------------------------------------------------------------------------------- help for

reghintfe() -------------------------------------------------------------------------------Titlereghintfe - Estimates a linear regression model with one interacted high dimens > ional fixed effect.

Syntax

reghintfedepvar[indepvar] [if] [in] ,id1(varname)intvar(varname)[options]

DescriptionThis command estimates a linear regression model with one high-dimensional interacted fixed effect. The command makes use of the Frisch-Waugh-Lovell result to avoid computing the dummy variables and the interactions for the fixed effects. Because of this it can be used with a very large number of groups for the fixed effect.

Options

id1(varname)fixed effect variable. This option is required.

intvar(varname)variables that interact with the fixed effect. This option is required.

cluster(varname)computes clustered standard errors.

verbosegives more information during estimation.

ExamplesExample 1: Estimates a model with interaction between a fixed effect i and the variable x3 > . Does the same as "regress y x1 x2 i.j##c.x3"

. regintfe y x1 x2, id1(j) intvar(x3) It creates two variables, __fe0 and __fe1. The variable __fe0 contains the esti > mates of the coefficients for the fixed effect and the variable __fe1 the estimates of the coefficients for the i > nteraction between "j" and "x3".

Example 2: Estimates a model with interaction between a fixed effect i and the variables x > 2 and x3. Standard errors are clustered on "j". Does the same as "regress y x1 i.j##c.x2 i.j##c.x3, cluster(j)"

. regintfe y x1, id1(j) intvar(x2 x3) cluster(j) Creates three variables, __fe0, __fe1 and __fe2. The variable __fe0 contains t > he estimates of the coefficients for the fixed effect, the variable __fe1 the estimates of the coefficients for the inte > raction between "j" and "x2" and the variable __fe2 the estimates of the coefficients for the interaction betwee > n "j" and "x3".

AuthorPaulo Guimaraes, Universidade do Porto, Portugal.

Email: pguimaraes@fep.up.pt

Comments welcome!

ReferenceIf you use this program in your research cite:

Paulo Guimaraes and Pedro Portugal. "A Simple Feasible Alternative Procedure to > Estimate Models with High-Dimensional Fixed Effects", Stata Journal, 10(4), 628-649, 2010.

{p_end}