Should Canada Overhaul the SLPC Test For Mergers?

(disponible uniquement en anglais)

By: Andy Baziliauskas and Margaret Sanderson

Abstract

Executive Summary and Recommendations

Governments throughout the world have been questioning whether their competition laws are sufficiently robust to police potentially anti-competitive conduct, including mergers. Of particular concern is whether the tools developed for addressing potentially anti-competitive mergers in traditional markets may fail in digital settings and, as a result, need to be overhauled. As more industries enhance their product offerings with digital innovations, these concerns may eventually cover a greater share of the economy from that which is currently held by so-called ‘Big Tech’ firms.

Competition laws in respect of mergers have tended to focus on price effects. In digital settings, adverse non-price effects, such as lower product quality and less innovation, may be the greater concern, particularly because digital products and services may have a “zero price” for consumers in that they are “free”. In addition, the link between share and concentration measures and competitive effects may be more tenuous in digital markets, particularly in cases involving incumbent firms acquiring nascent (or potential future) competitors with small, or even no, current share in relevant markets.

In response to these concerns, calls have been made to strengthen competition laws to better protect consumers and the competitive process. In Canada, the Competition Bureau (“Bureau”) is recommending numerous changes to the competitive effects evaluation of mergers, including adopting a US-style structural presumption, lowering the ‘substantiality’ threshold, and lessening or reversing the evidentiary burden to establish that a merger is likely to substantially lessen or prevent competition (“SLPC”).

A key motivation for the Bureau’s proposals is reducing the total resources it devotes under the current system to establishing and quantifying anti-competitive harm in the mergers context. With respect to quantification of harm, the Bureau recommends that the efficiencies defence currently included in section 96 of the Act – and the corresponding trade-off analysis that this defence entails – be repealed and that efficiencies instead be considered as one of several factors under section 93 when the Competition Tribunal (“Tribunal”) evaluates whether a merger is likely to result in an SLPC. Should the section 96 trade-off analysis be repealed, the requirement for the Bureau to quantify anti-competitive effects will no longer exist. Our commentary focuses on the proposals to reform the standards used to establish an SLPC. For the purposes of this commentary, we assume that the Bureau’s proposals with respect to the section 96 trade-off are adopted. 

1. Executive Summary and Recommendations

Governments throughout the world have been questioning whether their competition laws are sufficiently robust to police potentially anti-competitive conduct, including mergers. Of particular concern is whether the tools developed for addressing potentially anti-competitive mergers in traditional markets may fail in digital settings and, as a result, need to be overhauled. As more industries enhance their product offerings with digital innovations,¹ these concerns may eventually cover a greater share of the economy from that which is currently held by so-called ‘Big Tech’ firms.

Competition laws in respect of mergers have tended to focus on price effects. In digital settings, adverse non-price effects, such as lower product quality and less innovation, may be the greater concern, particularly because digital products and services may have a “zero price” for consumers in that they are “free”.² In addition, the link between share and concentration measures and competitive effects may be more tenuous in digital markets, particularly in cases involving incumbent firms acquiring nascent (or potential future) competitors with small, or even no, current share in relevant markets.

In response to these concerns, calls have been made to strengthen competition laws to better protect consumers and the competitive process.³ In Canada, the Competition Bureau (“Bureau”) is recommending numerous changes to the competitive effects evaluation of mergers, including adopting a US-style structural presumption, lowering the ‘substantiality’ threshold, and lessening or reversing the evidentiary burden to establish that a merger is likely to substantially lessen or prevent competition (“SLPC”).⁴

A key motivation for the Bureau’s proposals is reducing the total resources it devotes under the current system to establishing and quantifying anti-competitive harm in the mergers context. With respect to quantification of harm, the Bureau recommends that the efficiencies defence currently included in section 96 of the Act – and the corresponding trade-off analysis that this defence entails – be repealed and that efficiencies instead be considered as one of several factors under section 93 when the Competition Tribunal (“Tribunal”) evaluates whether a merger is likely to result in an SLPC. Should the section 96 trade-off analysis be repealed, the requirement for the Bureau to quantify anti-competitive effects will no longer exist.⁵ Our commentary focuses on the proposals to reform the standards used to establish an SLPC. For the purposes of this commentary, we assume that the Bureau’s proposals with respect to the section 96 trade-off are adopted.⁶

Our observations and recommendations follow.

• Courts and enforcers have imperfect information and imperfect foresight. As a result, there is inevitably some risk of false positives (wrongful prohibitions) or false negatives (wrongful approvals) in competition law enforcement. Optimal enforcement aims to minimize the risks of these errors, while also minimizing regulatory costs and maximizing the benefits from enforcement.
• The Bureau’s SLPC reform proposals emphasize the harms from false negatives (allowing harmful mergers) and high regulatory and litigation costs to support easing the Bureau’s prosecutorial burden. Increased risk of false positives, or potential harm to the economy arising from efficiency-enhancing or innovation-enhancing transactions not being pursued (from a fear of over-enforcement), are ignored. Yet the significant changes proposed by the Bureau to the SLPC test would make harms from false positives a greater concern assuming the ‘substantiality’ threshold is lowered, and/or the section 96 efficiencies trade-off is eliminated.
• While digital markets and concerns about the effects of mergers on innovation have been cited as motivation for the proposed reforms, any amended provisions will apply to all mergers in all industries. Merger transactions in Canada are predominantly in traditional or resource sectors, with very few involving ‘Big Tech’.⁷ In addition, the effects of mergers on innovation are seldom primary concerns for the Bureau, in either digital or traditional markets.⁸
• A too-stringent SLPC test is not the reason the Bureau has lost recent litigated merger cases.⁹ Instead, in all cases where the Bureau lost the efficiencies trade-off, the Bureau was able to successfully establish an SLPC under section 92. Further, we are not aware of any instances where the Bureau has chosen not to take enforcement action because it considered the SLPC test too difficult to meet. The Bureau has been clear in communicating it will take enforcement action on any merger that it finds will generate an SLPC even when the merging firms claim offsetting efficiencies under section 96. 
• The Tribunal has provided valuable guidance in recent litigated merger cases on its interpretation of the SLPC requirements, including the meaning of ‘substantial’ and ‘material’. The Tribunal’s current guidance establishes reasonable thresholds that the Commissioner can, and has, met in litigated mergers.
• The Bureau has successfully argued anti-competitive effects of mergers that include price and non-price factors. Indeed, the Tribunal has found non-price harms when considering an SLPC in several cases where the Bureau did not provide any quantification or direct evidence of non-price effects. Establishing an SLPC under section 92 does not require quantification of price or non-price effects, and proving non-price effects has not historically been a barrier to proving an SLPC. As discussed below, in Tervita the Supreme Court of Canada stated that the Act does not bar the finding of a substantial prevention of competition when the Commissioner has failed to quantify deadweight loss¹⁰, and the Commissioner has proven non-price anti-competitive effects in a number of litigated cases, including in the recent Secure case.
• While there are no litigated digital ‘killer acquisition’ cases in Canada, the Tervita case involved the acquisition of a prospective new entrant by a large incumbent that the Bureau successfully challenged on prevention of competition grounds. In that case, the Supreme Court of Canada affirmed that the Bureau must demonstrate a ‘discernable’ timeframe for entry by the acquired firm, as otherwise the timing of entry is merely speculative. This may, in some cases, make it more challenging for the Bureau to meet its burden, if the entry timeframe is highly uncertain or anticipated to be farther into the future.¹¹ On the other hand, the Federal Court of Appeal established what appears to be a flexible approach that permits the extension of the timing for consideration of entry when it can be demonstrated that entry is ‘poised’ over the extended timeframe. 
• While the Bureau recommends adopting a structural presumption in Canada to ‘simplify and expedite’ merger review,¹² proving an SLPC for both price and non-price effects has not required especially ‘complex economic tests’. Establishing market definition and the extent of competition between the merging firms can also be done without sophisticated economic analysis, whether a structural presumption or the current SLPC threshold is employed. Moreover, the SLPC framework (with some adjustments) is likely a better approach to addressing acquisitions of nascent competitors in digital markets than a structural presumption, precisely because nascent competitors will have small shares such that a structural presumption test may be unlikely to be met.
• In most litigated mergers, the Commissioner has satisfied the Tribunal with ‘sufficiently clear and convincing evidence’ of an SLPC, beyond the current legal burden of a ‘balance of probabilities’ standard. Replacing this standard with a ‘balance of harms’ approach may risk continued emphasis on quantification of harms as well as estimation of probabilities. 
• In sum, we find little support for making dramatic changes to the current SLPC test, assuming the Bureau’s section 96 efficiencies reforms are adopted. Instead, we recommend more modest revisions to address concerns about possible future underenforcement of digital economy mergers, without risking harms from increased false positive in traditional (and digital) markets. The reforms we support include:
  • As recommended by Professor Iacobucci:¹³ (i) broaden the section 93 factors to include consideration of whether an acquisition would entrench or enhance the market position of a leading firm, including firms in the digital sector; and (ii) expand section 93(f) to include consideration of the removal of ‘a potentially vigorous and effective competitor’.
  • Consider further increased funding for the Bureau to overcome its concerns about the resource costs of challenges related to digital economy mergers. This would reduce underenforcement while limiting the cost of false positives.
  • Consider lower notification thresholds for firms in digital markets if the number of increased notifications will not be overwhelming to Bureau resources. 
  • Extend the limitation period under section 97 from one year to the three years recommended by the Bureau, or longer.
  • Provide funding and investigative powers to enable more market studies and merger retrospectives, allowing for the refinement of analytical approaches and applying lessons learned. The government may consider making this mandatory, with such a mandate possibly only applying to digital economy mergers and/or mergers of nascent competitors. An impartial third party should conduct the retrospectives. 

We elaborate below. Section 2 describes the merger reform proposals that have been advanced by the Bureau and Industry, Science, and Development Canada (“ISED”). Section 3 describes the error costs inherent in enforcement. Section 4 analyzes the Tribunal decisions in merger cases in the context of the proposed SLPC reforms. In Section 5, we comment on the benefits and costs to adopting a structural presumption or a balance of harms standard to replace the current SLPC test, including some statistics on the number of digital transactions reviewed by the Bureau. Section 6 concludes and includes our recommendations. 

2. Canadian Merger Reform Proposals

Competition Bureau Merger Reform Proposals

The Bureau’s response to Professor Iacobucci’s consultation paper¹⁴ expresses several concerns, including that the efficiencies defence allows mergers that harm consumers. In respect of the section 92 competitive effects test, the Bureau is concerned that: (i) having it bear the burden to demonstrate an SLPC is not an efficient use of ‘judicial, business, or public sector resources’; (ii) the current competitive harm standards are not adequate for assessing acquisitions of nascent competitors; and (iii) the remedy standard, which only requires that a remedy eliminate the “S” from an SLPC, allows firms to accumulate market power. To address its concerns, the Bureau recommends the adoption of a US-style structural presumption to simplify and strengthen merger review. Under the US structural presumption, if the plaintiff shows that a merger significantly enhances concentration, the merger is presumed to substantially lessen competition and the burden shifts to the merging firms to rebut the presumption. It is then up to the merging parties to show that the merger is not likely to harm competition, which they typically attempt to do by demonstrating that: (i) the relevant market is ill-defined; (ii) the merger will generate significant efficiencies that will be passed through to consumers; or (iii) entry or expansion by competitors will mitigate post-merger price increases.

The recommendation to adopt a structural presumption in Canada is reiterated by the Bureau in its submission to the ISED consultation.¹⁵ In that submission, the Bureau claims that the requirement that it must demonstrate that the merged firms will likely be able to exercise materially greater market power “has been further interpreted as requiring proof that the merger is likely to result in materially higher prices or materially lower levels of non-price competition relative to those that would likely exist in the absence of the merger. In turn, this has been interpreted as requiring an assessment of the likely magnitude, duration and scope of these price and non-price effects.”¹⁶

The Bureau further asserts that these economic tests are complex, and that the Bureau is required to prove them on a balance of probabilities, citing the ISED consultation document’s assertion that these tests may be highly impractical.¹⁷ Repeal of section 92(2) of the Act, which prohibits the Tribunal from finding an SLPC solely on the basis of evidence of concentration or market share is a minimal initial step towards a structural presumption, according to the Bureau. The Bureau further recommends the structural presumption should be incorporated in legislation, with ‘defined thresholds’, which could be based on post-merger concentration or market shares and changes in those levels. 

The Bureau also recommends revisions to the Act to deal with the acquisition of ‘emerging competitive threats.’ The Bureau claims that jurisprudence in recent cases places too high a burden on the Bureau to prove anti-competitive harm, citing the Commissioner of Competition v Vancouver Airport Authority abuse of dominance case. The quote cited in a Bureau submission says that “the Commissioner has the burden to establish that new entrants would likely have entered or expanded in the relevant market, or would be likely to do so, ‘within a reasonable period of time, and on a sufficient scale, to effect either a material reduction of prices or a material increase in one or more levels of non-price competition, in a material part of the market’.”¹⁸ 

The Bureau claims that its burden to prove these elements on a balance of probabilities may be very difficult or impossible when the acquired firm is still developing its products that would compete with those of other competitors. The Bureau’s concern is that “(p)redicting the future is hard at the best of times, but it can be particularly problematic in industries characterized by rapid technological progress,”¹⁹ and contrasts the requirements under Canadian law with case law in the US, citing the DC Circuit Court of Appeals in United States v Microsoft Corp²⁰ and claiming that the Act does not provide the Tribunal with the same powers to protect the competitive process. 

Finally, the Bureau notes that the Australian Competition and Consumer Commission (“ACCC”), the UK government, and the US Senate have proposed amendments to ensure that competition laws protect emerging competitors. The Bureau’s recommendation 2.3 in its February 2022 submission is: “The standards established from analysis of more traditional industries are not suitable for assessing acquisitions of emerging competitors in the digital economy. A more workable standard would provide additional flexibility to protect the competitive process.”²¹ 

ISED Merger Reform Proposals²²

ISED’s main concern for merger review is focused on the adequacy of the current merger provisions to police acquisitions of potential innovators, some of which may not be detected because they do no cross notification thresholds, while others may not be blocked because the competitive harm that may be caused by such mergers “is too difficult to forecast with precision at the time of the acquisition”²³, and then may be too late to remedy once harm becomes apparent.²⁴ ISED has in mind that digital economy start-ups, often enabled by lower barriers to entry for technology firms, will be acquired by incumbents with the hope that this will ‘help them stay on the right side of any disruptive technology’ or with the intention of suppressing the acquired firm’s innovation. This would reduce competition for the incumbent in overlapping or adjacent markets. 

ISED notes that such acquisitions—often referred to as ‘killer acquisitions’, or acquisitions of ‘nascent’ competitors—may also provide consumer benefits, since the prospect of an eventual sale to an incumbent creates a financial incentive for the start-up to invest in product development in the first place. Further, the acquisition itself can lead to lower prices and faster product adoption because of the incumbent may have the financial means, distribution networks and other advantages that the start-up lacks. 

Identifying and stopping potential ‘killer acquisitions’ may be challenging if the acquired firm is small with limited assets in Canada, as it may fall below pre-merger notification thresholds. For transactions that are identified and reviewed, ISED notes two substantive challenges. First, the harms from these acquisitions often relate to non-price dimensions of competition such as innovation, which ISED claims may be given less weight by the Tribunal and appeals courts because such non-price effects are difficult to quantify. Second, the Bureau has the burden to demonstrate, on a balance of probabilities, that any harm to competition from a merger is ‘likely’ to happen, that such harm will occur within a ‘discernable’ timeframe, and that such harm will likely be ‘substantial’. ISED claims that because of the “complexity, dynamism and pace of change in many markets, especially digital ones, these specific tests may be highly impractical.”²⁵

One possible remedy to these challenges noted by ISED is ex post reliance on the abuse provisions, but for several reasons this is inadequate.²⁶ ISED notes some proposals made in other countries to better address ‘uncertain competitive harm before it happens’. These include:

• the balance of harms approach proposed by the UK Furman panel, which would take into account both the likelihood and magnitude of potential harms and benefits of mergers, rather than just the likelihood of potential harms as is the case under section 92;
• the proposals from Australia’s Digital Platforms Inquiry to make certain merger provision requirements more explicit, including the likelihood that a merger would take out a potential competitor, and the nature of the assets being acquired; and
• the proposal from the US Senate to lower the test for harm from a merger to ‘an appreciable risk of materially lessening competition.’ The US Senate proposal would also shift the burden of proof in some circumstances, such as significant increases in concentration, acquisitions by dominant firms, and mergers with value exceeding $5 billion. 

ISED suggests that “(s)imilar measures could be considered in Canada either for transactions or firms of certain sizes, or in particularly concentrated industries. Alternatively, a more stringent competition test, or threshold for notification, could be the state of affairs for designated sensitive sectors.”²⁷ 

ISED also questions whether the current efficiencies defence should be retained. As noted above, the Bureau recommends that the current efficiencies defence be repealed and that efficiencies instead be added to the list of factors in section 93 of the Competition Act. 

3. Error Costs in Law Enforcement

The error-cost framework was first proposed in law and economics scholarship by Richard Posner²⁸ with an extension made to antitrust by Paul Joskow and Alan Klevorick.²⁹ Frank Easterbrook also made an important contribution to this literature in 1984.³⁰ 

Under the error-cost framework, optimal legal rules are those that minimize total social costs, which include the costs of ‘false positives’ (incorrectly condemning conduct that is not harmful), the costs of ‘false negatives’ (incorrectly allowing conduct that is harmful), and the costs of enforcement. The costs of false positives consist of harms to consumers or the economy when efficiency-enhancing or innovation-enhancing mergers are blocked or not pursued by merging firms. The costs of false negatives consist of consumer harms from higher prices or less innovation when anti-competitive mergers are allowed. Enforcement costs include investigation and litigation costs for governments and private parties.

The Bureau’s submissions and the ISED consultation focus primarily on the harms from false negatives driven by perceived underenforcement, while largely ignoring the costs of potential overenforcement if all proposed reforms were to be adopted. This differs from some of the Bureau’s past policy statements. For example, in its ‘Big Data and Innovation’ paper in 2017, the Bureau noted that:

… an uninformed or overly interventionist enforcement approach risks chilling investment in the accumulation and use of big data through legitimate means, and losing out on significant benefits to competition and innovation. On the other hand, an approach that is too lax risks turning a blind eye to uses of big data that are harmful to competition and consumers. One goal of this discussion paper is to prompt discussion on how the Competition Bureau … should strike a balance in enforcing the Competition Act … in cases involving big data.³¹

A stricter merger law, while blocking more harmful mergers than might otherwise have been allowed (reducing harms from false negatives), is also likely to block more mergers that are benign and may have been beneficial if they were allowed (increasing harms from false positives). An error-cost framework attempts to find a balance. Assuming that the section 96 efficiencies trade-off is eliminated, a false positive is most likely to occur when a merger is blocked or not pursued by firms even though consumers would benefit on net, either because anti-competitive effects are overstated and/or efficiencies are understated. 

While the economic tools for assessing the effects of mergers on prices are well-developed, those available for estimating a merger’s effects on innovation – be they negative or positive – are likely to be less precise, or at the least require much stronger assumptions. The reform proposals neglect the possible incentives created through mergers to innovate, even in cases where market power might exist. Mergers internalize pricing and investment decisions of the merging firms. In cases where investments have significant spillover effects that benefit rivals, a merger can allow for the capture of a greater share by reducing spillovers of innovation returns, and thereby be expected to enhance the merged firm’s incentives to invest in innovations. Therefore, the net effect of a merger is, in general, a priori ambiguous.³² Given the challenges of estimating non-price effects, very strict merger laws are more likely to block some innovation-enhancing mergers. Similarly, mergers in traditional industries can generate significant efficiencies, including from scale economies (for example by reducing fixed costs through the sharing of underutilized common facilities and equipment), and these efficiencies can benefit consumers if producers reduce their prices in response to lower costs. Assuming the efficiencies trade-off is no longer a feature of Canadian merger review, the risk of harms from false positives would be compounded by the significant reforms to the competitive effects test that are proposed by the Bureau.

4. Canadian Merger SLPC Jurisprudence

The Meaning of ‘Substantial Lessening or Prevention of Competition’

“Substantial”

The Tribunal has identified two dimensions to the SLPC analysis. First, is the merged entity likely to have the ability to lessen competition relative to how the market would have evolved ‘but for’ the merger? Importantly, the Tribunal has found the Commissioner does not have to prove that there will likely be price increases with the merger to meet the SLPC test: it is sufficient that the merged entity has the ability to increase price or reduce quality, service, or product choice.³³ 

Second, is the difference between the level of competition with and without the merger substantial? Substantiality is determined by the likely effect on market power. Would competition likely be substantially greater ‘but for’ the implementation of the merger or proposed merger, through the merged entity’s ability to profitably influence price, quality, service, advertising, innovation, or other dimensions of competition?

In its assessment of substantiality, the Tribunal assesses three components. 

• The first component is degree or magnitude, which considers whether the merger enable or is it likely to enable the merged entity to exercise materially greater market power. Said differently, post-merger, would the merged firm have the ability to set prices that are likely to be materially higher, or are non-price dimensions of competition likely to be materially lower? 
• The second component is scope, which considers whether the lessening or prevention of competition affects the entire relevant market or a material part of it. 
• The third component is duration, which deals with whether the ability to set materially higher prices or reduce non-price dimensions of competition is likely to be maintained for approximately two years.

The Commissioner has the burden of demonstrating the SLPC. In determining whether the Commissioner has discharged this burden, the Tribunal assesses quantitative and qualitative evidence related to both the price and non-price dimensions of competition.³⁴ The Commissioner is not required to quantify any anti-competitive effects under section 92.³⁵ Instead, the Commissioner, must “always adduce sufficiently clear and convincing evidence, and he bears the burden to demonstrate, on a balance of probabilities, that the merger lessens or is likely to lessen competition substantially, as well as the basic facts of the “but for” scenario that are required to make that demonstration.”³⁶

Proof on a ‘balance of probabilities’ is the standard that must be met, which means the Tribunal must be of the view that it is “more likely than not” that the merger will result in a SLPC: “Mere possibilities are insufficient to meet this standard.”³⁷ The “more likely than not” standard was established by the Supreme Court of Canada in FH v McDougall, where it rejected the possibility of a higher standard: “while the concept of ‛51 percent probability’ or ‘more likely than not’ can be understood by decisionmakers, the concept of 60 percent or 70 percent probability cannot.”³⁸ In the same case, the Court also said evidence “must always be sufficiently clear, convincing and cogent to satisfy the balance of probabilities test.”³⁹ 

There are two recent concluded litigated merger cases in which the Commissioner did not meet his burden to establish an SLPC.⁴⁰ In Parrish & Heimbecker, the Commissioner proved the acquisition would cause ‘some’ lessening of competition for the purchase of wheat (but not for the purchase of canola),⁴¹ but he did not meet his burden to prove substantiality. 

In Parrish & Heimbecker, the Tribunal considered both relative and absolute levels of price variations. The Commissioner claimed predicted price variations of between 0.05% and 0.2% of the Cash Price for canola, and between 0.60% and 1.1% of the Cash Price for wheat, based on a merger simulation model, were material. The Tribunal disagreed, finding that “predicted price variations representing such a small fraction of the pre-Acquisition price for wheat or canola at the Moosomin or Virden Elevators are immaterial, especially in light of the fact that a merger simulation model will always predict a price increase”⁴²; for purchases of canola, price variations of “0.2% or less…are de minimus” and for the purchase of wheat, “price variations reaching at most 1.1%...are very minor and far from substantial in this market.”⁴³ In its conclusion on relative price variations, the Tribunal concluded that:

…the Commissioner has not presented any compelling argument nor any clear and convincing evidence regarding the materiality level (in terms of percentage) that should apply to the substantial lessening of competition analysis in this case… Similarly, the Commissioner has submitted no analysis nor any evidence to demonstrate that, in the particular circumstances of this case, the acceptable materiality level for a price decrease could be as low as around 1% or less.⁴⁴

The Tribunal also said it is not aware of any cases in Canada or elsewhere where a court has concluded that a predicted price increase of around 1% met the threshold of substantiality, and noted that since merger simulation models always predict price increases, “absent expert evidence allowing it to conclude differently, relative price variations predicted by a merger simulation model have to be more than 1% in order to have any significance or materiality.”⁴⁵

The Commissioner argued that the predicted absolute price increases matter to farmers. The Tribunal, however, did not find “clear and convincing evidence allowing it to conclude, on a balance of probabilities, that a few cents a bushel matter to farmers and to their behaviour in the market. On the contrary, the evidence from farmers is in respect of amounts much higher than… the predicted price increases of Dr. Miller.”⁴⁶ The parties’ expert, Ms. Sanderson, provided evidence that the predicted price variation was significantly lower than within-day daily fluctuations of the price of wheat and canola, and her analysis suggested that a price increase of less than 10 cents per bushel cannot be material.⁴⁷ The Tribunal agreed, noting that the absolute value of the price variations of the Commissioner’s expert were well under 10 cents. In summary, the Tribunal concluded that the price variations are “rather of a small magnitude and immaterial, consistent with the fact that P&H faces considerable competition from several rival Elevators and Crushers to constrain material price decreases after the Transaction.”⁴⁸

Similarly, in Rogers-Shaw, the Commissioner did not meet his burden of establishing an SLPC in part because he did not establish that his proposed ‘but-for’ price was valid.⁴⁹ The Commissioner’s expert proposed that, but for the merger, Shaw’s prices would have remained the same between March 13, 2021 and the present time but for the acquisition, but the Tribunal instead determined that Shaw did demonstrate with ‘clear and compelling evidence’ that its March 13, 2021 prices were introductory and would have increased independent of the acquisition in the ordinary course of business. ⁵⁰ The Tribunal concluded that the Commissioner “has not established that the November 2021 increase in some of Shaw Mobile’s prices was attributable to the execution of the Arrangement Agreement,”⁵¹ and “Shaw has demonstrated that the price increase it implemented in November 2021 occurred in the ordinary course of business.”⁵² The Commissioner would therefore have lost this point even if the merging firms had the burden of proof. 

Substantial Prevention of Competition

In Tervita, the Commissioner alleged that Tervita’s acquisition of Complete’s Babkirk facility would likely result in a substantial prevention of competition (“SPC”) because independent entry by Babkirk in competition with Tervita would have reduced Tervita’s market power.⁵³ The ‘prevention’ and ‘lessening’ branches are essentially the same because they both focus on whether the merger entity will likely be able to exercise materially greater power than it could without the merger⁵⁴, but there are differences because a lessening of competition involves increasing a firm’s market power, while prevention involves a firm retaining its existing market power.

In its appeal to the Supreme Court of Canada, Tervita agreed that a SPC analysis is forward looking, but argued projections into the future should consider the parties ‘as they are’, and the Tribunal made a “fundamental error” by focusing “not on the merger between Tervita and [the target firm], but rather on how competition might have developed looking years into the future.”⁵⁵ That is, the acquired firm must be “be a competitor based on the assets, plans and businesses of the party at the time of the merger.”⁵⁶ In reasoning that has important implications for the analysis of future ‘killer acquisitions’, the Supreme Court disagreed with Tervita and gave the Bureau and Tribunal wide latitude to project the acquired entrant’s likely conduct into the future. 

The Supreme Court of Canada confirmed the Bureau’s approach for assessing whether a firm with market power would use a merger to prevent competition that would otherwise arise. This involves comparing the likely competition in the market with the merger to competition in the market ‘but for’ the merger. To understand the “but-for” market involves consideration of: (i) the identity of the potential competitor; (ii) whether the potential competitor would likely enter the market; and (iii) whether the entrant’s effect on the market would likely be substantial. 

With respect to the identity of the potential competitor, the entrant could be either of the merging firms or a third party entrant.⁵⁷ With respect to the second and third elements, Tervita argued that the intention of section 92 is to establish a merger test that “provides certainty to Canadian businesses.”⁵⁸ The Supreme Court of Canada disagreed, finding that “likely does not require certainty” and there is only one civil standard of proof—on the balance of probabilities—citing the Supreme Court’s own decision in McDougall. The Supreme Court did not increase the Bureau’s burden to prove the likelihood of entry beyond the universal civil standard.

The Supreme Court of Canada’s decision sets out factors to be considered when assessing the likelihood of entry. First, any factor that could influence entry should be considered, including plans and assets of the merging party, current and expected market conditions, and other section 93 factors.⁵⁹ Second, if the evidence is only that there is a possibility of entry ‘at some point in the future’, there can be no finding of a substantial prevention of competition.⁶⁰ Third, the Supreme Court agreed with the Federal Court of Appeal (“FCA”) that the timeframe for entry must be ‘discernable’—and while timing need not be a “precisely calibrated determination…there must be evidence of when the merging party is realistically expected to enter the market in absence of the merger.”⁶¹ Absent such evidence, “the timing of entry is simply speculative and the test of likelihood of prevention of competition is not met.”⁶² Fourth, “(e)ven where there is evidence of a timeframe for entry, the farther into the future predictions are made, the less reliable they will be. The Tribunal must be cautious in declaring a lengthy timeframe to be discernible, especially when entry depends on a number of contingencies.”⁶³ 

Although Tervita sought to limit the Tribunal’s ability to look into the future to what can be discerned from the merging parties’ assets, plans, and business at the time of the merger, the Supreme Court of Canada disagreed, finding: “(t)here is no legal basis to restrict the evidence the Tribunal can look at in this way.”⁶⁴ The FCA found, and the Supreme Court agreed, that when assessing the likelihood of entry, how far into the future the Tribunal can look to assess likelihood of entry should be determined by the lead time required to enter a market due to barriers to entry,⁶⁵ and the ‘lead time’ is “the inherent time delay that a new entrant, facing certain barriers and acting diligently to overcome them, could be expected to experience when trying to enter the market.”⁶⁶ 

The Supreme Court of Canada further explained that ‘lead time’ “should not be used to justify predictions about the distant future,”⁶⁷ and in some cases lead time will be short—in which case the timeframe may be sufficiently definite to meet the ‘likely’ test. However, in other cases—for example when the lead time can be ‘some years’ because of product development or regulatory approval process— “the lead time may be so lengthy that a determination of the probability of market entry at the far end of that timeframe would be influenced by so many unknown and unknowable contingencies as to render such a prediction largely speculative.”⁶⁸ Finally, “(l)ead time is an important consideration, though this should not support an effort to look further into the future than the evidence supports,”⁶⁹ and “(f)actual findings about what a company may or may not do must be based on evidence of the decision the company itself would make; not the decision the Tribunal would make in the company’s circumstances.”⁷⁰

In a 2021 submission to the OECD,⁷¹ the Bureau wrote that the substantial prevention of competition analysis by the Tribunal and the FCA “offers support for SPC findings in Canada where the path to effective entry is, to some extent, both winding and protracted.”⁷² It added that “the Tribunal’s Tervita decision shows … that path to viable entry need not be short or direct.”⁷³

In its response to Professor Iacobucci’s discussion paper⁷⁴, the Bureau cites to a discussion in the Tribunal’s discussion in Vancouver Airport Authority that is based on the Supreme Court’s Tervita conclusions. The Bureau claims that:

To obtain a remedy, the Commissioner must have sufficient evidence to prove these elements on a balance of probabilities. While this may be possible in a traditional industry, such a task may be particularly difficult—or even impossible—when it involves the acquisition of a firm that is still developing the products that would challenge other competitors. Even when it is uncertain, or where there is only a low probability that an emerging firm would develop a competitive product, any acquisition of that business completely extinguishes this possibility. Predicting the future is hard at the best of times, but it can be particularly problematic in industries characterized by rapid technological progress. [footnote cite to Hemphill and Wu 2020 omitted]⁷⁵

The Bureau further claims in its response that “(t)his Canadian case law stands in stark contrast to case law in the U.S.” ⁷⁶, and cites a passage from the D.C. Circuit Court of Appeals in Microsoft that says: “it would be inimical to the purpose of [U.S. competition law] to allow monopolists free reign to squash emerging, albeit unproven, competitors at will” ⁷⁷. The Bureau seems to believe that Canadian case law does allow such squashing of emerging competition—notwithstanding the Supreme Court’s ruling in its favour in Tervita—and complains that “Canadian courts have no such overarching power or provision in the Act to protect the competitive process. Instead, the Act requires the Commissioner to identify, in the particular context of each case, the “concrete market opportunities” through which the emerging business would bring about greater competition.” ⁷⁸ 

It is not obvious that the Tervita prevention framework unduly constrains the appropriate enforcement of ‘killer acquisitions’. As the Bureau itself noted in its submission to the OECD, the Tervita entry framework supports enforcement of acquisitions when the path is both ‘winding and protracted’. The FCA establishes a flexible approach to the timing of entry that allows an extension of the timing for consideration of entry beyond the temporal dimension of barriers to entry as long as it can be demonstrated that entry is still ‘poised’ at a later date.

Furthermore, the Supreme Court of Canada’s finding that the likelihood of entry should not be established with ‘speculative’ evidence is consistent with an error-cost framework. Giving the Bureau the benefit of the doubt in a prevention case is appropriate when evidence or theory establishes that entry is more likely than not (and would erode the acquiror’s market power) in similar circumstances, as is the case in ‘lessening’ cases, where economic evidence points to a likelihood the merged entity will have the ability to exercise market power without needing to prove the merged entity will do so (under section 92). 

SLPC and Non-Price Effects

ISED cited a concern that harms to non-price dimensions of competition were given less weight and were harder to quantify. Therefore, reform is proposed. However, this misunderstands the law and the jurisprudence. The Tribunal has not required the Commissioner to quantify non-price effects as part of the section 92 competitive effects test. Any quantification of anti-competitive effects that are ‘quantifiable’ applies only to the section 96 efficiencies trade-off. This is confirmed by the Supreme Court of Canada in Tervita:

… the Tribunal was able to consider evidence as to the effect on the market of the emergence of likely competitors, whether acceptable substitutes existed, and so on. Section 93 expressly permits the consideration of these factors in and of themselves. Ordinarily, the Commissioner would also use the evidence bearing on those factors to quantify the net effect of those factors on the economy in the form of deadweight loss. However, the statutory scheme does not bar a finding of likely substantial prevention where there has been a failure to quantify deadweight loss, and thus the Commissioner’s failure to do so in this case was not fatal to the s. 92 determination. (emphasis added)⁷⁹

In all litigated merger cases, the Tribunal has explicitly and repeatedly stated that the Commissioner can establish an SLPC using evidence of the likely effects of a merger on price or non-price competition. For example, in Commissioner of Competition v CCS Corp, the Tribunal stated:

In assessing the intensity of price competition, the Tribunal focuses upon whether prices are likely to be higher than in the absence of the merger. In assessing the intensity of non-price competition, the Tribunal focuses upon whether levels of service, quality, innovation, or other important non-price dimensions of competition are likely to be lower than in the absence of the merger. This focus ensures that the assessment of the intensity of price and non-price dimensions of competition is relative, rather than absolute, in nature. (emphasis added)⁸⁰

The Commissioner has not lost a case before the Tribunal because of a failure to show that a lessening of non-price dimensions of competition was substantial. The Commissioner has alleged non-price effects under section 92 often as a secondary claim to price effects. In Secure, while the Tribunal rejected most of the quantified non-price effects alleged by the Commissioner under section 96, the Tribunal accepted the Commissioner’s SLPC evidence of anti-competitive harm of both price and non-price effects based on estimated price effects (price increases of at least 5%) and estimated loss of service (increased travel costs), high market shares and margins, along with supporting evidence for the section 93 factors (for example, the merger eliminated Secure’s closest competitor in the relevant markets which provided vigorous price and non-price competition to Secure,⁸¹ no acceptable substitutes, and significant barriers to entry or expansion).⁸² Although there have been no litigated cases where the Bureau sought to block a merger solely based on non-price effects, there is no reason to believe that the current law or jurisprudence make it too difficult for the Bureau to obtain an order in such circumstances if they were to arise in future.

In summary, there is no case for the competition test in section 92 to be strengthened on the basis that non-price effects are too difficult for the Tribunal to assess. The Tribunal has clearly stated the Commissioner does not need to quantify non-price effects under section 92, and the Tribunal’s decision in Secure demonstrates the existence of a substantial lessening of competition on the basis of non-price effects can be inferred from market shares and the section 93 factors, with no requirement that the Commissioner adduce evidence of a causal relationship between market power and non-price effects. At the time of writing, the Tribunal’s decision in Secure is under appeal.

The Bureau’s Claim that Establishing an SLPC Requires Complex Economic Tests 

The Bureau claims that establishing the likely magnitude, duration, and scope of price and non-price effects for an SLPC requires ‘complex’ economic tests. The Tribunal’s decisions in recent litigated cases suggest that ‘complex economic tests’, and in particular regression analysis and merger simulation models, were not necessary for the Bureau to establish an SLPC under section 92, although their use was likely influential to the decisions.

Regression analysis and merger simulations are regularly used by economists within enforcement agencies around the globe that are not required to quantify anti-competitive effects for an efficiencies trade-off. Such analyses are used because they are helpful for prosecutorial decisions. The data and information on factors like volume of sales, share of sales, variable margins, and prices required to execute these economic analyses are already collected through the ordinary course of an investigation. Further, even if these “complex” analyses were not used for assessing an SLPC, these same analyses are regularly used to implement the hypothetical monopolist test in market definition. If a structural presumption were adopted, market definition would require similar analysis and resources as those used in the economic analyses related to establishing an SLPC.

It is also the case that significant resources are already devoted by the Commissioner and the parties to market definition and share calculations. Consider, for example, Superior Propane, which pre-dates the Supreme Court’s Tervita decision setting out the current section 96 quantification framework. In Superior Propane, the Commissioner argued that “market power can be inferred from various factors such as high market shares and concentration, the high barriers to entry, the removal of ICG as a vigorous competitor, the lack of foreign competition and the fact that there is no effective remaining competition.”⁸³ The Commissioner’s market share expert, Professor West, calculated market shares in 74 local markets, and identified markets where combined shares were within certain bands. Two of the Commissioner’s other experts, Professors Schwindt and Globerman, prepared a joint expert report in which they defined product and geographic markets. Their opinion, based on their market definition, changes in concentration, and analysis of barriers to expansion, was that there was sufficient evidence that competition will be reduced in the relevant markets. Neither report undertook an econometric analysis of likely price (or non-price) effects of the merger. 

The parties’ experts, Professors Carlton and Bamberger, provided a joint expert report critiquing Professor West’s market share calculations, as well as the conclusions of Professors Schwindt and Globerman. The Tribunal also considered the Commissioner’s expert evidence and other evidence on barriers to entry and the other section 93 factors. 

Another Commissioner expert, Professor Ward, provided evidence of likely post-merger prices increases based on a demand estimation and merger simulation model that was the basis of the Commissioner’s empirical estimates of deadweight loss in the section 96 trade-off. To reach its conclusion that the merger would likely result in a substantial lessening of competition, the Tribunal “considered the evidence of market shares and concentration provided by Professors West, Schwindt and Globerman and the econometric evidence of Professor Ward on the ability of the merged entity to impose unilateral price increases.”⁸⁴ It is unclear whether the Tribunal would have reached this conclusion without econometric evidence of price increases. 

The Tribunal also concluded that the merger was likely to prevent competition in Atlantic Canada, based on evidence of ICG’s plans to expand in Atlantic Canada, as well as evidence of high market shares, the evidence of high barriers to entry and the lack of evidence that entry did occur in the past.⁸⁵ Notably, the Tribunal reached this conclusion without any econometric evidence of price effects of the prevention of competition in Atlantic Canada. 

In Tervita, the Commissioner presented substantial evidence on the boundaries of the relevant product and geographic markets. The Commissioner prevailed on product market definition. The Tribunal did not reach a firm conclusion on the precise boundaries of the geographic market, finding that the geographic market was at least as large as the Contestable Area as identified by Tervita’s expert. Ultimately the Tribunal found it would not matter if the geographic scope of the relevant market included additional customers beyond the Contestable Area, because CCS would remain the sole supplier of Secure Landfill services to any reasonably defined broader group of customers.⁸⁶ 

The Tribunal then used the analytical framework described above to determine whether the acquisition would likely result in a substantial prevention of competition. The Bureau’s expert, Dr. Baye, provided evidence based on a regression model that a decrease in prices of at least 10% was prevented by Tervita’s acquisition of the Babkirk facility. The Tribunal did not give “significant weight to that economic theory or to those regression models in assessing the magnitude of the likely adverse price effects of the Merger” for a number of reasons, including that his models did not control for costs, and that they were not based on CCS’s data.⁸⁷ The Tribunal did conclude that the merger likely prevented price reductions of 10%, which was relevant for the section 96 analysis, but this evidence was based on evidence from documents and fact witnesses, not the econometric evidence of Dr. Baye. The Supreme Court concluded that “(w)hile the Tribunal’s treatment of the asserted 10 percent reduction in prices that would allegedly have been realized in absence of the merger was flawed, there was sufficient other evidence upon which it could find a substantial prevention of competition as a result of the merger.”⁸⁸

In addition, in Rogers-Shaw, the Commissioner’s economic expert, Dr. Miller, provided evidence of price effects based on a merger simulation model, which was disputed by Rogers’ expert, Dr. Israel. The Tribunal concluded that “after adapting Dr. Miller’s model to address the shortcomings discussed [in the decision], Dr. Israel persuasively demonstrated that the model would not have predicted a material price increase in Alberta or British Columbia. In other words, the Tribunal finds that the Commissioner’s quantitative evidence of predicted price effects of the Merger and Divestiture was not reliable and substantially overstated. The Tribunal agrees with Dr. Israel that Dr. Miller’s predicted post-Merger price increases are highly doubtful, for the reasons set forth [in the decision].”⁸⁹

The Tribunal’s ultimate conclusion that the merger would not likely result in a lessening of competition (let alone a substantial lessening) was not based on the Commissioner’s failure to establish price effects through a merger simulation model, although it is possible that establishing price effects may have persuaded the Tribunal to reach a different conclusion. The Tribunal rejected the Commissioner’s key claims, including that the divestiture of Freedom to Videotron would result in Freedom being a less effective competitor, that Rogers’ acquisition of Shaw Mobile would likely give rise to anti-competitive unilateral effects, and that the merger would likely facilitate the exercise of collective market power.⁹⁰ In reaching these conclusions, the Tribunal appeared to rely mainly on qualitative evidence and testimony from fact witnesses.

5. No Need for Specific Reforms

Codifying a Structural Presumption in Canada

A core Bureau recommendation for strengthening the merger provisions of the Act is to adopt a structural presumption, under which the Bureau would establish that a merger is anti-competitive by demonstrating that it would result in a significant increase in concentration. The merging firms would have the opportunity to rebut this presumption. 

As described above, under the US structural presumption, the plaintiff can establish prima facie illegality of a merger in US courts by showing that the merger would result in a significant increase in, and high level of, concentration. Once established, the presumption of illegality can be rebutted by the merging firms by showing that market shares do not accurately predict competitive effects, which may be done by undermining the factual foundations of the presumption—for example by proving that the relevant market is not well defined or market shares overstate the potential for competitive effects—or by providing evidence of market conditions that offset the presumed effects of the increase in concentration. Such evidence may be that attempted post-merger price increases would be disciplined by entry, or that the merger would result in offsetting efficiencies that are sufficient to prevent prices from increasing or otherwise offset any anti-competitive effects. If the presumption is undermined, then it ‘bursts like a bubble’,⁹¹ but if the merging firms simply prove offsetting effects, the presumption remains, although in weakened form.⁹² 

Assuming the current SLPC test remains the same, adopting a structural presumption would mean that the Bureau no longer has the burden to prove entry or expansion by rivals would not eliminate the SLPC. Therefore, the ‘duration’ element of the Tribunal’s three-part ‘substantiality’ test would be removed from the Commissioner’s burden. However, even with a structural presumption, the Bureau would very likely still advance evidence, including economic analysis, to address the ‘magnitude’ and ‘scope’ aspects of the SLPC test detailed by the Tribunal. 

The economic basis of the structural presumption is founded on the proposition that the loss of a significant competitor in a concentrated market is likely to lead to an increase in market power, and concentrated markets often have significant barriers to entry.⁹³ These are the same foundations the Tribunal uses to determine whether a transaction is likely to enable the merged entity to exercise materially greater market power (the ‘materiality’ requirement) covering a material part of the relevant market (the ‘scope’ requirement). 

One potential benefit of a structural presumption is saving litigation costs by encouraging some firms to forego a merger that would otherwise be litigated. This is a social benefit in the error-cost framework because it reduces the costs of litigation, especially for mergers where the firms may have a weak case. Canada’s current merger framework with the section 96 efficiencies trade-off reduces the likelihood that firms will walk away from or not pursue a highly concentrating transaction as their transaction may prevail on the efficiencies defence even if a structural presumption were to exist. Assuming amendments are made to section 96, firms contemplating a transaction that is likely to meet the Tribunal’s SLPC requirements should realize they risk substantial litigation costs without having their merger allowed. This will be the case even without a structural presumption.

The Bureau’s recommendation to enact a structural presumption is described as a meaningful way to save enforcement costs the Bureau incurs under section 92. Such perceived savings may not come to pass. Any litigation – or work preparing for possible litigation – requires considerable investigative, legal, and economic work, at meaningful cost to the Bureau and merging firms. The Parrish & Heimbecker case is an example of a case where the central dispute related to relevant market definition, including both product market definition and geographic market definition. Most of the economic modelling and analysis undertaken in the Parrish & Heimbecker case was used to address market definition and competitive effects. This is not unique. 

It is not uncommon for the economic analysis of market definition and competitive effects to be intertwined. For example, merger simulation models are regularly used to implement the hypothetical monopolist test, and also provide an indication of the extent to which the merger provides the parties with the ability to exercise materially greater market power. Such issues are currently vigorously contested, and they are likely to remain so even with a structural presumption. In such circumstances, any savings in litigation costs for the Bureau may be small.

Litigation costs incurred by the Bureau are not the only relevant enforcement costs. In the error-cost framework, the costs to the merging firms are also relevant. These are high but may be reduced if quantification of efficiencies under the section 96 trade-off is no longer required in the future. The other important costs to consider are potential harms resulting from false positives. Replacing current Tribunal jurisprudence with an untested – and to be defined – structural presumption for Canadian merger review will increase uncertainty. Transactions are less likely to be pursued in a more uncertain regulatory environment, making it more likely some beneficial and benign transactions are not pursued. 

Finally, a structural presumption is not likely to be applicable for ‘killer acquisitions’ since the acquired firm, as a nascent competitor, would typically have minimal or even no market share. In other words, the structural presumption would not be relevant for most of the digital economy mergers over which the Bureau expresses the greatest concerns. For example, Facebook’s acquisition of WhatsApp and Instagram would not likely have met the market share thresholds under a structural presumption.

Balance of Harms Test Risks Continued Emphasis on Quantification

In its discussion paper, ISED described adopting a ‘balance of harms’ test for digital merger review to replace the current ‘balance of probabilities’ test. A balance of harms test was recommended by the Furman panel in the UK as a way to take account of the ‘likelihood’ and ‘magnitude’ of future harms.⁹⁴ In contrast, a ‘balance of probabilities’ test only considers potential harms—whatever their magnitude—if any harm is shown to be ‘more likely than not’. Under a balance of harms test, a merger would be blocked if the expected harm (equal to the likelihood of harm multiplied by the size of the harm if it occurs) is greater than the expected benefit (equal to the likelihood of the benefit multiplied by the size of the benefit). Assuming likelihood and magnitude of harm and benefit can be estimated, the balance of harms test is superior to the ‘balance of probabilities’ standard because it allow the Tribunal to block mergers that have low probability of very large harm, which would be allowed under the current standard. 

An important objection to this standard is that it can be very difficult—and costly—to estimate ‘likelihood’ and ‘effect’ with sufficient precision⁹⁵ to confidently establish that expected harms exceed expected benefits. Therefore, a ‘balance of harms’ test is likely to emphasize quantification, which the Commissioner claims is costly for the Bureau and which the Bureau’s proposed amendments seek to avoid. The magnitude of expected benefits and harms can be very sensitive to small changes in the values of the likelihoods and effects, which leads to the potential for significant false positives or false negatives. To illustrate, consider a merger where the Bureau estimates that the probability of harm is 6%, with an estimated consumer harm, if realized, of $100, (an expected harm of 6% x $100 = $6), and an estimated 50% probability of an estimated $10 in consumer benefit (an expected benefit of 50% x $10 = $5), which yields an expected net harm of $1 using a balance of harms approach. If the actual probability of harm is say, 4.9%, instead of 6% , then there would actually be a net expected benefit instead of a net expected harm. Thus, in this example if the expected harm is overstated by just over one percentage point, the balance of harms test will result in a false positive, which would lead to prohibiting a merger than has net benefits. It is unclear how the Bureau, the merging parties and the Tribunal would distinguish instances of 6% versus slightly less than 5% probability without resorting to extensive and costly quantification and “balancing” that the Bureau’s section 96 proposed reforms aim to avoid. Although this example was constructed to illustrate a point, the numbers may not be very different from those in real cases, and in particular the likelihood of harm will in often be low, since the test is mainly intended to capture low probability of harm/large magnitude of harm mergers, and the dispute will often be around small differences of low probability events. Such disputes are likely to be costly, time-consuming, and prone to error.

It is worth considering whether the Bureau investigates, or is likely to investigate in the future, a sufficiently large number of potentially problematic mergers in non-traditional industries to justify amendments of general application to deal with future digital mergers. As we discuss below, the Bureau has reviewed very few, if any, high-tech or digital economy mergers over the last several years, and we have seen no evidence that the Bureau has foregone an investigation or challenge of a high-tech merger owing of any challenges to meet the current SLPC jurisprudence standards.

The Bureau Reviews Very Few High-Tech Mergers

The merger transactions contested by the Bureau over the last two decades have almost exclusively involved firms in traditional industries. Superior Propane involved the sale of propane, Tervita involved oilfield waste landfills, Parrish & Heimbecker involved grain purchases at prairie elevators, Secure involved oilfield waste treatment, disposal and landfill services, and Rogers-Shaw involved mainly mobile phone services. None of these merger challenges involved firms in high tech industries.

The Bureau has released approximately 33 backgrounders on mergers since September 2017 (as of April 8, 2023). Very few of these mergers involved firms that could be considered high-tech, and none of the reviewed mergers involved an acquisition of a nascent competitor. The only mergers that appear to involve ‘high tech’ firms are S&P/IHS, Thoma Bravo/Aucerna, and United Technologies/Rockwell Collins. The other 30 backgrounders discuss mergers of firms in traditional industries, including pulp mills, retail gasoline and grocery stores. This is not surprising given the general make-up of Canadian industry.

In its monthly reports of concluded merger reviews, the Bureau lists 463 mergers between January 2021 and February 2023 that were either pre-notified and/or for which a request was made for an advance ruling certificate.⁹⁶ The monthly reports include the 4-digit NAICS code and the resolution of the matter (ARC, Consent Agreement, or No Action Letter). There appear to be very few mergers in any 3 or 4 digit NAICS code that appear to encompass ‘high tech’ firms. There were eighteen reviews in NAICS 541 ‘Professional, Scientific, and Technical Services’ accounting for 4% of the total. This code also accounted for one out of eleven consent agreements. This involved a proposed merger of S&P Global/IHS, which was a 2021 merger of credit ratings and analytics firms.⁹⁷ There were five notifications in NAICS 334 (‘Computer and Electronic Product Manufacturing’), including Google/Fitbit (which resulted in an NAL) and Jordan Company/CPI Intermediate Holdings (which resulted in an ARC).

6. Conclusions and Recommendations

Our review of litigated merger cases and recent Bureau reviews provides limited, if any, support for substantial amendments to the Act that would lighten the Bureau’s burden to prove an SLPC, especially if such amendments are motivated by concerns about underenforcement of the merger provisions in relation to digital economy mergers. The Act has proven to be flexible, and with two recent exceptions—which we would argue involve cases the Commissioner would have lost even with a less stringent SLPC test—the Commissioner has met the burden of proving an SLPC, including in the only merger case where the Commissioner alleged a substantial prevention of competition. 

We find little evidence that the current Act or the jurisprudence in the litigated cases will prevent the Commissioner from obtaining remedial orders for mergers that are likely to substantially lessen or prevent competition if a responsible case is brought. Furthermore, the Commissioner has consistently proven non-price effects under section 92 based on qualitative evidence, including the section 93 factors, and we find no reason to believe that the current law would not support a responsible application based on non-price effects against a digital economy merger. 

The Commissioner has not been burdened by the requirement to adduce complex economic evidence to prove an SLPC based on either non-price or price effects, and when such evidence has been adduced to prove an SLPC, it was mainly in support of meeting the Commissioner’s burden to rebut the merging firms’ evidence that efficiencies outweigh and offset the efficiencies claimed by the merging firms under the section 96 trade-off. Assuming the Bureau’s recommended reforms to the section 96 trade-off are accepted, the current quantification requirements on the Commissioner will be removed, although the Bureau may still find it valuable to use merger simulations and other “complex” economic analyses to persuade the Tribunal of a likely SLPC. 

We are unable to point to any harmful ‘Big Tech’ mergers that the Commissioner has failed to successfully challenge, or harmful mergers that it chose not to challenge, because the current law made it too difficult for the Commissioner to prove an SLPC. The Bureau has reviewed few, if any, ‘Big Tech’ mergers, and the Bureau has negotiated a consent order in the few mergers where the Bureau had concerns about the effects of the merger on innovation. Of course, the Bureau may need to review more digital economy mergers in the future, including mergers with non-traditional features such as platforms, two-sided markets, zero price goods, non-price effects including innovation, and ‘killer acquisitions.’

We note, however, that at this stage there is little information to establish whether acquisitions undertaken by digital firms are designed to ‘kill’ the target in Canada or elsewhere.⁹⁸ Many digital economy mergers that potentially harm Canadians are likely to involve global firms and will be reviewed by the US agencies, the European Commission, and other foreign authorities. The remedies obtained by these larger agencies to resolve their concerns are likely to have positive spillovers for Canadian consumers based on past experience. Most of the recent digital economy mergers that have drawn attention to alleged deficiencies in competition laws in policing non-traditional mergers have similar effects on Canadians as they have on Americans and Europeans, and remedies achieved in the US and Europe have extended beyond these jurisdictions. Given this context, it would not be prudent to amend Canada’s competition test for mergers that has worked reasonably well in traditional industries.

There may, however, be some scope for amending the SLPC test, along the lines suggested by Professor Iacobucci, which we include in our list of recommendations below.⁹⁹ We offer the following additional recommendations. 

Recommendations involving changes to Section 93:

• Amendments to section 93. We support the clarifying amendments to section 93 proposed by Professor Iacobucci. He suggests amending section 93 to include a factor that considers whether an acquisition would entrench or enhance the market position of a leading firm, including firms in the digital sector. He also suggests section 93(f) be expanded to include consideration of the removal of ‘a potentially vigorous and effective competitor’, to address concerns about nascent competitors in digital markets. Section 93(f) might also be amended to include a factor that invites the Tribunal to consider whether a merger would result in the degradation of product quality, including for products offered by firms in the digital sector. 

Other Recommendations:

• Increase Bureau funding. One of the Bureau’s main motivations for its recommendations is to limit the cost and complexity of proving cases. Lowering the Bureau’s burden will likely reduce enforcement and regulatory costs and reduce false negatives but is also likely to increase false positives. An increase in resources allocated to the Bureau would limit cost concerns and allow for a sharper resolution of cases.
• Enable merger retrospectives. As suggested in the ISED consultation paper, the enabling of more merger retrospectives would allow for the refinement of analytical approaches and applying lessons learned. Retrospectives could even be mandated to be conducted periodically—say every five years—with such a mandate possibly only applying to digital economy mergers and/or mergers of nascent competitors. Consideration could also be given to having an impartial third party—not the Bureau—conduct the retrospectives. There are many highly competent Canadian academics who would be qualified to conduct such reviews.
• Extend limitation period to three years, and possibly longer. The Bureau’s February 2022 submission noted that over the past decade, “only five out of the hundreds of acquisitions made by the largest tech firms—Google, Apple, Amazon, Facebook and Microsoft—were notified under the Act.”¹⁰⁰ Section 97 prohibits the Commissioner from challenging a merger more than year after it has been substantially completed. We agree that the limitation period should be extended to at least the three years recommended by the Bureau.¹⁰¹ This extended period may be particularly relevant for acquisitions of nascent competitors, since the Bureau would be able to monitor whether the acquiror has taken steps to minimize the target’s market roll-out or reduce innovation by either firm over a longer time period. This extended timeframe may cause some uncertainty for merging firms and may have the effect of freezing some beneficial investment, but this may be an acceptable economic cost given the possible benefits in terms of reducing the risk of false negatives in digital acquisitions.

Andy Baziliauskas, Principal, and Margaret Sanderson, Vice President, are economists specializing in matters of antitrust and competition at Charles River Associates. The opinions and conclusions expressed in this article are solely those of the authors and should not be attributed in any way to any other individual or organization.  The authors would like to thank Lisa Stockley, and the referees of the CCLR—Dimitri Dimitropoulos, Ian Cass and Chris Margison—for their valuable comments, as well as Rahim Lila for his assistance in editing draft proofs.  Any remaining mistakes are our own.