Intraday Return Skewness as a Stock Selection Factor
Summary
The report studies whether intraday return distributions contain stock selection information beyond conventional daily factors. Using one-minute stock prices, it constructs factors from intraday return variance, skewness, and kurtosis, then examines single-factor performance and adds the factors to a multi-factor model. It compares different sampling frequencies and calculation methods.
Skewness is the main useful signal in the reported tests. Its rank information coefficients are generally around 0.05–0.06, with monthly success rates near 80%; after orthogonalizing against industry, size, turnover, idiosyncratic risk, and reversal effects, it remains predictive, though weaker. Fama–MacBeth regressions also indicate incremental selection ability, stronger for one-minute than five-minute skewness.
Portfolio improvements are limited: adding the one-minute factor improves several reported risk and return measures, while some five-minute versions do not improve annualized return. The evidence covers Chinese stocks from 2010 through early 2017, so it is historical and may not generalize to other periods or markets.
Key ideas
- Intraday return distribution statistics can add information beyond daily stock selection factors.
- Skewness shows more consistent selection ability than the other examined distribution characteristics.
- Orthogonalizing skewness against common stock characteristics reduces its signal but does not eliminate it.
- One-minute skewness performs more strongly than five-minute skewness in the reported regressions.
- Portfolio improvements from adding the factor are modest and depend on the sampling frequency.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.