Chapter 33 Evaluating the predictive accuracy of models

The methods for evaluating the predictive accuracy of econometric models are discussed in this chapter. The use of these methods should allow to decide upon the model that best approximates the true structure of the economy and how much confidence to place on the predictions from a given model. The numerical solution of nonlinear models is reviewed in the chapter, including stochastic simulation procedures. The standard methods that have been used to evaluate ex ante and ex post predictive accuracy are discussed in the chapter. The main problems with these methods are that they (1) do not account for exogenous variable uncertainty, (2) do not account for the fact that forecast-error variances vary across time, and (3) do not treat the possible existence of misspecification in a systematic way. The comparison of the predictive accuracy of alternative models is not a straightforward exercise. The difficulty of evaluating alternative models is undoubtedly one of the main reasons that there is currently so little agreement about the model that best approximates the true structure of the economy.