Uncertainty and the Burden of Proof in Canadian Competition Law

The Study in Brief

• There are good reasons to take stock of Canadian competition law. The vulnerability of digital markets to market power stemming from network externalities and scale economies encourages reflection on whether the Competition Act continues to be suitable for present times.

• Recently, a number of statutory amendments have been proposed to amend the Act, some have been tabled in Parliament and still others already adopted. The federal government recently passed consequential amendments that grant the Minister of Innovation, Science and Economic Development (ISED) the power to initiate market studies, to include scrutiny of vertical agreements as possibly anti-competitive collaborations, to repeal the efficiencies defence to mergers, and to lower the burden of proof in abuse of dominance cases.

• Many of the government’s actions to date sensibly strengthen the enforcement powers of the Competition Bureau and make it easier for private actors seeking redress for allegedly anti-competitive behaviour.

• There are, however, other actual and proposed amendments that imply profound changes to the fundamental posture of Canadian competition law. In particular there are actual and proposed amendments that move away from the bedrock principle that the burden rests with the Bureau to prove, on a balance of probabilities, that a merger or practice by a dominant firm is likely to be or is anti-competitive.

• While enhancing enforcement is welcome, legislative amendments that lower the burden of proof are a mistake.

The author thanks Anthony Niblett, Ralph Winter, Paul Johnson, William B.P. Robson, Michael Trebilcock and four anonymous referees for helpful conversations and/or comments. The author also thanks the TSE Chair in Capital Markets at the Faculty of Law, University of Toronto, the C.D. Howe Institute and the Canadian Chamber of Commerce for financial support, and Matthew Holmes of the Canadian Chamber of Commerce for organizing online discussions with, and comments from, Anita Banicevic, Erika Douglas, Hilary Jennings, Jorge Padilla, John Pecman, Tom Ross and Danny Sokol, whose feedback was very helpful. The views expressed in this paper are the author’s alone.

1.Introduction

After decades of not having much in the way of political salience or attention, competition law now is front page news. There are several explanations for this. The rise of large technology firms has created unease and political hostility.1 Inflation in the recent past (in groceries, for example) has increased political attention on competitive conditions in Canada, whether or not this is justified empirically.2 Populist political movements in Canada and elsewhere have shifted their gaze to competition law.3 Various studies by the OECD4and others suggesting that Canada’s productivity growth has been weak are also politically relevant. Finally, and most substantively, there are studies abroad and in Canada suggesting that markets are becoming less competitive over time.5Recently, the Competition Bureau of Canada released a report on competition across the Canadian economy, Competition in Canada from 2000 to 2020: An Economy at a Crossroads,6 that concludes competition has worsened. It adds that the results “show how essential it is to modernize Canada’s competition law to respond to the realities of today’s economy.”7

While the Bureau’s study is a useful contribution to our understanding of the Canadian economy, I am skeptical about the relationship between these studies on competition trends and efforts to reform competition law. For one thing, studies of competition, including the Bureau’s, have critical shortcomings. For example, they rely on industry classifications rather than markets when examining concentration.8 As well, they exclude exports and imports, a particularly significant omission when examining the Canadian economy. The Bureau acknowledges many of these weaknesses,9yet also reasonably points to the report’s robust findings across various metrics about declining competition in pushing for competition law reform. But even if there were good evidence that competition has diminished in Canada, it does not follow that competition law reform is the appropriate remedy.

Assuming competition has worsened in Canada, there are several remedial policies that I suspect would be far more important than competition law reform. The OECD ranks Canada near the worst internationally in establishing regulatory barriers to competition.10To take one example of harmful regulation, supply management anti-competitively increases prices in agricultural staples by design and, for whatever reason, is not on the government’s reform table. Regulation, internal trade barriers, restrictions on international competition and ownership, and other policies are all important contributors to reducing competition in Canada and, certainly in their collective impact, are more important than competition law.

Moreover, even if there were problems with competition that relate to competition law specifically, it does not follow that law reform is appropriate. Rather, it could be that enforcement efforts have been inadequate.

That being said, whether or not some of the push for reform relies on questionable premises, there are good reasons to take stock of Canadian competition law. The vulnerability of digital markets to market power stemming from network externalities and scale economies encourages reflection on whether the Competition Act continues to be suitable for present times.11 The Bureau’s report, while not dispositive, suggests that markets may be becoming less competitive. Moreover, any piece of legislation can be improved. I am skeptical of the narrative that the law requires sweeping reform to address the digital economy or to reverse a strong, secular decline in competition caused by competition law, but I am not skeptical that there is room for improvement.

There have been several recent efforts to reform the Competition Act. A number of statutory amendments have been proposed, some have been tabled in Parliament and still others already adopted. The federal government recently passed consequential amendments that grant the Minister of Innovation, Science and Economic Development (ISED) the power to initiate market studies that treat vertical agreements as possibly anti-competitive collaborations, that repeal the efficiencies defence to mergers, and that amend the abuse of dominance provisions substantially.12 In its 2023 Fall Economic Statement, the government proposed a variety of significant changes to competition law, including expanding private rights of action for damages under the Act’s civil provisions, extending limitation periods on non-notified mergers and abolishing a provision that prevented the Competition Tribunal from disallowing a merger on the basis of market share alone.13 Bill C-59, if passed, would implement many of the proposals in the Fall Economic Statement.14 These and other amendments have the potential to change specific approaches to specific antitrust questions, with a significant emphasis on process. Many of the changes tend to strengthen the enforcement powers of the Bureau and make it easier for private actors seeking redress for allegedly anti-competitive behaviour. As I will explain in this Commentary, the government’s efforts to date to support enforcement are sensible and welcome developments.

There are, however, other actual and proposed amendments that imply profound changes to the fundamental posture of Canadian competition law. In particular, there is a family of amendments and proposals that move away from the bedrock principle that the burden rests with the Bureau to prove, on a balance of probabilities, that a merger or practice by a dominant firm is likely to be or is anti-competitive. For example, the amended abuse of dominance provisions no longer require a showing of a prevention or substantial lessening of competition (“SPLC”) prior to a Tribunal order, and the leader of the New Democratic Party (NDP), which is supporting the governing Liberals in Parliament, has also proposed significant changes to the Competition Act that would eliminate a competitive assessment in enforcement against mergers.15

Because of their significant potential to disrupt Canadian law and the continuing uncertainty about possible next steps for statutory reform in the short-to-medium term, these changes that reduce or eliminate the burden of proof on the Bureau to demonstrate anti-competitive effects are the focus of this Commentary.16

Many of the recent actual and proposed changes to the burden of proof were initiated by a 2022 discussion paper,17 in which ISED solicited feedback on a number of possible reforms to the Canadian Competition Act. That feedback is discussed in a follow-up report.18 ISED raises a variety of possible changes to the requirement that the Bureau must prove, on a balance of probabilities, that a merger or alleged abuse of dominance would cause a prevention or substantial lessening of competition. ISED states that the: “Bureau may not be able to take action against potentially harmful forms of conduct because of the specific legal tests to be met. While overenforcement is not desired, the field cannot be tilted too steeply against necessary intervention if an effective watchdog is to function.”19

ISED identifies some legitimate concerns about the status quo, but in my view changing the burden of proof is a mistake. I first consider mergers policy. Some of the specific concerns ISED raises about the status quo do not suggest reform to the burden of proof, but instead suggest reform to other aspects of the legal and enforcement landscape. Even if they were on point, ISED’s proposed changes present more drawbacks than they resolve. As I will argue, rather than lowering the burden of proof, the better calibrated approach is to build enforcement capacity, which the government has done, including through statutory amendments (e.g., extending the limitations period for some mergers), increasing the Bureau’s budget, and amending legislation to incentivize private actions by providing for damages for abuse of dominance. Lowering the burden of proof would predictably do more harm than good.20

ISED also contemplates changes to the burden of proof in abuse of dominance that I discuss and critique. The government recently amended the Act to abolish the requirement of a SPLC for a Tribunal order against a dominant firm in the abuse context, and to create the possibility of damages for aggrieved parties. While enhancing enforcement through money damages was appropriate, the change to the burden of proof was, in my view, a significant mistake. In light of the challenges of enforcement in dynamic markets, enhancing enforcement in general is appropriate; lowering the burden of proof is not.

The Commentary proceeds as follows. Section 2 discusses enforcement against mergers, while Section 3 examines abuse of dominance. Section 4 concludes the Commentary.

2. Difficulties of Proof in Merger Analysis

a) Uncertainty, Quantification and the Cost
of Proof

ISED raises a variety of possible disadvantages of the current requirement that the Competition Bureau prove, on a balance of probabilities, that a merger would prevent or substantially lessen competition. It proposes considering whether intervention ought to arise if a merger poses an “appreciable risk” of anti-competitive harms. ISED states:

“[I]t would seem that there are at least two possible substantive challenges to applying the merger provisions’ competitive effects test to acquisitions in fast-moving digital markets. The first concerns where harms to non-price dimensions of competition, such as innovation, may be difficult to quantify and are, accordingly, given less weight by the Competition Tribunal or appeal courts. The second challenge is the substantive requirement that the Bureau show, on balance of probabilities, that harm to competition is ‘likely’ to happen within a ‘discernible’ time frame, and that this harm would likely be ‘substantial.’ Given the complexity, dynamism and pace of change in many markets, especially digital ones, these specific tests may be highly impractical.”21

This section argues that both these challenges identified by ISED are tenuously connected to the burden of proof. I consider quantification first and then the general problem of proof in dynamic markets.

i) Quantification

ISED is justified in concluding that the courts and the tribunal display a preference for quantification of the anti-competitive harms of a proposed merger, and that quantifying such harms is especially difficult when the merger’s potential anti-competitive effects concern innovation and/or other non-price effects. In 2015, the Supreme Court in Tervita Corp. v. Canada concluded that if merging parties invoked the then-existing efficiencies defence to mergers found in Section 96 of the Competition Act, the Bureau would be required to quantify all reasonably quantifiable effects of the merger before the onus of proving offsetting efficiency gains would shift to the merging parties.22This implies that merging parties need only invoke the efficiencies defence to require the Bureau to quantify anti-competitive effects, which may be difficult to do convincingly. This particular concern is no longer relevant given that the government has abolished the efficiencies defence to mergers with the passage of Bill C-56 in 2023. But the Supreme Court’s reasons in that case clearly favour quantified evidence, denigrating non-quantified evidence as “more subjective,” for example.23

The importance of the distinction that the Court draws between quantitative and qualitative evidence in a non-efficiencies-defence context is not entirely clear, but it is reasonable for ISED to suggest that it potentially presents enforcement challenges for the Bureau, perhaps especially in dynamic markets. For example, if the Tribunal accepted the Supreme Court’s apparent view that qualitative evidence is problematically subjective, it may put less weight on qualitative evidence to show a SPLC, yet the Bureau would have a difficult time proving quantified deadweight losses from the absence of innovation.24

In light of the quantification difficulties in dynamic markets, and the preference for quantitative evidence expressed by the Supreme Court, ISED asks whether changing the burden of proof would make sense. For example, it cites a recent US legislative proposal that a merger should be stopped if there is an “appreciable risk” that the merger would be anti-competitive.25Following this principle in Canada would be a mistake. Aside from the substantive shortcomings of such a standard in its own right, that I discuss below, it is also disconnected from the preference for quantitative evidence.

To explain further, even if the standard is an appreciable risk of anti-competitive effects, it may be difficult to prove quantitatively the cost of that appreciable risk. That is, lowering the burden does not eliminate the quantification preference. Moreover, if quantification is the problem, then quantification ought to be the target of reform. My view is that quantification ought not to be required or even necessarily preferred.26For one thing, a principled boundary between what is quantifiable and what is not quantifiable does not exist, given that all effects may, in principle, be quantified.

Moreover, qualitative evidence may be every bit as objective as quantitative evidence. Indeed, the assumptions that underlie quantitative predictions (e.g., the shape of the demand curve) may be just as subject to judgment as many qualitative pieces of evidence (e.g., the qualitative similarity of the merging firms’ products). The trier of fact ought not a priori to prefer one kind of evidence over another, but ought to weigh all the evidence based on the circumstances.

The solution to a misplaced emphasis on quantification is not to change the burden of proof, but to address the quantification problem directly. I would support statutory reform to abolish any a priori preference for quantitative evidence in merger review or any other context.27 If the trier of fact concludes that qualitative evidence is not as persuasive in a particular case, then so be it, but it ought not to be treated that way as a categorical matter.

ii) Difficulties of proof in dynamic markets

Also in support of the “appreciable risk” standard, ISED observes that it may be difficult to prove that anti-competitive effects are more likely than not in dynamic, innovative markets that are inevitably laden with uncertainty.

It is worth considering two alternative understandings of ISED’s reasoning. It could be that the future effects of a merger are essentially unknowable. Or it could be that it is possible to reach reasonably accurate conclusions about the impact of a merger, but doing so is highly complex and, therefore, costly. I will return to the first possibility in the next sub-section, but consider here the plausible argument that it is costly for the Bureau to prove anti-competitive effects of a merger in a dynamic market.

If it is costly and challenging to prove anti-competitive effects in dynamic markets, it is a non sequitur to lower the burden of proof to something like “appreciable risk of” anti-competitive effects. If proof has become more costly to come by, given the considerable economic stakes in getting enforcement right, the logical approach is to increase enforcement resources.28 Indeed, the federal government has wisely increased the Bureau’s budget to address this situation.29 Moreover, the Bureau has moved to create a unit with particular expertise in digital markets, something that also makes sense.

ISED also discusses expanding the scope for ex-post remedies for mergers, perhaps extending the limitations period from one year, post-closing, to three years for the Bureau to challenge a consummated merger.30Bill C-59 follows up on this suggestion and would extend the limitation period to three years for non-notified transactions. This extension is especially relevant for acquisitions of nascent competitors, many of which are small transactions that are not notifiable. (The acquisition of nascent competitors provides a significant enforcement challenge; I return to this later.)

While there are clearly many considerations that affect the decision whether to adopt such an extension, including the challenges of imposing ex-post remedies on mergers, uncertainty about the competitive impact of a merger favours it. More information will be revealed over time after the merger takes place, diminishing uncertainty and the corresponding difficulties for the Bureau to meet its burden of proof.31Additionally (or perhaps alternatively), ISED considers expanding notification requirements to capture some mergers that are not currently notifiable. This would also provide the Bureau with more information that might be helpful to it.

Finally, the authority to order market studies, something the government established with the passage of Bill C-56, might also provide informational advantages. Market study powers, if exercised appropriately,32 would, among other things, increase the potential for the Bureau to learn more about the effects of past merger decisions, which would inform it better about future decisions, as well as about especially competitively fraught industries in which mergers might be proposed.33

In short, not only does reducing the burden to “appreciable risk” not follow from an increase in the cost of becoming informed about a merger’s competitive effects, but there are alternatives for change that would reduce the costs of gathering information.

b) Irresolvable Uncertainty About Competitive Effects

While some uncertainty about the competitive impact of mergers can be addressed by greater investment in enforcement, including ex-post enforcement, some residual uncertainty is inevitable. ISED is surely correct to conclude that dynamic markets in the digital age increase uncertainty. An increase in enforcement efforts would not necessarily resolve this uncertainty. Such an increase would, however, minimize residual uncertainty, and better target the enforcement challenges of market power and uncertainty than lowering the burden of proof.

i) Nascent competitors

An example of the thorny difficulties posed by irresolvable uncertainty in the mergers context concerns the acquisition of nascent competitors.34 In less dynamic markets, the acquisition of a small, new competitor by an established, prominent firm would typically not raise competitive concerns: the acquisition of a trivial competitor is generally not competitively problematic. In dynamic technology markets, in contrast, a small, insignificant competitor could become a prominent, significant competitor in a very short period of time. From a competition-law enforcement perspective, however, there is likely to be considerable uncertainty whether the nascent competitor would evolve into a competitive force and, thus, whether the merger is problematic.

Notwithstanding the enforcement difficulties involved in acquisitions of nascent competitors, lowering the burden of proof to “appreciable risk” or something analogous would be a mistake. The overwhelming problem with this standard is that it is too easy to meet and fails to distinguish anti-competitive from benign conduct, a concern that ISED acknowledges in its consultation report What We Heard.35To illustrate the point, it is helpful to consider a concrete case: the US Federal Trade Commission’s (FTC) recent challenge to Meta’s acquisition of Within. Newcomer Within produced a virtual reality (VR) fitness app, while Meta did not. But it was possible that Meta, which has identified VR as a priority, might have created a VR fitness app competitor to Within. It is also theoretically possible that Within could have grown from providing a fitness app to becoming a dominant player in VR applications (and hardware, for that matter) and competing with Meta.

In February 2023, the Federal District Court rejected the FTC’s request for a temporary injunction blocking the merger as resting on “impermissibly speculative” evidence about its harms to potential competition.36While the District Court did not accept that the FTC had shown that it was “reasonably probable” that the merger would harm competition – the US standard that governed the case37 – it is at the very least arguable that the FTC established that the merger raised an “appreciable risk” that the merger would be anti-competitive. This would likely be true of any acquisition of a VR app firm by Meta.

Meta/Within illustrates the shift in enforcement that would take place if “appreciable risk” were adopted as the standard: while the standard would require interpretation, its plain meaning (which drives statutory interpretation in Canada) implies that almost any merger involving a firm with market power in dynamic markets could be challenged. While I do not know any more than the District Court whether the merger will harm competition sometime in the future, stopping such mergers would predictably be socially costly. While some mergers caught by the standard may well have future anti-competitive effects, most would not since acquisitions of nascent firms typically benefit both firms and consumers. For example, the established firm often commercializes the technology of the acquired firm to provide even better products to its customers. This may be especially valuable if the startup has expertise in invention, but not in scaling.38 Stopping such acquisitions would also have harmful dynamic effects, as entrepreneurs who might innovate with the primary goal of selling to a larger firm would anticipate competition law issues from such mergers – even in the absence of likely anti-competitive effects – and thus have diminished incentives to innovate.39

For its part, ISED identifies a theoretically superior approach to the acquisition of nascent competitors. Under the “balance-of-harms” test, which was proposed by a 2019 UK Digital Expert Panel, enforcement authorities should conduct a full cost-benefit analysis of an acquisition.40Following this approach, the Competition Bureau would consider all the possible competitive outcomes from an acquisition, determine the social effects under each outcome and determine the probability of each outcome. Using these inputs, the Bureau would calculate the acquisition’s expected social impact.

While this is theoretically a reasonable approach – cost-benefit analysis is generally a sensible approach to public policy – there are practical problems that render it unsuitable. The balance-of-harms test does not require the future to be known, but it does require the distribution of outcomes to be known. That is, what states of the world might exist, what probability attaches to each state of the world and the costs/benefits associated with each.

In contrast, the conventional test simply asks whether a lessening of competition is more likely than not, which imposes much less of an informational burden on the Bureau.41 Moreover, there is a significant risk of error associated with the informational problems with the balance-of-harms test. It is too informationally intensive for practical implementation. Indeed, both the German and UK authorities have considered and rejected the balance-of-harms test because of its impracticality.42

Another important consideration weighing against lowering the burden of proof is the availability of alternative means of strengthening enforcement. For example, the government has increased the Bureau’s budget, which would better allow the Bureau to invest in reducing uncertainty about the competitive implications of a particular merger. Canada could follow the European Union and expand pre-notification requirements better to inform the Bureau about potentially anti-competitive acquisitions.43Requiring dominant platforms to pre-notify even more extensively than other firms could be one way to do this. As well, extending the limitations period beyond one year, something proposed in Bill C-59 for non-notifiable transactions (which would have a three-year limitation period), would also allow the Bureau to consider a merger ex post with better information about its competitive effects.44 Yet another approach to the acquisition of nascent competitors is to rely on abuse of dominance, if a dominant firm has made a practice of acquiring nascent competitors.45

While these other legislative and enforcement options to address uncertainty and the acquisition of nascent competitors exist, it is inevitable that uncertainty would continue to plague review of acquisitions of nascent competitors. In my view, this is acceptable because there is no better option.

ii) Uncertainty and the law’s conventional reluctance to intervene

Aside from the caution that results from the balance-of-probabilities burden-of-proof, the law offers many other examples of a reluctance to intervene when the social benefit of intervention is unclear. For instance, the law declines to intervene in the face of the so-called “oligopoly problem,” despite the social harms it creates. This refers to the high probability of anti-competitive pricing in a concentrated oligopoly. Even if a small number of firms recognize their interdependence and, as a consequence, fail to compete vigorously, Canadian law will not seek to punish this behaviour.46 This is because of uncertainty and the concern that intervention would do more harm than good. Fundamentally, there is uncertainty about what the appropriately competitive price would be, which makes both detection of problematically high prices and remedies for high prices difficult for the authorities to determine.47 Rather than attempting to impose price regulation in oligopolies, which is costly and error-prone,48the law simply accepts that some anti-competitive outcomes will arise because no sensible intervention is available.

This backdrop begs the question of those calling for intervention against mergers even in the face of uncertainty about the social benefits of intervention: What has changed so that the law’s presumption against intervention should change?

Technological, innovative and digital markets often have characteristics such as scale economies and network externalities that increase the risk and durability of market power.49 As noted above, setting aside the impact of uncertainty, an increase in the risk of market power tends to invite stricter antitrust enforcement. The government is right to have invested in the Bureau’s budget, for example, and the Bureau is to be commended for creating a unit specializing in technology markets. But an increase in market power does not necessarily say anything about why burdens of proof ought to change. If market power is more prevalent, then there is a greater risk that mergers will be anti-competitive and the authorities will have more cases to bring; this justifies increasing funding to the Bureau. But a shift in the frequency of anti-competitive mergers does not logically imply a necessary lessening of the standards to prove anti-competitive mergers.

It may also be true that dynamic markets increase uncertainty about the competitive and social impact of mergers. But this, too, does not necessarily imply that standards ought to change. It is a non sequitur to say that an era with more uncertain outcomes requires a different approach to uncertainty.

The strongest case for lowering the standards for intervention arises if there has been an increase in mergers with anti-competitive implications and more uncertainty. Even if both are true, such that the average merger is more likely to be anti-competitive yet it is hard to prove that any given merger is anti-competitive, it does not necessarily follow that reducing the burden of proof is appropriate. Even if more mergers are likely to be socially harmful in highly dynamic digital markets, it may also be that there are increased social benefits in the pro-competitive mergers that take place in such markets; on net, blocking more uncertain mergers may do more harm than good. For example, blocking the acquisition of a nascent competitor in the name of preventing uncertain, possibly anti-competitive effects in the future, would in many cases prevent the beneficial incorporation of a new product into an incumbent’s product suite.

Moreover, there are alternative and better calibrated means to address uncertainty and market power than reducing the burden of proof. A reduction in the burden of proof for mergers would presumably apply to all mergers whether or not they are in sectors in which market power and uncertainty have become more problematic. This implies more intervention with respect to mergers in markets that are not dynamic, uncertain or especially prone to market power. In contrast, taking steps to address market power and uncertainty targets problematic mergers in challenging dynamic contexts directly. If increased market power, dynamism and uncertainty pose the enforcement challenge, enforcement responses should focus on market power, dynamism and uncertainty.

For example, the increased enforcement resources that the government wisely allocated to the Bureau will mitigate uncertainty. The new authority to order market studies could also mitigate uncertainty, as could the Bureau’s establishment of a unit in the Bureau with specific expertise in technology markets. And there are several non-legal and legal factors, including recent changes to the Competition Act, that could mitigate harm from a merger that turns out to be anti-competitive. For example, as noted above, ex-post review, as proposed in Bill C-59 for non-notifiable mergers, could address mergers that reveal themselves to be anti-competitive over time. Ex-post review targets market power and uncertainty specifically, unlike lowering burdens of proof, which would apply across all mergers.

In addition, if the merger were somehow to increase dominance, the law is equipped to address anti-competitive conduct at that time.50While not all anti-competitive conduct is caught by the abuse provisions, as ISED rightly points out, the fact that a significant subset of such conduct is covered by abuse of dominance cautions against early intervention opposing mergers when there is significant uncertainty. Greater reliance on abuse also makes sense given the government’s recent investment in the Bureau’s budget, which will better equip it to bring abuse complaints, as well as Bill C-59’s establishment of the possibility of damages for abuse: both predictably will increase enforcement against abuse of dominance. (The government also unfortunately lowered the burden of proof for abuse; while the possibility of money damages makes sense, lowering the burden of proof does not, as I discuss further below.)

Unlike the blunt choice to make challenging all mergers easier by lowering the burden of proof, these enhanced enforcement alternatives specifically target and mitigate the harms of market power and uncertainty. For the residual concerns about uncertainty and enforcement that remain, it is not clear why the law should depart from its usual cautious approach to intervention. Aside from the costs of blocking socially beneficial mergers, it is important to appreciate that if a merger creates market power, there will be incentives for entry and more competition.51 To be sure, market power may be durable,52perhaps especially where firms with market power take anti-competitive steps to deter competition. Concerns about market power’s durability may be particularly strong in technology markets that are often associated with scale economies and network externalities. But, on balance, the threat of entry combined with legal alternatives, including ex-post review, continue to provide good reason for caution about intervention ex ante without a clear basis for intervention. These alternatives preserve potential gains while mitigating potential harms.53

There is another important consideration. If intervention were allowed even in the face of speculative harms, this would not sit well with the rule of law. Under the “appreciable risk” test, for example, competition authorities would have broad discretion to intervene in a very wide swath of merger cases. Such widespread intervention would not be practical, and the authorities would have to exercise discretion over which cases to bring. Aside from getting bad social outcomes in individual cases with lower standards, there would be a danger that political influence would play a role in case selection.

As noted above, in recent years political actors have taken a newfound interest in competition law, perhaps driven in significant part by the size and social influence of large tech firms, but the interest is not confined to tech. For example, Canadian parliamentarians across the political spectrum have recently scolded grocery executives for high grocery prices blaming, without careful evidence, inflation in this sector on the exercise of market power.54This example renders plain the political risks of lowering standards to allow easier intervention in the face of growing uncertainty: broad discretion would risk leaving enforcement subject to prevailing political winds rather than careful competitive analysis.55

To close the discussion on mergers, there are recent proposals to put much greater emphasis on market shares in evaluating mergers as a substitute for demonstrating anti-competitive effects. Bill C-352 would profoundly change enforcement in Canada by eliminating the requirement for the Bureau to demonstrate a prevention or substantial lessening of competition (SPLC) in a significant subset of cases.56 The Bill takes two approaches. The less radical is to provide that a merger that leads to a market share of between 30 percent and 60 percent would be presumptively subject to an order. Therefore, a merger in which a 30 percent competitor acquires a one percent competitor faces the threat of intervention, unless it can prove affirmatively that the merger is competitively benign, for example by leading to the same or lower prices.

There are many reasons why this legislative emphasis on market shares – something that the Competition Bureau supports – rather than a SPLC would be a mistake, including grossly misplaced confidence in the precision of market shares. Market definition is often highly contestable, as is the appropriate indicator of market share – is it capacity, for example, or production? Units sold or revenue? Over what timeline? It would be one thing for enforcement agencies to provide guidance to the market about how market shares and concentration may affect their enforcement presumptions in approaching a merger, it is another to enshrine highly contestable and often ill-defined market-share thresholds in a statute.

More fundamentally, the 30 percent-presumption approach fails to recognize that markets with few competitors may be intensively competitive and otherwise economically beneficial – the structure-conduct-performance paradigm has long been understood to be flawed. Highly successful competitors, for example, may have large market shares because they are so effective at competing. Or there may be trivial barriers to new entry, which implies discipline even on sellers with very high market share. The Bill C-352 approach, like most attempts to create rule-driven approaches to competition law enforcement, is badly over- and under-inclusive. For example, a concentrative merger could efficiently lower costs and create social benefits by intensifying competition. It is because specific context matters that enforcement against mergers is complicated, and reliance on simple rules, based on market share or not, is inappropriate.

To be sure, Bill C-352 relies on the 30 percent threshold simply to reverse the onus of proof; the parties could rebut a presumption of a SPLC in any given case. This is better than not allowing such a rebuttal, but remains problematic. For one thing, a positive correlation between concentration and anti-competitive outcomes may or may not hold in a given market, as discussed; market shares may not offer useful insight into competitive conditions in a given market. For another, one of the reasons there has been a push to eliminate or diminish the burden of proof on the authorities is that proving a SPLC may be difficult because of dynamic markets and uncertainty. But if it is difficult to prove a SPLC because of dynamism and uncertainty, it may be difficult for the parties to rebut a presumption by proving the absence of a SPLC. It is inappropriate to reduce by statute the burden of proof on the authorities because of the difficulties of proof only to shift by statute those difficulties to the parties on the basis of contestable and questionably relevant market share statistics. Again, the better policy response is what the government has done to date: strengthen enforcement capacity to address and mitigate uncertainty directly. To put market-share presumptions in the statute would be a mistake.57

Even worse would be the second approach in Bill C-352, which would remove discretion in the face of a merger that results in a market share of 60 percent or more: in such cases, Bill C-352 would require the tribunal to make an order against the merger. This removes entirely from the statute any possible analysis of what matters – competitive conditions pre- and post-merger – in favour of a focus on market share statistics that may or may not matter.58 The proposal is wrong-headed. The burden should remain on the Bureau to prove anti-competitive effects from a merger, and the legislation should continue to avoid reliance on contestable market share statistics that have a tenuous connection to competitive harm.

In Bill C-59, the government does not propose adopting these market share thresholds in the statute. It does, however, propose abolishing Section 92(2) of the Competition Act, which states that a merger shall not be disallowed on the basis of market-share data alone. This proposed amendment is a mistake for the reasons outlined above: market share is contestable and may not reflect uncompetitive conditions. Section 92(2) gets it right: mergers ought not to be rejected on the basis of market share alone.

My prediction, however, is that if this section is abolished, it will be a mostly harmless error, unlike the Bill C-352 approach. I doubt that – at least in the short-to-medium term – competition law adjudicators would make the mistake of relying solely on market shares to prohibit a merger while ignoring other evidence, such as the absence of barriers to entry or evidence about robust remaining competition. The Tribunal, in assessing the competitive effects of a merger, would naturally consider all relevant evidence that parties put before it, whether or not the statute requires it to do so. While the amendment to abolish Section 92(2) does not advance mergers law, neither does it set it back meaningfully.

3. Uncertainty and Abuse of Dominance

ISED also raises the difficulty of proving abuse of dominance in dynamic markets. It reviews the requirements for proving abuse and contends, without elaborating on its reasons, that lowering the burden of proof may be appropriate. It states:

The requirement for the Commissioner to prove that the anti-competitive practice is resulting in, or likely to cause, an SLPC may be unduly strict. For similar reasons that market dynamics in an evolving economy may complicate merger analysis (such as disruptive but small start-ups, zero-revenue or low-asset models), the assumptions behind competitive effects may need to be revisited.59

For example, ISED considers the approach of treating conduct as an abuse of dominance if it is “capable” of having anti-competitive effects. This standard suffers from the same kind of drawbacks as “appreciable risk” discussed above in the mergers context, but may even be worse. Virtually any practice by a dominant firm is “capable” of anti-competitive effects. To take an example, when a dominant firm expands its product offerings to compete aggressively in other markets, competition is more intense in the short run, but one could always argue that such conduct is capable of harmful effects in the longer run; aggressive competition could conceivably undermine innovation, for example, or provoke exit. Even if effects are highly likely to be benign in the short run and overall, expansion by the dominant firm would be problematic under the “capable of” test.

ISED does not emphasize the “capable of” standard in its What We Heard report, perhaps reflecting recognition of its disadvantages. ISED’s second proposed approach to the burden of proof when a dominant firm’s actions create uncertain competitive impacts is to avoid case-by-case analysis and adopt rules that ban specified conduct. This would sidestep problems of proving effects in any given case. A rule-based approach would require the Competition Bureau only to prove dominance and the conduct, itself, but not the effects of the conduct.60

There are two very significant problems with this approach. First, precisely defining what conduct is, and is not, caught by the rule would be challenging at best. To consider a prominent example, self-preferencing by a dominant firm/platform is far from self-defining.61 At one extreme, it could refer to situations where a dominant firm requires users to use only its vertically related product. At the other, it could apply to any situation where a dominant firm includes vertically related products in its product suite.

Second, even if definitions were precise, the rules will be over- and under-inclusive.62The very reason why it is challenging for the Bureau to prove anti-competitive effects of conduct is that a deep dive into context is required to determine social impacts. Moreover, practices that are socially beneficial in most contexts may be harmful in a narrow set of contexts, just as practices that are often harmful may be beneficial in some contexts. Even if the definitional problem were overcome and clear rules could be drafted, they would be only crudely related to the social good in any given case. As with Bill C-352’s push to consider only market shares in merger analysis, adopting rules to address nuanced, context-specific questions about abuse of dominance would be a mistake. If the Bureau can show that self-preferencing is likely anti-competitive, a remedy is appropriate, but not otherwise.

On a similar note, while review of abusive conduct takes place ex post – unlike merger review (at present) – without conducting an analysis of competitive harm, there will be considerable uncertainty about the social good associated with any remedy the authorities require. It would not be surprising if a remedy order would make things worse. For example, out of concern that Microsoft was improperly bundling Windows Media Player to its operating system, the EU required Microsoft to sell a version of Windows without Windows Media Player. This requirement imposed significant costs on Microsoft, while attracting almost no interest from consumers. The remedy created social costs without social benefits. Without identifying the precise reasons for social concern with a practice, appropriate remedies will be elusive.

For these reasons, and those outlined in the previous section on mergers about living with uncertainty generally, it is understandable why Canadian competition law has evolved over the years to define conduct as competitively problematic without further analysis in only one significant context, per se – illegal price-fixing. Treating conduct as intrinsically unlawful more broadly would deter socially beneficial activity.

While changes to the burden of proof are misguided, ISED identifies a genuine challenge for enforcement against abuse of dominance when there is uncertainty about a practice. As with the review of uncertain mergers, there are alternatives that strengthen the case against lowering thresholds for intervention. Funding enforcement, both with public money and by expanding private rights of action for abuse, would create better enforcement. The government has increased the Bureau’s budget, and proposes expanding private rights of action. Private rights of action for abuse – but not damages – are available at present, but Bill C-59 proposes adding the possibility of damages as a remedy.63This would be a good idea, in my view, as it would incentivize litigation that would better hold dominant firms to account.64 The threats of administrative monetary penalties (only recently meaningfully increased) and of private damages not only create deterrence, but also allow for greater ex-post enforcement; that is, more meaningful enforcement after the competitive effects of a practice become clearer.

If a practice takes time for its anti-competitive effects to become apparent, a regime that essentially only makes cease-and-desist orders may not deter anti-competitive activity: the parties can adopt anti-competitive conduct knowing that the worst case competition law outcome would be an order sometime in the future to stop. But if retrospective damages or meaningful administrative monetary penalties are ordered, enforcement authorities have the opportunity to wait and see what the effects are without undermining deterrence: dominant firms anticipating future financial payments for abusive conduct will have a disincentive to adopt anti-competitive practices. These alternative means of strengthening enforcement better calibrate enforcement to deter harmful practices and to allow beneficial ones than changing the burden of proof.65

That said, no matter what reforms are adopted, there will be residual enforcement shortcomings in the face of uncertainty. Living with these shortcomings, as the law has consistently done, is better than incurring the harms associated with lowering the burden of proof.

Considering a different tack, ISED currently identifies the conjunctive requirements of anti-competitive intent and effect under abuse of dominance as a possible subject for reform. If the requirement of intent or purpose were dropped, little would be lost. At present, case law appropriately infers intent from effects.66If a practice has anti-competitive effects, it is assumed that the firm adopting the practice intended those effects, just as if the practice has pro-competitive efficiency effects, it is presumed that the firm intended those effects.67As a matter of policy, it should never be the case that a practice that creates a SPLC is not found to be an abuse of dominance because the dominant firm did not intend to harm competition. If the practice has harmful anti-competitive effects, it should not be allowed, regardless of intent. The present emphasis on effects to infer intent avoids allowing anti-competitive conduct because of an allegedly benign intent.

On the other hand, focusing only on intent to find anti-competitive abuse would be a mistake. For one thing, there may be internally conflicting views at a dominant firm about what its intent is. Moreover, businesspeople may make cavalier statements to one another that reflect an aggressive, even martial posture to competitors, but that may say very little about whether the strategy is to harm competition rather than competitors, or about the prospects for successful anti-competitive behaviour. An older predatory pricing case, R. v. Hoffman-LaRoche,68 provides an excellent example. In response to entry by generics, an incumbent pharmaceutical company began aggressive giveaway promotions of its name-brand drug, with internal memoranda saying a variety of inflammatory things such as the importance of serving notice to “parasites” that the incumbent means business. The court convicted for predatory pricing, relying in part on these memoranda to establish mens rea (predatory pricing was criminal at the time), while at the same time observing that the result of the price wars was that the incumbent no longer participated in the market and generics had taken over. That is, the pricing strategies of Hoffman-LaRoche were not anti-competitive in effect nor, given the low barriers to entry, did they have a significant probability of being anti-competitive. Despite the internal statements, given the strong implausibility of anti-competitive effects in the circumstances, neither an abuse nor a predatory pricing claim ought to have been successful.

To be clear, evidence of a dominant firm’s intentional strategy to exclude competition may help the finder of fact infer likely anti-competitive effects caused by the dominant firm, but intent in itself is not germane to an analysis of competition.

While ISED’s discussion paper raises problematic approaches to the burden of proof in abuse of dominance cases, the government adopted clearly mistaken amendments to the burden of proof. Rather than requiring a practice of anti-competitive acts by a dominant firm that creates a SPLC, the government has amended s. 79 to allow the Tribunal to make an order against a dominant firm that engages in a practice of anti-competitive acts whether or not it creates a SPLC. As I have outlined, in my view, it is inappropriate to dispense with the SPLC test. Having a meaningful competition test here and globally has served as a sensible screen against unwarranted and harmful intervention by competition authorities. It is, however, especially perverse for Canada to abandon the SPLC test given Canadian case law on anti-competitive effects.

Following its reading of the statute, the Federal Court of Appeal in Canada Pipe (2006) held that an “anti-competitive act” is one that harms competitors and not necessarily competition. According to the court, the competition assessment should come only after the analysis of whether there has been a practice of anti-competitive acts.69 Indeed, Canada Pipe concluded that an act that is harmful to competitors may be classified as anti-competitive even if it is ultimately beneficial for competition.70 While calling a potentially pro-competitive act “anti-competitive” because it is harmful to competitors has always been incoherent, the additional SPLC requirement has mitigated the harm that the Canada Pipe interpretation has done in practice.

The recent amendment to the Act eliminated this mitigation and does significant damage to Canadian abuse law. It would be absurd to order injunctions against and even punish acts with administrative monetary penalties, or damages if Bill C-59 passes, that are pro-competitive, yet harmful to competitors and, thus, “anti-competitive” under the definition in Canada Pipe. That is nevertheless what the government has unfortunately invited with its revision to the Act. The amendment to lower the burden of proof by not requiring a SPLC, rather than simply enhancing sensible enforcement, is an unfortunate development.

4. Conclusion

The digitization of the economy has made life more complicated for competition enforcers. Determining competitive effects of practices, never entirely straightforward, has become even more complex with the rise of the digital economy in which markets may be especially prone to market power. The optimal response to this challenge, however, is to target market power and uncertainty directly by enhancing enforcement capacity, not simply to make it easier for the authorities to meet their burden of proof. The government deserves credit for strengthening enforcement, and credit for thus far resisting changes to the burden of proof in respect of mergers, but it is very unfortunate that it eliminated the requirement of a SPLC for abuse of dominance, especially given case law that holds that “anti-competitive acts” might include acts that benefit competition. In light of the steady stream of amendments to the law over the last months, perhaps it could revisit this decision in the near future.

References

Bawania, Ray, and Yelena Larkin. 2019. “Are Industries Becoming More Concentrated? The Canadian Perspective,” May 2019: papers.ssrn.com (accessed Nov. 30, 2023).

Baker, Jonathan. “Taking the Error out of ‘Error Cost’ Analysis: What’s Wrong with Antitrust’s Right,” Antitrust Law Journal 80 (2015–16): 1.

Baker, Jonathan, Andrew Gavil, Richard Gilbert, et al., 2023. “Comments of Economists and Lawyers on the Draft Merger Guidelines,” Sept. 15, papers.ssrn.com (accessed Dec. 1, 2023).

Bouras, Panagiotis, Christian Bustamante, Xing Guo, and Jacob Short. 2023. “The Contribution of Firm Profits to the Recent Rise in Inflation,” Bank of Canada Staff Working Paper, Aug. 1, 2023, www.bankofcanada.ca (accessed Dec. 1, 2023).

Competition Bureau. 2023. “Competition in Canada from 2000 to 2020: An Economy at a Crossroads,” Oct. 19, ised-isde.canada.ca (accessed Nov. 30, 2023).

Easterbrook, Frank. 1984. “The Limits of Antitrust,” Texas Law Review 63 (1984): 1.

Government of Canada. “Bill C-56: An Act to Amend the Excise Tax Act and the Competition Act,” https://www.parl.ca/legisinfo/en/bill/44-1/c-56, (accessed Jan. 18, 2024).

___________________. 2023 “Fall Economic Statement 2023,” www.budget.canada.ca (accessed Dec. 5, 2023).

____________________. 2023. “Bill C-352: An Act to Amend the Competition Act and the Competition Tribunal Act,” Sep. 18, 2023, www.parl.ca (accessed Nov. 30, 2023).

_____________________. 2023. “Future of Canada’s Competition Policy Consultation – What We Heard Report,” September 2023, ised-isde.canada.ca (accessed Nov. 30, 2023).

____________________. 2022. “The Future of Competition Policy in Canada,” ised-isde.canada.ca (accessed Nov. 30, 2023).

Government of the United States. 2021. Competition and Antitrust Law Enforcement Reform Act of 2021, SIL21191, 117th Congress.

Hemphill, C. Scott, and Tim Wu. 2020. “Nascent Competitors” (2020) 168 University of Pennsylvania Law Review 1879.

Iacobucci, Edward.. 2023. “Is the Canadian Competition Act Fit for Purpose in the Digital Era? What Purpose(s)?” Canadian Business Law Journal 67 (2023): 7.

___________________. 2021. “Examining the Canadian Competition Act in the Digital Era,” Senate of Canada consultation paper: sencanada.ca (accessed Nov. 30, 2023);

___________________.2015. “The Lessons of Tervita,” Canadian Business Law Journal 57 (2015): 217.

International Chamber of Commerce. 2023. “10 ICC Recommendations to Make Requests for Information in Competition Investigation More Efficient,” Oct. 15, 2023, iccwbo.org (accessed Nov. 30, 2023).

Johnson, Paul. 2021.“A Competition Conundrum: Winner-Take-All Markets,” C.D. Howe Institute, Intelligence Memo, November 29, 2021, www.cdhowe.org (accessed Nov. 30, 2023)

Krane, Joshua, and James Musgrove. 2021.“The Danger of Precautionary Principle Challenges to Nascent Mergers,” February 2021, C.D. Howe Institute Intelligence Memo, www.cdhowe.org (accessed Dec. 1, 2023).

Niblett, Anthony. 2023. “The Competition Act, Not the Competitors[sic] Act,” Canadian Business Law Journal 67 (2023): 83.

________________.2023. “Stronger Enforcement of the Competition Act Is Better than a Dramatic Overhaul,” MacDonald-Laurier Institute, Feb. 2, 2023, macdonaldlaurier.ca (accessed Dec. 1, 2023).

OECD. “Indicators of Product Market Regulation,” OECD, www.oecd.org (accessed Nov. 30, 2023).

OECD. 2023. “Canada: Strengthen Productivity to Boost Living Standards and Prioritise the Low-Carbon Transition,” March 6, 2023, www.oecd.org (accessed Nov. 30, 2023).

Ottaviani, Marco, and Abraham Wickelgren. 2023. “Approval Regulation and Learning, with Application to Timing of Merger Control,” Journal of Law, Economics and Organization (Jan. 9, 2023): doi.org/10.1093/jleo/ewac025 (accessed Nov. 30, 2023).

Padilla, Jorge. 2021. “Decision Theory and Legal Process in EU Competition Law,” June 4, 2021, papers.ssrn.com/sol3/papers.cfm?abstract_id=3859937 (accessed Nov. 30, 2023).

Pecman, John. 2022. “Toughening Canada’s Competitiveness,” November 2022, Canadian Chamber of Commerce, chamber.ca (accessed Nov. 30, 2023).

Pecman, John, Paul Johnson and Christine Reisler. 2020. “Essential Facilities Fallacy: Big Tech, Winner-Take-All Markets, and Anticompetitive Effects,” CPI Antitrust Chronicle 1 (Feb. 2020).

Rodal, Shuli, Michelle Lally, Kaeleigh Kuzma and Zach Rudge. 2023. “Significant Reforms to Canada’s Competition Act with Far-Reaching Implications,” Osler, Nov. 29: www.osler.com (accessed Dec. 5, 2023).

Shapiro, Carl. 2018. “Antitrust in a Time of Populism,” International Journal of Industrial Organization 61 (November: 714.

UK Digital Competition Expert Panel. 2019. “Unlocking Digital Competition: Report of the Digital Competition Expert Panel” (Furman Report), March. assets.publishing.service.gov.uk (accessed Nov. 30, 2023).

Winter, Ralph. 2015. “Tervita and the Efficiency Defence in Canadian Competition Law.” Canadian Competition Law Record 28 (2015): 133.

• 1 For an excellent overview and critical analysis of the recent political salience of US antitrust sentiment., see Carl Shapiro, “Antitrust in a Time of Populism,” International Journal of Industrial Organization 61 (November 2018): 714. For an example of the negative reaction to big tech, see Elizabeth Warren, “Break Up Big Tech,” 2020 Senate campaign document: 2020.elizabethwarren.com.
• 2 In June 2023, the Competition Bureau Canada released a market study on the grocery industry, concluding – on the basis of data the Bureau appropriately acknowledged was imperfect – that gross margins in groceries had increased only modestly in recent years: ised-isde.canada.ca. The biggest challenge for those who would blame grocery inflation on market power is why weak competitive conditions were not causing higher prices and margins earlier. There was no obvious break in competitive conditions that would lead to higher grocery prices around the time when inflation started to grow recently, but of course there were a variety of supply-side issues, including the Russian invasion of Ukraine, that could help explain rising grocery prices.
• 3 See Shapiro, “Antitrust in a Time of Populism.”
• 4 See “Canada: Strengthen Productivity to Boost Living Standards and Prioritise the Low-Carbon Transition,” OECD, March 6, 2023, www.oecd.org (accessed Nov. 30, 2023).
• 5 See Shapiro, “Antitrust in a Time of Populism” for review of US studies. In Canada, see Ray Bawania and Yelena Larkin, “Are Industries Becoming More Concentrated? The Canadian Perspective,” May 2019: papers.ssrn.com (accessed Nov. 30, 2023).
• 6 See “Competition in Canada from 2000 to 2020: An Economy at a Crossroads,” Competition Bureau, Oct. 19, 2023, ised-isde.canada.ca (accessed Nov. 30, 2023).
• 7 See Ibid.
• 8 The Bureau concluded that concentration has risen slightly in concentrated industries over the 2005–2018 period, based on its assessment of the Herfindahl-Hirschman Index concentration using the four-digit North American Industry Classification System (NAICS) codes. It consigned to an appendix its finding that concentration fell in assessing the sum of the market shares of the three largest firms in industries (CR3) as measured by three-digit NAICS codes, observing that the more aggregated three-digit NAICS classifications masked variation in concentration measures at the four-digit level. The same logic renders perilous any inference about concentration trends from industry data: data from four-digit codes at the national level undoubtedly mask variation at the product and geographic market levels.
• 9 As an example of an oddity that the report does not call attention to, the report relies on data from three-digit NAICS industries to assess scale economies, a level of aggregation the report also suggested was excessively crude when observing that the CR3 measure of concentration fell at the three-digit level.
• 10 See “Indicators of Product Market Regulation,” OECD, www.oecd.org (accessed Nov. 30, 2023).
• 11 For my part, I believe the Act remains suitable, relying as it does on flexible standards. See: Edward Iacobucci, “Examining the Canadian Competition Act in the Digital Era,” 2021, Senate of Canada consultation paper: sencanada.ca (accessed Nov. 30, 2023); and Iacobucci, “Is the Canadian Competition Act Fit for Purpose in the Digital Era? What Purpose(s)?” Canadian Business Law Journal 67 (2023): 7.
• 12 See “Bill C-56: An Act to Amend the Excise Tax Act and the Competition Act,” Government of Canada, https://www.parl.ca/legisinfo/en/bill/44-1/c-56, (accessed Jan. 18, 2024).
• 13 See “Fall Economic Statement 2023,” Government of Canada, 2023, www.budget.canada.ca (accessed Dec. 5, 2023).
• 14 https://www.parl.ca/DocumentViewer/en/44-1/bill/C-59/first-reading.
• 15 See “Bill C-352: An Act to Amend the Competition Act and the Competition Tribunal Act,” Government of Canada, Sep. 18, 2023, www.parl.ca (accessed Nov. 30, 2023).
• 16 There are different ways of describing the requirement to prove anti-competitive effects on a balance of probabilities (or some other formulation), including “burden of proof” or “standard of proof.” See, e.g., Jorge Padilla, “Decision Theory and Legal Process in EU Competition Law,” June 4, 2021, papers.ssrn.com/sol3/papers.cfm?abstract_id=3859937 (accessed Nov. 30, 2023). I prefer “burden of proof,” distinguished from “onus of proof,” which rests on the party that is obliged to prove a claim.
• 17 See “The Future of Competition Policy in Canada,” Government of Canada, 2022, ised-isde.canada.ca (accessed Nov. 30, 2023).
• 18 See “Future of Canada’s Competition Policy Consultation – What We Heard Report,” Government of Canada, September 2023, ised-isde.canada.ca (accessed Nov. 30, 2023).
• 19 See Government of Canada, “The Future of Competition Policy in Canada.”
• 20 Former Competition Commissioner John Pecman similarly concludes that many proposals to make Canadian competition law more aggressive on different dimensions would do more harm than good: John Pecman, “Toughening Canada’s Competitiveness,” November 2022, Canadian Chamber of Commerce, chamber.ca (accessed Nov. 30, 2023). Pecman considers a range of possible reforms, while I focus on ISED’s recent discussion of the burden of proof.
• 21 See Government of Canada, “The Future of Competition Policy in Canada.”
• 22 See Tervita Corp. v. Canada (Commissioner of Competition), SCC 3 (2015).
• 23 See Ibid., para. 124.
• 24 Illustrating the uncertainty about the legal importance of quantification, after Tervita, the tribunal accepted a harm-to-innovation theory without quantitative estimates in The Toronto Real Estate Board v. Commissioner of Competition 2017 FCA 236.
• 25 See Government of the United States, Competition and Antitrust Law Enforcement Reform Act of 2021, SIL21191, 117th Congress, 2021.
• 26 See Edward Iacobucci, “The Lessons of Tervita,” Canadian Business Law Journal 57 (2015): 217. See also Ralph Winter, “Tervita and the Efficiency Defence in Canadian Competition Law.” Canadian Competition Law Record 28 (2015): 133.
• 27 See Iacobucci, “Examining the Canadian Competition Act in the Digital Era” and Iacobucci, “Is the Canadian Competition Act Fit for Purpose in the Digital Era? What Purpose(s)?”
• 28 See Pecman, “Toughening Canada’s Competitiveness.”
• 29 See Ibid.
• 30 Note that an order against a merger ex post would potentially generate ex-ante deterrence by creating dissolution costs for the merged entity.
• 31 See Marco Ottaviani and Abraham Wickelgren, “Approval Regulation and Learning, with Application to Timing of Merger Control,” Journal of Law, Economics and Organization (Jan. 9, 2023): doi.org/10.1093/jleo/ewac025 (accessed Nov. 30, 2023).
• 32 Bill C-56 allocates authority for ordering a market study to the minister of ISED, which runs the risk of compromising the Bureau’s independence and of resulting in political motivations for market studies, rather than effective knowledge-building.
• 33 The International Chamber of Commerce recognizes the need for competition authorities to gather information from parties not directly subject to an investigation of competition law infringements, but calls for authorities to consider the costs to business in doing so. For example, it calls for the authorities to speak with industry participants to get a sense of the industry prior to sending information requests. See International Chamber of Commerce, “10 ICC Recommendations to Make Requests for Information in Competition Investigation More Efficient,” Oct. 15, 2023, iccwbo.org (accessed Nov. 30, 2023).
• 34 See, e.g., C. Scott Hemphill and Tim Wu, “Nascent Competitors” (2020) 168 University of Pennsylvania Law Review 1879.
• 35 To be sure, the precise definition of “appreciable risk” is not self-evident, but I take it to mean significantly lowering the burden of proof to catch mergers with relatively low risks of anti-competitive outcomes.
• 36 See Order Denying Plaintiff’s Motion for Preliminary Injunction, Fed. Trade Commission v. Meta Platforms, Inc., et al., No. 5:22-cv-04325- EJD, Dkt. No. 549.
• 37 This standard is itself arguably too easy to meet, depending on the interpretation of “reasonably probable.” The District Court in this case did not set the bar too low.
• 38 Stanford economist Stephen Tadelis puts the matter succinctly: “Unlike pharma, where acquisitions can lead to killing competition (@florianederer’s excellent JPE paper), in tech they often lead to large scale execution, something start-ups almost always fail at. Is that distinction too subtle for regulators?” Stephen Tadelis (@steve_tadelis), X, Sep. 15, 2021, 8:13 p.m. twitter.com/stevetadelis.
• 39 See International Chamber of Commerce, “10 ICC Recommendations to Make Requests for Information in Competition Investigation More Efficient.”
• 40 UK Digital Competition Expert Panel, “Unlocking Digital Competition: Report of the Digital Competition Expert Panel” (Furman Report), March 2019, assets.publishing.service.gov.uk (accessed Nov. 30, 2023). See also: Paul Johnson, “A Competition Conundrum: Winner-Take-All Markets,” C.D. Howe Institute, Intelligence Memo, November 29, 2021, www.cdhowe.org (accessed Nov. 30, 2023); and John Pecman, Paul Johnson and Christine Reisler, “Essential Facilities Fallacy: Big Tech, Winner-Take-All Markets, and Anticompetitive Effects,” CPI Antitrust Chronicle 1 (Feb. 2020). The latter two references discuss the importance of winner-take-all markets to an analysis of nascent competitor acquisitions, suggesting that perhaps in these markets the burden of proving anti-competitive effects should be lower. I am skeptical of this proposal, in part because identifying whether a market is truly winner-take-all is itself not at all straightforward.
• 41 Recall that the Bureau justifiably objects to the quantification preference set out in Tervita under the status quo. If quantification of likely anti-competitive effects is impractical, the balance of harms test is impractical.
• 42 Moreover, the balance-of-harms test would be jarring within the Canadian legal system, which generally relies on a balance-of-probabilities standard in civil matters without a full cost-benefit analysis. For example, it would likely be unnatural for Canadian adjudicators to allow a merger that had a 99-percent chance of being anti-competitive and socially harmful but a one percent chance of being spectacularly pro-competitive and socially beneficial.
• 43 Prominent jurisdictions, including the EU and the UK, have moved to expanding merger notification requirements for dominant platforms to address this problem. See International Chamber of Commerce Task Force on Competition Law and the Digital Economy, “ICC Report on Antitrust in the Digital Economy,” September 2023. I am sympathetic to such an approach to notification as another means of gathering information.
• 44 See Pecman, Johnson and Reisler, “Essential Facilities Fallacy: Big Tech, Winner-Take-All Markets, and Anticompetitive Effects.”
• 45 See Iacobucci, “Examining the Canadian Competition Act in the Digital Era.”
• 46 See Atlantic Sugar Refineries Co. Ltd. et al. v. Attorney General of Canada, 2 S.C.R. 644 (1980).
• 47 Bill C-352 would make it an abuse of dominance for a dominant firm to charge “excessive and unfair” prices. This would require the Bureau and tribunal to become shadow price regulators in any market involving a dominant firm. As anyone passingly familiar with utility regulation would appreciate, this is a complicated exercise that the Bureau and tribunal are not institutionally suited to undertake.
• 48 For example, setting prices so that investment in innovation is rewarded is a near impossible task.
• 49 See Iacobucci, “Examining the Canadian Competition Act in the Digital Era.”
• 50 See Joshua Krane and James Musgrove, “The Danger of Precautionary Principle Challenges to Nascent Mergers,” February 2021, C.D. Howe Institute Intelligence Memo, www.cdhowe.org (accessed Dec. 1, 2023).
• 51 See Frank Easterbrook, “The Limits of Antitrust,” Texas Law Review 63 (1984): 1.
• 52 Easterbrook, himself, notes that the “long run may be a long time coming…” Ibid. See also Jonathan Baker, “Taking the Error out of ‘Error Cost’ Analysis: What’s Wrong with Antitrust’s Right,” Antitrust Law Journal 80 (2015–16): 1.
• 53 To use the language of decision theory (see Padilla, “Decision Theory and Legal Process in EU Competition Law”), the alternatives reduce the cost of Type I errors by limiting intervention ex ante that prevents gains from acquisitions of nascent competitors, while also reducing the cost of Type II errors by allowing enforcement against mergers that turn out to be anti-competitive.
• 54 While not focused on the grocery sector, a 2023 study concluded that market power did not appear to play a meaningful role in causing inflation: price-cost markups and inflation had moved in opposite directions. See Panagiotis Bouras, Christian Bustamante, Xing Guo and Jacob Short, “The Contribution of Firm Profits to the Recent Rise in Inflation,” Bank of Canada Staff Working Paper, Aug. 1, 2023, www.bankofcanada.ca (accessed Dec. 1, 2023). It is jarring, though perhaps not politically surprising, that parliamentarians focused their ire on “greedflation” in groceries while seemingly uniformly supporting regressive and inefficient government-mandated cartels in agricultural staples such as dairy, poultry and eggs.
• 55 The government’s decision to authorize the minister, but not the Competition Bureau, to initiate market studies raises similar concerns.
• 56 See https://www.parl.ca/legisinfo/en/bill/44-1/c-352.
• 57 Bill C-352 creates a similar reverse onus should the merger “significantly increase” concentration. This is also wrong for the reasons outlined in the text and, in addition, would create great uncertainty about what counts as a “significant increase.”
• 58 In a submission to the Department of Justice about proposed merger enforcement guidelines, a group of US antitrust experts expressed their concern that the guidelines, like Bill C-352, would inappropriately regard high market shares or increases in concentration “as intrinsically harmful or, to similar effect, as conclusive indicators of harm.” However, increases in concentration could simply reflect growing economies of scale, for example, whose exploitation benefits consumers and other stakeholders. See Jonathan Baker, Andrew Gavil, Richard Gilbert, et al., “Comments of Economists and Lawyers on the Draft Merger Guidelines,” Sept. 15, 2023, papers.ssrn.com (accessed Dec. 1, 2023). Note that these experts state that they support stricter enforcement against mergers, but take issue with this overemphasis on market shares.
• 59 See Government of Canada, “The Future of Competition Policy in Canada.”
• 60 This is similar to the per se approach to anti-competitive conduct, or the EU’s anti-competitive “by object” approach.
• 61 See Anthony Niblett, “The Competition Act, Not the Competitors[sic] Act,” Canadian Business Law Journal 67 (2023): 83; and Anthony Niblett, “Stronger Enforcement of the Competition Act Is Better than a Dramatic Overhaul,” MacDonald-Laurier Institute, Feb. 2, 2023, macdonaldlaurier.ca (accessed Dec. 1, 2023).
• 62 See Niblett, “Stronger Enforcement of the Competition Act Is Better than a Dramatic Overhaul.”
• 63 It accomplishes this by authorizing the tribunal to order a firm to make a payment to the tribunal reflecting the gains from anti-competitive behaviour that the tribunal would then allocate to affected parties.
• 64 See Pecman, “Toughening Canada’s Competitiveness.” To be sure, the wisdom of monetary damages becomes less clear if the substantive standards for finding of abuse of dominance are inappropriate. As I discuss, recent amendments to lower the burden of proof were a mistake.
• 65 The prospect of very high administrative monetary penalties may raise constitutional questions, as well as concerns about over-deterrence. Appropriately calibrated public penalties and private damages have the potential to create appropriate deterrence, but are not without risk.
• 66 For example, The Commissioner of Competition v. Vancouver Airport Authority, Comp Trib 6 (2019) concerned the Vancouver Airport Authority’s (VAA) decision to grant certain airline catering firms exclusive rights at the airport out of concern that others would not be able to match their service. The tribunal decided that the VAA had not engaged in anti-competitive acts because the foreseeable effects of its conduct were on balance efficient, not anti-competitive. While at points observing that the intent of the exclusivity was efficient and not anti-competitive, the analysis clearly relied on an effects-based test to reach this conclusion. The tribunal concluded at para. 513: [513] Collectively, [VAA’s] concerns were and are linked to cognizable efficiency or pro- competitive considerations that are independent of any anti-competitive effects of the impugned conduct. Having regard to the conclusions reached in Section VII.E below in relation to paragraph 79(1)(c), the Tribunal finds that any such actual and reasonably foreseeable anti-competitive effects of the impugned conduct are not disproportionate to those efficiency and pro-competitive rationales. Indeed, the Tribunal is satisfied that, when weighed against the exclusionary negative effects of VAA’s conduct, these legitimate business considerations are sufficient to counterbalance them.
• 67 See Ibid.
• 68 See R. v. Hoffman-LaRoche, 28 O.R. (2d) 164 (Ont. H. Ct.) (1980).
• 69 See Canada (Commissioner of Competition) v. Canada Pipe Company Ltd., FCA 233 (2006).
• 70 A recent amendment to the definition of anti-competitive acts in s. 78 continues to treat acts that harm competitors as anti-competitive acts.