Enhancements in SAS/ETS® 12.1 Software


SAS/ETS 12.1 introduces many new estimation features, including Bayesian options, new variable selection methods, a new data access engine, and many enhancements to existing procedures. SAS/ETS has also developed p-values based on accurate high-performance simulation methods for many popular test statistics in the AUTOREG and PANEL procedures.

Bayesian Estimation in PROC QLIM

The QLIM procedure now provides users with Bayesian estimation methods for most univariate models supported by the procedure. PROC QLIM enables users to estimate models of limited dependent variables such as:

These models can now be estimated using Bayesian methods. The main features include:

Predictive Analysis of Ticket Sales

Heteroscedasticity- and Autocorrelation-Consistent (HAC) Covariance Estimators

The PANEL and AUTOREG procedures now offer convenient options to produce variance-covariance matrices for estimates in the presence of heteroscedasticity or autocorrelation of unknown form. The HAC options support the following five types of kernel functions:

The bandwidth parameter can be assigned or estimated with the Andrews method, Newey-West method, and sample size method. The HAC options also contain a prewhitening routine. A convenient HAC option, COVEST=NEWEYWEST, enables quick specification of the Newey-West standard errors presented in Stock and Watson (2002).

Variable Selection Algorithms

The COUNTREG and QLIM procedures now support the greedy search method of variable selection. In each step, Akaikeís information criterion (AIC) or the Schwartz Bayesian criterion (SBC) is evaluated, and evaluation continues until the selection criterion is met. The greedy search algorithm can be implemented with either forward or backward selection.

SEATS Method in PROC X12

The X12 procedure, the popular procedure for deseasonalizing data based on the U.S. Census X12 program, has been updated to include the SEATS decomposition method. SEATS is a polynomial-based seasonal decomposition method that extracts trend, seasonal, irregular, and cyclical factors from ARIMA models. With the SEATSDECOMP statement, a user can save the signal-adjusted time series in a SAS data set.

SASEXFSD Interface Engine

SAS/ETS 12.1 provides a new data access engine for FactSet data that are provided by the FactSet OnDemand service (formerly known as FASTFetch). This service provides access to a number of data libraries from economic and financial data sources, such as Compustat, Dun & Bradstreet, and FactSet.

New Specification Tests

A variety of new model specification tests have been added to the PANEL and AUTOREG procedures. These tests are used to check the statistical assumptions of the models concerning stationarity, cointegration, and structural change and include p-values generated by high-performance simulation methods. Many software packages report only selected critical values for these tests. The AUTOREG procedure now supports the following new test statistics:

For panel data models the PANEL procedure supports these new test statistics:

New ODS Graphics

Many new ODS graphics have been included with the SAS/ETS 12.1 release, including:

dependency details

Other Enhancements to SAS/ETS

The QLIM procedure now supports the Heckman two-step estimator for sample selection models. This is an alternative to the maximum likelihood estimates that the procedure already contains. The consistent standard errors for this two-step estimator are automatically generated for the second-stage OLS parameter estimates.

The SEVERITY procedure has been modified so that estimation algorithms execute in parallel, thereby taking advantage of all CPU cores available during estimation. The SEVERITY procedure has also been enhanced to provide standard errors and confidence intervals for empirical distribution function (EDF) estimates. In addition, PROC SEVERITY now supports regressors through its SCALEMODEL statement, enabling explanatory variables to scale the fitted distribution.

The MODEL procedure has a number of enhancements to improve simulation. The OPTIMIZE option has been added to the SOLVE statement to enable the simulation of models that include constraints on the solve variables in PROC MODELís systems of equations. Several new options also place bounds and restrictions on these solve variables. The new REPORTMISSINGS option helps to debug problems associated with missing data in FIT and SOLVE statements. The ANALYZEDEP option produces information about the nature of misspecification errors in simulations. New copula options have also been added that can be used with PROC MODEL's simulation features.

The COUNTREG procedure now supports multiple MODEL statements.

Download a pdf version.