## Wednesday, August 08, 2012

### How many prostitutes are there really?

As some of you have noticed, I am particularly fond of numbers. For years I was contemplating how to properly calculate things like: the total number of prostitutes in the Netherlands, the total number of prostitutes in the Netherlands at any moment, how many are forced, what percentage of the prostitutes are forced, etc...

I tried to find statistical means to do these calculations. I studied websites of brothels to count the number of prostitutes who worked there, so I could know approximately the average number of prostitutes per brothel, then count all the brothels, so I could make a estimate of the total number of prostitutes who work in brothels (not including window prostitution). I did the same with escort agencies. Several times I studied all the reviews on hookers.nl written by clients of prostitutes. This way I could know something about the ages of prostitutes, their countries of origin, even the colour of their hair. I also studied all media reports about forced prostitutes. I could compare the numbers derived from hookers.nl with what is known about victims of forced prostitution. Are there some patterns which can be discerned?

Or I could study research done by professional researchers who have interviewed many hundreds of prostitutes, what do these prostitutes tell about being forced? Or I could look at the eyewitness reports by social workers, policemen and prostitutes.

All these methods allowed me to make some crude estimates of the number of prostitutes or forced prostitutes in the Netherlands. However, it is difficult to find reasonably random samples. It can for instance be noticed that prostitutes who are interviewed by researchers nearly all say that they work voluntary in prostitution. But some say that they have been coerced to work in prostitution at least during some period(s) in the past. This is obviously a paradox if you assume that the prostitutes who were interviewed are part of a random sample. Probably the forced prostitutes simply are not interviewed, only the prostitutes who worked voluntarily in prostitution at the moment they were interviewed. Some however do indicate that they have been coerced in the past.

Adding this all up together I arrive at the conclusion that forced prostitutes are often very young, usually around 20. Prostitutes in general are 30 on average. I estimate that around 80% of the prostitutes are older than 22, and around 60% are older than 26. This means that probably most prostitutes work voluntarily in prostitution. I estimate that around 10% or 20% of the prostitutes in the Netherlands are coerced. I base this partly on estimates done by prostitutes themselves in a study done by Ine Vanwesenbeek (it is the report De sociale positie van prostituees in de gereguleerde bedrijven from the year 2002, which I mention later here in this post). The average estimate made by prostitutes themselves of the percentage of prostitutes who are forced in their surroundings was a little higher than 10%. And I also base myself on the notion that many prostitutes on De Wallen in Amsterdam are reported to be coerced, and the fact that many forced prostitutes in Amsterdam work on De Wallen. This, together with an estimate of the percentage of prostitutes in the Netherlands who work on De Wallen, and together with an estimate of the total number of forced prostitutes in the Netherlands who work in Amsterdam, leads me to a percentage of at least or around 10%, roughly. I also conclude - based on eyewitness reports -  that a great percentage of Eastern European prostitutes are coerced. About the Eastern European prostitutes, see this post on my blog:

But there are always very critical people  who challenge everything, and deny or minimize the problem of forced prostitution. Or who call forced prostitution a myth. I think they will never be satisfied by any evidence. On the other hand, I doubt the existence of forced prostitution myself sometimes. ( :-P )

By accident I stumbled upon research done by the RutgersWPF foundation. They have interviewed many thousands of Dutch men and women, and asked them all kinds of questions about sexuality. They didn't research prostitutes per se in the research, but they did asked the people about if they have ever paid for sex or if they have ever been paid for sex, if they have been forced to do that, etc... People gave some surprising answers. The research done by RutgersWPF clearly proves in my opinion that there really is a problem of forced prostitution in the Netherlands. I only have to give one warning. The researchers say that to make the study more representative of the Dutch population in general, they have invited special groups of people who would have otherwise been undersampled. They are vague about which groups they have invited. Could they be for instance…. prostitutes they have recruited at a shelter for battered women? On the other hand, the report says they have oversampled for Surinamese, Antillean, Turkish and Moroccan people, and for the age group 15-45. They didn't say that they oversampled for prostitutes or forced prostitutes. So I assume that the sample regarding prostitutes is reasonably random.

I find the reports important because all kinds of hidden prostitutes come into view, who would otherwise be hidden. As I have said before, there have been groups of prostitutes who have been interviewed, and they were asked if they were forced now or in the past, but former prostitutes and possibly also forced prostitutes have not been interviewed. So I find the reports of RutgersWPF important because they could give a more complete view of prostitution and prostitutes. Prostitutes who are interviewed by researchers almost never say that they have been coerced to work in prostitution, but some say they have been in the past. Usually the percentage of prostitutes who say that they have been coerced in the past ranges from 8% to over 40%. I can discern a declining trend however. Usually the higher percentages (30% or 40%) are from decades ago. I refer to research done by the sexologist Ine Vanwesenbeeck in the 80s and 90s (do a google search, hint: look at Prostitutes’ well-being and risk and Er gaat iets veranderen in de prostitutie which I mention later in this post). A more recent research from 2006 to study the social position of prostitutes indicates a percentage of 8% (De sociale positie prostituees 2006). So perhaps forced prostitution is declining over the years. I talk about this issue in this post on my blog:

Anyway, now the results.

The first research of RutgersWPF which I want to refer to is:
SEKSUELE GEZONDHEID IN NEDERLAND 2009 (written by many persons, 2009)
http://www.rutgerswpf.nl/sites/default/files/Seksuele%20Gezondheid%20in%20Nederland%202009.pdf

The title means: sexual health in the Netherlands 2009. They have interviewed 3221 men and 3207 women, aged between 15 and 70. On page 6 and 7 it is said that (translated by me from Dutch into English):
One fifth of the men has ever paid for sex, 4% also did it in the past 12 months. Women make use of paid sex considerably less often: two of every thousand women ever paid for sex, and one of every thousand did this the past year. […] In the youngest age categories the percentage of men who have ever paid for sex is understandably the lowest. Men with a Turkish, Moroccan, Antillean and Surinamese background say relatively often that they pay for sex.
Approximately an equal percentage of men (3,4%) and women (2,2%) have at any time received money in exchange for sex. One percent of the men and women have received money multiple times in exchange for sex. Having received money for sex multiple times can be viewed as an indication for sex work. Of the boys younger than 26 who go to school, lowly educated boys more often have experience with receiving money for sex than the highly educated boys. Concerning ethnic background, a relatively high prevalence emerges among Antillean men and Surinamese women.
A small group of men (2,5%) and women (1,4%) indicate to have had sex at any time in exchange for something other than money (so-called 'exchange sex'). During the past year four of every thousand men did this, and two of every thousand women. In most cases it concerned sex in exchange for a little present, such as clothes, shoes or jewellery. In addition to presents, also cigarettes, drugs and alcohol are named as a means of exchange for sex. In this research no differences arise between sociodemographic subgroups regarding the incidence of exchange sex.
Of all people who have at some time received money or another reward in exchange for sex (n=225) 28% was persuaded or coerced to do this at least once. For men it has more often been always their own choice (79%) than for women (61%). For women the one who persuaded or coerced them nearly always was a man. Among men this happened approximately equally often by a man and by a woman. In 4/5 of the cases the perpetrator was an acquaintance, mostly an (ex)partner, colleague, vague acquaintance or somebody from the neighbourhood.
One of every twenty men and almost a third of the women who have experience with paid sex or exchange sex, have at any time had a need for help concerning this. Within this group 47% of the men and 75% of the women have actually had contact with social work. Especially women who have experience with exchange sex relatively often indicate a need for care (44%).
Next report:

SEKSUELE GEZONDHEID IN NEDERLAND 2006 (written by many persons, 2006)

http://www.rutgerswpf.nl/sites/default/files/Seksuele%20Gezondheid%20in%20Nederland%202006_1.pdf

In this report 2072 men and 2075 women have been interviewed. In this research the researchers didn't ask how many of the (former) prostitutes have been forced to do this.
I will quote from it (translated by me from Dutch into English):
On pages 51-52:
A quarter of the men have paid for sex at any time, 5,1% also did it the during last year and 3,6% even did it during the last 6 months. Women, not surprisingly, make considerably less often use of paid sex: 0,4% of the women paid for sex at any time, 0,2% did this during the last year or half a year. The number of men who indicate to pay for sex sometimes seems to have been doubled within 15 years time: at that time 13,5% of the men had paid for sex at any time, and 2,6% during the last year. It is possible that the tendency to silence paid contacts has decreased, possibly also because of a different method of collecting data (interviews in 1991, compared to an internet questionnaire in 2006). There are no indications from the field that one can really speak of such a strong increase in the number of visitors of prostitutes.
In the recently finished research among young people younger than 25, the presentation of the question connected to paid sex was broader: it has been asked if one has given or received money or another reward in exchange for sex (De Graaf et al., 2005). This gives the expectation that the percentage that answers positively to these questions is higher than in the current research. Concerning paying for sex the percentages are higher here contrary to expectation, especially among men. In the current study 12,9% of the 19-25 years old men say that they have paid for sex at any time, in the current research among young people it was 8,1% for this group.
International figures about paying for sex are nearly all connected to men. Leridon, Van Zessen and Hubert (1998) have listed all the studies which have been done in the beginning of the 90s in multiple European countries. American (Michael et al., 1994), Swedish (Lewin et al., 1998) and Australian figures (Rissel, Richters, Grulich, De Visser and Smith, 2003) can be added to this comparison. The percentages of men who have indicated to have ever paid for sex vary enormously internationally: England 6,6%, Finland 9,9%, Norway 11,3%, Sweden 12,7%, Australia 15,6%, United States 16%, Switzerland 19,7% and Spain 38,6%. Only the Spaniards seem to surpass the Dutch in (openness about) the use of commercial sex.
Because of the very small group of women who have paid for sex at any time, the differences regarding demographic characteristics are only being presented for men (table 5.3). In the youngest age group (18-34 years) the percentage of men who have paid for sex at any time is understandably the lowest. When one looks at the last year, then it becomes visible that visits to prostitutes go down in connection with becoming older, albeit not significantly. Among lowly educated people paying for sex happens more often (approximately one of every three lowly educated men at any time paid for sex) than among highly educated people (approximately one of every five). Men with a Moroccan or Turkish background and especially Antillean or Surinamese men say relatively often that they pay for sex. Men with a Jewish-Christian faith say relatively less often that they pay for sex at any time, men with an Islamic background and men without a religion relatively often say it. In the strongly urbanised areas relatively often there is being paid for sex, possibly because of the bigger supply. Finally prostitution is practised the most by single men and the least often by married men.
An equally large segment of the men as well as of the women (3,2%) has at any time received money in exchange for sex. This is more than in other countries: in Sweden 0,5% of the men and 0,3% of the women received money in exchange for sex (Lewin et al., 1998), in Finland 1,5% of the men and 0,2% of the women (Kontula & Haavio-Mannila, 1995) and in Australia 0,9% of the men and 0,5% of the women (Rissel et al., 2003). Also, these percentages were lower in the Netherlands 15 years earlier: approximately 1% of both men as well as the women said at that time that they have been paid for sex at any time (Van Zessen & Sandfort, 1991). We see the relatively normalisation of the prostitution branch since then (and in the Netherlands in comparison with other countries) here as the most important reason for the differences.
In the recently finished research among young people, 1,2% of the 19-25 years old women say that they have received money or another reward at any time, in the current research 0,4% say that they have received money at any time (De Graaf et al., 2005). The difference is probably traced back to the more broader presentation of the questions in which also 'other rewards' have been admitted. Among the 19-25 years old men, despite the difference in presentation of the question, no differences are to be found: in both studies the percentage that at any time has received money (or another reward) is 2,4%.
On page 54:
For three quarters of the men and 85% of the women who have received money for sex at any time, this took place longer than a year ago. For men this was more often a once-only event than for women: of the men 0,7% (n=14) has received money multiple times in exchange for sex, among women this is 1,7% (n=35). The last thing (having received money multiple times for sex) can be viewed as an indication for sex work. Because the group of sex workers among men is too small for a further analysis, the demographic differences are only ascertained for the women in this group (table 5.3). In strongly urbanised areas there are more women who have done sex work. Aside from this, these women have relatively often a LAT-relationship [a living-apart-together relationship] and they are rarely married.
I find this information in those two reports very surprising. This means that men are more often involved in prostitution than women. And also men are very often involved (as victims) in forced prostitution as well!

WARNING: what I'm going to do now is very tricky from a statisticians' point of view. They will CRINGE!!! But I like to play with numbers anyway. Let's see what happens. I think it is a good illustration anyway of how difficult it is to find reliable ways to estimate the number of prostitutes or forced prostitutes in the Netherlands or in any other country, or worldwide.

I will assume that the situation doesn't change, and when I say 'at each moment' then I assume that this situation counts at all times. This is also true when I say 'during the last' year. I then assume that every year the situation is the same, so 'during one year' is identical to 'during the last year'.

Okay, okay, there are some beautiful tables in the 2009 report. Such as table 8 on page 42. These tables give some more detailed numbers, which I can use to make some calculations. Because I'm not very interested in men, I skip them. I want to know how many women have worked in prostitution, and who have worked in prostitution during the last year. It turns out that 3,2% (2006) or 2,2% (2009) of the women were at least once paid for sex, 1,7% (2006) or 0,8% (2009) have repeatedly been paid for sex, and 0,5% (2006) or 0,3% (2009) have been paid for sex during the last 12 months. The differences in percentages between the two studies makes you aware of the big error margins.

3207 women have been interviewed in the 2009 study. So that means approximately 71 women were at least once paid for sex, 26 women have repeatedly been paid for sex, and 10 women have been paid for sex during the last 12 months. The numbers are not very large, so this means the error margins must be large. If 39% of the women who have been paid for sex have been coerced or persuaded to do this, this means approximately 28 women. And thinking about error margins: when you think about 39% who have been coerced or persuaded, you must rather think about a percentage between 25% and 55% with 95% certainty, because 71 is such a low number. I calculate this using a formula which I figured out myself, but which is also confirmed in a book about statistics. Actually it is this formula:

error margin = 1,96/2 * √(1 – 4*(B/N – 0,5)^2 ) / √(N) * 100%

where N is the size of the population, and B is the size of the population with a certain characteristic that you have measured. Usually I ignore the 1,96/2 and round it off to 1. The 95%-confidence interval is approximately between:

B/N * 100% - error margin and B/N * 100% + error margin.

Actually, the statistics book uses a different notation for the error margin, but this means mathematically the same:

error margin = 1,96 * √( p*(1-p) / N ) * 100%

where p = B/N

Actually, a more accurate formula is:

error margin = 1,96 * √( p*(1-p) / (N+3) ) * 100%

where p = (B+1)/(N+2)

But I don't use this one, it only really differs for a B near 0 or N.

Despite the large margin of error, still this 39% (of the prostitutes who have been coerced or persuaded by a third person) is a surprising number, which does indicate that forced prostitution indeed is a problem in the Netherlands.

Now I want to extrapolate to the whole Dutch female population. I want to combine both studies to make the error margins smaller. The 2006 study had a sample of 2075 women and the 2009 study of 3207 women. If its true that 0,3% of the women in the 2009 study worked in prostitution during the last year and 0,5% of the 2006 study, then I calculate that of the combined number of 5282 women (assuming no overlap) 0,37% worked in prostitution during the last year, and I calculate the error margin to be: 0,17% (again, using the formula from the statistics book). This 0,37% represents approximately 20 women. So between 0,20% and 0,55% worked in prostitution during the last year. If you assume that approximately 7.000.000 Dutch women are older than 15, then approximately 26.000 Dutch women have worked in prostitution during the last year (if you take error margins into account then you arrive at between 14.000 and 39.000 prostitutes). I have tried to figure out how many women actually work in all those brothels, and in my opinion it cannot be many more than 7000. Perhaps some 4000 women work as escorts and homeworkers. So let's say some 10.000 women work in prostitution in the Netherlands according to my own earlier calculation. The RutgerWPF studies indicate that my estimate may be a bit too low. On the other hand, the number 10.000 counts at a single moment, and during one year there are some new recruits, or prostitutes who start working in prostitution again after they have stopped for a while. So perhaps on a yearly basis there could indeed be more than 14.000 prostitutes in the Netherlands.

Now I want to move my attention to the percentage of women who have been coerced or persuaded. Obviously, this group hasn't been coerced or persuaded throughout their whole career. I want to know how many prostitutes are forced at any given time. I do this by dividing the number of years which prostitutes are forced to work in prostitution by the number of years they have totally worked in prostitution. I have calculated once that forced prostitutes are approximately forced to work two years in prostitution (with an error margin of half a year, that means between 1,5 and 2,5 years). I have calculated this number after studying many hundreds of cases of forced prostitution. Unfortunately, I have no clear information of how long prostitutes in general work in prostitution. This varies heavily from study to study.

According to the following report from 2002 the average number of years that the 230 interviewed prostitutes have worked in prostitution is 5,1 years:

De sociale positie van prostituees in de gereguleerde bedrijven - Een jaar na wetswijziging (Vanwesenbeeck, I., Höing, M., Vennix, P., 2002), which you can download here:
http://wodc.nl/images/ewb02prostd_Volledige%20tekst_tcm44-57344.pdf

According to the following report from 2006 the average number of years that the 354 interviewed prostitutes have worked in prostitution is 2,3 years:

Evaluatie opheffing bordeelverbod - De sociale positie van prostituees 2006 (Dekker, H., Tap, R., Homburg, G., 2006), which you can download here:
http://wodc.nl/images/1204b-volledige-tekst_tcm44-59312.pdf

So which number is the right number? And another problem is that the prostitutes who have been interviewed in these previous two mentioned reports usually keep on working in prostitution after they have been interviewed. But I use a trick here, I assume that the distribution of the number of years that prostitutes have worked in prostitution is a negative exponential. If the distribution is a negative exponential then the average number of years that the interviewed prostitutes have worked in prostitution in one sample is also the average number of years that all prostitutes have worked and will work in prostitution in general.

But which number to choose? 5,1 years or 2,3 years? Perhaps just take a confidence interval of 2,3-5,1 years. Well... I just pick five years (I assume this to be also true for forced prostitutes). And now the trick, if it is true that forced prostitutes are on average forced to work in prostitution for two years, but they work for five years in prostitution as a whole, then I guess that 2/5 X 100% = 40% of the time they have worked in prostitution they were coerced to work in prostitution. So if you want to know how many prostitutes are coerced or persuaded to work in prostitution at each moment, take the 39% who have been coerced or persuaded at least once during their career, and take 40% of this 39%, and you will end up with approximately 16%. However, if instead of taking the five years that prostitutes work in prostitution, you take the 2,3 years, then you will end up with 39% X 2/2,3 = 34%. So, these numbers are just spielerei. But in absolute numbers, 16% of 26.000 women is approximately 4200, and 34% of 26.000 women is approximately 8800 women. But I repeat, this is just spielerei. It now becomes really vague. Too vague. Taking into account the full range of error margins then I arrive at between 7% and 55% of the total number of prostitutes who are forced at any given moment. In absolute numbers between 1000 and 21.500 prostitutes in the Netherlands might be forced at any given moment. So, I know nothing.

Oh, I have found a different way to calculate the percentage of forced prostitutes! If it is true that the average number of years in which a forced prostitute is forced to work in prostitution is two years (with an error margin of half a year, so that means between 1,5 and 2,5 years), and you know that the average woman in the Netherlands becomes 80 years old, then you know that of the 80 years that a forced prostitutes lives, 2 years she is forced to work in prostitution, that means 2,5% of the time. In the 2009 study 28 of the 3207 women were once forced or persuaded to work in prostitution. That's 0,873% of the total number of women, with an error margin of 0,313%. If 0,873% of all women have been forced or persuaded to work in prostitution once, and this happened 2,5% of the time during their lives, then I arrive at 0,873 X 0,025 = 0,022% of the Dutch women older than 15 years are forced to work in prostitution at one given moment (I estimate the 95% confidence interval to be around 0,0105%-0,0371%). In absolute numbers this is 7000.000 times 0,00022 = approximately 1500 women at each moment (that's between 735 and 2595 if you take error margins into account). If you want to calculate this on a yearly basis you simply have to add exactly 1 year to the estimated 2 years that forced prostitutes are forced to work in prostitution on average. Then I arrive at 0,033% of the total number of women who have been forced to work in prostitution during the last year (the 95% confidence interval is approximately 0,0175%-0,0519%). In absolute numbers this is approximately 2290 women on a yearly basis (with a 95% confidence interval of approximately 1225-3630). That's 9% (= 0,033%/0,37% X 100%) of the number of prostitutes who worked during the last year. If you take error margins into account, that's between 3% and 24%. You see, now the uncertainty becomes just too large if you want to know the relative percentage of forced prostitutes at any time compared to all prostitutes this way. But in any way, it doesn't seem that more than - let's say - 25% of the prostitutes are forced to work in prostitution. That's a sort of upper limit. You see how difficult it is to calculate the number of coerced prostitutes at any moment or during the last year.

I forgot to take into account that possibly the samples in these studies mainly involve Dutch nationals. Perhaps foreign prostitutes haven't been interviewed (such as Eastern European prostitutes working in prostitution in the Netherlands, without being Dutch citizens). It is known that most prostitutes in the Netherlands are foreign. In that case all the absolute numbers could be at least twice as large. So perhaps at least 30.000 women work in prostitution in the Netherlands on a yearly basis. And perhaps at least 1500 prostitutes are coerced or persuaded to work in prostitution at each moment. But these studies don't provide reliable data about foreign prostitutes, so I just leave it here.

I notice a paradox when you look at the 2,6% of the women who have worked in prostitution at all times, and 0,37% of women during the last year. If many prostitutes do it for such a short time, there must a very big replacement rate. So this 0,37% could be blown to very large proportions! These percentages suggests that prostitutes work in prostitution for approximately 11 years (= 80 years X 0,37/2,6). But in my opinion this cannot be true. I think prostitutes work on average less than five years in prostitution, taking in mind that most former prostitutes have only been paid for sex a couple of times. On the other hand, the error margins are big of course, and if I take error margins into account I calculate that prostitutes work on average between 5,3 years (= 80 X 0,20/3,04) and 20 years (= 80 X 0,54/2,15) in prostitution. So it could be possible. There is not yet a paradox. It is on the other hand possible that many women don't admit they have worked in prostitution. It is known that many people are hesitant to say they have done something illegal or socially unacceptable, even in an anonymous questionnaire.

I have the impression that many men lie about their visits to prostitutes. Perhaps not 5% of all men have paid for sex during the last year, but more than 10%. And not 25% of all men have visited a prostitute in total, but more than 50%. I will explain. If it really were true that 4% or 5% of the men have visited a prostitute during the last year, then on average these men would visit a prostitute every two weeks. I assume that there are 10.000 prostitutes who have approximately 4,3 clients a day (I base myself on the report about the social position of prostitutes in 2006, see page 27). But they typically work 4,5 days a week, so effectively they have 4,3 X 4,5/7 = 2,76 clients a day. And also assuming that they have 2 months vacation a year I assume they effectively have 2,76 X 10/12 = 2,3 clients a day. Then there would be about 2,3 X 10.000 X 365 = 8,4 million visits of prostitutes a year. Assuming that 5% of 7 million men have visited a prostitute during the last year, that would be 350.000 visitors of prostitutes each year. 8,4 million visits divided by 350.000 visitors would mean approximately 24 visits per year per visitor. I think that is too high. I think 12 times a year is more realistic. I think at least 10% of all men visit prostitutes each year. But taking error margins into account, perhaps there's also no paradox here. Perhaps I estimated the number of clients that prostitutes receive as too high, or I estimated the number of prostitutes as too high. It is known that many prostitutes actually struggle with attracting clients. They are reported by the Dutch prostitutes' pressure group The Red Thread to be lucky if they receive one client a day. Or perhaps the clients are indeed so horny. I don't believe it. I believe many visitors of prostitutes don't do it frequently, perhaps a couple of times a year.

Another point is I want to raise, is the 39% of the total number of (former) prostitutes who were forced mainly related to the prostitutes who have worked only once, or mainly related to the prostitutes who have worked multiple times in prostitution? This is a difficult issue. Another issue is if persuasion necessarily means coercion. In my opinion it is, but I’m sure diehard libertarians won’t agree. There are differences in opinion about what forced prostitution really constitutes, even among prostitutes themselves. It is known that many prostitutes of whom outsiders would say that they are forced, don't consider themselves to be forced.

Now I want to extrapolate to the whole world population, assuming that the situation worldwide is more or less the same as in the Netherlands. There have been estimates made of the number of prostitutes and forced prostitutes worldwide, and often they are not well substantiated. But let's see what my concoctions will show. Obviously, these concoctions will not be very reliable, but at least they will be substantiated! If it is true that 0,37% (between 0,20% and 0,54%) of all the women in the world (except women younger than 15) worked in prostitution during the last year, and I assume that there are about 3 billion women worldwide who are older than 15 (just a rough guess of mine, there are about 7 billion people in the world...), then there are about 11 million female prostitutes in the world who have worked in prostitution during the last year. They say that a number without an error margin is no number, so I would say 11 million plus or minus 5 million (between 6 million and 16 million female prostitutes). I calcuted earlier that perhaps between 7% and 55% of the prostitutes are forced. Then I calculate that between 420.000 and 9.000.000 prostitutes are forced worldwide. Yep, that's when you take error margins into account, it is like saying we know nothing other than that probably there are several hundreds of thousands or millions of women in the world who are being coerced to work in prostitution. If I use my second method (in which I calculated that between 0,0175% and 0,0519% of women are forced prostitutes during the last year), then I arrive at between 525.000 and 1.560.000 forced prostitutes worldwide during the last year. At each moment this is between 315.000 and 1.110.000 forced prostitutes (because I calculated earlier that between 0,0105% and 0,0371% of the women are forced during each moment). So, to be honest, I have no clue.

Now the next report, it is about young people who are younger than 25 (12-24 years old).

SEKS ONDER JE 25e (Hanneke de Graaf, Suzanne Meijer, Jos Poelman, Ine Vanwesenbeeck, 2005), which you can download here:
http://www.seksonderje25e.nl/images/boek_seks_onder_je_25e.pdf

2.382 boys and 2.439 girls have been interviewed. And now some quotes (translated by me from Dutch into English):
On page 108:
In chapter 3 it was already mentioned that 2% of the boys and 1% of the girls have at any time been given money or another reward for sex. Paid sex can be an own choice, but can also happen under pressure or force. In total, four of every thousand boys and eight of every thousand girls have at any time been coerced or persuaded into having paid sex. Of the girls who have at any time been given money or some other reward 57% were persuaded or coerced into doing this, and 14% came up with the idea themselves. Of the boys 30% came up with the idea themselves and for 53% somebody else came up with the idea, but it actually was found to be pleasant. Among a minority of the boys (17%) the paid sex took place under force (not in the table).
On page 175:
How many are faced with paid sex?
Six percent of the boys and almost none of the girls have at any time paid for sex or have given another reward for sex. Two percent of the boys and 1% of the girls have at any time been given something for sex (see table 3.6).
On page 176:
How many girls have experience with loverboys?
Of all girls who at any time have been given money or some other reward for sex, 57% have been persuaded or coerced to do this, possibly by loverboys. In total eight of every thousand girls have at any time been forced or persuaded to have paid sex. Converted to the total group of girls aged between 12 and 25 years in the Netherlands, then it would be almost 10.000 girls. We don't know which segment of these girls have actually been persuaded or coerced by a loverboy.
They say that 10.000 girls have been forced to work in prostitution in the age group 12-25. As I said earlier, I believe that forced prostitutes are forced to work in prostitution for on average two years. Let's divide two years by the number 25 minus 12, and multiply by 10.000. Then I arrive at approximately 1500 forced prostitutes at any moment who are younger than 25. And again, in my opinion this last research again hints that forced prostitution is indeed a problem! But again, the numbers are very small. 0,8% of 2439 girls is ... 20 girls who have been forced or persuaded to have paid sex. There must have been 35 girls who have been paid for sex in total (20 / 35 X 100% = 57%). I calculate using my statistics book's formula an error margin of 17%! So between 40% and 74% are forced or persuaded (with a 95% certainty of course).