{smcl}

{* *! version 0.3 12apr2017}{...}

{vieweralsosee "tabstat" "help tabstat"}{...}

{viewerjumpto "Syntax" "regife##syntax"}{...}

{viewerjumpto "Description" "regife##description"}{...}

{viewerjumpto "Options" "regife##options"}{...}

{viewerjumpto "Examples" "regife##examples"}{...}

{viewerjumpto "Stored results" "regife##results"}{...}

{viewerjumpto "References" "regife##references"}{...}

{viewerjumpto "Author" "regife##contact"}{...}

{title:Title}

{p2colset 5 18 20 2}{...}

{p2col :{cmd:reghdfe} {hline 2}}Linear models with interactive fixed effects{p_end}

{p2colreset}{...}

{marker syntax}{...}

{title:Syntax}

{p 8 15 2} {cmd:regife}

{depvar} [{indepvars}]

{ifin} {it:{weight}}

{cmd:,}

{opth ife(idvar timevar, ndmis)}

[{help regife##options:options}] {p_end}

{marker description}{...}

{title:Description}

{pstd}

{cmd:regife} fits a model with interactive fixed effects following Bai (2009). If you want to fit a model with interacted fixed effects, use {help reghdfe}. Optionally, it saves the estimated interactive fixed effects. Errors are computed following the regressions indicated in Section 6, but Monte Carlo evidence suggest bootstraps performs better in finite sample. The program requires {help reghdfe} and {help hdfe} to be installed (both are available on SSC).

{marker options}{...}

{title:Options}

{synoptset 30 tabbed}{...}

{synopthdr}

{synoptline}

{synopt :{opt ife(idvar timevar, ndims)}} specifies the id variable, time variable, and the dimension of the factor model. To save the estimated interactive fixed effects, write

{it:ife(ife_idvar = idvar ife_timevar = timevar, ndims)}. {p_end}

{synopt :{opt a:bsorb}{cmd:(}{help reghdfe##absvar:absvar}[...]{cmd:)}} identifiers of the fixed effects that will be absorbed. To save the estimated fixed effects, write {it:absorb(fe_absvar = absvar)}.{p_end}

{synopt:{opt vce}{cmd:(}{help reghdfe##vcetype:vcetype}[, {it:opt}]{cmd:)}}{it:vcetype}}

is {opt un:adjusted}/{opt ols} (default), {opt r:obust}, {opt bootrap} or {opt cl:uster} {it:clustervars}. Monte carlo evidence suggests that bootstrap performs better in finite sample{p_end}

{synopt:{opt tol:erance(#)}} specifies the tolerance criterion for convergence; default is {cmd:tolerance(1e-9)}{p_end}

{synopt:{opt max:iterations(#)}} specifies the maximum number of iterations; default is {cmd:maxiterations(5000)}. 0 corresponds to an illimited number of iterations{p_end}

{synopt :{opt res:iduals(newvar)}} save residuals {p_end}

{synopt :{opt bstart(matrix)}} start the iteration algorithm at a given value for b{p_end}

{synoptline}

{p2colreset}{...}

{p 4 6 2}{cmd:fweight}s, {cmd:aweight}s and {cmd:pweight}s are allowed but should be constant within idvar; see {help weight}.{p_end}

{marker examples}{...}

{title:Examples}

{pstd}Setup{p_end}

{phang2}{cmd:. webuse nlswork}{p_end}

{phang2}{cmd:. keep if id <= 100}{p_end}

{pstd}Factor model in id, year of dimension 1{p_end}

{phang2}{cmd:. regife ln_w tenure, ife(id year, 1)}{p_end}

{pstd}Model including id fixed effect, and a factor model in id, year of dimension 2{p_end}

{phang2}{cmd:. regife ln_w tenure, a(id) ife(id year, 1)}{p_end}

{pstd}Model including id fixed effect, year fixed effect, and a factor model in id, year of dimension 1{p_end}

{phang2}{cmd:. regife ln_w tenure, a(id year) ife(id year, 1)}{p_end}

{pstd}Save interactive fixed effects{p_end}

{phang2}{cmd:. regife ln_w tenure, ife(ife_id = id ife_year = year, 1)}{p_end}

{pstd}Save fixed effects and interactive fixed effects{p_end}

{phang2}{cmd:. regife ln_w tenure, a(fe_id = id fe_year = year) ife(ife_id = id ife_year = year, 1)}{p_end}

{pstd}Generate residuals{p_end}

{phang2}{cmd:. regife ln_w tenure, ife(id year, 1) residuals(newvar)}{p_end}

{pstd}Bootstrap standard errros{p_end}

{phang2}{cmd:. regife ln_w tenure, ife(id year, 1) vce(bootstrap)}{p_end}

{pstd}Block bootstrap with respect to id{p_end}

{phang2}{cmd:. regife ln_w tenure, ife(id year, 1) vce(bootstrap, cluster(id))}{p_end}

{marker results}{...}

{title:Stored results}

{pstd}

{cmd:regife} stores the following in {cmd:e()}:

{synoptset 24 tabbed}{...}

{p2col 5 24 28 2: Scalars}{p_end}

{synopt:{cmd:e(N)}}number of observations{p_end}

{synopt:{cmd:e(df_r)}}residual degrees of freedom{p_end}

{synopt:{cmd:e(df_m)}}model degrees of freedom{p_end}

{synopt:{cmd:e(F)}}residual sum of squares{p_end}

{synopt:{cmd:e(mss)}}total sum of squares{p_end}

{synopt:{cmd:e(rss)}}residual sum of squares{p_end}

{synopt:{cmd:e(iterations)}}number of iterations{p_end}

{synopt:{cmd:e(error)}}convergence error{p_end}

{synoptset 24 tabbed}{...}

{p2col 5 24 28 2: Macros}{p_end}

{synopt:{cmd:e(cmd)}}{cmd:regife}{p_end}

{synopt:{cmd:e(depvar)}}name of dependent variable{p_end}

{synopt:{cmd:e(indepvars)}}names of independent variables{p_end}

{synopt:{cmd:e(id)}}id variable {p_end}

{synopt:{cmd:e(time)}}time variable {p_end}

{synopt:{cmd:e(dimension)}}dimension{p_end}

{synopt:{cmd:e(congerged)}}did the algorithm converge?{p_end}

{marker references}{...}

{title:References}

{pstd}

{cmd:regife} implements the estimate proposed by:

{phang}

Jushan Bai. "Panel Data Models with Interactive Fixed Effects".

{it:Econometrica, 77.4 (2009): 1229-1279.}

{p_end}

{marker contact}{...}

{title:Author}

{phang}

Matthieu Gomez

{phang}

Department of Economics, Princeton University

{phang}

Please report issues on Github

{browse "https://github.com/matthieugomez/stata-regife":https://github.com/matthieugomez/stata-regife}

{p_end}