Yes, this is is something I think about, but I feel like I don't have a very good grasp on it for most businesses. I'm not very good at thinking about it deeply, and am mostly vibing--doing some sort of hand-waving composite involving historical ROI, moat, and gut feel for how hard I think it would be to reinvest the capital. (Note: I'm not saying this is a good process. It's atrocious, really.)
I think Fairfax should have OK returns on incremental capital these days because they are still small enough and do tend to be smartly opportunistic. I think Berkshire's returns on incremental capital won't be great. They're smartly opportunistic, but too big for that strategy to have really high returns because viable opportunities are too infrequent.