Interpreting Near-Zero Alpha in a Fama–French Three-Factor Regression

Summary

The document asks whether a near-zero intercept is plausible when regressing monthly excess returns for a dataset of stocks on the market, size, and value factors. The reported example has positive market and size loadings, a small negative value loading, and a high adjusted R-squared. The response explains that zero alpha can occur, particularly for mutual funds or ETFs, and is not inherently evidence that the regression is wrong.

The answer cautions that individual stocks may be less likely to show zero alpha and says the time period and example securities would be needed to replicate the calculation. It does not explain statistical significance, how the dataset was aggregated, or what the factor coefficients imply beyond the intercept question. Therefore, a near-zero estimate should be interpreted in the context of the return sample and regression setup rather than as a universal expectation.

Key ideas

• A near-zero intercept can be a valid result in a Fama–French three-factor regression.
• The answer notes that zero alpha is especially plausible for mutual funds and ETFs.
• Individual stocks may be less likely to have zero alpha, according to the response.
• The sample period and specific securities are needed to reproduce and assess the regression.

Tags

# Fama-French 3Factor Model alpha


# Fama-French 3Factor Model alpha












I have different large datasets consisting of 1000 stocks each. I want run a FF3 regression

I regress my monthly returns (minus riskfree rate) of the dataset against the Mkt-RF, SMB and HML factor.

But as alpha I always receive a value which is close to zero (0.001)

For Beta1 (MKT-RF) I receive 1.08 Beta2 (SMB) 0.268 Beta3 (HML) -0.069

Adjusted R^2 is 0.985

Can that be correct and if yes, how can i interpret this? I am very insecure because I always receive an alpha of 0

Thanks for your help

## Answer by phdstudent (score 1)

https://quant.stackexchange.com/a/75926

It is perfectly possible to get zero alpha (specially if you are looking at returns of mutual funds/ETFs). With individual stocks you are likely not to get zero alpha. If you edit your question to mention the time-frame and give an example of a stock or two, I can quickly replicate your regressions.

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.