How the DWP bends the rules on statistical data

Ministers in the Department of Work and Pensions have devised a new and cunning way of spinning the statistics in order to make political points.
 
Thanks to the watchful eye of the UK Statistics Authority, outright misrepresentation has become too risky a strategy. Statistical releases are carefully scrutinised to ensure that they meet acceptable standards.
 
But it’s quite possible to work your way round this obstacle if the information is issued in a more informal way, to a limited number of journalists, not as a departmental press release, but as what might be called a ministerial press notice.
 
The practice of “trailling” announcements ahead of their formal release is a well-established bad practice. Under the last Government, the Health Department was so adept at it that few of its announcements were ever properly reported by the daily papers. Snippets were reported by the journalists to whom they had been trailed, but when the speech was delivered or the report published, it was often not be reported at all by anybody else because, by then, it was old. How this served the department’s interests always escaped me.
 
It even employed press officers to wait for the first editions to drop and ring newspapers which had been trailed to cross-question the night desk should the item have failed to appear. “Because it was deadly dull” was the robust answer The Times’ night desk issued on my behalf on more than one occasion.
 
But I don’t remember ever being provided with my own, personalised statistical analysis of waiting times or cancer survival. And what was trailed to one paper was available to all the others the next day, to read and check, even if they could not raise the enthusiasm to report it.
 
The DWP now seems to have moved the process on, by issuing selective analyses of statistics that support the Government’s position, without the obligation to release them to the public. The process has been well-documented on the Full Fact website, and the Left Foot Forward blog.
 
One example was a briefing note on Housing Benefit issued on 7 November. This three-page briefing gave what appeared to be plausible numbers on the exploitation of Housing Benefit by landlords, who were accused of putting up rents paid by Housing Benefit faster than rents in the private sector and “taking advantage of a system that’s clearly out of control”, to quote Lord Freud, the Minister for Welfare Reform.
 
You won’t find this press notice on the DWP website, because it was not a departmental press release. It reads just like one, though, including the helpful quotes from the minister. 
    
Iain Duncan-Smith, the Work and Pensions Secretary, subsequently made the same claim in the Commons, attributing the figures to the Office for National Statistics – though the briefing note does not name any source. According to the British Property Federation, the data actually came from the website findaproperty.com.
 
Ian Fletcher of the federation said: “Ministers should not cut fast and loose with dodgy stats. It makes you wonder why taxpayers are paying for all these Government statisticians if the department would rather look up any old statistic off the web than use the professionals they have at their disposal.”
 
He also contested the conclusions drawn. Local authorities pay rents based on market data from the Valuation Office, but there is a time-lag before changes in rental values are reflected in the rents paid. That is why private rents can fall while Local Housing Allowance rents are still rising.
 
If the DWP’s claim had been made in a departmental press release, it could be challenged by the UK Statistics Authority. But it is not clear from the legislation under which the authority operates that statistics smuggled out in this way fall within its purview. Full Fact has made a complaint which the authority is considering.
 
The authority would certainly have leverage if the statistics issued in this way could be shown not to be readily available to the public. The DWP insists that they are. In answer to a question in Parliament asked by Ann Begg, chair of the Work and Pensions Select Committee, into the statistics used in another briefing note on the number of people who have never been employed, Chris Grayling replied that the department used a number of sources, including The DWP national benefits database; Nomis Web: www.nomisweb.co.uk; the Family Resource Survey; and the DWP Information Directorate: Work and Pensions Longitudinal Study.
 
In other words: these data are in the public domain, accessible to all, so we are not in breach of the Code of Conduct on official statistics by using them as we wish.
 
While the practice may not be against the letter of the code, it is certainly against its spirit. The selective leaking of analyses that are hard to check, not universally available, and unaccompanied by any statistical commentary, is clearly not what the code was intended to promote. Nor does it enhance public confidence in official statistics.
 
Can it be stopped? It’s disappointing that the Coalition Government is just as shameless as its predecessor in statistical spin. Close one loophole and they find another. The only remedy available at present is exposure, so congratulations to the two websites I have mentioned, and to The Independent, which has reported the story.
 
Iain Duncan-Smith has today been asked by his opposite number, Douglas Alexander, to explain himself. It will be interesting to see what he has to say.