Canada issued 192,240 employer-specific work permits in 2024, and the ordinary route to a different employer requires the new employer to first obtain a Labour Market Impact Assessment

What is happening?

The Temporary Foreign Worker Program of Canada runs on employer-specific work permits, sometimes called closed permits. 192,240 of them were issued in 2024. A closed permit fixes four things at once — the employer, the occupation, the wage and the work location. Moving to a different employer requires that the new employer first obtain a Labour Market Impact Assessment, after which the worker obtains a new permit.

In 2019 Immigration, Refugees and Citizenship Canada opened a second door. The Open Work Permit for Vulnerable Workers is a route created for workers facing abuse — physical, sexual, psychological or financial — or facing the risk of it, and it lets the holder start work elsewhere without the cooperation of the employer named on the current permit. 5,115 of these open permits were issued in 2024 and 9,625 in 2025, against 875 in 2021.

Three public bodies have examined the structure. The UN Special Rapporteur on contemporary forms of slavery visited Canada from 2023-08-23 to 2023-09-06 and said in a statement on 2023-09-06 that the program is a breeding ground for contemporary forms of slavery, and that the employer-specific work permit leaves migrant workers unable to report abuse for fear of removal. The Standing Senate Committee on Social Affairs, Science and Technology concluded on 2024-05-21 that employer-specific work permits make migrant workers inherently more vulnerable to abuse by bad actors and impose structural barriers on access to rights and protections, and it recommended phasing closed permits out within three years. The Office of the Auditor General of Canada audited the program in 2017-05 and audited health and safety for agricultural migrant workers again in 2021.

Whose problem is this?

The body that would have to change the form of the permit is the same body that issues the relief permit to people harmed under it. Nothing in the design of the ordinary route lets a worker move without a second employer applying first.

Where does this problem end?

The boundary matters here because the way out already exists and was built for exactly this situation. What is in question is not whether anyone thought of the problem but how far the remedy reaches.

What is the state now, and what should it be?

Now

Needs a new measurementthe target state: no source opened here gives a government target of any kind for this. Nothing names a date by which employer-specific permits would end, a share of workers who ought to be able to change employer without a new Labour Market Impact Assessment, or a level the open permit for vulnerable workers is meant to reach. A Senate committee recommended on 2024-05-21 that closed permits be phased out within three years, which is a recommendation addressed to government rather than a target government has adopted, and whether a written government response to that report exists is not stated in any source opened here.

How big is it?

192,240 employer-specific work permits were issued in 2024. That is a flow rather than a stock, and it does not convert into a headcount of people.

The reason is that an issuance count for one calendar year includes both first permits and extensions. It therefore cannot be read as a number of distinct people, and it is not the number of workers holding such a permit on any given day. No source opened here publishes that second number, and the arrival targets of 82,000 for 2025 and 60,000 for 2026 rest on a different counting basis and exclude seasonal workers, so they cannot be subtracted from it or divided into it.

The relief route is counted on the same issuance basis as the closed permits. 5,115 open permits for vulnerable workers were issued in 2024 and 9,625 in 2025, against 875 in 2021. The two counts are comparable as counts of permits, but they do not describe one cohort — everyone in the first number received a permit naming an employer, while the second counts people approved for release from such a permit, including people whose original permit was issued in an earlier year.

The rate at which those applications succeed is known for a single year. In 2023, 58 percent of applications for the open permit for vulnerable workers were approved and 80 percent were finished within 34 days. No later approval rate appears in any source opened here.

Under what conditions does it arise?

1. The relief asks for proof from inside the situation it is meant to end. The open permit for vulnerable workers is available to a worker experiencing abuse or at risk of it, and the application asks that worker to establish this while the permit naming the employer is still the only lawful basis for being in the country. The document that has to be escaped is also the document that has to be preserved until the escape is granted.

2. Mobility is decided about the employer, not about the worker. A Labour Market Impact Assessment is a finding about a job and an employer, so the ordinary route out of a bad placement runs through a second employer completing an application. No step in that sequence can be started by the worker alone.

3. A gap can open between the old permit and the new one. A public policy in effect since 2020-05 lets a worker begin work for a new employer before the new permit is finalized, which shortens the gap without altering the employer-specific structure that creates it.

4. Enforcement is sampled. In the fiscal year ending 2025-03-31 the responsible department inspected 1,435 employers and found 10 percent non-compliant, and in the following year it inspected 1,488 and found 12 percent non-compliant. No source opened here gives the number of employers participating in the program, so the share of employers reached by inspection cannot be computed from these figures.

5. The audit record predates the relief permit. The Office of the Auditor General of Canada reported in 2017-05 that in 40 percent of the files it reviewed the responsible officials did not challenge the labour shortage the employer had stated, that on-site and in-person inspection was insufficient, and that approving a sanction took months.

These five conditions point the same way. The person with the most to lose from a bad placement is the person who has to act first, and the action requires evidence gathered from inside the placement while it is still the only status that person holds.

What has been tried?

Two directions have been tried at once. One adds a way out for the worker who can establish what happened, and the other raises the penalty an employer faces when an inspection finds a violation. Neither changes whose name is written on the permit.

What was found?

Why is it still unsolved?

Incentive inversion — the one official way out of an employer-specific permit has to be requested by the person with the least room to request it.

A permit naming an employer is not only a work authorization. It is the basis on which the holder is lawfully present. So the request to be released from it comes from someone whose presence in the country currently rests on the very document being questioned, and it has to be supported by evidence about the employer written on that document. Nobody has to act in bad faith for that to bite. It follows from where the paperwork sits.

The ordinary alternative is an employer-side decision. A Labour Market Impact Assessment is a finding about a job and an employer, so a worker who wants to leave has to find a second employer willing to complete an application before there is anywhere to go. The open permit for vulnerable workers is the only route a worker can start alone, which is why the size of that route is the measure of how wide the exit actually is.

The record is not static. Issuances of the open permit rose from 875 in 2021 to 9,625 in 2025, inspections and penalties both rose, 36 employers were barred in a single year, and the intake target for 2026 is roughly a quarter below the one for 2025. What has not moved is the form of the permit. The share of inspected employers found non-compliant went from 10 percent to 12 percent across the two most recent reported years, measured on a sample whose coverage of the employer population cannot be computed from the published figures.

One public body recommended a deadline and no source opened here records a government adopting one. That is the part of this that stays stable. A recommendation with no date attached to it looks the same from the outside whether it is being worked on or not.

What observation would mean it is solved?

Candidates — a. work permits under the program stop naming a single employer, replaced by the sector-based or region-based permits the Senate committee recommended on 2024-05-21. b. the number of workers who change employer without a new Labour Market Impact Assessment rises, so that the open permit for vulnerable workers is no longer the only route a worker can start alone. c. the share of inspected employers found non-compliant falls while inspection volume is held steady or raised.

a alone is weaker than it looks. A sector-based permit still ties lawful presence to a category, and a worker who needs to move outside that category is back where the closed permit left them. A change in the form of the permit can also arrive with a transition period long enough that the people currently holding closed permits never reach it.

b alone is ambiguous in the way that matters most here. A rise in open permits for vulnerable workers can mean that more workers are reaching relief or that more workers need it, and no source opened here separates the two. The count rose more than tenfold between 2021 and 2025 with no source stating which of the two that movement measures.

c alone counts what was inspected. The non-compliance share is computed on employers who were inspected rather than on all employers in the program, so it moves with how the sample is drawn. It also measures employer conduct after the fact rather than whether a worker was able to leave.

The three have to be read together, and b has to be read against whether the underlying rate of abuse is being measured at all.

What is it connected to?

Fills with researchthe separate transfer provision inside the Seasonal Agricultural Worker Program, housing supplied by the same employer named on the permit, provincial labour standards enforcement for workers whose immigration status is federal, and the same employer-tied permit structure in other countries. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• How many employer-specific permits are in force. No source opened here publishes a stock. What exists is an annual issuance count of 192,240 for 2024 and arrival targets of 82,000 for 2025 and 60,000 for 2026, resting on different counting bases, so neither subtraction nor division between them is available. A federal statistical overview of temporary foreign workers was opened and carries no such stock figure either.
• How many workers fall out of status while an application for the open permit is pending. The 2023 figure says 80 percent of applications finished within 34 days and says nothing about the state of the original permit in the meantime. No source opened here counts the cases where it expired first.
• Whether the government responded in writing to the Senate committee report. No source opened here carries such a response, or any government statement adopting or declining the recommendation to phase closed permits out.
• Why employer-specific permits remain the form of the program. No source opened here carries a departmental explanation of it. An opening statement to a parliamentary committee dated 2023-11-07 did not open.
• The composition of the interview sample behind the 2025 report by an international human rights organization. The report page did not open, so only second-hand reporting of it was available, and the distribution of that sample by sex, industry and province appears in no source opened here.
• Approval rates for the open permit after 2023. Only the 2023 figure of 58 percent was found. Issuance counts are published year by year and approval rates do not appear beside them in any source opened here.
• What became of the penalties and the bans. Whether the 36 employers barred in the year ending 2025-03-31 contested the decision, and whether the penalties assessed were paid, reduced or overturned, is absent from every source opened here.

See the evidence

The evidence tier for this file is secondary. One federal document was opened directly, the question period note on employer-specific work permits, and it carries the 2024 issuance count of 192,240, the four conditions a closed permit fixes, the Labour Market Impact Assessment requirement, and the description of the open work permit for vulnerable workers created in 2019. Everything else rests on secondary reporting of federal releases, on a UN news page, or on a federal news release carried by a wire service. Four sources that would have been primary did not open — the Senate committee report is a scanned image, the report by an international human rights organization and the departmental penalty release returned HTTP 403, and the 2017 audit page returned HTTP 404. Those four rows are left in place with the URL column blank and the reason stated, rather than filled with a substitute address. Where sources overlap they agree: the Senate committee conclusion and the three-year recommendation appear in two independent reports, and the two most recent compliance years are reported consistently across two outlets. Where they do not overlap the gap is left visible — the 2023 approval rate of 58 percent has no counterpart in any later year, and no stock of valid employer-specific permits appears anywhere. This is a research-based definition, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below