[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: AW: F-statistic using Newey

From   "Martin Weiss" <>
To   <>
Subject   st: AW: F-statistic using Newey
Date   Tue, 19 May 2009 09:12:21 +0200


The F-stats coincide for -lag(0)- for -newey- and the sandwich VCV for
-regress-. As soon as you tell -newey- to use more lags for the standard
errors, it diverges from the robust -regress- standard errors. Given this
divergence, why do you find it surprising that the F-tests diverge as well?
After all, they jointly test for zero slopes of the covariates, so the
standard error has to have some bearing on that...

webuse idle2, clear
tsset time

reg usr idle, vce(robust)
newey usr idle, lag(0)

reg usr idle, vce(robust)
newey usr idle, lag(1)


-----Ursprüngliche Nachricht-----
[] Im Auftrag von Hernan Urcola
Gesendet: Dienstag, 19. Mai 2009 03:10
Betreff: st: F-statistic using Newey

I am running a series of regressions using the regress and newey
procedures. However, the F-statistic is different when I run regress y
x than when I run newey y x. I thought the newey procedure only
adjusted the VCV matrix leaving the residuals (used to compute the
F-stat) unchanged, but apparently that is not case.
Does anybody know the exact formulas that Stata uses to compute the
F-statistic under the newey procedure?

Thanks much,

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index