Gender differences amongst sex workers online

Two weeks ago the UK’s Office of National Statistics (ONS) announced that the UK economy had grown by 0.9% in the second quarter of 2014. This growth was partly due to the decision to include prostitution in gross domestic product (GDP) for the first time ever.

The ONS estimate that Britain’s 60,879 prostitutes contribute £5.314bn (0.4% GDP) to the UK economy.  But, this figure completely ignores male prostitution.  In a previous post we used data from AdultWork, a popular sex worker website, to determine that 42% of UK prostitutes are male, which (all other factors being equal) represents an additional £3.542bn of UK GDP!

Our initial analysis was very simple, but nonetheless raised a lot of interest from you all, along with some great questions about the gender differences between male and female sex workers.  We decided to look a little deeper into the gender differences amongst sex workers online and this is what we found.

Sex workers on AdultWork are heterosexual

sexual orientation

Sex workers on AdultWork are overwhelmingly heterosexual.  Only 1.3% of profiles on AdultWork specify sexual orientation as “gay”.  Homosexual sex workers both male and female are therefore missing from our previous revised estimates.  In order to get a good measure of the number of homosexual sex workers in the UK we would have to look at data from additional web sites that focus specifically on these communities.  It is interesting to note that women on AdultWork are much more likely to advertise themselves as bi-curious or bi-sexual than men.  Indeed bi-sexual is the most popular advertised sexual orientation for female sex workers.

Female sex workers are younger

age

Female sex workers are, on average 3 years younger than male sex workers.  The average age for female sex workers is 30, while the average age for male sex workers is 33.  The biggest age group across both genders is 24-year-old females.  As sex workers get older, it appears that women stop working at a younger age than men.

Female sex workers stop working sooner

account closure rate

Female sex workers close their accounts at a greater rate than male sex workers.  Over a period of 90 days, double the number of women will close their profiles compared to men.

years since joining

A female sex worker’s profile is most likely to be less than a year old, whereas a male sex worker’s profile is most likely to be between one and two years old.

Female sex workers are more active

days since last login

Women join AdultWork at a greater rate.  This is reflected in the fact that there are more female profiles overall despite the increased rate of female account closure.  In addition, while they are working, women are much more active than men.  Women are twice as likely as men to have logged in to AdultWork in the last day.

There is more demand for female sex workers

profile views

The total number of profile views for female sex workers is 38 times greater than the total number of profile views for male sex workers.

profile reviews

In addition, the total number of reviews for female sex workers is much greater than that of male sex workers.  But the difference in reviews is not as large as the difference in profile views.  The total number of reviews for female sex workers is only 5 times greater than the total number of reviews for male sex workers.

Another way of putting this result is to say that male sex workers get many more reviews than would be expected based upon the total number of profile views for male sex workers.

Female sex workers charge more

hourly rates

On average, female sex workers charge £120 for incall work and £150 for outcall work.  This is much greater than male sex workers who charge around £70 regardless of whether the work is incall or outcall.

Looking at these pricing distributions, it is very clear that they are not normally distributed.  They do not fit a bell curve.  But if we lay a normal distribution over the top of each pricing distribution, then an interesting pattern emerges,

hourly-rates-bell-curve

The pricing distributions are sort of normally distributed apart from spikes at certain price points.  It is most easy to see this in the distribution of female pricing, which is is evenly distributed around the average, apart from massive spikes at £100 and £120 with another smaller spike at £150.  These spikes are also present in the distribution of male pricing: at £80, £100, £120 and £150 although the effect is less strong.

hourly-rates-bell-curve-spikes

Sex workers appear to be deploying a psychological pricing strategy in order to maximise their revenues.  The classic example of psychological pricing is fractional pricing whereby the price of a product is set at £1.99 rather than £2.00 because £1.99 seems cheaper to a consumer, which drives greater demand.  What appears to be going on with sex workers is that they are taking advantage of consumers’ price indifference between, for example, £110 and £120 and rounding up services that might be charged at £110 to £120.

So what is the impact on GDP?

The ONS estimate for the contribution that prostitution makes to GDP is calculated as follows,

ons gdp estimates

Let’s look at each of these numbers in turn to see how our gender difference analysis can further improve our revised estimate for prostitution’s GDP contribution.

We know that the ONS figures are only for female sex workers.  In order to get an improved estimate of the contribution that prostitution makes to GDP we are going to have to add male sex workers into the calculations.

As for the number of clients per prostitute per week, 25 seems very high.  The source that the ONS uses for this number is a study conducted in the Netherlands, where prostitution is both legal and regulated.  Prostitution in the UK is not regulated and although legal, many of the activities associated with it are illegal.  The market for sex work in the UK is therefore likely to be very different to the market for sex work in the Netherlands.  However, given that this is the only number that we have got, we will have to stick with it.  But we do know from our analysis of AdultWork profiles that female sex workers are twice as active as male sex workers, so it seems reasonable to half the number of clients per week for male sex workers.

The average price per visit used by the ONS is much less than the average hourly rates for female sex workers advertised on AdultWork.  The ONS do not provide exact details about how they calculated average price per visit simply saying that it is “an estimate from Punternet supported by a 2004 journal article” – the journal article is not cited.

Punternet is a website where male sex work consumers write reviews of female sex work providers.  We decided to try and validate the average price per visit from Punternet ourselves.  We created an Import.io API for Punternet and extracted the last 12 months of review data.  We found that the average price per visit over the last 12 months was £138.  The average duration of a visit was 63 minutes which gives an average hourly rate of £131.  This is consistent with the average advertised incall and outcall rates on AdultWork.  It is closer to the incall rate (£120) than the outcall rate (£150), which makes sense as 95% of all reviews on Punternet are for incall work.

Taking these numbers and running them through the ONS calculations gives a total of £12.374bn sales of prostitution services.  Our estimated price per visit for female sex workers is double that of the ONS which in turn doubles the total sale of female prostitution services.  Our figure for the total sales of male prostitution services is down from our previous estimate due to the reduced rates that male sex workers charge and the reduced activity that they appear to display in the market.

import io gdp estimates

How we got the data

In order to do the initial count of sex workers we built an Import.io connector to the search function on AdultWork (WARNING linked materials contain sexual content) and used it to perform a search for sex workers in the UK who advertise “escorting” on their profile pages.  The connector paged through all the results (37,717) and for each result pulled down the profile summary information including gender, age, sexual orientation and number of reviews.  On AdultWork gender is listed as: Female, Male, Couple MF, Couple FF, Couple MM or Unspecified.  From this we were able to count the number of female (22,622) and male (16,326) sex workers on AdultWork.  We ignored profiles where the gender was Unspecified (355) and counted twice for Couple profiles (1,586).  42% of the sex worker profiles on AdultWork are male.

In order to look at gender differences in greater depth we needed more detailed information than is available on the search result pages.  AdultWork profile pages are not directly accessible via a URL.  All AdultWork profile pages sit behind a landing page that warns visitors that the site contains material of an adult nature and asks visitors to confirm the territory that they are from.  A simple extractor would not get past this landing page.  Instead we built an Import.io connector that takes a URL as an input (WARNING linked materials contain sexual content).  The connector first navigates to the homepage of AdultWork, clicks through the landing page and then navigates to the specified profile page URL.  In this way we were able to get profile views, joining date, last login date and hourly rates.  We performed the gender difference analysis on a random sample of 1,000 female and 1,000 male profiles from the entire population of AdultWork profiles advertising escorting services in the UK.

You can learn more about connectors that take URLs as inputs here.

In order to calculate the average price per visit on Punternet we created an Import.io extractor to the Punternet review pages (WARNING linked materials contain sexual content).  The URLs for Punternet reviews are sequentially ordered.  We found the most recent published review and a review from 12 months ago, generated the URLs for all the reviews in between and fed those URLs into the extractor to access 12 months of review data.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s