Perhaps it was in the air over at Statistics Canada, but an embarrassing misadventure underestimating jobs growth in the country this summer seems oddly appropriate for an agency that’s missing 20 per cent of its staff and $30 million of its budget.
Still, that was one doozy of a blunder.
Earlier this month, the federal numbers-crunching organization reported that the Canadian economy created a mere 200 positions in July, far fewer than economists had been expecting. Then, just last week, officials issued a statement confirming that the actual number was 42,000:
“An error has been detected in the processing of the August 8 Labour Force Survey release. This error impacts only the July 2014 estimates. The source of the error has been identified.”
What’s more, “Statistics Canada takes this matter very seriously and is immediately launching a review of the data verification processes in place. This does not affect other statistical programs. A report on the results of the review will be published on the Statistics Canada website as soon as it is available.”
It’s not then end of the world, of course. But the mistake caught many economists by surprise, given the agency’s well-earned, international billing for accuracy and verisimilitude.
“I struggle to think of a comparable foul-up anywhere in the world,” wrote Derek Holt, vice-president of Scotiabank Economics, in a memo to the investment community. “The revisions are broad sweeping and affected every major measure in a highly significant manner. Theories regarding how one single factor could be responsible for the revisions went straight out the window as StatsCan pointed to a systems error that affected everything.”
In an interview with the Globe and Mail, he added, “I think they (StatsCan) are still among the elite statistical agencies in the world. There’s no doubting that this was an uncharacteristic but rather large mistake. . .on this particular one that unfortunately blemishes what is otherwise a pretty solid reputation.”
Jim Stanford, an economist with Unifor, the largest private-sector union in Canada, goes further.
“I can’t say whether the funding cutbacks and the siege atmosphere that is evident at Statistics Canada contributed to this particular mistake, but they certainly have contributed to Statistics Canada tarnished reputation,” he told the Globe. “You’ve had a government now for eight years that’s often hostile to what I would call fact-based policy discussion and I do believe that has diminished Statistics Canada’s standing.”
He has a point.
In 2010, when the federal government announced it was scrapping the mandatory long-form census in favour of a “voluntary” household survey, editorials across Canada screeched their opprobrium. The nation’s two top statisticians, Munir Sheikh and Phil Cross, actually resigned their posts at StatsCan in diaphanously concealed protest.
In a news release, Mr. Sheikh wrote that while he could not “reveal and comment on (the) advice” he gave the government “because this information is protected under the law,” he wanted to “take this opportunity to comment on a technical statistical issue which has become the subject of media discussion. This relates to the question of whether a voluntary survey can become a substitute for a mandatory census. . .It can not.”
Indeed, last summer Robert Gerst, a partner in charge of operational excellence and research and statistical methods at Calgary-based Converge Consulting Group Inc., had choice words for the federal government’s evident preference for voodoo science over rigorous research.
“Take the first data releases from the national household survey of Statistics Canada,” he wrote in a commentary for the Waterloo Region Record. “The quality of the results has come under criticism because the voluntary survey replaced the compulsory long-form census questionnaire. In effect, this replaced a random sample with a non-random sample. Non-random samples have their place, but making conclusions about the population isn’t one of them.
Naturally, then, “no conclusions about the Canadian population can be drawn from the national household survey.”
The monthly Labour Force Survey is not the same type of beast. Given the rigorous process on which it depends, it should be virtually immune to errors.
That it is not is a troubling sign that, thanks to limited resources and battered morale, mistakes might well become a statistically meaningful trend at Canada’s numbers-crunching agency.