The case against the size premium
This Research Affiliates research note argues that the size premium does not exist. They acknowledge that long-term US data suggests a size premium of 3.4%, while global data suggests a size premium of 1% since the effect was first documented in the early 1980s.
However, they then argue that (1) there is an upward bias in size premium estimates due to inaccurate returns on delisted stocks in major databases, (2) indices and hypothetical portfolios ignore trading costs, (3) the statistical significance of the size premium estimates is likely overstated due to data-mining and reporting bias, (4) the statistical significance is very weak, and (5) there is no outperformance on a risk-adjusted basis.
The authors emphasize that their results do not imply that investors should stop investing in small-cap stocks, only that investors should not expect these stocks as a group to deliver superior returns.
Our researchers publish many whitepapers based on their own empirical studies; they also follow quantitative research done by others.
