I have a binary outcome variable and a binary treatment variable. Using margins, it's easy enough to see the estimated probability of Y for T=0 and T=1. But is there a stored estimate for the percentage change in probability of Y for T=1, relative to T=0? For example:

Code:
webuse nlsw88
probit union i.south
margins south
reports that people from the south (south=1) have a probability of .1702128 of joining the union, compared to a probability of .3012048 for south=0. But what I want to extract is that people from the south have a 43.49% lower probability of joining the union, compared to people not from the south, which is computed from:
Code:
di ((.1702128 -.3012048)/.3012048)
Is there a command to extract this directly from margins?