The Meese-Rogoff Puzzle: What Puzzle?

Abstract

The Messe-Rogoff puzzle has been a debatable topic since 1983 when Richard Meese and Kenneth Rogoff demonstrated that no exchange rate model can outperform the random walk in out-of-sample forecasting. This finding been taken to imply the weakness of international economics and finance and raised the question as to why firms spend money on exchange rate forecasts and use them as an input in the financial decision making process when these forecasts are that bad. In this study we resolve the puzzle by examining a number of propositions, including the following: (i) we should expect nothing but that exchange rate models cannot outperform the random walk in out-of-sample forecasting if forecasting accuracy is measured by metrics that depend on the magnitude of the forecasting error only; (ii) the use of dynamic models to outperform the random walk is inappropriate because it is tantamount to beating a random walk with a random walk; and (iii) it is possible to outperform the random walk in terms of other metrics such as direction accuracy and profitability.