Fraud, dishonesty, and exaggeration

It’s depressing. By now I should be used to it, but it’s still depressing. The past week has seen another couple of high-profile programmes dedicated to demonising benefit claimants by concentrating on different types of fraud – by the BBC of all people, obviously keen to shed the reputation ‘left-wing bias’ that is thrown at by right-wing politicians (and indeed some ex-journalists).

The first programme was The Future State of Welfare, by one of the flagship BBC radio current affairs presenters, John Humphrys – more of which next week. This week I want to concentrate on the documentary series Panorama‘s expose ‘Britain on the Fiddle’, about the most outright forms of benefit fraud.

So what was news?

This being the UK, it’s hardly news to focus on benefit fraud, which is a regular preoccupation of the right-wing press. What’s new about this is their estimate about the level of fraud in the public sector. The exchange in the programme goes like this (about 10mins in):

“Expert: We estimated that £22bn of public expenditure was being lost to fraud and error.

Interviewer: That’s way more than people imaginged isn’t it? 

Expert: Yes it is.  And that’s because we applied the best available data, so we’re confident in that figure.”

The expert in question here is Jim Gee, ex head of counter-fraud at the NHS, and now working for PKF accountants as well as ‘advising ministers’. The estimates come from work PKF have done with the Center of Counter Fraud Studies at Portsmouth University, primarily this report which is based on this data.

There are two things that are wrong with this figure.

• Firstly, by implication, the programme is saying that this is about benefits fraud (broadly defined, including unemployment/disability, tax credits and housing benefits/social housing). This is why the Daily Mirror – by far the biggest-selling left-wing newspaper – led with stories that ‘benefit fiddles’ were costing the country£22bn. 

This is actually a complete misreading of the data – although an understandable misreading in the context of a programme that focused on this kind of fraud. In fact, £22bn refers to the entire figure for public sector fraud, most of which is in the rest of central government (the NHS, schools, bin-cleaning, civil service offices etc). Only £4bn is claimed to be from the benefits system.

Moreover, no-one as any idea where even this £4bn figure comes from, as the admirable Factcheck organisation have promptly chased up (incidentally, please give money to Factcheck, they’re fantastic). The Department of Work and Pensions (DWP) are described by the Centre for Counter Fraud Study’s report as “measuring their losses accurately since the late 1990s and undertaking a range of work to reduce them over an even longer period” [11]. The DWP should be heartened at receiving such glowing support for their fraud measurement, and their estimate of fraud is only £1.2bn in the latest period – not £4bn.

Judging from Factcheck, it appears that the £4bn is achieved by adding another £500m for tax credits fraud (reasonably enough), plus £2bn for fraud in social housing. Except that the value for housing benefit fraud is for the total value of the houses that are being fraudulently occupied, not the value of renting them at below-marked rates for a year. So even the £4bn is far too high, let alone the £22bn.

• Secondly, and more simply, the estimates for public sector fraud outside the benefit system are a lot more fragile than the programme makes out. The exchange reported above makes it sound like these figures are solid, dependable – ‘we’re confident in that figure’, they say.

Well, actually they have no new data whatsoever on UK public sector fraud.  Instead, they review ‘the latest global data’, which shows 3-9% fraud levels. “If we accept for a moment that the UK is probably no worse or no better than most countries, then we are losing £22.407 billion in respect of public sector expenditure” [10].

This isn’t unreasonable in itself. What’s unreasonable is to pretend that these data are solid and dependable, rather than speculative. (£22.407 billion is a hilariously inappropriate level of precision!).

As an aside, there’s a black comedy in a report about fraud and dishonesty being based on data that entirely lacks transparency – they seem to provide no details whatsoever on where these ‘global surveys’ come from, so we can’t analyse them ourselves. Except that this is the idea, because the figures come from an accountancy firm (and associated university centre) that are clearly touting for business.

The other new claim they come up with is that fraud has increased by 30% since 2007. Factcheck chase this up, and while it comes from this data again, it’s impossible to tell exactly where the data come from, nor how robust it is.

Benefit fraud in context

Benefits fraud violates the spirit of give-and-take underlying the welfare state, and obviously I have no sympathy for most of the people featuring in the programme – the man claiming incapacity benefits while having a house in France and a yacht, or the guy in social housing who goes to the cash-and-carry for the pub he owns in his Bentley.

Yet it’s important to put these figures in context. The latest DWP evidence is that only 0.8% of the total benefits bill is due to fraud – an impressively low figure that most organisations would probably be proud of. A further 0.8% (£1.2bn) is lost in customer error, plus 0.5% (£0.8bn) in official error.

This can be set against (i) 0.8% (£1.3bn) in underpayments in the same year – this ignores people who are entitled to claim and don’t, and is simply the figure due to the DWP screwing up its payments to people; (ii) £15bn in uncollected tax cited (HMRC official estimate) and an estimated £1bn in uncollected local government tax from the Centre for Counter Fraud Studies report; and (iii) legal tax *avoidance* (rather than fraud), where a Guardian editorial noted that “no one has the vaguest clue” how much this is but that a Government stab in the dark was £4-14bn, while the Trades Union Congress puts it at £25bn.

But still…

So behind the headlines here there’s little new, and in fact a glowing endorsement of the DWP’s benefits fraud measurement. Yet there is another claim here – that rather than being outright fraud, there are people who are genuinely sick/unemployed, but are refusing to work in the worst jobs in the economy. It’s this claim that lies at the heart of John Humphrys’ programme, and it’s this that I’ll focus on next week.