2.5 Fama-MacBeth Regressions
Why Quantitative Finance Stack Exchange recommends it
Quant Stack Exchange answerers point people to it as part of Cochrane's video course when explaining how time-series and cross-sectional risk premium estimation work.
Recommended by 1 people across 1 different months, 2019 to 2023.
What it is
A short lecture segment from John Cochrane's asset pricing course at UChicago Booth, covering Fama-MacBeth regressions as a variant of the cross-sectional approach to estimating risk premia. It is module 2.5 in Part II on classic linear models, following segments on time-series regressions, GRS and cross-sectional regressions.
What people said
Strong rec Kevin · 2022-08 · Quantitative Finance Stack Exchange · on “Cross Sectional vs. Time-Series Risk Premia Estimate”
No one is better at explaining asset pricing than John Cochrane [LINK] . He explains it in detail in his brilliant textbook [LINK] . The following videos from his video course are pure gold 2.2 Time-Series and GRS [LINK] 2.3 Cross-Sectional Regressions [LINK] 2.4 Comments [LINK] 2.5 Fama-MacBeth Regressions [this video] The notes for Week 5b Notes on empirical methods of his course Business 35150 Advanced Investments [LINK] are also fantastic. A few notes though Time series approaches only apply to traded assets. You can't use time series regression to find the price of consumption risk (because consumption growth isn't a traded return). Fama-MacBeth is a variant of the cross-sectional approach (
Cites skoestlmeier · 2019-03 · Quantitative Finance Stack Exchange · on “Question about Fama Macbeth Regression (Confusion about paper)”
Fama-MacBeth procedure (Step 1): So if my understanding is clear, first we would use the cross sectional regression to estimate 4*5,000=20,000 betas? That is not right, because betas (and other risk-factor loadings) are estimated by a time-series regression: $$ R_{i,t}^e = \alpha_i + \beta_{i, MktRf} MktRf_t + \beta_{i, SMB} SMB_t + \beta_{i, HML} HML_t + \epsilon_{i, t}$$ where $R_{i,t}^e$ is the excess return of stock $i$ (i.e. in excess of a risk-free rate of return) and the right-hand variables are the well known risk-factor premiums of the Fama/French 3-factor model. For your example with 5,000 stocks, you would run the above regression a minimum of 5,000 times if you consider full-samp
Cites Richard Hardy · 2023-01 · Quantitative Finance Stack Exchange · on “Clustered vs. GMM-based standard errors: which ones to use in asset pricing?”
Consider estimating an asset pricing model such as the CAPM or a multifactor model using monthly data. Petersen (2009) [LINK] section "Asset pricing application" suggests use of standard errors clustered by time, as this addresses the problem of cross-sectional correlation. (Meanwhile, serial correlation is not much of an issue.) Petersen (2009) [LINK] discusses how time-clustered standard errors fare against alternative approaches such as Fama-MacBeth (works fine) and several other. However, he does not seem to consider GMM estimation that is discussed extensively in Cochrane "Asset Pricing" [LINK] (2005) Part II, e.g. Chapter 11. On the other hand, Cochrane does mention clustered standard