a b s t r a c t
This paper proposes forecast combination based on the method of Mallows Model Averaging (MMA). The
method selects forecast weights by minimizing a Mallows criterion. This criterion is an asymptotically
unbiased estimate of both the in-sample mean-squared error (MSE) and the out-of-sample one-step-
ahead mean-squared forecast error (MSFE). Furthermore, the MMA weights are asymptotically mean-
square optimal in the absence of time-series dependence. We show how to compute MMA weights in
forecasting settings, and investigate the performance of the method in simple but illustrative simulation
environments. We find that the MMA forecasts have low MSFE and have much lower maximum regret
than other feasible forecasting methods, including equal weighting, BIC selection, weighted BIC, AIC
selection, weighted AIC, BatesGranger combination, predictive least squares, and GrangerRamanathan
combination.