I have a dataset where the dependent variable has 4 categories namely P1, Q1, R1, S1. There is a possibility of correlation among the error terms. Therefore, I am using SUR model. My variable of interest in the regressor is 'ethnicity'. I would like to impose the restriction that the impact of 'ethnicity' on P1,Q1,R1,S1 is the same. How can I impose this cross equation restriction for the 4 equations.