Uncovered Interest Parity Puzzle: Cross-sectional and Time-series Analysis

Abstract

This paper proposes a new explanationfor the UIP puzzle by analyzing a large number of cross-country bilateral exchange rates in two dimensions, cross-sectional and time-series. The exchange rates analyzed here includes a broad spectrum of developed and developing countries. The UIP relationship holds up well in cross-sectional analysis and the slope estimates remain largely between zero and one throughout the sample periods, with a few exceptions. There does not appear to be a well-publicized UIP puzzle for cross-sectional UIP. For time-series UIP, short-term (one month) UIP holds up well and UIP puzzle is largely confined to the key currencies. We introduce the key currency bias to explain the empirical failure of UIP in these cases. The key currency concept is a similar to the home bias for portfolio holdings. UIP seems to fail more often when a key currency is involved in the bilateral exchange rate relationship than when only non-key currencies are involved, especially when the key currency offers higher return on capital. This paper presents an empirical evidence for a state-dependent asymmetric response in exchange rate changes depending on the direction of the forward premium.