Kevin Carmichael: Concerns the Canadian economy is in trouble are based on a set of data blind to much of the wealth being generated in the digital realm

Stephen Poloz, the Bank of Canada governor, has been telling us for a while now that he’s skeptical that statistical methods developed for the old economy are telling us everything we need to know about what’s happening here in the new economy.
When he’s in front of a crowd and the subject comes up, Poloz tends to ask everyone who buys stuff on Amazon to put up their hands. Then he informs the audience that none of those purchases are captured in Statistics Canada’s monthly tally of retail sales because that report only monitors retailers with a physical presence in Canada.
This is important because the case that the Canadian economy is in trouble is based on a set of data that is blind to much of the wealth that’s being generated in the digital realm. For example, Bank of Nova Scotia’s real-time forecast of gross domestic product, which is based on historical patterns of high-frequency data, dipped into negative territory after StatCan reported on Jan. 7 that merchandise exports declined in November.
StatCan knows it has some catching up to do, but the testing required to prove new methodology is accurate takes time. To the agency’s credit, it’s been releasing the results of research efforts that show that its high-frequency releases are missing a material chunk of the real economy. Last year, StatCan published an estimate that put the value of investment in “data, databases, and data science” at $40 billion in 2018, or 12 per cent of all non-residential investment. The value of the stock of such investment was as much as $217 billion, or about 70 per cent of the value of bitumen reserves at the end of 2017.
Oil is still a pillar of Canada’s economy, but it’s now data that drives growth. And yet it’s oil and other tangible goods that get all the press.
So how do you track economic growth when there are no gauges attached to the primary engine? You go back to the basics and rely to a far greater degree on your instincts. For Poloz, that means putting more weight on measures that he can reliably count in real-time, and less on those that only tally the value of stuff that matters less today than it did two decades ago.
“The labour market data are telling us more than the GDP data,” Poloz said after a speech at the Federal Reserve Bank of San Francisco in November. “The labour market, that’s easy right? We can ask firms, how many people are working for you? How much are they making? Those are real. The survey part, the household part, of course it’s a survey, but even so, it has stood the test of time.”
The newest data from the “household part” of StatCan’s monitoring of the labour market were released Jan. 10. The numbers were inconsistent with an economy that’s grinding to a halt, suggesting the Bank of Canada’s leaders needn’t panic over weaker merchandise trade when they gather to reset policy later this month.
Extrapolating from the 60,000 households it contacted last month, StatCan estimates that Canada’s economy added about 35,000 positions in December, compared with an outsized drop of more than 70,000 jobs the previous month. The jobless rate was 5.6 per cent, near the lowest on records that date to the mid-1970s.
You might recall Pierre Poilievre, the Opposition finance critic, using the November hiring numbers to support his contention that we were at risk of a “made-in-Canada” recession. “In November, 71,000 Canadians went home and looked their family members in the eye and said, ‘I lost my job,’” he said at a press conference.
Given the volatile nature of the Labour Force Survey (LFS), anyone without an agenda knew that the weaker number likely signalled a slower pace of hiring, not devastation. StatCan prefers its trend measure of hiring, which increased by 1,800 positions, the fewest since November 2015. That’s kind of what you’d expect from a labour market that’s been performing at a high level for a long period of time.
Canadian employers created 320,300 jobs in 2019, the second most since 2007. The labour participation rate of Canadians aged 25 to 54 who are working or seeking employment is around 87 per cent, near the highest on record. The youth participation rate — one of Poloz’s favourite indicators — is around 65 per cent, essentially the highest in a decade. Employment growth must slow because we are running out of people to put to work.
“Canada’s strong LFS out today suggests recent job losses were merely a blip or statistical noise, not a more worrying long-term trend,” said Julia Polk, an economist at ZipRecruiter Inc., which operates a digital jobs marketplace.
Just to be clear, nothing you’ve read here is meant to make you feel great about the economy.
The latest hiring numbers caused Scotiabank’s nowcast of fourth-quarter GDP to reset to a 0.03-per-cent increase from a 0.03-per-cent decrease. Overall, the economy is weaker than most expected it would be. In October, the Bank of Canada predicted GDP would grow at an annual rate of 1.3 per cent over that period.
Canada is benefiting from having a number of different economic engines, but several of them are sputtering. Employment in Alberta was essentially unchanged from December 2018, while in Ontario, there were some 243,000 new positions, the biggest year-over-year increase for the month of December since 1987, according to StatCan.
If GDP growth continues to fall short of the central bank’s estimates, policy makers could be persuaded to respond. But predicting the path of interest rates in the years ahead won’t be as simple as running the numbers through a sophisticated model. Poloz has made clear that the Bank of Canada will be applying a lot of judgement to take into account the rapid growth of the digital economy.
“We don’t assume it,” he said in San Francisco. “We’ve got to wait to see it. Meantime, you act as if it could be happening.”
Financial Post
Email: kcarmichael@nationalpost.com | Twitter: carmichaelkevin