Steele, Alessandra --- "Too Much Power? Is Facebook's Threat To Deny Local Publishers Access A Misuse Of Market Power?" [2021] UNSWLawJlStuS 13; (2021) UNSWLJ Student Series No 21-13

---

TOO MUCH POWER? IS FACEBOOK’S THREAT TO DENY LOCAL PUBLISHERS ACCESS A MISUSE OF MARKET POWER?

ALESSANDRA STEELE

Assuming this draft code becomes law, we will reluctantly stop allowing publishers and people in Australia from sharing local and international news on Facebook and Instagram. This is not our first choice – it is our last. But it is the only way to protect against an outcome that defies logic and will hurt, not help, the long-term vibrancy of Australia’s news and media sector.^[1]

In August 2020, Facebook’s Australia and New Zealand Managing Director, Will Easton, made this threat (Threat) in response to a Commonwealth government proposal to make digital platforms pay Australian media companies a share of their Australian revenue in exchange for their access to the media companies’ online news content. On its face, this threat resembles a “refusal to deal” which could be prohibited by section 46 of the Consumer and Competition Act 2010 (Cth),^[2] the prohibition on misuse of market power.

This essay explores whether the Threat itself, or the carrying out of the Threat (Conduct), could constitute a contravention of section 46 of the CCA, and in doing so, explores some important questions regarding the application of the recently amended section 46 to digital platforms. In particular, market definition and assessment of market power for digital platforms remain relatively unchartered in competition law, and while it is clear that the ACCC and other national regulators are firmly of the view that the digital platforms do have substantial market power, the analysis on which those views are based may not survive the judicial process.

I MARKET DEFINITION AND MARKET POWER

In the Digital Platforms Inquiry, the ACCC found that Facebook has substantial market power in three separate but related markets: the social media market, the display advertising market and news referral markets.^[3] This finding gives an indication of how the ACCC would structure a section 46 case against Facebook.

Facebook criticised these findings on the basis that they were not based on sufficiently rigorous competition analysis supported by economic evidence.^[4] Facebook also disputed that there was a market for “news referral services” as found by the ACCC,^[5] and even if it did exist, that the ACCC had not correctly performed the analysis required by section 4E of the CCA.

The market definition relied upon by the ACCC is critical in the determination of Facebook’s market power: first, “too narrow a description of the market will create the appearance of more market power than in fact exists; too broad a description will create the appearance of less market power than there is”.^[6] Secondly, the ACCC will be confined to its case as pleaded; the court is not able to substitute a new or better market in lieu of the ACCC’s choice unless it has been pleaded in its defence by Facebook, which will argue for the broadest possible definition of the market to avoid liability.

The choice of market is determined by the court by following a purposive approach, with the first step being to articulate the alleged contravening conduct and the second step to ask “what market identification best assists the assessment of the conduct and its asserted anti-competitive attributes?”.^[7] In the context of section 46, “the object is to discover the degree of the defendant’s market power”,^[8] although market definition is relevant to the newly introduced “substantial lessening of competition” (SLC) test also. For this reason, it is explored in depth in the following sections.

A The Alleged Contravening Conduct

Threats can be conduct that attracts the operation of the CCA; in Rural Press v ACCC, Rural Press’ threat to enter a newspaper market where its rival had a monopoly if its rival did not retreat from a market in which Rural Press had a monopoly, was accepted by courts of all levels including the High Court of Australia as conduct attracting the operation of section 46.^[9]

For the purposes of market definition however the Threat and the Conduct relate to the same aspects of the market, namely behaviour threatening to restrict (in the case of the Threat) or actually restricting (in the case of the Conduct) access by Facebook users to certain content, that from news media publishers, as well as access by news media publishers to upload content to its platforms and present their content to users.

Section 4E of the CCA defines “market” as follows:

a market in Australia and, when used in relation to any goods or services, includes a market for those goods or services and other goods or services that are substitutable for, or otherwise competitive with, the first-mentioned goods or services.

1 News Referral Services

Relevantly to the Conduct, the ACCC has indicated that it considers there will be a “news referral services” market, although in the face of strident criticism from Google and Facebook, it ultimately declined to reach a conclusion on this issue.^[10]

Duke notes that the ACCC has a tendency to favour narrow market definitions whereas the Australian Competition Tribunal and the courts are more inclined to favour broader markets: in ACCC v Flight Centre,^[11] the narrowly defined markets proposed by the ACCC were rejected by the High Court as being artificial and lacking in commercial reality.^[12] In that case, Kiefel and Gaegeler JJ cautioned against an overly contrived market definition that did “violence to commercial reality”.^[13] It may be the case that the identification of the “news referral market” is another such example^[14] – Facebook makes this argument in its submissions on the Preliminary Digital Platforms Inquiry report, albeit without much in the way of analysis.^[15]

In Re QCMA the Trade Practices Tribunal defined a market as:

the field of actual and potential transactions between buyers and sellers amongst whom there can be strong substitution, at least in the long run, if given a sufficient price incentive.^[16]

There is a real question as to whether the “news referral services” market suggested by the ACCC fits this definition, given that there is no evidence that the publishers who might participate in this market engage in any form of actual or potential transaction with the platform in relation to the content that they post to Facebook. While Facebook permits media outlets to post content, so it does with any user who holds an account; if anything, arguably the market for news media referral posts is part of a broader market for user-generated content on the user side of the platform. There is available an argument that Facebook competes in the market for attention, not content. Having said that, the Australian authorities suggest that a lack of transactions in the news referral market is unlikely be an impediment to finding the existence of a market.^[17]

It is not within the scope of this essay to resolve the question of whether there is a market for “news referral services” as suggested by the ACCC; quite clearly it would be a contested question in any litigation in relation to the Conduct or the Threat, and the outcome of that analysis could well be make the difference in the outcome of such litigation. Ultimately, the matter will be determined on the balance of probabilities by reference to expert evidence or anecdotal evidence which the ACCC will need to bring.^[18] The challenges that may be faced by the ACCC in bringing such evidence are discussed in the following sections.

According to the ACCC’s analysis, Facebook also operates in the markets for social media services (to users) and the market for digital display advertising (to advertisers) and if Facebook has market power in those markets, then the question of whether a “news referral services market” exists would become otiose for the assessment of market power, particularly in light of subsections 46(1)(b) and (c) of the CCA.

2 Social Media Services

The ACCC notes that Facebook’s unique audience (via Instagram and Facebook) dwarfs that of its next rival, Snapchat and “accounts for a large share of the market for social media services”.^[19] Not surprisingly, Facebook considers the market to be broader and include other websites and apps including Google, YouTube, Microsoft, TikTok, and Skype, which share common features with Facebook.^[20] While many commentators view Facebook and the other digital platforms as having a monopoly or near-monopoly position, these assessments tend to be bald assertions with little factual evidence to support them.^[21] As Sanchez-Cartas and Leon note,

A common myth about multisided platforms is that monopolies are common. This thought has become popular nowadays because of the rising of social networks, such as Facebook, Twitter, or LinkedIn. However, there is no clear evidence supporting this line of thought...^[22]

While the courts are likely prepared to accept anecdotal evidence from industry participants^[23] as to the definition and power of Facebook, mere assertions from regulators and academics without evidence in support is unlikely to be sufficient to meet the ACCC’s burden of proof.^[24]

3 Digital Advertising Display Services

The ACCC found that Facebook (via both the Facebook and Instagram platforms):

has substantial market power in the supply of display advertising in Australia, with a market share of around 51 per cent... [with no other firm having] a market share of more than 5 per cent.^[25]

The market for display advertising services was considered distinct from online search advertising, eliminating Google as a potential competitor from the market definition.

4 The Social Media and Digital Advertising Services Markets are Interrelated

News media organisations supply news content to audiences in a two-sided market that sells those audiences’ attention to advertisers. There is a market for consumption of the news content, which is often free, and a market for the sale of advertising, which is typically priced at a level to cover the cost of supplying the platform to audiences (whether it is physical newspapers, a website or television or radio broadcast facilities). The two markets are interrelated: the quality of the audience (be it size, demographic or engagement), the higher the price that can be charged to the advertisers. ^[26] This interrelationship results in “network effects” or externalities whereby the size of one side of the market increases the value to either other participants in that market (direct network effects), or to customers in the interrelated market (indirect network effects).^[27]

Social media platforms like Facebook and Instagram are similar to news media organisations in that they are also two-sided,^[28] connecting users who utilise the platform at a zero price with advertisers who pay the platform to have the users see their content. Unlike news media organisations however, the social media platform does not produce significant quantities of content itself, rather the users themselves create and share the content. News media organisations are like many businesses, for the purposes of this market, users, creating content that is shared on Facebook platforms and also advertisers, paying Facebook to promote their brands in targeted advertising that may include content.

Digital platforms present a unique challenge in the assessment of market definition. The OECD has noted that:

Identifying the relevant markets inside the Big Data ecosystem can be a particularly daunting task, as a result of the many different players involved that may take multiple roles, as well as the complex relations that link them.^[29]

The OECD summarised the economic consensus on multi-sided attention matching platforms such as Facebook as follows:

as long as consumers, advertisers, content providers and any other agents involved do not directly transact with each other, law enforcers should define a different market for each side of the platform (see Filistrucchi et al, 2014 and Wright, 2004). The underlying reasoning for identifying multiple markets is that products may be perceived with different degrees of substitutability in the different sides of the platform – for instant, social networks and search engines may be regarded as substitutes for advertisers, but not for consumers.^[30]

Thus, while Facebook operates in “display advertising” and “social media” markets and those markets are interrelated, the ACCC has followed the economic consensus in separately identifying two sides of the market, rather than either conflating the two or considering only the one side.

While it would be open to Facebook to argue that only the user market (“social media services”) is germane to the section 46 inquiry at hand, section 46(1)(b) and (c) mean that a finding of “substantial market power” in the other market (“display advertising services”) would satisfy the market power criterion of section 46.

5 Product Dimension of the Market

Having identified the three possible markets in which Facebook might be argued by the ACCC to have market power, the next step is to determine the dimensions of the market to a sufficient extent that an assessment of market power can be made. The relevant dimensions are function, product and geography,^[31] although the functional and geographical dimensions are less relevant to this analysis due to the nature of the online platforms. The ACCC’s analysis does not appear to be a comprehensive examination of the product dimension of the markets. It was criticised for failing to adequately perform this step by Facebook and Google and the Global Antitrust Institute,^[32] although the Digital Platforms Inquiry was not an investigation of either of Facebook or Google that such a determination would have been necessary.^[33]

If a news media referral services market can be said to exist, the question would be, what products are close substitutes for Facebook’s services to news outlets? The ACCC, albeit not in the context of market definition, identified four substitutable channels for news referrals: search engines, social media, news aggregators and direct access.^[34] After considering the mitigating effect of these alternative channels, the ACCC found that news referral services supplied by Google Search and Facebook are “must have” products for news media businesses, due largely to the percentage of referrals driven by those platforms to news media organisations’ own platforms.^[35] The ACCC considered the impact on referrals if Google and/or Facebook ceased to provide these services to news media organisations, finding that the news media organisations would suffer substantial losses.^[36] It is therefore arguable that the product dimension of the market would include at least the search engine, social media and news aggregator channels as substitutes.

If that is so, Google would be a competitor to Facebook in news media referral services and given its size and dominance, presumably would operate as a competitive constraint on Facebook’s ability to “give less and charge more” ^[37] in that market; a factor not considered by the ACCC in its assessment.

In the digital display advertising market, the ACCC indicated that offline advertising, search advertising and classified advertising are not close substitutes for display advertising.^[38] In this market, it was found that the news media organisations’ own websites and platforms competed with Facebook.^[39] Display advertising on social media was said by the ACCC to be differentiated from other advertising because of its unique characteristics of allowing users to interact with the advertisers on the social media platform.^[40] Yet the question that must be asked is whether there would be strong substitution between these and other forms of advertising in the long run, given sufficient price incentive.^[41] As a price-based market, the Small But Non Transitory Increase in Price test (SSNIP Test)^[42] could be apposite^[43]: this would involve the presentation of econometric evidence. Anecdotally, on the demand side, there is a reasonable argument that, given sufficient price incentive, these other forms of advertising could be strongly substitutable even in the short to medium term: display advertising on Facebook does not require lock-in contracts that would prevent advertisers from shifting to other forms of online display advertising or indeed offline advertising. On the supply side, in the long run and noting a potential increase in price, we may see news media outlets themselves set up targeted display advertising or Google or another social media competitor enter the field of rivalry. It follows that Facebook’s objection that the display advertising market may have been defined to narrowly to exclude other forms of advertising may, if supported by econometric evidence, be successful.

Finally, the “social media market” in which Facebook competes for users to deliver to its advertisers, as a non-price market, a SSNIP Test to determine the extent of the product dimension of the market would be inappropriate since no prices are charged.^[44] In such markets, a “small but significant non-transitory decline in quality” (SSNDQ Test) is frequently used,^[45] whereby an assessment is made as to whether a significant decline in quality would result in users switching to another platform or form of attention. While Facebook and Instagram are currently very popular in Australia, platforms exist to which users could switch if its quality declined, such as Twitter, Tiktok, Youtube, Snapchat and LinkedIn, and where, on the supply side, technological adjustments could be made to entice such switching. Also on the supply side, relatively low startup costs mean that a new entrant could be drawn into the field of rivalry without great expense or investment. Facebook made this argument in its submissions, pointing to the transition from Friendster to Myspace before Facebook became the most popular social media network.^[46] Again, it is very possible that the product dimension of the market has been understated by the ACCC.

B Market Power

Once a market has been determined, the court would consider evidence as to the power of Facebook in that market. Typically, analysis of market power commences with an assessment of the firm’s market share, although this is not determinative.^[47] The OECD notes the difficulty in using market share as a measure of market power in multi-sided markets, in particular that it may not be possible to compute value shares on both sides if one side does not pay for using the platform, and that customers on one side of the market can be competitors on the other side.^[48] The ACCC notes in its guidelines that it does not impose a threshold for market share,^[49] however it certainly appeared to place weight on its assessment of market share in the Digital Platforms Report, as noted above with respect to Facebook’s share of the digital advertising market. However, this assessment was heavily qualified and most importantly, did not include Google as a potential competitor. If a court ultimately finds Google and Facebook to be competitors in digital display advertising against the ACCC’s assessment, this could impact the assessment of market power.

Interestingly, in relation to the news referral market, the ACCC found that Google and Facebook supply news referral services which it considers “probably constitute a market” in which Google and Facebook are each likely to have substantial market power.^[50] The ACCC’s reasoning centres around the determination that Google and Facebook are “unavoidable trading partners” for media businesses.^[51] Yet no detailed analysis of the indicia of market power discussed below is provided by the ACCC in support of this “probable conclusion”, nor is there any consideration of the impact of competitive constraints by each company on the other in that market.

In its guidelines, the ACCC notes that more than one corporation can have a substantial degree of market power, and points to the focus of the courts on the existence of effective competitive constraints, including the existence of competitors that may constrain the firm’s ability to act to increase prices or reduce output below the competitive level, the extent to which the products are characterised by product differentiation, the extent of vertical integration as well as contractual arrangements.^[52]

Importantly, the ACCC accepts that dynamic competition imposes some degree of competitive constraint on Facebook, but that this effect is tempered by barriers to entry and expansion and its advantages of scope and its acquisition strategy.^[53] Evans however argues that dynamic competition in multi-sided markets also give rise to positive feedback effects from the resultant innovation that is required to maintain user interest.^[54] Facebook^[55] and Google^[56] also emphasise the effect of dynamic competition on their businesses.

Critical to this analysis is the market definition, and in that regard the ACCC has unsurprisingly chosen markets that do not include Google (on the advertising side), or Google owned platforms like Youtube^[57] (on the social media side). If the ACCC is found to be wrong in this regard, it will have to convince the court that Google and its platforms do not impose a competitive constraint on Facebook in the display advertising and social media markets.

In relation to barriers to entry and expansion, the ACCC notes the existence of network effects, branding, and economies of scale and scope as being significant barriers to entry, and these aspects should be taken into account.^[58] However, the OECD recommends the starting point should be a solid understanding of the nature of competition in the market under consideration, and then to proceed with an analytical framework that takes account of the multi-sided nature of the market.^[59] In particular, while it is true and widely accepted in the literature that network effects can create barriers to entry and expansion, this should be against the context of the absence of the more traditionally understood structural barriers to entry; little is needed to launch a new platform other than some web developers, trade marks, data storage and domain names. How these factors are weighed against each other will be a matter for the courts, though it is clear that the ACCC and other competition regulators are of the view that the network effects will outweigh these countervailing factors.

Facebook has relied on the fact that it was able to enter the social media market and displace MySpace and Friendster to argue that barriers to entry are not high in the social media market. Against that argument is the theory of tipping, whereby network externality feedback loops result in the dominance of one platform over others as a multi-sided market matures.^[60] The US Antitrust Subcommittee viewed evidence from Facebook that it believed it had tipped the social networking market as compelling evidence of market power in that market.^[61]

C Conclusion on Market Power

The above analysis has raised more questions than it has answered. A myriad of potential markets could be pleaded by the ACCC and their extent will determine the question of whether Facebook has substantial market power. Critical to that question is whether Google operates as a competitive constraint on the advertising side of the market, and whether its social media-like platforms like Youtube operate as a constraint on the social media/networking side of the market; the ACCC in its analysis excluded them from the market definition. The economics of multi-sided markets is relatively new and still developing, with many scholars presuming the market power of platforms like Facebook and Google and appearing to use network effects as a justification for this presumption. Whether this analysis will withstand the high standard of proof were the ACCC to commence proceedings against Facebook is a live question.

Notwithstanding these uncertainties, for the purposes of the remainder of this essay, a finding of substantial market power in each of the three markets articulated by the ACCC will be assumed. Given the attitudes of regulators worldwide including the ACCC and the US Subcommittee on Antitrust as well as the body of scholarly work on platform market power, this is not an unrealistic assumption.

II THE NEW SECTION 46

The misuse of market power prohibition found in section 46 of the CCA has undergone dramatic reshaping over the years since the concept was introduced at the turn of the twentieth century, with five successive statutes and nine amendments since.^[62] Prior to the most recent amendment in 2017, section 46 prohibited a corporations with substantial market power from taking advantage of that power for three specified purposes. As the cases were heard under this provision, it became clear that the requirement of “taking advantage” of market power had become an impediment to successful prosecutions.^[63] The Harper Review noted these challenges and recommended removing this wording as well as the prescribed purposes and introducing an “effects” test to better focus section 46 on the aim of the legislation on protection of consumers (via the promotion of competition in markets) rather than individual competitors.^[64]

The new section 46 has yet to be tested before the courts, but it is clear that some conduct that would previously have been prohibited (for example, conduct aimed at harming competitors) will no longer contravene if the conduct does not have the purpose, effect or likely effect of substantially lessening competition (SLC), narrowing the impact of section 46. The scope of section 46 is also narrowed by the requirement that the conduct be in a market in which the corporation or a related body corporate trades rather than any market. In other ways, the operation of section 46 has been broadened: first, beyond the three previously proscribed categories to any conduct that would meet the SLC test; and secondly, the extension beyond an anti-competitive purpose to anti-competitive likely effects, picking up unintentional contraventions previously untouched by the legislation.

This change of focus in section 46 arguably has significant impact on the type of unilateral conduct that is the subject of this essay. Refusals to deal typically harm individual competitors, but the impact on competition may in many cases not be substantial.

The introduction of the SLC test also brings section 46 back to market definition. The fraught process of market definition highlighted in particular in relation to digital platforms in the previous section becomes relevant again at this stage and the extent of the impacted market will again be critical to the outcome of the section 46 case.

A Purpose Test

It is clear from its text that a primary purpose of the Threat was to apply pressure to the ACCC and the Commonwealth Government, and perhaps implicitly those media organisations lobbying for it, not to proceed with legislating the draft mandatory bargaining code. No doubt, Facebook would argue that this purpose was merely to prevent legislation from being enacted that would harm its interests. Such a purpose does not appear consistent with a subjective intention to SLC.

However section 4F permits a consideration of any purpose of particular conduct, provided that purpose is substantial. Much will turn on internal evidence from Facebook surrounding the publication of the Threat.

The purpose of a corporation may be inferred from its conduct and the relevant circumstances,^[65] and it may be inferred that the Threat and the Conduct was designed to apply commercial pressure to those media companies to cease their efforts in seeking a mandatory bargaining code. The fact that the mandatory bargaining code has been recommended to address an imbalance of market power between news media and digital platforms would be relevant, though perhaps would have been more influential prior to the 2017 amendment when harm to competitors and deterring or preventing a person from engaging in competitive conduct was prohibited without reference to the SLC test.

Facebook’s stated intention to launch Facebook News, an aggregator product, in Australia could also be a relevant factor,^[66] if for instance it can be established that the true purpose of the Threat or the Conduct was to transition news media from its free platform to a new business model. Such a purpose would need to result in SLC in a market in which Facebook operates to meet the threshold in section 46. If the purpose of the Facebook News product was to redirect user attention away from news referral and towards its own platform where advertisements could be targeted to them, a SLC argument in the news referral market (if it exists) is arguable.

However if the court agrees with the arguments of Facebook and Google that there is no market for news referral services, then the larger markets of social media and display advertising involved could potentially mean that the substantiality threshold will not be met. Under the new section 46, the prohibited purpose is not just to lessen competition, but to substantially lessen competition, and as the market becomes broader the requisite impact of the conduct must be greater. In the social media market, the news media are user-customers (albeit important user-customers^[67]) and their content competes with the content of billions of other users for attention. In the display advertising market, as noted above, the news media’s online platforms are considered by the ACCC to be competitors of Facebook. By removing the news media’s content from its platforms, Facebook could be seen to be harming the news media’s interests, but this is no longer the test under section 46. A SLC is required, and again, given the news media’s limited importance in the display advertising market, Facebook could argue that the impact on the display advertising market of removing traffic to the news media platforms would not be substantial.

B Effects Test

The effects test in s46 means even if the purpose element is not made out, the ACCC may still succeed under s46 if the conduct would have the effect or likely effect of SLC. To assess the effect or likely effect of the Threat or the Conduct, it is necessary to perform a counterfactual analysis by comparing the likely state of competition with the Threat/Conduct, and the state of competition without the Threat/Conduct.^[68]

Much will turn on the relevant markets as properly defined, however if there is a “news referral services” market, and Facebook or a related body corporate is found to have substantial market power in any market, it is difficult to escape the conclusion that removing all news content from Facebook’s platform would have the effect of lessening competition; “news referral services” would no longer be offered by Facebook and only by its competitors.

However Facebook may argue that if the news referral services market includes Google, the news outlets themselves, and other social media platforms, the impact of Facebook refusing to deal with the news media will not result in SLC. Indeed, Facebook may argue that in the longer run, low barriers to entry could mean that the news media outlets could set up their own platform (such as a news aggregator platform for instance) to generate their own referrals. Indeed, in its submissions to the Digital Platforms Inquiry, Facebook noted a “natural experiment” that occurred when a Facebook outage resulted in publishers receiving fewer referrals. Facebook argued that this outage resulted in a 22% increase in direct traffic through the publishers’ own apps, thereby mitigating the impact on the publishers of the absence of referrals from Facebook.^[69] Such an argument would have been more compelling under the previous section 46 when the focus was on harm to competitors; the fact that competitors receive increased referrals would tend to suggest that concentration would increase following the departure of Facebook from the news referrals market and therefore that SLC would be likely to occur.

If however, as per Facebook’s primary argument, there is no “news referral market” that is distinct from the broader markets identified by the ACCC, as discussed in the previous section, the substantiality threshold becomes more difficult. In relation to social media services, the impact of a refusal to permit news media users to publish, and other users to share, news content becomes difficult in the context of a multi-sided platform like Facebook. It would tend to mean a reduction in quality of the Facebook service, a point obliquely noted by the ACCC in its response to the Threat, with Rod Sims saying “39% of Australians use Facebook for general news”,^[70] pushing those users onto competing social media platforms such as Twitter and Youtube, news aggregator platforms and perhaps onto the news media’s owned digital platforms. Potentially, rather than lessening competition in the social media market, Facebook could argue that in strengthening its competitors, the Conduct would improve competition by creating more viable competitors in that broader market.

C Analysis

What this analysis also reveals is perhaps a concerning corollary of the new legislation: whereas refusals to deal were historically caught by the former section 46(1)(c) of the CCA, now the requirement that they be subject to the SLC Test requires the ACCC to meet several onerous hurdles: the conduct must lessen competition in a particular market and do so substantially. While Duke has suggested that section 46’s application to refusals to deal has been extended by the 2017 amendments (by removing the “take advantage” requirement),^[71] its application to this category of unilateral conduct may well also be limited by the requirement that the refusal to deal have the effect or likely effect of SLC.

This corollary is consistent with the intent of the 2017 amendments; as the Harper review stated:

Ordinarily, competition law is not concerned with harm to individual competitors. Indeed, harm to competitors is an expected outcome of vigorous competition. Competition law is concerned with harm to competition itself — that is, the competitive process.^[72]

However, in refusal to deal cases the effect on the market can be negligible. For instance, in ACCC v Ticketek, in which Ticketek was required to pay a $2.5 million penalty, Bennett J noted that there was no ongoing effect on the market from the conduct.^[73] The only relevant question under the previous section 46 was the purpose of Ticketek (which was agreed between the parties) in seeking to deter or prevent its competitor Lasttix from supplying discount tickets in competition. However, given the actual effects on the market, such a purpose would not necessarily constitute a purpose of SLC under the new section 46.

It is apparent from the analysis above that for the ACCC to establish SLC in refusal to deal cases in all but a monopolistic or “potential” market where competitors have been denied entry will be difficult, particularly given its track record on defining an appropriate market for the purposes of the analysis.

Further, as the analysis in this essay has demonstrated, multi-sided platforms introduce problems in analysing the concepts introduced in the new section 46. First, defining the market, challenging for the ACCC even in one-sided markets, becomes even more complex and the economic literature is still developing. Secondly, the use of SLC in the purpose element does not obviously point to or prevent unilateral exclusionary practice such as refusals to deal. Thirdly, depending on the size of the market chosen, establishing substantiality in purpose or effect creates a further challenge.

III CONCLUSION

This essay has used the Threat as a lens to highlight both aspects of the new section 46 that will impose additional evidentiary burdens on the ACCC as well as the difficulties faced by authorities in examining the conduct of multi-sided digital platforms like Facebook and Google.

Whether the Threat or the Conduct will contravene section 46 is driven by the appropriate market definition. Based on the ACCC’s analysis in the Digital Platforms Report and Facebook’s and Google’s responses, the three markets defined by ACCC would be hotly contested in any litigation relating to the Threat or the Conduct. A key consideration will be whether Google acts as a competitive constraint in any of the identified markets; the ACCC did not consider that it did in the broader social media or digital display advertising markets. Economic thinking on digital platform market power is still evolving and while many economists consider that the digital platforms have market power because of network effects, it is not clear that dynamic competition and low start up costs have been adequately factored in to those assessments as would a court.

Having said that, if the ACCC’s analysis in the Digital Platforms Report were accepted by the court, then it is very possible that the Conduct and the Threat could constitute a contravention of section 46. However, as discussed in the last part of this paper, the recent amendments to section 46 arguably make the task of the ACCC in investigating and prosecuting such a contravention more difficult than would have been the case under the previous drafting of section 46.

IV BIBLIOGRAPHY

A Articles and Books

Alleman, James , Edmond Baranes and Paul N. Rappoport, 'Multisided Markets and Platform Dominance' in James Alleman, Edmond Baranes and Paul N. Rappoport (eds), Applied Economics in the Digital Era: Essays in Honour of Garry Madden (2020) 303

Bamberger, Kenneth A. and Orly Lobel, 'Platform Market Power' (2017) 32 BERKELEY TECH. L.J. 1051

Duke, Arlen, Corones' Competition Law in Australia (Thompson Reuters, 2019)

Evans, David S., 'Multisided Platforms, Dynamic Competition, and the Assessment of Market Power for Internet-Based Firms' (University of Chicago Coase-Sandor Institute for Law & Economics Research Paper No 753, 10 March 2016)

Filistrucchi, Lapo et al, 'Market Definition in Two-Sided Markets: Theory and Practice' (2014) 10 Journal of Competition Law and Economics 293

Hanley, Daniel, 'A Topology of Multisided Digital Platforms' (2020) 19(2) Conneticut Public Interest Law Journal 272

Kemp, Katherine, 'Misuse of market power in australia and abuse of dominance in canada: Two legislated effects tests for unilateral conduct' [2018] UNSW Law Research Series 85

Nadler, Jerrold and David N. Cicilline, Antitrust Subcomittee of the US House Committee on the Judiciary, Investigation of Competition in Digital Markets, 6 October 2020)

Sanchez- Cartas, Juan Manuel and Gonzalo Leon, Multisided Platforms and Markets: A Literature Review (2019)

B Cases

ACCC v Flight Centre Travel Group [2016] HCA 49

ACCC v Ticketek [2011] FCA 1489

Air New Zealand v ACCC [2017] HCA 21

Australian Gas Light Co v ACCC (No 3) (2003) [2003] FCA 1525

Boral Besser Masonry v ACCC [2003] HCA 5

NT Power Generation Pty Ltd v Power and Water Authority [2004] HCA 48

Outboard Marine Australia v Hecar Investments [1982] FCA 265; (1982) 66 FLR 120

Queensland Wire Industries Pty Ltd v Broken Hill Proprietary Co Ltd (1989) 67 CLR 177

Re Queensland Co-operative Milling Association Ltd (1976) 25 FLR 169

Rural Press & Ors v ACCC [2003] HCA 75

Stirling Harbour Services v Bunbury Port Authority [2000] FCA 38

C Legislative Instruments

Competition and Consumer Act 2010 (Cth)

D Reports

ACCC, Digital Platforms Inquiry Final Report, (June 2019)

Australian Competition and Consumer Commission, Guidelines on misuse of market power, 2018)

Harper, Ian et al, Competition Law Review, (2015)

OECD, 'Big Data: Bringing Competition Policy To The Digital Era: Background Note By The Secretariat' No DAF/COMP(2016)14,

OECD, 'Measuring market power in multi-sided markets' No DAF/COMP/WD(2017)35/FINAL,

Yun, John M. et al, Comment of the Global Antitrust Institute, George Mason University School of Law, on the Australian Competition & Consumer Commission's Digital Platforms Inquiry, Preliminary Report, 22 January 2019)

E Other

Easton, Will, 'An Update About Changes to Facebook’s Services in Australia', (Blog Post, 31 August 2020) <https://about.fb.com/news/2020/08/changes-to-facebooks-services-in-australia/>

Facebook Australia Pty Ltd, Response to the ACCC’s Preliminary Report by Facebook Australia Pty Limited, (3 March 2019)

Google Australia Pty Ltd, 'SUBMISSION IN RESPONSE TO THE ACCC’S PRELIMINARY REPORT', 18 February 2019)

Sims, Rod, 'Statement on Facebook' (1 September 2020) <https://www.accc.gov.au/media-release/statement-on-facebook>

---

^[1] Will Easton, 'An Update About Changes to Facebook’s Services in Australia', (Blog Post, 31 August 2020) <https://about.fb.com/news/2020/08/changes-to-facebooks-services-in-australia/>.

^[2] Competition and Consumer Act 2010 (Cth) ('CCA').

^[3] ACCC, Digital Platforms Inquiry Final Report, (June 2019) 9 ('Digital Platforms Report').Although the ACCC, in the face of criticism from Facebook and Google of its suggestion that there was a market for news referral services, indicated that “it was not necessary for the purposes of this Report to reach a firm conclusion” on the question, it nevertheless said that “there is probably a market for news referral services”: at 103.

^[4] Facebook Australia Pty Ltd, Response to the ACCC’s Preliminary Report by Facebook Australia Pty Limited, (3 March 2019) 36.

^[5] Ibid.

^[6] Queensland Wire Industries Pty Ltd v Broken Hill Proprietary Co Ltd (1989) 67 CLR 177 ('Queensland Wire'). Per Mason CJ and Wilson J, 187-188.

^[7] Air New Zealand v ACCC [2017] HCA 21, [57]-[59] ('Air New Zealand v ACCC'). Citing Queensland Wire (n 6) per Deane J, 195.

^[8] Queensland Wire (n 6) per Mason and Wilson JJ, 187.

^[9] Rural Press & Ors v ACCC [2003] HCA 75 ('Rural Press').

^[10] Digital Platforms Report (n 3) 27.

^[11] ACCC v Flight Centre Travel Group [2016] HCA 49, [70]-[71] ('ACCC v Flight Centre Travel Group').

^[12] Arlen Duke, Corones' Competition Law in Australia (Thompson Reuters, 2019) 65.

^[13] ACCC v Flight Centre Travel Group (n 11).

^[14] Although it should be noted that the Bundeskartellampt considers Facebook to be operating in an even narrower market than that found by the ACCC, with LinkedIn and Twitter considered to be in different markets to Facebook: Digital Platforms Report (n 3) 78.

^[15] Facebook Australia Pty Ltd (n 4) 36.

^[16] Re Queensland Co-operative Milling Association Ltd (1976) 25 FLR 169, 190 ('Re QCMA').

^[17] Queensland Wire (n 6); NT Power Generation Pty Ltd v Power and Water Authority [2004] HCA 48, [109] ('NT Power').

^[18] See e.g.Boral Besser Masonry v ACCC [2003] HCA 5 ('Boral Besser Masonry v ACCC'). McHugh J said that “the views and practices of those within the industry are often the most instructive on the question of achieving a realistic definition of the market...those they perceive to be their competitors and those whose conduct they seek to counter is always relevant to the question of market definition”

^[19] Digital Platforms Report (n 3) 77.

^[20] Ibid.

^[21] See e.g. Daniel Hanley, 'A Topology of Multisided Digital Platforms' (2020) 19(2) Conneticut Public Interest Law Journal 272; Kenneth A. Bamberger and Orly Lobel, 'Platform Market Power' (2017) 32 BERKELEY TECH. L.J. 1051.

^[22] Juan Manuel Sanchez- Cartas and Gonzalo Leon, Multisided Platforms and Markets: A Literature Review (2019) 15.

^[23] Boral Besser Masonry v ACCC (n 18), [257] per McHugh J.

^[24] The burden of proof in a section 46 case would likely be higher than the civil standard due to seriousness of the issues and the size of the penalties under the Briginshaw principle: Briginshaw v Briginshaw [1938] HCA 34; (1938) 60 CLR 336.

^[25] Digital Platforms Report (n 3) 97.

^[26] David S. Evans, 'Multisided Platforms, Dynamic Competition, and the Assessment of Market Power for Internet-Based Firms' (University of Chicago Coase-Sandor Institute for Law & Economics Research Paper No 753, 10 March 2016) 8.

^[27] Lapo Filistrucchi et al, 'Market Definition in Two-Sided Markets: Theory and Practice' (2014) 10 Journal of Competition Law and Economics 293, 296.

^[28] Indeed, most including Facebook are in fact multi-sided, though the other markets are not discussed here as they are not germane to the section 46 analysis.

^[29] OECD, 'Big Data: Bringing Competition Policy To The Digital Era: Background Note By The Secretariat' No DAF/COMP(2016)14, 15.

^[30] Ibid 16. In fact, Filistrucchi et al state “In two sided markets, competition authorities should take into account both sides of the market when defining the relevant market, even in the case of two-sided non-transaction markets, when they define two interrelated markets”: Filistrucchi et al (n 27) 322.

^[31] Australian Gas Light Co v ACCC (No 3) (2003) [2003] FCA 1525, [378] per French J ('AGL v ACCC').

^[32] John M. Yun et al, Comment of the Global Antitrust Institute, George Mason University School of Law, on the Australian Competition & Consumer Commission's Digital Platforms Inquiry, Preliminary Report, 22 January 2019).

^[33] The terms of reference for the Digital Platforms Inquiry Report specified the following matters to be taken into consideration: “the extent to which platform service providers are exercising market power in commercial dealings with the creators of journalistic content and advertisers”, “the impact of platform service providers on media and advertising markets” and “the impact of information asymmetry between platform service providers, advertisers and consumers and the effect on competition in media and advertising markets.”, but expressly stated “this is not to be an inquiry into supply by any particular persons”: Digital Platforms Report (n 3) 537.

^[34] Ibid 102. Although it is unclear to this writer how “direct access” can be a referral service.

^[35] Ibid 102-103.

^[36] Ibid.

^[37] Re QCMA (n 16) at ALR 517.

^[38] Digital Platforms Report (n 3) 92.

^[39] Ibid 89.

^[40] Ibid 98.

^[41] Re QCMA (n 16) 190.

^[42] The SSNIP test is an iterative assessment of the impact in a hypothetical monopoly market of an increase in price. Substitute products are included in the market definition if the increase in price of 5-10% above the competitive level can be sustained for a period of 12 months: ACCC Merger Guidelines.

^[43] James Alleman, Edmond Baranes and Paul N. Rappoport, 'Multisided Markets and Platform Dominance' in James Alleman, Edmond Baranes and Paul N. Rappoport (eds), Applied Economics in the Digital Era: Essays in Honour of Garry Madden (2020) 303, 310.

^[44] Ibid.

^[45] Evans (n 26), 26, noting that this approach was used informally in Qihoo 360 v Tencent by the Chinese Supreme People’s Court.

^[46] Facebook Australia Pty Ltd (n 4) 34.

^[47] Queensland Wire (n 6) per Mason CJ and Wilson J.

^[48] OECD, 'Measuring market power in multi-sided markets' No DAF/COMP/WD(2017)35/FINAL.

^[49] Australian Competition and Consumer Commission, Guidelines on misuse of market power, 2018) 6.

^[50] Digital Platforms Report (n 3) 58.

^[51] Ibid 100.

^[52] Australian Competition and Consumer Commission (n 49).

^[53] Digital Platforms Report (n 3) 58.

^[54] Evans (n 26) 1.

^[55] Facebook Australia Pty Ltd (n 4) 36.

^[56] Google Australia Pty Ltd, 'SUBMISSION IN RESPONSE TO THE ACCC’S PRELIMINARY REPORT', 18 February 2019) 30.

^[57] Google + was considered and then disregarded as it was closed down as a platform in April 2019: Digital Platforms Report (n 3) 78. Note also that Youtube was considered not to be competitive with Facebook based inter alia on Facebook’s own internal documents in the US Antitrust Subcommittee report on competition in digital markets and the UK Competition and Markets Authority: Jerrold Nadler and David N. Cicilline, Antitrust Subcomittee of the US House Committee on the Judiciary, Investigation of Competition in Digital Markets, 6 October 2020) 142 at footnote 800.

^[58] OECD, 'Measuring market power in multi-sided markets' (n 48) 5.

^[59] Ibid.

^[60] Hanley (n 21) 289.

^[61] Nadler and Cicilline (n 57) 138-139, 140.

^[62] Katherine Kemp, 'Misuse of market power in australia and abuse of dominance in canada: Two legislated effects tests for unilateral conduct' [2018] UNSW Law Research Series 85, 66.

^[63] See e.g. Kirby J’s famously trenchant dissent in Rural Press (n 9), where his Honour said “This is the third recent decision of this Court (Melway and Boral Besser Masonry Ltd v Australian Competition and Consumer Commission being the other two) in which a majority has adopted an unduly narrow view of s 46 of the Act. ...Section 46 might just as well not have been enacted for cases like these where its operation is sorely needed to achieve the purposes of the Act...Once again, I dissent” [footnotes omitted].

^[64] Ian Harper et al, Competition Law Review, (2015) 348.

^[65] Queensland Wire (n 6).

^[66] Easton (n 1).

^[67] The Daily Mail and the ABC each have 4 million Facebook followers, Buzzfeed Australia has 3 million, the Sydney Morning Herald, the Daily Telegraph and news.com.au have over 1 million; yet no news media feature in the top 10 Facebook pages in Australia; with AC/DC and fitness influencer Kayla Itsines each having 28 million users. Source: www.socialbakers.com

^[68] Outboard Marine Australia v Hecar Investments [1982] FCA 265; (1982) 66 FLR 120 ('Outboard Marine Australia v Hecar Investments'); Stirling Harbour Services v Bunbury Port Authority [2000] FCA 38 ('Stirling Harbour Services v Bunbury Port Authority').

^[69] Facebook Australia Pty Ltd (n 4) 39.

^[70] Rod Sims, 'Statement on Facebook' (1 September 2020) <https://www.accc.gov.au/media-release/statement-on-facebook>.

^[71] Duke (n 12) 568. Ironically, the much lamented requirement of “taking advantage” would have been unlikely to present an issue to the ACCC in this case; it is hard to conceive of a legitimate business rationale for making the Threat and the Threat and the Conduct could not be engaged in without substantial market power for fear of losing users and advertisers to a competitor.

^[72] Harper et al (n 64) 339.

^[73] ACCC v Ticketek [2011] FCA 1489, [54]-[55] ('ACCC v Ticketek').