Dear Statalist,

First post here!

I am working with a panel of exports from Spain to 170 countries. These are divided into three categories, homogeneous goods, heterogeneous goods and food and beverage. I would like to now if the effects for my independent variables change for homogeneous and food.

My question is: should I estimate a model with y x x*H x*F (where H and F are dummies that take value of 1 if good is homogeneous or food respectively), or should I instead estimate two separate equations, one with each interaction term?

I have searched my relevant literature, but all papers simply discriminate between homogeneous and heterogeneous goods, while I would also be interested on 'food'.

Thanks

Matt