This paper examines return predictability of the U.S. stock market using portfolios sorted by size, book-to-market ratio, and industry. A novel panel variance ratio test is proposed and employed to evaluate time-varying return predictability from 1964 to 2011. It is found that the stock returns...

We evaluate the empirical validity of popular asset-pricing models in explicit consideration of statistical power, by employing the adaptive significance level and equal-probability test. Past studies often use samples from a large cross-section of portfolios over a long time period, conducting...

We conduct the extreme bounds analysis (EBA) to evaluate the robustness or fragility of a range of stock market anomalies, using U.S. daily data from 1960. The EBA is a large-scale sensitivity analysis, able to isolate the effects of potential data-mining or p-hacking under model uncertainty....

We study return predictability of the Dow Jones Industrial Average indices from 1900 to 2009. We find strong evidence that time-varying return predictability is driven by changing market conditions, consistent with the implications of the adaptive markets hypothesis. During market crashes, no...