Inference for time-varying alpha of U.S. equity market and its implications

Abstract

This study aims to examine significance and time-variation of alpha in the asset pricing models by adopting kernel smoothing regression and simultaneous confidence bands. The five-factor model of Fama and French (2015) and two other models are considered, and monthly returns of stocks and active mutual funds in the United States, grouped into six benchmark portfolios and seven mutual fund portfolios, are used for empirical analysis. After confirming significance and time-variation of alpha, determinants of time-varying alphas are explored using standard fundamentals such as yield spreads, inflation rate, and industrial production index as explanatory variables. The results show that in all model specifications alphas vary substantially over time, implying that none of the models is successful in explaining the returns of benchmark portfolios and mutual fund portfolios. They also show that term spread and default spread are significant determinants of time-varying alphas. In line with previous studies, the yield spreads turn out to have a role in explaining asset prices. It implies that business cycle fluctuations and irrational market expectations may cause the anomalies.