I think it's more likely that the interview process would just have another step to suss out how much commuting time I'd be paying you for, and then figuring out how to divide up your day rate of $400 (higher in reality, of course) across the buckets. So Laura Long Commute gets $40/hr * 10 hrs/day and Sally Short Commute gets $50/hr * 8 hours/day.

You'd have to be careful about that. Employers might end up, in data, showing a negative impact against some protected class (married people, certain races) if they weigh hiring or compensation decisions based on proximity to the office.

It's my understanding that current jurisprudence doesn't require proof of intention to discriminate, just evidence of disparate impact. Single and married people live in different neighborhoods. Different races do as well, depending on the city. Non-heterosexual people tend to cluster in certain neighborhoods, too, for that matter. If a class action suit can show that married people are paid 10% less than single people for equivalent roles, it could be problematic.