Death and injury frequencies. Anecdotes of preventable deaths, e.g. where one stairwell was not enough versus where a corridor separating two stairwells became, itself, an obstacle. Drill simulations.
We should have have some basis for the rule. (If we don't, I'd support scrapping it.)
You'd need to weigh that against the costs and lost opportunities. If no one can buy an affordable apartment and overengineering is part of the cause, or if many die because of a car-centric culture because you can't built densely enough, that can be a problem.
I'd wager most of the US deaths occur in single-family homes--in basements (illegally housed), or even while sleeping in rooms with plenty of opportunity for egress.
Is there any empirical data on the safety benefits of a second stairwell?