No, Big Data firm, the UK isn't teeming with UBER-FRISKY GIGOLOS

Prostitution + official GDP figures = buckets of FAIL

Big Data versus prostitutes: Hookers 1, number-crunchers 0

The relationship of GDP to national debt involves the ability of the government to tax economic activity to pay the debt burden. So maybe we don't want to include grey and black markets (respectively, legal in themselves but not paying tax, and illegal and also not paying tax) in GDP.

Prostitution: Legit economic activity at work?

We also have the issue of quite how accurate these estimations of these under-reported activities are. Which brings us to big data upstart import.io, which strongly disputes that the Office of National Statistics has got things right here.

ONS has estimated the number of prostitutes, assumed £50 for each instance of a happy ending and multiplied that by 25 smiling clients a week to reach their total. However, they are only measuring the output, incomes and labour of female prostitutes. Import.io attempt to add in the number of male prostitutes to give an even fuller picture. And while their method seems, on the face of it, to be entirely valid, their conclusion is so at odds with observable reality that there's got to be something cock-eyed about it.

Using their (umm, big data-ey?) tool to scrape an escort site, they find that 42 per cent of those offering services are men. So, they take the ONS number for women, multiply it by 42 per cent, add that sum to the ONS one and we get almost £9bn for the contribution of prostitution to the economy.

And we also get a vast problem in believability. Think it through for a moment: male prostitution to heterosexual women does exist: but it's a vanishingly small part of the sex trade as a whole. The gigolo is more a fantasy coming from low-grade literature than an outstanding feature of the economy. So this male prostitution isn't of that form: it's of the male-to-male form. But there again we have a believability problem.

Gay men make up, depending on whom you want to believe, between one and three per cent of the population, while self-reported bisexuals (obviously slightly biased as it is self-reporting) comprise another one to three per cent. And there'd have to be a really very large difference indeed between heterosexual male use of paid sexual services and gay male use for our numbers to now add up.

That there's rather more gay sex going on isn't much of a surprise: adding male sexuality to male sexuality is going to lead to rather more male-style sexual activity going on. But as it happens (I have a colleague who has studied exactly this point: it's amazing what academia will try to understand, sometimes) we're reasonably sure that gay demand for paid sex is only about 10 per cent higher than heterosexual. In the sense that if 10 per cent of men (usually what it's pegged at) are users or hirers of female companions, then 11 per cent or so of gay men are.