Using Regression to Separate Stock, Sector, and Market Returns

Summary

The document asks how to divide an asset’s growth into a broad-market component and an asset-specific component. Its example compares a stock’s return with the market’s return, while also raising a similar product-versus-market question in retail. The sole proposed approach is to regress the stock’s returns on market returns and, if appropriate, sector returns. The regression coefficients can describe the stock’s estimated exposure to those factors, while residual variation represents movement not explained by the chosen predictors.

The suggestion is explicitly preliminary and does not establish a standard attribution method or guarantee an intuitive split such as subtracting the market return from the stock return. Results depend on the return measure, sample period, benchmark and sector definitions, and model specification. A regression describes statistical co-movement, not necessarily causal contributions to growth. The exchange provides no empirical results, diagnostics, or guidance for translating estimated exposures into an additive percentage attribution, so further methodological choices are required before using it for performance analysis.

Key ideas

• A regression of stock returns on market returns can estimate broad-market exposure.
• Adding sector returns may help account for industry-level co-movement.
• The choice and definition of benchmarks affect the resulting decomposition.
• Regression coefficients describe statistical relationships and do not prove causal contributions.
• The proposed approach is crude and comes without empirical validation or attribution details.

Tags

# Technical analysis - Market vs Stock/Product Growth Decomposition


# Technical analysis - Market vs Stock/Product Growth Decomposition












I am looking for an authoritative source for the standard methodology for decomposing growth between market and individual units.

For example, in retail sales, decomposing growth between market level and product level (e.g. if market grew 5% and product grew 7%, then 5% due to market and 2% due to product).

In the stock market, this would be analogous to the growth due to market vs due to individual stocks (e.g. stock price grew by 7% but market was up 5%, then 2% due to individual stock).

Some weak references like this are all I've been able to find, but seems it should be a common problem in things like technical analysis (I probably just don't know how to refer to it).

## Answer by user42108 (score 1)

https://quant.stackexchange.com/a/60114

Assuming I understood your question: a crude way to do this would be to regress the returns of the stock on the market and the sector (however you define those, which is not always obvious). No guarantee you get sensible results.

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.