The uncovered interest parity puzzle and time-varying risk premium

Uncovered interest parity puzzle is one of the most prominent puzzles in international finance that has remained unsolved for over 30 years. This theory stipulates that the currencies of the countries where risk-free interest rates are high should depreciate relative to the currencies associated with low interest rates. However, empirical findings of last three decades report the opposite effect, and none of the proposed theoretical models have been yet able to explain why. In the meantime, a popular trading strategy that exploits the puzzle and is commonly referred to as carry trade has been gaining popularity among traders and speculators. It on average generates significant positive excess returns that cannot be explained by any existing asset pricing models, but it also is susceptible to sudden crashes. Growing popularity of this strategy, and the persistence of the uncovered interest parity puzzle motivates us to analyze possible causes of the failure and to propose a model that corrects for the bias. Most of the literature in the field rely on the use of forward rates, but we argue that futures data provides the means to analyze evolution of the foreign exchange risk premium over time. In this thesis we use currency futures rates to perform the tests of the uncovered interest parity condition. We then test the hypothesis of futures rates unbiasedness and we link the two concepts together with a risk premium component. We find that both deviations from rational expectations and a presence of the risk premium term are possible causes for the failure of the UIP tests. Using futures rates at daily and weekly frequencies we are able to provide evidence that the risk premium varies over time. Incorporating this nature, we propose a modified UIP model that corrects for the bias. We find that the inclusion of the risk-premium component mitigates the puzzle. We also find that carry trade portfolios have high abnormal returns due to higher exposure to volatility and funding liquidity risks. During periods of market turmoils, the uncovered interest parity condition holds better because the trades that normally exploit the puzzle lose money. Overall, we conclude that there should be some non-traditional state variables that have their innovations related to the risk-premiums in a cross-section of currency returns