Multivariate rank-based forecast combining techniques

: We analyze macroeconomic data using univariate and multivariate forecast combining techniques. We simulate forecast errors with different variance-covariance structures. The simulations are used to compare the performance of univariate and multivariate combining techniques. Keywords: Variance-covariance structure, simulation, combination of forecasts, multivariate combination of forecasts. Acknowledgement: Supported by Deutsche Forschungsgemeinschaft, Graduiertenkolleg "Angewandte Statistik". AMS 1991 Subject Classification: 62G30 1. Introduction Previous research showed that if we combine several forecasts of the same event we get a combined forecast that is better than the individual forecasts or a simple average, e.g. Newbold and Granger (1974), Russel and Adam (1987), Klapper (1998a), and Klapper (1998b) to name a few. Usually several variables are forecasted at the same time e.g. economic indicators, indices of stock markets or election results of several parties. In the past,...