Bivariate ordered probit using Roodman's cmp command

[Ernie Zampelli]

I have estimated a bivariate ordered probit model in the following way:

cmp (y1=x1 x2 x3...) (y2=x1 x2 x3...), ind($cmp_oprobit $cmp_oprobit)

where both y1 and y2 are ordered variables coded as = 0, 1, or 2. Explanators are common to both equations and consist of both continuous and dichotomous variables. Parameter estimation was not problematic,
Now, I would like to use "margins" to estimate marginal effects of all variables on some salient joint probabilities, e.g., Pr(y1=j, y2=k), j,k=0, 1, or 2.
Simple question--can the cmp command handle this, and if so, how?

Thanks in advance for the help.

[Mustafa Coban]

Hi Ernie,

there is a workaround to get marginal effects on the joint probabilities of a bivariate ordered probit model.

Here, a short example to compute the marginal effects on the joint probability Pr(y1 = 2, y2 = 3)

cmp (y1 = x1 x2 x3) (y2 = x3 x4 x5), ind(5 5)
replace y1 = 2
replace y2 = 3
margins, dydx(*) expr(exp(predict(lnl))) force