selmlog procedure - problems with reproducing the same results

[Aleksandra Anic]

Dear all,

I use user-written command selmlog in order to estimate wage equation when selection is modelled by mlogit (reference Bourguignon F., Fournier M. and Gurgand M., Selection Bias Corrections Based on the Multinomial Logit Model: Monte-Carlo comparisons, mimeo Delta, 2004 (download from http\\:www.pse.ens.fr\senior\gurgand\selmlog13.htm)).

I do not get the same estimated coefficients every time when I consequently run my do fail. I tried with set seed option before selmlog command, but again the results slightly differ.

What should I do in order to get same results every time I run my do fail?

Any comments/suggestions are wellcome.

Thank you
Best wishes,
Aleksandra

[William Lisowski]

The link you give does not open in my browser. The following link from Professor Gurgand's web page on the PSE web site (English language version)

http://www.parisschoolofeconomics.co...selmlog13.html

does open, and it concludes with the following.

We will be glad to receive any notice of the problems you may encounter (write to [email protected]).