LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous (more recent) messageNext (less recent) messagePrevious (more recent) in topicNext (less recent) in topicPrevious (more recent) by same authorNext (less recent) by same authorPrevious page (February 1996, week 5)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Thu, 29 Feb 1996 22:14:34 GMT
Reply-To:     Alex Pearson <pearson@SAMIAM.COLORADO.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Alex Pearson <pearson@SAMIAM.COLORADO.EDU>
Organization: University of Colorado at Boulder
Subject:      Simultaneous regressions


Can anyone out here in SASland help one of our professors with a simultaneous regression problem? I'm afraid I'm not familiar enough with SYSLIN and Logistic to be of much help. Here is the problem:

Thanks for your help with this problem. I am writing a paper with Ed Nelling of Georgia Tech, and we need to run a three equation simultaneous equation regression model where the first equation is either a Poission regression or a multinomial PROBIT regression (we could do either) and the next two regression equations are OLS.

SYSLIN in SAS seems to only allow us to do three OLS equations simultaneously. Due to the count nature of the dependant variable in one of the equations, we really need to have one of the three equations be a regression that assumes that the dependant variable is of a count nature, such as the Poisson regressions or multinomial Probit regressions.

Although the perfect solution is a SAS proceedure that will accept one Poisson regression and then two OLS regressions in a simultaneous framework, we could use multinomial Probit (or Logit) instead of Poisson. We would also be willing to try another programming language instead of SAS, like SPSS or Shazam!, although I would prefer SAS if possible. We would also be willing to try a three equation simultaneous regression where all three of the regressions are Poisson or multinomial probit, although this is less beneficial.

Thanks for your help in this matter!

***************************************************************************** Michael A. Goldstein tel: (303) 492-1242 Asst. Professor of Finance fax: (303) 492-5962 Univ. of Colorado at Boulder email: Campus Box #419 Boulder, CO 80309-0419 *****************************************************************************

Alex Pearson Alex.Pearson@Colorado.EDU (303) 492-3879 Statistical Computing and Consulting Office University of Colorado

Back to: Top of message | Previous page | Main SAS-L page