What Fama–MacBeth Regressions Say About Return Prediction

Summary

The document asks how to use estimated betas and risk premia from a Fama–MacBeth regression to predict returns. Its answer draws a distinction between testing an asset pricing model and building a forecasting model: Fama–MacBeth regression is presented as a way to assess whether the model's pricing relationships hold unconditionally, rather than as a direct prediction procedure. The response points readers to a standard asset pricing reference for further explanation.

The exchange offers no equations, empirical results, or detailed account of either regression step, so it does not show how to construct a separate forecasting model or evaluate predictions. Its useful lesson is limited to clarifying the stated purpose of the method: estimated betas and lambdas should not automatically be treated as a return forecast. The answer is brief, and a reader seeking implementation details or a fuller discussion of conditional versus unconditional tests will need additional material.

Key ideas

• Fama–MacBeth regression is described as a test of asset pricing relationships rather than a direct forecasting model.
• The exchange does not explain how to turn estimated betas or lambdas into a separate return forecast.
• The response points to an asset pricing textbook for a fuller treatment.

Tags

# For a Fama-Macbeth regression , How does one predict the returns based on the model?


# For a Fama-Macbeth regression , How does one predict the returns based on the model?












Fama-Macbeth does a two-step regression i.e a time-series and cross-sectional regression and we estimate betas and lambdas, so how does one predict based on these parameters, which one to choose?

## Answer by phdstudent (score 4)

https://quant.stackexchange.com/a/21827

Those are not predictive models. Fama-MacBeth just checks whether the model is valid unconditionally. Chapter 12 from Cochrane book is a good reference.

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.