10
$\begingroup$

Fama-MacBeth (1973) introduce a two stage cross-sectional regression method (http://en.wikipedia.org/wiki/Fama%E2%80%93MacBeth_regression).

  1. If I was to regress stock prices (or returns) on a conditioning set using the Fama-MacBeth regression method, what are the econometric assumptions behind this model?

  2. Is the model still used today or is another model now preferred?

$\endgroup$

2 Answers 2

5
$\begingroup$

2) Alternative to Fama-MacBeth is Fama-French approach. Explanation of difference see, for example, here: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1271935 Fama-French approach was used by Carhart (introduced momentum), Pastor-Stambaugh (introduced liquidity), Fama-French themselves (used it to build 5-factor model), and many other (elsevier or google for "fama french factor model").

Fama-MacBeth approach was used in Chen, Roll, Ross, 1986, which is believed to be quite important paper for APT.

There is also PCA approach to modelling asset returns. See, e.g., Luedecke, 1984, Connor and Korajczyk, 1988.

In industry, still, scholar models are not used. Practitioners develop their own models, sometimes kept as closely held secret, sometimes - made available to public (BARRA, CSFB, Morgan Stanley, Salomon-Smith-Barney, Bloomberg). Sometimes these models are just overcomplicated versions of scholar models, built with the same approaches. Sometimes they try to use some combination, for example Bloomberg family of models combine PCA with macro- and fundamental approaches.

$\endgroup$
5
  • $\begingroup$ Thanks for your answer - will definetly look into your suggested alternatives. Do you have any comments to 1)? $\endgroup$
    – Sunv
    Jan 27, 2014 at 10:29
  • $\begingroup$ Frankly, I don't get the question in 1). What is meant by "assumptions"? $\endgroup$ Jan 27, 2014 at 10:38
  • $\begingroup$ Well, what I mean is what is the assumption on the distribution of the error term for instance (in both stages). $\endgroup$
    – Sunv
    Jan 27, 2014 at 10:41
  • 2
    $\begingroup$ The assumptions on which the procedure is based are the same of the simply linear model. You can find that also here: en.wikipedia.org/wiki/Linear_regression#Assumptions. In the case you need for some more deepened source about this topic, I suggest you to study that on some introductory econometrics books. $\endgroup$
    – Quantopik
    Jan 27, 2014 at 17:36
  • 1
    $\begingroup$ I wanted to double-check before answering, but yes, currently I agree with user21040 - Fama-MacBeth is usual story of regression. $\endgroup$ Jan 27, 2014 at 19:23
4
$\begingroup$

The key assumption is that there is no time-series correlation between the error terms. Fama-MacBeth can deal with cross-sectional correlations.

See Samuel Thompson's "Simple formulas for standard errors that cluster by both firm and time" in the Journal of Financial Economics (2011) for a treatment of different regression methods for testing equity pricing models.

$\endgroup$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.