Empirical Exchange Rate Models of the Nineties: Are Any Fit to Survive?
IMF Working Papers, April 1, 2004
Source details
- Canonical URL
- Empirical Exchange Rate Models of the Nineties: Are Any Fit to Survive?
Other formats
Bibliographic details
- Authors: Yin-Wong Cheung, Menzie David Chinn
- Published: April 1, 2004
- Series: IMF Working Papers
- DOI: https://doi.org/10.5089/9781451849493.001
Summary
- Reassessment of exchange rate prediction using a wider set of models proposed in the last decade.
- Performance compared against two reference specifications: purchasing power parity and the sticky-price monetary model.
- Models estimated in first-difference and error-correction specifications.
- Forecast horizons evaluated: 1, 4, and 20 quarters.
- Evaluation metrics: mean squared error, direction of change metrics, and the "consistency" test of Cheung and Chinn (1998).
- Overall conclusion: model/specification/currency combinations that work well in one period do not necessarily work well in another period.
Methodology
- Models examined include a wide set of proposals from the literature of the previous decade (1990s).
- Reference specifications used for benchmarking:
- Purchasing power parity.
- Sticky-price monetary model.
- Estimation approaches:
- First-difference specification.
- Error-correction specification.
- Forecast horizons:
- 1 quarter.
- 4 quarters.
- 20 quarters.
- Performance metrics:
- Mean squared error (MSE).
- Direction of change metrics.
- "Consistency" test of Cheung and Chinn (1998).
Key Findings
- No single model/specification/currency combination consistently outperforms others across periods.
- Performance is period-dependent: methods that perform well in one period may fail in another.
- Evaluation across multiple horizons and metrics necessary to assess robustness of exchange rate models.