GOVCON WEEKLY
Canada spent the early summer congratulating itself on internal trade, and it had reason to.
In Charlottetown, the first ministers pledged to build "one Canadian economy," naming open interprovincial procurement a priority and targeting an agreement in principle by year end.
Days earlier, the CFIB handed out its best internal-trade report card yet. The federal government jumped from a C to an A+, ten jurisdictions earned an A, and only British Columbia (A-) and Nunavut (C-) trailed.
On paper, the internal market has never looked more open, and the two events fit together neatly: governments made commitments, and the report card rewarded them for making them.
That is the catch. The report card grades what governments promise and legislate. It does not grade what they actually buy, from whom, or whether an outside firm can even see the contract. CFIB says as much itself: its grades, it notes, "reflect commitments and policy intentions more than progress felt on the ground," and it is rewriting its methodology for 2027 to measure outcomes instead of intentions.
We spent the spring on the outcomes. In May we published a paper with the Macdonald-Laurier Institute scoring all fourteen of Canada's procurement systems on how transparent they actually are, and for this issue we went further, into the award data itself. The promises and the grades line up with each other. What neither lines up with is the procurement.
The 30-Second Version
Procurement is the biggest lever, getting the least attention: 13.4 percent of GDP, roughly $350 billion a year.
Only four of fourteen governments let you get a downloadable list of who won each contract and for how much. In the other ten, you copy it out of web pages by hand, pay a private portal to reach it, or it was never published at all.
You cannot measure whether contracts stay local across most of Canada, because most governments do not publish who wins. The blackout is not a side effect of the barrier. It is the barrier.
Where we can see it, local capture is real and uneven: half of Newfoundland's money stays on the Rock, 70 percent of Toronto's stays with Ontario firms, and physical work stays home while technology leaves.
In the same year they earned A's for openness, nearly every government runs a buy-local preference, a US-supplier ban, or both.
The transparency is already required, and ignored: the Canadian Free Trade Agreement obliges governments to publish who won every contract over roughly $130,000, at what price, for what. Most do not, and no one at Charlottetown promised to start.
The Two Report Cards
CFIB grades whether a government has opened its rulebook. We built a different scorecard, one that grades whether you can actually see what a government buys. In the MLI paper we scored all fourteen systems on four pillars, each out of four, for sixteen in total:
Ease of use: can a normal person navigate the portal and find awards without a manual.
Data quality: does a record actually carry the price, the vendor, and ideally the losing bidders.
Coverage: does it capture all award types, including the pre-qualified standing-offer contracts that are easiest to hide.
Accessibility: can you export in bulk or through an API, or are you stuck copying HTML tables by hand.
The spread is enormous, from open data you can download in one click to portals you cannot get the awards out of at all.
Tier | Jurisdiction | Score |
|---|---|---|
Strong | Nova Scotia | 16/16 |
Quebec | 15/16 | |
British Columbia, Government of Canada | 14/16 | |
Moderate | Alberta | 11/16 |
Saskatchewan, Newfoundland and Labrador | 9/16 | |
Ontario, PEI, NWT, Nunavut | 7/16 | |
Weak | New Brunswick | 6/16 |
Manitoba, Yukon | 5/16 |
Strong, 14 to 16 of 16: a real transparency regime, reachable today.
Nova Scotia, 16 of 16: the only perfect score. An open-data portal with an API, multiple export formats, and more than 31,700 contracts including submitted bid information, all downloadable in minutes. A journalist can audit its sole-source awards without a single access-to-information request.
Quebec, 15 of 16: the only Canadian jurisdiction publishing in the international Open Contracting Data Standard, with bulk JSON exports that make cross-provincial comparison possible. Its one failing is a slightly clunkier interface.
British Columbia and the federal government, 14 of 16: broad open-data programs with real history, CanadaBuys reaching back to 2012. Both lose points only for publishing little about who bid and lost, and CanadaBuys tucks some documents behind a separate SAP Ariba login.
Moderate, 7 to 11 of 16: the data exists, but you cannot get it out.
Alberta, 11 of 16: the most user-friendly portal in the country and the richest fields anywhere, including RFP documents, sole-source justifications, and plan-taker counts, and then no bulk export or API at all, so none of it can be pulled at scale.
Saskatchewan, 9 of 16: every award back to 2009 and rich detail, locked behind the same wall.
Newfoundland, 9 of 16: complete data since 2018, published only as HTML tables a researcher has to copy by hand.
Ontario, 7 of 16: competitive tenders almost completely covered, but the awards made through its standing arrangements go undisclosed, and there is no API.
PEI, Northwest Territories, and Nunavut, 7 of 16: they meet their obligations on paper, with weak portals and no real way to extract anything.
Weak, 5 to 6 of 16: closed to anyone outside the system.
New Brunswick, 6 of 16: aware of its CFTA thresholds, but the portal is confusing and the coverage has gaps.
Manitoba, 5 of 16: hands disclosure to a private portal that claims more than 16,000 awards while surfacing only about a thousand.
Yukon, 5 of 16: hands disclosure to a private portal that requires registration and payment just to see what is open.
The Ontario tell is the one to keep: Ontario earned an A on the CFIB report card this month for opening its rulebook, and it scores 7 of 16 on ours, because you cannot see what it buys through half its channels. A government can remove every formal barrier it has written down and still run procurement no outsider can watch. The report card only measures the first half.
Everyone's Buying Local Now
In the same year Canada handed out its best internal-trade grades, nearly every government kept or built a procurement preference. The CFIB grade sits beside each one.
Jurisdiction | CFIB | Standing local / Indigenous preference | 2025 US measure (status) |
|---|---|---|---|
A+ | Buy Canadian: 10% discount + up to 25% content score | US ineligible, non-defence over $10K (in force) | |
A | Buy Ontario: 10% edge for ON firms under ~$139K | US excluded unless ~90% Canadian staff (in force) | |
A | Language / local-establishment rules act as preference | 25% penalty on US-establishment bidders (in force) | |
A- | None standing | US barred by directive (in force) | |
A | None; champions open trade | US ban paused April 2025 | |
A | Priority SK: best-value, ~90% local, no margin | Guidance rescinded June 2025 | |
A | Buy Canadian Act passed, not yet proclaimed | Same Act, dormant | |
A | Discretionary local preference | US barred, contracts cancelled (in force) | |
A | NB First strategy, local preference | Halt on new US contracts (in force) | |
A | None standing | US-contract review | |
Not graded | 10% local price preference since 2020 | None codified | |
A | First Nations policy: up to 15% bid reduction | None | |
A | BIP: 15% NWT + 5% local on first $1M | None | |
C- | NNI: 5%, up to 25% with Inuit firm/labour | None |
Three different kinds of preference are stacked in that table, and the report card measures only one of them.
Layer one: the old local preference, strongest in the North
These are not tariff reactions. Most predate the trade war by years, and they are the most explicit preferences in the country.
Northwest Territories: the Business Incentive Policy adjusts a registered NWT company's bid downward by 15 percent, plus another 5 percent for a local community, on the first million dollars of a contract. On a competitive tender, a 20 percent adjustment is close to decisive.
Nunavut: the NNI Policy gives a registered Nunavut business a 5 percent adjustment and stacks further adjustments for Inuit-owned firms and Inuit labour up to a 25 percent cap, enforced by an independent tribunal.
Yukon: reduces the bid of a Yukon First Nations business by up to 15 percent, and awards up to a fifth of an RFP's points for hiring locally.
Newfoundland: has given firms with a provincial place of business a 10 percent price preference on everything below the trade-agreement thresholds since 2020.
Every one of these lives below the CFTA's thresholds, which is exactly where the median award lives.
Layer two: the 2025 US-supplier wave, most of it already unwinding
When US tariffs landed in early 2025, almost every government reached for procurement as a lever, and almost as quickly, several put it back down.
Quebec was the most aggressive and the most durable: a 25 percent penalty applied to the price of any bidder with a US establishment and no Quebec one, across four categories, computer hardware and software, medical supplies, pharmaceuticals, and scientific instruments, all in force by decree.
British Columbia bars US suppliers by directive unless the deputy minister to the premier signs off, explicitly including software subscriptions and licences.
Ottawa's reciprocal-procurement policy makes US firms ineligible for non-defence federal contracts over $10,000, and existing supply arrangements had to comply by mid-July 2026.
Nova Scotia and New Brunswick barred or halted US contracts outright, cancelling existing ones where they could.
But the wave crested fast. Alberta, which has no standing local preference and openly champions free internal trade, announced a US ban and paused it by April. Saskatchewan issued tariff-response procurement guidance in March and quietly rescinded it by June. Manitoba passed a Buy Canadian Act and left it unproclaimed. A lot of this layer was signal, not policy.
Layer three: the one Ontario just built, and the most consequential for internal trade
The Buy Ontario Directive, in force since April 2026 under the new Buy Ontario Act, is a tiered preference by contract value:
Below roughly $139,000: Ontario businesses get a 10 percent evaluation advantage, or a buyer can restrict the bid to Ontario firms outright.
Roughly $139,000 to $368,000: the preference widens to Canadian firms.
Above that: domestic content becomes a weighted criterion, capped at 35 percent of the score on capital infrastructure.
An "Ontario-made" good means at least 51 percent of its direct costs were incurred in the province, and municipalities came under a parallel directive weeks later, with mandatory Ontario-made vehicle purchasing on top. Read the bottom tier again. In the same year Ontario earned an A for removing its interprovincial barriers, it handed its own firms a formal 10 percent edge over a Manitoba or Nova Scotia competitor on every small contract. Both are true, because the CFTA's thresholds leave the small-contract lane wide open.
Between us, this is why the grades and the reality diverge: the report card rewards the barriers a government removes at the top, above the trade-agreement thresholds. The preferences that actually decide who wins sit below those thresholds, where the CFTA does not reach. Saskatchewan learned the hard way what happens when you cross the line: it built an overt local-labour preference into a highway tender, a losing out-of-province bidder challenged it, and an arbiter found it breached the Canadian Free Trade Agreement. Ontario took the other lesson. Its Buy Ontario edge is deliberately capped at contracts under about $139,000, just below the threshold, where preferring local is allowed. Same instinct, drawn a few dollars inside the rules.
Why You Can't See It
The opacity is not an accident of old software. It has specific, current causes.
The portals went private, and the price of seeing public tenders jumped: after private equity firms bought them, access to Ontario municipal data on Bids and Tenders quadrupled from about $449 to $2,000 a year, and Bonfire climbed to around $1,500.
In Ontario, seeing the whole market costs at least $5,000 a year: a business chasing every Ontario public-sector tender has to subscribe to four separate private portals, MERX at $80 a month, Bonfire at around $1,500 a year, Bids and Tenders at $2,000, and Biddingo at about $250, which carries a set of hospitals and municipalities you cannot find anywhere else. That is a tax on bidding that governments and taxpayers ultimately pay in higher prices, and it locks smaller firms out entirely.
Ontario is the worst offender, and it does not have to be: in Alberta, Quebec, Nova Scotia, and BC, procurement is aggregated to the provincial level, so a tender has to be posted and made available on the central system even when the buyer runs it through a private portal. Ontario has no such aggregation, so the cost of finding the work falls on the bidder.
Standing offers are the black hole: call-ups under standing offers, supply arrangements, and Ontario's Vendor of Record lists are frequently never disclosed as individual awards. It is the most opaque corner of Canadian procurement, where established vendors keep winning work that firms outside the network never even see.
Single-bid contracts cost more: our paper cites research on 3.5 million European contracts finding single-bid tenders run over 7 percent more expensive than competitive ones, and that when Mexico opened its contracting data, opportunities drew up to 25 percent more bids. On a $350 billion base, that gap is billions.
What It Looks Like Where We Can See It
We cannot measure the country, so we measured the markets we could excavate and classify by hand. Same pattern in all three.
Market | Window | Local capture | The tell |
|---|---|---|---|
Newfoundland and Labrador | since 2019 | 50% of dollars stay with NL firms (77% Canadian) | Roads 98% Canadian, the whole tech market just 29% |
Toronto | 2021-2025 | 70% of dollars go to Ontario-based vendors | Construction is 65% of spend; 19% goes foreign |
Quebec (technology only) | FY20-26 | 37% stays with Quebec firms (49% Canadian) | The foreign half is the platform layer, with no domestic substitute |
The pattern is the reframe worth keeping: local capture is not one dial. It runs high in physical, place-based work and low in tradeable technology, in every market we can see, regardless of policy. And we can only say "every market we can see" because everywhere else, the contracts are dark.
The Rules Already Exist. Governments Just Ignore Them.
Here is the uncomfortable part: none of this needs a new law. The Canadian Free Trade Agreement, which every government in this piece has signed, already requires them to publish the awards. Every contract above the thresholds, roughly $130,000 for services with a separate limit for goods, is supposed to be posted with the winning vendor, the price, and what was bought. Most governments simply do not, the standing-offer channel makes it worse, and no one at Charlottetown promised to fix it. So the three reforms below do not reopen the agreement or write a new one. They make governments do what they already agreed to do.
A common data standard: the Open Contracting Data Standard is open-source and used by 50-plus governments. Ottawa can adopt it unilaterally and invite the provinces in.
A Canadian procurement passport: modelled on the EU's single procurement document, so a Nova Scotia firm qualified at home stops re-proving itself to bid in Alberta.
A genuine single point of access: one searchable window where a business in any province can find every open opportunity, the way Tenders Electronic Daily does for the EU.
Your Takeaways
If you run one of these systems: the A+ measures your rulebook, not your market. If you cannot export your own awards, you do not know who is winning your contracts, and neither can the firms you say you want to attract. Fix the standing-offer disclosure gap first; it is where the risk concentrates.
If you sell across a provincial line: the hardest markets to read are the ones with the least competition, because most of your rivals quit at the portal. And every market we measured buys foreign platforms and Canadian services, so if you sell technology, the services layer is where the door is open.
Quick poll: Have you ever walked away from a government opportunity in another province because you could not get a clear read on the market?
Yes, more than once
Once, and it was enough
No, we push through the portals
We only sell in our home province for exactly this reason
Have you ever won a contract in another province, not for the federal government?
Our Take
The report card is not wrong. It is answering an easier question than the one that matters. Grading whether a government has written the right rules is fair, and this year's grades reflect real progress. But procurement is where internal trade is either open or it is not, and on that question four of fourteen governments pass and ten do not. The barrier is no longer mainly a rule that bars out-of-province firms. It is a system so opaque that an outsider cannot find the work and the public cannot check who won. We can see Newfoundland, Toronto, and Quebec's tech market clearly because we did the excavation the portals would not. Every contractor in the country should be able to see every market that clearly, and every government should want them to. The rules are getting the attention. The contracts are still in the dark.
Know a contractor who should see their market this clearly? Forward them this issue, or send them to GovCon Weekly.
GovCon Weekly is written by the team at Publicus. We built the AI agent system that helps businesses find and win Canadian government contracts, and it worked well enough that the federal government came to us to use it on their own data. If you want to see opportunities like these before your competitors do, subscribe, or reach out for a demo.Canadian Procurement Pulse: Your Weekly Contractor Insider
Date
xx
