Below, we have provided the full transcript of our panel discussion Defining the Market & Assessing Competition Dynamics in the Digital Platform Industry. Read below to see the timely discussion where a panel of experts deepened the discussion regarding this topic, and how it specifically relates to South Korea.
Ki Jong LEE:
Hello, I am Professor Ki Jong Lee of Sookmyung Women’s University. My heartfelt thanks to Competition Policy International for inviting me to today’s session, which I view is rather timely and important.
Today’s topic is an in-depth analysis of market definitions and competitive dynamics in the digital platform industry. As you are all well aware, most digital platforms have a history of expanding their businesses into multiple markets. This has to do with the fact that once a digital platform secures its customer base or data in one market, it can enter another market easily by utilizing such existing customer bases or data. This is a pro-competitive aspect because it accelerates the platform’s entry to the market. Of course, there are anti-competitive sides, too. In particular, when the platforms are vertically-related, it is said that this could lead to anti-competition. And a platform, upon entering multiple markets, can provide a series of products with enhanced complementarities; in other words, a product ecosystem. This product ecosystem increases efficiency and improves consumer experience. This is also a pro-competitive aspect.
On the other hand, if a platform, while providing its product ecosystem, also plays a gatekeeper role, there could be concern regarding anti-competitive actions. Due to this duality of platforms, we must undoubtedly pursue a balanced approach. However, there is an obstacle to this endeavor. Just as I said, the actions of a platform affect not only a single market, but also multiple ones. However, the methodology to analyze such a cross-market effect has yet to be sufficiently developed. It is because the traditional methodology focuses only on the inside of a relevant market.
Therefore, personally, I have been suggesting a new methodology, which is that if one market affects another relevant market, such a market should be defined as a secondary relevant market. In that way, a court will be obliged to look into the interactions between the two markets, and the parties will be able to freely prove both pro-competitive and anti-competitive elements that exist in this secondary relevant market.
However, today, I will not go into more details about my suggestion. My suggestion is simply a stepping stone to enter today’s discussion. Today’s program has been devised to ask two questions of the best experts in their respective fields.
Firstly, what kinds of challenges has the advent of digital platforms brought to the traditional competition analysis methodology?
Secondly, how can we cope with these challenges?
In order to answer these questions, CPI has assembled a distinguished group of panelists.
Let me introduce today’s speakers. First, Professor Dae-Sik Hong of Sogang Law School. Prof. Hong is the director of the ICT Law and Economy Institute, Sogang University, with many research projects involving the digital platform industry under his belt.
Next, Prof. Sang-Seung Yi of Department of Economics, Seoul National University. Prof. Yi has published numerous papers on competition analysis in competition law. He has also offered his expert opinions in numerous classic cases concerning competition law.
Thirdly, Professor Hak-Soo Ko of Seoul National University School of Law. Prof. Ko specializes in Law and Economics, data privacy, and artificial intelligence law while providing expert advice to the Korean government and many public institutions.
Last but not least, President and CEO Bruce Gustafson of the Developers Alliance. President Gustafson has been the leading advocate for the global developer workforce and the companies that depend on them. He is also the founder of the Loquitur Group, a consulting firm.
Today’s session will be largely divided into three parts. The first part is an overview of overall issues.
First of all, I will ask each speaker to take about 5 minutes to explain what kind of challenges these digital platforms impose on the traditional competition analysis approaches. Then, the second part will be about discussing market definitions among individual issues.
And in the third part we’ll be discussing how to analyze the competitive dynamics in the digital platform industry.
Now we’ll begin the first part. As you will have time to talk about specific issues later, in this first part, please give us an overall picture of what’s going on.
On my screen, since Prof. Dae-Sik Hong is positioned farthest on the left, I’d like to ask Prof. Hong first. What kind of challenges do you think the digital platform industry poses to traditional competition analysis?
Hello, I am Prof. Dae-Sik Hong of Sogang University. Considering many panelists here are native Korean speakers and as I am not so fluent in English yet, please allow me to proceed in Korean. I ask for the other panelists’ and Mr. Gustafson’s understanding. If any meaning gets a bit lost during interpreting, I will make up for it later on. Thank you for the opportunity.
As I understand that the overview should conclude within 5 minutes, I will try to be succinct. Later on, I will be discussing specific issues with this case that was sanctioned by the Korea Fair Trade Commission. Prior to that, I would like to point out that because the KFTC holds an important position on the world map of competition law implementation and handles many important cases, the commission faces the same, common challenges that other countries’ authorities struggle with.
Since South Korea’s fair trade law has differences as well as similarities as compared to such laws of the U.S. or Europe, problem solutions cannot be identical, either. Having said that, for now, I will address how regulations are enforced concerning the abuse of market dominance. EU has already enforced its regulations against Google. I understand that even an interim judgment has been made and regulations are being enforced against very large BigTech companies in another digital economy.
The KFTC also has been enforcing its regulations in a similar manner, albeit slightly lagging behind. However, Korea’s market situation is a little different from those of the U.S. and Europe. For example, there is the Google Shopping case, the first one among all Google cases, which KFTC failed to sanction. It was because Google is not number one in South Korea’s search market when it comes to the market share, the most significant indicator. Therefore, following this logic, South Korean authority targeted Naver, the country’s leading search engine.
What these cases tell us is this; we do advocate changes in competition law analyses based on certain features of the digital economy and we are keenly aware that there are issues. Still, from the enforcement authority’s viewpoint, their tendency is to conduct regulatory enforcement in the same old, path-dependent manner instead of developing a new framework or new methods of proof.
That’s what I want to say.
Such an attitude is also detected in EU’s Google cases. Let’s take a look at the current cases that KFTC is handling. Throughout their processes and in their written resolutions, it is clear that KFTC well understands the characteristics of the digital economy and digital platforms. However, although analyzing previous markets and their characteristics is laudable, I doubt if they have become truly progressive in such areas as analysis tools, analysis results, used indices or proofs, and the degree of verification.
Consequentially, aren’t results derived in the same way as we have practiced in the other economy, i.e. the traditional one, the non-platform economy? KFTC acknowledges there are hidden obstacles against deriving such results. For example, as I will tell you later, there are difficulties in figuring out relevant markets, identifying market dominance positions, and evaluating agreements among competitors. In order to overcome such difficulties, KFTC asserts that they need to apply new methods as follows; defining markets more narrowly; or rather dismissing counter-evidence; or, for example, highly valuing the interaction or correlation between the two sides of a given platform market. However, in reality, these methods are not considered.
Also, as for evaluating agreements among competitors, results of such restriction may not be visible. In such instances, KFTC’s analyses would either state that potential results are sufficient or overemphasize yet-to-be-realized threats. All these might be inevitable choices because competition authorities find it difficult to reach a proper conclusion by inserting the existing frameworks into the platform economy. Notwithstanding, I believe that the authority, being on the forefront of regulatory enforcement, should do more research and develop their analysis tools in order to operate within a persuasive regulation framework that everyone accepts.
Yes, let me finish my first part here. Thank you.
Thank you. It seems that Prof. Hong is quite interested in Google Android cases. He has compared the Android cases of Europe with those of South Korea, citing the differences in market positions of Android between Europe and Korea; such differences also need to be considered; though the Korea Fair Trade Commission has in-depth understanding of the digital industry, it is desirable for KFTC to develop a new methodology more proactively.
Next speaker visible on my screen is President Bruce Gustafson. President Gustafson, please answer the same question.
Thank you very much, Professor Lee. I’m a little outclassed by the panel here today. Three professors, it’s very intimidating. I am more horizontally integrated. My background is in engineering, business and law, as opposed to sort of vertically into the legal profession.
The Android case is very interesting. I should sort of start my remarks by saying that the Developers Alliance was an intervener in the European courts on behalf of Google in that case and one of our primary commentaries was about characterizing the market and how the European Commission had kind of, we believe, made errors in that space.
But I’m going to talk about four items today. So first off, I’m going to talk a little about digital platforms, which I think have real world analogies, but are different from ecosystems, and ecosystems is really where I think we need to focus.
Number two, those differences will impact competition analysis.
Number three, ecosystems, I think, compete as ecosystems as well as market participants in many cases. I think that’s significant.
And then number four, I think there are policy implications for competition law reform, based on what we have learned so far. My personal viewpoint, I guess, is based on my participation in the industry, alongside those of my members. We have been involved in litigation and I certainly talk to policy makers all the time on how policy can evolve to anticipate the digital market, and it’s sort of different dynamics versus the real world. And so my focus has been, are the current laws adequate? How should they change? And, are some of the options being contemplated by policy makers appropriate, or are they missing some of the nuance?
So first off, platforms versus ecosystems. I think one way to look at digital platforms, facilitating transactions between multiple markets, I think there are some real world analogs that we can kind of talk to. Certainly, you can look at the Amex case and others to be able to see how those markets are characterized.
An ecosystem, rather than just connecting multiple markets, to my mind, creates a stable market by facilitating one way transactions in a market where incentives and the rents are complex and are not necessarily in balance. So markets that would not necessarily form spontaneously without some steward who is driving them. Analogs outside of the digital space would include open source projects, where there has to be sort of somebody who leads and brings the parties together. Standards organizations, where we have competing objectives, and yet they need to come together, sort of in one harmonious group.
Ecosystems then, I would submit, have some attributes that are maybe more analogous to a vertically integrated firm in that they have internal markets as well as outward facing markets. So for each element, an ecosystem owner has to sort of create a market. They have to encourage participants to come in and join, or they have to, as a vertically integrated firm, well, they would have to build those contracts somehow facilitate their action to create a complete system that produces a product for consumers. I think that stewardship role is key.
If you look at a vertically integrated company, there is no obligation for the independent elements inside to be profitable, to behave in sort of any particular manner, and I think when you sort of externalize that in an ecosystem model, it gets a little more challenging.
Android is an excellent case in point. So the role that Google plays is both as a participant in that market, but also as a steward of the overall ecosystem, to make sure that the various parties are able to benefit from the transactions that are sort of being incurred inside the market. So when we talk about competition analysis, a participant in an ecosystem, just like inside an integrated company, could operate below cost in one market and be overcompensated in some other portion of the ecosystem, such that they come out whole. I think that dynamic highly complicates the analysis of the overall marketplace, if the behavior was independent of the ecosystem, it may be seen as anti-competitive. Further, where ecosystems touch non-ecosystem markets or where the stewards compete themselves inside the ecosystem, I think we end up in a very complex space.
I’ll use app developers, who we represent, as an example. An application developer is not directly compensated, typically, for their work, or often is not. They rely on advertising as a monetization mechanism. The advertising ecosystem is part of the larger Android ecosystem. App developers do not produce phones. They don’t produce operating systems, that’s handled by other parties, but nobody buys phones without apps. So there is a web of interactions and inter-reliance that sort of makes the system work.
In the EU Android case, the Commission did a poor job, I think, of identifying those various markets and the activities in each. The context does matter, and we stepped in as representing developers to say that the impact on them, the system on them, was overlooked.
Third, ecosystems compete as ecosystems as well as markets. I think one of the characteristics of the big ecosystem players is they have a competence in bringing market participants together. They have the digital tools, they know how to manage the complexity. And so you will often see the biggest competition with a platform company or a large ecosystem company is another ecosystem company entering into their market, a lot of their competencies overlap.
And so in looking at marketplaces, I think, ecosystem as a whole, is a market or a market function.
I’ll conclude with my thoughts on sort of the policy implications of this and when I talk with policy makers, this is what I try to highlight. Reaching into an ecosystem and trying to identify market failures or characterized behavior is difficult, especially if you only look at one segment of an ecosystem, the market inside an ecosystem. I would submit that there has been more success to date in regulating the inputs and the market externalities that go along with an ecosystem. So regulating privacy or copyright or patents or content is a very effective way to shift the incentives within the ecosystem. And I think that’s sometimes the most direct way to get at behaviors or promote behaviors that are positive for the marketplace as opposed to potentially dangerous.
So, I will leave it with that. Thank you very much.
President Gustafson also seems most interested in Android cases. According to President Gustafson, this ecosystem itself forms an autonomous order through stewardship; without such an operating system, it is impossible to do revenue-generating business amongst the system’s participants; and when such an ecosystem exists, the market itself becomes very complicated. In short, President Gustafson thinks that though there are separate product markets, there is room for an ecosystem itself to be able to function as another market. So, all these complicated dynamics should be taken into account, which is lacking at the moment. If a given ecosystem is regulated, such regulations for other areas beyond competition law would be quite effective, as he said.
Thank you for your overview, President Gustafson.
Next, I’d like to ask Prof. Ko for his own general overview. Thank you.
I’ll make three brief points, perhaps with a broader perspective.
First is about data. I mean, everyone nowadays says data is important. Having access to data or having possession of data is significant and everybody talks about it, but at the same time, it’s very, very difficult to make the assessment as to how data makes an impact on competition dynamics. So it’s, in that sense, it’s a very early stage to have come up with a coherent theory about data’s impact on competition or competition dynamics. That’s the first point.
The second point is related to the data. One major role these days that people talk about data is that, data has a tremendous impact on artificial intelligence; developing a model of artificial intelligence. And if a company has much more data than others, then they can come up with a better AI model or AI algorithm. And in that sense, they are given a competitive advantage. But there, again, having data may possibly be a necessary condition, or even- it may not even be a necessary condition, and combined with data, there may well be many other factors that are required to have a competitive edge. So again, at this stage, talking about data and an AI-related competitive edge, that’s an area that’s very underdeveloped, and we are still a long way to go in terms of understanding how data and AI modeling work together.
The third point is more about platforms. A platform could serve many different functions, and one of the functions that a platform could possibly serve, would be as your gateway, Gateway meaning providing a connection between a platform and their consumers. So large platforms can possibly serve as your gateway between these two; between a service provider and consumers. So for example, there are many platforms, but just as an example in Korea, there’s a service provider named Kakao. Everybody in Korea would know this name and Kakao is basically a messenger app, so people use Kakao for sending and receiving messages. But, at the same time, as a big platform, Kakao serves as a conduit between, on the one hand, consumers or users, and on the other hand, various types of service providers. And here, the service providers include not just private business entities, but also public institutions and even government agencies. And if you live in Korea, you would get messages from government agencies through Kakao, and moreover through Kakao, consumers these days can make a financial payment or do money transfers. And this means that Kakao serves as a main gateway for some types of financial transactions. So overall, some platforms function as a gateway, as we saw through Kakao. But other platforms do not serve as a gateway. And even if some platforms serve as gateways, at this point, it’s really not clear whether they show a significant anti-competitive impact. So there are many scenarios possible in terms of market dynamics. So again, in this area of seeing a platform as a gateway, the analysis is still at an early stage, I would say.
And I’ll stop there. Thank you.
Prof. Ko has neatly summarized three points. The first point was that we hardly know either how influential data are or how to analyze data, i.e., the influence of data.
Secondly, data have become the critical key to developing artificial intelligence.
Thirdly, there are platforms who play the role of a gateway and we need to pay attention to such a role. For example, South Korea’s Kakao plays a gateway role in our financial services; from business people and consumers to the government or public institutions, many access financial services through Kakao.
Prof. Ko adeptly pointed out these three issues.
Prof. Sang-Seung Yi, you are next. Please give us your general overview.
Yes. I’m in broad agreement with the other panelists that there are many challenges in assessing the competitive effects of the digital platform industry. Market definition, assessing dominance and evaluating potential anti-competitive effects, and balancing them against pro competitive effects.
So first, on the market definition, on the consumer side, we all know that the services are often provided for free. The platform operator typically monetizes his services by selling advertisements. So, on the consumer side, assuming that you defined two separate markets, one cannot apply the common technique of defining the relevant market, that is the SSNIP test, because increasing the price by 5% or 10% has no meaning when the starting price is zero.
So the SSNDQ test, the small, but significant non-transitory decrease in quality test has been proposed as an alternative, as a conceptual framework. I agree that the SSNDQ test makes intuitive sense, but it is really difficult to operationalize the SSNDQ test because often, the data is not available.
And also as many panelists said, the platform operators, or ecosystems, provide many, many different services. And also the services evolve all the time. So it is really hard to come up with a single measure of quality.
Just for one example, think about targeted advertisements. I personally prefer to see targeted advertisement that suits my needs instead of seeing random ads, but people differ in their tastes. So some people may say that, are they surveilling me? Or are they looking at my online behavior? So people have different preferences, so it is hard to come up with a single measure of quality and also, the appeal of the SSNIP test for defining the relevant market is simplistic. You only need the price to increase 5 or 10%, and then how many consumers will switch from your service to the potential competitors. That’s the only information you need, in addition to the margin. But even if you assume that you come up with a single measure of quality, then if you reduce 5% or 10% the quality, you have to come up with the cost reduction. How much cost is reduced when you reduce quality. So it is very difficult to apply the SSNDQ test.
Second, about assessing market dominance. Because these platforms provide multiple services, which evolve continuously and often rapidly, it is difficult to identify who the competitors are and what their strengths are. I think these difficulties arose in the recent advocacy’s lawsuit against Facebook, where the trial judge dismissed, or granted Facebook’s motion to dismiss the case precisely on these grounds. That is, what are the so-called personal social service networks.
In the FTC’s original complaint, the judge pointed that out. In the amended complaint, the FTC now has identified Facebook, Instagram, Snapchat, and MeWe; only four, excluding all other services, which at least some aspects of the personal social network services are provided.
And also, calculating market share is difficult because, precisely for this reason, that is, think about YouTube. Now they have some features of the social networks services, like commenting and so on, following. So then how much time do people spend on YouTube, on the personal social network service? It is really difficult to come up with reliable measures.
And third, on assessing the potential competitive effects, I think there is broad agreement that potential competition and nascent competition should receive far more weight in competitive assessment of the digital platform industry. But again, it is easier said than done. That is, the FTC is currently, in 2020 and 2021, challenging Facebook’s acquisition of Instagram in 2012 and WhatsApp in 2014. But the natural question is, the FTC looked at all the documents that Facebook submitted on the second request about the internal documents and so on. At that time, the agency approved them, but then it changed and now challenges the acquisitions as an anti-competitive scheme in violation of section two. But then, the FTC has currently been silent on why, what, for what reason did it change its mind. So these are some of the difficulties, and I do not have the answer yet, so I think we should discuss it more.
First, Prof. Yi addressed the fact that in market definitions it is hard to apply the SSNIP test, which is based on the traditional pricing, to the digital platform markets. However, he continued that the SSNDQ test does not work well, either as an alternative. He said that, in fact, there are considerable difficulties in implementing the SSNDQ test.
Also, in evaluating market dominance, it is hard to figure out who competes against whom in the digital platform industry. For example, social media and YouTube can be in a competitive relationship, and YouTube also could play a role in social media, said Prof. Yi.
He also pointed out the importance of evaluating potential competitors in the digital platform market, mentioning the KFTC’s recent re-review of business consolidations that it had already approved in the past.
Now we have heard everybody’s overview. Though this first part is essential for today’s discussion, it seems that we went over the originally allocated time. Therefore, I am afraid that we need to spend less time on the remaining two parts.
The second part is about discussing the market definition in detail. I am very sorry but please try to spend only about 2 to 3 minutes per speaker.
Prof. Sang-Seung Yi, since you just spoke about the market definition in such detail, it would be great if you could continue on further.
Sure, given the time shortage, let me just say that there are two common approaches, one is defining two markets that are interrelated.
Secondly, is defining a single market based on the difference, whether the platform can be characterized as a transaction platform or non-transaction platform. I think this approach makes sense, but instead of spending too much time and energy on the precise boundary of the relevant market, I think we all agree that we should begin with the competitive assessment, that is, we should keep in mind the first principle that we define a market for the purpose of assessing the competitive effects of conduct at issue.
So let me stop here. We can come back to, I think, to the Google case on this issue, but later.
Earlier, Prof. Dae-Sik Hong pointed out the problem of defining a market too narrowly in the market definition. Prof. Hong, please speak a little more about the market definition.
Sure. Since the time is of essence, I will address a specific case. It is the Google Android case in South Korea. Although it is similar to the Google Android cases in EU, I shall explain it briefly for the overseas audience who are not familiar with this case.
The EU cases involve, firstly, tying of search apps, secondly, certain payments requiring pre-installation of search apps, and thirdly, the anti-fragmentation agreement.
Korea’s case was focused on that third issue, the anti-fragmentation agreement. It was because there was no visible exclusion of competition in Korea even though both the first and second cases of EU were applied to Korea. As the result, the KFTC couldn’t challenge it.
The third case had to do with Google’s imposing its anti-fragmentation condition while package-licensing its own apps pursuant to the MADA, i.e. mobile application distribution agreement. KFTC only took issue with the third case, whose issue was about how to define the market.
In reality, the ecosystem competition is between Apple’s ecosystem and Google’s own. Apple wasn’t investigated as it adopts a closed setting, with no licensing offer. Thus, KFTC said that it had only targeted the ecosystem that offers licensing. As the result, Google’s market share or influence ended up looming large. In this case, everything from the market definition to how to define a market was drawn quite narrowly, which in turn influenced the next analysis.
Another issue arose, which was that, in South Korea, not only smartphones but also other mobile devices became issues. Those devices such as mobile watches, smart watches, smart TVs, and smart speakers also need an OS in the end. Yes, those devices do form a new market.
When South Korea’s OEM manufacturers first developed such products, they utilized Android because it is an open source software. Another related issue was some apps’ claiming difficulty in manufacturing due to lack of coordination with Google. Then, KFTC launched the concept of “other smart device OS market”, saying that Google impeded innovation in such a market.
The question I have about the market definition is this excessive broadness of the “innovation market”, a concept that the KFTC came up with, while defining the mobile OS market narrowly. Other mobile devices refer to such items as watches, TVs, speakers, etc. After all, each of these devices forms a separate ecosystem, however, the KFTC bundled them together, designating the cluster as “other mobile devices”, and then introduced an OS market specific to the cluster.
This constantly fluctuating definition of a market size from narrow to wide by the KFTC seems too unprincipled in my opinion.
Prof. Hong is understandably most focused on South Korea’s Google Android case. He’s expressed his frustration regarding the KFTC’s market definition, which changes from too narrow to too wide depending on a case.
In particular, considering Google’s ecosystem is competing against Apple’s ecosystem, he stated how lacking it was that this reality was not factored in the decision.
President Gustafson also addressed ecosystem’s market and ecosystem’s competition earlier. President Gustafson, do you have anything else to add to the competition within an ecosystem?
I will be very brief.
I think that at a fundamental level, companies know what markets they compete in, and they know who their competitors are. And so for a court to sort of second guess what the participants already know is always a surprise to me when the answers can be determined. But the reality is in an ecosystem, there will be more than two sides to a market. There will be more than one participant. They will overlap. And when it comes time to sort of define how markets are, whether they’re competitive or not, I think you have to look at the impact one market to another behavior in one market, how it affects another market within the ecosystem, or you’re going to get sort of a faulty answer because of the trade offs between the various components.
Yes, thank you for your succinct remark.
In fact, President Gustafson already addressed this issue sufficiently; within this ecosystem are quite diverse players as well as multiple competitive relations. However, such diversity is yet to be reflected in competition law enforcement, as he added. Perhaps he could elaborate more later on while addressing competition analysis.
Prof. Hak-Soo Ko, we are discussing market definition. How should we solve the market definition issue in the digital industry?
Well, that’s a very challenging and difficult question, and I’ll give a couple of examples, which is beyond the competition law perspective, but more from a regulatory jurisdiction viewpoint, I would say.
The first case is the Kakao that I mentioned already. Kakao is the dominant messenger app in Korea and a few other countries. And I was told each DAU, daily active user, the number of DAU is 50 million, that is close to the whole population of the entire country of Korea.
So basically, pretty much everyone in Korea uses Kakao service on a daily basis every time, I mean every day, constantly. So in that sense, Kakao is essential if you live in Korea. And Kakao, obviously as a messenger service, people use the service as a messenger or social media in a broad sense. But as I mentioned, Kakao is being used for purposes of making payments and money transfers. So in that simplified sense, Kakao serves as a messenger app, plus serves as a financial conduit for financial transaction service.
So if you simplify that way, Kakao serves two main different functions. Well, in actuality, there are many other functions that Kakao provides, but depending on how you distinguish, or if you want to categorize Kakao’s services, at least partially, as providing financial services, you’re now subject to, you should, well, at least theoretically, be subject to a whole new set of regulations coming from the financial industry. And financial industry in itself is a very heavily regulated industry. And if you categorize it as the messenger service, then the whole different set of regulations should be in place. So depending on how you categorize your service, you’re in a totally different ball game.
The second example is sbout the my-data scheme. My-data is coming from what’s called, right to data portability. Data portability was enacted, perhaps, first in the EU, under the GDPR, and in Korea, credit information act, which governs financial data introduced the scheme of data portability, and colloquially it’s called “my data.” And my data means different financial services companies can ask their customers to request transferring customers’ data from one platform to another platform. And there are currently dozens of financial institutions that are involved in this scheme, which is still in the very early stages, but depending on how the market dynamics work out, there may well be a significant impact at least within the financial industry. And the problem of delineating the different types of players is out there. And the main difficulty is distinguishing between the traditional market institutions, such as banks and credit card companies, insurance companies, and that’s one group.
And the other group is a new commerce into the financial market. So-called FinTech companies, or BigTech companies. And the question is, is it worthwhile distinguishing between these two groups of companies? And even within FinTech and BigTech companies, is it meaningful to distinguish between these two groups? I mean, FinTech tends to specialize and they tend not to have too many different business lines. And as opposed to that, BigTech companies tend to be large and tend to provide whole different sets of diverse services, and depending on how you view the whole market dynamics, you can come up with different types of ways of delineating market players, and this is an open question and for now it’s still challenging and hard to make an assessment.
Prof. Ko spoke about platforms’ activities beyond the market boundaries. For example, Kakao not only plays a messenger role but also offer multiple services including financial one. There is even a system where a user’s data could move from this platform to that platform.
He also asked how we could easily distinguish FinTech from BigTech or vice versa. After all they have a lot in common. Therefore, I think that Prof. Ko pointed out that the dynamics of the digital platform industry makes it harder to lump together businesses within its narrow relevant market and then analyze them.
Regarding discussing relevant markets, I think we should stop here. It is because the panelists appear more interested in competition analysis within a relevant market than such a market itself.
Well, in what order would you like to go? Earlier, I felt bad that I gave too little time time to Prof. Sang-Seung Yi. Since Prof. Yi said, “let’s focus more on competition analysis than on market definition”, I would like to offer him a little more time.
Prof. Yi, please proceed.
Well, actually I think it’s more productive to focus on an actual case as Prof. Hong mentioned, the Google case.
Let me first disclose that I was the testifying economics expert on behalf of Google before the KFTC, so I’m not a neutral observer, but let me try to update the international audience what the KFTC’s case is about, and what’s the difference between the KFTC’s and EC’s case. So as professor Hong explained, the EC challenged Google’s entire business model on Android, consisting of a web of contracts.
First, the anti-fragmentation agreements, where a smartphone maker agrees that if a smartphone maker installs Google apps, then the maker promises that it will not fragment the Android system by modifying the open source code in an incompatible way.
That is, apps that operate on the compatible devices, whether they are Google apps or non-Google apps, like Kakao, they should work smoothly on all devices.
The second is the distribution agreements where the EC characterizes as time, that is, if he installed one Google app, then you have to install all others, or the certain apps that Google designates.
Third is the revenue sharing agreement, which Google entered into smartphone makers like Apple and Samsung, and also all the network operators where by exclusively pre-installing Google search on the mobile phones, then they get a cut of Google’s search revenue on those devices.
Now, the important, I think, point is that the EC’s theory of harm is that, through all these agreements, Google tried to maintain and solidify its dominance on the general search.
It happens that in 2011, Naver and Daum, Daum is the predecessor to Kakao, they filed a complaint before the KFTC challenging these agreements. That is, by pre-installing the, especially the Mada, pre-installing the Google search, Google is trying to extend its dominance from desktop to mobile. The problem was, the story that I heard was that Google lamented that the most searched word on the mobile Google search engine was “Naver.” They wanted to download Naver app because that’s what they wanted to use. And also on the desktop, the Naver share is about 80%. So in 2013, KFTC dismissed the complaint.
Now fast forward to 2021, this fact did not deter the KFTC from challenging the AFA, anti-fragmentation agreements. And what’s remarkable is that the EC only looked at smart devices, like smart phones and tablets, but the KFTC extended them into all smart devices, including smart watches, smart speakers, smart TVs, as Professor Hong mentioned, based on the innovation market theory, but it’s quite, as Professor Hong said, quite vague, so I really look forward to what the court is going to say on the KFTC’s ruling.
Our Prof. Yi also appears most interested in the Google case. Well, as the topic is on the table, let me ask Prof. Dae-Sik Hong some more questions as Prof. Hong earlier wanted to discuss the same Google case a little more.
Please understand that I am unable to summarize what Prof. Yi said. It is because we are pressed for time. We ask you to understand our situation.
Dear panelists, please keep your speech short from now on.
Prof. Hong, please go ahead.
As Prof. Yi shared his involvement in South Korea’s Google case, I think I should address mine as well. The omission was due to my unfamiliarity with the protocol here. In fact, neither am I an objective observer in this case as I had also submitted my expert report on behalf of Google in the Google Android case. It’s news to me that Prof. Yi provided economic analysis for the case. We appeared to have been unaware of each other’s work… until today.
As someone who was asked to provide a legal opinion on the case, I am not qualified to discuss the case’s factual grounds. That is why I said that the KFTC should make a decision based on certain legal principles and standards.
I’d like to make a few points regarding the KFTC’s decisions; first, in that specific case related to mobile OS licensing, KFTC came to its decision by applying the traditional market foreclosure theory. In the past, the foreclosure theory was typically applied when there was evidence that either a competitor had been ousted from the market or their competitiveness was greatly reduced. However, no such evidence was presented in this case. So, what KFTC brought up was other similar cases raised by the EU. However, KFTC mentioned Amazon’s Fire OS and Alibaba’s Aliyun OS as evidence of market foreclosure… yes, they are indeed Android forks, saying that neither could enter the Korean market properly due to the entry barriers set up by Google Android against its potential competitors.
Some parts in KFTC’s theory make sense, but it is doubtful if Amazon or Alibaba ever tried to enter South Korea’s digital market. After all, South Korea’s ecosystem is Android’s in addition to Apple’s. Within the Android ecosystem, Korean OEM manufacturers such as Samsung were able to pass compatibility tests by making variants based on what Google had provided. That way, Samsung Electronics was able to make their own devices and then inserted their own app market into such devices.
Also, Samsung develops their own apps. In that aspect, you can say that South Korea sees competition among those compatible Android variants. Nonetheless, KFTC excluded the fact in its ruling, saying that there are no such factors that suppress competition amongst such variants. In a way, the ruling by KFTC didn’t sufficiently reflect what people really felt. Even though KFTC only saw certain Android forks, and Mr. Gustafson is with us here, the reason that Google emphasizes compatibility is because from app developers’ point of view, a multiple-fork ecosystem is like a nightmare.
Why? You have to redesign an app to fit the new fork continuously. In that aspect, multiple compatible versions of Android OS do offer clear benefits. However, such an upside was not considered in the KFTC’s ruling, which I find somewhat problematic.
And in the other smart device OS market where KFTC had raised an issue of innovation market, its ruling turned out even more hypothetical. Even though those other smart devices such as watches, TVs, etc. have not seen their own OS markets sufficiently formed, KFTC’s approach is that Google completely blocked emergence of such markets. Therefore, Google impeded innovation, says KFTC.
Speaking of innovation in this context, let’s talk about what’s happened to Microsoft. When regular consumers largely transitioned from PC to mobile devices, MS OS could no longer assert its power in mobile markets because their PC OS’s strengths didn’t transition to mobile devices.
This is a good example of new, utterly disruptive innovation. However, the KFTC views that such disruptive innovation is unthinkable in new mobile devices such as smart watches; only Android variants are possible; blocking development of Android variants is harming innovation. These are how they approach, and I doubt how this attitude could fit in with any concept of innovation.
Thank you. Both Prof. Yi and Prof. Hong addressed the Google Android case. I guess that President Gustafson may have something to add to the discussion.
Please be succinct.
Certainly, thank you.
Apparently, we were all involved in one way or another.
So our testimony in the European Android case was quite clear that, from the developer point of view, fragmented operating systems drive up costs and complicate the market for developers. Operating systems compete with each other, we’ll use Apple and Google as examples for a developer community that is capable of making third party apps. And developers will develop for the platform that is the most profitable for them. And that fragmentation is a factor in that decision. And so, certainly, there is a strong market reason why an operating system would seek to limit fragmentation because of its value in another portion of the ecosystem.
I will make a couple of comments just in response to Professor Ko and Professor Yi. So on data, data portability, data ownership, our view is a little more nuanced from my common law colleagues. We view rights and data as more of a bundle of sticks. There are joint rights by many of the people that are involved in the value chain; the person who collects the data, people who process data, the people who produce the data. And so I think there’s a much more sophisticated discussion will need to take place there before we define a legal framework that is operable.
And then finally on sort of my stewardship comment, there is a cost to an ecosystem to support a brand, and in the consumer’s mind when they are interacting with an ecosystem and all of the participants in it, it is that brand that is foremost in their mind, and so that is a service that is rendered unto the ecosystem, in addition to sort of bringing the parties together. And so what is sometimes termed as a gatekeeper role, I think is also important to protect the general brand value of all of the participants that are in the ecosystem. So we do a service when we make sure that bad actors cannot enter the system when we make sure that there’s a certain quality level. And so that role, likewise, can be a bit of a policeman role. It can be seen as negative, and certainly is painted that way by policymakers, but I think it’s necessary for the ecosystem to function, so that has to be reflected in law somewhere.
Though today’s topics are market definition and competition analysis in digital platforms, this session’s focus is specifically on South Korea throughout the entire CPI series. Understandably that our panelists have zeroed in on the in-country Google Android case.
Prof. Ko, you are not directly involved in the Google case, are you?
Prof. Ko, I will be grateful if you have anything else to add to the Google case, or if you can suggest a new direction that competition analysis of digital platform markets should take.
Yeah, I’d like to make a short concluding remark: I was not involved in the Google case. I feel isolated here, just a joke.
Well, I don’t intend to talk about the Google case, but, and rather than saying there’s a direction for the future, I would say we are currently at a loss as to what the future directions should be. I mean, everyone knows large platforms tend to show a network effect and I mean, network effects naturally would point toward a less of competition. I mean, theoretically, that’s what the textbook says.
What’s unclear is what is really taking place in the marketplace. I mean, some of the large platforms tend to become bigger in Korea, in the US, in the EU, wherever, but that’s not a universal phenomenon in every market segment. So in some markets, some of the platforms are becoming dominant players, but in some other market segments, there’s a very fierce, very dynamic competition going on. And also, platforms, particularly in the past couple of years, some of the platforms are trying to go into the markets that have been dominated by other competitors. So in that general sense, there’s been a tremendous market pressure to go into other fields and become a meaningful competitor.
So in Korea, for instance, there’s a large bit of competition. For example, in the e-commerce market or in the messaging service market or in the food delivery market, in many different markets, there’s a lot of competitive pressure and not dominated by a single player, but several large players that are competing against each other very fiercely. So we’ll have to see what happens in these markets.
Due to the network external effect of a digital platform any growth of a large platform is indeed a competition-threatening factor. However, as we can see in the South Korean market, a large number of new entries are made in various markets such as delivery apps, messaging apps, etc. competing ferociously amongst themselves, as Prof. Ko said.
Subsequently, as he observed, it would be interesting to see how these two contradicting forces would interact with each other and what kind of result would come out of it.
Naturally, with our time, we couldn’t expect to come up with a solution in competition analysis of the digital platform market. Nonetheless, I believe that through today’s discourse we were able to find some clues for a future direction.
I wish all the panelists greater accomplishments and contributions in this area in the future.
I wish we had more time, but today’s discussion has reached its end point.
Dear panelists and online viewers, have a wonderful weekend, thank you.