Fama MacBeth regression and dummy variable

[Lisa Arthur]

Hello everyone,

I am using the Fama MacBeth Regression approach in my analysis.
What I am trying to do, is to conduct the Fama MacBeth regression using xtfmb for two subsamples.
In the help file, there is the note that xtfmb can be combined with "by", however I get the following error message "Date is not regularly spaced".
Using newey with an indicator variable and the option "force" gives me the Newey-West t-statistics however does not conduct the 2 step procedure that is captured by xtfmb.

Is there any way to either use a dummy variable with xtfmb without receiving the error message above? And how can I compare coefficient estimates across subsamples using xtfmb?


I appreciate any help! Thanks!

[Phil Bromiley]

You'll increase your chances of a useful answer by following the faq on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Also, assume we're not in your area. Fama MacBeth is strictly a finance and accounting approach.

xtfmb is user written so you cannot count on the routine operating precisely like standard Stata routines. However, without the extra information and sample data, it is hard to guess what is going on. There appears to be a problem with your date variable - is it annual? Are you using lags? etc.

You can often trick a routine into doing essentially subsample estimates by interacting a dummy for the subperiods with the x variables. This is easiest with the factor variable notation, but I don't know if xtfmb handles that.