help ueve-------------------------------------------------------------------------------

Title

[R] ueve-- Unbiased errors-in-variables estimator (UEVE), Errors-in-variables estimator (EVE) and Efficient Wald estimator (EWALD) regressions on grouped data

Syntax

uevedepvar[indepvars] [if] [in] [weight],group(groupvar)[estimatorlevel(#)]

estimatordescription -------------------------------------------------------------------------ueveUnbiased errors-in-variables estimator (UEVE), the defaulteveErrors-in-variables estimator (EVE)ewaldEfficient Wald estimator (EWALD) -------------------------------------------------------------------------

xiis allowed; see prefix.pweights,iweights,aweights andfweights are allowed; see weight.

Description

uevefits a linear regression ofdepvaronindepvars, using one of the three estimators for grouped data: Devereux (2007) errors-in-variables estimator that is approximately unbiased (UEVE); Deaton (1985) errors-in-variables estimator (EVE) that was shown to be equivalent to Jackknife Instrumental Variable Estimator in Devereux (2007); and Efficient Wald estimator (EWALD) (Angrist 1991). Input data should be in individual-level (not grouped) format, as the program uses them to compute estimates of variance of sampling errors necessary to correct the bias in the grouping estimator. Since the variance of the sampling errors is estimated using group-level sampling variances with (groupsize-1) in denominator, groups cannot have less than two observations. If any group has one observation in it, the program automatically drops such groups from estimation, and a warning message is displayed. Variance-covariance matrix of the estimator is calculated using formula following Deaton (1985).

Options+-------+ ----+ Model +------------------------------------------------------------

group(groupvar) specifies the name of the grouping variable. The data must be divided into a set of mutually exclusive and exhaustive groups indexed bygroupvar.+-----------+ ----+ Reporting +--------------------------------------------------------

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

ExamplesSetup

. sysuse nlsw88. egen grp=group(south smsa c_city industry)Fit UEVE regression in which observations are grouped by groups

grp, which are allocated into mutually exclusive and exhaustive supersets indexed by categorical variablessouth, smsa, c_city, industry.. ueve wage race married grade tenure, group(grp)Fit EWALD regression.

. ueve wage race married grade tenure, group(grp) ewald

Saved results

uevesaves the following ine():Scalars

e(N)number of observationse(G)number of groupse(r2)R-squared (defined using data grouped into means)Macros

e(cmd)uevee(depvar)name of dependent variablee(indepvars)indepvarse(properties)b VMatrices

e(b)coefficient vectore(V)variance-covariance matrix of the estimatorsFunctions

e(sample)marks estimation sample

Angrist, Joshua D., 1991. "Grouped-data estimation and testing in simple labor-supply models," Journal of Econometrics, Elsevier, vol. 47(2-3), pages 243-266, February.

Deaton, A. 1985. "Panel data from a time series of cross-sections," Journal of Econometrics, 30, 109- 126.

Devereux, Paul J., 2007. "Improved Errors-in-Variables Estimators for Grouped Data," Journal of Business & Economic Statistics, American Statistical Association, vol. 25, pages 278-287, July.

AuthorAliaksandr Amialchuk, University of Toledo, USA aamialc@utnet.utoledo.edu

Also seeOnline:

[R] eivreg;[R] ivregress;[R] regress