had a significant impact on the UIP. A detailed case study on Poland, using granular survey data, indicates that the directly measured exchange rate expectations do not seem to be informed by the UIP relationship. Employing data on option-implied risk reversals, we reveal that the limited resilience of CEE economies to rare disasters may plausibly explain deviations from the UIP.