Canada’s Economy and Job Market Are On Fire

• Canada grew 75k jobs and the unemployment rate dipped

• Hours worked point to sustainably strong GDP growth in Q3

• Wage growth eased after a torrid prior gain

• Ontario and BC led the way

• Maybe start asking the sickies for notes from the doctor

• Market reactions were restrained by concerns about the US job market

• Canadian jobs m/m 000s / UR %, July, SA:
• Actual: 75.1 / 6.4
• Scotia: 50 / 6.4
• Consensus: 20 / 6.5
• Prior: 18.2 / 6.5

Canada’s job market met Scotiabank Economics’ expectation for a strong outcome. Jobs increased by 75k with Scotia’s estimate of 50k at the top end of consensus, and the unemployment rate dipped to 6.4%, also matching Scotia’s off-consensus call. The details were broadly positive. It’s another notch in favour of our BoC year-end hike call.

The market reaction was relatively muted. As US 2-year Treasury yields rallied following the disappointing US payroll numbers, Canada’s 2-year yield moved up by about 3bps post-data and narrowed the 2s spread to the US. The Canadian dollar appreciated to the US but is a middle of the pack performer relative to other currency pairs this morning. OIS pricing for Bank of Canada policy rate changes was unaffected with a steady 17bps of a quarter-point hike priced for December.

Why little market reaction? One possibility is that if the US labour market is stumbling —and by extension, the US economy— then Canada may be negatively impacted through the back door in isolation of domestic drivers of Canada’s economy and job market. Another possibility is that markets dismissed the big gain as unreliable statistical noise but I’ll come back to why that’s wrong in a moment.

Broadly Solid Details

Chart 1 shows some of the details to start.

On a provincial basis, the gain was driven by Ontario followed by BC’s distant second place showing. Mild gains were recorded in a few other provinces and the only notable dip was in Alberta (-7k). Chart 2.

The gain was led by private payrolls (+58k) and the self-employed (+44k) as public sector payrolls fell by 27k. The latter was significantly driven by fewer civil servants as public administration jobs fell by -14.5k.

By sector, chart 3 shows high breadth to the gains. Services led the way with 64k more jobs as goods sectors added about 11k. Within goods, construction added 16k and manufacturing added 11k jobs while agriculture (-10k), natural resources (-5k) and utilities (-1.6k) dropped.

Within services, the gain was almost equally driven by wholesale/retail (+21k), the FIRE finance sector (+18k) and the professional, scientific and technical services sector (+16.6k).

Can the large job gain be dismissed as statistical sampling noise? Not really. chart 4 shows the reported change in jobs in red bars by month and the 95% upper and lower confidence bands that reflect the noise factor. Even the lower band would have been a mild gain and the top band would have been explosive. There is only a small chance in repeated sampling that the number could have been below the lowest band and the same chance it could have been above the top band.

The seasonal adjustment factor was high, but not as high as the prior two years (chart 5). The pattern of pandemic-era SA factors being higher than pre-pandemic still holds, but less so.

The seasonally unadjusted change in jobs is negative in July more often than not but this was among the weaker months when comparing like months of July over time (chart 6). That too is the pattern of the pandemic era as all of the weaker NSA changes have been in the pandemic period which is what higher SA factors are seeking to control. The pandemic messed up a lot of things including seasonality effects years after the fact.

Hours lost due to sickness and weather remained higher than usual for like months of July (charts 7, 8). Maybe I should go licking hand rails in shopping malls because I just don’t see much evidence of sickness out there. Perhaps some folks need doctor’s notes. If true, and if hours also mean that hiring might have been held back by the effects of sickness and weather on the process of applying, interviewing, accepting and starting a job, then perhaps that reinforces a strong job gain in July if not a stronger one.

Women rocked the job market in July (chart 9). They led the gain and also did so two months ago.

Hours worked were up by 0.6% m/m SA which is a strong indication for July GDP growth given that GDP is hours times labour productivity (chart 10). Hours are tracking an explosive 3.8% q/q SAAR expansion in Q3 which suggests continued strong GDP growth in Q3 relative to our already strong tracking for Q2 (chart 11).

Wages cooled as expected after the torrid pace of increase the prior month. Wages slipped by 1.8% m/m SAAR in July after a downwardly revised 8.3% rise in June (chart 12).

The unemployment rate dipped a tenth partly because of the strong job gain but the amazing thing was the 60.5k rise in the size of the labour force which was the biggest since December. That lifted the participation rate a tenth to 65.1%. Charts 13, 14.

In closing, something’s asunder in the population counts. I know that the LFS lags the actual population counts provided in separate quarterly surveys partly because the LFS applies a rolling 12-month smoothed approach to estimating the temporary resident category which is the one that has been in freefall. But LFS is still reporting about a 200k rise in population over the past year and another 15k increase this July over June. So, is population rising or falling? Is the labour force disconnected from the population counts?