Industry Fixed Effects in Cross-Sectional Asset-Pricing Regressions

Summary

The document explains industry fixed effects in cross-sectional regressions used in asset pricing. The method adds industry-specific constant terms, often implemented as indicator variables, to the regression. These terms capture average return differences associated with the industry groups represented in the data.

With industry effects included, coefficients on other explanatory variables, such as market or value factors, describe return sensitivities after accounting for the average return associated with a stock’s industry. The answer also notes that results depend on the industry classification applied. It provides an intuitive explanation but no regression example, estimation details, or evidence about how industry effects change particular factor estimates; those depend on the dataset and model specification.

Key ideas

• Industry fixed effects add an industry-specific intercept to a regression.
• These intercepts account for average return differences across industry groups.
• Other factor coefficients describe sensitivities conditional on the industry-level averages.
• The industry classification chosen determines how stocks are grouped.
• The explanation gives intuition but no empirical example or model-specific results.

Tags

# What are industry fixed effects?


# What are industry fixed effects?












I have come across the term "industry fixed effects" in some papers in relation to cross sectional regressions in asset pricing. I know what "fixed" regression models are, but not what is meant by "industry fixed"... Can someone explain to me in a clear way what exactly this is?

Thanks for the answers.

## Answer by fes (score 2, accepted)

https://quant.stackexchange.com/a/57122

You apply the fixed effects regression (https://en.wikipedia.org/wiki/Fixed_effects_model) model with industry specific individual effects. Essentially add industry specific constant terms (/dummies) to a regression model. In the context of a cross-sectional asset pricing regression this intuitively captures the mean return of all stocks in the same industry. The other betas will then represent sensitivities of a stock return to factors such as market return or value factor controlling for the mean return of all stocks in the relevant industry. There are different industry classifications of stocks, you can find them online.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.