Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Re: two-way fixed effects

From   "Austin Nichols" <>
Subject   Re: st: Re: two-way fixed effects
Date   Wed, 16 May 2007 12:47:48 -0400

The data on which Abowd and others work has many persons per firm and
many firms per person over time; see

Abowd J., Creecy R., Kramarz F., Computing person and firm effects
using linked longitudinal employer-employee data, March 2002.

so I do think it is a type of 2WFE model, though -a2reg- does estimate
models with nested effects as well ( la -xtmixed-). In theory, the
-a2reg- program could be handy for those who want to estimate panel
regressions with 3 flavors of FE, e.g. county, firm, and year, by
specifying county and firm IDs as FE to difference out, and dummies
for the year effect, but the output is weird, compared to, say,
-areg-, and the error-handling is subpar:

webuse grunfeld, clear
gen ind=company<4 if time<4
replace ind=company<6 if time>3
qui tab time, gen(dyr)
a2reg k inv dyr*, individual(com) unit(ind)
a2reg k inv mval dyr*, individual(com) unit(ind)

On 5/16/07, Kit Baum <> wrote:
The use of 'two-way fixed effects' to describe this model is
certainly not standard in the econometrics literature. 2WFE is
usually understood to mean an effect for each individual and an
effect for each time period in panel (xt) data. As I understand it
Ouraad's approach involves nested effects, where an individual is a
member of a unit, and there is an effect on each level: more like
xtmixed or gllamm, which would treat these nested effects as random.
I suppose there is ample reason to consider a nested FE model with
two-level nesting, but I would not call it 2WFE.

On May 16, 2007, at 2:33 AM, Scott said:
> A program to estimate two-way fixed effects is available from:
> - - net from
> stata/ -
> See also:
> and
> M. Andrews, T. Schank, and R. Upward.  2006. Practical fixed-effects
> estimation methods for the three-way error components model.  Stata
> Journal
> 6(4) 461-481.
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index