Stata 15 help for xtdpd

[XT] xtdpd -- Linear dynamic panel-data estimation

Syntax

xtdpd depvar [indepvars] [if] [in] , dgmmiv(varlist [...]) [options]

options Description ------------------------------------------------------------------------- Model * dgmmiv(varlist[...]) GMM-type instruments for the difference equation; can be specified more than once lgmmiv(varlist[...]) GMM-type instruments for the level equation; can be specified more than once iv(varlist[...]) standard instruments for the difference and level equations; can be specified more than once div(varlist[...]) standard instruments for the difference equation only; can be specified more than once liv(varlist) standard instruments for the level equation only; can be specified more than once noconstant suppress constant term twostep compute the two-step estimator instead of the one-step estimator hascons check for collinearity only among levels of independent variables; by default checks occur among levels and differences fodeviation use forward-orthogonal deviations instead of first differences

SE/Robust vce(vcetype) vcetype may be gmm or robust

Reporting level(#) set confidence level; default is level(95) artests(#) use # as maximum order for AR tests; default is artests(2) display_options control spacing and line width

coeflegend display legend instead of statistics ------------------------------------------------------------------------- * dgmmiv() is required. A panel variable and a time variable must be specified; use xtset; see [XT] xtset. depvar, indepvars, and all varlists may contain time-series operators; see tsvarlist. by, statsby, and xi are allowed; see prefix. coeflegend does not appear in the dialog box. See [XT] xtdpd postestimation for features available after estimation.

Menu

Statistics > Longitudinal/panel data > Dynamic panel data (DPD) > Linear DPD estimation

Description

xtdpd fits a linear dynamic panel-data model where the unobserved panel-level effects are correlated with the lags of the dependent variable. The command can fit Arellano-Bond and Arellano-Bover/Blundell-Bond models like those fit by xtabond and xtdpdsys. However, it also allows models with low-order moving-average correlation in the idiosyncratic errors or predetermined variables with a more complicated structure than allowed for xtabond or xtdpdsys.

Options

+-------+ ----+ Model +------------------------------------------------------------

dgmmiv(varlist [, lagrange(flag [llag])]) specifies GMM-type instruments for the difference equation. Levels of the variables are used to form GMM-type instruments for the difference equation. All possible lags are used, unless lagrange(flag llag) restricts the lags to begin with flag and end with llag. You may specify as many sets of GMM-type instruments for the difference equation as you need within the standard Stata limits on matrix size. Each set may have its own flag and llag. dgmmiv() is required.

lgmmiv(varlist [, lag(#)]) specifies GMM-type instruments for the level equation. Differences of the variables are used to form GMM-type instruments for the level equation. The first lag of the differences is used unless lag(#) is specified, indicating that #th lag of the differences be used. You may specify as many sets of GMM-type instruments for the level equation as you need within the standard Stata limits on matrix size. Each set may have its own lag.

iv(varlist [, nodifference]) specifies standard instruments for both the difference and level equations. Differences of the variables are used as instruments for the difference equation, unless nodifference is specified, which requests that levels be used. Levels of the variables are used as instruments for the level equation. You may specify as many sets of standard instruments for both the difference and level equations as you need within the standard Stata limits on matrix size.

div(varlist [, nodifference]) specifies additional standard instruments for the difference equation. Specified variables may not be included in iv() or in liv(). Differences of the variables are used, unless nodifference is specified, which requests that levels of the variables be used as instruments for the difference equation. You may specify as many additional sets of standard instruments for the difference equation as you need within the standard Stata limits on matrix size.

liv(varlist) specifies additional standard instruments for the level equation. Specified variables may not be included in iv() or in div(). Levels of the variables are used as instruments for the level equation. You may specify as many additional sets of standard instruments for the level equation as you need within the standard Stata limits on matrix size.

noconstant; see [R] estimation options.

twostep specifies that the two-step estimator be calculated.

hascons specifies that xtdpd check for collinearity only among levels of independent variables; by default checks occur among levels and differences.

fodeviation specifies that forward-orthogonal deviations be used instead of first differences. fodeviation is not allowed when there are gaps in the data or when lgmmiv() is specified.

+-----------+ ----+ SE/Robust +--------------------------------------------------------

vce(vcetype) specifies the type of standard error reported, which includes types that are derived from asymptotic theory and that are robust to some kinds of misspecification; see Methods and formulas in [XT] xtdpd.

vce(gmm), the default, uses the conventionally derived variance estimator for generalized method of moments estimation.

vce(robust) uses the robust estimator. For the one-step estimator, this is the Arellano-Bond robust VCE estimator. For the two-step estimator, this is the Windmeijer (2005) WC-robust estimator.

+-----------+ ----+ Reporting +--------------------------------------------------------

level(#); see [R] estimation options.

artests(#) specifies the maximum order of the autocorrelation test to be calculated. The tests are reported by estat abond; see [XT] xtdpd postestimation. Specifying the order of the highest test at estimation time is more efficient than specifying it to estat abond, because estat abond must refit the model to obtain the test statistics. The maximum order must be less than or equal to the number of periods in the longest panel. The default is artests(2).

display_options: vsquish and nolstretch; see [R] estimation options.

The following option is available with xtdpd but is not shown in the dialog box:

coeflegend; see [R] estimation options.

Examples

Setup . webuse abdata

Arellano-Bond estimator with two lags of dependent variable included as regressors and strictly exogenous covariates . xtdpd l(0/2).n l(0/1).(w ys) k, dgmmiv(n) div(l(0/1).(w ys) k) . xtdpd l(0/2).n l(0/1).(w ys) k year yr1980-yr1984, dgmmiv(n) div(l(0/1).(w ys) k year) div(yr1980-yr1984) nocons hascons

Arellano-Bond estimator with two lags of dependent variable included as regressors, strictly exogenous covariates and robust VCE . xtdpd l(0/2).n l(0/1).(w ys) k year yr1980-yr1984, dgmmiv(n) div(l(0/1).(w ys) k year) div(yr1980-yr1984) nocons hascons twostep vce(robust)

Arellano-Bover/Blundell-Bond system estimator with two lags of dependent variable included as regressors and strictly exogenous covariates . xtdpd l(0/2).n l(0/1).(w ys) k, dgmmiv(n) lgmmiv(n) div(l(0/1).(w ys) k )

Arellano-Bond estimator with two lags of dependent variable included as regressors, endogenous covariates and a robust VCE . xtdpd L(0/1).(n w k) year yr1979-yr1984, dgmmiv(n w k) div(year yr1979-yr1984) nocons hascons vce(robust)

Arellano-Bover/Blundell-Bond system estimator with two lags of dependent variable included as regressors, endogenous covariates and a robust VCE . xtdpd L(0/1).(n w k) year yr1979-yr1984, dgmmiv(n w k) lgmmiv(n w k) div(year yr1979-yr1984) nocons hascons vce(robust)

Stored results

xtdpd stores the following in e():

Scalars e(N) number of observations e(N_g) number of groups e(df_m) model degrees of freedom e(g_min) smallest group size e(g_avg) average group size e(g_max) largest group size e(t_min) minimum time in sample e(t_max) maximum time in sample e(chi2) chi-squared e(arm#) test for autocorrelation of order # e(artests) number of AR tests computed e(sig2) estimate of sigma_epsilon^2 e(rss) sum of squared differenced residuals e(sargan) Sargan test statistic e(rank) rank of e(V) e(zrank) rank of instrument matrix

Macros e(cmd) xtdpd e(cmdline) command as typed e(depvar) name of dependent variable e(twostep) twostep, if specified e(ivar) variable denoting groups e(tvar) variable denoting time within groups e(vce) vcetype specified in vce() e(vcetype) title used to label Std. Err. e(system) system, if system estimator e(hascons) hascons, if specified e(transform) specified transform e(diffvars) already differenced variables e(datasignature) checksum from datasignature e(properties) b V e(estat_cmd) program used to implement estat e(predict) program used to implement predict e(marginsok) predictions allowed by margins

Matrices e(b) coefficient vector e(V) variance-covariance matrix of the estimators

Functions e(sample) marks estimation sample

Reference

Windmeijer, F. 2005. A finite sample correction for the variance of linear efficient two-step GMM estimators. Journal of Econometrics 126: 25-51.


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index