Bayesian Model Averaging and exchange rate forecasts
Introduction
Out-of-sample forecasting of exchange rates is hard. Meese and Rogoff (1983) argued that all exchange rate models do less well in out-of-sample forecasting exercises than a simple driftless random walk. Although this finding was heresy to many at the time that Meese and Rogoff wrote their paper, it has now become the conventional wisdom. Mark (1995) claimed that a monetary fundamentals model can generate better out-of-sample forecasting performance at long horizons, but that result has been found to be very sensitive to the sample period (Groen, 1999, Faust et al., 2003). Claims that a particular variable has predictive power for exchange rates crop up frequently, but these results typically apply just to a particular exchange rate and a particular subsample. As such, they are by now met with justifiable skepticism and are thought of by many as the result of data-mining exercises.
However, in many contexts, researchers have recently made substantial progress in the econometrics of forecasting using large datasets (i.e. a large number of predictors). The trick is to combine the information in these different variables in a judicious way that avoids the estimation of a large number of unrestricted parameters. Bayesian VARs have been found to be useful in forecasting: these often use many time series, but impose a prior that many of the coefficients in the VAR are close to zero. Approaches in which the researcher estimates a small number of factors from a large dataset and forecasts using these estimated factors have also been shown to be capable of superior predictive performance (see for example Stock and Watson (2002) and Bernanke and Boivin (2003)). Stock and Watson, 2001, Stock and Watson, 2004 obtained good results in out-of-sample prediction of international output growth and inflation by taking forecasts from a large number of different simple models and just averaging them. They found the good performance of simple model averaging to be remarkably consistent across subperiods and across countries. The basic idea that forecast combination outperforms any individual forecast is part of the folklore of economic forecasting, going back to Bates and Granger (1969). It is of course crucial to the result that the researcher just averages the forecasts (or takes a median or trimmed mean). It is, in particular, tempting to run a forecast evaluation regression in which the weights on the different forecasts are estimated as free parameters. While this leads to a better in-sample fit, it gives less good out-of-sample prediction.
Bayesian Model Averaging (BMA) is another method for forecasting with large datasets that has received considerable recent attention in both the statistics and econometrics literature. The idea is to take forecasts from many different models, and to assume that one of them is the true model, but that the researcher does not know which this is. The researcher starts from a prior about which model is true and computes the posterior probabilities that each model is the true one. The forecasts from all the models are then weighted by these posterior probabilities. It has been used in a number of econometric applications, including output growth forecasting (Min and Zellner, 1993, Koop and Potter, 2003), cross-country growth regressions (Doppelhofer et al., 2000, Fernandez et al., 2001) and stock return prediction (Avramov, 2002, Cremers, 2002). Avarmov and Cremers both report improved pseudo-out-of-sample predictive performance from BMA.
The contribution of this paper is to argue that BMA is useful for out-of-sample forecasting of exchange rates in the last fifteen years although the magnitude of the improvement that it offers relative to the random walk benchmark is quite small.
One does not have to be a subjectivist Bayesian to believe in the usefulness of BMA, or of Bayesian shrinkage techniques more generally. A frequentist econometrician can interpret these methods as pragmatic devices that may be useful for out-of-sample forecasting in the face of model and parameter uncertainty.1 The plan for the remainder of the paper is as follows. Section 2 describes the idea of BMA. The out-of-sample exchange rate prediction exercise is described in Section 3. Section 4 concludes.
Section snippets
Bayesian model averaging
The idea of Bayesian Model Averaging was set out by Leamer (1978), and has recently received a lot of attention in the statistics literature, including in particular Raftery et al. (1997), Hoeting et al. (1999) and Chipman et al. (2001).
Consider a set of models . The th model is indexed by a parameter vector . The researcher knows that one of these models is the true model, but does not know which one.2
Exchange rate forecasting
I consider prediction of the bilateral exchange value of the Canadian dollar, pound, yen and mark/euro, relative to the US dollar, using both direct forecasting and iterated multistep forecasting. Iterated forecasts are more efficient if the model is correctly specified, but direct forecasts may be more robust to model misspecification (see Marcellino et al. (2006)).
Conclusion
In this paper I have considered a specific approach to pooling the forecasts from different models, namely Bayesian Model Averaging, and argued that it gives promising results for out-of-sample exchange rate prediction. For most currency–horizon pairs, with a sufficiently high degree of shrinkage, the Bayesian Model Averaging forecasts yield mean square prediction errors modestly lower than one obtains from a random walk forecast. Though the improvements are modest, this nonetheless stands in
Acknowledgment
I am grateful to Ben Bernanke, David Bowman, Jon Faust, Matt Pritsker and Pietro Veronesi for helpful comments, and to Sergey Chernenko for excellent research assistance. The views in this paper are solely the responsibility of the author.
References (35)
Stock return predictability and model uncertainty
Journal of Financial Economics
(2002)- et al.
Monetary policy in a data-rich environment
Journal of Monetary Economics
(2003) The out-of-sample success of term structure models: A step beyond
Journal of International Economics
(2003)- et al.
Exchange rate forecasting: The errors we’ve really made
Journal of International Economics
(2003) - et al.
Comparing dynamic equilibrium models to data: A Bayesian approach
Journal of Econometrics
(2004) - et al.
Why is it so difficult to beat the random walk forecast of exchange rates?
Journal of International Economics
(2003) - et al.
A comparison of direct and iterated multistep AR methods for forecasting macroeconomic time series
Journal of Econometrics
(2006) - et al.
Empirical exchange rate models of the seventies: Do they fit out of sample?
Journal of International Economics
(1983) - et al.
Bayesian and non-Bayesian methods for combining models and forecasts with applications to forecasting international growth rates
Journal of Econometrics
(1993) - et al.
The combination of forecasts
Operations Research Quarterly
(1969)
Characterizing predictable components in excess returns on equity and foreign exchange markets
Journal of Finance
Stock return predictability: A Bayesian model selection perspective
Review of Financial Studies
Expectations and exchange rate dynamics
Journal of Political Economy
Long swings in the dollar: Are they in the data and do markets know it?
American Economic Review
Cited by (166)
Combining probabilistic forecasts of intermittent demand
2024, European Journal of Operational ResearchExchange rates and fundamentals: Forecasting with long maturity forward rates
2024, Journal of International Money and FinanceA hybrid approach of wavelet transform, ARIMA and LSTM model for the share price index futures forecasting
2024, North American Journal of Economics and FinanceForecast combinations: An over 50-year review
2023, International Journal of ForecastingEconomic policy uncertainty and the ECOWAS exchange rate mechanism
2023, Scientific AfricanModel averaging for asymptotically optimal combined forecasts
2023, Journal of Econometrics