Three Essays in Forward Rate Unbiasedness Hypothesis

• Main Author: Chatterjee, Devalina
• Format: Others
• Published: DigitalCommons@USU 2010
• Subjects: essays; forward rate; unbiasedness hypothesis; Economics; Finance
• Online Access: https://digitalcommons.usu.edu/etd/644; https://digitalcommons.usu.edu/cgi/viewcontent.cgi?article=1640&context=etd

The objective of this dissertation is to verify and explain the forward exchange rate unbiasedness hypothesis in the foreign exchange market. Since in most of the cases the unbiasedness hypothesis fails to hold, we try to provide three different explanations of this puzzling behavior in the three essays. The first essay tries to resolve the forward premium puzzle by addressing the model misspecification issue and thereby adding a time-varying risk premium term in the percentage change specification. The risk premium term is modeled using the GARCH-M representation and the model is estimated by applying a GARCH (1, 1) specification. The second essay attributes the failure of the unbiasedness hypothesis to hold to the nonstationarity of the spot and forward exchange rate. It verifies the existence of a cointegrating relationship between the spot and the forward exchange rates and thus specifies an Error Correction Model to better capture the relation between the spot and the forward rates. Further, a cointegrating or the existence of a long run relationship between the spot and forward exchange rates and the domestic and foreign interest rates is tested. It can be viewed as a robustness check where we ensure whether the cointegrated exchange rates are still related in the long run with the inclusion of the interest rates. The objective of the third essay is to apply the generalized method of moments (GMM) to test the unbiasedness hypothesis in the foreign exchange market. Empirical evidence suggests that the spot and forward rates are nonstationary with unit roots and are cointegrated. Cointegration further suggests that the changes in the spot rate can be modeled by an Error Correction Model. The third essay explicitly derives an ECM from the levels specification and uses the GMM estimation technique to test the unbiasedness hypothesis.