PROC QLIM Statement :: SAS/ETS(R) 14.1 User's Guide

Data Set Options

DATA=SAS-data-set: specifies the input SAS data set. If the DATA= option is not specified, PROC QLIM uses the most recently created SAS data set.

Output Data Set Options

OUTEST=SAS-data-set: writes the parameter estimates to an output data set.
COVOUT: writes the covariance matrix for the parameter estimates to the OUTEST= data set. This option is valid only if the OUTEST= option is specified.
CORROUT: writes the correlation matrix for the parameter estimates to the OUTEST= data set. This option is valid only if the OUTEST= option is specified.

Printing Options

NOPRINT: suppresses the normal printed output but does not suppress error listings. If NOPRINT option is set, then any other print option is turned off.
PRINTALL: turns on all the printing-control options. The options set by PRINTALL are COVB and CORRB.
CORRB: prints the correlation matrix of the parameter estimates.
COVB: prints the covariance matrix of the parameter estimates.
ITPRINT: prints the initial parameter estimates, convergence criteria, and all constraints of the optimization. At each iteration, objective function value, step size, maximum gradient, and slope of search direction are printed as well.

Model Estimation Options

COVEST=covariance-option

specifies the method to calculate the covariance matrix of parameter estimates. The supported covariance types are as follows:

OP: specifies the covariance from the outer product matrix.
HESSIAN: specifies the covariance from the inverse Hessian matrix.
QML: specifies the covariance from the outer product and Hessian matrices (the quasi-maximum likelihood estimates).

The default is COVEST=HESSIAN.

HECKIT <(heckit-options)>

requests that the selection model be estimated by Heckman’s two-step estimation method. You must specify exactly two MODEL statements when you use the HECKIT option. One of the models must be a binary probit model; therefore, you must specify the DISCRETE option in the MODEL or ENDOGENOUS statement. You base the selection on the binary probit model for the second model; therefore, you must specify the SELECT option for this model.

You can specify one or both of the following heckit-options:

SECONDSTAGE=OLS | ML

specifies the estimation method of the second stage of Heckman’s two-step method. SECONDSTAGE=OLS requests the ordinary least squares method for the second stage. If you specify SECONDSTAGE=OLS, then the model of interest—that is, the model that uses the SELECT option—must be linear and contain a continuous dependent variable. Therefore, you cannot specify the DISCRETE, CENSORED, or TRUNCATED option along with the SELECT option for the model of interest. When you specify the SECONDSTAGE=OLS option, you cannot test or restrict the parameters of the model of interest. However, you can test or restrict the parameters of the selection model—that is, the model that defines the selection rule.

If you specify SECONDSTAGE=ML, then PROC QLIM uses the maximum likelihood method in the second stage, as it does in the first stage. When you specify SECONDSTAGE=ML, the model of interest can be nonlinear. Moreover, you can also use the TEST or RESTRICT statement to test or restrict the parameters of the model of interest.

By default, SECONDSTAGE=OLS.

UNCORRECTED

requests the conventional OLS standard errors when the second-stage estimation method is the ordinary least squares method. If you do not specify the UNCORRECTED option, PROC QLIM reports the corrected OLS standard errors. For more information about the corrected standard errors, see the section Heckman’s Two-Step Selection Method.

If you specify both the UNCORRECTED and SECONDSTAGE=ML options, PROC QLIM ignores the UNCORRECTED option, because the UNCORRECTED option is related to the OLS standard errors.

NDRAW=value

specifies the number of draws for Monte Carlo integration.

SEED=value

specifies a seed for pseudo-random number generation in Monte Carlo integration.

Optimization Process Control Options

PROC QLIM uses the nonlinear optimization (NLO) subsystem to perform nonlinear optimization tasks. You can use any of the NLO options in the NLOPTIONS statement. For more information, see Chapter 7: Nonlinear Optimization Methods.

METHOD=value

specifies the optimization method. If this option is specified, it overwrites the TECH= option in NLOPTIONS statement. Valid values are as follows:

CONGRA: performs a conjugate-gradient optimization
DBLDOG: performs a version of double-dogleg optimization
NEWRAP: performs a Newton-Raphson optimization combining a line-search algorithm with ridging
NMSIMP: performs a Nelder-Mead simplex optimization
NONE: specifies that no optimization be performed beyond using the ordinary least squares method to compute the parameter estimates
NRRIDG: performs a Newton-Raphson optimization with ridging
QUANEW: performs a quasi-Newton optimization
TRUREG: performs a trust region optimization

The default method is METHOD=QUANEW.

Plotting Options

PLOTS<(global-plot-options)> = plot-request | (plot-requests): controls the display of plots. By default, the plots are displayed in panels unless the UNPACK global-plot-option is specified. When you specify only one plot-request, you can omit the parentheses around the plot-request.

Global Plot Options

You can specify the following global-plot-options:

ONLY: displays only the requested plot.
PRIOR: displays the prior predictive graph that is associated with the requested posterior predictive plot BAYESPRED. This option is available only for Bayesian analysis.
UNPACKPANEL UNPACK: specifies that all paneled plots be unpacked, meaning that each plot in a panel is displayed separately.

Plot Requests

You can specify the following plot-requests:

ALL: specifies all types of available plots.
AUTOCORR<(LAGS=n)>: displays the autocorrelation function plots for the parameters. This plot-request is available only for Bayesian analysis. The optional LAGS= suboption the number (up to lag n) of autocorrelations to be plotted in the AUTOCORR plot. If this suboption is not specified, autocorrelations are plotted up to lag 50.
BAYESDIAG: displays the TRACE, AUTOCORR, and DENSITY plots. This plot-request is available only for Bayesian analysis.
BAYESPRED: displays the predictive analysis. The predictive analysis takes into account the variability of the error term, whereas the PREDICTED plot-request does not. The BAYESPRED plot-request is available only for Bayesian analysis.
BAYESSUM: displays the posterior distribution, the prior distribution, and the maximum likelihood estimates. This plot-request is available only for Bayesian analysis.
CONDITIONAL: displays the conditional expected values for continuous endogenous variables. Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
DENSITY<(FRINGE)>: displays the kernel density plots for the parameters. This plot-request is available only for Bayesian analysis. If you specify the FRINGE suboption, a fringe plot is created on the X axis of the kernel density plot. This plot-request is available only for Bayesian analysis.
ERRSTD: displays the error standard deviation versus observed regressors when you also specify a HETERO statement. This plot-request is not available for Bayesian analysis.
EXPECTED: displays the expected values for continuous endogenous variables. Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
MARGINAL: displays the marginal effects. Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
MILLS: displays the inverse Mills ratio. Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
NONE: suppresses all diagnostic plots.
PREDICTED: displays the model predicted values. Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
PROB: displays the predicted response probability. Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
PROBALL: displays the predicted probabilities for each level of the response. Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
PROFLIK: displays the profiled log likelihood. Each profiled graph is obtained by setting all the parameters to their maximum likelihood estimate except for the profiling parameter. The profiling parameter takes values on a predefined grid that is determined by the maximum likelihood estimate of the corresponding standard deviation. When a restricted optimization is requested, the profiled log likelihood plots depict the behavior of the profiled log likelihood around the restricted MLE without imposing the actual restrictions.
RESIDUAL: displays the residuals versus observed regressors. This plot-request is not available for Bayesian analysis.
TE1: displays the technical efficiency for the stochastic frontier model as suggested by Battese and Coelli (1988). Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
TE2: displays the technical efficiency for the stochastic frontier model as suggested by Jondrow et al. (1982). Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This plot-request is not available for Bayesian analysis.
TRACE<(SMOOTH)>: displays the trace plots for the parameters. This plot-request is available only for Bayesian analysis. The SMOOTH suboption displays a fitted penalized B-spline curve for each TRACE plot.
XBETA: displays the structural part on the right-hand side of the model. Each contributing regressor is set equal to its mean, except for the parameter that is reported on the X axis. This is not available for Bayesian analysis.

The QLIM Procedure