Vital information on the health of the economy is “arbitrarily” altered using “dark art” to avoid awkward questions from journalists, a former government statistician has alleged.
Mr S Onyeche, who worked for the Office for National Statistics (ONS) until 2014, has described how a UK public expenditure figure was reduced by £2 billion so that it didn’t look like spending had risen when it was meant to be falling.
His allegations have prompted accusations that economic data produced by ONS – described as “the UK’s largest independent producer of official statistics” – has been “manipulated” for political reasons. The Scottish National Party (SNP) and an economic expert have called for an investigation.
Onyeche made the allegations in evidence to a government review of economic statistics conducted by Sir Charles Bean, a former deputy governor of the Bank of England. Onyeche’s submission, dated 7 August 2015, has been published online.
He highlighted “balancing meetings” that regularly took place between senior ONS officials to review the figures that make up the UK’s gross domestic product (GDP). GDP is used as a key measure of economic activity.
The purpose was to “arbitrarily ‘smooth’ out any peculiarities in the sub-data,” he said. “It really brings home how much of the economic statistical production system is as much a ‘dark artform’ as anything approaching a science.”
Onyeche, who worked in the national accounts office at the ONS headquarters in Newport, South Wales, recalled one meeting on a GDP estimate for three months in 2012. The general government branch of ONS was “told to go back and adjust expenditure for the quarter down by £2bn”, he wrote.
The reason for the adjustment given by a senior ONS official was that he “didn’t want to stand in front of the journos at the press release lock-in on publication day and try to explain why general government expenditure had increased by over two per cent on the previous quarter”.
Onyeche pointed out that this was when the government was meant to be cutting public spending. “General government statistics increasing would run against that narrative and hence would appear contradictory to any journalists scrutinising the statistics,” he said.
“It would be interesting to ascertain to what extent the final GDP figures published are down to senior statistician’s discretion/judgement – given the magnitude of some of these fairly arbitrary adjustments, and whether that impacts on the magnitude of some later revisions.”
The SNP MP George Kerevan, who is a member of the Treasury Select Committee, contended that compiling GDP figures involved a high degree of subjectivity and professional judgement. “That makes it even more important that the ONS is free from any hint of political pressure from the Treasury,” he told The Ferret.
“Sir Charles Bean must investigate any allegations that political pressure – whether direct or indirect – has resulted in data being altered arbitrarily. Above all, future reports should lay out measures to guarantee the future independence of the ONS.”
Kerevan also suggested that the ONS might be too close to government. “As long as the ONS reports to government ministers rather than directly to Parliament (through the Treasury Select Committee) its independence – and the quality of the statistics published – will always be open to question.”
Richard Murphy, a political economy professor at City University in London and a campaigner for financial reform, was similarly critical. “The chance that the data was manipulated to assist the messaging around these decisions is profoundly worrying,” he said.
“The very fact that it has been suggested by someone who should know would seem to justify an inquiry.”
Murphy argued that all statistics were subjective. “Anyone who believes that they’re just a matter of adding up is seriously mistaken,” he said.
“All involve assumptions, estimates and reporting methods that are chosen to suit particular purposes, most of which will be intended to flatter the person delivering the message.”
The whole reason for having an independent body like ONS responsible for statistics was to keep judgements as far removed from government as possible, he added. “If that has failed we need to worry.”
The ONS and the Treasury were defended, however, by Diane Coyle, an economics professor from the University of Manchester. “The annual balancing of the three separate ways of calculating GDP certainly requires judgement calls – ‘dark arts’ is a colourful way of saying this,” she said.
“GDP isn’t a real thing, but a statistical construct, with income, output and expenditure measures calculated separately. It isn’t surprising that any statistician making such a judgement call would avoid doing so in a way bound to attract difficult questions.”
Coyle “very much” doubted whether the treasury put any pressure on the ONS. “The Treasury plays no part in the process and gets figures just a short time before publication,” she said.
According to ONS director-general for economic statistics Jonathan Athow, a large number of data sources were used to produce GDP estimates. “Adjustments have to be made to balance its three measures – output, income and expenditure,” he said.
“We are transparent about these adjustments and now publish them alongside GDP. The Office for National Statistics is independent of Government, and our interpretation of our statistics is based on our own professional judgement.”
A spokeswoman for the Treasury said it was a matter for the ONS and declined to comment further. The Labour Party did not respond to a request to comment.
Onyeche’s submission in full
Photo thanks to Colin and Kim Hansen/ Wikimedia Commons, CC BY-SA 4.0.
On a positive note, good to see the ONS wholeheartedly confronting the need to elevate staff skills levels following the verdict of the Sir Charlie Bean Review
http://www.telegraph.co.uk/business/2016/08/06/newport-braced-for-wave-of-economists-moving-in-to-town/
https://www.theguardian.com/uk-news/2016/aug/11/ons-recruit-data-analytics-apprentices-skills-criticism