Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Problems with exogenous variables in Heckman model

Subject   st: Problems with exogenous variables in Heckman model
Date   Fri, 18 Oct 2013 17:52:51 -0500

Dear Statalist,
I'm running a Heckman selection model to calculate price elasticities
of credit cards but I'm having problems to obtain them. I'm including
three exogenous variables in my selection equation and all of them are
categorical variables. When I run -margins- to obtain elasticities,
the coefficients value, standard errors, z-score and p-values are the
same for all the categories of my variables:

 margin compart, eyex(PC_INT_REVOLVENTE)atmean noatlegend

Conditional marginal effects                      Number of obs   =      46738
Model VCE    : Conventional

Expression   : Linear prediction, predict()
ey/ex w.r.t. : PC_INT_REVOLVENTE

                  |            Delta-method
                  |      ey/ex   Std. Err.      z    P>|z|     [95%
Conf. Interval]
          compart |
               0  |  -.7632775   .3828873    -1.99   0.046
-1.513723   -.0128323
               1  |  -.7632775   .3828873    -1.99   0.046
-1.513723   -.0128323

Does anyone know if there is a restriction of using only
non-categorial variables as exogenous in the selection model (the
classical examples of the selection models includes married status and
having children or not as exogenous variables, which are similar in
nature to the ones we are using)? In Stata Manual -margins- is listed
as a valid postestimation option after -heckman- but it does not
mention any restriction or additional option regarding categorical

Thank you in advance
*   For searches and help try:

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