Skip to main content

Jeff Renaud works on the engine of a Chrysler minivan at Chrysler Group's assembly plant in Windsor, Ont., in this file photo.Jeff Kowalsky/Bloomberg

The question of whether Canada slipped into a recession in the first half of 2015 may depend on our definition of "recession." And most of us have been using the wrong one, says one of the country's top experts on measuring business cycles.

Many commentators pointed at the Bank of Canada's revelation this week that the Canadian economy likely contracted in the second quarter as evidence that the country met the "technical" definition of a recession: two consecutive quarters of declining gross domestic product. Indeed, much was made of Bank of Canada Governor Stephen Poloz's unwillingness to even utter the word "recession" after the central bank's interest-rate cut Wednesday.

But Philip Cross, a former chief economic analyst at Statistics Canada who was the architect of the methodology for tracking recessions – first at the statistical agency and now at the C.D. Howe Institute – says the two-quarters definition is "simplistic nonsense," based on little more than a misunderstanding from a 1967 research paper that noted recessions in the United States typically lasted at least six months.

"People took it as a definition rather than an observation," he said. "I don't think there's anything 'technical' about it."

While there are no precise criteria for officially defining a recession in Canada, most experts agree that the notion of two straight quarters of GDP contraction is merely a starting point for the discussion, not a sole determinant. And the current "mild contraction," as Mr. Poloz preferred to call it, doesn't look like a recession in some key ways.

"If it is a recession, it would be a very unusual one," Mr. Cross said.

In the past, the Bank of Canada described a recession as "a generalized and sustained decrease of economic activity." There are variations of this definition out there, but they all hit on the same points: Recessions are substantial slowdowns, they last for a while and they cut across a broad swath of the economy.

The U.S. National Bureau of Economic Research (NBER), whose Business Cycle Dating Committee is considered the gold standard for determining recessions, relies on a wide range of economic indicators beyond a contraction in GDP. They include gross domestic income, employment, hours worked and personal income, industrial output and manufacturing, and wholesale and retail sales.

Statistics Canada used to track business cycles through its Composite Leading Indicator (CLI), a collection of 10 diverse economic indicators, and by extension was considered the arbiter of Canadian recessions. But it discontinued the CLI in 2012, amid budget cuts at the agency, and around the same time got out of the recession-analysis business.

"There is no universally accepted definition of a recession," a Statscan spokesman said in an e-mail. "Statistics Canada does not and will not make a determination regarding the business cycle and whether or not the economy has entered into a recession."

The C.D. Howe Institute, an independent economic think tank, has worked to fill that void by forming its own Business Cycle Council, a committee of 11 top economists, in 2012 to provide a Canadian version of the NBER's recession analysis.

Mr. Cross helped design the Business Cycle Council and remains a member. In papers both for Statscan in 2011 assessing the 2008-09 Great Recession and for C.D. Howe in 2012 upon the launch of the Business Cycle Council, he said that at a bare minimum a recession requires a sustained decline not just in GDP but also in another critical indicator: employment.

"If the slump in output led firms to reduce employment, it would confirm that the cyclical mechanism of the economy had been engaged," Mr. Cross and C.D. Howe policy analyst Philippe Bergevin wrote in the 2012 paper. "If, however, GDP fell slightly but employment rose steadily, that would point to the restraining of output by irregular factors, not deliberate cuts by firms."

This is the most compelling argument against labelling the 2015 contraction as a recession. In the two quarters in which GDP is estimated to have declined modestly, Statscan's Labour Force Survey shows that employment didn't shrink in concert but rather grew by almost 100,000 jobs. Meanwhile, average hours worked – another key recession indicator identified by Mr. Cross – rose steadily in the first half of the year.

"The country is not in recession in any meaningful or broadly defined way at this point," Bank of Nova Scotia economist Derek Holt said in a recent research report, pointing to the employment data.

Beyond the key aggregate countrywide statistical measures such as GDP and jobs, another critical element distinguishing an economic slump from a recession is its breadth. In a true recession, there is a sustained downturn not just in one or two sectors or regions but across the economy.

As Mr. Poloz pointed out in his post-rate-cut news conference Wednesday, most of Canada's economy isn't in contraction at all. He stressed that while the energy sector (which makes up 10 per cent of the Canadian economy) and to a lesser degree other commodity sectors (another 8 per cent) have suffered sharp declines, the remaining 82 per cent of the economy, the non-resource sectors, are growing at a healthy estimated annual rate of about 2 per cent.

Meanwhile, two key sectors of the economy that are typically highly sensitive to true downturns in the business cycle – housing and manufacturing – have held up well despite the GDP contraction, another indicator that the downturn may be something less than a recession.

Regardless, Mr. Cross cautioned not to get too hung up on the small GDP pullback in the first quarter – just 0.1 per cent on a quarter-over-quarter basis, according to Statscan's estimates – and a similarly small decline that the Bank of Canada expects to see for the second quarter. Allowing room for statistical error, moves that small may mean the economy didn't shrink at all.

"As a statistical agency, [Statscan] can't tell the difference between a contraction of 0.1 per cent and growth of 0.1 per cent," he said. "If people think statistics are that precise, they're naive."