> Are men really less performant than women on aggregate?
I don't think that's the conclusion being reached here. What I'm getting at is performance is a VERY hard thing to quantify to everyone's satisfaction (after all, I doubt anyone thinks that they aren't a valuable employee, and everyone has numerous examples of "well this other guy sucks, I'm much better than him"), but it has a major impact on pay.
But for this specific case, there's one of two things happening:
1. As another poster said, "men" is code for "men working under H1-B visas."
2. In an effort to not look sexist, Google let the pendulum swing the other way and pay women more out of general principle.