astly, digital markets offer significant opportunities for large and small companies to benefit from

integrating businesses which allows these companies the access to new markets and distribution
channels, without funding the full costs of a platform business up front.

Given features such as network effects, economies of scale and opportunities of data sharing, mining
and analysis for scaling services, many platform operators strategize to increase their size and user
base by mergers and acquisitions (M&As). Google, Facebook and Amazon have acquired various
younger companies in recent years. For instance, Facebook acquired Instagram in 2012, when the
latter was only 2 years old providing its users a free mobile photo application, while Facebook was a
digital social networking platform with photo sharing as one of its features. Google acquired Flutter,
an early stage start-up for gesture recognition for advancing its artificial intelligence cluster.

For platform operators, such M&As increase revenues, both through direct payments as well as
through advertising revenues, and also give them access to large volumes of customer data to
develop new products and services, thereby increasing their user base further.71 For consumers, the
synergies resulting from such transactions lead to improvement in the product quality, increased
choice and convenience. For instance, after its acquisition by Facebook in 2012, Instagram
introduced fully-fledged social network functionalities, such as direct messaging, photo tagging, and
allowed advertisers to place their advertisements on the platform.

However, it is also possible that the transacting parties use these very features of the digital
markets, especially the access to free consumer data, design and algorithms, to their advantage,
which in turn might impact competition and / or consumers adversely. The providers of this free
data, namely the platform users are disaggregated and have limited collective bargaining power
against platform owners, who can use the freely acquired data for targeted advertising to improve
and eventually monetize their services. For consumers however, this might lead to a compromise of
privacy by making available their personal data to the merged parties for free.

Furthermore, the combination of datasets (user information) of the merging parties could also
increase the merged entity’s market power in the use of the relevant data to raise barriers to entry
for competitors, as the merged entity would now have access to richer information of their users. By
reducing potential competition, the merged entity might also have lower incentive to innovate going
forward.

Hence, while there are efficiencies from M&As in digital markets, expectations or concerns about
adverse effects on competition and consumer welfare resulting from these transactions, are
legitimate. Key consideration in such transactions is to determine how, as providers of free data to
digital platform owners, consumers can be compensated and hence given greater bargaining power
against the latter.

In the next section, we discuss these concerns in the context of challenges that the nature of the
market imposes on competition assessment in the sector, using some of the recent transactions to
substantiate how advanced jurisdictions around the world have addressed these challenges. Our
review suggests that although the advanced jurisdictions are using a flexible approach to address
some of these concerns, no clear solution has emerged.

The competition assessment of a merger or an acquisition, which crosses the thresholds mandating
antitrust review, comprises three basic steps (a) defining the relevant product and geographic
markets within which competition impact is to be assessed; (b) assessing market power in the
defined relevant market; and (c) analyzing the effect of the transaction on competition in the
relevant market. In the case of mergers in the digital market however, the complex characteristics of
the digital market discussed earlier, and the potential positive and negative impact of mergers in this
space, make some of the tried and tested antitrust theories and procedures unsuitable for
competition assessment.

The dilemma of competition authorities in assessing such mergers is therefore multi-fold. One, how
do they anticipate the synergistic effect of such mergers, which may seem inconsequential and
might even fly under their radar because it is common for digital platform operators to provide
services for “free”. (Even though the platform owners get access to free consumer data in return,
which allows them to gain scale before being able to monetize their services.)

Second, how to define the relevant markets and assess the impact on consumers and competition
because of the dynamic nature of the sector. Third, what corrective action to take without being
able to predict the effect such actions could have on dampening innovation. Fourth, how to deal
with the problem of consumers not being compensated for data that is eventually monetized by
these companies.

Identifying digital mergers requiring antitrust scrutiny:

Typically, revenue or turnover thresholds are used to assess whether a merger is notifiable.
However, with the advent of digitization, the adequacy of turnover-based thresholds to identify
certain mergers is becoming questionable as often digital platforms offer services for “free” to begin
with and later start monetizing their products or services. This impacts the revenue streams of digital
companies and thus allows some of the high valued digital mergers to escape competition
assessment. For instance, the turnover of WhatsApp when Facebook acquired it in 2013, was lower
than the threshold requiring competition assessment in the European Union (EU). This meant that
based on turnover thresholds, the transaction did not need a review in the EU. The European
Commission (EC) however, opined that the transaction amounted to concentration and was capable
for review under the competition laws of three of its member states, thereby mandating a detailed
competition assessment of the transaction.

More recently, when Apple acquired Shazam, a developer and distributor of music recognition
mobile applications for smartphones, tablets and personal computers (PCs), in 2018, the parties did
not satisfy the EC’s turnover thresholds. However, the EC assessed the merger for potential
anticompetitive conduct based on referrals from member states. These transactions suggest that the
EC is adopting a flexible approach in bypassing the traditional thresholds for identifying digital
mergers which might require further scrutiny from an antitrust standpoint. The issue with this
approach, however, is that it is discretionary rather than rule based. In other words, this approach
gives the regulator, the power to investigate any merger which in its opinion warrants investigation.
Secondly, it raises the issue of false positives, as the regulator might find market concentration
where none exists in the future. This can in turn have a negative impact on innovation.

Defining the relevant product and geographic markets

From an economic perspective, the definition of relevant markets is largely driven by the principle of
demand and supply substitutability.74 Demand (supply) substitutability implies that consumers
(producers) can and do switch to another product, or purchase (sell) the product or service from
another geographic area when the price of a product increases in one market.

Most of the tools used to define the relevant market are based on this principle and consider price
as the main reason consumers or producers switch between products or geographies. The
commonly used “Small but Significant and Non-Transitory Increase in Price (SSNIP)” test for instance,
focuses on defining the relevant market on the smallest set of products/services, such that a
hypothetical monopolist would not find it profitable to increase prices by 5 to 10 percent as the
consumers can substitute between the products / services within that set. However, in case of digital
markets, the price-based tools become redundant as platform owners often provide products or
services for free in exchange for data. In such markets, non-price factors such as quality or privacy,
are more important to consumers.

Another consideration in defining markets in case of multi-sided platforms, such as Facebook, where
the stakeholders using the platform includes consumers, advertisers, and content providers is that
substitutability must be reviewed with respect to each stakeholder group, each operating system
and underlying functionality.

For instance, in the merger between Facebook and WhatsApp, the EC defined the relevant markets
at three levels – consumer communication services, social networking services and online
advertising services. For consumer communication services, the EC further assessed the possibility of
delineating the market based on the platform used (smartphone, PC or tablets), operating system
(Windows, Mac, Android or iOS) and functionalities (text messaging, SMS, MMS, e-mails etc.). Based
on market investigations and factors such as (a) consumers’ and producers’ perception about
consumer communications applications on different operating systems, (b) the overlap of consumer
communications applications with the parties’ functionalities and (c) the pricing conditions across
functionalities, the EC opined that the market delineation based on functionalities and the EC opined
that the market delineation based on functionalities and operating system was inappropriate. The EC
ultimately defined the relevant product market as the market for consumer communications
applications for smartphones. For social networking applications and online advertising, the EC left
the market definitions open.

While the EC has so far relied on substitutability across these parameters to define the relevant
markets, according to recent literature on digital markets, the SSNIP can be modified to take into
consideration non-price factors such as quality, i.e., defining the relevant market by measuring the
effects of change in quality of service (known as Small but Significant NonTransitory Change in
Quality Test (SSNIQ)).75 Even though change in quality is more difficult to measure and quantify
compared to price, consumers’ conduct might provide an indicative measure about their
preferences when quality changes. It is also possible to consider change in costs to consumers in
terms of the attention that the consumers pay while buying the free product or the information to
be provided to use the product/service, as a parameter to undertake the SSNIP test.76

Assessing market power and the impact of the transaction on competition in the relevant market

Structural indicators of market power, such as market shares can be misleading in case of assessing
impact of competition from digital mergers because of the dynamic nature of these markets, which
makes it possible for market participants to displace market leaders relatively quickly. Secondly,
often the products / services offered by digital platforms are provided free of cost which makes the
estimation of market shares based on revenues, an incorrect indicator of the merging parties’
market power.

This implies that market power assessment in case of digital mergers should focus on the
exclusionary power resulting from the merger, which is not necessarily associated with higher
market share or profits and requires examination of other factors. For instance, in the acquisition of
Skype by Microsoft in 2011, the EC opined that market shares gave limited indication of competitive
strength because of the dynamic nature of the market and that volume-based market shares were
better than value-based market shares. The EC’s assessment of the two relevant product markets -
the consumer communications services and the enterprise communications services, was based on
the following:

Another important consideration for competition assessment of digital mergers is the possibility that
greater access to and control of data might translate into market power if it reduces the competitive
constraint from other firms. In other words, access to important data insights related to customer
behaviour may allow companies to use data to squeeze out competitors. Access to unique data
insights may also be construed as barriers to entry for potential new rivals.

For instance, an important issue assessed by the EC in its review of Apple’s acquisition of Shazam
was to determine the ability of the merged entity to undertake any anticompetitive conduct based
on the concentration of data.

In particular, the EC assessed whether Apple could improve the performance of Apple Music’s
customer acquisition channel using the data acquired through Shazam, by engaging in targeted
advertising for customers of rival providers of music streaming services such as Spotify. The EC’s
assessment found that there existed various factors which ensured that the parties could not use the
combined datasets to engage in any anticompetitive conduct. First, the EC found that there would
be some legal limits for Apple to use the information about the customers of its competitors.
Second, it was possible for the parties’ competitors to gather information similar to that collected by
Shazam, which gave Apple limited incentive to engage in any anticompetitive conduct using
Shazam’s data. Therefore, the Commission concluded that it was unlikely that the data increment
brought by Shazam would provide Apple with competitive advantage to cause any anticompetitive
effects by reducing the ability and incentives of other digital music streaming providers to compete.

The discussion above indicates that in Europe, antitrust enforcement is becoming receptive to
adopting a forward-looking approach to assess competition in digital mergers. This is evident from
the tools such as market investigations, heed to complaints regarding concentration of consumer
data, or reliance on factors other than market shares to indicate market power, that the EC has used
in its review of these mergers. Though the FTC judgement on many of these transactions is not
available for public review, evidence such as the FTC’s notice to Facebook and WhatsApp to ensure
consumer privacy, indicates that the FTC has also shown willingness to raise consumer protection
concerns in the merger context, when needed.

While this shows the intent of the regulators to address competition issues in this dynamic market
by adopting a dynamic approach, it also reflects an approach to competition assessment that is not
rule based and hence susceptible to being challenged.

Moreover, while in the EU, the Commission has recourse to other mechanisms to review mergers
which do not meet the threshold requirements, the CCI does not have the regulatory authority to
assess non-notifiable transactions.

Even when transactions in the digital market have been notified, like the merger between Flipkart
and eBay87 or Walmart and Flipkart,88 the CCI has applied standard measures of assessing market
power and competition such as market shares, barriers to entry, extent of vertical integration, extent
of competition likely to remain after the combination etc. The CCI has however not assessed the
transactions with respect to the dynamic nature of the digital markets or the potential
anticompetitive conduct arising from combinations of data held by the parties, plausibly since most
of these transactions involve only a small element of digitization through the e-commerce platforms
of the parties.

With digital advancement, it is likely that the transactions in the digital space will get more complex.
This requires that the CCI follows a framework to be able to effectively assess mergers in the digital
market. Factors to examine are whether mergers can lead to dominance by merged entities and as a
result increase the likelihood of anticompetitive abuse of such dominance, or whether they promote
efficiency, innovation, competition and consumer protection, and finally how the CCI can identify the
possibility of such market power and its abuse or efficiency generation, ex-ante. Taking lessons from
the judgements on some of the global transactions in this space, we list the following
recommendations for the CCI to assess digital mergers:

Since turnover-based thresholds sometimes are unable to capture mergers in the digital space, it is
critical to revise the definitions of thresholds to suit the characteristics of digital markets. To address
these concerns,89 especially after the acquisition of WhatsApp by Facebook fell short of the
notification thresholds, Germany amended its competition laws in 2017 to include size of transaction
based thresholds in order to subject start-up acquisitions to greater scrutiny.90 Transaction value-
based thresholds thus might provide a reasonable alternate to estimating the potential impact of a
transaction when the turnover of the target is not significant enough to raise competition concerns.

Since the existing tools to define relevant markets may have shortcomings if applied in case of digital
mergers, it is critical that while evaluating mergers in this space, existing tools such as SSNIP are
modified – for instance to account for the forward looking and dynamic nature of the market. It is
also critical to assess whether the platform’s activities as a whole can be characterized as a stand-
alone market or further delineation based on functionalities, platforms, operating systems etc. is
possible.

Some of the factors which must be assessed for assessing the impact of competition include: a. The
presence of competition which mitigates the role of network effects in conferring market power to
the parties b. The concentration of data with the parties post the transaction and whether there
exist mitigating factors to ensure the parties do not create barriers to entry, foreclose competition or
harm consumers in anyway. c. The role of innovation in bringing efficiencies which benefit both the
consumers and service providers using the merging platforms. The discussion in this paper shows
that the digital economy world over as well as in India, is creating channels for different users to
connect with each other, thereby making life easier, more productive and enjoyable for people. In
such a scenario, there is a need for the government to adopt a flexible yet balanced competition
framework that reflects the market changes and provides a sound foundation for competition,
innovation and consumer protection

Digitalisation, new technologies and scientific breakthroughs are unfolding on many fronts.
Advances in communication and data processing are not just profoundly affecting existing industries,
but also rearranging global value chains, thereby allowing for entirely new products and services and
disrupting traditional ones. These trends can deliver benefits and stimulate economic growth;
however, they can also give rise to competition concerns as well as create needs for new regulation.
Since the effects on society extend far beyond the digital technology context alone, concerns arising
from digitalisation have become increasingly relevant for both policy makers and stakeholders.

In view of the dynamic nature of competition, and especially the need to preserve incentives to
invest and innovate, finding the appropriate scope of competition enforcement in digital markets is a
controversial issue. Some believe that competition authorities should not intervene in fast-moving,
technology-driven industries, as the cost of intervention would generally outweigh the potential
benefits. At the opposite end, others argue that a timely intervention is needed; although in certain
cases this cannot be enough to prevent or discipline market power of digital giants.

When applying competition law, competition authorities may face several challenges affecting both
the traditional analytical framework that they apply and the empirical tools that they use. The
following part will review some of the main challenges in turn.

The assessment of market power and more broadly, the analyses of competitive effects, are typically
undertaken with reference to the boundaries set by the relevant market. Under the traditional
analytical framework, the relevant market is a concept that must be defined first, and then this
forms the background for the analysis that follows. In digital markets, however, the traditional
approach does not necessarily work because – as highlighted before – digital firms continuously
redefine the boundaries of a market or create new markets. Filistrucchi et al. (2013) analysed several
European and national antitrust cases involving two-sided markets, related to both digital as well as
non-digital markets.15 They conclude that in two-sided markets, competition authorities should take
into account both sides of the market when defining the relevant market. Moreover, they should
distinguish two-sided transaction markets and twosided non-transaction markets. This distinction is
essential because in the case of transaction markets, one relevant market, while in the case of non-
transaction markets, multiple relevant markets should be defined. In line with this, when services
are offered at a zero price, for example in the case of a social network, traditional analytical tools,
which tend to generate a rather static perspective of the relevant market, can fail to function.
Several authors suggest, for example, employing a modified version of the SSNIP-test,16 that takes
into account the two-sided nature of the market and incorporates the indirect network effect
between the two sides.17 In particular, in a two-sided non-transaction market, one should check the
profitability of a rise of the price on each side of the market, while in a twosided transaction market,
one should instead check the profitability of an increase in the total price level (i.e. the sum of the
prices paid for the transaction by the two parties). Ideally, in both cases one should allow the
hypothetical monopolist to adjust the price structure

All these considerations question the use of market shares or profit margins as relevant variables,
and point instead in the direction of a better understanding of the business model and how digital
platforms generate turnover and profit. This approach requires a closer look at indicators that
inform about contestability, such as the presence of entry barriers, the availability of alternative
routes to reach end-users, and the extent to which both incumbents as well as challengers are trying
to create new markets. These indicators may help to identify potential rivals, which can steal away
profits and also reflect to the notion that competitive pressure does not only come from demand-
side and supply-side substitution, but more generally from disruptive entry and innovation.

Apart from assessing dominance, it can also be challenging to assess whether market power may
give rise to competition concerns through exclusionary or exploitative conducts. Similarly to
vulnerable and contestable market positions, digital business models may easily blur the lines
between pro- and anticompetitive practices, therefore competition authorities need to find ways to
distinguish anticompetitive behaviour from normal business strategies.

A widespread use of algorithms, for example, may raise concerns of possible anticompetitive
behaviour as it become easier for firms to achieve and sustain collusion without any formal
agreement or human interaction. While algorithms might be used to implement virtually any
anticompetitive conduct that is typically observed in traditional markets, a particular concern
highlighted in the literature is the risk that algorithms may work as a facilitating factor for collusion
and may enable new forms of co-ordination that were not observed or even possible before. This is
referred to as ‘algorithmic collusion’. One of the main risks of algorithms is that they expand the grey
area between unlawful explicit collusion and lawful tacit collusion, allowing firms to sustain profits
above the competitive level more easily without necessarily having to enter into an agreement.

Another possible challenging area relates to the question of leveraging of market power, i.e. using
the market power in the primary market to gain grounds in other related or neighbouring market,
which is very common practice in digital markets. Such behaviour, however, can expand the
innovation frontier in those industries by introducing new technologies and new (platform-based)
business models in other markets. Digital companies that are familiar with both the technology and
the new business models are looking to leverage their qualities into other industries. They compete
on the basis of their merits as well as on the basis of market power. By doing so, they urge
incumbents in those industries to innovate as well. When monopolies are controlling bottlenecks
that are contested by new business models, there is a risk for defensive leveraging, i.e. trying to
prevent others from gaining ground in your market. Defensive leveraging, however, is not just about
gaining additional profits from a second market, but it is an attempt to defend the primary monopoly
position.

Was antitrust doctrine—developed primarily in a bygone era of smokestack industries—

appropriately designed for the digital age?

If the antitrust enterprise is to play a meaningful role in years to come, it must evolve to address the
unique challenges posed by digital markets.

Moreover, the proper focus is not merely on whether some type of rudimentary entry can occur.
Instead, the question is whether the type of entry that would provide a meaningful competitive
check on dominant firms can occur.93 In many digital markets, such meaningful entry is surprisingly
difficult.

Consider, for example, Google Maps, the leading online map application. Google developed the
present iteration of Maps over a period of several years by acquiring several smaller firms at
considerable cost. These acquisition targets included Waze, a direct horizontal rival with access to a
unique treasure trove of self-reported user data.94 Developing Maps also required creating specially
outfitted camera cars; collecting more than 21.5 billion megabytes of street-view imagery from
around the world; employing computer-vision techniques to transform satellite and aerial imagery
into three-dimensional building shapes;95 combining multiple sources of place data to identify the
locations of bars, restaurants, shops, and even clustered “areas of interest”;96 leveraging
proprietary user location data to determine how busy a given bar or restaurant is in real time;97 and
much more. In theory, it may be possible for a small team of programmers to rapidly develop a
rudimentary online mapping service that would “compete” with Google Maps. But developing a
meaningful constraint on Google Maps would be—to put it mildly—no small task. As of 2018, Apple
had more than $285 billion in cash on hand,98 as well as unique access to millions of individuals’
personal data via its own proprietary mobile OS. Yet even Apple, with all of its distinct competitive
advantages, struggled mightily to gain traction against Google Maps. The launch of Apple Maps in
2012 was widely derided as a “debacle.”99 Four years later, nearly 70% of Apple’s own smartphone
users still identified Google Maps as their preferred map application.

Network effects often constitute a particularly salient barrier to entering digital markets.101
Network effects pose relatively little difficulty to first movers. It is subsequent rivals who must offer
not only a product that is better ceteris paribus, but a product that is so clearly better as to outweigh
the incumbent’s network advantage. Nonetheless, anti-enforcement commentators often downplay
the importance of network effects by pointing out high-profile examples of disruptive entry in
networked markets.

These markets are often characterized by positive direct and indirect network effects. Social
networks attract new users by presenting them with the opportunity to interact with other
users.105 The value of a given network to users thus increases along with the size of the network, an
example of direct positive network effects.106 Some social networks allow third parties to develop
compatible applications, thereby introducing the possibility of indirect network effects. Online
search engines may enjoy a type of indirect network effect if “users [do] not consider, when deciding
whether to run another query, that the results of their query and subsequent clicking behavior on
suggested links are stored by the search engine.”107 Thus, each “next” user benefits from the
behavior of past users. Due in no small part to the entry barriers described above, digital giants cast
long shadows. Even the mere presence—and certainly the activities—of an incumbent like Google or
Facebook in a given market can hinder entry and stifle innovation. This dynamic may not be entirely
unique to, but does appear to be particularly acute in, digital markets. Yet, perhaps because
collecting sufficiently rigorous empirical evidence of consumer-welfare harm is difficult in this
context,109 it has gone largely overlooked. The evidence gathered to date suggests that the
presence of Google or Facebook in a market can hinder innovation in that market. Recent empirical
work indicates that after Google vertically integrates into the market for an app that runs on its
Android mobile OS, the developers of existing apps in that market reduce their own efforts to
continue innovating. As market concentration continues to rise, in part due to relatively lax antitrust
enforcement, start-up rates are declining across all sectors of the economy.

A particular type of strategic conduct by an incumbent—even if legal—can also disincentivize entry

and innovation. Dominant digital firms are in a unique position to clone, or mimic, small startups’
features.113 Over time, such free-riding may dissuade startups from even attempting entry. A
hypothetical illustrates the problem. Suppose a new platform, E, enters the social networking space
with hopes of attracting users via an attractive, unique feature. If the social networking space were
characterized by vigorous competition, E might stand a good chance of success. Even if an existing
rival were to mimic E’s feature, E would remain the first mover as to that feature.114 The would-be
copycat has no unique strategic advantage to exploit, leaving E free to compete on the merits.

The saga of Snapchat, a multimedia messaging app, provides a ready example. Noticing the traction
Snapchat was gaining among teenage. users, Facebook offered to buy it.116 When that attempt
failed, Facebook turned instead to mimicking Snapchat’s features.117 Google reportedly offered to
buy Snapchat as well,118 then similarly pivoted toward mimicry.119 Multiple analysts credited these
tactics with depressing Snapchat’s user growth and share price.120

Emerging empirical evidence suggests this is not an isolated example.121 After surveying dozens of
investors and entrepreneurs, one technology reporter concluded that Facebook’s free-riding “is
having a profound impact on innovation in Silicon Valley, by creating a strong disincentive for
investors and start-ups to put money and effort into creating products Facebook might copy.”122
According to a founder, Amazon casts a similarly long shadow: “People are not getting funded
because Amazon might one day compete with them.”123 At a University of Chicago panel
discussion, venture capitalist Albert Wenger depicted the shadows around digital giants like Google,
Facebook, and Amazon as “Kill Zones,” that is, “areas not worth operating or investing in, since
defeat is guaranteed.”124
Nevertheless, anti-enforcement scholars and stakeholders contend that digital markets should evade
antitrust scrutiny because “competition is just [a] click away.”125 The claim is, essentially, that
demand is extremely elastic in digital markets because low switching costs allow customers to
substitute easily among competing products. In a technical sense, of course, a user can physically
click (or tap) her way from one search engine, social network, or online retailer to the next. But in
reality, the cost of that click can be much higher than orthodox antitrust analysts have previously
imagined. If an incumbent has created the lowest-cognitive-load ecosystem, a user will find it
relatively easy to click from (for example) Google’s search engine to Google’s email service to
Google’s videosharing platform to Google’s map application, and so forth. But those are not the
“clicks” anti-enforcement commentators invoke to defend their pro-defendant position.

The sort of click that would matter—away from using one search engine and toward using another—
entails a level of cognitive burden much higher than what is required to simply click around within
Google’s ecosystem.

Lack of data portability may raise users’ switching costs higher still. To illustrate, consider a given
social network user.128 At the outset, the user makes her choice among the available networks
based on a range of quality and price considerations.129 But once an individual starts to use a
particular service, that service becomes a repository for her photos, conversations, status updates,
contacts, and more.130 Unless her data is portable across platforms—and it generally is not131—
she cannot easily switch to a different social network, even if she would otherwise prefer to do
so.132 Moreover, because most digital products can be improved by personalization based on past
user experience, the passage of time makes it increasingly difficult for rival networks to offer an
equally valuable product.133 The argument that “competition is just a click away” in digital markets
is overly simplistic, bordering on naive.

Even if switching costs were exceedingly low, the potential for anticompetitive conduct and effects
would remain.

Despite the fact that digital markets frequently exhibit high barriers to entry, skeptics of antitrust
enforcement have one card left to play: they portray digital markets as nonetheless being
characterized by intense innovative rivalry.

As a result, the argument runs, antitrust would move too slowly to correct any problems and is
unnecessary because the relevant markets will quickly correct themselves.136 Under this view, the
lure of monopoly profits will inevitably attract disruptive upstarts seeking to replace dominant
incumbents—and monopoly is actually good and desirable because it is necessary to spur
technological progress.137 This unorthodox vision traces its roots to Schumpeter’s decades-old
invocation of “creative destruction,”138 which became a favorite trope among those associated with
the Austrian and Chicago schools.139 For empirical support, proponents of this digital
creativedestruction narrative commonly point to Facebook’s “disruption” of MySpace and Google’s
“disruption” of Yahoo.

The anecdotes of MySpace and Yahoo, still commonly cited by those who argue that digital markets
are epicenters of creative destruction,150 look increasingly creaky with age. The relevant markets
have been characterized not by the “gale” of creative destruction described by Schumpeter, but by
entrenched and unchecked dominance. It is high time to abandon the “romantic but naïve
Schumpeterian [notion] that giant” monopolists and concentrated oligopolies are necessary for
technological progress.151 In fact, a more sophisticated reading of Schumpeter suggests that he was
not nearly so opposed to government intervention—particularly in the form of antitrust
enforcement—as his modern-day adherents tend to be.152 An antitrust enterprise that somehow
came to view monopoly as good and necessary has rather clearly lost its way.1

The societal cost from false negatives is substantially higher than pro-defendant analysts have
previously assumed. Normatively, this militates in favor of an invigorated approach to digital
markets.

Digital markets are susceptible to at least two (and likely many more) unique anticompetitive
strategies. These are referred to herein as “no escape” and “split-the-rents.”154 To date, these
strategies have gone unnoticed, or at least unremedied, by the antitrust enterprise. This failure
suggests that the consensus view regarding the balance of error costs in this context rests on yet
another flawed assumption. The likelihood of false negatives occurring is higher than previously
imagined.

When a digital-product provider attains a dominant position in its own market, it may be able to
steer its users to a favored counterparty operating in a different market. If such steering causes
consolidation of that distinct market, the dominant firm and its favored counterparty may be able to
share in the resulting rents. This ability to split the rents from a different market would make the
steering strategy rational for the platform.

providers (like Google) similarly bring together advertisers and users—but also indirectly bring
together sellers and buyers of other products. Thus, for example, a search user might search for
“local restaurants” and be shown a display ad next to the search results. But that search user is likely
also seeking to engage in an offline transaction: the purchase of a meal from a restaurant. The
search provider facilitates this offline transaction, though the latter occurs in a market distinct from
the provider’s core business.

A search provider could, of course, simply provide “neutral” search results to its users. But it could
also extract payment from advertisers in exchange for a more favored spot in the results,178 alter a
“reputational system” to favor certain advertisers,179 or otherwise tilt the offline playing field by
“steering” users toward favored advertisers.180 The ultimate effect is foreclosure of non-favored
sellers from—and increasing concentration in—the offline market. In other words, digital steering is
likely to cause consolidation of markets that appear to be distinct from the core digital market(s).
The resulting rents, if split between the dominant provider and its favored counterparties, make the
scheme rational.

Traditional antitrust analysis could easily overlook the possibility of such harm. To illustrate how this
type of false negative might occur, suppose two search providers propose to merge. The relevant
market(s) would not likely include local restaurant markets and the like, because neither of the
merging parties would be treated as a direct market participant in such markets. As a result,
traditional analysts would presumably ignore the possibility of split-the-rents harm.

The acquisition of digital real estate portal Trulia by its direct rival, Zillow, provides a possible real-
world illustration of such harm occurring due to a false negative on the part of an enforcement
agency. In 2014, Zillow and Trulia announced plans to combine into Zillow Group (“ZG”). “At the
time, the two firms were the largest and secondlargest online real-estate portals, respectively.”182
After conducting a pre-merger review, the FTC cleared the deal without condition.183 During a
subsequent earnings call, ZG’s Chief Executive Officer, Spencer Rascoff, announced that “Zillow
Group represented greater than 67 percent of the total online real estate category . . . and 78
percent of the category on mobile only.”184 Those market shares would likely be enough to warrant
a presumption under U.S. antitrust law that ZG possesses monopoly power185 and would almost
certainly have made the proposed deal presumptively anticompetitive ex ante.186 By its own
admission, then, ZG appeared to have gained a dominant market position.

Additional statements from ZG executives suggest a subsequent shift in strategy to steer users
toward “Premier” real estate agentadvertisers and away from non-Premier agents. As Rascoff put it,
“[W]e will continue to encourage lower performing agents to leave”187 while helping Premier
agents “grow their market share in their respective cities.”188 More specifically, ZG’s strategy was
intended to have the effect of “accelerating the larger trend across the real estate agent population
of higher producing agents gaining market share from those who are less competitive.”189 Suppose
that ZG’s strategy had its intended effect of increasing concentration in offline real estate agent
markets. For this strategy to be rational, ZG would need a means of splitting the resulting rents with
its favored realtors. As it turns out, the “Premier” agents that benefit from ZG’s steering appear to
be those who pay fees to ZG.190 These payments could be viewed as a mechanism for splitting the
rents between realtors and ZG.

Split-the-rents harm is uniquely facilitated by the attributes of digital markets. The same features
that drive users to digital portals also render users uniquely susceptible to this type of steering.192
In fact, there are some parallels to the European Commission’s Google Search (Shopping) decision.
The Commission’s basic theory of harm was that Google used its dominant general search engine to
steer users toward its own comparison-shopping service, leveraging its power over general search to
increase its power over comparison shopping.193 Users access portals like Google, Zillow, and the
like as a means of cutting through the fog of information overload, which is felt most acutely in
digital contexts.194 It is this unique backdrop that makes users particularly vulnerable to steering—
and makes steering strategies more attractive—in digital markets.195 When steering crosses the line
into outright deception, it may violate consumer-protection laws. But if and when it is used to
facilitate market consolidation, it is an antitrust problem.

Anticompetitive product design (and redesign) is a relatively well-accepted theory of antitrust

harm.196 The archetypical designconduct claim alleges that a dominant firm redesigned its core
product so as to favor its own complementary product197 and/or disfavor rivals’ products.198 Such
design- and redesign-related strategies have spawned a rich body of antitrust precedent.199 For a
variety of reasons, leading cases and scholars advocate for heavily pro-defendant rules in this
area.200 But, as the following discussion demonstrates, their arguments are inapposite in digital
markets. Instead, several unique characteristics make digital markets an ideal context in which to
deploy anticompetitive product design strategies.

Plaintiffs have brought multiple cases alleging anticompetitive product design in digital markets.
Microsoft III was the earliest—and remains the most prominent—of these.201 There, a dominant
firm (Microsoft) issued a new version of its core product (the Windows operating system) that was
designed so as to maximize interoperability with its own complementary product (Internet Explorer)
and minimize interoperability with a rival’s product (Netscape Navigator, a competing web
browser).202 The plaintiffs’ allegations in In re Apple iPod iTunes Antitrust Litigation, if taken as true,
provide another example.203 According to the complaint, Apple issued software updates to a core
product (iPods) that were designed to block interoperability with a rival’s product (RealNetworks’
low-price music files).20

The modern antitrust enterprise employs a relatively laissezfaire approach to conduct involving
product design. Some go so far as to argue that courts and enforcement agencies should treat all
productdesign strategies as per se legal.205 Proponents of such extreme positions argue that
product design is uniquely unattractive as an exclusionary strategy. In brick-and-mortar markets,
theorists posit that “product innovation is extremely costly and time consuming to develop, design,
manufacture, and place on the market”206 and that product redesigns done purely to disfavor rival
products would likely prompt a negative customer reaction.207 If the would-be monopolist were
thereby forced to reverse course, it would forfeit any sunk costs invested and perhaps incur
additional reversal costs by switching back to its former product design.208 As a result,
anticompetitive product design was thought to be quite rare. Redesigning code-based products,
often done through issuing updates to existing software or altering HTML or algorithms, can
generally be accomplished at far lower cost than redesigning physical products. A software update,
however, can be created by a single programmer or small team working at their desks (or even at
home).213 Moreover, digital distribution is generally much less costly than offline distribution. On
the other side of the scale, monopoly profits are often higher in digital markets. As the leading
treatise recognizes, strategically designed incompatibility can cause “serious anticompetitive
consequences, particularly in ‘network’ industries where compatibility itself is often an essential
ingredient to product success.”215 This is so because the resulting monopoly power is uniquely
durable. Whether labeled “implicit tying,”219 “technological tying,”220 “predatory innovation,”221
or something else, design-related exclusionary strategies are uniquely attractive in digital markets.

On October 10, 2019, the Belgian, Dutch, and Luxembourg competition

authorities published a joint memorandum on "challenges faced by
competition authorities in a digital world."[6] Among the main issues
identified for further consideration in the EC's report on competition in the
digital era, the three agencies recommended the introduction of a non-
punitive, ex ante instrument providing for binding commitments without the
establishment of an infringement. According to the Benelux authorities, this
tool would allow the EC and Member State regulators to impose
proportionate remedies on dominant companies to prevent competition
problems, rather than relying on ex post enforcement, which could be too
late to keep the fast-moving digital markets competitive.
It may also mean the development of appropriate ex-ante tools to intervene
even before an infringement takes place, or the design of regulations to set
out clear rules about what dominant platforms can do and what they
cannot."

In view of the dynamic nature of competition, and especially the need to preserve incentives to
invest and innovate, finding the appropriate scope of competition enforcement in digital markets is a
controversial issue. Some believe that competition authorities should not intervene in fast-moving,
technology-driven industries, as the cost of intervention would generally outweigh the potential
benefits. At the opposite end, others argue that a timely intervention is needed; although in certain
cases this cannot be enough to prevent or discipline market power of digital giants.