I have almost no data to back this up, but it's my suspicion that the current workweek is something of a lowest common denominator: across all job types, company needs, etc., it maximizes productivity while keeping most HR polices identical.

I mean, imagine the administrative overhead to create a system like you propose: you might want to work 25% less, but I really think a decent number of young and single software engineers could actually work something like 40% less and be OK at a 40% reduction in salary. Somewhere there is someone that is working on an interesting problem and, in spurts, wants to work 30% more. This variability increases as you look at, say, the sales team, the accounting team, etc, and all other positions that have different optimal working styles.

Now, imagine creating a system to make sure you pay people fairly in that world (remember, not even everyone with the same role/job title starts with the same "100%"). Or a way find out who your real best/most essential employees are... and so on. I honestly think that the workweek, as it is, exists mostly because as a company becomes sufficiently complex a few assumptions need to be made in order to keep it functioning properly, and '5 days, 8-10 hours a day' (or whatever) is an assumption that everyone can more or less stomach - even if they don't enjoy it.