2
$\begingroup$

When calculating a panel data regression with multiple fixed effects using the felm() (of the lfe package), no constant / intercept is generated in the summary results.

Why would there be no constant generated?

When using Stata (xtreg, fe), for example, an intercept is generated by default. My understanding is that the "constant" is not a real constant, but rather a grand mean of all constants. Is that correct?

(I apologize in advance for not providing a reproducible example...)

Here is an excerpt from Simen Gaure's article in The R Journal (Vol. 5/2, December 2013) https://journal.r-project.org/archive/2013/RJ-2013-031/RJ-2013-031.pdf

"The careful reader has noticed that the behaviour of summary() on a ’felm’ object with respect to degrees of freedom and R2 is the same as that of on an ’lm’ object when including an intercept. There is no explicit intercept in the result of felm(), but the factor structure includes one implicitly."

Thank you!

$\endgroup$
  • $\begingroup$ Welcome to CV. If your question is about how to do something in R, it is probably off topic here. If you are asking about the statistics behind this, then it is on topic. If that is so, then please edit your question to reflect this. $\endgroup$ – Peter Flom Mar 19 '18 at 11:09
0
$\begingroup$

Any regression package will either (1) include an intercept and drop one dummy to avoid collinearity or (2) not include an intercept but not drop any dummies. In case (1) the interpretation of any dummy coefficient is "in respect to the average of all others" and in specification (2) it is in respect to the one omitted where the intercept is shared for all.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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