Edlund and Korn  (EK) proposed that prostitutes are well paid and that the wage premium reflects foregone marriage market opportunities. However, studies of street prostitution in the U.S. have revealed only modest wages and considerable risks of disease and violence, casting doubt on EK’s premise of an unexplained wage premium. In this paper, we present evidence from high-end prostitution, the so called escort market, a market that is, if not entirely safe, notably safer than street prostitution. Analyzing wage information on more than 40,000 escorts in the U.S. and Canada collected from a web site, we find strong support for EK. First, escorts in the sample earn high wages, on average $280/hour. Second, while looks decline monotonically with age, wages follow a hump-shaped pattern, with a peak in the 26-30 age bracket, which coincides with the most intensive marriage ages for women in the U.S. Third, the age-wage profile is significantly flatter, and prices are lower (5%), despite slightly better escort characteristics, in cities that rank high in terms of conferences, suggesting that servicing men in transit is associated with less stigma. Fourth, this hump in the age-wage profile is absent among escorts for whom the marriage market penalty is lower or absent: escorts who do not provide sex and transsexuals.
One should also read Lena Edlund’s earlier paper: “A theory of prostitution”
Some assumptions and implications in these papers:
1) Men will not marry women who are prostitutes. Corollary: Prostitutes are either single or married to losers who didn’t mind their wives are prostitutes.
2) Women who are prostitutes would otherwise be compensated $280 (or some equivalent of this in consumption) if they had gotten married instead.
3) If these women were out being prostitutes, they would choose to get married.
The part I don’t get about is this: if Edlund, Engelberg, and Parsons (henceforth, EEP) are right about wage premium amongst escort represents forgone opportunities in the marriage market then one of the follow is true:
1) These women would have to marry millionaires who can compensate them in some ways equivalent to being paid $280/hour.
2) The combination of marriage and childbearing holds the same utility on the women’s preference curve as being paid $280/hour.
3) If neither 1) nor 2) are true, then women who choose marriage in their mid-to-late twenties are probably making an irrational choice: they should probably be escorts till they are past 30 years old for travelling businessmen, and then lie to their future husband about their previous career choice.
There is also a funny bit in there about transsexual escorts. According to EEP’s logic, they act as a control group. “Transsexuals provide a convenient comparison: while their jobs as escorts are very similar, their marriage prospects are not. According to the logic underpinning EK, if they married as husbands, they would be on the paying side, and if their gender role was that of the wife, they are clearly barren and therefore the basis for compensation would be moot.”
Using the same methodology, they found that transsexual escorts make 10% – 22% less than cisgender escorts. Those “barren” transsexuals still seems to get paid a lot. EEP doesn’t seem to explain why these “low skilled”, “commoditized” workers still gets paid so much.