Patrick Rey is a professor of economics at Toulouse 1 Capitole University in France and serves as a consultant to the World Bank and the Organization for Economic Co-operation and Development (OECD). Professor Rey's primary research interests include industrial organization, competition policy, and information economics. He has led various European Union policy reports and published over 80 articles in top academic journals like the American Economic Review, Econometrica, and Review of Economic Studies. He has also served on the editorial boards of the Review of Economic Studies and the Journal of Public Economics. In the summer of this year, Professor Patrick Rey delivered the Jiang Xuemu Economic Lecture series on "Vertical Relations" for Fudan students, and we had the honor to conduct an exclusive interview with him.

Chief Planner:

Kou Zonglai (Professor and Associate Dean of the School of Economics, Fudan University)

Produced by: Office of Disciplines and Talents, School of Economics, Fudan University

Reporters for this Issue:

Hu Bo (Assistant Professor, School of Economics, Fudan University)

Sun Rui (Ph.D. Candidate, School of Economics, Fudan University)

Xu Xiaohe (Master's Student, School of Economics, Fudan University)

I. Academic Background and Perspectives on Research Fields

Hu Bo: Could you tell us what the main message is that you want to convey through the course on "Vertical Relations"?

Patrick Rey: If there's one word that can encapsulate the main theme of this course, it's "competition." Companies engaged in vertical relations need to interact and cooperate with each other, but at the same time, competition may exist. This competition can come from common competitors, the individual competitors of the companies themselves, or even competitors of the cooperating firms. From the perspective of maximizing social welfare, there is a delicate balance between the benefits of promoting cooperation between companies and the potential adverse effects of competition. In this course, I mainly analyze how to manage this relationship—competition needs to be monitored and regulated to achieve a balance between cooperation and competition among companies.

Specifically, for companies with vertical relationships, they provide complementary products or services to each other. Incentives are needed for companies to supply sufficient complementary products or services. To coordinate cooperation among companies, many of them need to reach agreements and contracts. This cooperation is initially achieved by adjusting the profits of companies, and thus, there may be various forms of vertical restraints, such as price restrictions like resale price maintenance, as well as non-price restrictions like exclusive dealing. At the same time, these provisions may distort competition, such as exclusive dealing limiting the entry of potential competitors, which could harm consumers and overall social welfare; resale price maintenance might also be used to eliminate competition.

As economists, we need to carefully identify behaviors that may harm competition. For example, when companies have significant market power, non-price restrictions like exclusive dealing are generally harmful. Price restrictions like resale price maintenance are not a big issue unless this practice is widespread throughout the industry. When individual manufacturers use resale price maintenance to encourage dealers to offer new products, it usually does not create problems. However, if all manufacturers do this, it can reduce social welfare.

Hu Bo: Thank you for your introduction. We will come back to vertical relations later. Now, I'd like to ask a more general question about industrial organization. Compared to 20 years ago, industrial organization research is now more empirical in nature, and from what I know, you are more involved in theoretical research in industrial organization. I would like to hear your thoughts on the future research directions in this field.

Patrick Rey: In summary, industrial organization as a branch of economics has indeed seen a growing emphasis on empirical research. The trend of empirical research in industrial organization started primarily in the United States, with Europe progressing somewhat more slowly in this regard. Nevertheless, it cannot be denied that this trend is gradually emerging in other parts of the world as well. I welcome this trend because the question of which theory is more relevant to reality is an empirical one. Furthermore, I highly appreciate the cooperation between empirical researchers and theoretical researchers. It's important to note that this is a collaborative relationship rather than a competitive one. I personally enjoy collaborating and exchanging ideas with colleagues who conduct empirical research because it helps me better understand specific industries, which, in turn, aids in the construction of theoretical frameworks.

However, I tend to believe that American economists may be leaning a bit too heavily toward empirical trends in industrial organization. Although empirical research is useful and necessary, you cannot let data speak for itself; you also need theoretical guidance. When looking at data, having some theoretical frameworks in your mind is very helpful. This is why I consider data important, but we still need theory. The purpose of building a theoretical framework is to help us better understand what might happen in specific industries. The theoretical research I am discussing is still of an applied nature.

Hu Bo: You mentioned that you like collaborating with empirical economists. What you mean is that having a theoretical framework to guide empirical research is good because theoretical models can guide the direction of empirical research, and at the same time, theory also needs empirical evidence to validate it. Personally, I find collaborating with colleagues who do empirical research very challenging. I think it might be because the mechanisms or intuitions behind most regression analyses are straightforward. If you build a theoretical model based on empirical research results, most of the time the model ends up being redundant because you don't need a model to explain straightforward economic intuitions. Do you have any experiences or advice in this regard?

Patrick Rey: For example, if you want to estimate the demand conditions of a specific industry using the BLP method, you might have good data on industry sales and prices, but you don't have data on product quality and costs. You can construct a theoretical model of how the industry operates, such as the competition mode, whether it's price or quantity competition, and whether there is a dominant firm. You would build different competition models and derive equilibrium conditions regarding costs and demand, which you need to empirically test. Using these equilibrium conditions and data on prices and sales quantities, you can infer potential costs, and the model can help you validate your estimates. Furthermore, if you have many such models, you can find out which model best matches the empirical results. Therefore, empirical research can help you discover which type of model better characterizes a particular industry, and theory can help you address difficulties in observing certain variables. It's a two-way match. If no model matches the empirical evidence well, it might mean that you've overlooked an important variable about the industry. This is what I mean by the interaction of theoretical and empirical research. For example, some of my colleagues use the model framework I developed to study vertical relations, and they need to propose many models. Some models assume price competition among companies, while others assume quality competition, and there are questions about whether pricing is linear or non-linear. So, there's a range of possible models to consider. We examine the combinations of different assumptions, seeking the one that might work.

Host :

Thank you for your guidance. Next, we would like to learn more about your academic journey. How did you embark on the path of industrial organization research?

Patrick Rey: I became interested in economics because it's the science of managing the allocation of scarce resources, and I believe this is crucial for society. I thought that my research in economics could make a real contribution to society. I was also fascinated by how different mechanisms operate, which led me to become a microeconomist rather than a macroeconomist. Building on this, if you ask which area of applied research is more valuable, I believe it should be competition policy. Imagine competition as a game, and markets are where this game unfolds. You need to set the rules of the game, and you also need a reason to enforce these rules. So, what are the useful rules? How do you implement them? I think this is a fantastic research field. If you can do it and improve social welfare, you become more passionate about doing better. I'm particularly interested in the interactions between firms and their responses to incentives, which is why I entered this field.

Host: When did you start studying economics, during graduate school, college, or did you have exposure to economics even earlier?

Patrick Rey: In fact, I was quite uncertain at that time. In France, after high school, you need to take an entrance exam, and then you have to decide what major to choose. At that time, I was not sure about what I wanted to do in the future. Fortunately, in France, we have comprehensive universities, as well as some specialized schools, like engineering schools and business schools. To enter these schools, you need to complete a two-year foundational program. For example, if you want to enter an engineering school, you need to study mathematics and physics, and that's what I chose to do. So, it actually delayed my decision-making. I really liked mathematics, and that allowed me to delay choosing a major and enjoy two pleasant years of studying.

After that, I was fortunate to be admitted to the Ecole Polytechnique in Paris. As the name suggests, the Ecole Polytechnique exposes students to various disciplines. Here, I once again postponed my decision-making and explored different fields. It wasn't until my third year there that I came into contact with economics. I suddenly realized that this was what I was looking for. So, I didn't make any decisions until it became obvious to me. I was lucky in that regard.

Host: You are a highly accomplished economist, but at the same time, you hold positions in many government regulatory agencies. Can you share some of your experiences related to policy-making? Communicating with the personnel in these organizations, I assume, is quite different from presenting academic papers at university seminars.

Patrick Rey: This actually goes back to the point I discussed earlier about the interaction between theoretical and empirical research. I wanted to contribute to social development, so I dedicated myself to theoretical research, hoping to better understand the development of specific industries. If you don't know what the real world is like, how can you understand the key issues in those industries? If you stay in an ivory tower, you might end up building models that are completely irrelevant to the real world, which doesn't help society. So, just as the interaction between empirical and theoretical research is important, you need to be driven by facts from the real world or specific industries to inspire you to build theoretical models. This is why it's crucial to communicate with practitioners. It's not just about communicating with scholars who do empirical research but also with practitioners who truly understand industry developments. These people can be regulators, policy makers in non-competitive agencies, or industry participants. The research today is notably different from what it was 30 years ago. Nowadays, we have more communication with the industry. We have some corporate partners who inform us about the issues they face. We need to engage with them to understand which issues need to be researched and which areas are worth exploring further. As you mentioned, we may need a different way of communicating. It's not like presenting your work in an academic setting. Listening to the views of the industry and regulators is very important; it can inspire and help improve our research.

Host: So, a good model should be inspired by the real world. What do you consider to be a good model?

Patrick Rey: Yes, a good model is first and foremost relevant to the real world. You need to understand the constraints of the real world. In Toulouse, we usually discuss with the industry and policy makers. We organize competition policy workshops and invite industry professionals and policy makers to participate. We use their industry experience to help us understand relevant issues. Second, a good model must be concise and insightful; it must be able to uncover what matters. After interacting with industry professionals, you need to identify what the key points are in your problem and incorporate those key points into your model. The model is concise; it doesn't need anything beyond the key points. Once you include many other irrelevant factors in the model, making it complex, you will quickly get lost. In other words, if the model doesn't truly need certain factors, those factors shouldn't be included in the model. Of course, sometimes, to make the model more realistic, you need to add various constraints. But you still need to start with the simplest model that can derive the core of what you want to express. Once you've built that simplest model, you can further enrich it.

2. Platform Economy

Hu Bo: Some scholars believe that the core feature of a platform is the two-sided market, which seems to be the mainstream research direction. But another possibility is to look at the platform economy from the perspective of vertical relationships, where the platform company is positioned upstream in the industry, and the sellers within the platform company are positioned downstream in the industry. How do you view these two research perspectives?

Patrick Rey: Platforms do play a very important role in the modern economy. You can view a platform as a two-sided market or even a multi-sided market, connecting sellers and buyers, and the platform decides where to provide advertising, where to provide content services, and so on. This is one characteristic of a platform. But it's not the only characteristic. Platforms also exhibit economies of scale, club effects, and network effects, among other things. Take club effects, for example; you join a social platform if your friends and family are already on it. These effects are very important for understanding platform markets.

This has an important impact on how we think about platform regulation. When we construct models of firm competition in our minds, the benchmark is perfect competition. So, when you think about how to design regulations, you want regulations that can make the industry tend towards perfect competition. For platform economy, I think there's a consensus that perfect competition is not a good benchmark model because of increasing returns to scale, network effects, and more. The benchmark of research has disappeared, and therefore you need a new way of thinking about competition, policies, and regulations. You still need to consider the role of competition in the market. For example, even if there's currently only one firm in the industry, if another firm comes up with a better idea or organizes more efficiently, how can it enter the market and replace the incumbent firm that is lagging behind? This is where scholars need to think more deeply, and interesting research may start from here.

Another characteristic of the platform economy is that the lines between competition and cooperation among firms have become blurred. Typically, products, whether they are complementary or substitute, often have a clear boundary. For competition regulators, the first step is to determine whether the relevant market is a substitute market or a complementary market. However, for platforms, this may not even be the first step to consider. If there are already several social software companies in the market, how can a new company enter it? What is the optimal strategy for entry? You need to consider the network effects of incumbent companies. If I offer a product with features similar to the incumbent companies, users have no reason to switch to my product since they can benefit from the network effects of the original product. If I slightly improve the features of the product I provide, I may attract some users. Therefore, for me, the best entry strategy might be to offer a product that is completely different from existing products so that I can attract some users. As I gradually add other features, I can gradually attract more users. Perhaps, in the end, I also introduce social features and eventually surpass the incumbent social software companies. In this example, the main competitors are uncertain, and the closest substitute manufacturers are not necessarily the main competitors. So, I think the boundary between substitute manufacturers and complementary manufacturers has become increasingly blurred, which is another aspect I want to talk about. So, going back to what I said at the beginning, in the platform economy, before, if firms produced substitute products, it was more likely to be a competitive relationship, and if they produced complementary products, it was more likely to be a cooperative relationship. But in the platform economy, this is uncertain.

胡博(Hu Bo): Alright, I'd like to delve deeper into this issue. We know that Amazon is a shopping platform, and on this platform, there are not only other sellers but Amazon itself also provides products. People might think of this as a form of vertical integration. How do you view this perspective? Do you think we can gain some new insights from the perspective of vertical integration? I think people might be concerned about Amazon's behavior; you have user data, you are the platform manager, but at the same time, you are also competing in the product market with other sellers.

Patrick Rey: Yes, I completely agree that this is a form of vertical integration. I'd like to ask further, in the case of Amazon's self-operated business, is Amazon upstream or downstream?

胡博(Hu Bo): My understanding is that if Amazon is just a platform, similar to eBay, then it should be upstream, and the sellers within the platform are downstream because the sellers are closer to consumers. But if Amazon itself also sells products, then Amazon is just like a traditional retailer; therefore, Amazon is both upstream and downstream.

Patrick Rey: This answer is absolutely correct! You mentioned the two roles that Amazon plays in transactions. In fact, there are three roles here, and Amazon is involved in each one of them. First, Amazon is a third-party trading platform, which is the Amazon marketplace. As a platform, Amazon is an upstream company, and the sellers on the platform are downstream companies that use Amazon's marketplace to directly transact with consumers. Second, Amazon is a retailer, which is consistent with the role of a traditional supermarket. Amazon buys products from manufacturers and sells them to consumers. This is the retail part of Amazon's business. So, at this point, we know that Amazon has the platform business as an upstream company and the retail business as a downstream company. But what you didn't mention is that Amazon is actually building its private labels, which reflects vertical integration. In this case, Amazon is a producer and brand creator. Amazon designs its own products and controls the production process. This is different from traditional manufacturers. Amazon is an interesting case. In recent years, there has been a lot of research in this area. For example, we have compared traditional retail models with the agency models commonly used in online platforms like the Amazon marketplace, and we have also studied the impact of vertical integration. However, we haven't been able to look at all three models together, and I think this is an interesting direction for the future.

Section Three: Regulation and Policies

Hu Bo: We are trying to understand why some countries have a more successful development of the platform economy. For example, in the United States and China, there are many highly successful internet companies, whereas in European countries, there are fewer internet companies with global influence. Do you think this is related to the regulations of the European Union, or perhaps it is related to network effects, like Facebook, where if your friends are already using Facebook, there is no need to develop another version of Facebook in Europe? What is your opinion on this?

Patrick Rey: In fact, Europe has some successful platform companies. For instance, BlaBlaCar, a ride-sharing platform, operates in over 30 countries, and it's a French company. There are other examples as well. But when you focus on global market share, it's challenging to find European companies that have achieved that level of success. You mentioned European regulations. This could be a part of the reason. However, in reality, Europe doesn't have a single regulatory policy. Each country within Europe, such as France, Germany, Italy, has its own regulatory policies. So, even though the European Union is working to promote the integration of the goods market, there are different regulatory policies in European countries. As mentioned earlier, scale effects are crucial in the development of the platform economy. Overcoming the differences in regulatory policies between countries is often a significant challenge in Europe. This is why many internet companies are founded and thrive in the United States. In the United States, platform companies can develop and achieve economies of scale, enabling them to operate efficiently. However, the situation is different in Europe. The variations in regulatory policies between European countries are something you can't overlook. It will be a long road for European countries to coordinate their regulatory policies. While there is some consensus on antitrust policies within Europe, each country often pursues its fiscal policies. Europe has many difficulties to overcome if it wants to develop further.

Section Four: Academic Advice

Hu Bo: For young scholars, the early stages of their academic careers can be quite challenging. They've just transitioned from being students to researchers, especially for scholars engaged in theoretical research. What advice do you have for them?

Patrick Rey: I believe it's an exciting time for industrial organization theory researchers right now. As you mentioned earlier, the rise of platforms has introduced new business models. As discussed earlier, platform competition blurs the lines between substitute and complementary effects. Previously, these distinctions were clear-cut, with different models to explain substitute and complementary effects. But suddenly, this no longer holds, so you need to think actively, and this is an interesting development. All competition theories are facing challenges, and we need entirely new ways of thinking. This is a golden age for researchers in industrial organization theory, so come and join us.

Hu Bo: That's very exciting! Earlier, you mentioned the platform economy. Do you have any further insights into what interesting research questions there might be in the field of industrial organization?

Patrick Rey: You need to distinguish between market competition under regulation and unregulated market competition. This relates to the distinction between substitute and complementary goods, but as I mentioned earlier, the line between them is blurring. Some products are substitutes for some users and complements for others. By the way, this isn't a new problem, but it's becoming increasingly important. Let's look at some empirical studies on British supermarkets. We're not discussing platforms right now, but this model is quite similar to what we observe in the platform economy. Some consumers go to a supermarket once a week and buy everything they need for the week. Others buy items from different supermarkets, so they don't shop for everything at the same place. This resembles online shopping, where some customers prefer to buy from one place, and others like to buy different products from different vendors. Going back to the article on supermarket competition, it does an excellent job of answering a crucial question: what is the main source of competition among supermarkets? Is it about retaining customers who shop around? Their research found that the real competition between supermarkets comes from one-stop shoppers, those who are loyal to a particular store. I find this conclusion quite persuasive, and you can apply it to analyzing the platform economy. In platforms, there are both one-stop and multi-stop shoppers. Returning to supermarket competition, if consumer choices between stores depend on their own preferences and prices, the situation can differ. For example, if there are two supermarkets, A and B, and a consumer buys meat from A and bread from B, if A lowers the price of bread, the consumer may start buying both meat and bread from A. At this point, the customer shifts from being a multi-stop shopper to a one-stop shopper at supermarket A. So, as consumption patterns or the way people use online shopping change, products can sometimes shift from being complements to substitutes. This opens up an interesting avenue for research. As I mentioned earlier, this isn't a new problem, and we've done research on supermarkets in the past. It might provide some inspiration for our research in the platform economy and internet user studies. This issue is very important for platform economy research. We need more work to advance our understanding of this issue, including how to effectively regulate platforms in the future.

Hu Bo: Thank you very much for participating in this interview. Your insights have been truly valuable!