Predictability in Equity Markets: A Quantile Approach

In this thesis we examine the predictability of equity premiums using international lagged excess returns. Rapach et al. (2013) shows that lagged excess returns of different countries, especially the U.S., have significant forecasting power when predicting equity premiums. This paper extends on the results of Rapach et al. (2013) by showing, with the use of quantile regression, that the relationships found by Rapach et al. (2013) also hold in other parts of the distribution besides the mean, especially in the lower tails. Furthermore we present that the U.S. continues to be one of the countries with the most predictive power regardless of the specific part of the distribution. However the out-of-sample predictive power, measured with adjusted versions of the Diebold-Mariano statistic, is less visible for the tails of the distribution than it is for the mean.