author:
- Muxin Li, Research Fellow, Economics Department, Università Bocconi, Milan
__________
This blog article is derived from the author’s paper entitled Self-Preferencing Across Markets, a project of the Economics of Digital Services (EODS) initiative led by Penn’s Center for Technology, Innovation & Competition (CTIC) and Warren Center for Network & Data Services. CTIC and the Warren Center are grateful to the John S. and James L. Knight Foundation for its support of the initiative.
__________
The prevalence of digital platforms has endowed us with more accessible information and great convenience. Recently, the number of worldwide Internet users has reached 5 billion, accounting for 63 percent of the world’s total population. As distinguished from other market types, the network effects in digital platforms make the dominant firms more attractive than others, allowing them to further accumulate a larger market share. Consequentially, we find the advent of digital platforms also leaves us increasingly dependent on those dominant companies. Take Alphabet (the owner of Google and YouTube) and Amazon as examples. Each day, there are around 9 billion searches made on Google[1], 122 million active users browsing on YouTube[2], and around 1.6 million packages shipped by Amazon[3] (not taking into account third-party sellers).
In addition to their key services, those monopoly platforms have further ventured into or acquired companies in other industries. For example, Alphabet runs the marketing platform Display & Video 360 (DV 360) for managing online display ads, the real-time advertising technology Google Ad Manager for buying and selling ads, and numerous ad exchange platforms connecting publishers and advertisers. Amazon has over 40 subsidiaries, including its logistics network Amazon Air; various distribution systems with fulfillment centers for storage, packaging, and delivery of products; and the Whole Foods Market supermarket chain. Most of these industries are not mutually independent. On the contrary, the majority of them are within the same supply chain and complementary to each other. Thus, by practicing self-preferencing across industries, gatekeepers can easily exploit their dominant position and users’ trust in one layer of the supply chain to gain competitive advantages in other layers. In other words, they are extending their dominance across the whole supply chain.
In recent years, several influential policy reports have discussed this phenomenon, with the majority of scrutiny to be directed toward Amazon and Google. For instance, the European Commission (EC) recently published the commitment on Amazon which aims to diminish its self-preferencing in logistic services[4]. In 2021, the EC also started a similar investigation against Google’s self-favoring in ad tech services[5]. In the same year, the Italian Competition Authority (ICA) imposed a fine of 1.1 billion euros on Amazon as well as behavioral remedies for self-preferencing in its logistic services. In the United States, the Senate Committee on the Judiciary voted in January 2022 on a bill called the “American Innovation and Choice Online Act,” which forbids platforms from practicing discriminatory treatments in providing services, i.e., self-preferencing[6].
Despite the enormous attention and heated discussion by regulators, there is surprisingly little economic research about whether and how self-preferencing across the supply chain affects market outcomes, and even less guidance on related regulatory intervention. Strictly speaking, this conduct also does not fit in the narrow classification of existing self-preferencing in the hybrid platform research, wherein a monopoly platform facilitates transactions among users and simultaneously retails within the same market. One well-known example is Amazon, a company that was frequently criticized for steering shoppers from third-party sellers to its own products. As pointed out in the existing research (see Etro (2022)), the platform’s dual role in the hybrid business mode creates inherent conflicts of interest and permits self-preferencing through systematic recommendations of its products. Its tradeoff would then involve monetization through seller referral fees or direct product sales margins.
Besides this self-preferencing on the hybrid business platforms, little has been explored about other types of self-preferencing. Note that competing firms in vertically integrated markets do not necessarily rely on the platform to reach potential purchasers. As a consequence, the platform may not be able to charge referral fees from rivals, and the tradeoff in the hybrid business mode no longer applies to self-preferencing in such a market. For instance, Google could practice self-preferencing in ad ancillary technology by restricting YouTube’s interoperability with rivals’ ad servers, but it cannot charge competing companies for making transactions with YouTube publishers. Likewise, Amazon self-favors its logistic services by assigning its purchasers higher chances of winning Amazon BuyBox, but it cannot charge competing firms for shipping orders on Amazon.
Amazon’s self-preferencing toward the logistics industry has an even stronger impact than its self-preferencing toward its retailer offers (see Raval (2022)). As long as users purchase the ancillary products together with the intermediation services, the platform can easily practice self-preferencing by restricting competing firms’ product value achieved in its marketplace, either through manipulating interoperability or steering user attention. In this context, the imperfect rent extraction can motivate the platform to practice self-preferencing (see Motta (2023)), and additionally, the tradeoff of the marketplace platform now arises from its underlying complex relationship with other firms. On the one hand, they are competitors as users view their ancillary products as close substitutes. On the other hand, users consume the platform service and ancillary products together, resulting in a complementary relationship between the platform and other firms.
To comprehensively address these aspects, this paper seeks to establish the first canonical and tractable model, embodying a set of key market features. First, a monopoly platform (gatekeeper) exists as the sole means for users from two distinct groups to interact with each other. Second, users in one group must purchase an ancillary product, in addition to the intermediation service, to accomplish interactions with the other group on the platform. For example, sellers need to acquire delivery services for transactions on Amazon, YouTube publishers have to require advertiser servers to monetize their content, and advertisers must use advertising technologies to manage ads on Google. Third, although the intermediation service is monopolized, its ancillary product is provided by both the platform itself and other competing firm(s). Moreover, the platform can potentially restrict the value of users who purchase ancillary products from competing firms. Lastly, the model must account for cross-group network effects in digital markets, i.e., a user’s surplus on the platform depends on the participation of the other group. These factors guide our selection of framework, prompting us to merge a duopoly model with vertical product differentiation and multi-sided market features. Unlike the literature on hybrid platforms suggesting the conditions for harmful entry and self-preferencing are rarely met, our analysis reveals that platforms are inclined towards self-preferencing in a highly competitive market, leading to the foreclosure of rivals. These findings support existing regulatory concerns.
Within the framework of our model, the implementation of self-preferencing initiates a dual reduction in both the price and demand for ancillary products offered by competing firms. Consequently, the monopoly platform’s practice of self-preferencing always imposes detrimental effects upon its rivals within the ancillary product markets. Additionally, our research also reveals an intriguing insight: implementing self-preferencing does not necessarily guarantee benefits for the platform or lead to negative effects on user surplus. While constraining the product value of competing firms enables the platform to charge higher prices for its ancillary product and generate greater demand, it has to lower the platform fee due to the inherent complementary relationship of intermediation services and ancillary products. When the two ancillary products exhibit high similarity or users derive significantly lower value from purchasing third-party products, self-preferencing greatly redirects users from rival to the monopoly platform, resulting in increased profits. The substantial increase in the platform’s market share of ancillary products empowers the platform to extract more profit from each user. Consequently, there is a stronger motivation for the platform to boost user engagement and curtail platform fees, leading to a corresponding increase in user surplus. When the two ancillary products lack horizontal similarity, however, the opposite outcome emerges.
The theoretical findings underscore the importance of market competitiveness and product value in determining the overall effects of self-preferencing, as these parameters dictate the relative strengths of conflicting impacts on the platform’s profit and user participation. More importantly, we show that self-preferencing does not always reduce social welfare. From a policy standpoint, the research illustrates the underlying incentives for a monopoly platform to favor its own ancillary product in a vertically integrated market. It also points out the conditions under which the platform avoids doing so. More importantly, our findings confirm existing antitrust concerns, as we show that the profit of rivals is generally diminished by self-preferencing. The research also underscores that any subtle variations in remedy design can influence the outcomes of interventions in digital markets.
[1] https://earthweb.com/how-many-google-searches-per-day/
[2] https://www.globalmediainsight.com/blog/youtube-users-statistics/
[3] https://www.algrim.co/3031-how-many-orders-does-amazon-get-every-day
[4] https://ec.europa.eu/commission/presscorner/detail/en/ip_22_7777
[5] https://ec.europa.eu/commission/presscorner/detail/en/ip_21_3143
[6] https://www.congress.gov/bill/117th-congress/senate-bill/2992/text
Bibliography
Etro, F. (2022). The Economics of Amazon. Available at SSRN 4307213. URL: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4307213
Motta, M. (2023). Self-preferencing and foreclosure in digital markets: theories of harm for abuse
cases. International Journal of Industrial Organization, 102974. URL: https://www.dropbox.com/s/q85wgikf4zib2gn/Theories-of-harm-for-digital-abuse-cases-v2.pdf?dl=0
Raval, D. (2022). Steering in One Click: Platform Self-Preferencing in the Amazon Buy Box. Unpublished manuscript. URL: https://deveshraval.github.io/buyBox.pdf