Testing Predicitive Ability of Business Cycle Indicators for the Euro Area

We analyze the predictive power of seven leading indicators for economic activity in the Euro Area developed by different banks, institutions and research centers. Our comparison is conducted in a bivariate vector autoregressive framework. Indicators are compared by means of an in-sample and an out-of-sample forecasting experiment. Predictive accuracy is compared by recently proposed tests for superior predictive ability. Our results suggest that nearly all indicators have good in-sample properties and that a majority of them is able to outperform a naive univariate autoregressive model out-of-sample. Additionally, we find that indicators perform better in boom periods than in recessions. The OECD and FAZ indicators are both composite indicators and deliver the most accurate forecasts.