PREDICTIVE EVALUATION OF ECONOMETRIC FORECASTING MODELS IN COMMODITY FUTURES MARKETS

The predictive accuracy of various econometric models, including rando m walks, vector-autoregressive and vector-error-correction models, are investigated using daily futures prices of four commodities (the S&P 500 index, treasury bonds, gold, and crude oil). All models are estima ted using a rolling-window approach, and evaluated by both in-sample a nd out-of-sample performance measures. The criteria considered include system criteria, where we evaluate multiequation forecasting models, and univariate forecast-accuracy criteria. The five univariate criteri a are root mean square error (RMSE), mean absolute deviation (MAD), me an absolute percentage error (MAPE), confusion matrix (CM), and confus ion rate (CR). The five system criteria used include the trace of seco nd-moment matrix of the forecast-errors matrix (TMSE), the trace of se cond-moment matrix of percentage-forecast errors (TMAPE), the generali zed forecast-error second-moment matrix (GFESM), and a trading-rule pr ofit criterion (TPC) based on a maximum-spread trading strategy. An in -sample criterion, the mean Schwarz information criteria (MSIC), is al so computed. Our results suggest that error-correction models perform better in shorter forecast horizons, when models are compared based on quadratic loss measures and confusion matrices. However, the error-co rrection models which we consider perform better at all forecast horiz ons (one to five steps ahead) when models are compared based on a prof it-maximization loss function. Further, our error-correction model, wh ere the error-correction term is constructed according to a cost-of-ca rry equilibrium condition, outperforms our alternative error-correctio n model, which uses the price spreads as the error-correction term.