Monday, 11 April 2011

How trafficking is counted

There is an excellent article by investigative journalist Nick Davies, a primer on how the UK trafficking numbers were blown out of all proportion: Anatomy of a moral panic. (For commentary about US-based numbers, here's a blog about the topic.)

To summarise Davies's article, a paper which estimated a very small range - between 142 and 1,420 trafficked sex workers in the UK per year - was misreported and misinterpreted, ending up with people claiming 4,000 (or even as many as 80,000!) trafficked women entering the UK sex trade every year.

Part of the problem with these kinds of numbers is that while they're very, very wrong, they are also difficult to disprove. And the idea, once it enters mainstream media, is difficult to dislodge, even with facts.

In many social research fields, exact numbers can be hard to come by. Even seemingly straightforward calculations are fraught with error. Let's take a simple example. Imagine you were trying to count the number of people living in the world (and that you, like Father Christmas, are able to get to every household in an unfeasibly short amount of time). It would be a hard job. By the time the count was finished, loads of those counted would have died, and even more would have been born. An actual number that represents the real number of living people on earth at any one time? It's impossible. So, the world population is actually an estimate made based on some facts known about the countries of the world, their last population estimate, and their birth and death rates.

Making this kind of an estimate is a “Fermi problem”. Enrico Fermi, one of the physicists who worked on the Manhattan Project, was reputedly able to make accurate guesses based on limited information.

Here’s an example of a Fermi problem in action. I was at pub quiz one week, and our team was tied for the lead. The tiebreaker was the question “How many performances did Yul Brynner have as the King of Siam in The King and I on Broadway?” As the only former drama geek in our team, it came down to me.

I calculated that Brynner probably did 8 performances a week ("once a day and twice on Sundays", as the saying goes). It’s a full-time job, so minus a two-week holiday, Brynner was probably performing 50 weeks a year. I wasn’t sure how many years it ran but knew he had been in at least one revival of the popular musical, so let’s say ten years of being the King total. That makes an estimate of:

8 shows a week x 50 weeks a year x 10 years = 4,000 shows

Sounds pretty high, right? We won the tiebreak (and the quiz) because, as it turned out, the real answer is 4,525. I was off by over 10%, which would be terrible for science, but was good enough for the quiz. The other team guessed 600... way too low. Picking a number out of thin air, as the other team probably did, is fraught with error. It’s hard to make good guesses with no information. Apply a few basic assumptions, however, and your accuracy goes up rapidly.

Fermi problems are great for pub quizzes, less so for evidence-based reporting. Common-or-garden estimates are not the stuff on which good research is built. At the very least, applying a set of assumptions to estimate a number should meet two major criteria:
1. The assumptions must have some foundation in reality. Eight Broadway performances a week is reasonable; 80 wouldn’t be.

2. The method of calculation needs to be explained. If an assumption turns out to be wrong, the calculation can then be adjusted. I don’t think the other person on my team would have bought 4,000 as an answer if he hadn’t seen my reasoning.
What does this have to do with the trafficking estimates?

The people who claim there are thousands, or even tens of thousands, of sex slaves in Britain are claiming an unrealistically high number. So unrealistic in fact that if it were true, that would mean the vast majority of prostitution in Britain was undertaken by trafficked people. That violates the first principle - basis in reality. Some people involved in sex work have encountered people who may have been trafficked; the vast majority have not. So either there's a whole other sex industry going on that no one in the sex industry knows about, or... the calculations are wrong.

Part of the difficulty with fighting such unrealistic claims, however, is getting good estimates to counter them. There is no comprehensive UK mapping of sex workers, much less trafficked ones, but there are some estimates. As part of the European Network for HIV/STD Prevention in Prostitution (EUROPAP), Hilary Kinnell contacted projects providing services for sex workers. [pdf] She had 17 responses. The average number of prostitutes per project was 665. She then multiplied that figure by 120, the total number of projects on her mailing list, to get an estimate of 79,800. This total includes women, men, and transgender women and men sex workers in the UK.

Kinnell notes there are obvious problems with this particular Fermi problem: the centres responding might be larger than most, some sex workers might use more than one centre. She finds it strange that number - ten years old, a huge estimate, and taken out of context - is still quoted. "The figure was picked up by all kinds of people and quoted with great confidence but I was never myself at all confident about it. I felt it could be higher, but it also could have been lower."

Meanwhile data from the UK Network of Sex Work Projects (UKNSWP) records an estimate of 17,081 sex workers in some kind of contact with centres. Of these 4178 - about 24% - work on the street. A larger total for all sex workers was 48,393. More recent, and rather lower, than the 1999 estimate. So if the trafficking hype is correct, that would make anywhere from one in 12 to as high as one in 2 sex workers in the UK the victim of trafficking.

Let's go back to the paper which kicked this all off, the one that estimated a range of 142 to 1,420 trafficked sex workers in the UK. Now, a note about that number: it included not only women who were trafficked against their will, but also women who willingly arrived (perhaps illegally) to the UK for sex work. In other words, Kelly and Regan’s total included both unwilling and willing sex migrants.

Part of the problem is how different groups define “trafficked”. Some assume that if someone is not British and is working in the sex trade, she must be trafficked. That’s quite a leap in logic! Hold on a sec - I was born abroad. And I worked in the sex trade. Does that mean they count me as "trafficked"? WTF?

The Poppy Project reported in 2004 that 80% of prostitutes in London flats were foreign-born. But there is no evidence that those women were trafficked or that this high proportion of foreign sex workers to natives is true of the entire UK. (In fact, evidence puts the UK-wide proportion closer to 37%.)

‘Foreign-born’ also includes citizens of other EU countries, who have the automatic right to live and work in the UK. Eaves, the organisation that includes the Poppy Project, did an interesting nip-and-tuck on reporting the origins of women working in the sex trade in London. In their 2004 report Sex In the City [pdf], they claimed 25% of women working in London were from Eastern Europe. But look closer - they have classified Italy and Greece as “Eastern European” countries.

Why? Well, the reason is given that “because these ethnicities are often used to code women from the Balkan region, advised by pimps and traffickers to lie about their ethnicity to avoid immigration issues.” Hey, my dad is Italian... if I said this to a researcher, does that mean they would assume I'm really Eastern European? That violates the second principle of the realistic estimate: show your work clearly. It’s the kind of sloppy calculation that throws all subsequent conclusions into question. It's bad Fermi.

So if some people who come here voluntarily can be erroneously called “trafficked,” then what is “trafficking”, exactly? The Palermo Protocol to Prevent, Suppress and Punish Trafficking in Persons Especially Women and Children, part of the 2000 UN Convention against Transnational Organized Crime defines ‘trafficking’ as
…the threat or use of force or other forms of coercion, of abduction, of fraud, of deception, of the abuse of power or of a position of vulnerability or of the giving or receiving of payments or benefits to achieve the consent of a person having control over another person, for the purpose of exploitation.
In other words, illegal migration for purposes of economic advantage, if undertaken willingly, is not trafficking. If nothing else, it's worth remembering the excellent analogy offered by Charlie Glickman:

Sex work is to Trafficking as Consensual Sex is to Rape.

Just because rape exists (and is rightly both reviled and illegal), that doesn't mean banning sex would solve any problems. What Glickman's statement encapsulates so brilliantly is that while trafficking occurs within sex work, that in and of itself is no good reason to either equate the two, or to ban sex work. Pumping up the trafficking numbers might be great for getting media attention, but it does nothing to solve the real problems of people who are really trafficked.

Claiming huge numbers of trafficked sex slaves where they do not exist distracts attention and resources from the (far smaller) number of people forced into sex who genuinely need assistance. And I for one think inflating a problem is not only unethical, it's dangerous to real victims. Let's get our terminology right, at the very least. Let's start with realistic research and maybe someday we'll get realistic results.