Despite funding boost, Statistics Canada jobs-data upgrade will take time

Share

New funding for better jobs data won't immediately lead to Statistics Canada producing the same kind of fast labour information that is available in the United States and that economists want to see here, the agency says.

Officials at Canada's statistics agency are planning how to spend an additional $14-million a year that the Conservative government announced in June to fund two large new employer surveys.

That announcement followed months of criticism that the government was making policy decisions on everything from training to immigration without reliable job vacancy statistics. Since then, Statscan's two existing labour market surveys have come under closer scrutiny and criticism.

The agency told The Globe and Mail on Wednesday the new money won't have any immediate impact on those existing surveys, but Statscan isn't ruling out changes down the road.

The agency is expected to release a report soon that will explain what led it tto make the unprecedented decision this month of pulling its flagship jobs report, a survey of households called the Labour Force Survey, after the discovery of an error in its July numbers.

Attention on the LFS – which was already facing criticism for its appearance of high volatility – led to debate over whether Canada should follow the U.S. in releasing both a household survey and a survey of employers simultaneously to provide a more complete and timely picture of the labour market.

Statistics Canada currently produces the monthly Labour Force Survey of about 56,000 households. Then two months later, it releases its Survey of Employment, Payrolls and Hours, which is based on a survey of 15,000 businesses as well as payroll data from the Canada Revenue Agency. The payroll survey is considered to be more accurate than the LFS, but receives less attention because it comes out later.

Sylvie Michaud, the agency's director-general of education, labour and income statistics, said the payroll survey's release cannot be moved up because it is a mix of information from a survey of businesses and payroll data from the CRA, which takes time to come in. Ms. Michaud said the U.S. can release its information more quickly because it is based solely on a large survey of about 144,000 businesses and government agencies, representing about 554,000 individual work sites.

The new $14-million from Ottawa will allow Statistics Canada to perform a much larger survey of businesses as 100,000 will be contacted for a new job vacancy survey and a new wage survey that are scheduled to launch next year. But Ms. Michaud said at this time, those surveys will be separate and will not allow the agency to expand and move up its payroll survey.

"Right now, the purpose of the [new] survey is very different from the business payroll survey," she told The Globe and Mail in an interview. "Whether in a few years we could examine something, it's something that we'll have to see as the survey evolves."

Ms. Michaud said Statscan is in regular discussions with the CRA about ways of speeding up the release of payroll data, however no changes are imminent.

Statscan says as more businesses report payroll data online to CRA, that could speed up the process of compiling statistics. However a spokesperson for CRA told The Globe that only 36 per cent of Canada's 1.7 million employers currently submit payroll data electronically.

Benjamin Tal, deputy chief economist for CIBC World Markets, said Ottawa should focus its resources on producing the most accurate and timely labour report possible given the frustration with the current jobs data. Moving up the release of business survey data should be a priority, he said.

"I don't understand why we shouldn't use this opportunity to really increase our labour force information coming from employers, which I think statistically is a more accurate measure of employment," he said. "If an economist has a fantasy, it's a survey of employers. That's what we want: a timely survey of employers like what is done in the U.S. … If they do it in the U.S., we probably can do it here."

Editor's note: an earlier version of this story online and in Thursday's paper, which noted that Statistics Canada's Survey of Employment, Payrolls and Hours surveys 15,000 businesses, incorrectly said a similar survey in the United States surveys 554,000 businesses. In fact, the U.S. Current Employment Statistics program surveys about 144,000 businesses and government agencies, representing about 554,000 individual work sites.