--- title: "Temporary foreign workers and Canadian innovation: what the automation evidence shows, and the study nobody has run" kind: "Article" topic: "Canadian economy & policy" date: "2026-10-10" url: "https://canopticon.com/articles/temporary-foreign-workers-and-canadian-innovation" author: "Chris Telford" license: "CC BY 4.0" --- The claim that immigration drives innovation is well evidenced. It is also evidence about a different kind of immigration than the kind Canada has mostly been running for the past decade, and the economics of the kind it has been running points the other way. ## Why the "immigration drives innovation" evidence does not transfer The strongest case for immigration and invention is American, and it is about a particular kind of immigrant. Using records of inventors who died young, researchers estimate immigrants are responsible for approximately 32% of aggregate US innovation, over half of which is human-capital externalities on US-born collaborators. The headline is robust. Its scope is narrow: The Kerr literature is exclusively about high-skilled immigration (H-1B holders, STEM PhDs, university-trained scientists and engineers, and entrepreneurs). It does not speak to TFW agricultural workers, fast-food cooks, food-counter attendants, or Tim Hortons supervisors — the categories that dominate Canada's actual 2020s intake. The American programme also has a guardrail the Canadian low-wage stream lacks: The US H-1B program imposes a minimum wage and a labour-certification ceiling that the Canadian TFW low-wage stream does not. Read correctly, the American evidence points at a policy Canada has been moving away from: Even if Kerr's externalities were operating in Canada, they would imply that the policy lever is high-skilled selection — exactly the points-based system Canada has been quietly dismantling ## What changed was the composition, not the number The familiar research on immigration and invention studies points-selected skilled arrivals. Canada's intake moved elsewhere. Temporary work authorisations grew from approximately 67,000 in 2000 to roughly 1.5 million in 2024, and within that the low-wage stream grew fastest — Low-wage TFWs specifically grew from 15,817 in 2016 to 83,654 in 2023. By the end of the period most of the growth was temporary: By 2023, non-permanent residents accounted for roughly two-thirds of the immigrant contribution to Canadian population growth, up from one-third in 2019. The surge ran through every temporary channel at once. The number of international students with active permits rose to 1,040,985 in 2023, a 29% year-over-year increase. Food and retail grew fastest: TFW approvals for food and retail jobs rose 211% between 2019 and 2023. The temporary route also became the main route to staying: Three-quarters of principal economic-class permanent residents in 2023 had previously held temporary work or study permits in Canada, which the research describes as the points-based front door being replaced by a temporary side door. ## The central bank measured what that did to wages The Bank of Canada's staff research is the most authoritative number on the question, and it is a statement about skills rather than about people. Had the temporary workforce of 2023–24 matched the qualifications of its 2006–14 equivalent, their wages would have been approximately 7.5% higher. The distance from Canadian-born pay widened accordingly — The wage gap between temporary workers and Canadian-born workers more than doubled, from 9.5% to 22.6% — and the effect shows up in the national aggregate, where the shift reduced nominal wages across the Canadian economy by approximately 0.7% in both 2023 and 2024. The central bank's leadership has framed weak productivity as a national problem without blaming immigration. The governor said that if productivity growth since 2000 had matched other G7 countries, GDP would be 9% higher today — about $7,000 per person. The cause he names is the one the automation research targets: weak business investment per worker (especially in machinery, equipment, and intellectual property). ## Cheap labour measurably slows automation — elsewhere The mechanism connecting that to innovation has been tested abroad with the kind of design that supports a causal claim. In one study, a 1% increase in low-skill wages causes a 2–5% increase in automation patenting in the relevant field, and the reverse holds when wages fall. A German quasi-experiment pushes the same lever directly: an exogenous increase in the low-skilled workforce reduces automation innovation as measured by patents, concentrated in large manufacturers and in process automation. When labour is cheap and plentiful, the machine that would replace it does not get built. Other work points the same way. One estimate finds an elasticity of substitution between automation capital and labour of 3.8 among actively automating firms, meaning cheap labour strongly suppresses adoption. A Waterloo economist offers a comparison rather than a proof: The Norwegian fish-processing industry is highly automated because Norway does not import low-wage labour for it. Canada's equivalent plants have unlimited access to temporary workers, and are not. ## Employers are not describing a shortage Canadian firm-level data suggests the attraction is not availability. Matched employer-employee records find temporary workers work longer hours than comparable Canadian workers, are absent from work less often, and put out more effort for lower earnings than domestic employees. The federal government's own programme evaluation concedes the consequence, which is notable because it is an admission against interest — some key informants and focus group participants indicated that worker displacement may be occurring in some sectors (trucking, construction, food industry, beauty parlors). The same evaluation describes how the pattern locks in: once employers have hired from a particular origin country, they get information about productivity and commitment to the job and tend to keep hiring from the same origin. The federal auditor saw the business model in 2017. The department relied on employers' own accounts of need and did not check its own records, which could have shown that over 80% of laid-off Canadian workers in some sectors using TFWs were claiming EI at the same time those companies were employing TFWs. The auditor concluded that some companies have "effectively built a business model on the program," which could be causing wage suppression and discouragement of capital investment— the productivity channel, named by Ottawa's own watchdog. In food service the pattern is now structural. A food-industry researcher calls the programme "a structural labour strategy aimed at suppressing wages, lowering turnover, and sidestepping long-term investments in human capital." One chain illustrates the scale: Tim Hortons increased its TFW hires by an estimated 1,131% between 2019 and 2023. And a market grew around the permits themselves, with jobs worth immigration points sold for tens of thousands of dollars. A consultant described the hiring plainly: "They are not advertising jobs to Canadians in any way. Canadians come last for sure." Enforcement has not kept up. Ottawa failed to collect nearly 40% of fines levied against TFW-program rule-breakers, and of approximately 20,000 employers granted TFW approvals between April–September 2024, only 649 were inspected. ## The two-step pathway keeps the shortage going The route from temporary permit to permanent residence turns a shortage into a cycle. Leading labour economists describe the problem: the two-step pathway is structurally myopic because workers who do become PRs are then as likely as any other Canadian to refuse low-wage, unattractive work — meaning employers must continually re-import disposable labour to sustain the model. Workers accept low wages because permanent status is the real prize; once they have it, they leave the job; the employer imports the next group. On that reading: The implication is that the system's apparent "labour shortage" is endogenous to the program design. ## The youth numbers, and what they can and cannot show The most visible correlate is young Canadians. Youth unemployment climbed from 10.0% in 2022 to 13.8% in 2025, teenage unemployment reached 20.8% in Q3 2025, and the overlap is hard to miss — Five of the top six industries that employ the most youth are also among the top users of the TFW Program. The department's own briefing recorded the drop where it would be expected: youth employment declined significantly in retail trade (-8.3%) and accommodation and food services (-8.0%). But attribution is where the research applies its brake, and so should anyone quoting it: the reading that blames immigration principally should be reported as interpretive, not a settled causal finding — minimum-wage increases and weaker aggregate demand are in the same frame. The scale of the youth change is large: the number of unemployed youth rising 57% from 290,000 to 437,000 in three years. Part of the youth population growth was itself temporary residents: Between 2022 and 2024, Canada's youth population (15–24) grew 9.9% — versus 6.0% overall — driven primarily by international students and TFWs. And some study permits were work permits in practice; Statistics Canada found that approximately 19% of international students with study permits in 2019 had no record of studying at any college or university in Canada. The competing explanation is real too: the same period saw a 124.5% average provincial increase 2005–2024 in minimum wages. ## The skilled are leaving while the cheap labour arrives At the other end of the labour market the flow runs out. Statistics Canada net emigration reached 65,372 in 2024–25, the highest in the 50-year series. Skilled immigrants leave too: one in five departs within 25 years, and the most educated go first: Highly educated immigrants leave at twice the rate of less-skilled arrivals. The strongest predictor is pay: Stagnant earnings are the strongest predictor of departure The research joins the two flows with an argument rather than a measurement: If high-skilled native-born and immigrant Canadians are leaving because of wage stagnation, and wage stagnation is being driven (in measurable part) by the temporary-worker compositional surge, then the cheap-labour immigration policy is causally producing the brain drain it nominally seeks to offset. It is a chain of inferences, each with support; the whole has not been tested. The startup gap is consistent with it: The U.S. went from producing 11 times more high-potential startups than Canada in 2015 to 45 times more in 2024 ## The Canadian study has not been done For a question argued this confidently, the central test is missing. The research says so in the plainest terms: There is no published study directly comparing innovation output between TFW-heavy and TFW-light Canadian firms or industries. The German work is the nearest analogue, not a substitute, and a definitive Canadian replication of the Hémous or Ganglmair findings has not yet been published. Everything above is a chain of strong links with the last one unmeasured — which is a reason to run the study, not a reason to assume its result. The research lists the studies a serious answer needs, and none exists yet: a Canadian repeat of the German design using regional temporary-worker inflows and patents; A firm-level study of capital expenditure and automation adoption between TFW-using and non-TFW-using Canadian firms in matched sectors.. Another: A causal study of LMIA-driven promotion crowd-out on Canadian-born entry-level workers. ## Ottawa has started to reverse course Policy has begun to move. From September 2024 there is 10% cap on the share of any employer's workforce that can be hired through the low-wage TFW stream (20% in select sectors); ban on low-wage TFW hiring in census metropolitan areas with unemployment ≥6%. Student permits were capped, and the current government's plan, in the research's summary: Targets economic immigrants at 64% of all admissions by 2027–2028. How far the cuts have gone is disputed: the department claims large reductions, while the opposition alleges 105,000 new TFW permits in H1 2025 alone, a figure the research flags as coming from a partisan source. > **Objection: This is an argument against immigrants** > > It is an argument about a programme's incentives, and the people it treats worst are the workers inside it — which is why the same evidence is cited by labour economists and by the United Nations Special Rapporteur on contemporary forms of slavery. The remedy the economists propose is not fewer people but a different instrument: phasing out the low-skill streams over 3–5 years to give viable businesses time to make productivity-enhancing training and technological investments. A worker admitted permanently, with the right to leave an employer, is not the subject of this finding. > > The harm to the workers is documented at the highest level. The UN's Special Rapporteur on slavery called the programme a "breeding ground for contemporary forms of slavery" that "institutionalises asymmetries of power that favour employers and prevent workers from exercising their rights." Closed permits tie a worker's status, housing and healthcare to the employer, and the Rapporteur met with some workers who had been instantly dismissed for exercising their trade union rights. ## What this explains It explains why the innovation argument and the immigration argument keep talking past each other: one is built on studies of selected skilled migration, the other on a programme that selects for compliance and low cost. It explains why Canada's productivity problem and its labour-supply solution point in opposite directions — the cheaper the labour, the weaker the case for the machine. And it marks what would settle it, which is a study of Canadian firms rather than German ones. Until someone runs it, the strongest honest statement is that every link in the chain is evidenced except the one that bears the conclusion.