{smcl}
{* 18april2005}{...}
{hline}
help for {hi:xtfevd}{right:authors:  {hi:Troeger, Vera E. and Pluemper, Thomas}    }
{hline}

{title:Panel Fixed Effects Regression with Vector Decomposition}

{p 8 15 2}{cmd:xtfevd} {it:depvar} [{it:varlist}] [{it:weight}] [{cmd:if}
{it:exp}] [{cmd:in} {it:range}] {cmd:,}
{cmdab:invariant:(}{it:varlist_invariant}{cmd:)} 
[
{cmdab:nocon:stant}
{cmdab:sk:ew} {cmdab:corr} {cmdab:ar1} {cmdab:robust} {cmdab:pcse} 
{cmdab:cluster:(}{it:varname}{cmd:)} {cmdab:s2iv_endog:(}{it:varlist_s2endog}{cmd:)}
{cmdab:s2iv_exog:(}{it:varlist_s2exog}{cmd:)}]


{p 4 4 2}
where {cmdab:invariant}  is mandatory and {it:varlist_invariant} contains all explanatory variables 
that are purely time-invariant, almost time-invariant or cross-sectionally dominated and the researcher 
want to include them into the second stage of {cmd:xtfevd}.


{p 4 4 2}
{cmd:xtfevd} is for use with panel data.  You must {help tsset} your data
before using {cmd:xtfevd}; see help {help tsset}.


{p 4 4 2}
{it:depvar} and {it:varlist} may contain time-series operators; see help
{help varlist}.


{p 4 4 2}
{cmd:xtfevd} shares the features of all estimation commands; see help
{help estcom}.


{p 4 4 2}
The syntax of {help predict} following {cmd:xtfevd} is

{p 8 16 2}{cmd:predict} [{it:type}] {it:newvarname} [{cmd:if} {it:exp}]
[{cmd:in} {it:range}] [{cmd:,} [ {cmd:xb} | {cmd:stdp} ] ]

{p 4 4 2}
These statistics are available both in and out of sample; type
"{cmd:predict} {it:...} {cmd:if e(sample)} {it:...}" if wanted only for the
estimation sample.


{title:Description}

{p 4 4 2}
{cmd:xtfevd} estimates a three stage panel fixed effects vector decompostion model 
that allows for the inclusion of time-invariant variables and efficiently estimates 
almost time-invariant explanatory variables within a panel fixed effects framework. 
the first stage estimates a pure fixed effects model to obtain an estimate of the unit effects, 
the second decomposes the fixed effects vector into a part explained by the time-invariant and 
almost time-invariant variables and an unexplainable part - the error term of the second stage -, 
and the third stage re-estimates the original model by pooled OLS, including the time-invariant 
variables and the error term of the second stage, eta. 
This third step assures to control for collinearity between time-varying and invariant right hand side 
variables, and adjusts the degrees of freedom.



{title:Options}

{p 4 8 2}{cmd:invariant(}{it:varlist_invariant}{cmd:)} specifies all right hand side variables
that are time-invariant or quasi time-invariant and have to be included into the second stage
of the model.

{p 4 8 2}{cmd:noconstant} suppresses the intercept term in all three stages.

{p 4 8 2}{cmd:ar1} controls for serial correlation of the error term, ar1 Cocrane-Orcutt transformation
on third stage.

{p 4 8 2}{cmd:robust} specifies a Huber-White Sandwich estimator to control for cross-sectional
heteroskedasticity.

{p 4 8 2}{cmd:pcse} specifies panel corrected standard errors in stage 3.


{p 4 8 2}{cmdab:cluster:(}{it:varname}{cmd:)} robust cluster option.

{p 4 8 2}{cmdab:s2iv_endog:(}{it:varlist_s2endog}{cmd:)} specifies an instrumental regression on stage 2, the z-variables 
that should be instrumented in stage two must be specified in varlist_s2endog. this option has to be used 
together with {cmdab:s2iv_exog:(}{it:varlist_s2exog}{cmd:)}.


{p 4 8 2}{cmdab:s2iv_exog:(}{it:varlist_s2exog}{cmd:)} specifies the list of variables uncorrelated with the unit spedific
effects that serve as instruments for the z-variables that are correlated with the unit effects in stage 2.

{p 4 8 2}{cmd:skew} provides skewness, kurtosis and normality tests for the unit fixed effects (u) 
and the error term (eta) of the vector decomposition in stage 2. 
Note: the stata command {cmd: normtest} has to be implemented beforehand.

{p 4 8 2}{cmd:corr} provides a correlation matrix for all right hand side variables with the
unit fixed effects.


{title:Options for {help predict}}

{p 4 8 2}{cmd:xb}, the default, calculates the linear prediction.

{p 4 8 2}{cmd:stdp} calculates the standard error of the linear prediction.


{title:Examples}

{p 4 8 2}{cmd:. tsset c_code year, yearly}{p_end}
{p 4 8 2}{cmd:. xtfevd govcon gdppc federalism pres referenda turnout open fdi infl unemp old left, invariant(federalism pres referenda left)}

{p 4 8 2}{cmd:. xtfevd govcon gdppc federalism pres referenda turnout open fdi infl unemp old left, invariant(federalism pres referenda left) nocon}

{p 4 8 2}{cmd:. xtfevd govcon gdppc federalism pres referenda turnout open fdi infl unemp old left, invariant(federalism pres referenda left) ar1}


{title:Also see}

{p 4 13 2}
Manual:  {hi:[U] 23 Estimation and post-estimation commands},{break}
{hi:[U] 29 Overview of Stata estimation commands},{break}
{hi:[XT] xtreg}

{p 4 13 2}
Online:  help for {help estcom}, {help postest}, {help tsset},
{help xt}, {help xtgls}; {help newey}, {help prais}, {help regress},
{help svy}, {help xtgee}, {help xtpoisson}, {help xtprobit}, {help xtreg},
{help xtregar}