OTTAWA — The agency silently embroiled at the centre of the census debate has long been viewed as the best of its kind in the world, but observers worry government intervention could damage the impeccable methodology and autonomy for which Statistics Canada is renowned.
“StatsCan is thought of as one of the best,” says Kevin Milligan, an associate professor of economics at the University of British Columbia, who has worked extensively with the agency’s data. “The institutional culture is one of independence. That is at the heart of why they’ve been so distressed by this turn of events, because they really felt that independence was threatened.”
Over the last month, opposition has mounted to the Conservative government’s plan to turn Canada’s mandatory long-form census into a voluntary survey — a move critics say will produce a skewed or useless national demographic record. The government says it made the change because the long form was an invasion of privacy and it was coercive to force Canadians to complete it.
On Friday, the parliamentary industry committee will meet to discuss hearings on the census issue that will likely take place next week.
“StatsCan has a world-class reputation for its methods, for the reliability of its arithmetic and the credibility of the institution and it would be a huge tragedy if that Canadian model of excellence is sacrificed on the altar of Conservative [politics],” Liberal House leader Ralph Goodale said at a Thursday news conference.
In the early 1990s, a team of statisticians from 10 OECD countries ranked international statistical agencies for The Economist’s “good statistics guide.” Statistics Canada took the top spot, with the magazine lauding the reliability of its figures, its methodology and autonomy, noting that “British and American number-crunchers lack the formal independence” enjoyed by their Canadian counterparts.
The agency participates in international statistical conferences and publishes its own top-level peer-reviewed journal, Survey Methodology, says Don McLeish, president of the Statistical Society of Canada and a professor at the University of Waterloo.
Statistics Canada has been unable to speak about the potential impacts of the changes to the census, and as a senior bureaucrat, Mr. Sheikh was not permitted by law to reveal what advice he gave the minister.
His blunt resignation statement made clear that he did not support the government’s decision and that will help shore up the agency’s reputation, Mr. McLeish says, but the planned changes to the census are still anathema to statisticians who believe in the “sanctity of the data.”
“If I’m a statistician and I develop a methodology for analyzing some data, and someone comes along and for what I believe to be political reasons or any other vested interest says that I should change my methodology, then that strikes at the heart of our discipline,” Mr. McLeish says, emphasizing he speaks for himself and not the society. “If forced to do it, that would be very demoralizing.”
In a statement released Wednesday, after Mr. Sheikh’s resignation, Industry Minister Tony Clement, who oversees the agency, said: “The government made this decision because we do not believe Canadians should be forced, under threat of fines, jail, or both, to disclose extensive private and personal information.”
The mood among agency employees was sombre on Thursday, with many saying they were shocked by Mr. Sheikh’s resignation but thought he made the right decision to step down on principle.
“Obviously, it’s a bit disheartening. It was a surprise for everyone here,” Roberto Casagrande said of the news, which came hours after the cancellation of a town-hall meeting intended to address staff concerns about the census. “It’s unfortunate the way things have transpired. Anytime government gets too involved in specific departments and some of these critical programs, and we can’t resolve them without getting to this point, it is a bit disheartening for staff here.”
“It’s very sad that a man of that quality quit his job,” Alain Despatie said. “It’s not understandable that we have a government that won’t listen to its highest civil servant in a very specific field.”
When the previous chief statistician, Ivan Fellegi, retired in 2008, assistant chief statistician Michael Wolfson — now retired himself — gave a speech that touched on the delicate balancing act of maintaining the agency’s independence.
Mr. Wolfson recalled an instance a decade earlier when he was to present a paper on policy options for Canada’s tax-transfer system at an international meeting and a deputy minister intervened with Mr. Fellegi to stop him.
After determining the paper met “reasonable standards for impartiality and objectivity,” Mr. Fellegi gave the deputy minister his answer: No.
Earlier this week, a Tory senator attacked the veracity of new statistics showing crime is on the decline.
“The data is indeed agnostic, but there are many messages carried by the data to various constituencies. They can be social messages or economic messages and most of those messages probably have a political stripe to them, but you cannot blame that on the data,” Mr. McLeish said.
Ernie Boyko, director of census operations from 1991 to 1996 and a Statistics Canada employee until 2004, says it was “absolutely shocking” to hear Clement say in an interview that some people at the agency “like to think” it’s an independent body, but in fact it reports to him.
“Every relationship we had with a government, this has always been kind of established right at the beginning in a meeting between the chief statistician and the minister, and the minister generally agreed we need to have objective information and the agency should not be seen as being under political influence,” he says of his time at the agency.
Some national statistical agencies operate autonomously, similar to the Bank of Canada, while others function more as an arm of the government, Mr. Milligan says, but the situation in this country is “ambiguous,” depending on how the Statistics Act is interpreted and who are the politicians and bureaucrats involved. One good that could come out of the census controversy and Mr. Sheikh’s resignation would be to clarify that relationship and cement the agency’s ability to do its work “independent from the political pressures of the day,” he says.
“It’s distressing to me that the minister is playing politics with the hard-earned reputation of StatsCan,” Mr. Milligan says.
July 22, 2010
- with files from the Ottawa Citizen