Canada’s economy is booming, as long as you ignore the actual data and upcoming revisions. Statistics Canada’s (StatCan) latest Labour Force Survey (LFS) shows seasonally adjusted employment climbed by 75,000 jobs in July. The unseasoned reality was a little less savoury, showing a 6-digit decline in the same month, along with rising unemployment. The picture may be even worse than presented, with a Big Six bank projecting substantial upward revisions to population estimates in just a few weeks.
Canadian Unemployment Falls To 2-Year Low*
Canadian seasonally adjusted unemployment rate.
Source: StatCan.
Canadian seasonally adjusted (SA) employment rose 0.4% (+75k) to 21,215,000 in July, adding 0.1 points to the employment rate. This helped trim 0.1 points from the SA unemployment rate, which fell to 6.4%, its lowest level since July 2024.
Two provinces that have been lagging got a boost from the World Cup last month. Ontario made the most gains, seeing its SA employment rise 0.6% (+52k), pushing its unemployment rate to 6.8%—its lowest point since July 2024. BC also saw employment rise 0.6% (+18k), pushing its unemployment rate to 6.2%, shedding 0.6 points in two months.
What do you mean we’re saying seasonally adjusted too often? This is a totally normal number of times to say it—when seasonally adjusted. But since you asked, there were a few odd details worth noting if the data fails to match reality.
Unadjusted Employment Data Shows Canada Lost 127,700 Jobs
Seasonally adjusted data accounts for predictable seasonal variation. For those who missed everything since 2020, Canada hasn’t had much of that. Consequently, those adjustments risk skewing the actual data. It’s virtually impossible to separate seasonal variations from shock until after the fact. This is a problem that experts have previously warned risks overstating growth, providing the wrong insight at the wrong time, and extending pain.
The unadjusted data looked a bit different. While Canada added 75,000 SA jobs in July, it lost 127,700 that same month when unadjusted. The SA unemployment rate fell to a multi-year low of just 6.4% in July, but it climbed a mind-boggling 0.6 points to 6.7% when unadjusted. Then there’s that darn population growth quirk.
Canadian Population Revisions To Reveal Higher Unemployment
Canada’s latest population estimates revealed a historic decline, but there’s one problem—it may not have existed. Last week, CIBC economists explained that StatCan assumed a non-permanent resident (NPR) left when their visa expired. It didn’t factor in the hundreds of thousands who transitioned to different visas or received extensions that overrode the previous policy of cutting them, legally remaining in the country. We won’t bore you with the bank’s full estimates again, but the economists noted the undercount has artificially improved employment data.
It’s important to note that while StatCan uses hard data for employment, it projects its population and labour force. This means it may not have counted a person in the population or labour force, but that same person is counted in the employment column. As a result, this skews unemployment.
How much does it change the picture? Let’s be generous and assume just 60% of those undercounted were in the labour force, though this demographic tends to skew higher. Using CIBC’s projections, that would add roughly 222,000 workers, pushing the 6.7% unadjusted unemployment rate to 7.6%. For context, that would mean only two months have reported higher rates since the end of the pandemic.
StatCan itself warned it will be making an “unusually” large revision to population estimates in September. The shock will be real, but seasonal adjustments will slowly introduce those revisions dating back to 2025, according to CIBC.
The data revisions don’t change reality; they change the narrative around it. If the narrative doesn’t match what you’re seeing, it’s probably not just you. Unfortunately, you’ll have to wait 12 to 24 months for reality to catch up.
But what about cats?
poorly written, pathetic article.
Poorly written, pathetic comment
comment lacks wit & originality. 2/10
He is correct though, the article is trash.
Nah, they’re the only ones that talked to StatCan about the actual modeling of the numbers (few weeks back), and therefore are the only ones reporting on the actual data.
StatCan is knowingly cranking out trash while Canada issues record bonds, and if you understand that and think the author’s grammar or something is the problem, you should stop banging your cousin for a minute and learn some finance basics.
I totally see that perspective for anyone that can’t do math.