Choosing Daily or Monthly Data for Fama-French Regressions

Summary

The document considers why estimated Fama-French three-factor coefficients and their statistical significance may differ when portfolio returns are sampled daily or monthly. It does not recommend one frequency universally. Instead, it frames the choice around the purpose of the estimate and the amount of data available, noting that daily observations can support a shorter estimation window while monthly observations are often collected over a longer one.

The cited empirical asset pricing guidance gives an example of requiring a substantial sample of daily observations over a year and contrasts it with using monthly excess returns across several years. The answer treats the problem as an ordinary regression estimation issue: sample size and estimation-window choices affect coefficient precision and inference. The document offers a rule of thumb rather than a comparison of frequencies or a formal adjustment for differing return horizons, dependence, or market microstructure effects. For a task requiring daily data, the user should still ensure an adequate sample and interpret estimates in light of the chosen window.

Key ideas

• Fama-French coefficient estimates can vary with the sampling frequency of portfolio returns.
• The appropriate frequency depends on the estimation goal and available sample length.
• Daily estimation generally calls for enough observations within the selected window.
• Monthly excess returns are commonly estimated over a longer historical period.
• The guidance is a regression sample-size rule of thumb, not a universal frequency prescription.

Tags

# Should Fama-French coefficients be calculated with daily or monthly returns?


# Should Fama-French coefficients be calculated with daily or monthly returns?












I noticed when I regress the return of a portfolio on the Fama French 3 factor model that the value and the statistical significance of the coefficients vary when I use daily versus monthly portfolio returns. I would like to know what is the best frequency to estimate these coefficients?

For my purpose I do need to use higher frequency data (daily)

## Answer by shoonya (score 2)

https://quant.stackexchange.com/a/60106

In this context, I refer to the book on Empirical Asset Pricing by Bali, Engle and Murray (2016). They state on page 124 that

> A minimum number of data points are usually required to ensure the quality of the values estimated by the regression. In the case of daily data over a one-year period, a reasonable requirement may be that the regression be fit using at least 200 data points. While using one year’s worth of daily data to calculate beta is common, other estimation period lengths and data frequencies are also used. Another common approach is to use monthly excess return data from the past five years.

Its basically a regression and usual regression thumb rules are applicable.

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.