Predicting returns with financial ratios: evidence from Greece

This article investigates whether financial ratios (dividend yield, price earnings ratio, book to market value) can predict aggregate stock returns. We report a forecasting competition between single and multiple OLS, GARCH and ECM-GARCH regressions of the Greek return series. First, we test the out-of-sample forecasting accuracy, and then we compare the forecasting techniques based on the symmetric error statistics under both static and dynamic methods. The results show that ECM-GARCH(1,1) model has significant coefficients. Both static and dynamic forecasts confirm that ECM-GARCH(1,1) is the most appropriate model for forecasting returns during the period January 2003 - December 2004.