If they pay 100% to the blonds, then they eliminate any benefit to them from hiring blonds in particular
Yes. (Maybe they get some benefit from being able to get all the blonds, though; see below.)
and reduce the rate at which this inequity is wiped out.
That's not at all clear to me, especially in the case where the number of blonds is large and especially in the case where the employer in question can use lots of them. If Employer A is paying blonds much more than everyone else, then Employer A gets the pick of the blonds, other employers have a severely reduced range of possible employees, and they therefore have reason to offer more if they ever find a blond person they want to employ. Which increases salaries for blond people and thereby reduces the inequity.
refusal to buy from sweat shops
I understand the situation you're drawing parallels with, but I don't see that there actually is a parallel. An employer who pays currently-exploited workers more isn't like a consumer who won't buy things made in sweat shops, but like a consumer who buys things made in sweat shops but sends extra money to the sweat-shop employees when s/he does. (Which isn't really possible, but it's not my fault you picked that analogy.)
not ethically superior to other forms of charity
Oh, quite likely. But the same goes for any form of good behaviour that has a personal cost, whether it's an act of heroic virtue or just a matter of common decency. Instead of being polite to other people I could be brusque and rude for the sake of efficiency, make a bit of extra money with the time I gain, and send the proceeds to AMF. Instead of being honest I could shoplift, sell the stuff I steal, and send the proceeds to AMF. All true, and it's (uncomfortably) quite possible that the best thing really is to behave in whatever way maximizes money and send it all to the most effective charity one can find -- even if "whatever way maximizes money" would generally be considered odious.
But (1) it does seem that the vast majority of people behaving in that sort of way aren't actually doing it in order to send more money to effective charities, and (2) the heuristic of treating others decently even if it means passing up some chances of making extra money to feed starving people or prevent malaria does have its advantages.
And (this is of course the point) it's not at all clear that not paying women 20% less than men for equivalent work doesn't fall into that category of "treating others decently".
If Employer A is paying blonds much more than everyone else, then Employer A gets the pick of the blonds,
If Employer A is paying blondes any more than the others, not 'much more'. They get the pick either way.
other employers have a severely reduced range of possible employees, and they therefore have reason to offer more if they ever find a blond person they want to employ.
One employer isn't usually going to make that big a difference, but yes, that's how it works: since Employer A is bidding up the price of blondes, the other employers will need t...
LW readers have unusual views on many subjects. Efficient Market Hypothesis notwithstanding, many of these are probably alien to most people in finance. So it's plausible they might have implications that are not yet fully integrated into current asset prices. And if you rightfully believe something that most people do not believe, you should be able to make money off that.
Here's an example for a different group. Feminists believe that women are paid less than men for no good economic reason. If this is the case, feminists should invest in companies that hire many women, and short those which hire few women, to take advantage of the cheaper labour costs. And I can think of examples for groups like Socialists, Neoreactionaries, etc. - cases where their positive beliefs have strong implications for economic predictions. But I struggle to think of such ones for LessWrong, which is why I am asking you. Can you think of any unusual LW-type beliefs that have strong economic implications (say over the next 1-3 years)?
Wei Dai has previously commented on a similar phenomena, but I'm interested in a wider class of phenomena.