Off-campus WSU users: To download campus access dissertations, please use the following link to log into our proxy server with your WSU access ID and password, then click the "Off-campus Download" button below.
Non-WSU users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
Access Type
WSU Access
Date of Award
January 2023
Degree Type
Dissertation
Degree Name
Ph.D.
Department
Economics
First Advisor
Liang Hu
Abstract
Exchange-rate models based on macroeconomic fundamentals have consistently proven unable to outperform a simple, atheoretical random walk in out-of-sample forecasting. This is known as the Meese-Rogoff puzzle. The present work attempts to resolve this puzzle from two directions. Under the first approach, we introduce oil price volatility as a new macroeconomic fundamental for exchange-rate forecasting. We focus on oil-exporting countries with small, open economies for whom oil price and volatility shocks are likely exogenous. After estimating our model and generating one-step-ahead forecasts, we find that both oil price and volatility contain out-of-sample predictive content for the exchange rates of oil-exporting countries. We also find predictability for the direction of exchange-rate changes even where point forecasts do not outperform the random walk. These results are robust to different forecast evaluation methods, time periods, and volatility measures. Predictive content disappears when moving from daily to monthly frequency, suggesting that past failures to overturn the Meese-Rogoff puzzle may be due to temporal aggregation.
Under the second approach, we investigate the time-series properties of exchange rates. Since a Markov-switching (MS) process can easily be mistaken for a random walk and a long-memory (or fractionally integrated) process could account for the observed long swings in exchange rates, we estimate both types of models for ten U.S. dollar exchange rates over the period 1999 to 2021.We find good in-sample fit for the two-state MS model, with a dominant regime featuring low volatility and mean appreciation and a less common regime featuring higher volatility and mean depreciation. By contrast, although we find evidence of long memory for some currencies, in-sample fit is poor. We show that a MS process in exchange rates could be confused for a random walk and that this behavior is not better explained by fractional integration.
Recommended Citation
Breen, John, "Two Approaches To The Meese-Rogoff Puzzle" (2023). Wayne State University Dissertations. 3777.
https://digitalcommons.wayne.edu/oa_dissertations/3777