Additional Evidence on Information Variables and Equity Premium Predictability

This paper presents the most comprehensive out-of-U.S.-sample examination of information variables and equity premium predictability by focusing on Canada to reassess the growing U.S.-based evidence casting doubt on predictability. Using monthly data for 36 variables from 1950 to 2013, we test their individual predictive ability and provide a new empirical assessment of the related econometric issues. We find conclusive and robust evidence of in-sample, out-of-sample and economically meaningful predictability of the Canadian equity premium, providing guidance on each variable as a market indicator. Our results nevertheless raise questions on some variables that have been successful U.S. predictors.