By Anthony Niblett
February 2, 2023
What should the government do if it perceives that a current law is not achieving its objectives? Instinctively, one might say the law should be changed and new legislation should be introduced. This instinct is understandable. But this instinct is not always correct.
It may be that the current legislation is suitable for achieving its objectives, but the way in which that law is being administered needs to be improved. It may be that the current law should be enforced with greater vigour, or the agency that enforces the law needs more resources and more staff. Further, it may be that proposed changes to the legislation result in significant adverse effects. A different toolkit may not only deter bad behaviour; it could deter the good.
The question of whether we need a change in the law or a change in how the law enforced has renewed relevance given the recent debate over whether Canada’s Competition Act is “fit for purpose” in today’s digital economy. Some have argued that the way the Competition Act is enforced does not achieve its objectives. But this does not necessarily mean there is a problem with the legislation.
Following the meteoric rise of “big tech” companies such as Amazon, Apple, Google, and Facebook, some critics of Canada’s competition policy have suggested a dramatic change in the law. They argue we should follow in the footsteps of other jurisdictions, such as Europe, that have enacted sweeping legislation to ex ante (or preventively) regulate the unilateral conduct of dominant players in the digital economy.
Such a dramatic overhaul of the Competition Act would not be in Canada’s best interests. The Competition Act is – for the most part – sufficiently flexible to deal with dominant companies who engage in anticompetitive behaviour in digital markets.
To be clear, I am not suggesting that the Competition Act is perfect. Far from it. But improvements to Canada’s competition policy can come through incremental changes to the Act and through improvements in how we enforce the law. Tearing down the existing structure of the law and building a new law may prove counterproductive.
If the problem is that the Competition Act is not being enforced properly, then the solution is to enforce the law properly. The government should ensure that the Competition Bureau has sufficient funding and staff to adequately investigate anticompetitive conduct. It should provide the agency with resources to enforce the law vigorously and effectively. It should ensure that the agency is able to provide firms with sound economic guidance about the types of conduct that are not allowed.
Some recent developments point in the right direction. The increase in funding for the Competition Bureau is a positive step. Private enforcement in abuse of dominance cases will offer an additional avenue for enforcement. Further, the administrative monetary penalties were previously too small to provide effective deterrence, though one questions the constitutionality and potential for over-deterrence of the new penalties.
But other – more dramatic – changes to the Competition Act are being explored by the federal government. The Minister for Innovation, Science and Economic Development Canada (ISED) recently announced a public consultation process. The accompanying ISED consultation paper offered a wide array of potential changes. With regards to unilateral conduct, the ISED consultation paper, released in November 2022, has an explicit focus on “big tech.” The government is exploring some potential avenues that appear to mirror Europe’s Digital Markets Act. While these prohibitions will no doubt deter anticompetitive conduct, such regulations will also likely deter companies from engaging in conduct that can be beneficial to Canadians.
A good example is the call to prohibit “self-preferencing.” Self-preferencing is difficult to define. But, generally, self-preferencing encompasses conduct where a vertically integrated firm that sells the products of others on a platform also sells its own products on the same platform and gives its own products an advantage. Some have suggested that it is concerning that Apple gives preference to its own apps on their iOS App Store, that Amazon ranks its own Amazon Basics line higher than other competing products, and that Google prioritizes their own Google Shopping price comparison service over competing comparators.
There is no doubt that self-preferencing can be anticompetitive. Such conduct can hurt Canadian consumers. For example, a platform could use its market power in an upstream market to foreclose opportunities for more efficient competitors in a downstream market. If these competitors provided better quality products and the competition would have lowered the prices in the downstream market, then foreclosing this competition is harmful. Further, platforms with market power in an upstream market could also use their upstream power to eliminate nascent threats to that power that may come from downstream markets. Lawmakers in Europe have responded by enacting legislation that ex ante prohibits large dominant players in the digital space from self-preferencing.
But Canada should not implement an ex ante prohibition on self-preferencing. A blanket rule would likely be problematic. A test that focuses on the likely anticompetitive effects of self-preferencing is better. The argument proceeds in three parts:
- Self-preferencing is difficult to define. But a rule prohibiting self-preferencing would need to be clear as to what is allowed and what is not allowed.
- A blanket rule against self-preferencing would likely capture – and deter – beneficial conduct. It would likely stifle innovation.
- The Competition Act already prohibits anti-competitive self-preferencing under section 79.
Ultimately, if anticompetitive self-preferencing is a problem in Canada (and this is a big “if”), it is not necessarily a problem with the law; it is most likely a problem with the way the law is being enforced.
I focus on self-preferencing here not because it is the most important potential change in the consultation. Rather, I offer it as an example of how implementing ex ante rules and prohibitions in competition policy can be detrimental. Unilateral conduct that may be prohibited in future amendments may or may not be problematic. Implementing an ex ante prohibition against ill-defined conduct would likely diminish the importance that we place on competition and innovation – and this would be counter-productive to Canada’s interests.
“Self-preferencing” is difficult to define
Consider the problem of what conduct will be caught by a proposed ban on self-preferencing. There does not appear to be an agreed upon definition of “self-preferencing” in the competition and antitrust literature. In a forthcoming paper in the Canadian Business Law Journal, I highlight some of the many definitions that have been put forward by commentators and scholars. There is a high degree of variability in the types of behaviour that the different definitions capture.
Some commentators offer a very broad definition of self-preferencing. They assume that the term incorporates any conduct that offers preference to its own products. If self-preferencing is defined too broadly, then desirable behaviour will inevitably be caught up in a ban.
To others, there is almost a presumption that self-preferencing under the law would be limited to the narrower anticompetitive conduct. For example, a panel of economic experts in Europe were strongly in favour of a total ban on self-preferencing, noting that “if the platform’s own supply provides a better match with consumer preferences, then there is no self-preferencing, as the outcome is based on maximization of consumer value” (Cabral et al. 2021, 21).
Why does this matter? Put simply, it is difficult to ex ante prohibit something that is difficult to define. When writing a law, the boundaries of the law need to be very clear. What is permitted? What is prohibited? Simply saying that “self-preferencing” is prohibited is unclear and confusing. The rule of law demands that the law be capable of guiding the behaviour of its subjects. If the lawmaker, the enforcement agency, and the platforms each have different definitions of self-preferencing, then this would be problematic. Vague prohibitions on self-preferencing would likely extend well beyond the types of behaviour of concern.
Some confine their definitions of self-preferencing to digital markets. The Digital Markets Act in Europe, for example, limits its prohibition on self-preferencing to a small number of large digital “gatekeepers.” But giving preference to one’s own products is not limited to the digital space. Retailers such as Loblaws sell in-house products and compete with more established products, all on the same physical platform (the supermarket.) In-house products are typically sold at lower prices. For the most part, this is good for consumers. But these in-house products may get preferential treatment. They may, for example, be advertised more favourably. They may be placed on more prominent or more easily accessible shelves.
What is the difference between a more traditional platform and that of digital platforms? The argument, it seems, is that the digital players are presumed to have greater ability to foreclose competition. But if Loblaws were the only place in town to purchase groceries, one might raise a concern that self-preferencing behaviour may have anticompetitive effects. Whether a digital platform has the ability to foreclose competition is something that should be shown rather than presumed. Indeed, one might argue that switching costs may be higher in traditional outlets than on digital platforms. There are higher costs in physically travelling to another store, rather the simply clicking on a competitor’s website.
A blanket rule prohibiting self-preferencing could deter conduct that benefits Canadians
Prohibitions on self-preferencing, where the proscribed conduct is not neatly confined to anticompetitive conduct, could very well capture all sorts of conduct that is either benign or beneficial.
What are the benefits from self-preferencing? How can it be good for consumers? First, consider some of the potential benefits of vertical integration. Platforms that sell the products of others may be able to identify gaps in the market that are not currently being met by other downstream sellers. These platforms may offer new and innovative products to fill a gap or offer similar products to compete with products sold at supra-competitive prices. The competition may generate lower price alternatives. Lower prices may also be a consequence from the fact that fewer parties are adding their own margin to sales. But a ban on preferencing your own products in ranking, for example, may diminish the incentive to engage in this type of conduct.
Firms may also integrate different products for a better overall product. For example, when you search on Google for “hairdressers near me,” the website includes not only search results, but also includes a Google map so that you can see how far away each of the hairdressers are. This, under some definitions, would constitute self-preferencing. Google is giving preference to its own maps over the maps of others.
A per se prohibition on self-preferencing could stifle this form of innovation. Such a prohibition would likely harm the very interests that competition policy is intended to protect. Would the law prohibit Google from placing a Google map of the nearest hairdresser when you search for hairdressers in your area? If so, would Google have invested to create Google maps?
Consider another example. A website may decide to hire journalists as employees, as well as publishing articles written by freelance writers. When a website publishes an article written by an employee, they are, under some definitions, preferencing their own product over the products of other freelance writers. Would this fall afoul of the self-preferencing prohibition? From a competition perspective, this is not concerning at all. But would the law be written in a sufficiently clear way such that the operators of the website know this? Would it sufficiently clear such that the enforcement agency knows this?
Per se prohibitions of conduct are good when the conduct has no redeeming qualities. They should be used when the conduct is always or almost always harmful to society. Naked price fixing, for example, is always or almost always harmful to society. It simply results in lower output at higher prices. But we are not even close to self-preferencing being always or almost always harmful to society.
It would be difficult to write a law that spells out ex ante what types of self-preferencing are beneficial and which are anticompetitive because it is difficult to ex ante identify what type of self-preferencing conduct is harmful to society and what conduct is beneficial. The Australian Competition and Consumer Commission (ACCC) is grappling with this problem. The ACCC recently called for a prohibition on digital gatekeepers engaging in self-preferencing (ACCC 2022). But in doing so, the ACCC expressly recognized that self-preferencing can have legitimate efficiencies or other business justifications that would need to be considered when developing obligations. There is a recognition by the ACCC that self-preferencing is “complex,” and assessments of whether self-preferencing is anticompetitive will require “detailed information and technical expertise.”
The Competition Act already prohibits anticompetitive self-preferencing under section 79
Anticompetitive self-preferencing – that which forecloses competition and excludes competitors – is an abuse of dominance. This conduct falls within the purview of section 79 of the Competition Act. Section 79 prohibits firms from abusing a dominant market position. It prohibits dominant firms from engaging in anticompetitive conduct that is having – or is likely to have – the effect of preventing or lessening competition substantially in a market.
Section 79 is by no means perfect. For example, questions remain over how courts have defined anticompetitive conduct. But section 79 is a flexible law that prohibits anticompetitive self-preferencing conduct by dominant digital platforms. To the extent that there is any disagreement that self-preferencing falls within the ambit of section 79, then “anticompetitive self-preferencing” could be added to section 78 as an example of conduct that constitutes an abuse of dominance. This solution is more incremental than implementing ex ante regulations. This solution would still suffer from the same problem of definition discussed above, but it would only prohibit conduct that has – or is likely to have – an effect on competition. The effects on competition should continue to be a central part of the Competition Act.
If the problem stems from the fact that section 79 is not being enforced properly and digital platforms are engaging in anticompetitive self-preferencing, then the solution is to improve enforcement of this law. More resources should be devoted to investigating these anticompetitive abuses of dominance. Further, resources could be devoted to developing stronger advisory guidelines of when self-preferencing may fall afoul of section 79. Lawyers, economists, and other technical staff should work together to help the Bureau better understand when such conduct is likely an abuse of dominance.
But a broad prohibition on self-preferencing is not the answer. There is no doubt that it would cut down on the time and money required to investigate section 79 abuses. And there is little doubt that a prohibition combined with sufficiently high penalties would deter anticompetitive self-preferencing. But such a ban would likely be over-inclusive. It would not just deter anticompetitive conduct by digital platforms; it would likely deter innovative conduct. It would likely capture practices that are beneficial for consumers and for society at large.
In sum, Canada’s Competition Act prohibits firms with a dominant market position from abusing that position. Section 79 prohibits large digital platforms from engaging in anticompetitive self-preferencing conduct. If the government perceives that this law is not being enforced properly, then there are steps they can take to improve how the law is administered. The government can ensure that the enforcement agency has sufficient resources to administer the law in line with its objectives and provide clearer guidance to businesses.
Our competition law does not currently prohibit self-preferencing that is beneficial or procompetitive. But an ex ante prohibition could have this effect. This will likely harm Canadian consumers with higher prices, fewer product choices, and stifling innovation. It would be a step in the wrong direction. The likely effects on competition and innovation matter. The likely impact on Canadians is what is important. This should continue to be focus of our competition policy in Canada.
About the author
Anthony Niblett is an Associate Professor and Canada Research Chair in Law, Economics, & Innovation at the University of Toronto’s Faculty of Law. He also holds the Canada Research Chair in Law, Economics, and Innovation. Professor Niblett holds a Ph.D. in economics from Harvard as well as degrees in law and commerce from the University of Melbourne.
Australian Competition and Consumer Commission [ACCC]. 2022. “ACCC calls for new competition and consumer laws for digital platforms.” ACCC, November 11. Available at https://www.accc.gov.au/media-release/accc-calls-for-new-competition-and-consumer-laws-for-digital-platforms.
Cabral, Luis, Justus Haucap, Geoffrey Parker, Georgios Petropoulos, Tommaso Valletti, Marshall Van Alstyne. 2021. The EU Digital Markets Act: A Report from a Panel of Economic Experts. European Commission. Available at https://publications.jrc.ec.europa.eu/repository/bitstream/JRC122910/jrc122910_external_study_report_-_the_eu_digital_markets_acts.pdf.