I am running ZINB (zero-inflated negative binomial) models, and after which
running MFX to calculate marginal effects.
Although several coefficient estimates are NEGATIVE, Stata reports POSITVE
marginal effects.
My question: Is this possible and why? Is this the case since the marginal
effect is for the overall model (which includes the zero-inflation equation)
and therefore if both coefficents from the two equations are negative, then
the resulting marginal effect if positive (i.e., it's a product of the two)?
One point on my models: The do not include instruments; both equations
contains the same variables, thus nonlinearities are identfying the
zero-inflation.
Thanks for your help.
*****************************
M. Christopher Roebuck, MBA
Health Economist
AdvancePCS
chris.roebuck@advancepcs.com
410-229-8382 office
410-785-8140 fax
*****************************
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/