Abstract:
Diffusion indexes based on dynamic factors have recently been advocated by Stock and Watson (1998), and further used to perform forecasting tests by the same authors on US data. This technique is explored for the euro area using a multi-country data set and a broad array of variables, in order to test the inflation forecasting performance of extracted factors at the aggregate euro area level. First, a description of factors extracted from different data sets is performed using a number of different approaches. Conclusions reached are that nominal phenomena in the original variables might be well captured in-sample using the factor approach. Out-of-sample tests have more ambiguous interpretation, as factors seem to be good leading indicators of inflation, but the comparative advantage of the factors is less clear. Nevertheless, alternative indicators such as unemployment or money growth do not outperform them
