[Omar Shaher]

Dear Prof. Clyde,

Thanks very much for your answer. Greatly appreciated.

I understood the case now, so if the margins calculate the predicted outcomes according to the regression command, then that means all points in the blue curve for all years reflect what leverage would have been if the policy had not already been adopted by everyone. On the contrary, all points in the red curve for all years reflect what leverage would have been if the policy had already been adopted by everyone.

But, do you think that we can get the curve for the actual values (averages) instead of the predictive values by using the following command:

by Event Year, sort: summ Leverage

But if the above command is correct, I don't know how to graph it?

[Clyde Schechter]

To graph the actual values (which will not be adjusted for your other covariates):

collapse (mean) Leverage, by(Event Year)
reshape wide Leverage, i(year) j(Event)
graph twoway line Leverage* Year

[Omar Shaher]

Dear Professor Clyde, I can't thank you enough for all your great efforts. I will never ever forget your help. It is an honor for me to know a great person like you.

My deepest respect to you and to your great efforts.

Million thanks.