Five Things I’ll Be Looking for in Friday’s Labour Force Survey
At 8:30am on Friday May 8th, Statistics Canada will be releasing the April edition of the Labour Force Survey. The data will reflect the state of employment from April 12–18, so it’s not a perfect reflection of where we are, but it will be the most timely employment data we have.
Here’s what I’ll be looking for.
1. The number of people employed in Canada, relative to March
Last month the number of Canadians with a job fell by over 1 million. It was by far the largest drop in the publicly available Labour Force Survey data (which goes back to 1976).
So how bad could it get? I asked Brett House, Deputy Chief Economist at Scotiabank. He was kind enough to send me Scotia’s numbers. Turns out, Bay Street is predicting a loss of 4 million jobs. Scotia is predicting a loss of 5.5 million and an unemployment rate of 20%!
Note the U.S. forecast: A drop of 25 million jobs and an unemployment rate of 15%. I don’t know what’s going to happen on Friday. I’m skeptical of the ability of macro models to forecast in an unprecedented period of crisis, but I’m even more skeptical of my own prognostic abilities. So we’ll have to wait and see.
Brett was also kind enough to send to me Scotiabank’s Global Forecast Tables, which gives forecasts to the end of 2021.
2. Changes in hours worked and “employed with no hours”
Job loss is not the only negative employment outcome from this crisis; we’re also seeing people working reduced hours and some still technically employed but working no hours at all. Brendon Bernard, an economist at Indeed was kind enough to send me this chart showing last month’s data breakdown by industry. Not surprisingly, “educational services” experienced a massive drop in hours:
If you’re on Twitter, be sure to follow Brendon and Brett.
My particular interest is in the drop in employment, as these jobs will bounce back in a smaller proportion when we re-open the economy. But we cannot forget about those who are working reduced hours or no hours at all! In many industries, they will make up the majority of the reduction in labour output.
Brendon’s chart is a perfect segue to the next thing I’ll be looking for…
3. Changes in employment by industry
I wrote about this extensively in: This isn’t your older brother’s recession: 2020 vs. 2008. Check out the piece if you haven’t already.
Last month almost all of the job loss was in service industries. This is unusual (though everything about the current situation is unusual), as in a typical recession goods-producing industries, such as construction and manufacturing, see substantial declines in employment. Here’s one visual, comparing last month’s data to the first six months of the Great Recession. The percentages beside industry indicate each industry’s proportion of total job losses:
Here’s another way to visualize the data, which compares the first 3 months of the Great Recession to last month’s employment drop:
Note that during the Great Recession, construction and manufacturing accounted for more than half of the lost jobs; last month they were barely a blip. I expect to see a higher proportion of goods-producing jobs lost this month (it would be difficult to be lower than last month’s 6.5% of the total). That said, expect to see this continue to be a services-driven recession.
4. Changes in employment by sex
Because the affected industries are substantially different than they were during the Great Recession, so too are the affected people. In 2008–09, 80% of the net job loss was experienced by men. Last month, women led the way:
My guess is that this will trend closer to a 50/50 split. Even if that holds true, this recession looks remarkably different than the last one. It will require a new set of stimulus tools to return the economy to working order.
5. Changes in employment by province
Last month central Canada saw the largest employment drops (relative to their previous month’s employment):
I suspect this has more to do with the timing of provincial responses to COVID-19 than it does anything else. I’m particularly curious to see if we’ll see bigger proportional employment drops in oil producing provinces, given the sharp downturns we saw in oil prices in April (all the way to negative prices).
Unfortunately, we still won’t have a great handle on the municipal level impacts, as the data at the CMA-level is averaged over 3 months, due to sample size. So that’s one thing I won’t be looking for.
That’s what I’ll be watching for. Would love to hear what you’ll be watching for.
