Multivariate Out-of-Sample Tests for Granger Causality.pdfVIP

Multivariate Out-of-Sample Tests for Granger Causality Sarah Gelper and Christophe Croux K.U.Leuven, Faculty of Economics and Applied Economics, Naamsestraat 69, 3000 Leuven, Belgium Abstract A time series is said to Granger cause another series if it has incremental predictive power when forecasting it. While Granger causality tests have been studied exten- sively in the univariate setting, much less is known for the multivariate case. In this paper we propose multivariate out-of-sample tests for Granger causality. The perfor- mance of the out-of-sample tests is measured by a simulation study and graphically represented by Size-Power plots. It emerges that the multivariate regression test is the most powerful among the considered possibilities. As a real data application, we investigate whether the consumer confidence index Granger causes retail sales in Germany, France, the Netherlands and Belgium. Key words: Consumer Sentiment, Granger Causality, Multivariate Time Series, Out-of-sample Tests 1 Email: {sarah.gelper,christophe.croux}@econ.kuleuven.be, Fax: +32(0)16/326732 This research has been funded by the ’Fonds voor Wetenschappelijk Onderzoek’ (contract number G.0594.05). Preprint submitted to Elsevier Science 23 December 2005 1 Introduction Suppose we have a multivariate time series yt of length T containing k compo- nents and would like to investigate whether another time series x , consisting t of l components, Granger causes y . The series x is said to Granger cause y t t t if the past of xt has additional power in forecasting yt after controlling f