Reclaiming Canada's Labor Data

This post is part of a series of contributor perspectives and analyses called "Democratic Accountability in the Shadow of US Tech Power—How Should Canada and Australia Respond?" Learn more about the call for contributions here.

If Ottawa wants to know how many machinists Saskatoon will need next spring, it waits for Statistics Canada to process another monthly Wage and Vacancy Survey. Meanwhile, Microsoft-owned LinkedIn already knows, because the recruiters paid for their ads last night. We have allowed a handful of American platforms to become the undisputed stewards of Canada’s real-time labor-market pulse, even though that data now informs immigration targets, reskilling funds, and billion-dollar industrial bets. The worst part is we’ve come to accept a system where private platforms hoard core economic intelligence and treat it as business as usual.

The data sucks. Public labor market information (LMI) is patchy, outdated, and spread across multiple unconnected surveys and portals. Private players like Burning Glass (now Lightcast) and Vicinity Jobs (recently acquired by the Conference Board of Canada) scrape job postings at scale, offering insights that Canadian policymakers must pay to access. This is inherently reactive because Canada has ceded critical infrastructure to private actors rather than modernizing its own tools. How fit for purpose is government labor data today?

We once had more transparent and accessible alternatives to these platforms, although job ads were always somewhat dependent on ad placement (such as, in a newspaper, or a shop window). Job boards used to be physical, and search costs were high. Now, the labor market feels infinite and global, which is remarkable, but also inundates recruiters with applications. Increasingly, AI tools are deployed to screen applicants, which can alienate and frustrate job seekers. The data confirms it: it’s getting harder to find a job.

A labor market mediated by platforms

Canada spends millions on labor surveys and runs a federally backed “Job Bank” that was once a digital marvel. But we never rebuilt it for the era of search indexes and API feeds. It is no surprise that job seekers drifted to Indeed, Glassdoor, and LinkedIn. As a result, we’ve effectively privatized the country’s labor graph. Resumes, vacancies, and wage signals sit on servers in California and get monetized by pay-per-click ads and recruiter subscriptions. Should Microsoft really have more insight into Canada’s labor market than Statistics Canada?

It’s not just nuanced labor market information that we’re surrendering. These platforms are highly surveillant. LinkedIn mines everything a user types, clicks, views, or searches to determine what appears in their feed, who they are prompted to connect with, and which jobs are surfaced. We’ve ceded curatorial power to the platform, and assume that all relevant opportunities will be presented to us, but that’s not necessarily the case. We’re placing trust in a digital system that isn’t accountable to us; rather, it’s motivated by data extraction and engagement in an attention-based economy.

Statistics Canada still bases all official labor-market indicators on its own surveys and administrative files. However, a federally backed inter-governmental body, the Labor Market Information Council, has a standing data-sharing agreement with LinkedIn’s Economic Graph (at no cost), and several provincial governments have purchased LinkedIn “Talent Insights” licenses for access to real-time labor analysis. Why are we paying for access to information that we should already own?

Canada’s online-hiring market is also highly concentrated: Recruit Holdings’ Indeed and Glassdoor dominate Canada’s online recruitment market, while LinkedIn Jobs pushes the Big-Three share past two-thirds. Following them are platforms like Ceridian’s Dayforce and UKG’s Ultipro (enterprise HR portals whose employee log-ins inflate “job-board” rankings), the federal Job Bank (the only major domestic player), and smaller sites like ZipRecruiter, Eluta, and the now-legacy Monster. In practice, most real-time data on Canadian vacancies and wages resides on servers owned by two US firms and one Japanese conglomerate, leaving Ottawa’s own Job Bank as the only significant platform under national jurisdiction.

What role should public job postings play in a healthy labor ecosystem? Provinces once invested in their own labor market information centers: Ontario, Alberta, and others maintained regional job boards and labor observatories, but these efforts have largely atrophied. The federal Job Bank remains, but its interface, matching algorithm, and data feeds feel like a time capsule. A modernized Job Bank could serve as Canada’s “public labor API,” transparent, comprehensive, and built for interoperability while complementing local or sector-specific portals.

In other domains, we’ve allowed gig platforms like Fiverr, Upwork, Uber, and others to create closed ecosystems of curated work opportunities; a situation that can be described as being monopsonistic. A public job board could also facilitate matching for fractional and gig workers, albeit in a less punitive manner.

Add it to the list of digital sovereignty challenges. When real-time skills data is behind a paywall (both in terms of employer demand and workers’ skills), Canada is left working with second-hand numbers. Provinces design training programs months after demand spikes. What if we could develop insights that predict demand in advance and inform training initiatives instead? We could also better integrate worker capability assessments in the dataset. Workers in a city might accept stagnant pay because they are unaware that wages are rising in one city over. If Saudi Aramco owned our pipeline gauges, we’d probably say that was a security risk. Labor data should absolutely trigger the same reflex.

Learning from other global models

Canada wouldn’t be starting from scratch. Other jurisdictions have already shown how public employment platforms can evolve into real-time, data-driven infrastructure that supports workers, employers, and policymakers alike. Canada can build on the tools it already has.

Take Australia, for instance. The federal government has developed a pair of powerful tools: JEDI (Jobs and Education Data Infrastructure) and NERO (Nowcast of Employment by Region and Occupation), which combine live job postings with traditional labor market surveys. These systems help policymakers anticipate where skills will be needed, region by region, month by month, and estimate how many local job seekers in the area currently possess those skills. It’s a public-private architecture designed to close the gap between what’s happening on the ground and what the government sees, and it works.

The European Union is also moving in the same direction. Through Cedefop’s Real-Time Labour Market Information initiative, EU institutions are scraping and analyzing vacancy data from public and private sources across member states. The result is a highly detailed picture of skill demand across sectors, which helps countries not only respond to employer needs but also forecast where public investment in training will have the greatest impact.

In the Basque Country in Spain, the Basque Talent Observatory aggregates insights from over a dozen data sources (both online and offline) to produce a dynamic map of local labor demand. It serves as a model for using workforce intelligence to guide regional planning, rather than waiting for outdated stats to catch up to economic change.

Sweden has taken a different approach, focusing on public-service delivery. Its employment agency, Arbetsförmedlingen, has built a full-stack API-based platform that enables integrated digital case management. Using tools like MuleSoft and Pega, the system delivers “next best action” prompts for caseworkers, allowing them to provide personalized, data-informed recommendations to job seekers. It’s not just a job board, it’s an adaptive, public-facing hiring infrastructure.

On the common job board front, the US National Labor Exchange (NLx) is a public-private partnership that serves as the most comprehensive national database of job listings. It connects job seekers with employers by aggregating postings from state and local job boards into a centralized federal system. This tiered structure: from local to state to national, ensures that listings are both wide-reaching and locally relevant, offering a coordinated alternative to fragmented commercial job platforms.

Labor data as national infrastructure

One promising vehicle is legislated interoperability and data escrow. Legislated interoperability uses laws or regulations to require major players to open up their systems, thereby turning private control into a public right, allowing others to connect and compete without waiting for voluntary standards or antitrust action. Mandatory data escrow would require any platform with a certain proportion of the job ads market to deposit anonymized, machine-readable postings and wage data in a public vault daily. Privacy would be preserved, but policy-makers would gain access to better signals. And job ads should be accessible via an interoperable API. That way, dominant job boards would be required to publish an open API to smaller sites, or even a system overseen by a government, that can syndicate listings in real time. This would diminish the power of network effects as a competitive moat.

A modernized Job Bank would ensure that Canada can better capitalize on digital data about the labor market. Rewriting the functioning of the market is only half the job; the other half is building a credible public alternative that the Canadian people collectively own.

We should relaunch Job Bank as a Crown asset, fusing StatCan’s job vacancy and wage survey, and the feeds from job platforms into a single, open API. Other groups could build interfaces, but Ottawa would (and should) control the infrastructure. Alternatively, provinces and sector councils could form a Labor Graph Co-operative, pooling regional data while sharing one API, to ensure both local control and national reach.

The current Job Bank isn’t up to the task. Employers don’t trust it, job seekers find it clunky and ineffective, and its job-matching algorithm has been widely panned. Evaluations show low employer awareness and poor engagement, with many relying instead on private platforms like Indeed or LinkedIn. The underlying data is outdated or irrelevant for many modern roles, and users frequently report that the platform’s matches miss the mark entirely. It’s a federal service in name only and not a serious player in the labor market or as a core source of labor market information.

Still, bolstering the mechanisms that facilitate talent matching in the country won’t solve the ‘invisible’ labor market – those openings that are filled through informal or closed channels.

Canada could radically improve labor market matching by making postings more transparent and sharing real-time data on emerging skills with universities and colleges, allowing curricula to keep pace with demand. Today’s system is clogged with “ghost jobs” that employers never intend to fill, distorting the signal for job seekers and policy makers alike. Hiring processes already feel random and futile, demoralizing job-seekers. Without more transparent information and oversight, labor market matching risks becoming dysfunctional and friction-filled.

The pledge to allocate 5 percent of GDP on defense includes billions for cyber and talent pipelines. A sovereign labor data utility is a national security infrastructure. Canada already operates a passport office and a weather service - among many other things. It can absolutely run a labor-data information utility that isn’t trapped behind a premium subscription.

Consider weather apps: Environment Canada provides the data through their WeatherCAN app, and anyone can build a forecast interface on top. Crucially, the public source doesn’t surveil users or optimize for clicks; it simply serves the public. Why should our labor market infrastructure not work the same way? A rebuilt Job Bank without relying on platform monopolies to mediate what workers see is within reach.

We could treat labor-market intelligence like clean water: a public good, delivered through shared pipes and priced at cost. Part of a digital sovereignty agenda needs to include reclaiming our labor market dashboard.