Universitat Pompeu Fabra , LBS and CEPR This draft

This paper compares the forecasting performance of some leading models of inflation for the cross section of G-7 countries. We show that bivariate and trivariate models suggested by economic theory or statistical analysis are hardly better than univariate models. Phillips curve specifications fit well into this class. Significant improvements in both the MSE of the forecasts and turning point prediction are obtained with time varying coefficients models which exploit international interdependencies. The performance of the latter class of models is independent of the sample, while it is not the case for standard specifications.