A Reconception of Power in Competition Law with Its Application in the Digital Realm

Abstract

As a fundamental concern of competition law, the concept of power remains inadequately understood. The limitations have become increasingly evident in the digital economy realm where new challenges have often not been addressed promptly, coherently, and effectively. Although new theories and analytical tools have been proposed, they frequently result in overlaps and gaps. There is a pressing need to establish a more coherent theoretical foundation that integrates traditional and emerging theories of power. Tapping into legal philosophy and global empirical evidence, this article proposes a reconceptualization of power as an overarching theoretical framework. This reconceptualization differentiates between power as capacity and its exercise, classifies the various forms of power exertion, and identifies several distinct types of power. The reconceptualization elucidates how different types of power—such as economic, political, informational, and technological—all contribute to power dynamics in markets. By broadening the dimensions of power analysis in competitive assessment, the proposed understanding of power frees legal discourse from the limitations of a purely economic perspective. The Facebook/WhatsApp merger is analysed as an illustration. However, the proposed conception is also relevant for assessing other competitive issues, such as the power interplay between economic and noneconomic factors like privacy and personal data protection, and democracy.

I. INTRODUCTION

The digital economy has led to significant disparities in competition law globally.¹ On the one hand, identical cases have been approached in starkly contrasting ways in different regions. A prime example is the Google Search (Shopping) case. In the EU, Google’s self-preferencing conduct incurred a record-breaking fine of €2.42 billion.² Conversely, the United States concluded that there was no problem with Google’s self-preferencing conduct, voting unanimously to refrain from further action.³ These two prominent jurisdictions maintain fundamentally different perspectives on whether Google’s self-preferencing conduct, which exploited market power across different markets, should be classified as ‘competition on the merits’.⁴

On the other hand, divergences have arisen when a competition authority substantially alters its position on the same or similar issue, or when previous investigations or analyses have been found to be inadequate. For example, in one of the earliest cases regarding free online services in 1997, the European Commission (EC) expressed scepticism in the Bertelsmann/Burda-HOS Lifeline case, questioning whether ‘online health information that anyone [could] view for free on the Internet [could] even be considered a market.’⁵ In this case, there was neither a market definition nor competitive assessment of the free online services.⁶ However, the EC changed its perspective before long, defining the relevant market and conducting a competitive assessment of free online music delivery services.⁷

The divergences are not only historical. A representative example is the EU’s shift in its perspective regarding whether competition law can address privacy or personal data protection issues. Initially, both the Court of Justice of the European Union (CJEU) and the EC asserted in their cases that privacy and data protection matters should be addressed by data protection regulations rather than competition law.⁸ However, this viewpoint has shifted, and currently, both the CJEU and the EC acknowledge that privacy and personal data protection issues can fall within the purview of competition law.⁹

The divergences are not confined to specific regions. For example, apart from the EU, mergers approved in the past decade in the US have now become targets for divestments. In 2012 and 2014, Facebook (currently Meta) acquired Instagram and WhatsApp, respectively. Back in 2012, the US Federal Trade Commission (FTC) concluded its nonpublic investigation into Facebook’s acquisition of Instagram without taking any action, with a unanimous 5–0 vote.¹⁰ The acquisition of WhatsApp by Facebook received approval with an FTC notification highlighting the obligation to protect users’ privacy.¹¹ However, in 2020, a shift occurred as the US FTC and 48 attorneys general from various states separately filed two lawsuits, both advocating for the divestment of WhatsApp and Instagram from Facebook.¹²

Nor are the divergences limited to mergers. In Germany, the Facebook personal data exploitation case experienced dramatic changes. Initially, the German competition authority found Facebook’s personal data processing terms in violation of German competition law, citing both user data exploitation (exploitative abuse) and the creation of an illegal competitive advantage for Facebook (exclusionary abuse).¹³ However, in the first trial, the Higher Regional Court of Düsseldorf disagreed, stating that Facebook’s personal data processing did not exploit users nor result in an illegal exclusion of competition.¹⁴ The next reverse occurred when the case was appealed to the German Federal Court of Justice Bundesgerichtshof. The Federal Court disagreed with the Higher Regional Court and upheld the enforceability of the competition authority’s decision with a different theory of harm.¹⁵ The Düsseldorf Higher Regional Court took the disputed issues further to the CJEU for a preliminary ruling with several questions.¹⁶ The questions include, among others, whether consent could be effectively given freely to a dominant undertaking like Facebook at all.¹⁷ This question is related to the exploitation accuse—if the answer is no, then dominant undertakings would not be able to provide effective choices to avoid exploiting their users in competition law. The CJEU ruled that holding a dominant position itself does not prevent users from validly giving their consent.¹⁸ What matters is to ensure the freedom of choice.¹⁹ This opinion supports the rule of the German Federal Court of Justice.

In the meantime, competition law has also made significant progress, albeit with time lags and through separate avenues. New theories and analytical tools have been put forward to address emerging challenges, such as from the perspective of dynamic competition,²⁰ two-sided or multisided markets,²¹ attention rivalry,²² data,²³ gatekeeping power,²⁴ and business ecosystem,²⁵ among others. However, the theories may overlap, such as the attention rivalry theory and the two-sided market theory.²⁶

Although the advancements in competition law are widely recognized, controversy and debate persist. For example, although some advocate for using competition law to address personal data protection issues,²⁷ others believe sector-specific regulations are more appropriate than competition law.²⁸ Furthermore, there is growing attention on the interplay between competition law and noneconomic issues, such as whether social and political factors can shape competition law,²⁹ and whether the law’s objectives should extend beyond economic considerations.³⁰

Above all the advancement of theories and tools for competitive assessment of the digital sector, there still lacks an overarching theory to explain (1) what impeded conventional competition law from addressing novel competitive issues promptly, coherently, and effectively; (2) where are the gaps in conventional theories, and why do they exist; (3) how can the distinct new theories be comprehended more cohesively; and (4) how can competitive analysis be improved when confronting a novel issue in the future. It is crucial to clarify from the outset that this research has no intention to overturn all the established theories and tools of conventional competition law or undermine the significance of economics. Instead, its purpose is to highlight that competition law is not only about economics and to show what is important beyond economic considerations.

With the clarification and revolving around the questions, this research not only draws upon legal philosophy but also uses international materials to conduct comparative analysis. The hope is that, given the commonalities in competition law across different jurisdictions (and without prejudice to the differences), the analysis can provide inspiration to different countries and regions. The first section of this article establishes the cornerstone by scrutinizing the necessity for a reconception of power. This includes demonstrating why a more precise comprehension of power is crucial, conducting an examination of current power conceptions, and highlighting their inherent limitations. The second section revisits the nature of power and offers a reconception. It examines power from different perspectives, encompassing its nature as capacity, the exercise of power, various forms of power, notable types of power, and the interplays of different power types. It also considers the power of competition law. After the reconception, the third section proceeds to demonstrate the benefits of the proposed power conception to competitive analysis in practice. The Facebook/WhatsApp merger is employed as an illustration. The concluding section summarizes and references other potential issues that the proposed power conception may help address.

II. THE NEED FOR A RECONCEPTION OF POWER

Before providing a reconception of power, two premising questions should be answered. First, in addressing novel issues and challenges in competition law, whether a deeper understanding of power is crucial. Second, with all the conventional and newly proposed power conceptions, whether there is still a need to have another reconception of power in competition law. This section aims to address these questions.

A. Why a Deeper Understanding of Power Is Crucial

The conceptualization of a subject mirrors one’s understanding of it and, in turn, shapes perspectives when exploring and explaining the world with the concept.³¹  Power is of paramount importance in competition law.³² As will be presented below, its significance manifests in various dimensions—facilitating the articulation of competitive benefits and harms, fostering the development of competitive tools, and demarcating the boundaries between competition law and other laws.

1. Power and the Articulation of Benefits and Harms

Power plays a fundamental role in articulating competitive benefits and harms. In this process, two aspects of power are important. First, the degree of power. Take the EU as an example. In abuse of dominance cases, the establishment of significant power, a dominant position, is the premise for further competitive assessment.³³ In merger control, evaluating the extent of consolidated power is a pivotal consideration, as outlined in the merger regulations.³⁴ The degree of power is also essential in analysing anticompetitive agreements, as shown by the De Minimis Notice, which establishes safe harbours for undertakings whose combined power is below certain thresholds.³⁵

Second, how power is exercised is also of great significance. The essence of competitive analysis is to assess how the exercise of existing or potential future power may result in benefits or harms. Power can either be exercised within the same relevant market, such as predatory pricing, or leveraged into another relevant market, as observed in tying and bundling practices.³⁶ The exercise of power may either directly target consumers, constituting exploitative abuse (which is controversial regarding whether it should be prohibited, as the approach differs between the EU and the US), or directly affect competition and consequently indirectly influence consumers, constituting exclusionary abuse.³⁷

2. Power and Competitive Tools

Power is closely related to the tools for identifying, measuring, and controlling harms. To begin with, the understanding of power directly shapes the development of tools for identifying and measuring the extent of power. If one relates power of businesses to pricing, then the design of competitive tools for identifying and measuring power would revolve around the price parameter. This is reflected in the conventional conceptualization of business power—market power—and the established tools for identifying and measuring it. Market power is defined as the ability to raise prices above a certain level, with different views on what the ‘certain level’ is.³⁸ In alignment with this, conventional tools were formulated with pricing considerations at their core. For instance, the Hypothetical Monopolist Test (also known as the SSNIP test) is employed for relevant market delineation by implementing a small but significant and nontransitory increase in price.³⁹ Traditionally, turnover-based pecuniary market share serves as a crucial indicator of the extent of market power.⁴⁰

If understanding power based on price, when the price parameter is absent, doubts may arise regarding whether a free product can form a relevant market, possess market power, and receive competitive assessment. Such doubts were exemplified by the EC's Bertelsmann/Burda-HOS Lifeline case, which questioned ‘whether online health information that anyone can view for free on the Internet can even be considered a market’.⁴¹ Subsequent developments in theories and tools have, to some extent, addressed these doubts, including the market definition and market power assessment tools for free services.⁴² However, this historical review underscores how the conceptualization of power can affect the identification and measurement of power.

Moreover, the understanding of power also significantly shapes the tools for controlling competitive harms. Competition law can use behavioural remedies and structural remedies to address potential harm.⁴³ Behavioural remedies operate by constraining the exercise of power through the imposition of restrictions on specific business behaviours.⁴⁴ Structural remedies directly address the magnitude of power, such as preventing power consolidation in merger reviews, dismantling an undertaking, or divesting assets.⁴⁵ Which remedy to adopt depends on the assessment of how power can be effectively controlled to alleviate competitive concerns.⁴⁶

3. Power and the Capacity of Competition Law

Beyond the power of businesses, the understanding of power holds significance in clarifying whether competition law can address a specific issue and, if so, how. Naturally, competition law does not extend to cover every economic concern. Other legal frameworks, such as sectoral regulations in telecommunications, personal data, and the digital sector, also contribute to fostering competition and enhancing consumer welfare.⁴⁷ The eligibility of competition law to address a particular issue hinges on two factors: (1) whether competition law has the capacity, namely the power, to have an influence on the issue, and (2) decisions in both legislation and practice regarding the utilization of this capacity.⁴⁸

The capacity of competition law lies in its remidial tools. When scrutinizing these tools, understanding power is vital due to the intricate interplay between the factor directly affected by a tool and the ultimate desired effect. For instance, consider a remedy for a merger that involves opening an undertaking’s Application Programming Interface (API) access. The undertaking’s technological power is directly affected, as it has to change its API rules. As a result, other undertakings that obtain access to the API will be able to increase their technological powers. This shift can alter the comparative technological power dynamics between the focal undertaking and others, which can further contribute to maintaining effective competition in the market.

The capacity of competition law may change following different circumstances of power interplay. This is because, due to the existence of power interplay, when competition law directly influences one type of power, it may consequently affect another. Personal data protection serves as an illustrative example. Due to the information that personal data can carry, access to personal data not only alters technological power but also transforms the ability to ‘know individuals’. The latter power can be wielded to inflict harms, such as manipulating the choices of individuals, including noneconomic decisions, as evidenced by the Facebook-Cambridge Analytica data scandal.⁴⁹ Debates frequently revolve around whether a specific harm falls within the purview of competition law. The nature, characteristics, and interplay of power are crucial in elucidating how harms are generated, how to control them, and whether by competition law.

Up to this point, none of the content stated above should be alien to any competition law expert. Nevertheless, a closer examination will reveal issues within the established power theories. The following section will delve into the current understanding of power in competition law and highlight the areas of concern.

B. Current Power Conceptions and Associated Issues

1. Price-Based Power Conception and Its Challenges

The dominant concept widely embraced in competition law is market power. Despite being the central concern of competition law,⁵⁰ there exists an apparent imbalance in the attention dedicated to its conceptual understanding versus its application in practice.⁵¹ Academic discussions predominantly concentrate on the pragmatic aspects of market power, such as delineation of relevant markets,⁵² methods for measuring market shares,⁵³ assessment of entry barriers,⁵⁴ considerations of gatekeeping power,⁵⁵ and evaluations of market power in business ecosystems.⁵⁶

For the conception, lawyers usually adopt that provided by economists, which defines market power as the ability to raise prices above a certain level, with different opinions on what the certain level is (hereinafter the ‘competitive level’).⁵⁷ This conception appears to be influenced by the basic factors and analytical methods in economics, primarily rooted in the demand and supply relationship based on price. The concept is widely adopted. For example, the OECD defined market power as ‘the ability of the firm to keep the price above the long-run competitive level’ in its policy roundtable report.⁵⁸ Similarly, in the United States, where modern competition law originated, the Supreme Court defines market power as ‘the ability to raise prices above those that would be charged in a competitive market.’⁵⁹ In both academic discourse and judgements, the definition of market power is often a mere reference to the economics-oriented conceptualization. When applied, the concept frequently appears as a self-defining term.⁶⁰

The limitations of the price-based market power conception have become evident in the digital economy, where many services are offered free of monetary charge.⁶¹ The logic is straightforward—if market power is a necessary condition for causing competitive harm and if the power is defined as the ability to raise prices above a competitive level and is considered a necessary condition for causing competitive harm, then free services, which do not charge a price, lack the ability to raise prices above a competitive level. Therefore, they are perceived to possess no market power, thus resulting in no competitive harms. Moreover, traditional legal tools based on price parameters, such as the SSNIP (Small but Significant and Non-transitory Increase in Price) test, are not applicable to free services.⁶²

The shortcomings are not merely theoretical but also manifest in practice by significantly affecting competition law’s ability to promptly address issues in the ‘free economy’. A notable example is the scepticism expressed by the EC regarding whether free services could constitute a market in the Bertelsmann/Burda-HOS Lifeline case.⁶³ A closer examination of this case provides further insights. Although the market power conception may not pose a significant barrier to developing tools for measuring power, it profoundly influences perceptions of the existence of power and competitive harm. To continue with the EC’s example, there are different ways to explain the EC’s hesitation.⁶⁴ The first is that technically, there was no method to define a market in the absence of price. This was unlikely to be the reason, as the EC stated in the same year that it followed an open approach to empirical evidence when defining a relevant market, utilizing all available information.⁶⁵ In other words, the EC did not rely solely on quantitative methods such as the SSNIP Test.

Another explanation is that there was no need to define a relevant market for the free service. There are two potential reasons—the absence of market power identification or the absence of harm. To begin with, according to the price-based market power conception, no market power is associated with a free online service with zero price. Therefore, there was no need to define a relevant market for the service because market power assessment is the primary purpose of market definition.⁶⁶ Alternatively, the rationale could be that there was no harm in the provision of free services. Consequently, there was no imperative to consider these services. In essence, either the perception of power or harm was adversely influenced.

Minor improvements in the conception are unlikely to work properly. For example, it could be argued that because profits are made in another market, such as advertising, the price parameter still exists to make the market power conception work when considering the two markets collectively. Although the assertion appears plausible, it is grounded in the frequent classification of the ‘free plus advertising’ business model as a two-sided market,⁶⁷ a classification that the author does not fully agree with.⁶⁸ Under any circumstance, the degree of power within subsidizing and subsidized markets may differ. For instance, while a social networking services provider like Facebook could possess significant market power in the social networking services market, this dominance might not translate to the same level of influence in the online advertising market. This is because although other online services, such as short-form video hosting platforms like TikTok, may not directly rival Facebook’s social networking services, their respective targeted online advertising services may compete with each other. Consequently, the power disparity between their respective content services and targeted online advertising services may be different in magnitude, let alone their nature.

Given this clarification, strictly applying a price-based definition of market power would suggest that free services like Facebook’s social networking and Google’s online search services lack market power. However, such a conclusion is highly questionable, especially considering legal precedents such as the Google Search (Shopping) case and Facebook’s abuse-of-dominance case, which found that free service providers can indeed wield dominant market power.⁶⁹ This discrepancy between conception-based conclusions and legal realities suggests that at least one of them is flawed, and it is unlikely to be the reality.

2. Economic Strength–Based Power Conceptions: Advantages and Shortcomings

a. Economic strength–based power conceptions and their advantages

Apart from market power, there are other power conceptions in competition law. For example, the EU adopts the dominance conception, which is defined as ‘a position of economic strength enjoyed by an undertaking, which enables it to prevent effective competition being maintained in a relevant market, by affording it the power to behave to an appreciable extent independently of its competitors, its customers and ultimately of consumers’.⁷⁰ The US case law employs the monopoly power conception, which is defined as ‘the power to control prices or exclude competition’.⁷¹ In the landmark U.S. Microsoft case, it was elaborated that ‘a firm is a monopolist if it can profitably raise prices substantially above the competitive level’.⁷² This conceptualization continues to be applied, as in the FTC v. Facebook case.⁷³

The dominance conception and the monopoly conception are not replacements but rather developments based on the market power conception. In practice, these two conceptions are often used together alongside the market power conception.⁷⁴ It was pointed out that the EU’s dominance conception was related to the estimation of market power.⁷⁵ Others also contended that both the EU’s dominance conception and the US’s monopoly power conception focused on the price dimension of competition.⁷⁶

That said, both the dominance conception and the monopoly conception (hereinafter referred to together as the ‘economic strength-based power conception’) show clear deviations from the price-based market power conception. They do not simply see the power of a business as a price-related ability. Instead, they encompass additional evidence that substantiate the presence of power. This can be traced back to their definitions—to exclude or prevent (effective) competition. The EU also clarified that price and other economic factors could all be important parameters for evaluating power.⁷⁷ In other words, price and other economic parameters are regarded as evidence of the existence of power rather than power itself.

The impact of such conceptualization is not just linguistic but has practical significance. For example, compared with the price-based power conception, conceptions based on economic strength afford greater flexibility in establishing and evaluating the power of an undertaking when price parameters are absent. This flexibility stems from the diverse manifestations of economic strength across different services, allowing for the utilization of suitable parameters tailored to the nature of the service. Such conceptions accommodate changes seamlessly, avoiding theoretical incoherence. The EU’s Google Search (Shopping) case is an example of this. In the case, the EC observed that a price-based market share evaluation was impractical because (1) general search services were provided free of charge, (2) precise and verifiable values of ‘Revenue Per Search’ for general search services were unattainable, and (3) advertisers looked at search volume when deciding search adverting placement.⁷⁸ Consequently, instead of using price-based parameters such as revenue, the EC estimated Google Search’s market shares based on search volume, which were calculated according to other parameters like page views and site visits.⁷⁹

b. Shortcoming I: gaps arising from path dependence

Although advancing beyond the market power conception, economic strength–based power conception still maintains close ties with market power and often adhere to established analysis frameworks originating from the latter. A prime example is the prevailing analytical approach in cases, the ‘relevant market definition—market power analysis—competitive assessment’ approach.⁸⁰ Competitive assessment entails scrutinizing horizontal, vertical, and conglomerate effects, aligning with legal precedents. This approach is reasonable as it ensures legal consistency and coherence.

Nevertheless, path dependence may hinder the competition law’s capacity to address novel issues promptly, coherently, and effectively. For example, the power dynamics and competitive relationships in digital services may significantly differ from those in traditional sectors, which may lead to potential gaps when applying conventional analyses to the digital sector. This is reflected in Facebook’s merger assessments. When Facebook sought to acquire Instagram in 2012, the U.S. FTC approved the concentration with a unanimous 5-0 vote, taking no further action.⁸¹ A decade later, the FTC filed a lawsuit, contending that Facebook’s acquisition of Instagram should have been blocked. The claim asserts that Facebook acquired Instagram to stifle its potential growth as a strong competitor, and the FTC now proposes the divestiture of Instagram from Facebook (case pending).⁸² This substantial shift and inconsistency in the FTC’s stance suggest oversights or misanalyses during its 2012 merger review. The authority failed to promptly identify and effectively address the emerging issues.

Another influence is less obvious but no less important. Take the Facebook/WhatsApp merger as an example. When Facebook proposed to acquire WhatsApp, the EC investigated and permitted the proposed acquisition.⁸³ WhatsApp’s consumer communications services (WhatsApp CCS) belonged to the consumer communications services market, and Facebook’s social networking services (Facebook Network) belonged to the social networking services market.⁸⁴ In this case, a crucial nonhorizontal concern revolved around the potential for cross-platform integration through user matching—specifically, whether Facebook could use the merged entities’ data to match and recommend users between WhatsApp CCS and Facebook Network.⁸⁵ The conclusion was that Facebook could not engage in such activities, a conclusion that later proved incorrect as Facebook was found to have provided misleading information.⁸⁶ The story seems to have a happy ending. The competition authority was still the righteous protagonist—it was only misled by Facebook’s false information. Facebook, found in violation of the law, was condemned and fined.⁸⁷

However, from a different perspective, the merger’s influence on innovation and competition can be more serious than shown in the EC’s competitive assessment. There was an oversight regarding the potential for WhatsApp to incorporate social networking features and evolve into a strong competitor to Facebook.⁸⁸ A pertinent example can be drawn from the competition between WeChat, which was developed by Tencent Inc., and Renren Network, which was developed by Renren Inc. (now Moatable Inc.) in China.⁸⁹ The former started as a consumer communications services provider like WhatsApp, and the latter began as a social networking services provider like Facebook.⁹⁰ The two did not compete with each other until WeChat added the Moment feature, which had similar functions to Renren Network—sharing texts, photos, and videos.⁹¹

The technological changes have competitive significance. The changed competitive dynamics marked a turning point in the number of Renren Network’s monthly unique log-in users, from previous steady growth to continuous decline, leading to its eventual failure.⁹² Although other factors might also lead to WeChat’s ascent and Renren Network’s decline, the example illustrates how digital services from disparate markets can become competitors by introducing similar features. After the granted merger, WhatsApp never developed social networking features or grew into a potent competitor to Facebook Network like WeChat did to Renren Network.

As will be elaborated later,⁹³ a significant factor contributing to the oversight of this potentiality can be traced back to the power conception, specifically, the path dependence in competitive analysis stemming from the prevailing understanding of power. Currently, competition law does not recognize technological power as a distinct category but rather evaluates technological factors within the framework of economic power. One might contend that there is no substantial distinction as long as the technological element is taken into account.

The differences lie in the details of perspectives. With evolving technologies, competition may manifest in novel forms, introducing additional factors for scrutiny. Therefore, competitive analysis may need to add new perspectives. However, if technological power is amalgamated with economic power, it is more challenging to break free from path dependence and introduce supplementary perspectives. As shown in the Facebook/WhatsApp case, the European Commission did indeed consider many new technological factors in the digital sector, such as the impact of data portability on consumers’ platform-switching ability, and the potential integration of Facebook Network and WhatsApp through cross-platform communication.⁹⁴ Nevertheless, these considerations still adhere to conventional frameworks, encompassing aspects like inputs, lock-in effects, and the integration of final products. Although they addressed some issues, they do not encompass the entirety of how technological power can be formulated and exercised uniquely in the digital sector. This left out the ‘compete by adding features’ perspective.

Although this is not to say it is absolutely impossible to capture unique technological characteristics by making developments on conventional methods, certain challenges arise. On the one hand, novel competitive practice such as ‘compete by adding features’ can may be less conspicuous and easily overlooked when approached from an economics-based power framework. On the other hand, even in attempts to enhance conventional methods, technological power remains pivotal. For example, the discussions on how a digital platform’s control over data can lead to market power and competitive advantage revolve essentially around how data may serve as a source of technological power and how this technological power transforms into economic power.⁹⁵ In essence, there is still a pressing need for a better understanding of the nature of technological power and how it interacts with economic power in the digital sector.

c. Shortcoming II: the noneconomic predicament

The second shortcoming of economic strength–based power conceptions demands more attention—these conceptions still originate from economics and resist considering the noneconomic power wielded by undertakings. This becomes problematic when competition law faces challenges involving noneconomic power. A contentious realm is personal data protection/privacy protection (PDP/PP). Conventionally, issues related to PDP/PP were deemed outside the purview of competition law. In the EU, for instance, the Court of Justice asserted in Asnef-Equifax v. Ausbanc that ‘any possible issues relating to the sensitivity of personal data are not, as such, a matter for competition law, they may be resolved on the basis of the relevant provisions governing data protection.’⁹⁶ The EC also clarified in the Facebook/WhatsApp case that ‘[a]ny privacy-related concerns flowing from the increased concentration of data within the control of Facebook as a result of the Transaction do not fall within the scope of the EU competition law rules but within the scope of the EU data protection rules.’⁹⁷

However, perspectives on the role of competition law have changed over the years, with more agreeing the law can and should address noneconomic concerns. The present approach involves not altering the conception of power in competition law but rather regarding noneconomic issues as economic parameters, such as product quality, to align with the existing power framework. A representative example is the EC’s Microsoft/LinkedIn case, in which privacy protection was regarded as a ‘parameter of competition’.⁹⁸ Although this perspective was not new in academic discussions, it gained popularity after the case.⁹⁹ Similarly, in the United States, in the 2020 antitrust lawsuit by the DOJ against Google, the DOJ asserted that ‘[b]y restricting competition in search, Google’s conduct has harmed consumers by reducing the quality of search (including on dimensions such as privacy, data protection, and use of consumer data)…’.¹⁰⁰

It is questionable whether incorporating PDP/PP issues under the category of product quality and addressing them within the economics-oriented analysis framework provides a satisfactory or correct solution. The PDP/PP was never considered as product quality and was even explicitly excluded from competition law concerns. It is still not clear what changed the view. Did the understanding of product quality undergo a transformation? Or, was there another factor driving competition law to address PDP/PP issues, with product quality serving as a convenient avenue? Furthermore, if privacy is considered a facet of product quality, should other noneconomic issues be similarly regarded and subject to competition law? Actually, the suggestion of incorporating more noneconomic issues into competition law has been advocated by the U.S. DOJ Assistant Attorney General Makan Delrahim. He contended that, beyond privacy, aspects like free speech and openness could also be dimensions of quality in the digital realm.¹⁰¹ However, this raises the crucial query of where the boundary of competition law lies, if there is one.

Actually, where economic power interplays with noneconomic power, a competition law decision may have far-reaching (noneconomic) influence. The influence may go beyond the capture of the law’s current economics-based analysis framework. This concern is substantiated by instances such as Facebook’s influence on democracy. Approved by competition law, Facebook had a considerable part of its power built up and maintained through mergers and acquisitions, leading to a significant concentration of personal data.¹⁰² The Facebook-Cambridge Analytica scandal revealed the misuse of Facebook’s user data to predict voters’ behaviour and design targeted political advertisements to influence the voters, affecting the 2016 U.S. presidential elections.¹⁰³

The notion that other laws would address the harms may not materialize, as these laws might not exist or even might not be able to come into existence. An illustrative example is the Australian parliament’s early 2021 proposal to mandate online platforms like Facebook and Google to compensate news publishers for content.¹⁰⁴ Upon the announcement of this legislative intent, Facebook responded by blocking all news content in Australia, effectively halting the legislative initiative.¹⁰⁵ The intention is not to solely attribute Facebook’s actions to competition law, but rather to acknowledge that competition law decisions can affect and have affected not only Facebook’s economic power but also its noneconomic power.

To sum up, the gap between the current conceptions and reality, exemplified by the PDP/PP issues, suggests that something that was not readily observable in the past has become observable now, or that something novel has taken place, or both. In seeking a solution, if all noneconomic issues were to be considered as part of product quality in competition law, the concept of quality would extend beyond economic utility. Consequently, the comprehension of potential harms and their mitigation should still go beyond economic analysis.

3. Progress in Understanding Power and the Need for Further Examination

Based on the conventional power conceptions—market power, dominance, and monopoly—progress in competitive theories has been made to better understand power. It is essential to clarify from the outset that the objective is not merely to enumerate significant theories, but rather, by delineating their main propositions, to highlight the lacunae in these theories and underscore the necessity for a more cohesive understanding of power. The advancements can be classified into three categories: (1) general developments applicable to the digital sector, (2) developments specifically targeting the digital sector, and (3) recent reflections on the market power conception.

a. General developments applicable to the digital sector

For this category, notable developments encompass dynamic competition and innovation in the high-tech sector, direct and indirect network effects, and attention rivalry. Among these, indirect network effects are also intertwined with another extensively discussed theory, the two-sided market theory.

The dynamic competition theory has sparked debates over whether monopoly or competition serves as a better catalyst for innovation. The founder of the theory, Schumpeter, and his followers argued that, in high-tech markets, a monopoly might be more inclined to bear the risks and costs of innovation, and that monopoly power in high-tech markets was likely short-lived, as competing new products might swiftly capture the market through a ‘gale of creative destruction’.¹⁰⁶

However, different views emerged as more evidence accumulated, indicating that monopoly power in the high-tech sector was not always contestable.¹⁰⁷ It was also observed that competition, rather than monopoly, encouraged neck-and-neck firms to innovate.¹⁰⁸ At least, the innovation potential should be examined case-by-case.¹⁰⁹ Regarding competitive assessment tools, there were proposals to shift from identifying substitutes to other aspects, such as examining the abilities and incentives to innovate,¹¹⁰ paying more attention to the key resources for innovation (such as human intelligence rather than capital),¹¹¹ and exploring the feasibility of defining an innovation market.¹¹²

The network effects theories are related to barriers to entry, encompassing both direct network effects and indirect network effects.¹¹³ Direct network effects occur when the utility of a product for consumers on one side depends on the number of other users on the same side, whereas indirect network effects occur when both sides bilaterally enjoy additional surplus as the number of consumers on the other side increases.¹¹⁴ A two-sided market represents a market structure leading to indirect network effects.¹¹⁵ It was asserted that strong network effects, coupled with other economic characteristics like high switching costs as well as economies of scale and scope, could result in high entry barriers in the digital sector.¹¹⁶ For competitive assessment tools, although the importance of considering both sides has been acknowledged, there are different opinions on how to approach this, such as the debate on whether to define a single market or multiple markets for competitive assessment.¹¹⁷

The attention rivalry theory revolves around theories of harm and power measurement associated with the ‘free/low price plus advertising’ business model (hereinafter referred together as the ‘free price plus advertising’ model). The core proposition is that attention rivals attract user attention by offering products and services, subsequently selling this attention to those who value it.¹¹⁸ This business model is prevalently adopted by online services and can also be observed in traditional sectors such as newspapers. It was pointed out that the attention rivals imposed competitive constraints on each other, including services that belong to different product markets, so it was crucial for competitive analysis to focus more on the differences in the attention produced and sold, rather than solely adhering to the nature of the products and services attracting attention.¹¹⁹ Regarding competitive assessment tools, it was advised that market definition tools be revised for ‘attention markets’, and to define relevant markets based on the ‘time spent’ and use the substitution concept to delineate relevant markets for further competitive analysis.¹²⁰

These developments offer valuable insights into the distinctive competitive characteristics of the high-tech sector, including the digital sector. It is also noteworthy that although the term ‘technological power’ was not explicitly used, there is a recognition of the importance of better understanding and measuring technological power, as reflected in proposals to consider abilities and incentives to innovate and to take into account factors such as human (intelligence), rather than focusing solely on capital.

However, they are not fully effective in overcoming the path dependence in the competitive analysis described in Shortcoming I, the novel cross-market competitive influence. In fact, the proposed solutions in these theories exhibit a clear path dependence, evident in the reliance on market definition, such as the propositions to define relevant markets for innovation, two-sided markets (either defining the two markets separately or as a whole), and attention. Furthermore, they do not address Shortcoming II at all—the oversight of noneconomic aspects—therefore, they cannot contribute to overcoming it.

b. Developments specifically targeting the digital sector

Shifting focus from general power theories applicable to the digital sector to more sector-specific power discussions within the digital realm, four notable aspects emerge: (1) the relationship between data and power, (2) the interplay between business ecosystems and power, (3) the relationship between innovation in the digital sector and power, and (4) the calls for considering noneconomic dimensions when evaluating the power of tech giants. It is important to note that the third aspect is distinct from the broader developments in general innovation theories.

To begin with, the discussions on data and market power revolve around two key aspects: first, whether data are pertinent to market power, and second, if so, how to accurately assess its influence. In the early stages of discourse, certain viewpoints asserted that data could not increase or entrench market power and would not serve as an entry barrier due to its ease of collection and nonexclusive and nonrivalrous nature.¹²¹ However, as practices have evolved, there is a growing consensus that data can indeed serve as a source of market power.¹²² In measuring the competitive influence of data, debates arose regarding whether to define a putative relevant market for data when it is solely used as an input and not traded as a commodity in the market.¹²³ An alternative perspective proposed is to examine data’s relevance to market power in the digital realm through the concept of ‘tracking power’—the collection of users’ personal data and tracking users using that data.¹²⁴ Critics argued that traditional tools based on revenues or sales market shares are inadequate in capturing tracking power, and instead suggest measuring market power based on ‘the reach of a tracker on popular websites and apps’.¹²⁵

In addition to data, significant attention has been devoted to understanding the relationship between business ecosystems and power. There are two main approaches to analyse this relationship. One approach emphasizes the gatekeeping power—the dominant power of a digital platform derived from its pivotal role in a business ecosystem, particularly in providing a key online service that facilitates matching between a user and other services, other users, or (informational) contents.¹²⁶ The attention on gatekeeping power arose from the fact that certain powerful ‘digital gatekeepers’, through their online services, could control the access between third-party services and users.¹²⁷ Scholars have also put forward conceptions related to gatekeeping power, such as gatekeepers, economic dependence, bottlenecks, digital gatekeepers, unavoidable trading partners, and economic dependence.¹²⁸ A notable example of gatekeeping power is the EU’s legislation on gatekeepers outlined in its Digital Markets Act (DMA), complementary to competition law, aimed at addressing competitive issues in the digital realm.¹²⁹ On a global scale, competition authorities have proposed several new concepts concerning gatekeeping power. Representative examples include the United States’ superior bargaining power, the EU’s power of gatekeepers, the United Kingdom’s strategic market status, and Germany’s paramount cross-market significance.¹³⁰ The other approach focuses on comprehensive business ecosystem power—the integrated power stemming from an undertaking’s ownership of diverse online services crossing various relevant markets.¹³¹ It is argued that, facilitated by digitalization, different digital services could have more sophisticated networks of interdependencies and distinct competitive dynamics, and that traditional competition law tools may be inadequate in properly evaluating the competitive influence of cross-market power dynamics.¹³² To address the influence, proposals have been made to conduct market definition and integrated market power evaluations specifically tailored for digital ecosystems.¹³³

Gatekeeping power and ecosystem power are closely related yet different. The two are related because, given that a business ecosystem typically revolves around a central service, the ecosystem’s owner often assumes the role of gatekeeper.¹³⁴ However, the perspectives are different, as the gatekeeping power focuses on the power associated with a core gatekeeping product, whereas the business ecosystem power emphasizes the collective power of a business ecosystem.

Regarding the relationship between innovation in the digital sector and power, a noteworthy power theory is the moligopoly theory. This theory proposes a more comprehensive consideration of innovation’s influence when analysing the power and competitive effects within the digital sector, particularly concerning tech giants.¹³⁵ The Moligopoly theory emphasizes competitive pressure, asserting that such pressure in the digital sector may arise from innovative startups and in-house competition, encompassing a broader spectrum than what conventional theories recognize.¹³⁶ It was suggested that methods for evaluating market power be enhanced by adding the ‘assessing tipping from pressure’ method and utilizing R&D expenditures, among others, as a key indicator.¹³⁷ Additionally, it is worth noting that killer acquisitions fall under the newly proposed test only if a target is shut down,¹³⁸ excluding mergers like Facebook/Instagram and Facebook/WhatsApp from its scope.¹³⁹

All the aforementioned power theories for the digital sector offer diverse perspectives to enhance the understanding of power dynamics within the sector, particularly concerning tech giants. These theories have the potential to mitigate certain path dependencies, such as transitioning from defining relevant markets based on the substitutability of final products to considering business ecosystems systematically.

However, taking a comprehensive perspective, there remains space for enhancement. Economic theories exhibit overlaps and leave gaps. The overlaps stem from the recognition that data holds significant importance within a business ecosystem, and the ecosystem owner often assumes a gatekeeping role.¹⁴⁰ Consequently, theories centred on data and gatekeeping may intersect with those pertaining to the broader business ecosystem. Regarding gaps, firstly, the proposed theories fail to address the path dependency issues outlined in Shortcoming I. None of the data, business ecosystem, or moligopoly theories adequately address the competitive dynamics like that between WeChat and Renren, as highlighted earlier. Secondly, these theories do not adequately tackle Shortcoming II, which pertains to noneconomic power and related concerns.

c. New theories on power conception

In recent years, noting the absence of a clear legal conception of economic power within competition law, novel economic power concepts and metrics have been proposed.¹⁴¹ It has been argued that the traditional market power conception, aimed at quantifying the extent of horizontal competition, shows inadequacies when applied to zero-price markets and multisided platforms.¹⁴² Additionally, it has been observed that without a proper theoretical foundation, the various newly proposed concepts aimed at addressing the evolving economic realities in the digital sector may result in conceptual inconsistencies and legal uncertainties.¹⁴³

In an effort to establish a comprehensive theoretical framework for a multidimensional conception of power, insights have been gleaned from noneconomic literature. Several new categories of power have been identified,¹⁴⁴ as outlined in Table 1.

Economic metrics were further developed for two power conceptions—the social exchange theory, which posits dependency as its source of power, and the panopticon power, which derives its power from unequal information gathering—as other concepts are deemed either too broad for translation into a metric or too contextual.¹⁴⁵

The alternative economic power conceptions offer a stronger theoretical and conceptual foundation for capturing power derived from the economic value chain and business ecosystem. The introduction of two new metrics may facilitate a more accurate measurement of a business’s economic power. More importantly, these propositions demonstrate how different conceptions of power can lead to new insights into understanding emerging competition issues and developing innovative tools to address them.

While the contributions of these conceptions in competition law have been acknowledged, several aspects warrant further exploration. Initially, the conceptions were introduced to enhance understanding of power dynamics within business ecosystems. Subsequent applications of these conceptions have predominantly focused on power measurement rather than elucidating potential harms. Furthermore, although factors such as innovation and technological path dependence have been assessed, the emphasis has primarily been on their relationship with industry architectures.¹⁴⁶ For instance, considerations revolve around whether to adopt a system of open innovation or a walled garden, and how processes like first-mover advantage may lead to a lock-in effect and technological dependence.¹⁴⁷ These observations suggest that the conceptions have limited applicability to deal with other path dependence issues in Shortcoming I, such as assessing the competitive circumstances between WhatsApp and Facebook. Moreover, the focus on economic power without addressing noneconomic power implies that these conceptions cannot adequately deal with Shortcoming II. These gaps warrant further research.

In conclusion, there is a pressing need for further examination of the power conception for competition law. A well-defined conception not only facilitates the articulation of crucial elements that would otherwise remain overlooked, but also helps to prevent unnecessary disputes arising from differing interpretations of the same term across various contexts.¹⁴⁸ The introduction of newly proposed power conceptions underscores that understanding power is not solely a conceptual endeavour—it significantly influences the articulation of harms and the design of competition law tools. However, the current power conceptions do not provide precise answers regarding the true nature of business power. Consequently, competition law may struggle to address emerging competitive issues promptly, coherently, and effectively. Moreover, given the overlaps and gaps among existing power theories, there is a clear need for an overarching theory. To achieve a deeper understanding of power for competition law, it is imperative to transcend economic perspectives and conduct a more fundamental examination of the conception of power. Acknowledging this, the forthcoming examination will approach the power conception from a philosophical standpoint, complemented by empirical observations.

III. A RECONCEPTION OF POWER

This section aims to examine and provide a reconception of power through the lens of legal philosophy. It is worth mentioning that legal philosophy has been employed to underscore the importance of maintaining competitive process within the market.¹⁴⁹ Revolving around a different pillar, the following analysis delves into the nature of power in the market.

A. Power as Capacity

Amidst the debates in philosophical discourse regarding how to define power, the most prevailing understanding is ‘A in some way affects B’.¹⁵⁰ The affect is construed to signify ‘to have an effect on, either materially or otherwise’.¹⁵¹ It is further expounded that ‘[i]f something affects a person or thing, it influences them or causes them to change in some way’ (emphasis added).¹⁵² According to this definition, power is independent from its source and vehicle (agent) and from what is influenced by power (patient). This aligns with the prevailing understanding that an agent, such as an undertaking, can wield power, whereas a patient, such as a consumer, can be subject to its influence.

Philosophers have drawn the distinction and clarified that power is a capacity rather than an event (the exercise of power).¹⁵³ The adage in competition law, stating that holding market power is not inherently problematic, but rather the misuse (exercise) of it matters, aligns with this distinction between power and its exercise.¹⁵⁴ Even the narrowest definition of power in competition law—the ability to raise prices above competitive levels—also suggests that power is a capacity distinct from its exercise.

That being said,  capacity cannot be directly observed but needs to be perceived from its exercise. For instance, one understands that a person wielding an axe has the capacity to chop wood based on our accumulated knowledge of similar past occurrences. If such an action never occurs, which means the power is never exercised, the power still exists, but its existence may remain unknown. Therefore, to comprehensively examine power in social context, including the business context, it is necessary to move beyond analysing power solely as a latent capacity and to consider its actual exercise.

B. Forms of Power Exercise: Will, Action, and Evidence

Philosophers have attempted to distinguish different forms of power. An early categorization of power forms was made by Russell, who held that an individual might be influenced by (1) direct physical power over his body, (2) inducements such as rewards and punishments, and (3) influence on opinion, encompassing propaganda in its broadest sense, including the cultivation of desired habits in others.¹⁵⁶ A more detailed categorization was made by Wrong, as shown in Figure 1, which identified four forms of power: (1) Force, including physical force (both violent and non-violent) and psychic force; (2) Manipulation; (3) Persuasion; and (4) Authority.¹⁵⁷

These attempts are valuable for a better understanding of power. However, for the purposes of this research, they are not exhaustive, lack sufficient detail, and can be confusing. For instance, within the power tree provided by Wrong, Authority overlaps with the other three forms of power—Force, Manipulation, and Persuasion. Wrong himself acknowledged that one subcategory, the Coercive Authority, represents a threat of using force.¹⁵⁸ This makes Coercive Authority fit in the Force category. Consequently, Coercive Authority could also be considered a subset of Force.

An alternative categorization of power forms can be made based on will and action, which are fundamental factors in the examination of power exercise. Action is the external manifestation of power.¹⁵⁹  Will, being abstract and internal, pertains to the inner state of mind.¹⁶⁰ In the context of businesses, the will is the internal decision of an undertaking. A subject’s will regarding a matter can assume different statuses—having no will, being in favour of it, or being against it. In cases where there is no will, upon its formation, it immediately transitions into a ‘for or not for’ status. It is worth pointing out that apart from having no will, a will is unlikely to be purely ‘neutral’, particularly when action or inaction is required. For example, if an individual is ‘indifferent’ about voting and chooses not to vote, the will regarding voting can be categorized as ‘not for’.

For a given agent, the rational aim of power exercise is to affect its patient in a manner that maximizes the agent’s will. Following this, the circumstances can be further divided into the following categories. When the patient’s will (and consequently the actions, if any) aligns with an agent’s will, the agent’s power can be discerned by observing how the agent can reinforce the patient’s will (Elevation). When the wills of an agent and its patient are divergent, the agent’s power can be observed by its ability to compel the patient to conform to the agent’s will (Conversion). Conversion can be achieved through two ways. The first entails influencing the patient’s action against the patient’s will (Conversion-Act).¹⁶¹ Conversion-Act can be achieved through violence or nonviolent means like the limitation of action choices. Violent means fall outside the scope of this research due to their distant relationship with business activities and competition law. The second approach involves altering the patient’s will, consequently influencing their actions (Conversion-Will).¹⁶²

Recall that the most common philosophical understanding of power is ‘A in some way affects B’. Taking capacity, will, and action into consideration, a more precise definition of power in the social context is A has the capacity to affect B’s will or action according to A’s will. This conception covers the established understanding of power as a social phenomenon. Taking a few influential views as examples, Russell held that power ‘may be defined as the production of intended effects’; Max Weber regarded power as ‘the probability that one actor in a social relationship will be in a position to carry out his own will despite resistance, regardless of the basis on which this probability rests’; Robert Dahl saw power as a relationship among people and that ‘A has power over B to the extent that he can get B to do something that B would not otherwise do’; Dennis Wrong viewed power as ‘the capacity of some persons to produce intended and for[e]seen effects on others’.¹⁶³

Logically speaking, there is another scenario when a patient has no will, and consequently takes no action, regarding a particular matter, and an agent may affect the patient to form a will from nonexistence to existence. However, once the will is formed, it falls directly into the three aforementioned categories. For example, consider an individual who initially lacks the will, and therefore takes no action, to participate in an election. A news push about the election may prompt the generation of a will related to voting. However, whether the individual’s will is to vote or abstain, and whom to vote for, can be sorted into the three categories discussed above. The mere generation of a patient’s will without aligning it with the agent’s will falls beyond the scope of competition law and the interest of this research.

Evidence plays a pivotal role in the observation of power. Although evidence helps to identify power, it is not power itself. Where there is a change in how a type of power is exercised or how different power types interact with each other, the evidence for observing and evaluating power changes correspondingly. For example, price is important evidence for observing and evaluating power where monetary price is involved. This aligns with the conventional exercise of economic power through the imposition of pecuniary charges. However, within the digital sector, numerous services are rendered free of monetary charge. This paradigm shift indicates changes in the realization of economic power. Consequently, the evidence for power observation should be changed accordingly. Using search volume instead of revenue as a market share evaluation parameter is an example of this.¹⁶⁴

Evidence is also relevant for observing the influence or potential influence of power exercise (hereinafter ‘influence of power exercise’). The relevance rests both in screening the influenced scope and analysing the influence. Changes in the exercise of a type of power and the interaction of different power types may influence what should be scrutinized and how to assess the influence properly. Failing to consider this could hinder competition law’s ability to identify and address emerging issues effectively. A more detailed analysis will be presented later.¹⁶⁵ Before that, an examination of power types is necessary to provide a sounder analytical foundation.

C. Types of Power: A Further Anatomy

Instead of packing everything up under market power, further categorization of different power types can help to show the characteristics, interactions, and dynamics of the power of businesses. It is important to clarify from the outset that the aim here is not to compile an exhaustive list of power types but to highlight distinct categories and their interactions, particularly based on the evidence of the digital realm. The purpose is to demonstrate that economic power is not and should not be viewed as the sole or standalone type of power. Instead, economic power both influences and is influenced by other power types.

It is also well noticed that measuring non-price-related power would pose significant challenges. The intention below is not to design parameters for power measurement but rather to argue that competition law should not dismiss the consideration of other types of power. This is crucial because an exclusive focus on economic parameters in competition law may permit the emergence of business power beyond the control of other legal frameworks, as exemplified by the case where Facebook influenced the Australian government to withdraw a proposed legislation.¹⁶⁶ A summary outlining the prominent power types and their interrelationships will be provided at the end of the categorization.¹⁶⁷

1. Economic Power (P_Economic)

a. A reconception of economic power

An economic transaction between a business and its customer,¹⁶⁸ either an individual consumer or a business customer, can be abstracted as an exchange of economic interests, as shown in Figure 2.

Economic Interest 1 is an interest related to a product or service provided by a business and received by its customer. Economic Interest 2 is often pecuniary interest, money. However, it can also be other economic interests, for example, before money was used, the bartered goods,¹⁶⁹ and in the digital era, personal data.¹⁷⁰ Economic Interest 2 may also be a package of different interests. For example, a user may both pay money and allow the user’s personal data to be collected in exchange for using a product.¹⁷¹

Economic interest is at the core of P_Economic. A rational market participant would aim to give out less and gain more economic interest from the other party in an economic transaction. As a result, the rational wills and actions of the trading parties inherently go against each other. Consequently, the rational will and action of one side are to affect the other party to give out more or gain less economic interest. This can be achieved either in the form of Conversion-Act or Conversion-Will. Based on these and the basic conception that power exists when A has the capacity to affect B’s will or action according to A’s will,¹⁷² the P_Economic  of a business is the business’s capacity to affect the will or action of the other side to give out more or to gain less economic interests.

This definition prompts questions about its relevance to competition law and how it aligns with established competition law theories. Such questions may arise because, according to this definition, both sides of a transaction hold a certain degree of P_Economic. It is important to note that competition law’s understanding of power should be based on, rather than be the basis of, what the true nature of power is. Bearing this in mind, competition law becomes concerned when there is an unacceptable significant imbalance in P_Economic.

The unacceptable significant imbalance has two elements. One element is the unacceptableness that connects with competition law’s theories of harm, which is related to exploitative abuses under the business–consumer relationship. It is important to note that the focus here is on the business–consumer relationship. The business–competitor relationship will be discussed later. The other element, the significant imbalance, pertains to the degrees of power held by different businesses, such as in scenarios of dominance or monopoly.¹⁷³ The second element might lead to another question: significant power imbalance compared with what? At this point, the wisdom from conventional competition theories, especially economic theories, can be applied, as the baseline could be the economic interests that would be transacted in a competitive market.

Besides the business–consumer relationship is the power relationship between a business and its competitor(s). Before applying the definition to the business–competitor relationship, it is important to clarify one point. Referring to Figure 2, when saying a business has more power than its customer in the business–customer relationship, it is to say that the business has a greater capacity to gain more from a customer or give out less economic interest to the customer concerning Product A. However, in the business–competitor relationship, when saying a business holds more power than a competitor concerning Product A, it means that the business holds a higher degree of P_Economic compared with that of its competitor.

A more complex circumstance is depicted in Figure 3. It shows that a business may, through another product it holds (Product B), have the capacity to affect its competitors’ P_Economic concerning Product A.

Although appearing unique, this situation remains within the framework of the proposed P_Economic concept. The key is to understand different power relationships. Under this circumstance, there is a P_Economic over P_Economic. The first is the P_Economic held by the Business through Product B in the business–competitor relationship. The second is the P_Economic concerning Product A held by the Competitor in its business–customer relationship. Unlike the power degree comparison in the business–customer relationship discussed above, the P_Economic over P_Economic is about an influence from a business on its competitor’s capacity to gain economic interests from the competitor’s customers. For example, when a search engine excludes a webpage link in its search results or an app store excludes an app, the former influences the latter’s capacity to reach the latter’s consumers and, consequently, the latter’s capacity to gain economic interests. The related theories in competition law are exclusionary conducts.

b. A comparison of different conceptions

The proposed P_Economic conception covers and extends beyond the conventional power conceptions and has advantages. The price-based market power conception is covered in that price is a prominent economic interest—the ability to raise prices above competitive levels is the capacity of a business to affect consumers to give out more money than in a competitive market. Compared with the price-based marker power conception that relates economic interests simply to pecuniary interest, the proposed conception provides more flexibility and perspectives for observing, understanding, and evaluating the power and competitive issues in businesses, especially when a new business model emerges. For example, under the ‘free plus advertising’ business model, the price-based power conception struggles to recognize and evaluate the power and competitive issues related to the ‘free’ services, as shown by the doubt on whether online services that were free of monetary charge ‘[could] even be considered a market’ back in the 1990s.¹⁷⁴ If following the same logic, it should also be doubtful whether general search services like Google Search, which is also provided to users free of monetary charge, should also be considered a market and receive power and competitive analysis. However, subsequent case law acknowledged that power could exist in ‘free’ services, like Google’s general search services.¹⁷⁵ The proposed P_Economic can help to avoid this logical inconsistency. Moreover, the measurement of power is not confined to price-based parameters such as revenue but also includes parameters related to other economic interests, such as search volume.¹⁷⁶

Compared with the price-based power conception, the economic strength–based conception aligns more closely with the proposed P_Economic conception. This is because where there is a change in a business’s capacity to affect the will or action of its customer to give out more or gain less economic interest, there is typically also a change in the comparative power between the business and its competitor in the market. When both power and its exercise are at a distinguishing level, it may show the competitive image of preventing or excluding effective competition.

However, there are still differences between the two. The proposed P_Economic conception may help to shift the mindset concerning how competition is carried out and to deal with new challenges better. For example, with economic strength–based conception stresses a business’s ability to influence the competition being maintained in a relevant market,¹⁷⁷ when new challenges arise, it is likely to direct attention to improving relevant market definition methods, such as the propositions for defining putative relevant markets for data,¹⁷⁸ innovation,¹⁷⁹ attention,¹⁸⁰ and so on. However, the core solution may not lie in the market definitions. For example, it has been argued that defining a putative relevant market for data would not sufficiently capture the influence of data.¹⁸¹ What is needed is a better identification of which market(s) may be relevant, which further requires a better understanding of (1) how different power types are related to each other in new ways concerning a product, and (2) how the power interplay across different products has changed.¹⁸² Although it is not absolutely impossible to develop new theories and tools under the economic strength–based conception to better capture the influence of data, the proposed P_Economic conception can do a better job. A detailed example will be provided later.¹⁸³ Before that, based on the proposed P_Economic conception, the next subsection will give a closer look at P_Economic from an empirical perspective.

c. An empirical observation of economic power in the digital sector

For any for-profit business, a core will is to gain more economic interest in economic transactions. Generally speaking, there are two important stages related to achieving this goal—product fabrication, which involves innovation and production, and product distribution.

To begin with, concerning product fabrication, businesses rely on their products and services to participate in transactions. The quality of products and services has always been an important factor in competition.¹⁸⁴ However, different consumers may have different opinions on which product is better.¹⁸⁵ For conventional businesses, discerning what constitutes a superior product from the perspective of individual customers poses a significant challenge.¹⁸⁶ Even if such insights were attainable, personalizing the products would be arduous or too costly. Consequently, the prevailing approach often revolves around enhancing overall quality with some degree of product segmentation, assisted by market research.¹⁸⁷

In the digital realm, the use of personal data makes massive personalized production possible, a phenomenon unprecedented in previous eras. Online service providers have the capacity to acquire more personal preference information in the form of digital data on a large scale and scope, in more detail, and at a higher speed than ever before.¹⁸⁸ Consequently, competition no longer revolves solely around offering products of superior overall quality, but delivering better and more relevant products on a personal level. According to surveys, nearly all customers prefer to shop with companies that could recognize and remember their preferences, and provide them with personalized offers.¹⁸⁹ For example, TikTok gains great success by providing personalized short videos to its customers.¹⁹⁰ The mechanism of TikTok is to send users personalized video content based on user data, which includes not only basic information such as age and gender but also dynamic data such as searching data, thumb-up data, comment data, and contact data such as friends, families, and colleagues.¹⁹¹ Customisations like these are prevalent in the digital realm. Actually, online service providers try to provide users with personalized answers even before questions are asked. For instance, a scholar shared his experience of being alerted to head off for a meeting earlier because of the heavy traffic although he did not set such an alert.¹⁹² Service providers like Google appear to be very proud of such ‘answer before being asked’ capacity.¹⁹³

Among massively personalized products, targeted online advertising is particularly worth noting. As a part of the prevalent ‘free plus online advertising’ business model in the digital realm, online advertising is on the service list of virtually all tech giants.¹⁹⁴ The advertising industry predates the advent of computers and the Internet and even personalized advertising itself is not novel in the predigital era. A local shopkeeper, for instance, would advertise to his customers according to his knowledge of their preferences.¹⁹⁵ However, the scope of such personalized targeting was significantly limited compared with the capabilities afforded by digital technologies.

Then, massive advertising emerged, but it was not until the late twentieth century that personalized marketing gained prominence.¹⁹⁶ This shift was driven by the advent of online advertising, facilitated by the extensive processing of digital personal data made possible by advancements in computer and Internet technology.¹⁹⁷ Over the past decades, online advertising has evolved into two main forms. The first is search advertising, which targets consumers based on their search queries and provides personalized search results based on the search terms entered by a user and other data the business holds about the user.¹⁹⁸ The second is display advertising (nonsearch advertising), such as ads in banners in plain text, audio, and video forms.¹⁹⁹ Nonsearch advertising, used to only display indifferent content, has evolved over time into targeted display advertising based on user data such as browsing history.²⁰⁰

Apart from product fabrication, another important aspect is product distribution. Ensuring a product reaches its intended consumers is pivotal in gaining economic interests. Distribution encompasses either direct exchange with consumers or intermediated distribution channels.²⁰¹ The latter is prevalent in the digital realm—the use of digital platforms such as online marketplaces and app stores. In cases where a product is distributed through an intermediary, the P_Economic related to distribution (P_{Economic(Distribution)}) is involved. It is a power concerning whether and how to distribute a product that wants to use the platform.²⁰² Several key aspects of this power merit attention.

First, the exercise of P_{Economic(Distribution)} becomes evident when an intermediary has power over whether to distribute a product. A representative example is Apple’s App Store. In March 2019, the EC received complaints from Spotify, a competitor of Apple’s music streaming service, regarding two rules of App Store license agreements: (1) companies distributing their products through the App Store were obligated to use Apple’s in-app purchase system, which charged developers a 30-percent commission on subscription fees; and (2) developers were restricted from informing users of alternative purchase methods, potentially offering cheaper alternatives.²⁰³ Similar concerns were raised in another complaint concerning Apple’s e-book and audiobook distribution in March 2020. Notably, iPhone and iPad users were constrained to download native non-web-based apps exclusively from the App Store.²⁰⁴ Refusal to comply with these terms means a business would lose access to Apple’s App Store and the users that use Apple’s operational system.

Second, the exercise of P_{Economic(Distribution)} manifests when an intermediary that affects how a product is distributed. In product distribution, it is crucial not only to have access to a platform but also to effectively disseminate product information to reach users better. A better reach can be achieved through preferred access arrangements, such as preinstallation. For instance, in the Google Android case investigated by the EC, it was found that in 2016, 95 percent of search queries were made through Google Search on Android devices where Google Search and Chrome were preinstalled. However, on Windows mobile devices where the Bing search engine was preinstalled but Google Search and Chrome were not, less than 25 percent of search queries were made through Google Search, whereas more than 75 percent were conducted on Microsoft’s Bing.²⁰⁵ This significant disparity was not attributable to differences in service quality but rather to which app was preinstalled and set as the default. Although consumers were not restricted from downloading alternative search engines, the lack of motivation to do so was evident, given the additional time and effort required.

When numerous products and services are given initial access to an intermediary, a better reach also concerns how product information is organized and presented to users, as the order of information can significantly influence users’ choices. For example, in the Google Search (Shopping) case, it was found that Google listed and showcased its own comparison shopping services in a higher position with richer features than those of its rivals within its general search results.²⁰⁶ Users tended to click on more visible links with distinguishing features presented in general search results.²⁰⁷ Consequently, Google’s actions resulted in the foreclosure of competing comparison shopping services and reduced consumers’ ability to access the most relevant comparison shopping services.²⁰⁸

2. Political Power (P_Political)

The types of power that businesses pursue and hold may extend beyond economic power.²⁰⁹ This sub-section examines the P_Political held by businesses. Politics constitutes a fundamental dimension of public life, intersecting with culture and social progress, and exerting influence on the daily lives of citizens.²¹⁰  P_Political is a power related to politics and can be wielded by businesses. Political interest is at the core of P_Political. Based on the basic conception that power exists when A has the capacity to affect B’s will or action according to A’s will,²¹¹ the P_Political  of a business is the business’s capacity to affect the will or action of others in a way that can better achieve the business’s political will.

The fact that businesses may have political influence is no news. P_Political of a business can be obtained in different ways. One way is not related to the nature of products. For example, in the late nineteenth and early twentieth centuries, when Standard Oil was at the peak of its P_Economic, the company was reported and criticized for significantly influencing political affairs through partisan participation and important political position appointments.²¹² Although its P_Economic was an important basis for obtaining its powerful P_Political, its P_Political was not exercised directly through its oil products. Subsequent changes in businesses’ political power have been witnessed in the era of globalization, primarily through the expansion of campaigning, financing political parties, being more active in rule-setting such as through providing rating services, and the taking over of political power in the privatization process.²¹³ Conversely, P_Political can influence the P_Economic of a business. For example, governmental decisions to offer state aid can bolster the P_Economic of the subsidized businesses.²¹⁴

Another way to acquire P_Political is related to the nature of products, especially those with information dissemination functions. In this context, P_Political closely interplays with P_Economic and informational power (P_{Informational}).²¹⁵ The media sector is a representative example. Media like newspapers, magazines, radios, and televisions are pivotal platforms upon which P_Political is based and exercised. The capacity lies in their nature—passing information to the public and influencing the latter’s belief, an element of will. As noted by John Kenneth Galbraith, the principal instrument through which media power is enforced is belief, namely a dimension of the influence on the mind.²¹⁶ Rod Sims, former chairman of the Australian Competition and Consumer Commission, also underscored that ‘[a]ll governments are responsive to media in some way or another, because in all countries media is the filter through which things are seen’.²¹⁷

Instead of just making offline information available online, the P_Political in the digital sector is characterized by new players, new roles, and new influences. Many traditional media outlets have developed online websites to make their news content available online. However, the exchange of information increasingly relies on online platforms capable of disseminating information (hereinafter ‘online information platforms’), such as online search and social networking services. For example, Facebook’s social networking services show news from traditional media outlets such as BBC News, The Guardian, and Sky News. Apart from serving as media of traditional media, online information platforms also play a pivotal role in disseminating political campaigns.²¹⁸ Moreover, online information platforms serve as direct channels for transmitting voices directly. For example, former U.S. President Donald Trump famously made use of Twitter to engage with the general public.²¹⁹ Consequently, online information platforms have become crucial intermediaries for disseminating political information.²²⁰ Users heavily rely on information rankings provided by online information platforms, with research indicating that users typically prioritize links featured on the first few pages of search engine results, or even just the first few links on the initial page.²²¹

It is worth reiterating that the demonstration of such power interplay does not necessitate that all issues pertaining to P_Political should fall under the purview of competition law. However, it does underscore the importance of acknowledging that simply excluding all P_Political considerations from competition law would be inappropriate. This is because competition law decisions could potentially create a powerful undertaking whose P_Political influence exceeds the bounds of legal control, as exemplified by Facebook successfully pressured the Australian government to drop the legislation that would harm Facebook’s interests.²²²

3. Informational Power (P_{Informational})

P_{Informational} is another prominent power type wielded by businesses. Informational interest is at the core of the P_{Informational}. Based on the basic conception that power exists when A has the capacity to affect B’s will or action according to A’s will,²²³ the P_{Informational}  of a business is the business’s capacity to affect the will or action of others concerning information in a way that is in line with the business’s informational will.

P_{Informational} encompasses two prominent aspects, information acquisition and information dissemination. On the one hand, it is important for a business to acquire information to gain insights into other market participants, including both competitors and customers. On the other hand, it is also crucial for a business to disseminate information to influence its consumers to achieve its will, not only economic will but also noneconomic will, such as political will.

a. Power concerning information acquisition (P_{Information(Acquisition)})

‘Ipsa scientia potestas est’ (knowledge itself is power). This saying of Sir Francis Bacon about the difference between knowing and further action was taken out of its theological context and became the well-known ‘knowledge is power’.²²⁴ Though the contexts between theology and economic activities are different, the distinction made between knowing and acting is nonetheless important.

Two subpower types of P_{Informational(Acquisition)} merit attention. One is the power to acquire business information (P_{Businessinfo(Acquisition)}). For example, Amazon may acquire nonpublic business information of other sellers, such as sales data, through its online platform Marketplace.²²⁵ It is reported that Amazon further used the information to calibrate the offers and strategies of its retail services, such as by tailoring offers to focus on the best-selling products.²²⁶ By doing so, Amazon mitigates the typical business risks associated with retail competition.²²⁷ In essence, its P_{Businessinfo(Acquisition)} based on its Marketplace was further transformed into the P_Economic of its retail services.

The other prominent subtype of P_{Informational(Acquisition)} in the digital sector is the acquisition of users’ personal information (P_{Personalinfo(Acquisition)}). Such information acquisition commonly occurs in the form of personal data. It can be reasonably inferred that, with other criteria being the same, it is a common will of users to better protect their personal information. Conversely, it is in the interest and will of a business to acquire personal data in a greater amount, in a wider scope, in more detail, and at a higher speed.²²⁸ In practice, businesses in the digital sector have leveraged technological advancements to augment their P_{Personalinfo(Acquisition)}. It was put forward as early as 1998 that the Internet’s interactive nature presented businesses with opportunities to glean more information about their customers.²²⁹

As the availability of digital products and online services increases, more individuals are conducting their economic and social activities in the digital realm, with their data being collected in greater detail. For example, online marketplaces have the capability to collect data on items browsed but not purchased, providing insights into consumer preferences and behaviours. Furthermore, advancements in sensor technologies have enabled the datafication and collection of individuals’ daily activities, ranging from their location to physiological metrics like heart rate.²³⁰ The information acquired may concern users’ privacy, such as undisclosed pregnancy.²³¹ Consequently, there has been a notable increase in digital businesses’ P_{Personalinfo(Acquisition)}, with a corresponding decrease in users’ capacity to withhold their personal information.

b. Power concerning information dissemination (P_{Information(Dissemination)})

The digital era is characterized by information explosion. With the widespread use of digital devices and the Internet, individuals have access to an expanding volume of information, encompassing not only economic information but also social and political information. Because the time and energy of an individual are limited, how to effectively disseminate information to individuals has become a critical issue for those seeking to advance their wills. Consequently, the P_{Informational(Dissemination)} assumes increasing significant importance.

Some services inherently hold P_{Informational(Dissemination)} due to their nature, with advertising services being particularly noteworthy in this regard. In the business realm, information is pivotal for achieving success. In this aspect, the P_{Informational(Dissemination)} is closely related to P_Economic. A prevalent and effective method to disseminate business information is online advertising. ‘Doing business without advertising is like winking at a girl in the dark. You know what you are doing, but nobody else does.’²³² This aphorism of Steuart Henderson Britt may be hyperbolic and contentious,²³³ but it nevertheless underscores the importance of advertising. Due to the importance, digital service providers often offer free services to attract user attention and use the attention to support their revenue-generating online advertising services.²³⁴ For example, Facebook’s initial public offering registration statement highlighted its vast advertising market opportunities and noted that most of its annual revenue was from third-party advertising, reaching as high as 98 percent.²³⁵ According to Facebook’s latest annual financial report, online advertising still constituted more than 97.8 percent of the company’s total revenue.²³⁶

In addition to online advertising services, online search services also inherently hold significant P_{Informational(Dissemination)}. P_{Informational(Dissemination)} closely interplays with P_Economic and P_Political. A notable example of transferring P_{Informational(Dissemination)} into P_Economic is Google’s general search services. When Google decided to promote its own comparison shopping services, it developed special algorithmic ‘guidelines’ to adjust search result rankings, among other measures.²³⁷ According to its policy, websites with duplicated or less original content were downgraded in rankings.²³⁸ Consequently, comparison shopping services, which often fell under these categories, were typically relegated to the second page or beyond of search results.²³⁹ However, Google exempted its own comparison shopping services from this policy and deliberately positioned them prominently on the first page.²⁴⁰ Empirical studies have demonstrated a strong positive correlation between a comparison shopping service’s ranking in search results and its traffic volume—the greater the visibility, the more traffic.²⁴¹ Evidence showed that Google’s conduct resulted in a sharp decline in the traffic of its competing comparison shopping services.²⁴² Meanwhile, there was a simultaneous increase in the click-through rates of Google’s own comparison shopping services.²⁴³

The services-based P_{Informational(Dissemination)} can also be transferred into P_Political. The Facebook-Cambridge Analytica data scandal is an example. Cambridge Analytica was a political consulting company that provided data-based campaigning services.²⁴⁴ In 2014, it had acquired the personal data of several hundred thousand Facebook users.²⁴⁵ The data were used to predict and influence voters’ choices in the 2016 U.S. elections by pushing personalized political advertisements.²⁴⁶

4. Technological Power (P_{Technological})

A foundational type of power underlying the aforementioned power types is P_{Technological}.²⁴⁷  P_{Technological} is of paramount importance in industries marked by high demand for innovation and technologies, such as pharmaceuticals and high-tech sectors, including the digital realm. In the digital sector, it is through the products and services with technologies that a business wields other types of power. Unlike other types of power, the patients that P_{Technological} points at are not consumers or rival businesses but a business’s own products and services. In this regard, it is a power to empower the business itself. Based on the basic conception that power exists when A has the capacity to affect B’s will or action according to A’s will,²⁴⁸ the P_{Technological}  of a business is the business’s capacity to affect its product in a way that can better achieve its technological will. The technological will depends on what a business wants a product or service to achieve—integrating certain features, enhancing personalization, collecting more user information, and so forth. The conception addresses both product quality and innovation concerns within conventional competition law. As will be elucidated later, when technological power exists or is exercised in new ways, the proposed conception provides greater adaptability and a more comprehensive framework for understanding the power, power exercise, and subsequent influence.²⁴⁹

Under conventional competition law, P_{Technological} is analysed together with P_Economic under market power. This is because P_{Technological} serves as a fundamental basis for P_Economic. In the digital sector, both production and distribution heavily rely on the exercise of P_{Technological}. For example, Apple develops and has technological control over its operating system and app store. It makes the Apple Store the exclusive source for iPhone and iPad users to download native non-web-based apps.²⁵⁰ This control, enabled by Apple’s P_{Technological}, allows Apple to mandate the App Store dependants to use its in-app purchase system and charge a 30-percent commission on all subscription fees.²⁵¹ This demonstrates the transformation from P_{Technological} to P_Economic to achieve further economic interests. Conversely, P_Economic can also influence P_{Technological}. For example, P_Economic may contribute to the internal build-up of P_{Technological} through investment in research and development. In addition to being developed internally, P_{Technological} can also be acquired through economic mergers and acquisitions. Facebook, for instance, has acquired many other products and services through mergers, including Instagram and WhatsApp.²⁵²

Although closely related, P_Economic and P_{Technological} are distinct in nature. Considering P_{Technological} as an independent power type enables a clearer observation of technological changes and their interplay with P_Economic and P_Political, especially when there is a novel change in how P_{Technological} interplays with the other power types. For an example of the interplay between P_{Technological} and P_Economic, general search services and search advertising services are usually delineated into different markets, as in the Google Search (Shopping) case, in which the EC did not consider online advertising when defining the relevant market for general search services.²⁵³ However, from a technological perspective, when a search is conducted, both general search results and search-based display online advertising appear in the same result list. Therefore, when a rival’s link is deliberately pushed down in default search results, the rival would need to purchase online advertising services to regain a higher rank. In this regard, due to the technological overlap between general search and online advertising, their economic power may also influence each other. For an example of the interplay between P_{Technological} and P_Political, in the Cambridge Analytica scandal, the personal data collected were used to predict and influence voters’ choices in the 2016 U.S. elections through pushing personalized political advertisements.²⁵⁴ In this regard, changes in how a business holds and exercises P_{Technological} can lead to changes in P_Political, including the degree of P_Political, the ways to exercise P_Political, and the scope and extent of the influence of the power exercise.

5. Summaries of prominent business power types and their interplay

Table 2 and Figure 4 provide summaries of the prominent power types and their interplay.

D. The Power of Competition Law

After presenting the prominent power types of businesses, the power of competition law is also worth revisiting. Understanding the nature of competition law’s power is important because competition law does not address all business power issues. The power of a law decides which problems it can address and how its tools can be developed and modified to tackle challenges, especially novel issues. These are particularly important to the digital realm, where whether a problem should be addressed by competition law is frequently questioned.²⁵⁵ To begin with, for the power of competition law, the patients are the businesses in the market, and the will is the law’s goal(s). Based on the basic conception that power exists when A has the capacity to affect B’s will or action according to A’s will,²⁵⁶  the power of competition law is competition law’s capacity to affect the will or action of businesses in a way that can better achieve competition law’s goal(s).

There are two prominent elements to note about the power of competition law. The first is capacity. Competition law’s capacity lies in both its legitimacy as a guide and in the legal tools it can employ. The former facilitates voluntary obedience (namely Elevation), and the latter can be used to convert a patient’s will or action in case of disobedience (namely Conversion-Will or Conversion-Act). The legal tools of competition law encompass the behavioural and structural remedies that can be imposed on businesses. Behavioural remedies aim to control power exercise, whereas structural remedies seek to control power itself. Conventionally, there is a priority in applying the tools—if controlling power exercise is sufficient to address a problem, there would be no need to diminish the power itself. This is reflected by the well-known (though not unquestionable) saying in competition law that holding power itself is not a problem, but rather how the power is exercised matters.²⁵⁷ Where behavioural remedies are inadequate, structural remedies may be implemented, including blocking mergers and breaking up undertakings, as exemplified by the breakups of Standard Oil and AT&T in the United States.²⁵⁸

The second element to note is the will element, namely the law's goals. Determining the goals of competition law, or what it can and should serve, is a complex and controversial topic. Perspectives on this matter may evolve over time and can vary by region.²⁵⁹ For example, studies in legal history have highlighted that when crafting the Sherman Act, the U.S. Congress considered not only economic goals but also social and political factors stemming from concentrated private power, such as ensuring opportunity, efficiency, prosperity, justice, harmony, freedom of trade and choice, access to markets, liberties, and democracy.²⁶⁰ Subsequently, particularly with the ascent of the Chicago School, competition law shifted its focus towards economic welfare and efficiency.²⁶¹

There was also a shift in focus within EU competition law. The origin of EU competition law is significantly influenced by the Ordoliberalism, a school of thought that vigilantly addresses the potential illegitimacy of private economic power and the devastating effects it might produce.²⁶² The specific goals of EU competition law evolved from preventing public obstacles to interstate trade to controlling mergers and ensuring the liberalization of the public economy sectors.²⁶³ Some contended that the ultimate goal of EU competition law was to ensure that the internal market met consumers’ wishes for competition, not only economic wishes such as lower price, better product quality, more product variety, and more innovation, but also noneconomic ones such as privacy protection.²⁶⁴ Likewise, it has been argued that EU competition law is grounded in values such as consumer welfare, democracy, and fairness.²⁶⁵

Instead of delving into what specific goals a particular region’s competition law has or should have, the aim here is to examine from the power interconnection and interplay perspectives to see what the goals could be. Several points are worth noting:

First, the law only aims to address certain power issues of businesses. Competition law has its boundaries. The well-established ones are (1) inappropriate power combinations in the form of horizontal or vertical agreements among different undertakings, as seen in collusions; (2) inappropriate power exercise of the undertakings that hold significant power, as observed in abuses of dominant position cases; (3) inappropriate creation of significant power through concentration, as in blocked mergers; and (4) inappropriate transfer of political power into economic power, as in the control of state aids or regional aids. Among these, the control of dominant undertakings’ inappropriate power exercise deserves special attention. This is because other laws may also control dominant undertakings’ inappropriate exercise of power. If it is a conduct that should be banned among all the undertakings in a sector, then other sector-specific laws or regulations might be more appropriate than competition law, like the control of data processing in the EU General Data Protection Regulation (GDPR).²⁶⁶ That said, there may still be opportunities for cooperation between competition law and other laws. For instance, both the German Competition Act and the GDPR were applied in the German Facebook case series.²⁶⁷

Second, the power issues that competition law is able to address may extend beyond purely economic. With its legal tools, the direct influence of competition law is on the P_Economic and its exercise as well as P_{Technological} and its exercise, such as preventing collusions that increase prices, prohibiting abusive tying and bundling, blocking harmful concentrations, and the recovery of unlawful state aids. However, it has been demonstrated that P_Economic and P_{Technological} can interplay with P_{Informational} and P_Political through digital products.²⁶⁸ In cases where such power linkage exists, a change in P_Economic or its exercise, or a change in P_{Technological} or its exercise, may also lead to a change in P_{Informational} or P_Political. This influence is noteworthy, as it means competition law may affect noneconomic interests such as personal and social lives. Just as Senator Sherman made about trusts in the congressional discussion concerning the Sherman Act, ‘[i]f the concentered powers of this combination are intrusted to a single man, it is a kingly prerogative, inconsistent with our form of government, and should be subject to the strong resistance of the State and national authorities.’²⁶⁹

Further to the second point, it is also noteworthy that when addressing an issue with competition law, it would be inappropriate, or even problematic, to view noneconomic interests solely through an economic lens and only employ economic standards like efficiency and consumers’ economic welfare. This is because such an approach overlooks the different goals and values inherent in noneconomic interests, which may diverge or even conflict with economic interests. For instance, promoting privacy, autonomy, and democracy would necessitate restricted access to personal data,²⁷⁰ whereas economic advancement often requires more extensive data mining.²⁷¹

IV. THE PRACTICAL ADVANTAGES OF THE PROPOSED POWER CONCEPTION

With the reconception of power, this section aims to exemplify how the proposed power conception can address Shortcoming I (the gaps in the established analysis paths) by filling existing gaps and moving beyond path dependence. The analysis will be based on the cases related to the Facebook/WhatsApp merger. Merger cases often involve future predictions, and ex ante assessments usually require significant effort. Both the EC’s decisions and the U.S. FTC’s new litigation raised in 2021 (pending) will be analysed. Additionally, the discussion will incorporate the competition between WeChat and Renren in China as a comparative example, offering insights from a distinct geographic market.

A. Following the Evidence: A Discrepancy between Theories and Practice

The EC’s analysis followed the conventional competitive assessment theories and methods. In the absence of the vertical relationship between Facebook’s social networking services (Facebook Network) and WhatsApp’s consumer communications services (WhatsApp CCS), the EC assessed horizontal and conglomerate effects.²⁷² The decision outlined three reasons why the merger did not raise serious concerns. First, if Facebook Network and WhatsApp CCS were considered competitors in a broader relevant market for social networking services, there would be many alternative service providers available to consumers.²⁷³ Second, in terms of the competitive landscape, Facebook Network offered features such as personal homepages, newsfeeds, and timelines, which were not provided by WhatsApp CCS, indicating they were not close competitors.²⁷⁴ Third, potential conglomerate integration between Facebook Network and WhatsApp CCS was hindered by significant technical hurdles, including user profile matching.²⁷⁵ Additionally, even if an integration were to take place, it would be mitigated by the fact that a large portion (70–90 percent) of WhatsApp’s active users were already Facebook users.²⁷⁶

It is also noteworthy that the decision stated ‘[n]o indication was found of WhatsApp’s plans to become a social network which would compete with Facebook absent the merger’.²⁷⁷ However, the EC did not further explain the meaning of ‘become’—whether it implies (1) transforming WhatsApp CCS entirely into social networking services like Facebook Network, (2) developing an independent social networking app, or (3) adding additional social networking features to the WhatsApp CCS. The decision did not examine that, if WhatsApp were to alter its plan, how difficult it would be for WhatsApp to add social networking features and become a competitor of Facebook Network.

The last point is significant because, despite the appearance that WhatsApp CCS was not a competitor of Facebook Network, evidence on P_{Technological} shows that the possibility indeed existed. In the digital sector, consumer communications services and social networking services can turn from noncompete to compete by incorporating additional features. To illustrate this, consider the example of WeChat and Renren Network in China. Renren Network was the flagship product of Renren company and was referred to as the ‘Facebook of China’.²⁷⁸ Renren Network reached 31 million monthly unique log-in users and 117 million activated users when the Renren company made its initial public offering on the New York Stock Exchange in May 2011.²⁷⁹ The first version of WeChat was launched in January 2011, 4 months before Renren was listed.²⁸⁰ Similar to WhatsApp, WeChat initially focused on consumer communications services, such as sending texts and voice messages, images, videos, and audio and video chat, without providing any social networking functions.²⁸¹ By the time Renren filed its first annual report for the fiscal year ended on 31 December 2011, WeChat had already reached 100 million registered users with its consumer communications services.²⁸² However, Renren did not identify WeChat as a competitor in its annual report for the fiscal year 2011.²⁸³

The competitive relationship changed swiftly after WeChat introduced social networking features. In April 2012, WeChat made a notable change by introducing the Moments feature, integrating social networking functions into its version 4.0 interface, as shown in Figure 5, and swiftly followed by great success.²⁸⁴ The Moments feature enables users to share text, pictures, and videos with friends, with posting and viewing functionalities similar to those of Renren Network.

It was starting from this year that Renren began to list WeChat as a major competitor in its fiscal reports.²⁸⁵ The financial statements of Tencent, WeChat’s developer, revealed that from 2012 to 2013, WeChat’s monthly active users increased from 160.8 million to 355 million, followed by a steady increase in the following years.²⁸⁶ Meanwhile, 2012 marked a downturn in Renren Network’s monthly unique log-in users, from 56 million to 45 million, followed by a significant decrease in the next few years, leading to the eventual sale of Renren Network by Renren company.²⁸⁷

Given that both WeChat and WhatsApp started as consumer communications services and Renren Network and Facebook Network were primarily social networking services, the competition between WeChat and Renren Network appears to contradict the EC’s perspective that WhatsApp was unlikely to be or become a competitor of Facebook Network. Realizing this discrepancy, the following subsections will attempt to identify and address this gap using the proposed power conception.

B. Locating the Gaps in Conventional Competitive Theories

The evidence regarding the P_{Technological} of WeChat and Renren helps to show that in the digital sector, a highly competitive rival can rapidly emerge from a neighbouring market through feature development (hereinafter ‘neighbouring competition’). Based on the proposed power conception, especially the P_{Technological} and the will element in power exercise, this subsection aims to elucidate why conventional theories have proven inadequate in assessing the competitive influence of neighbouring competition.

1. Following P_{Technological}: The Difference between Neighbouring Competition and Conventional Cross-Market Competition

Although conventional competition law theories encompass cross-market competition and power leverage, such leverage typically hinges on P_Economic. Examples include pure tying and bundling without integrating products or services, exclusive dealing, and imposing other unfair conditions. In the digital sector, however, tying and bundling may be based on P_{Technological}. For example, the Microsoft/LinkedIn case examined the feasibility of tying, bundling, and preinstalling LinkedIn’s professional social network services to Microsoft Windows and Office.²⁸⁸ New theories of technology-based cross-market power leverage were also put forward and developed with disputes, such as self-preferencing in the Google Search (Shopping) case, in which the EU and the United States had totally different opinions on whether Google’s self-preferencing conduct—preferred treatment of its own comparative shopping services in its general search results—constituted competition on the merits.²⁸⁹

Neighbouring competition is different from conventional cross-market competition. Through the lens of P_{Technological}, a further anatomy can be conducted on the tying and bundling practices typically addressed by conventional competition law. The underlying power structure involves two different kinds of services with their respective P_{Technological} A and P_{Technological} B. To facilitate power interplay, namely cross-market competition, an extra P_{Technological} C is needed to integrate the two services technologically. This framework serves as the foundation for the EC’s conglomerate effect assessment in cases like the Facebook/WhatsApp merger. The EC cited technological impediments to integrating Facebook Network and WhatsApp CCS,²⁹⁰ indicative of the absence of P_{Technological} C. However, neighbouring competition operates under a distinct power structure. From the perspective of P_{Technological}, the requisite technologies involve the development of additional features based on an existing product, namely P_{Technological} A(+B). This fundamental difference underscores why neighbouring competition eludes capture by conventional conglomerate effect assessments.

2. Following the Will Element: A More In-Depth Analysis Is Needed for the Potential Competition Theory

Following the will element in the proposed power conception, another theory of harm is notable—lessening or eliminating potential competition. The potential competition theory existed long before the Facebook/WhatsApp merger was proposed.²⁹¹ There has been growing recognition of the need to improve the potential competition theory, including better considering the competitive influence of the acquisitions made by tech giants. For example, the OECD held a roundtable on the concept of potential competition in 2021, which examined the concept of potential competition, theories of harm, and the key parameters to identify potential competition.²⁹²

With the acknowledgement of the advancements made in this area, the primary objective of this research is different. Rather than aiming to deliver a sophisticated analysis of the potential competition theories, the focus is to demonstrate how the proposed power conception can be applied to help identify and address the issues that might be overlooked following conventional theories and methods.

This further exploration is significant. Take the Facebook/WhatsApp merger as an example. Despite the availability of the potential competition theory at the time of the merger proposal in 2014, neither the EU nor the U.S. regulatory bodies opposed the concentration.²⁹³ It was not until December 2020, nearly 7 years after the merger, that the U.S. FTC took legal action, contending that the merger should have been blocked because it allowed Facebook to neutralize perceived competitive threats,²⁹⁴ resulting in the elimination of potential competition. The aforementioned evolved potential competition theories and related theories like nascent competition were developed after a time lag.²⁹⁵ The analysis presented in this subsection and the subsequent one aims to show how this lag could have been mitigated had the proposed power conception been adopted earlier. Furthermore, this example analysis underscores the advantages of the proposed power conception by demonstrating its capacity to identify and address analytical gaps that may arise when applying general theories to new situations.

A closer look at the Facebook/WhatsApp merger cases back in 2014 reveals that potential competition did not receive sufficient analysis. At that time, the U.S. FTC did not disclose the details of its investigation, leaving uncertainty regarding whether it considered potential competition. On the EU’s side, the EC considered potential competition in a paragraph in its decision:

‘As regards the first claim concerning potential competition, the Commission collected and assessed relevant evidence, […]. No indication was found of WhatsApp’s plans to become a social network which would compete with Facebook absent the merger. Indeed, the focus of WhatsApp has traditionally been on offering a light and simple communications service on smartphones only.’²⁹⁶

Despite the EC’s acknowledgement of the potential competition issue, following the proposed power conception, it can be demonstrated why the brief analysis provided in the decision was inadequate. Recall that (1) power is a capacity, distinct from its exercise, (2) will is an important element in the exercise of power, and (3) P_{Technological} is an important power type in the digital sector, and it closely interplays with P_Economic.²⁹⁷ The EC’s analysis focused on a specific status and time—WhatsApp’s plan by the time of EC’s investigation. However, the decision failed to address several critical aspects: (1) whether WhatsApp had the capacity to provide social networking services and become a competitor of Facebook, such as WhatsApp’s technological capacity (namely P_{Technological}) to develop social networking features and its economic capacity (namely P_Economic) to support the realization of P_{Technological}; (2) whether and, if so, how soon might WhatsApp change its will to consider providing social networking services without Facebook’s acquisition, and how would Facebook’s acquisition influence such will; and (3) what would the competitive influence be if WhatsApp’s will was negatively influenced by Facebook. These leave gaps in the evaluation of potential competition. Addressing these gaps is crucial, particularly given the focus on ‘potential’ competition in the examination.

C. Filling the Gap with the Proposed Power Conception

Following the power’s nature as a capacity, the will element of power exercise, P_{Technological}, and the interplay between P_{Technological} and P_Economic,²⁹⁸ several additional factors warrant consideration when gathering evidence and conducting competitive assessments regarding potential competition.

1. The Capacity Perspective

From the perspective of capacity, both technologically and economically, the feasibility of WhatsApp offering social networking services should be assessed. This evaluation should encompass the following key considerations.

First, the magnitude of technological barriers should be considered. As shown by initial evidence, the ways to provide social networking services include not only developing independent apps but also adding new features to existing online services. Given the prevailing global trend of enhancing messaging services with supplementary features,²⁹⁹ particular attention should be directed towards how difficult it would be for WhatsApp to integrate social networking functionalities into WhatsApp CCS.

Second, from the perspective of the power interplay between P_{Technological} and P_Economic, the assessment should consider how much would it cost to develop and add the social networking features, and whether WhatsApp had the economic capacity to support the development of the features. If WhatsApp had or could have the economic capacity to support the development of the additional social networking features, the analysis can then proceed to the next perspective, the potential degree of power.

Third, from the perspective of the potential degree of power, it is noteworthy that the extent of the P_{Technological} that could be potentially developed would have a significant influence on whether WhatsApp CCS could become a strong competitor of Facebook Network if the social networking features were added. Regarding this, the evidence on potential power degree is different from the evidence collected in a conglomerate effect assessment. For example, the EC examined the potential influence of integrating WhatsApp CCS with Facebook Network. It was found that even if such integration were to occur, its effects would be mitigated by the significant overlap in users, with a large percentage of WhatsApp’s active users (estimated between 70 and 90 percent) already being users of Facebook.³⁰⁰ However, in the context of potential competition, attention should switch towards whether this substantial user overlap might make it more convenient for users to migrate from Facebook Network to WhatsApp’s potential new social networking feature, especially as users have already built their networks on WhatsApp. To assess this, evidence needs to be collected and assessed regarding users’ perceptions of the lock-in effect associated with social networking applications, as well as their preferences for the convenience of integrated communication and social networking functions within a single app.

2. The Power Exercise Perspective

From the perspective of power exercise, more evidence on the will element should be collected and examined. Recall that, based on will and action, the forms of power can be classified into Elevation, Conversion-Will, and Conversion-Act.³⁰¹ Because the addition of social networking features is only potential—no action has been taken yet—the focus should be on Elevation (when the patient’s will is in line with the agent’s will) and Conversion-Will (when the patient’s will is not in line with the agent’s will).³⁰² Following these two aspects, three key points emerge: (1) the will of the patient, WhatsApp; (2) the will of the agent, Facebook; and (3) how Facebook’s proposed acquisition would enable Facebook to influence WhatsApp’s will. Moreover, given that a business’s will is manifested through its internal decisions,³⁰³ internal documents serve as crucial evidence. With the key points and evidence, a more detailed analysis can be conducted, focusing on the following aspects:

First, whether and, if so, how soon might WhatsApp change its will to consider providing social networking services without Facebook’s acquisition. This further entails considering the following aspects: (1) gathering evidence not only on WhatsApp’s publicly declared will but also its internally expressed will, such as those expressed through internal emails and meeting records; (2) the search for evidence needs to include not only whether WhatsApp has indicated it might add social networking features, but also whether it expressed it is reluctant to add the features and the underlying reasons; and (3) if evidence shows WhatsApp was unaware of such features—an unlikely scenario given the prevailing global trend of messaging apps integrating social networking features³⁰⁴—further evidence should be gathered and examined on how likely WhatsApp might generate a will on whether to add the features.

An example of the evidence for the third point is how WhatsApp’s subscription-based revenue model might be adversely influenced by the growing trend of the ‘free plus advertising’ profit model. In fact, WhatsApp’s subscription fee model incurred significant net losses for years—it incurred a net loss of 232,508 USD in the first half of 2014 alone.³⁰⁵ Meanwhile, the ‘free plus advertising’ model was growing increasingly prevalent and profitable.³⁰⁶ In contrast to consumer communications interfaces designed for private conversations, the social networking interfaces are less private and serve as platforms for connecting people. Therefore, the social networking interfaces present a more conducive environment for integrating online advertising, as exemplified by the advertising business success of WeChat Moments.³⁰⁷ In addition, as illustrated above in Figure 5 with WeChat, there would not be any problem if a user wanted to simply stick to the messaging function, as social networking can be integrated with a different table. Actually, integrating social networking features and other features with existing consumer communications features may well facilitate and promote users’ usage of WhatsApp, as shown by the user growth of WeChat.³⁰⁸

Second, it is rational to assume that Facebook’s will would be to prevent other undertakings from emerging as its competitors, especially those that could potentially threaten its flagship social networking services. This will can be achieved either through improving its own services to become more competitive, which could benefit consumers, or by suppressing or eliminating competition from its rivals, which could lead to competitive harms. Internal evidence regarding Facebook’s will is crucial in understanding its strategies. For example, Facebook’s internal document disclosed by the U.S. FTC showed that Facebook noticed the global trend of messaging apps incorporating social networking features and engaging in fierce competition with social networking apps.³⁰⁹ Meanwhile, the WhatsApp company grew rapidly and emerged as a leading provider of messaging app (WhatsApp CCS) in many regions, including Europe and the United States.³¹⁰ In response, Facebook initially attempted to launch and improve its messaging app Messenger to compete with WhatsApp CCS, but failed.³¹¹ With the failure, Facebook turned to its next strategy—affecting WhatsApp’s will.

Third, it is essential to consider evidence of Facebook’s potential influence over WhatsApp’s will. After the failure of Facebook Messenger to compete effectively with WhatsApp CCS, Facebook turned to acquire WhatsApp to make sure WhatsApp’s will would stay in line with Facebook’s will. After the merger, Facebook’s decisions would turn into WhatsApp’s will. Consequently, despite WhatsApp possessing the P_{Technological} to add new features to its consumer communications services, it would no longer have the will and would not exercise its power to develop social networking features to compete with Facebook Network. In this regard, Facebook’s acquisition served to either (1) elevate WhatsApp’s will, if WhatsApp’s will was not to add the social networking features yet (namely Elevation), or (2) convert WhatsApp’s will, if WhatsApp began to consider adding the social networking features (namely Conversion-Will). In either scenario, through the acquisition of WhatsApp, Facebook could ensure the achievement of its own will.

3. Significance Assessment on Power

Further analysis of power status is crucial in assessing competitive influence, particularly in determining whether a loss of potential constraint is significant. First, it is essential to evaluate the extent of the potential constraint reduction if the merger were to be approved. This question is the counterpart of how strong the potential competition would be if the proposed merger were prohibited. The assessment closely aligns with capacity analysis, the third point of which is the potential degree of the P_{Technological} that could have been developed. For example, if the social networking features were added to WhatsApp CCS, how strong would it be as a competitor of Facebook Network?

Second, it is necessary to consider the remaining potential competition as competitive constraints. The assessment should focus on, if the proposed merger were approved, whether there would be other potential competitors and their competitiveness. Although case-by-case analysis is needed, the aforementioned points regarding capacity, power interplay, and power exercise remain relevant. For example, it is important to examine the technological and economic capacity of other undertakings to introduce social networking features to compete with Facebook Network, and their wills to do so. The remaining potential competition should be compared with the potentially lessened competition, as it helps to show how significant the lessening of competition would be. It is noteworthy that the ‘significant’ here has a different meaning from the ‘strong’ mentioned in the last paragraph—the potential competition from other undertakings could also be strong. If so, the lessening of competition resulting from the proposed concentration might not be significant.

V. CONCLUSION

With the rapid development of the digital realm, many novel competitive issues emerged, encompassing not only those that entail economic harms but also noneconomic harms such as those related to personal data protection and democracy. The response to these issues frequently sparks debate regarding whether competition law should address them and, if so, how. In practice, many novel issues were not addressed promptly, coherently, and effectively. Although many new theories of harm and tools have been proposed after the problems arose, there still lacks an overarching theory to hold various novel theories together, explain why conventional competition law theories fell short, what can be done at a more fundamental level to help better articulate harms and adopt appropriate tools, and, in the future, to help deal with other potential novel issues.

This research posits that a deeper understanding and a more precise conception of power could be the key. Compared with power’s central importance in competition law, its conceptual exploration within competition law appears to be insufficient. By revisiting the nature of power and conducting a more thorough anatomy—exploring power as a capacity, the elements and forms of power exercise, different power types, and their interplays—a deeper understanding of the power and its dynamics can be attained. This reconception also helps better understand the capacity and limitations of competition law in addressing various forms of harm.

As demonstrated by the application example, rather than attempting to overturn the established conventional theories entirely, the proposed power conception can enhance competitive analysis by offering more detailed assessment methods and addressing gaps at an earlier stage. Furthermore, although not extensively explored through example analyses in this article, the proposed power conception has the potential to be applied in analysing other contentious competitive issues. For example, the interplay among P_{Technological}, P_Economic, and P_{Personalinfo(Acquisition)} can shed light on the role that competition law might play in personal data protection matters. The interplay among P_{Technological}, P_Economic, and P_Political can provide insights into how competition law decisions may affect democracy. Overall, the proposed power conception may serve as a framework for integrating competitive theories and analyses more cohesively.

Footnotes

Author notes