The education premium for prostitutes

The Economist ran a long feature story, full of data on the world’s oldest profession.  Here is one bit of interest (WWBCS?):

A degree appears to raise earnings in the sex industry just as it does in the wider labour market. A study by Scott Cunningham of Baylor University and Todd Kendall of Compass Lexecon, a consultancy, shows that among prostitutes who worked during a given week, graduates earned on average 31% more than non-graduates. More lucrative working patterns rather than higher hourly rates explained the difference. Although sex workers with degrees are less likely to work than others in any given week (suggesting that they are more likely to regard prostitution as a sideline), when they do work they see more clients and for longer. Their clients tend to be older men who seek longer sessions and intimacy, rather than a brief encounter.

Are there general lessons here for the rate of return to education?  Here is another bit, when it comes to disintermediation one sex worker complains:

Moving online means prostitutes need no longer rely on the usual intermediaries—brothels and agencies; pimps and madams—to drum up business or provide a venue. Some will decide to go it alone. That means more independence, says Ana, a Spanish-American erotic masseuse who works in America and Britain. It also means more time, effort and expertise put into marketing. “You need a good website, lots of great pictures, you need to learn search-engine optimisation…it’s exhausting at times,” she says.

The full story is here.


Another factor may be that the smarter and hence sexier working girls (at least to me, I don't like dumb girls tho it's the modern trend) in fact have the ability and aptitude to go into higher education.

Trivia: porn star Belle Knox went to Duke University while debuting in the porn industry (her movie with Skylar Green at Nubiles Casting is pretty interesting, though I don't like looking at Ms. Green so much).

Belle Know is unattractive.

or Knox

...Or Not, as the case may be (let Bing Images be your guide).

The invisible hand of the market disagrees.

That's obvious. A high IQ pays in any profession. Smart girls are sexy. They can also discuss international trade, derivatives or banking regulation with their well-off clients.

If your theory is correct, then prostitutes with econ degrees should get top dollar. Something like chemistry or microbiology should be way down the list. We need to see the breakdown by major subject.

Chemistry opens up a channel of auxiliary income from producing complement drugs...

I wonder what percentage of their clients are doctors?

But bankers have alternatives- they could try harder or invest more time and sleep with women without money changing hands.
Chemists may be more captive as clients.

I agree that 'education in what, exactly?' is the fascinating question.

I see an opportunity for a website that groups listings for "escorts" by degree major.

"I get all of the johns that want to talk about civil engineering."

Education premiums or class privileges in the sex-for-pay market?
Or premiums from informal education?
If a young lady has been raised among the right sort of people, she will be in demand whether she has a diploma to show or not. Compare to the term "high-class prostitute". Presumably the necessary skills are imparted more from the general environment than from the class room.
I would not call this a signaling problem since I assume customers pay for the premium experience and not the credentials but the decision-making process is obviously not very transparent.
Average is always over for the class-based sectors.

Would it be fair to translate "class privileges" to "will not embarrass you in public"?

Many high-end escorts have their own web sites. Presumably they are able to signal their educational level on those sites, if not listing their degrees outright.

College degree as weak IQ signal with IQ being positively correlated with beauty?

In this case, we think its a kind of sorting story more than signaling. Escort markets appear to be vertically differentiated into high and low quality markets. The low quality markets tend to provide "mere transactional sex". The high quality markets tend to provide sex bundled with other intangibles that are correlated with the education variable -- compansionship related factors like conversation. I think that the education variable is proxying for a variety of unobserved factors that simply make the higher quality sex worker more productive. These men in the high quality sector are often looking for someone they can be seen out with, or talk with, as well as have sex with. But interestingly these women were surveyed from a website called the erotic review, and education doesn't show up there. I've never heard of it being something that clients screened on either. So the education variable here if it is a signal is not operating through resumes, as the kinds of resumes that exist in these markets are largely reputational resumes captured by scores on their online profile. They don't seem to advertise their education as much.

This is fascinating. Did you look at the sites of escorts themselves? The Erotic Review is a review site and often links to the provider's main site if one is available. Plus, the text reviews often talk about the relative intelligence/sophistication of the providers.

Yes we have. The "survey of adult service providers" was what we termed this particular survey. But it was a sample of the erotic review. In other work, Todd Kendall and myself (as well as recently Manisha shah and me) have studied the erotic review in more detail, but this particular paper was done using our sasp data.

Proxy for parents' socioeconomic status.

High-priced call girls are problem a Veblen service. Spitzer wasn't hiring 40 dollar hookers.

probably, not problem. damn no-edit

We control for parental education actually.

A link to Researchgate where the article "Moonlighting: Skill Premia in Commercialized Sex Markets" from Cunningham & Kendall is found:

Assuming the researchers data is good enough, there is a significant correlation between sex workers education level and weekly and hourly income. However, there are other two interesting variables: "having kids" and "being independent" that means not working for brothel, massage parlor, pimp. Being independent has a stronger effect on weekly income than college education. In discussion and conclusion sections, the latter variables are forgotten. I know the researchers were looking for the college premia, but you should report what you find not only what you're looking for.

Well we weren't "trying" to find anything. We found those patterns in the data, and wrote out a moonlighting labor supply model to explain it. The real intuition is found in the theoretical model which helps explain the role of fixed and variable costs of production in prostitution markets. We do not say that this is a true causal effect of education on earnings; rather education is (we argue) likely correlated with those fixed and variable inputs.

As for the other things you note -- if I reported everything I found in one paper, the paper would be a thousand pages long. I could write papers on this data til I'm an old man and still not mine it all. The correlations we found when we finished fielding the survey for adult service providers was absolutely fascinating. I think its the only survey of labor market practices of sex workers ever done so a lot of it is just novel empirical associations. We have other papers using the data but this is probably my personal favorite. The moonlighting model helped explain the associations very well.

Education level is correlated with being independent?

There is also selection bias, I assume partially controlled for. Women with degrees make more money than others elsewhere and so will only hook for a premium due to that option.

That's sort of what is predicted from the model that we work out in the paper. We are finding evidence for this too because the premium that we find is largely coming from variable labor supply moreso than a big premium for the educated worker. The educated worker works less often -- meaning in the survey she reported less work as a sex worked recently. But when she did work, she tended to see fewer clients but saw them for longer periods of time. These men were more likely to be regulars. The women also were more likely to have better outside options than the non educated workers evidenced by their stated second "legal" jobs.

It's all selection as we explain it, so this is not one of those education papers where we are trying to instrument for education to separate its effect from IQ or ability and so forth. Rather, what we are finding seems to be more about occupational sorting, fixed and variable costs of production, and education is correlated with the fixed and variable factors.

"when they do work they see more clients and for longer."
Your statement appears to contradict T.C.'s

Sorry - typo. The economist has it right. Typing on my phone makes me sloppy.

(1) The 'education premium' is really an 'intelligence premium'. Smart people are just better at maximising benefits, reducing costs, and evading risks, whatever their occupation, because they're able to see and exploit opportunities that others can't.
(2) Educated prostitutes were probably socialised among people of higher socio-economic status. They better understand and relate to wealthier clients, and so they're in a better position to target them as customers.

First, sex workers *may* start work as early as 13. This obviously virtually eliminates a large fraction of them from having any degree.
Second, my stereotype of sex workers includes an above 'average' number of:
a) severely emotionally distressed
b) drug or alcohol addicted
c) having mental dysfunctions severe enough to limit their options for gainful employment
d) young men and women - turn-over must be huge in this 'profession' (if one excludes the legal varieties {spouses, mistresses, supported girl-friends (and boy)}).
How do you control for the 2-4-6 years getting a degree (full time, possibly) takes? Obviously, some hook during college, but once one gets to 22, there has been a LOT of choices made, both for sex workers who are continuing their practice, and people entering the profession. It seems to me that self-reporting is going to be very, very unreliable measure of how active they've been prior to the date they claim to have become 'professionals'.

Great article. Liked the stats and graphics. Will definately make a post on our blog about it. They mentioned peppr. We are a kind of competitior site: International sex database.

Sometimes, it is the automated posts that provide the most entertainment.

This is not new at all. There has always been a tradition of more educated prostitutes commanding higher fees. The examples of ancient Greece and medieval Japan are quite well known, at least ot educated people.

In fact, the education premium should be higher or at least more consistent for prostitution than in other trades. For prostitutes, education is directly relevant to how well they do their job, since clients tend to like to talk with them as well as have sex. Where education requirements are not really necessary for the work, but imposed by regulatory, bureaucratic, or status (signalling!) reasons, we can expect the premiums to crash when these barriers are removed.

Aside from a simple data dump (our survey says "... ") the first question I have is: is this an interesting question? Is it education, or socio-economic background? Isn't it executive function that will drive income, on average? In low status professions, such as this one, I think the confounding factors virtually eliminate the possiblity of getting data clean enough to make good causal conclusions. How do you separate parenting/home life as a significant contributor to both degree attained and executive function? Is it nature or nurture? I should have more faith in "social science", (and I should actually read the article), but I really seriously doubt that in a low status, non-equilibrium (short term workers abound) trade like sex, much meaningful information is going to be easily deduced. Not to mention the coercive aspects of this work.

You'll have to read the paper more carefully and then decide if you think this is an interesting question I suppose. It's decidedly not a simple data dump though, nor do we claim it is. The paper uses a canonical labor supply model on moonlighting (ie taking on a second job) which with some modifications does a good job of explaining the patterns in our data, and which independently of that shows the important roles that fixed and variable costs of production play in this particular labor market. It is also one of the only studies in crime showing a positive association at the intensive margins for education and earnings. (It also predicts at the extensive margins a negative association between education and labor supply consistent with Becker though and that is also what we find).

Ancient Rome, too.

Stupid question perhaps, but wouldn't you expect that women working fewer hours would on average have more lucrative working patterns than those who do not, and that on average those women would also earn more?

The first clients accepted are probably the clients with the highest value to the prostitute and as more hours of a person's time are allocated to prostitution, clients with a lower value are probably accepted and this lowers the average.

This seems to square well with that for women with a higher education prostitution is a sideline, but for all women for whom prostitution is a sideline, the hours will probably be more productive on average than for those women for whom it is not a sideline. It seems to me that for those women for whom prostitution is a sideline, the marginal cost curve is probably higher. Education is one reason for this, but I can imagine that there are other reasons as well.

1. Education is correlated with conscientiousness and IQ, the latter of which is correlated with physical attractiveness, the number one feature men seek in women.
2. Higher IQ and higher conscientious women will be better at non-sexual aspects of the job, like setting up infrastructure for scale (e.g. websites) and better at scheduling/logistics, or better at hiring and managing third parties for these things.
3. Higher IQ women will be better at charming and/or manipulating their clients into seeing/paying them more.

>Are there general lessons here for the rate of return to education?

Yes, and it's the same one that's been well-known for centuries.

The education itself is largely worthless. It is, however, a good signalling mechanism that shows the recipient is capable of producing value.

"Are there general lessons here for the rate of return to education?"

You misspelled "IQ."

I'm more interested if this has anything to do with "those new service sector jobs" we hear so much about :)

Many are rushing to try and explain a direct education effect, but there's a much simpler explanation. More educated prostitutes have better alternative and thus only the ones who can command higher prices (for whatever reason, which may be totally unrelated to education) work in the field.

My book on the economics of prostitution explains all of this in much more detail. It's entitled 'Hookernomics: The Business of Sex', and is available on Kindle only (including KindleUnlimited).

Why do I get the feeling the MR commentator set is intimately familiar with this subject.

I don't know how this discussion occurs without the words "opportunity cost."

A college graduate can make good money in other professions, so prostitution would only attract those women who can make very good money as prostitutes.

A high school dropout might get into prostitution even if she made very little because her lack of education would hinder her ability to make money elsewhere and make other options less appealing.

It's not necessarily that education makes prostitutes earn more. It may just be that education prevents all but women with very high earnings potential go into prostitution.

Comments for this post are closed