Canadian labour market: Could we really run out of room before year-end?

While the implied odds dropped after last week’s rebound in the unemployment rate, financial markets are still pricing in the possibility of an interest rate hike before the end of the year, suggesting that some investors still believe Canada will quickly absorb its room for non-inflationary growth. Given that the jobless rate now sits a full percent higher than its pre-pandemic level, running out of room by then would need either a rapid improvement in that measure from here, or for the non-inflationary (NAIRU) rate of unemployment to be persistently higher than it was pre-pandemic.

We would disagree firmly on the first count, and there are signs that the NAIRU rate, which may indeed have been more elevated in recent years, could now be on a path back to a lower rate. Measures of labour market slack that are broader than the unemployment rate alone suggest that there is more room for non-inflationary gains in employment ahead, and cooling compensation data also support that assessment.  The Beveridge curve relationship between job vacancies and unemployment appears to be moving back closer to pre-pandemic norms, partly due to improved job prospects for newcomers into the country. And while trade barriers could leave some structural unemployment in place, demographic changes should help alleviate the task of matching people to jobs. If the NAIRU rate moves back towards, or even below, 6%, returning to full employment this year would require the workforce to shrink or job growth to run twice as fast as that seen during 2025 — neither of which should likely be a base case assumption.

While we do see a period of above trend quarterly growth rates later in 2026 and into 2027, our forecast suggests an extended period in which labour market slack will be in evidence (Table 1). In turn, that should be sufficient to keep the Bank of Canada on hold this year (Table 3).

Hiding in plain sight

Admittedly, the rather muted market reaction to the recent jump in unemployment was perfectly justified, as the overall amount of slack within the labour market was little changed relative to the prior month. It’s just that instead of hiding in the form of a low participation rate, it was showing up in plain sight through higher unemployment (Chart 1).

However, the monthly climb in participation doesn’t mean that there isn’t scope for further improvement, or that we should expect a downtrend from here driven by the aging population. If December’s move sticks, prime aged (25-54) participation is now back close to where it stood in 2023, although the employment ratio still hasn’t recovered much due to higher unemployment.

Where labour force participation remains weak, and where there could be scope for improvement, is among young people (Chart 2, left). This age group has seen the largest increases in unemployment in recent years, which appears to have disincentivized people to look for work. While the rapid increase in international student population may have distorted these figures initially, that doesn’t appear to have been the case more recently. Labour force participation among domestically born 15-24 year-olds is much lower than was typical pre-pandemic.

Because fewer young people are currently engaging in the labour market, the aggregate participation rate remains lower than an aging population alone would suggest (Chart 2, right). If more young people engage in the labour market as the economy recovers, it will be easier to fill job vacancies without sparking an acceleration in wage inflation.

It would also slow the unemployment rate’s recovery back to levels which we feel are consistent with a 2% inflation target. Assuming average employment growth of 20K (in line with the average of 2025), the workforce would need to contract by an average of 15K a month from here for labour market slack to be eliminated in time to support a rate hike before year-end (Chart 3). Using the Bank of Canada’s own assumption of 5K average labour force growth, slack wouldn’t be eliminated until well into 2027, unless employment growth accelerates to average 40K a month.

What’s new with NAIRU?

Admittedly, this also depends on how low the jobless rate can go before wage pressures emerge. And there is plenty of debate over where that non-accelerating inflation rate of unemployment (NAIRU) lies, and where it might be headed as the economy improves and government immigration policies alter Canada’s demographic picture.

Our prior study into a labour-market based output gap placed NAIRU at a bit below 6% prior just ahead of the pandemic. However, NAIRU likely moved higher subsequently, as structural changes caused by the pandemic as well as a big surge in the population resulted in a weaker relationship between skills and job availability. That showed up in a weaker job finding rate, although this has started to recover and should continue to improve if the economy gets closer to full employment.

The trade war, and the permanent loss of manufacturing plants, does lengthen the timetable for finding replacement positions, thereby creating some structural unemployment, but there’s less new in that phenomenon than many think. The auto sector, for example, had been steadily losing ground to plants in the US south and Mexico well before tariffs were imposed. Our base case forecast for 2026 assumes that no new trade barriers emerge, and we might see some progress in sectors like aluminum.

The move higher in NAIRU during the post-pandemic years did see a less-favourable Beveridge curve relationship between job vacancies and unemployment (Chart 4). If higher demand for labour is needed to achieve the same level of unemployment, that would signal more employers competing for fewer available workers, which would bid up wages and result in greater inflationary pressure at any given unemployment rate.

However, it is notable that the modest decline in unemployment relative to its mid-2025 peak hasn’t coincided with higher overall demand for labour. In fact, job vacancies (as measured by the generally weaker SEPH survey) have if anything continued to drift lower. While we’ve seen only a handful of observations, for now we appears to be back on the pre-pandemic Beverage curve relationship. If that holds, it implies that the reduction in joblessness since mid-2025 has represented a structural improvement in job-matching rather than just a cyclical upturn in demand for labour.

Part of this improvement has represented a reduction in unemployment for newcomers into Canada, whose skills may not (initially at least) have lined up well with job openings. The contribution of newcomers to the aggregate unemployment rate has been much higher than normal since 2023, reflecting both the sharp rise in their population as a proportion of the total, alongside an increase in unemployment within the group itself. However, that contribution has started to dip recently (Chart 5).

It would be easy to think that this is purely because the non-permanent resident population (NPR) is falling, as shown by the quarterly population count. However, that doesn’t appear to be the case. Within the labour force survey, the population of NPR’s is actually still rising, albeit more slowly than in prior years (Chart 6).

The purpose of this article is not to debate which population count is correct. While the LFS data could still be overstating population growth, it is also possible that the quarterly count could be understating it, due to the assumption that NPR’s immediately leave the country when visas expire. However, what the continued rise in NPR population within the LFS data does suggest, is that the reason for reduced unemployment within this segment is not simply a decline in its population. Rather, it appears to be another sign of improved job-matching. Government immigration policy changes are also aimed at tilting additions to the permanent resident population, including places granted to those who are currently NPRs, towards those with greater employability, which should again help reduce structural unemployment and lower the NAIRU rate over time.

There are other structural changes in the economy that should help push the NAIRU rate lower again. An ageing population, for example, could structurally lower unemployment rates as there are fewer young people (where joblessness tends to be higher) relative to older people (whose unemployment rates are generally lower).

Another factor that could support a return to a lower NAIRU rate over time is the reduction in short-term unemployment. Apart from during the first month of a recession, the rate of short-term (1-4 weeks) unemployment is typically driven by structural factors. These could be demographic (i.e. the number of young people entering the labour market in any given month), but also could be linked to sectoral changes. A smaller proportion of work is now undertaken in seasonal sectors or those where contract work is common, which reduces the propensity for short-term unemployment. As a result, the proportion of short-term unemployment is much lower today than it was 15-20 years ago (Chart 7), supporting a structurally lower aggregate jobless rate when the economy is healthy.

Seeing is believing

In reality, we will only really know where NAIRU lies when we see it — when broad wage inflation starts to accelerate again. At the moment, private sector earnings growth still appears relatively tame. Business sector hourly compensation and unit labour costs are running at levels that should support a further easing in underlying inflation ahead (Chart 8). And we would expect wage growth to remain only modest ahead, if we are correct in our assumption that there’s still plenty of room for the non-inflationary job gains, including those filled by Canadians who are currently “out of the labour force” rather than counted among the ranks of the unemployed.

That would be particularly true if the most recent relationship between job vacancies and the unemployment rate is maintained, easing previous fears of skills mis-match and a less efficient labour market coming out of the pandemic. We expect that the Bank of Canada will only need to tap the breaks slightly to slow labour demand around the middle of 2027, rather than before the end of this year.