There is academic work studying this. The superlinear return on time invested does occur in a variety of fields, and pay tends to follow the productivity gains.
Here is a paper which cites a lot of the literature (note: primary topic of the article is flamewar inducing):
I didn't read the paper in its entirety, but is the argument something like
competition introduces a superlinear factor in the time/pay equation
superlinear factors on average hurt women more than men since the latter are more likely to be on the upper end of workload (assuming a linear correlation of time/pay, you'd already get from a 77% "wage gap" to a pay gap of around 92%, which would tighten further or disappear if we account for "winner-takes-it-all" distribution of pay)
I find it relatively easy to believe that today's working environment leads to an incentive structure similar to all-pay auctions, where the employer reaps huge benefits due to everyone's fiction of attainable winning, and which hurt a majority of people, but women more visibly than men.
Law firms are a prime example of that. When the client calls at the end of a 60-hour work week, he wants to talk to the guy who is primary on the case. Even if he's burned out a bit, the answer he gets from him is still better than the answer any other person in the law firm could give him
Here is a paper which cites a lot of the literature (note: primary topic of the article is flamewar inducing):
http://www.aeaweb.org/aea/2014conference/program/retrieve.ph...