Empirical Exchange Rate Models of the Seventies: Do They Fit Out of Sample?
Read the paperopens doi.org in a new tab
What they found
Meese and Rogoff took the leading structural models of exchange rates (based on money supplies, output, interest rates, and inflation) and tested how well they forecast the dollar against the mark, yen, and pound at horizons of one to twelve months. None of them beat a naive random walk, even when the models were given the actual future values of the economic variables they depend on. The 'Meese-Rogoff puzzle' remains one of the most durable results in international finance.
What you can use
- Even with perfect knowledge of future fundamentals, economic models could not forecast currencies better than assuming no change.
- If you trade FX on macro fundamentals at horizons under a year, the research says your model is unlikely to beat a coin flip.
- Exchange rates over short horizons are dominated by flows, positioning, and risk sentiment, not by textbook fundamentals.
Caveats
1970s data on three currency pairs; subsequent research finds some predictability at horizons of several years and in panels of many currencies. The result concerns point forecasts, not risk premia like carry.
Tags: forex, forecasting, random-walk, fundamentals
Summaries are our own reading of the paper, not the authors' words. Educational only, not advice. Discuss it in Book Club.