Medium-Term Exchange Rate Forecasting: What Can We Expect?
IMF Working Papers, January 1, 2003
Source details
- Canonical URL
- Medium-Term Exchange Rate Forecasting: What Can We Expect?
Other formats
Bibliographic details
- Authors: Guy M Meredith
- Published: January 1, 2003
- Series: IMF Working Papers
- DOI: https://doi.org/10.5089/9781451843934.001
Summary of research question and models
- Examines the medium-term predictability of exchange rate movements using three models of fundamentals:
- purchasing power parity
- the monetary model
- uncovered interest parity
Main empirical findings
- Purchasing power parity and the monetary model:
- Yield favorable in-sample results.
- These in-sample results largely reflect finite-sample estimation biases.
- After adjusting for finite-sample biases, there is little evidence of predictability.
- Out-of-sample forecasting performance:
- There is a lack of systematic improvement in out-of-sample forecasting performance relative to a random walk.
- Uncovered interest parity:
- Fares better at long horizons.
- Reflects information already embodied in market prices and therefore may not be useful as an indicator of exchange rate misalignment.
Interpretation and implications
- The apparent predictability from simpler fundamentals models can be driven by finite-sample biases rather than true forecasting power.
- Models that appear successful in-sample should be assessed for finite-sample biases before inferring medium-term predictability.
- Even when a model (uncovered interest parity) performs better at long horizons, its usefulness as a misalignment indicator is limited if it merely reflects market prices.
Cautions and recommendations for future work
- More elaborate models of fundamentals might have better medium-term forecasting properties.
- Careful attention must be paid to finite-sample biases in assessing predictability.
Content in this bundle
- Medium-Term Exchange Rate Forecasting: What Can We Expect? - WP/03/21