Detecting Hidden Size and Liquidity Bias in Equity Factors
Summary
The article explains why a factor’s strong historical returns may come from an unintended exposure to market capitalization or liquidity. It uses large-order sell value as an example: because the measure is an absolute amount, it tends to be larger for bigger companies, and a negative signal can therefore behave like a small-cap strategy. The article argues that this hidden size effect may explain why buy and sell order-value signals perform similarly during a period when small-cap stocks did well.
It also describes testing modified versions of a trading-flow factor. Separating buy and sell activity and adding periodicity and price-volume features raised reported returns, but increased correlation with market capitalization. The author then favors a composite weighted toward less size-sensitive components, reporting lower size and liquidity correlations alongside improved returns. These figures are examples from the author's tests, not evidence of durable performance. The proposed explanation linking institutional trading patterns to volume periodicity is acknowledged as tentative. The main lesson is to measure style exposures during factor research and distinguish a factor's intended signal from incidental exposures.
Key ideas
- Absolute order-value measures can proxy for company size because larger firms tend to generate larger transaction amounts.
- A factor's returns may reflect a style exposure, such as small-cap performance, rather than the market behavior its name suggests.
- Modifying a factor can unintentionally increase its correlation with size or liquidity.
- Composite factors can be designed to emphasize components with lower unwanted style exposure.
- Style correlations and historical returns should be assessed together, while treating proposed mechanisms and backtest results cautiously.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.