The Risk Is Not The Algorithm, It Is The Data It Is Fed

A Canadian business that licenses pricing software is usually buying a margin improvement, not a legal position. But the Competition Bureau spent 2025 examining exactly this category of product, opened and then closed an investigation into its use in one Canadian market, and has made it an enforcement priority. The distinction that determines whether a business is on the right side of this is neither obvious nor technical, and most buyers never ask about it.

The Distinction That Organizes Everything

Before the regulatory history, the analytical point that makes the rest usable.

There are two entirely different things sitting under the phrase "algorithmic pricing," and conflating them produces both unwarranted alarm and unwarranted comfort.

The first is a business using software to optimize its own prices from its own data: demand patterns, inventory, capacity, historical conversion, publicly observable competitor list prices. This is a sophisticated version of what pricing managers have always done, and its competitive character is generally benign or positive.

The second is a shared system into which multiple competing firms contribute non-public information, and which returns pricing recommendations to each of them. That is not primarily a pricing tool. It is an information exchange with a pricing interface, and information exchange among competitors has been a core competition law concern for a century, long before anyone called it an algorithm.

The Bureau's own framing of its rental housing concern lands precisely on this line: the concern with these types of pricing tools is that they allow users to set rents based on non-public, competitively sensitive data[1]. The operative words are non-public and competitively sensitive, not algorithmic.

Everything practical in this article follows from that distinction, and it is the reason the diligence questions at the end are about data provenance rather than about model architecture.

What Actually Happened In Canada

The sequence, which is more advanced than most Canadian businesses realize.

On June 10, 2025 the Competition Bureau opened a consultation as part of its inquiry into algorithmic pricing, seeking feedback on the prevalence of algorithmic pricing in Canada and the source of the data used in pricing algorithms[2]. The Bureau's own account states the consultation took place from June 10, 2025 to August 4, 2025, that it published a discussion paper for public consultation in June 2025, and that the purpose was to strengthen the Bureau's understanding of algorithmic pricing and how it might impact competition and consumers[3].

We note one discrepancy: a regulatory tracker records the consultation as closing on July 22, 2025[2], while the Bureau's own publication states feedback was accepted until August 4, 2025[3]. We treat the Bureau's own statement as authoritative.

The response volume is worth recording because it indicates who is paying attention. The Bureau received 103 responses: 77 comments from individuals and 26 longer submissions from other stakeholders including businesses, industry associations, members of the academic and legal communities, and consumer interest groups[3].

On market prevalence, the Bureau's discussion paper reports that in Canada, over 60 companies offer services that use algorithms and that claim to help companies optimize pricing, and defines algorithmic pricing broadly as the process of using automated algorithms to set or recommend prices for products or services, often in real time, based on a set of data inputs[4].

Enforcement posture matters too. The Bureau's 2025-2026 annual plan identifies that it will be launching a consultation and researching the use of algorithmic pricing to examine its impact on competition and consumers, alongside continued priority on greenwashing and drip pricing, and continued focus on the competitive aspects of consumer data including data portability[5].

The Rental Housing Investigation

The Canadian case that took this from theory to enforcement, and the only concrete domestic application available.

The Bureau commenced its inquiry in January 2025 in response to public concerns about housing affordability and whether landlords were colluding on rental prices by using algorithmic pricing software. It investigated whether landlords and property managers used algorithmic pricing tools supplied by RealPage Canada, Inc. and Yardi Canada, Ltd. in a manner that resulted in tenants paying higher rents than would otherwise have been the case, and considered whether this behaviour raised concerns under both the civil and criminal provisions of the Act[1].

Two features of that description deserve attention from any business using pricing software.

The description of how the software works is generic rather than specific to housing: algorithmic pricing software combines multiple data points, including users' competitively sensitive data, and supply and demand data, to calculate recommended pricing[1]. Nothing in that description is unique to rental property. The same architecture exists in hospitality, transportation, professional services scheduling, and any market where a vendor aggregates client data to improve recommendations.

And the Bureau considered both civil and criminal provisions. Canada's criminal conspiracy provisions carry consequences of a different order from civil reviewable conduct, and the fact that a pricing software arrangement was assessed against them is the single most important signal in this file for a business evaluating its own exposure.

Discontinued Is Not Exonerated

The outcome, and the correct reading of it.

On November 10, 2025 the Competition Bureau issued its findings following its investigation into the use of algorithmic pricing software in the rental housing market[2], and the Bureau discontinued the algorithmic pricing investigation and issued guidance[1].

It would be easy, and wrong, to read that as the Bureau concluding algorithmic pricing is fine. Three observations cut against it.

A discontinuance reflects a conclusion about the evidence in a specific investigation involving specific firms and a specific market. It is not a general ruling, and it establishes no safe harbour for a different business using a different tool with a different data architecture.

The Bureau issued guidance alongside the discontinuance[1]. Regulators do not publish guidance about conduct they consider unproblematic. Guidance is how an agency signals where the line sits, which implies there is a line and that conduct on the wrong side of it remains actionable.

And the broader workstream continued independently. The consultation ran through 2025 and the What We Heard report was published subsequently[3], with algorithmic pricing named in the annual plan as a research and enforcement priority[5]. A file the Bureau had finished with does not get a consultation, a report and a plan entry.

We have not obtained the text of the guidance issued, which is a real limitation of this article, and any business materially exposed should read it directly rather than rely on this summary.

The Bureau's Two Theories

The Bureau has articulated two distinct routes by which algorithmic pricing can become a Competition Act problem, and they are worth separating because they implicate different businesses.

The first is coordination without agreement. The Bureau has previously said algorithmic pricing could become an issue under the Competition Act if multiple companies rely on the same model to set pricing, creating a form of collusion on pricing even without explicit collaboration[6].

That theory does not require anyone to have met, spoken or agreed. It requires only that competitors converge on a common pricing engine. A business that never contemplated coordinating with rivals can find itself inside the theory purely by choosing the market-leading vendor, which is also the vendor its rivals chose for the same commercial reasons.

The second is targeted predation. The Bureau said the systems could also be used for predatory pricing if a company uses them to target specific customers of rivals, rather than lowering prices overall[6].

This theory is the mirror image of the first and implicates a different actor: a firm with market power using granular targeting to discipline competitors while leaving general price levels intact. It is worth noting because conventional predatory pricing analysis looks at overall price levels, and a tool that permits surgical discounting to a rival's specific customers may produce predatory effects that aggregate pricing data would not reveal.

Why The Law Is Older Than The Technology

A framing point on which the industry submissions and, in our reading, sound legal analysis agree.

The International Center for Law & Economics argues that the competitive effects of algorithmic pricing depend on three key factors, market structure, software design and implementation methods, that algorithmic adoption can facilitate coordinated outcomes but also offers consumer benefits through improved capacity utilization and dynamic pricing, that the same technology which intensifies price competition in some contexts can dampen it in others, and that these findings underscore that technology amplifies existing market characteristics rather than fundamentally altering competitive dynamics[7].

The Information Technology and Innovation Foundation makes a parallel argument about legal adequacy, submitting that the Bureau should adapt enforcement tools without stretching legal thresholds, resist calls to attack tacit collusion, require sufficient evidence of an anticompetitive agreement, and recognizing that the Competition Act is already equipped to address algorithmic collusion and abuse effectively, especially with improved detection tools, enhanced data access and increased analytic capacity that AI can help facilitate[8].

Both organizations advocate positions congenial to technology vendors and we flag that interest. But the underlying observation is analytically sound and useful to a business: the legal exposure here is mostly not novel. Agreements among competitors, information exchange, abuse of dominance and deceptive marketing are long-established categories. What has changed is the ease with which conduct can fall into them, not the categories themselves.

That has a practical consequence. A business cannot defend itself on the basis that the law has not caught up with the technology, because the relevant law was not written about technology at all.

The Vendor Is The New Vector

The genuinely new element, and this is our own analysis rather than a sourced finding.

Classical price-fixing required competitors to communicate. That requirement was itself a natural constraint: meetings leave witnesses, correspondence leaves records, and the coordination problem is hard. Algorithmic pricing removes the need for competitors to communicate with each other, because each communicates only with a vendor, and the vendor performs the aggregation.

From each participant's perspective nothing improper occurred. They licensed software, uploaded their own data as the contract required, and accepted or declined recommendations. No competitor was contacted. Yet the aggregate structure, non-public competitively sensitive data from rival firms pooled and returned as pricing guidance, reproduces the informational conditions of a cartel without any of its conventional evidence.

This is why we would argue the diligence burden has shifted onto the procurement decision. The moment of legal significance is not when a price is set; it is when a business signs a contract that contributes its non-public data to a pool it does not control and cannot see.

Most Canadian businesses evaluating pricing software assess accuracy, integration and price. In our assessment they should be assessing data provenance with equal seriousness, and the questions to ask are set out below.

Personalized Pricing Is A Different Problem

A second strand entirely, sharing the word "algorithmic" and almost nothing else.

The Bureau's discussion paper notes that personalized pricing is not restricted to online behaviour or the use of algorithms, but that advancements in AI have enabled businesses to implement personalized pricing more effectively by collecting and analyzing vast amounts of consumer data[4].

The academic submission by Chapdelaine, Larouche and Quaid focuses on algorithmic personalized pricing, defined as the use of algorithms to set individualized prices based on consumers' maximum willingness to pay. The authors argue it represents a transformative shift in market dynamics, analyze its potential to facilitate first-degree price discrimination, and contend that the pursuit of this goal by firms, even if not fully achieved, raises significant concerns for consumer welfare, market transparency, and the normative foundations of competition law. They examine how the strategies firms use to implement it, collusion or coordinated competitor conduct, abuse of dominance, and deceptive marketing practices, are likely to fit within the ambit of Canada's Competition Act, particularly given recent amendments[9].

The economics deserve stating plainly because the term "price discrimination" sounds like a legal conclusion and is actually a technical description. Under uniform pricing, buyers who would have paid more than the posted price capture the difference; that difference is consumer surplus. First-degree price discrimination means charging each buyer their own maximum willingness to pay, which transfers that surplus entirely to the seller.

This is why the authors' reference to the normative foundations of competition law is not rhetorical. Competition law has generally been understood to protect a process that delivers surplus to consumers. A technology whose logical endpoint is the complete extraction of that surplus, achieved through competition rather than despite it, poses a question the framework was not built to answer.

The Ethics Nobody Has Settled

The genuinely unresolved question, which a business adopting these tools should confront rather than delegate.

Personalized pricing means two customers pay different amounts for an identical product based on what a model infers about them. The inference inputs are typically behavioural and demographic proxies: device, location, browsing history, purchase patterns, time of day.

Several distinct objections arise and they are not the same objection. There is a transparency problem, because a customer cannot know they were charged more and therefore cannot respond. There is a proxy discrimination problem, because inputs correlated with protected characteristics can produce differential pricing along those lines without any variable naming them. There is a regressive-outcome problem, because willingness to pay can correlate with urgency and constrained alternatives, meaning the customer with fewer options may be charged more. And there is a consent problem, since the data enabling the inference was rarely provided for pricing purposes.

Our own view, offered as analysis and not as a legal position, is that businesses adopting personalized pricing frequently have not decided which of these they consider acceptable, because the capability arrives as a vendor feature rather than as a strategic choice requiring a policy. A business that would not consciously adopt a policy of charging more to customers with fewer alternatives should establish whether its pricing engine is doing so as an emergent property, and that is an empirical question its own data can answer.

We would also note the reputational asymmetry. Differential pricing that a customer discovers is materially harder to defend publicly than most pricing decisions, because it looks personal rather than commercial, and detection is becoming easier rather than harder.

The Case For The Defence

The counterargument deserves equal prominence, and it is not weak.

ITIF argues that distributional risk should be addressed through safeguards rather than bans, and that personalized pricing can expand access for price-sensitive consumers[8].

That argument has real force. Uniform pricing is not neutral; it excludes everyone whose willingness to pay falls below the single posted price. A seller able to price discriminate can profitably serve customers a uniform price would have shut out, which expands access and can increase total welfare. Student discounts, seniors' rates and regional pricing are all crude price discrimination, and they are generally regarded as socially positive.

The Bureau's own consultation heard the efficiency case too: dynamically setting or recommending prices creates several market efficiencies[3].

The honest position is that personalized pricing's distributional effect depends on which direction it runs. Charging less to the price-sensitive expands access. Charging more to the desperate extracts surplus from those least able to resist. The same technology does both, the difference is implementation, and a firm optimizing purely for margin will not naturally select the first.

The empirical evidence is not one-sided either. ICLE cites Buchholz et al. (2025), studying the European ride-hailing platform Liftago, whose auction-based mechanism lets consumers choose among drivers offering different combinations of price and wait time, and describes the resulting evidence as offering a more cautionary perspective on personalized pricing[7]. That an industry-aligned submission reports cautionary empirical findings on personalized pricing is worth noting precisely because of the source's orientation.

What The Bureau Heard

The four themes the Bureau summarized from 103 responses, which map the debate compactly.

The report records that dynamically setting or recommending prices creates several market efficiencies; that algorithmic pricing can lead to anti-competitive behaviour; that lack of data transparency could harm consumers, workers and competition; and that regulations should address anti-competitive conduct without stifling innovation. The report summarizes what stakeholders said and does not necessarily reflect the Bureau's views[3].

Two observations. The inclusion of workers alongside consumers and competition is notable and under-discussed, since the same architecture that sets prices for buyers can set rates for suppliers and labour, and the analytical concerns transfer.

And press coverage of the consultation reported that the Bureau heard affordability and privacy concerns[6], which locates this file at an intersection Canadian regulators have generally kept separate: competition policy and privacy law.

The Privacy Overlay

A dimension a Canadian business should not treat as belonging to a different department.

Personalized pricing works by collecting and analyzing vast amounts of consumer data[4], and privacy concerns were among what the Bureau heard[6].

The general point, which we state as observation rather than legal analysis, is that Canadian privacy law is built on purpose and consent: personal information is collected for identified purposes and used consistently with them. Data a customer provided to complete a transaction, receive support or use an account was not obviously provided so that a model could infer their price tolerance.

Whether a particular personalized pricing implementation is compliant is a question for privacy counsel and depends on the applicable statute, which may be federal or provincial depending on the business and the province, and on what the organization actually told people. This publication has examined the residency and cross-border dimensions of Canadian privacy law elsewhere.

The transferable point is one of governance rather than law: a pricing initiative that never passed a privacy review has an unassessed exposure, and the fact that the competition regulator heard privacy concerns in a pricing consultation suggests these files will not stay in separate boxes.

The Tacit Collusion Problem

The hardest legal question in the area, and one worth understanding because its resolution determines how much risk a business carries.

ICLE's submission identifies the crux as the distinction between conscious parallelism and actual agreement in algorithm-mediated markets[7], and ITIF urges the Bureau to resist calls to attack tacit collusion and to require sufficient evidence of an anticompetitive agreement[8].

The background is that conscious parallelism, firms independently observing each other and converging on similar prices, has generally not been unlawful, because it involves no agreement. Cartel prohibitions target agreements. Algorithmic pricing strains this because algorithms can achieve convergence faster, more precisely and more stably than human rivals watching each other, arriving at coordinated outcomes with no agreement in any conventional sense.

Regulators face a genuine dilemma. Requiring proof of agreement may leave real consumer harm unaddressed. Abandoning the requirement risks penalizing firms for independently rational responses to observable conditions, which is what competition is supposed to look like.

For a business, the practical reading is that this is unsettled and that the settled part is not. Whatever happens to tacit collusion doctrine, the exchange of non-public competitively sensitive information among competitors is already actionable under existing law. A business whose exposure runs through the unsettled doctrine has a genuine argument; one whose exposure runs through data pooling does not.

A Worked Case: Two Tools, One Question

A Canadian mid-market business in a concentrated regional sector evaluates two pricing platforms. The comparison is constructed to isolate the variable rather than reported from a specific engagement.

Tool A ingests the company's own transaction history, inventory, capacity and conversion data, plus publicly posted competitor list prices scraped from websites. It returns recommendations optimizing the company's own margin against observable market conditions. Every input is either the company's own or genuinely public.

Tool B offers materially better performance. Its advantage comes from a benchmarking database built from participating clients across the sector, including realized transaction prices, discount levels and occupancy figures that none of those firms publish. Several of the company's direct regional competitors are participants. The contract requires the company to contribute its own equivalent data.

On the Bureau's stated concern, that such tools allow users to set prices based on non-public, competitively sensitive data[1], and its theory that multiple companies relying on the same model can create a form of collusion on pricing even without explicit collaboration[6], the two tools sit in different places despite being sold in the same category and evaluated in the same procurement.

Nobody at the company would contemplate calling a competitor to discuss pricing. Signing Tool B's contract produces a comparable informational result through a vendor, and the performance advantage the buyer is paying for is substantially the value of that pooled non-public information.

That last observation is the uncomfortable one and it generalizes: where a pricing tool's edge comes from competitor data rather than from better modelling of your own, the thing you are buying is the thing that creates the exposure.

The Diligence Questions

What data sources feed the recommendation? Own data, genuinely public data, or pooled data from other clients. Get the answer in writing from someone accountable, not from marketing material.

Are any of those other clients my competitors? If the vendor cannot or will not say, treat that as an answer.

Is the pooled data public? Posted list prices are public. Realized transaction prices, discount rates, occupancy and margin are generally not.

Does my contract require me to contribute data, and where does it go? Contribution is the mechanism by which you join the pool rather than merely observe it.

Does the tool set prices or recommend them, and do we deviate? Documented independent judgment is worth more than a policy nobody follows, and a business that accepts every recommendation has weaker evidence of independent conduct.

Are we personalizing, and on what inputs? If prices differ between customers, know which variables drive it and test whether the outcome correlates with anything you would not defend publicly.

Has privacy reviewed this? If personal data feeds pricing, the purpose and consent question is live regardless of the competition analysis.

Read the Bureau's guidance directly. It was issued alongside the November 2025 discontinuance and is the most authoritative available statement of where the line sits.

The Limits Of This Analysis

Several caveats matter. We did not obtain the text of the guidance the Bureau issued alongside its November 2025 discontinuance, which is the most important primary document in this area, and readers should read it directly. Sources disagree on the consultation closing date, reported as both July 22 and August 4, 2025; we have followed the Bureau's own statement. The What We Heard report summarizes stakeholder views and expressly does not necessarily reflect the Bureau's own position, so nothing in it should be read as regulatory guidance. Two of the submissions cited, from ICLE and ITIF, are from organizations advocating positions generally favourable to technology firms, and we have flagged this; the academic submission by Chapdelaine, Larouche and Quaid is cited from its abstract rather than its full text. The distinction between own-data optimization and pooled-data exchange, the vendor-as-vector argument, and the ethical framing in this article are our own analysis rather than sourced findings or legal conclusions. This article does not address specific Competition Act provisions, penalties, the private right of action, the criminal conspiracy threshold, or how any particular arrangement would be assessed. Nothing here is legal advice; a business using or considering algorithmic pricing tools should obtain Canadian competition law advice on its specific arrangements.

Related Reading

Compliance Competition Act Greenwashing And Drip Pricing Rules The Bureau's other standing enforcement priorities, and the substantiation standard behind them. AI & Automation In Finance When The Algorithm Becomes The CFO: Ethics And Automated Risk The governance question underneath any model that makes commercial decisions on your behalf. AI & Automation In Finance Does Your Data Have To Stay In Canada? PIPEDA, AI Tools, And The Residency Myth The privacy framework that personalized pricing engages whether or not anyone consulted it.

Frequently Asked Questions

Is algorithmic pricing legal in Canada?

Using algorithms to set your own prices from your own and public data is ordinary business practice. The Bureau's concern is narrower: that such tools can let users set prices based on non-public, competitively sensitive data, and that multiple firms relying on one model can produce collusion-like outcomes without explicit collaboration. The distinction is data provenance, not automation.

What happened with the rental housing investigation?

The Bureau began inquiring in January 2025 into whether landlords using RealPage Canada and Yardi Canada tools caused tenants to pay higher rents, assessing both civil and criminal provisions. On November 10, 2025 it issued findings, discontinued the investigation and issued guidance. A discontinuance addresses specific facts and firms; it is not a general clearance, and regulators do not publish guidance about conduct they consider unproblematic.

Can I be liable without ever contacting a competitor?

That is the Bureau's stated theory: if multiple companies rely on the same model to set pricing, it could create a form of collusion on pricing even without explicit collaboration. This is why we argue the moment of legal significance is the procurement decision rather than the pricing decision, since that is when a business joins a data pool it does not control.

What is first-degree price discrimination?

Charging each buyer their individual maximum willingness to pay, which transfers consumer surplus entirely to the seller. Academic submissions to the consultation argue that pursuing this, even without fully achieving it, raises concerns for consumer welfare, market transparency and the normative foundations of competition law.

Isn't personalized pricing sometimes good for consumers?

Yes, and that argument was made in the consultation: personalized pricing can expand access for price-sensitive consumers, and dynamic pricing creates market efficiencies. Uniform pricing excludes everyone below the posted price. The distributional effect depends on direction, charging less to the price-sensitive expands access while charging more to the constrained extracts surplus, and margin optimization alone will not select the first.

What should I ask a pricing software vendor?

What data sources feed the recommendation; whether any contributing clients are your competitors; whether the pooled data is genuinely public rather than realized transaction prices, discounts or occupancy; whether your contract obliges you to contribute; and where that data goes. Get it in writing from someone accountable. If the vendor's performance edge comes from competitor data rather than better modelling of your own, that edge is the exposure.

References

1. Norton Rose Fulbright. Competition Bureau Discontinues Algorithmic Pricing Investigation And Issues Guidance, on the January 2025 inquiry, RealPage Canada and Yardi Canada, the non-public competitively sensitive data concern, and consideration of civil and criminal provisions. nortonrosefulbright.com/en/knowledge/publications/a793cfb2
2. Digital Policy Alert. (2025, November 10). Canada: Competition Bureau Inquiry Into Algorithmic Pricing, on the June 10, 2025 consultation opening, the consultation scope, and the November 10, 2025 findings. Note: records the closing date as July 22, 2025, which differs from the Bureau's own statement. digitalpolicyalert.org/change/15047
3. Competition Bureau Canada. (2026). Consultation On Algorithmic Pricing And Competition: What We Heard. Government of Canada, on the June 10 to August 4, 2025 consultation period, the 103 responses and their composition, and the four summarized themes. competition-bureau.canada.ca/.../consultation-algorithmic-pricing-and-competition-what-we-heard
4. Competition Bureau Canada. (2025). Algorithmic Pricing And Competition: Discussion Paper. Government of Canada, on the definition of algorithmic pricing, the finding that over 60 companies in Canada offer algorithmic pricing optimization services, and personalized pricing enabled by AI and consumer data. competition-bureau.canada.ca/.../algorithmic-pricing-and-competition-discussion-paper
5. Norton Rose Fulbright. Canadian Competition Bureau Releases 2025-2026 Annual Plan: Signals More Aggressive Enforcement Ahead, on algorithmic pricing as a research priority alongside greenwashing, drip pricing and consumer data. nortonrosefulbright.com/en/knowledge/publications/a178c41c
6. The Canadian Press. (2026, January 22). Competition Bureau Hears Of Affordability, Privacy Concerns Over Algorithmic Pricing, on the Bureau's collusion-without-collaboration and targeted predatory pricing theories. cp24.com/news/canada/2026/01/22/competition-bureau-hears-of-affordability-privacy-concerns-over-algorithmic-pricing
7. Manne, G. A., Auer, D., Albrecht, B., Gilman, D. J., & Radic, L. (2025, August 4). ICLE Comments To The Canadian Competition Bureau On Algorithmic Pricing And Competition. International Center for Law & Economics, on market structure and software design, conscious parallelism versus agreement, technology amplifying existing characteristics, and Buchholz et al. (2025) on Liftago. Note: an organization advocating positions generally favourable to technology firms. laweconcenter.org/resources/icle-comments-to-the-canadian-competition-bureau
8. Information Technology and Innovation Foundation. (2025, August 8). Comments To Competition Bureau Of Canada Regarding Algorithmic Pricing And Competition, on resisting tacit collusion theories, the adequacy of the existing Act, evaluating vertical pricing tools by effect, and safeguards rather than bans. Note: an organization advocating positions generally favourable to technology firms. itif.org/publications/2025/08/08/comments-competition-bureau-of-canada-regarding-algorithmic-pricing-competition
9. Chapdelaine, P., Larouche, P., & Quaid, J. (2025). The Anti-Competitive Effects of Algorithmic Personalized Pricing and the Big Data Economy. SSRN. Submission to the Competition Bureau consultation, on first-degree price discrimination and the normative foundations of competition law. Cited from the abstract. doi.org/10.2139/ssrn.5399219

This article discusses competition policy and consultation submissions and is provided for general informational purposes. It is not legal advice. The Bureau's guidance issued in November 2025 was not obtained for this article and should be read directly. The What We Heard report summarizes stakeholder views and does not necessarily reflect the Bureau's position. Obtain Canadian competition law advice on any specific pricing arrangement.