Hi all,

I'm using a gravity model to predict migration through PPML. The dependent variable is Migration in stocks (EMI). On the RHS, I have a number of variables (below I'm just showing the most relevant ones for this question): country Fixed Effects (DUM_COUNTRY*) and the Migration Rate (MR). The Migration Rate is the variable that is causing trouble. This is defined as: log(1+(EMI_lag/(EMI-EMI_lag))). The important thing to notice here is that EMI also appears on the RHS of the equation.

In terms of the estimation through PPML, this does not cause any troubles, of course, since EMI is known for the whole historical period.

ppml EMI MR DUM_COUNTRY*

However, the problem arises when I want to predict migration given that EMI appears on both sides of the equation. Is there any way to get Stata to solve this through PPML? It seems necessary to add a loop where MR is defined, but I'm not quite sure how to get it to work.

Any assistance would be greatly appreciated.

Ainhoa