What percentage of sex workers in the US are trafficked?

This is a really hard question. Many numbers about sex trafficking are actually about general trafficking, and the rest of them seem to bubble up inexplicably from some void of “experts say”. How did you get this number? I wanted to know, but it’s hard to find any concrete explanation anywhere. I started this post out with a quick tweet thread, but rapidly found myself in a deep, dark rabbit hole, so now I’m bringing you with me.

The United States Department’s Trafficking In Persons reports from the early 2000s are a common one, including citations on Wikipedia.

This Estimating Human Trafficking Into The United States addresses it. They say the Trafficking In Persons report said around 50,000/year were sex trafficked in 2001 and 2002, but then in 2003 and 2004, after they included “regional site visits” to what I assume are local orgs that track sex trafficking, the number drops to the commonly cited 14,500-17,500. After this the report series stops altogether (except for a 2010 report which did not include any estimates, and briefly mentions that the US has not yet instituted uniform data collection about trafficking).

The writers are pessimistic about this, saying the government provided no details about how they did these estimates and did not release their methodology.

They cite a 2006 GAO report which says depressing things like “the U.S. government’s estimate was developed by one person who did not document all his work” and “There is also a considerable discrepancy between the numbers of observed and estimated victims of human trafficking”

They say definitions are inconsistent across agencies, and there’s high risk of duplication; if a victim of sex trafficking gets help from two different government agencies, then they’re separately reported and end up in the data twice.

The writers cite Tuller 2005, which claims there’s 4,600 people under forced prostitution in the US at any given time. Tuller 2005 actually cites a Human Rights Center report, which got its number by counting people from instances reported in forced labor from 131 newspaper articles. This isn’t great, but is at least clear and concrete, which is more than what I can say for the Trafficking In Persons report.

(Interestingly, Wikipedia cites this report by saying “46% of people in slavery in the United States are forced into prostitution”, without mentioning the report’s actual number.)

And finally, there’s this 2019 Ohio study which I admittedly half-read, half-skimmed, but it looks pretty good (like, they checked for duplicates, looked at a variety of decent sources and seemed to minimize wild speculation), which concluded 1032 verified cases of labor trafficking in one year, 87% of which were sex, which is around 900. Probably most people who get sex trafficked are unverified, so this number is probably really conservative and we should multiply it by 5 or something. But also 85% of their data was from minors, and it seems “minors engaging in paid sex” qualifies as “sex trafficking” for this study. Minors engaging in paid sex is not good, but I’m not sure it qualifies as sex trafficking in the way most people conceive of it.

Squinting at this, I’m gonna shrug and multiply their number by 3, which is 0.023% of the population. Extrapolating this to the US, it gives us an estimate of ~76,000 currently-existing sex-trafficked people.

What about my prevalence/visibility survey?

I thought it was weird I hadn’t met anybody who had ever heard of anybody who’d gotten sex trafficked. What if we could measure social networks to check frequency of things?

My full writeup is here, but to summarize: I asked a bunch of people if they knew anyone who’d been sex trafficked - and if so, was it an acquaintance, a friend, or themselves? The closer to “it was me”, the higher the score. An average score of “4” means “it happens to literally everybody”, an average score of “0” means “it happens to literally nobody”

“Sex trafficking” scored 0.01. Of course maybe it’s just vastly underreported. So to check this, I also asked about a bunch of other things to figure out how actual prevalence corresponds to how much people report hearing about it in their personal spheres. I measured 19 different things - stuff like lung cancer, being a doctor, orthodox judaism, being a sex worker, dying from covid, etc. - things we know the actual prevalence for. The correlation between the reported visibility and the actual prevalence was r=0.6 (when I cut it down to 7 “low visibility” ones, where people have an incentive not to share, such as rape, homelessness, or escorting, the correlation rose to r=0.84)

Using regression and help from a math friend, this info predicted total sex trafficking at ~290,000. This estimate is a lot higher than the highest total existing estimate from Ohio, but is also a lot noiser; a lot of the data points I entered came from googling rates for other things to compare, and the number might change a lot depending on how accurate the samples were that I gathered.

But it also predicts it over a much longer period of time; in the survey, people answered yes if they knew anyone in the last five years who’d ever been sex trafficked; I don’t know how long a span this covers. If we assume it’s something like 15 years, with an average of 2 years spent actively sex trafficked (which I think I remember reading in one of the billion studies I read today but don’t feel like digging through again), then we get an estimate of ~39,000 people in the US currently sex trafficked at any time.

So in the US: the Trafficking In Persons report says ~16,000 newly trafficked per year (if we assume 2 years spent trafficked, this is ~32,000 currently sex trafficked. The Human Rights Center report says 4,600 currently sex trafficked, the Ohio study says (maybe?) ~76,000 currently sex trafficked, and mine says ~39,000 currently sex trafficked.

I don’t know if averaging these is the right thing to do, but I did it anyway, which leaves us with 37,900 (or ~0.01% of the US population, roughly equivalent to the amount of people who die while canoeing)

There. That’s my official estimate. It’s very rough - you could easily bump the number lower or higher by changing some parameters - but most sources on sex trafficking are even rougher, so I’m going with this for now.

So: How many sex workers are there?

Procon lists a bunch of sources that seem to mostly cite the National Task Force On Prostitution, which says “1 million” But the original source appears to be in a physical book somewhere, published in the 80s, and given the history of government agencies making really terrible number estimates about sex work, I don’t trust this source very much. Their estimate claims a 0.42% sex work prevalence.

(One deviation is a study in Colorado Springs over the 70s and 80s, which found 0.023% of women were sex workers, translated to 84,000 nationally. But they seemed to only look at low-class prostitutes, and claimed to know 95% of all of the prostitutes in the city(!!), so I am not sure I would generalize this to the rest of the US.)

So I did my own estimate!

I ran a survey of 165 in-person mostly-US-based female sex workers, in which 8% reported being arrested over an average of 4.5 years of work. My source was mostly (but not entirely) women who advertised online, which I would guess results in lower arrest rates (it seems likely street prostitution is easier for law enforcement to target). But probably some women are arrested many times, inflating the numbers; we can imagine a regular group of street prostitutes being subject to more frequent arrests, while online escorts go almost entirely unnoticed. I don’t know which way I should update this number, so I’m going to take it as is. This gives us an estimated arrest rate of 1.8% per year.

We do know how many prostitution arrests there are - Crime In The US reports an average of roughly 15,000 women per year in the last three years.

If 1.8% of in-person sex workers are arrested per year, then this implies around 830,000 sex workers in the US, or 0.25% of women.

edit: I forgot I could also check the predicted rates based on my visibility survey I used to determine sex trafficking frequency; this gave an estimate of 1,200,000 in-person sex workers, or 0.29% of the US population, which is pretty close to the other estimates. I’m mildly annoyed that the average of this estimate and my catch-and-release-style arrest-based estimate is basically 1 million, the same as the original mysterious source from the 80s.

So: given my estimated sex trafficking prevalence, I estimate about 3.2% of active, in-person sex workers in the US are currently being sex trafficked. This is higher than I anticipated.

I updated this number after including the visibility survey estimate; the previous number was 4.6%

I went through a lot of steps for this, and read through a lot of surveys myself, so of course the number is nowhere close to definite, and I welcome feedback and suggestions that might help. But the current literature on this is pretty spectacularly bad, so I’m hoping this can be at least one more little stone in the evidence jar.

Knowingless

What Percentage of Sex Workers in the US are…