Abstract:
This article attempts to examine whether the equity premium in the United States can be
predicted from a comprehensive set of 18 economic and
financial predictors over a monthly
out-of-sample period of 2000:2–2011:12, using an in-sample period of 1990:2–2000:1. To
do so, we consider, in addition to the set of variables used in Rapach and Zhou (2013), the
forecasting ability of four other important variables: the US economic policy uncertainty,
the equity market uncertainty, the University of Michigan’s index of consumer sentiment,
and the Kansas City Fed’s
financial stress index. Using a more recent dataset compared to
that of Rapach and Zhou (2013), our results from predictive regressions show that the
newly added variables do not play any significant statistical role in explaining the equity
premium relative to the historical average benchmark over the out-of-sample horizon,
even though they are believed to possess valuable informative content about the state of
the economy and
financial markets. Interestingly, however, barring the economic policy
uncertainty index, the three other indexes considered in this study yield economically
significant out-of-sample gains, especially during recessions, when compared to the
historical benchmark.