Lies, Damn Lies, and Department for Work and Pensions Statistics

So much of what Iain Duncan Smith and his department is doing is based on empirical sand.

 

Things have come to a pretty pass when the Economist - not a publication renowned for its strident radicalism - is moved to describe the statistical output of a Government department with a simile that contains the words “raw sewerage”. But then hell hath no fury like a centrist (in this case, Daniel Knowles) frustrated by the misuse of statistical evidence.

The claim which sparked his reaction came from the Department for Work and Pensions: “Around one million people have been stuck on a working-age benefit for at least three out of the past four years.”

Knowles was, of course, too polite to say so in such terms, but this is all but a barefaced lie: “[the] one million includes single mothers who have children too young to go to school, people who are seriously ill but may eventually get better, and people who may be ill, but have yet to be tested.” His piece goes on to question the figures in a number of other DWP press releases, including the claim that the benefits cap has encouraged 8,000 people to get jobs, and the 878,000 people who apparently dropped their claims for disability benefits when faced with a test.

And at risk of reproducing all of the post, his argument as to why we should take issue with these dubious statistics was unimprovable: "the whole point about government statistics is that they are meant to be at least sort of objective. Ministers can quote the ones which support their case — but they shouldn’t manipulate them and distort them to tell stories that aren’t actually true. There is plenty of evidence to support welfare reform without resorting to such disgraceful abuse of numbers.”

Look, like Knowles, I’m broad-minded. I can live with the idea of benefit reform. Maybe I can even live with the really ugly stuff - the misery, the bailiffs champing at the bit to cash in - actually I can’t, but anyway: the real problem is that I think there’s even more to this issue than he highlighted.

For a start, as this blog shows, these were actually two of three misleading claims in the space of four weeks: in a Daily Mail article on Personal Independence Payments, we hear: “The decision to introduce new tests has produced an extraordinary ‘closing-down sale’ effect, with rocketing claims as people rush to get their hands on unchecked ‘welfare for life’ [by claiming Disability Living Allowance] before [Esther] McVey’s axe falls on April 8.” True? Well, you can find out more about the veracity of this claim here. Yet despite the lack of credibility, Iain Duncan Smith repeated it on 8 April, in all the major newspapers.

Perhaps we should start at the beginning - about five or six years ago. Back when the Tories were in opposition I became - and I type these words while pummelling myself viciously about the face with a rolled up copy of the Spectator - a believer in Iain Duncan Smith. I read about the research he was doing in Easterhouse in Glasgow with some approval. Here’s a 2002 report on a visit which crept somewhat under the radar. Duncan Smith’s rhetoric chimed with many of the things I was discovering at the time about crime and poverty - and indeed his think tank, the Centre for Social Justice (CSJ), would later produce a report on gangs which contained barely a word with which I could find fault.

The trouble was that as the years went by, the more it seemed like the by-now-minister’s beliefs were based on nothing more than my own: gut instinct and a few statistics that appeared to back them up. Now journalists, by the nature of their trade, generally have little more than this with which to work. But Government departments? Policy-makers? Surely they need a little more. And that’s why I worry about the DWP. Because it strikes me that so much of what it’s doing is based on empirical sand.

When the left complain about the cruelty of Duncan Smith’s various reforms, his usual reaction to get particularly angry about how unjust the “benefits trap” is for those families trapped in it in the long term: a symptom of the left’s own peculiar brand of cruelty; one that echoes down the generations. Common sense dictates there’s some truth to the claim. But the question is, how many families are are in this situation? Given that one of us is a government minister enacting reforms with evangelical zeal and the other a semi-sober hack at a keyboard shrugging his shoulders, I find it concerning to think he’d know no more than I do.

Consider, then, the CSJ’s response to a report by a number of churches entitled “The Lies We Tell Ourselves: Ending Comfortable Myths About Poverty”. The idea is to fisk these ecclesiastical do-gooders right up their cassocks and prove that these myths are, in fact, not myths. So. “On intergenerational worklessness, the [Church] report argues that there is ‘no credible evidence that such families actually exist’. The report is right to highlight that the Government does not collect data on this, but this does not mean that the problem itself does not exist. The CSJ frequently speaks with poverty-fighting charities who comment that they come across it regularly.”

Well, that’s put the lie to bed hasn’t it? I mean, might it not be worth mentioning one of the few studies of intergenerational worklessness - this one - and note the researchers’ findings that “even two generations of complete worklessness in the same family was a very rare phenomenon, which is consistent with recent quantitative surveys of this issue”? Was there no space to mention Bristol University’s study of Labour Force Survey figures which found only only 0.3 per cent of UK households have two generations that have never worked? I guess if you “frequently speak to charities” you don’t need to draw on such things.

And so it goes on - the churches rightly argue that the level of benefit claimants on the fiddle is lower than people think - the CSJ responds: “The report argues that the level of benefit fraud is low. Yet it fails to mention that the level of error is very high... However, rather than blame claimants for the levels of fraud and error, the CSJ places much of the blame with our complex benefits system.”

To clarify, since the CSJ doesn’t:: 0.7 per cent, or £1.2bn, of total benefit expenditure is overpaid due to fraud - 0.9 per cent, or £1.4bn, of total benefit expenditure is overpaid due to claimant error, and this is offset by £1.3bn underpaid due to error. So when the CSJ doesn’t blame claimants, that’s because it’s bloody right not to.

What it should say is: “Agreed: the overwhelming majority people on benefits aren’t on the fiddle, according to the statistics.” But that wouldn’t fit with the pledges to go after the “bogus disabled”, would it? Instead the best we can manage is this patronising, come one come all rhetoric farted out on a weekly basis, the weasel words that imply that this could only ever be about helping people to help themselves.

I’m not going to go through every point of this response. Needless to say, there’s a huge elephant in the corner, one unacknowledged by either of our main political parties, nor the yammering baked-wind merchants who belch this stuff out on their behalf. They have singularly failed to find an answer to the problem of low wages. It’s impossible to make this point any clearer: six out of ten benefits claimants are in work. You want to get angry? Forget the scrounging folk demons - get angry at our supine political class.

And speaking of the jobs market, there’s a whole other dimension to the DWP's statistical horseplay worthy of contemplation. The most recent Private Eye used a FOI request to reveal that ministers attempted to massage the results of the Work Programme, by pushing for “simple” figures to be publicised instead of the official statistics. Mark Hoban apparently told Kirsty McHugh, head of the Employment Related Services Association, that he was keen on the ERSA’s figure of “200,000 Job entries”, which would be “more understandable to the media/public than discussion around Job Outcomes.” As the magazine reported: “Baldly stating that 200,000 people started jobs means little without counting the far larger official number who did not start or hold on to a job.”

In the end it didn’t help - the DWP’s press release avoided mention of minimum performance failure, but it was still the dominant headline. How indicative a little tale this is. To some it might suggest a department floundering, frantically spinning, its beliefs and its practises founded on the shakiest of statistical foundations. If this is the soil upon which our political right wants to plant its flag then fine. I wish it luck. But It shouldn’t be surprised if it suddenly sinks out of sight.

David Cameron and Iain Duncan Smith visiting A4e in 2011. Photograph: Getty Images

Alan White's work has appeared in the Observer, Times, Private Eye, The National and the TLS. As John Heale, he is the author of One Blood: Inside Britain's Gang Culture.

Photo: Getty
Show Hide image

The New Statesman 2016 local and devolved elections liveblog

Results and analysis from elections across the United Kingdom. 

Welcome to the New Statesman's elections liveblog. Results will be coming in from the devolved legislatures in Scotland and Wales, local elections in England, and the mayoral contests in London, Salford, Bristol and Liverpool. Hit refresh for updates!

00:34: For an alternative view on Zac Goldsmith, Andrew Boff, a Conservative member of the London Assembly, said this earlier today on Newsnight:

"I don't think it was dog whistle because you can't hear a dog whistle. Everyone could hear this"

00:31: Duncan Smith droning on about how Zac Goldsmith's campaign is not at all racist, oh no. I'm not getting paid enough for this. 

00:29: Iain Duncan Smith has appeared on screen. He says he is "hopeful" that Zac Goldsmith will be elected tonight. In Wales, the Conservatives have walked out of the count in marginal Delyn. Labour are sounding fairly pleased about that, as you'd expect.

00:27: I have made two discoveries. The firsts is that the lights in the New Statesman offices are motion-sensitive. The second is that sitting and typing is not quite enough motion. (It's just me here tonight.)

00:26: Council seats so far: Labour have 59, the Liberal Democrats have four, Ukip have none, the Greens have none. The SNP are hopeful of picking up all the Scottish Parliament seats in Motherwell and Glasgow, but Edinburgh is trickier territory. 

00:25: Speaking of bets...I look likely to owe Wings Over Scotland £20 (I bet on a clean sweep for the SNP in the constituencies), as Labour are buoyant about Edinburgh Southern and the Liberal Democrats are hopeful in Edinburgh Western.

00:19: John McDonnell doing a good job putting a brave face on some grim early numbers for Labour. This line about needing only do better than a general election is nonsense, psephologically speaking but he's making it sound like good sense. A validation of Jeremy Corbyn's decison to ignore even some of his closest allies and put him in as shadow chancellor. And still only 9 to 1 on Betfair as Labour's next leader. 

00:10: People on the BBC and keep talking about 2012 as a "high point for Labour". Is this true? Well, sort of. It was Ed Miliband's best year. However, that doesn't mean that Labour doesn't still have room to gain seats tonight - governments tend to lose seats in opposition and Labour lost seats pretty consistently in the areas up for election tonight throughout their 13-year-stay in government. So they still can and should make gains. And bear in mind, even Ed's good years were padded out with gains in safe Labour seats, which went from Labour strongholds with say, 40 Labour councillors and 20 Liberal Democrats to 58 Labour councilors and three Greens. In the places Labour needs to win at Westminster to get back into government, there is real room for growth. Which is why I wouldn't worry overmuch about losing some* seats in safe seats if when the marginals report Labour is making headway there. 

*Some is key. Going from a majority of 10,000 to 5,000 in Labour heartlands is fine if Corbyn is putting on 5,000 votes in seats Labour lost by that kind of margin. Going from a majority of 10,000 to -1,000 in Labour heartlands, less so. 

00:06: Labour look likely to lose Crawley

00:02: Labour have kept control of Newcastle Council, taking a seat from the Liberal Democrats. (I knew that would happen the second I typed the words "Liberal Democrat revival"). 

00:00: For those of you just joining us: welcome. Labour is projected to lose seats but remain the largest party in Wales, where the Conservatives seem to be gaining ground. In England, the Liberal Democrat revival appears to be a thing and not just a Twitter meme. In Scotland, the SNP are sounding buoyant while the Conservatives believe they may beat Labour into third. London won't count until tomorrow but everyone - Labour, Tory, Cannabis is Safer Than Alcohol - is getting jittery over low turnout. 

23:55: That early worry I heard from Wales has vanished completely from the Tory side. Vale of Glamorgan is rumoured to be close - a close to six point swing to the Conservatives. So we have biggish swings away from Labour so far tonight. 

23:52: Labour are down 17 per cent in the six seats we've had so far (from 2012 when last contested). Still not very much data, but that would put the party in the mid to low 20s in terms of nationwide share. Personally I think it's unlikely to be that bad when all the results have rolled in. 

23:48: How about that Liberal Democrat fightback, huh? The Liberal Democrats have won a seat in Sunderland from Labour. 

23:47: The knives are already out for Kezia Dugdale in Scotland, where Labour may come third. 

23:42: Bad news for Labour from Wales. Clywd South is in play and the Tories may well win it. Cardiff North, which is Conservative-held at Westminster, looks likely to go the same way in the Assembly having been Labour-held since 2011. Newport West and Llanelli are worth looking out for too. 

23:39: Good news for Labour - they've held the first seat to declare out of Newcastle, and the Liberal Democrats, their main opposition, have privately conceded that Labour will remain large and in charge in Newcastle. 

23:35: Speaking of the Liberal Democrats, they are feeling cautiously optimistic about winning a seat in Edinburgh Western from the SNP, while they expect to recover a bit from 2015. (Things could hardly get worse, I suppose.)

23:32: The first Labour gain of the night, as a Liberal Democrat councilor in Stockport defects. 

23:30: Labour sources are gloomy about their chances of holding onto Exeter Council, where Ben Bradshaw is the party's only remaining MP in the South West. Looks like it will slip into no overall control. Party is also nervous about holding Derby. 

23:25: Tory mole in Wales tells me that things look bad for them - potentially worse than the losses shown in YouGov's poll. The election has become "a referendum on steel", apparently. 

23:20: Early results from Sunderland show Labour doing fairly badly (you know, for Sunderland) and Ukip doing very well. But one swallow doesn't make a summer and we need more data before we know anything. 

23:15: We should get our first result from Scotland in 45 minutes or so. Rutherglen, Labour-held since the Scottish Parliament's creation in 1999, and highly likely to go to the SNP. 

23:13: And what the results mean so far, according to ace numbercruncher Matt Singh:

23:07: Those numbers from Sunderland, where Labour have held in St Anne's ward. Labour down 15 points on 2012, when these seats were last fought, Tories down 3. It's Ukip who are making the headway (they didn't stand last time and expect them do post performances like this throughout the United Kingdom tonight and as results roll in over the weekend). 

23:04: Back to Wales - YouGov's poll "looks about right" according to my Plaid Cymru source. What does that mean? Labour could go it alone and do deals on a vote-by-vote basis - they govern alone now with just 30 seats. If the poll is even a little out - let's say either Labour or the Liberal Democrats get one more seat - they might do a deal if they can get a majority with the Welsh Liberal Democrats. 

23:01: Pallion Ward in Sunderland is the first to declare, and it's a Labour hold! More on percentages as I get them. 

22:58: Why isn't it an exit poll, I hear you ask? Well, an exit poll measures swing - not vote share, but the change from one election to the next. People are asked how they've voted as they leave polling stations. This is then projected to form a national picture. Tonight's two polls are just regular polls taken on the day of the election. 

22:57: The Sun's poll - again, not an exit poll, I'm not kidding around here - of Scotland has the SNP winning by a landslide. (I know, I'm as shocked as all of you) But more importantly, it shows the Conservatives beating Labour into second place. The Tories believe they may hold onto Ettrick as well. 

22:55: What news from Scotland? Labour looks to have been wiped out in Glasgow. Liberal Democrats think they might hold at least one of Orkney or Shetland, while the seats in Edinburgh are anyone's game. 

22:52: Hearing that turnout is low in Waltham Forest, Lewisham, Hackney and my birthplace of Tower Hamlets (the borough's best export unless you count Dizzie Rascal, Tinchy Stryder or Harry Redknapp, that's me). Bad news for Labour unless turnout is similarly low in the Tory-friendly outer boroughs. 

22:47: YouGov have done a poll (note: not an exit poll, it should not be taken as seriously as an exit poll and if you call it an exit poll I swear to god I will find you and kill you) of the Welsh Assembly. Scores on the door:

Labour 27

Plaid Cymru 12 

Conservatives 11

Ukip 8

Liberal Democrat 2

There are 60 seats in the Assembly, so you need 30 seats for a majority of one. 

22:40: In case you're wondering, how would closing a seven point deficit to say, six, compare to previous Labour oppositions, I've done some number-crunching. In 1984, Neil Kinnock's Labour turned a Tory lead of 15 per cent at the general election to a Conservative lead of just one per cent. In 1988, one of 12 per cent went down to one per cent. (He did, of course, go on to lose in both the 1987 and 1992 elections). In 1993, John Smith's Labour party turned a deficit of eight points at the general to a Labour lead of eight points in the local elections. William Hague turned a Labour lead of 13 points to one of just six in 1998, while Iain Duncan Smith got a Tory lead of just one point - from a Labour lead of nine. In 2006, new Tory leader David Cameron turned a 3 point Labour lead to a 13 point Tory one. Ed Miliband - remember him? - got from a Tory lead of seven points to a two point Labour one. 

22:35: John McDonnell is setting out what would be a good night as far as the party leadership is concerned - any improvement on the 2015 defeat, when the party trailed by close to seven points. Corbyn's critics say he needs to make around 400 gains.

I've written about what would be good at length before, but here's an extract:

"Instead of worrying overmuch about numbers, worry about places. Although winning seats and taking control of councils is not a guarantee of winning control of the parliamentary seat – look at Harlow, Nuneaton, and Ipswich, all of which have Labour representation at a local level but send a Conservative MP to Westminster – good performances, both in terms of increasing votes and seats, are a positive sign. So look at how Labour does in its own marginals and in places that are Conservative at a Westminster level, rather than worrying about an exact figure either way."

22:31: Oh god, the BBC's election night music is starting. Getting trauma flashbacks to the general election. 

22:22: A few of you have been in touch about our exit poll. Most of you have been wondering about that one vote for George Galloway but the rest are wondering what happens - under the rules of the London mayoral race (and indeed the contests in Salford, Bristol and Liverpool), 2 votes would not be enough for Sadiq. (He needs 2.5). However, all the other candidates are tied - which makes it through to the second round. What happens then is the second preferences are used as a tie-break. Of the tied candidates, Sian Berry has the most second preferences so she goes through to face Sadiq Khan in the final round. Final round is as follows:

Sadiq Khan: 3

Sian Berry: 2

3 votes is above the quota so he is duly elected. An early omen? 

22:19: Burnham latest. A spokesperson for Andy Burnham says:

"Approaches have been made to Andy Burnham to give consideration to this role. It is early days and no decision as been taken. Whatever the decision, he will continue to serve the leader of the party and stay in the shadow cabinet."

22:17: Anyway, exit poll of the office. We've got:

Sadiq Khan: 2

George Galloway: 1

Caroline Pidgeon: 1

Sian Berry: 1

22:15: Update on Andy Burnham. He has been asked to consider running. More as we get it. 

22:13: People are asking if there's an exit poll tonight. Afraid not (you can't really do an exit poll in elections without national swing). But there is a YouGov poll from Wales and I am conducting an exit poll of the four remaining members of staff in the NS building. 

22:11: It's true! Andy Burnham is considering running for Greater Manchester mayor. Right, that's it, I'm quitting the liveblog. Nothing I say tonight can top that. 

22:09: Rumours that professional Scouser Andy Burnham is considering a bid for Greater Manchester mayor according to Sky News. Not sure if this is a) a typo for Merseyside or b) a rumour or c) honestly I don't know. More as I find out. 

22:06: Conservatives are feeling good about Trafford, one of the few councils they run in the North West.

22:03: Polls have closed. Turnout looks to be low in London. What that means is anyone's guess to be honest. There isn't really a particular benefit to Labour if turnout is high although that is a well-worn myth. In the capital in particular, turnout isn't quite as simple a zero-sum game as all that. Labour are buoyant, but so are the Tories. In Scotland, well, the only questions are whether or not the SNP will win every single first past the post seat or just the overwhelming majority. Both Labour and Tory sources are downplaying their chances of prevailing in the battle for second place at Holyrood, so make of that what you will. And in Wales, Labour look certain to lose seats but remain in power in some kind of coalition deal. 

22:00: Good evening. I'm your host, Stephen Bush, and I'll be with you throughout the night as results come in from throughout the country. The TV screens are on, I've just eaten, and now it's time to get cracking. 

Stephen Bush is special correspondent at the New Statesman. He usually writes about politics.