Fact Check: Will the return of the long-form census save money?
Minister of Families, Children and Social Development Jean-Yves Duclos looks on as Innovation, Science and Economic Development Minister Navdeep Bains responds to a question in the Foyer of the House of Commons on Parilament Hill in Ottawa Thursday, November 5, 2015. (THE CANADIAN PRESS/Adrian Wyld)
Jordan Press, The Canadian Press
Published Thursday, November 12, 2015 8:02AM EST
OTTAWA -- "The voluntary process actually cost an additional $22 million. Making it mandatory will actually make it less expensive, (and) it will be on budget and on time." -- Economic Development Minister Navdeep Bains
The Trudeau Liberals, in one of their first acts in government, brought back the long-form census.
Bains proclaimed the decision wouldn't cost taxpayers anything more than what is already budgeted for the 2016 census. Indeed, the minister, who oversees Statistics Canada, said taxpayers are likely to save some cash with the return of the mandatory, long-form census.
Will Statistics Canada save money with the return of the mandatory questionnaire?
Spoiler alert: The Canadian Press Baloney Meter is a dispassionate examination of political statements culminating in a ranking of accuracy on a scale of "no baloney" to "full of baloney" (complete methodology below).
This one receives a finding of "a little baloney" -- it is largely accurate about the mandatory census costing less than the voluntary survey, but there are details missing that would help determine the actual savings figure.
In 2011, Statistics Canada sent out the national household survey to one in every three homes to get as high a response rate as possible. It was a voluntary survey meant to replace the mandatory long-form census of the past. In 2006, the long-form census went to one in five homes, complementing the shorter mandatory census that went everywhere. (According to the Huffington Post, the new long-form census will go to one in four homes in 2016.)
The government added $30 million to the census budget to handle the extra administrative costs that came with sending out a separate questionnaire. Statistics Canada spent $22 million and returned $8 million to the federal treasury. That brought the total cost for the 2011 census to about $652 million.
Some of that extra cost was because the Conservatives decided so close to census day to go with a voluntary questionnaire. "Everything had to be rejigged and so a good deal of the extra cost was due to that," said Ivan Fellegi, who retired as Canada's chief statistician in 2008 after 23 years on the job.
Statistics Canada estimates the total budget for the 2016 census at about $700 million, but that is under the premise the voluntary survey was coming back.
Statistics Canada spreads the costs over a seven-year period that includes time to prepare, collect, analyze and distribute results. The final cost isn't known until two years after census day when Statistics Canada releases the data for public consumption.
The government will save money because the mandatory survey is less costly to administer than the voluntary version, experts say.
"There are at least a couple of significant reasons why there would be savings" Fellegi said. "I mean, not gobs of money, but savings."
The savings start with the number of people hired to help with the census. For the 2016 census, Statistics Canada is hiring 1,400 people for the data centre that processes responses.
Statistics Canada needs all those bodies to handle answers and followup with Canadians who don't fill in the form either because they didn't want to, forgot to, or misplaced it. Doug Norris, who spent nearly 30 years at Statistics Canada, said the agency will need fewer bodies with a mandatory survey going to fewer homes.
"Followup is extremely expensive," Norris said.
Statistics Canada could also see some savings on the back end with statisticians inside the agency having to spend less time supplementing shaky data from the national household survey with information from other data sets such as income tax returns. That quality assurance work is easier with the better response rates that come with a mandatory census.
"The statisticians inside Statistics Canada...spend a huge amount of time and effort doing their best to figure out ways to improve the data quality," said Michael Wolfson, a former assistant chief statistician. "That kind of work will be much less necessary now that the long-form (census) is a mandatory part."
The questionnaire itself won't change much: The questions will be the same as those on the voluntary survey, Norris said. The instructions up front will have to change to remind people the survey is now mandatory.
But Statistics Canada could also incur costs to bring in the mandatory form about seven months before Statistics Canada sends out census forms on May 2, 2016. Right about now, Fellegi said, the agency is ramping up its logistics and hiring plans. Changing plans isn't an easy task; the census is like an ocean liner in that both take time to change course.
The return of the mandatory long-form census is likely to cause some savings over at Statistics Canada. Just how much is tough to say. For that reason, the statement of savings has "a little baloney" -- the statement is mostly accurate, but more information would be helpful to determine how much the agency will save.
There could also be wider economic benefits from the return of the long-form census: Better quality data could reduce costs for municipalities and school boards, for instance, when it comes to planning for future growth, said Kevin Milligan, an economist at the University of British Columbia.
And Charles Beach, a retired economics professor at Queen's University, said that kind of data is worth it even if the 2016 census ends up costing more than the 2011 version.
The Baloney Meter is a project of The Canadian Press that examines the level of accuracy in statements made by politicians. Each claim is researched and assigned a rating based on the following scale:
No baloney -- the statement is completely accurate
A little baloney -- the statement is mostly accurate but more information is required
Some baloney -- the statement is partly accurate but important details are missing
A lot of baloney -- the statement is mostly inaccurate but contains elements of truth
Full of baloney -- the statement is completely inaccurate