This FTC hearing examines how antitrust agencies should evaluate acquisitions of nascent competitors in technology markets, where new firms may have the potential to become significant competitive threats. The presentations discuss the current analytical framework, which includes concepts like perceived potential competition (where incumbents' fears of future entry affect current behavior) and actual potential competition (where future entry would increase competition). Key challenges include defining relevant markets in zero-price digital environments, assessing innovation effects, and determining appropriate probability thresholds for recognizing nascent competitors as market forces. The panelists debate whether existing merger guidelines adequately address these issues or require modification to better protect competition in dynamic digital markets.
Nascent Competition: FTC Hearing on Tech Firm Incentives & Antitrust
Added:okay I think we'll get started we have three panels and two presentations on nascent competition this afternoon to illustrate the importance that the Chairman specifically and the Commission collectively is putting on this issue first we're gonna hear from Susan Athey who was on an earlier panel this morning and as those of you who are here or watching earlier she's economics of Technology professor at Stanford Graduate School of Business she'll do a presentation which will talk a little bit about the incentives business model that a tech firms then we'll turn it over to Paul Denis to do a presentation really describing the current analytical framework for analyzing acquisitions of nascent competitors or arguably acquisitions occurring in nascent nascent markets once they're both done we'll have a panel discussion and I will introduce the panel when we turn to the panel discussion as I think everyone knows if you have questions you either have a card already or you can signal for a card from some of the FTC staff that's circulating among amongst the room and it'll be passed up passed up to us for possible or even likely inclusion in our discussion so with that I will turn it over to Susan who will then turn it over at full great thanks so much for for having me here today and what I would like to do in my remarks is to really connect some of the issues from the morning and the afternoon and talk about some of the reasons that I think they are connected that one of the main reasons I was concerned about vertical manipulation and sort of vertical integration is that in fact that is a nation competition issue for certain types of platforms well and I and another theme that I want to continue from the morning is that in order to understand what the role of regulation should be and understand the welfare consequences of behavior it's important to first understand the strategies of the firms because firms are most likely to do something that's a bit of the antitrust risk as well as potentially harmful for consumer welfare if they're facing certain kinds of threats and so understanding what firms would view as really existential threats is sort of a critical input to knowing where we should be be concerned or at least to understand what's going on so I want to start out just by thinking about how I would talk about this to my business school classes because that's really I think where you start thinking about business strategy and business incentives and so when I teach about platforms to my business school classes I start by by posing some questions they want to think about are these platforms gonna tend towards monopoly um when would you see competition when are they going to be profitable and what I'll really want to focus on today is is it possible for a new startup or a new company to enter and succeed against established incumbents and if so how would you do that and if you understand how you would do that of course you could that also flips back to if you were the incumbent course I always the economists like to take the side of the entrant when when teaching my classes but of course the flip side of the advice for what the introdu is what the incumbent should block and understanding those strategies can help guide policy so we talk a lot about basic platform you cannot mix and the key point there of course is a chicken-and-egg problem if an entrance gonna come in and get started they need to solve that more buyers more sellers more sellers more buyers and also the broader types of economies of scale that are incredibly important and tech companies and it's not just you know having more data which is important but what one thing I really like to focus on is learning by doing the fact that if you're one one of the benefits of being an incumbent is that you have lots of users you can experiment and learn on the user's what works and use their use the data but not just have historical data of any type that exists Oracle experiments that have allowed you to understand how to make your product better when I think about you know more I ask questions about market structure and profitability how do you get started how do you scale and grow and of course for startups a lot of times they're they're actually using platforms initially to get started so startups that I work with or you are advertising on Google to acquire their customers initially they also have questions about how do you how do you grow from there do you expand horizontally or do you go into different verticals so to start with I want to pick an example and actually a mirre also used this example earlier today but it's one that I that I think is useful to go back to from before you know the tech the the modern tech era although this was also in a way a tech case and this is the case of the of American Airlines and Sabre and there's a newspaper article from 1982 and and it's talking about a Justice Department investigation to see if computerized scheduling and reservation services might have been manipulated to give an advantage to American here's some quotes in that same article from the American CEO he starts out by saying well the best guarantee that the Sabre system doesn't advantage American Airlines is that we have to sell them to travel agents so travel agents by them that disciplines us to be fair but then he goes on immediately to say well we're mostly fair but actually we Advantage American Airlines so the fact that travel agents choose put some discipline on the system but not enough to keep them from advantaging America and you know as and then he'd go and then a final point that he makes is well we invested in the Sabre system we built it so therefore you know we shouldn't we should be allowed to use it to advantage our own airline and of course anybody who's been following say the the Google antitrust case will recognize all of these is our same arguments that Google used to talk about their vertical manipulation so when I when we think about this cases and as an economist in some ways I think this case feels a little easier because we we feel like we have like an allergic reaction to somebody not showing customers prices like that feels immediate and harmful and we can immediately map that into the impact on prices but in fact you know when I think about this this case really the thing that that bothers me more about it is the fact that it made it hard for low-cost carriers to come in so if you know more about the history they say well core all the details today but not only did this manipulation soften price competition and mislead consumers but it also in was out alleged led to the exit of low-cost airlines because those low-cost airlines weren't able to compete on price and get the traffic away from American and of course we know that a low-cost airline can enter in some routes gain scale and then grow into a larger competitor and so that that the using your platform to keep someone from getting a toehold and eventually reducing competition is really what what concerns me so if I put that in context of course as we talked about this morning there are lots of things platforms do and in the vertical space that are actually good for consumers and in the interests of consumers so you know Walmart and sumed insurance suppliers are competitive Airbnb make sure that hosts that respond quickly and provide high quality or ranked more highly even search engines demote irrelevant ads and don't show them all of these are things that maybe the advertisers or the suppliers don't like but are to the benefit of consumers and so those are types of vertical behavior that I might not worry as much about so the kinds of things that I'm going to be more concerned about are things that that might be really threats to a dominant platform and so when I started thinking about this one one question I started looking at more than 10 years ago was you know how could a search engine compete a general horizontal search engine compete and so one way that that a smaller search engine can compete is to get really good in a narrow category and of course the chicken and egg problem of platform more users more advertisers more advertisers more users can actually be solved within a vertical and so one way you can compete is to get really good at something so I was advising Microsoft today they made a no they went out and found the very best travel search provider ITA at the time integrated that into the backend of Bing and had better travel services then Google did which then was allowing them to steal share away from Google and the travel vertical and the hope was that from Microsoft's perspective that that would then allow them to grow and get the suit these consumers to do adjacent verticals now what happened was Google bought the ITA travel search engine which ended the Bing relationship and flipped it to where Google was better and travel than Bing was and so this is interesting because you know first of all the the the travel search itself it could have spawned into something completely distinct and not integrated with with either platform but it also was something that helped a very small the Bing was around seven or eight percent market share at the time it wasn't called Bing it could have helped it grow into a a larger general competitor and broadly if you're if you were gonna compete with Amazon or compete with Google or compete with eBay in what they do you know you it would be silly to come in and try to think that you could be better at everything all at once instead generally you might start in a vertical and so they recognize those threats and try to make sure that in a vertical nobody gets too big or too strong another way that this can play out which is very related is that a company like say Amazon can start out they're not actually in search at all they're not in general search they're just doing shopping but over time they acquire more users and then they they enter as an ad platform which then is is more of a competitive threat to Google so we have to worry about these vertical things because that actually is the entry path it's really the main viable entry path to compete with a big platform that has more buyers and more sellers we have to watch for those Nate the the verticals because they are nascent competitors now one challenge in making these kinds of antitrust arguments the Sabre case in some sense would be an easier one to make as an economic expert you could come in and you could show that consumers were getting different prices and the harm would be like very immediate and apparent and I think one of the challenges in in the policy arena is that you know the harm from the vertical manipulation how am I going to show that if you hadn't put this shopping comparison engine out of business they might have grown into the next Amazon in say Europe and then from there they might have grown into a competitive ad platform you know maybe that would have happened and now we see that it did happen when we were talking about it we didn't know whether Amazon would be successful in advertising and we still don't know how big it will be and so as an economic expert to talk about these longer-run impacts I think is really challenging even though that's the welfare effects from that are just so much larger than the short-term effects another way that you can enter is to find to get a lot of traffic from intermediaries and so again in search that was another way that we wanted to get from an unprofitable and never possibly profitable at 7% market share to something bigger so you look out and you say who else has a lot of consumers and maybe I could get those consumers as a block to switch platforms and of course in turn that's a threat to the incumbent platform so just as an example if Google had not it had its own operating system in mobile and hatton made a deal with the iPhone we might have seen very different dynamics in the search business so say the the mobile phones a mobile operating system if it makes a search engine a default can swing very big chunks of traffic all at the same time and so that means two things first of all that intermediary that can shift a whole bunch of consumers in a block risk one is they might shift those to a competitor and risk two is they're just gonna extract all the surplus so Google has to pay billions of dollars a year to Apple in order to make sure that Google is the search default and this also can happen with browsers a very popular browser can put up to auction what's the default search engine and extract a lot of the surplus and so on and so we then if we once we understand that that the the big risk the firm's as any other intermediary who could shift big blocks of take they're they're they're partners with them to another platform we can think about the incentives that the platform has to make that not happen and there's lots of examples of these different what we call them referring services they may not be able to exactly bring their customers with them but they can shift customers from one platform for the other and they typically have a lot of power and are also potentially very ripe for vertical integration that may or may not be beneficial now one another big theme that's important is that when we think about all the things that a company can do to try to to in in the face of a potential threat what we would like a company to do when they when they're faced with a threat from a new entrant is to make a better service or you know in and if it's a it's a vertical competitor they might say ah I'm if I'm worried that shopping is gonna be really big I'm gonna make my own shopping and it's gonna be really awesome and if I even if I don't manipulate everybody's gonna use my awesome shopping service that's one like welfare enhancing thing you might do another welfare enhancing thing you might do is just make your search engine or your platform even better so that even if that shopping engine gets big people still want to use your main central platform but there's also things that you might do that are welfare harming like make or buy an adequate vertical service and then promoted an advantage it to take away customers for the competing vertical service make sure that you own that thing so that if you've now squashed the nation threat of a shopping infant getting big and actually cult Amelie turning into a horizontal competitor I think these types of things are particularly harmful if their innovative services scale driven services or services with network effects so you know things like ad driven news media or review websites these are all examples of things that can be really harmed and and then and therefore never get it get the chance to grow into something that might become a more horizontal competitor now this type of thing also happens in other types of contexts as well so we we haven't talked about ad tech today and it's it's actually kind of hard to talk about cuz there's so much jargon and it's really complicated but just as an example we can think about what what's what what's a way that someone might enter and compete in a in ad exchanges well if you if you have to advertising exchanges where publishers and advertisers buy and sell ads what you and a smaller ad exchange might hope that they can get advertisers and publishers to multi-home and that might give them a chance to grow into a larger Ad Exchange and a more effective competitor and when you think from the publisher side of this market it's actually like in principle pretty tough competition because the ad the ad exchanges are giving the publishers money so that's like perfect substitutes it's just you know which Ad Exchange is gonna give me more money and so if there's two equally sized ad exchanges all the revenue should get go to the publishers because you know I'm gonna add one one exchange says I'll give you you know this many cents per impression and then the other publisher says I'll give you a little bit more and they would cut out their take rate and have to give all the surplus the publishers or else the publishers would just go to the other ad exchange so it's really scary if you're a big Ad Exchange to worry about a small or ad exchange growing because if they go get to their equally-sized this is kind of a zero profit business basically you have to give all the all the all the revenue to the publishers is that my time yep so so so let me finish my let me finish my thought here so the so the so in order to prevent that type of growth that you can use something like a tool a software tool that helps people compare across at exchanges and if it's vertically integrated those tools can actually advantage one ad exchange over the other and so in particular this but this particular practice was ended in 2017 the the software would basically collect the bids from all the other ad exchanges and then give Google a chance to come in at the end and outbid the others just by a penny and get the traffic which helped which prevented the the smaller nation competitor from growing so the the big theme is that in these types of markets we have to look at what are the tactics that are used to try to keep the nation competitors from growing and be very brought very cautious about all the different things that you can do so I'm out of time but I just will will highlight that I that I if you the ways that you compete with different services really are different from business to business and the way that a nascent competitor would compete against a social network may be different than how it competes in a search engine or an operating system or e-commerce and so for each of these industries you need to first think about the business strategy of how a nascent competitor comes in and then look for for exclusionary conduct or other types of squashing or integrations or acquisitions that stop the particularly threatening path thank you so Paul will Paul Dennis will talk a little bit about the current analytical framework a neglected to mention that Paul is a partner at Deckard he was the principal drafter of the 1992 merger guidelines and the relevance that may be clear as he speaks and Paul is the developer of damit which really analyzes how long it takes for the agencies to clear large complicated merger transactions thank you go out thank you for the opportunity to be part of this conference tough act to follow Susan fascinating presentation mine will be much more prosaic focusing on the legal framework but Susan has set it up well in giving you some context for the type of problems that we've devised these legal framework to handle I will focus on the acquisition context I'll stay away from you know the issues of predation or vertical I'll be difficult enough in the time available to get through just talking about the analytical frameworks we have in place for analyzing mergers that involve nascent potential competition issues if you're believer in fate you might actually think I was fated to do this presentation the fall of 1982 I was in Ann Arbor halfway through course of study in law and economics and despite having just finished a lucrative summer clerkship big law firm needed money so I needed a part-time job and we had a visiting professor in that fall guy named Joe Bradley and he was gonna teach my antitrust class so I figured he must be trying to write some papers maybe he needs a research assistant I should go see him about a job so I went to see him and I didn't know much about Joe Bradley at the time but if I'd done even a modicum of research I would have learned that he had written the seminal article on what was called the potential competition doctrine it was a 89 page 300 some odd footnote tome in the Yale Law Journal cumbersome title of potential competition mergers structural thins synthesis was a mouthful but as fate would have it he was working on another article he'd been asked to contribute to a California loss suppose iam on mergers under the then brand new 1982 merger guidelines and his topic was not surprisingly potential competition mergers I'm not sure that he really needed a research assistant to help him with that but he hired me what was supposed to be a job turned into a pay tutorial for which I was the principal beneficiary with my short-term cash flow problem in check I started thinking maybe I need another job next summer so I you know proudly put on my resume that I was the research assistant for the distinguished professor Joe broadly working on this important article on potential competition mergers and signed up for a bunch of interviews with law firms that were doing what I thought was important and I trust work then one of the firms I had targeted Skadden sent two interviewers to campus and one was a tax lair one was an antitrust lawyer it's a mild by luck or design I ended up on the list of the antitrust guy turned out that was Bill palster we just tried the Grand Union case one of the FTC's most important potential competition decisions by the time I interviewed with Bill I knew a little bit about the potential competition doctor not a great deal it was a good thing I did because he spent the whole interview telling war stories about the case and drilling me on various aspects of the doctrine so Joe broadly saved me there bill never really interviewed me but it managed to get me a job anyways I had the benefit of working with Bill and a number of other extraordinary colleagues including a guy who was just about my age well I think all of you know it's Joe Simon's now our FTC chairman a great group of people so why do I tell you these stories well I tell them in part because I thought I could get away with it before Belisle would give me the hook and I could honor some people who I think we're important in my career but well I tell you these stories because they illustrate what I think is a central point we need to focus on this afternoon and these hearings and that's the these concepts of nascent and potential competition really are pervasive in US antitrust merger law and merger enforcement practice it's something well embedded in in our legal framework and the panels that are gonna follow we you know will debate the efficacy of some of those frameworks but I don't think there's any debating that these concepts are well entrenched into our antitrust thinking I'll set the stage for the panel discussion by outlining really just at a high level what some of these concepts are but first before delving into that we'll touch on some definitional issues so that we can hopefully make the discussion a little bit more tractable then after reviewing the framework I'll offer a couple thoughts on ways in which the framework might be filled out and the panel discussion will allow us to dig into those issues and others in in more detail so the terminology in this area is to say not the greatest some of the terminology is not well defined some that is well defined but it's defined in ways that most users do not find to be intuitive at all nascent competition doesn't really have a formal legal definition the word itself implies some degree of competition that's present but maybe not yet fully realized in common usage the term nascent competition is sometimes used to refer to competition that we've yet to see that's incorrect and I concede I've fallen into that usage at times myself so my suggestion for this afternoon as we try to focus on nascent competition being limited to competition that's presently being felt but not yet fully realized used in this sense the acquisition of a nascent competitor by one of its rivals would be seen as extinguishing not only current competition between those firms but also either extinguishing or perhaps amplifying the prospect for greater competition in the future the term potential competition by contrast really is well defined in the law the courts have spent considerable amount of time parsing what's called the potential competition doctrine and they developed the bifurcated concept of potential competition distinguishing between perceived potential competition on the one hand an actual potential competition on the other hand the essence of these two types of competition is not immediately obvious from their names so spend just a minute focusing on you know what the courts have left us in this area seed potential competition is focused on the present competitive effect that's thought to result from incumbents perceptions about the prospects for future entry the acquisition of a perceived potential entrant therefore is thought to lead to a reduction in current competition as that constraint that's being imposed by the threat of future entry is eliminated so the potential for the acquisition to increase competition through the realization of synergies or otherwise is really not part of the perceived potential competition doctrine but as we'll discuss over the course of the afternoon remains part of the the standard antitrust merger analysis actual potential competition by contrast is focused on the future competitive effect that's thought the result from future entry don't ask me why they call it actual them the acquisition of an actual potential entrant doesn't change current competition in any way but it's seen as a matter of concern because it eliminates the increase in future competition that's expected to result from that future entry as with the proceed potential competition the prospect of increased competition from the merger is not part of the doctrine but it remains part of the analysis in practice and part of our discussion so as I said these definitions are not intuitive the most people that probably aren't intuitive to you they never were to me but that's what the courts have given us and my suggestion is that we use that as our guide for this afternoon so that we don't find ourselves talking past each other as I noted at the outset these issues really are pervasive throughout the analysis the impact of nascent and potential competition really begins that first steps of merger analysis when we determine the base price to which we're going to apply the so-called snip the small but significant non transitory increase in price that's applied the hypothetical monopolist paradigm continues through the identification of market participants the assignment of market shares and measurement of market concentration in defining the competitive effect of concern that's where we're most profoundly influenced by notions of nascent potential competition it's here where you know I trust merge analysis takes us into issues like the potential competition doctrine but it was dissatisfaction with that potential competition doctrine that led to the development of other alternatives to the traditional market definition that were thought of Brep perhaps being better ways of incorporating concepts of nascent potential competition into the analysis we could talk about how well we've done with those those are things like innovation markets technology markets or R&D markets entry analysis of course is inherently about potential competition as is efficiency analysis and efficiency analysis doesn't make page here reflecting the short shrift that usually gets in in merger analysis but it is most assuredly part of the consideration of the impact of nascent potential competition the treatment of efficiencies and other forms of dynamic what I call dynamic response such as rapid entry you're committed entry raises a point of practical application that you know I believe warrants further discussion this afternoon well guidelines and analytical frameworks you know generally purports to be burden free and try to avoid giving you relative weights of evidence there has been a decided drift in in how we look at these issues and it's been adrift in the direction of what I regard is somewhat asymmetric treatment of potential competition and nascent competition and that drift was reflected you know first in 2006 DOJ and FTC commentary on the rigid guidelines later in the agency's revision of the 2010 guidelines I think the agencies are properly focused on protecting nascent and potential competition it is something that warrants protection panel discussion will will explore the appropriate scope of that protection but it is certainly warranted in some circumstances where there's been a decided reluctance has been in recognizing or at least fully crediting nascent and potential competition as market forces that can be relied upon to ensure continued competitive performance and markets that are affected by mergers among incumbent firms firms that are well-established in the market that are not either nascent or potential competitors so as the agency sharpened their focus on nascent potential competition my suggestion is that the burdens of proof and evidentiary standards that are imposed on that analysis be imposed in a symmetric way so there were equally likely to consider and recognize credit a nascent of potential combination or potential competitor as a market participant as we are to look at it as a market force of interest in you know traditional horizontal merger analysis having introduced the primary ways in which the issues of nascent potential competition analysis affect our merger analysis let me go into each of them in a little bit more detail and then go into how the framework might be filled out of it so in implementing the hypothetical hypothetical monopolist paradigm and the agencies typically apply this snip to the current market price but nascent potential competition in a market may mean that future prices are going to be quite different than current market prices and it may mean that we can reliably predict those prices to be at a lower level the guidelines recognize this effect and suggest that in those circumstances that anticipated future prices be used for applying the snip so all the things equal using these lower anticipated future prices will lead to definition of more narrow markets the identification of fewer market participants and the recognition of fewer other entrants it's just one way in which nascent and potential competition is slipping into the analysis that most people may not be paying attention to in in the identification of of market participants you know separate and apart from treatment of the benchmark price the guidelines are recognizing that market participants are not limited to firms that are currently producing and selling the relevant product all right the guidelines explicitly recognize that new entrants firms that are committed to entering but you haven't done so will be counted as market participants right they will also include so-called rapid entrance the firms that are expected to be likely to respond to non-competitive performance by supply responses that don't involve the expenditure of significance not cost those firms will also be counted as market participants so the guidelines is already looking at a number of these nascent potential competitors and thinking about ways to include them in the analysis in some sense the guidelines have converted horizontal merger analysis so they're just subsuming some aspects of what we call potential competition analysis both these new entrants and these rapid entrants look and feel a lot like the so-called actual potential entrants which is why I suggest that you know horizontal merger analysis has subsumed a great deal of what was previously thought of as a separate doctrine it's become an issue an objective test really of the timing the likelihood and the sunk costs associated with directory so even after these market participants have been identified nascent potential competition issues come into play in how we assign market shares the shares of the incumbent firms may be discounted based on reliable predictions of the impact of nascent potential competition and because share is a zero-sum game then some portion of the share has to be attributed to those nascent competitors the use of projected shares necessarily affects the measurement of market concentration because market concentration is itself a function of share but even if there's not some quantitative adjustment in shares and therefore some quantitative adjustment and concentration the guidelines recognize that there is a significant qualitative difference in the analysis when an incumbent proposes to acquire potential entrant and in practice you know we see a comfortable adjustment made and analyzing the merger of an incumbent firm with a nascent competitor particularly when we're looking at coordinated affects analysis the nascent competitor is likely to be regarded as a so-called maverick defining the competitive effect of concerns suggested earlier really is the hotbed for the inclusion of nascent potential competition issues in through our analysis this is felt in horizontal merger analysis potential competition analysis you know and in vertical analysis in horizontal analysis most often we talk about price effects the guidelines go beyond price effects because Nason potential competitors may affect product quality I mean effect product variety and may affect the level of innovation in the relevant market I pooped all these together as output effects because output really is the best way of measuring what's going on particularly when you have price and quantity moving simultaneously guidelines have also you know explicitly focused on innovation effects as a competitive effect of concern yeah these innovation effects are increasingly the focus of the agencies in practice what innovation effects are of concern there's a concern about the reduced incentive to continue innovations that may be started by the acquired firm be a reduced incentive to initiate development of new products but there's also you know potentially an increased incentive and ability to innovate that might derive from the combination of complementary capabilities between the incumbent firm and the nascent or potential entrant the guidelines recognize this as well and take it into account in efficiencies analysis here again this asymmetry issue that I mentioned earlier comes into play with agency practice seeming to reflect an expectation that reduced innovations incentives are the more likely outcome resulting from mergers rather than an increase in innovation the source of this asymmetry is perplexing to me because despite all the focus on innovation we do not have a generally applicable theory of innovation that links innovation to mergers or links innovation to market structure economists have written countless models that attempt to predict innovation and within the confines of the assumptions of those models they work but determining which of these countless models to apply in a given real-world situation where the real world situation doesn't conform to the assumption of any of the models precisely remains a dark heart at best the potential competition doctrine is perhaps you know the most focused embodiment of these issues in merger analysis in the u.s. over the years the courts and the Commission have imposed significant but appropriate evidentiary requirements on making out a case on a potential competition doctrine I think that those requirements reflect you know considerable uncertainty we all have over how reliably we can look forward and predict what's going to happen in the future well the Supreme Court's accepted the notion that perceived potential competition states a claim under Section 7 the twice reserved on this issue when thinking about the actual potential competition doctrine I think that's simply the inherent conservativism of the court addressing only issues they absolutely have to address and not doing other issues that they can duck section seven after all is focused on whether the acquisition is likely substantially to lessen competition but left unstated in the statute is lessen competition relative to what in practice we've all filled in the answer there and it's lessen competition relative to what would happen absent the acquisition so that this counterfactual is inherently forward-looking it requires us to consider what competition would be in the future both with and without the acquisition so just as General Dynamics teaches that we have to consider factors that can reliably be said to predict you know diminished future competitive significance for incumbents we need to apply the same sort of thinking to nascent and potential competition and ask where we can reliably predict what the future affects what they see and potential competitors might be again this this symmetry problem comes into play and how you know are going in by C's effect we look at that issue on each side now I'm running well past my time here but we'll try to wrap this up just receive that's a competition doctor and three primary elements here market structure uniqueness and effect right the market has to be structured in such a way that entries likely would have a bro competitive effect normally we look at concentration as being the indicator their uniqueness as a second requirement this acquired company that the potential entrant has to be one of few comparable and potential entrants and finally an effect the prospect of entry by this firm has to actually have an effect that alters income and behavior in some pro competitive sort of way the actual potential competition shares the first two elements with a perceived potential competition doctrine market structure and uniqueness but adds two others I mean the actual potential entrant actually has has to have a plan right there has to be a subjective intent and and objectively the Trier of fact has to conclude that they have the capacity to enter finally that entry has to be likely you know the Commission's b80 decision is probably one of the more focused opinions on this point and unlikelihood requirement Commission applied an elevated standard of proof they had to be clear proof that entry was in fact likely so because those elements of proof for difficult for plaintiffs there arose a number of alternative doctrines to try to get around the potential competition doctrine yet incorporate concepts of nascent potential competition into the analysis you know they are themselves innovations and these are the concepts of innovation markets technology markets and R&D markets give them where we are on time we'll leave the details of these to the discussion on on the panels so filling out the framework where do we need to go with this we we have a well-developed framework it's been applied for years and we'll talk about how well it's been applied but there certainly are points that could be filled out a bit the empirical foundation is probably the first and most obvious so use the term the Chairman Simon's is used you know we have limited empirically grounded economic analysis of the effects of nascent potential competition and this is certainly an area of research success we have professors in the room encourage your best and brightest graduate students to to pursue this part of getting there maybe by doing more in the way of retrospective analysis the Commission has certainly pioneered these efforts in the past we have a much greater need for retrospective analysis of our predictive tools in this area than in other areas this is not a real damning criticism of this area of the law it's true throughout what we do in merger analysis we have a number of tools that we use that we haven't quite tested out yet I mean think about upward pricing pressure where we spend you know a considerable amount of time without having any strong evidentiary foundation for knowing that upper price and pressure theory is actually predictive of what's going to happen in mergers you even have some reasonably good empirical evidence that it's not predictive so this is an area where actual nascent potential competition you know share shares a weakness with horizontal merger analysis and you know a little more in the way of retrospective analysis and testing of our predictive tools is in order probability you know we're talking about two things that are inherently probabilistic events you know a nascent competitor becoming you know more competitive in the future a potential entrant coming into a market in the future these things are not black and white they may happen they may not we don't have a generally accepted threshold of what the probability of success must be before we need to protect this nascent potential competition or before we recognize it as a market force in looking at horizontal merger analysis I think we need one there actually may be a bit of an implicit one that the Commission has already adopted for those of you who are familiar with dammit while mentioned earlier the dexhart and I trust merger investigation timing tracker you know that our firm we have a habit of digging into what the agency is actually doing to try to infer what's going on behind the curtain of nonpublic investigations so with the help of my colleague costum ed Petoskey we took a look at the Commission's enforcement practice in pharmaceutical mergers see whether that might tell us a little something about you know what what is the probability success that you have to have before you're really going to count these nascent potential competitors and we can talk later about the details of this but our preliminary view suggested that the answer might be something north of 60% and we get there by looking at the Commission's enforcement practice involving pipeline branded pharmaceutical products but certainly when you look at the pharmaceutical area and you know I'd seen Mike Boise ups here and he knows this better than anybody Commission has considerable experience as a well-developed reputation for thoughtful enforcement in this area and does not go chasing after low probability events last point here is that the temporal dimension of the analysis you know we don't have clear guidance on the timeframe if you remember back to the 92 guidelines we used to talk about supply responses occurring within a year in response to a snip that lasted a year we talked about entry occurring within two years from initial planning to significant market impact but in the 2010 guidelines we moved away from bright line tests towards something that was far more nebulous we talked about whether rapid entrance would occur in the near future we talked about whether committed entry would be rapid enough whatever rapid enough what happened to me so we lost track of you know a firm temporal dimension for anchoring our consideration of nascent potential competition that that's something that I think need to be revisited in order to give us a little more clarity about how these doctrines are going to be applied when it comes to platform acquisitions and other related concepts so just to sum up you know I think we have an analytical framework in the u.s. that you know provides fairly rich consideration of nascent potential competition issues how well it's considered is something that I hope you'll join us in discussing in the panel discussions to follow thank you Paul so I will now introduce the panel each each member of the panel will will give some opening remarks that I think will focus on you know how robust how complete how sufficient is this analytical framework that Paul laid out and potentially offer alternatives or addition additional considerations that the agency needs to take account of so I've already introduced Susan and Paul Lena Khan is on the panel she is presently a fellow at Columbia University Law School John Newman is an assistant professor at the University of Memphis School of Law bill Rogerson is Charles and Emma Morrison professor of economics at Northwestern University Steve - Dallas was been on a panel earlier in this in this three-day set of hearings is a professor of economics business and public policy at the University of California Berkeley High School of Business and will Tom is partner at Morgan Lewis there's more information on their bios or background on the web right now I'll turn it over to Leena to begin and we'll we'll go right down the right down the road we'll give Susan and Paul a chance to respond and then we'll have each panelist respond to each other and of course we'll take some questions from the from the audience great thank you thank you for the FTC for inviting me into OPP for putting these together and I know it's taken a lot of work so I'm gonna discuss potential competition in the context of digital markets specifically discussing how safeguarding potential competition in these markets is especially important there's been significant debate in the last few years about the growing dominance of a small number of tech platforms and the role they now play as key arteries of Commerce and communications I think a key fissure in that debate is whether any of the dominant platforms are already using their dominance in ways that undermine competition such that it should be within the purview of the antitrust laws I think wherever you fall within that debate steps taken by these firms to solidify their positions through eliminating future challengers should pose a huge concern to everybody that is even if you believe that the current dominance of these firms is nothing to worry about because we're gonna see the you know inexorable forces of creative destruction swoop in and dislodge their dominance that can only be true if tomorrow's innovators are not blocked by today's incumbents so in light of this I think preventing mergers that entrench the positions of leading incumbent tech firms by eliminating future challengers should be a key priority for the antitrust agencies when thinking about potential competition challenges I think there are a few areas that deserve significant attention one is entry barriers so entry barriers are important to this analysis because as we heard from Paul the potential competition framework includes an analysis of whether there are some limits on the entrants that are positioned to enter in digital markets data and analytics capabilities can be a significant barrier to entry some would argue that aggregation of data does not pose a competition problem because data are non rival risks but I think in practice data that is significant for competition purposes might be costly and difficult to obtain so there's gonna be a little incentive to share this is not new to the FTC the FTC recognized that data can serve as a significant entry barrier so in the Nielsen Arbitron case it determined that proprietary data held by the firms would be key inputs for downstream services that were so nascent and the consent decree the consent order included divestiture of certain data assets adding another reason that it's important to consider the role of data as a barrier is that data advantages can be self reinforcing which means that for a new entrant gathering enough data contestant incumbent will be a significant challenge so what does this mean for potential competition analysis think the fact that data does serve as an entry barrier suggests that in many digital markets there will be a significant limit on the potential entrance which is what renders potential competitions analysis so salient and by extension of that because potential competitors are less likely to emerge it is especially crucial that when they do emerge antitrust safeguards that potential competition so second challenges is what I call the another problem so Nava is a company that Facebook acquired in 2013 it's a VPN provider that grants users heightened security but it also allows Facebook to track an extremely close detail which rival apps are diverting attention from Facebook which means the facebooking detect at the very earliest stages of a company's growth which competing apps might pose competitive threats this information then shapes Facebook's acquisition strategy and enables it to purchase apps at very earliest stages such as tbh and moves which are presumably identified as as quickly growing so novo is one example of the significant information asymmetries that exists between platforming intermediaries and enforcers for whom the competitive significance of a deal may be less obvious if it doesn't have access to the same granular information about the usage level of rival apps and so I think this means that agencies should be more willing to issue second requests for even seemingly small acquisitions and then make sure that information collected and second requests include competitive intelligence gathered through devices like Inaba I think it also means that there may be acquisitions that don't significantly undermine competition in the relevant market but that do structurally position the incumbent to detect nascent rivals much earlier information that they can then go and use out to to make early acquisitions and so I think these these acquisitions that don't affect the relevant market but do structurally improve the position of an incumbent to make early acquisitions is something that should also be relevant to the agencies I want to quickly address two arguments that are sometimes made to caution against aggressive enforcement in these markets so one is the idea that there's a risk of short-term immediate consumer harm given that an incumbent that acquires a start-up may offer an offer the quickest path to market I think that's a very reasonable consideration I think business literature and experience suggests that while incumbents may be more successful at delivering innovation that continues on established research paths it's really the startups and the new firms that are more likely to account for the truly breakthrough paradigm shifting innovations this is for at least two reasons one that incumbents may not be eager to invest in innovations that are likely to lose them value on their existing investments and the second is that even in instances when it's incumbents that are making the breakthrough in innovations in order for them to do that they need some outside competitive pressure so even if there is some short-term consumer harm here I think we need to be careful about weighing that against the long term gains and innovation the second argument that sometimes made is that aggressive enforcement could result in a negative effect on the capital markets for startups if acquisition by an ax company is an exit strategy that's motivating startup funding then limiting these acquisitions could lead to a few or startups I think that's also a very fair argument but it's somewhat incomplete I think new research by Ian Hathaway and how singer shows that venture capital first financing and seed stage activities contracting more rapidly or growing more slowly in sectors where say Facebook Google and Amazon are likely to enter corroborating the Killzone story that we've heard reported other places so I think concerns about declines in venture capital are very legitimate there so far there are other sources of that decline and I guess I'll close by an identifying a few paths forward for the agencies that could help address concerns about potential competition so one is as Paul also mentioned more merger retrospectives merger enforcement of course is rife with uncertainty and as merger enforcement has become more more fact-specific oftentimes the information this most relevant won't be available for a few years after the mergers been consummated and if that's the case I think it makes sense to do more merger retrospectives to identify when acquisitions did in fact stifle important competition and learn from that and consider undoing those mergers and the second is to review acquisitions by monopolistic firms as potential section 2 violations so this is something that former Commissioner McSweeney also proposed and there's precedent for this so in 2017 the FTC challenged the Mallinckrodt ADR Zack juez ition of certain assets from Novartis under Section 2 on the theory that this acquisition was a defensive move to extinguish a nascent competitive threat to its monopoly the settlement therefore involved a license to third parties to develop the relevant sets and so this didn't involve a digital market but I think it is a good model for how the agencies might consider evaluating acquisitions by by dominant firms um so I'll close there I think generally you know antitrust has been haunted by this fear of false positives and I think in in the context of potential competition we should be rebalancing towards more comfort with false positives with the recognition that often times that's necessary in order to prevent false negatives you can applaud I didn't thank you well we'll turn to John yeah all right so I'm gonna start off by thank you making a claim that I think would have been really uncontroversial five years ago it's starting to feel a little risque though these days and that is that our current basic legal framework at least as applied to sort of core issues like horizontal mergers and acquisitions is not totally broken so how do i how do i base that claim i'd say that we can look at the types of cases that agencies have been bringing i'm gonna focus on digital markets since a lot of the talk about nascent competition has to do with digital spaces the agencies have been bringing a surprisingly broad variety of cases different types of cases in different digital markets along a spectrum right so you had FanDuel draftkings where you there were two actual competitors that were merging you had cdk automate where you had an actual rival merging with a sort of nascent rival and then he had nielsen Arbitron which was a combination of two potential competitors neither of which competed in the relevant market of concern at the time of the challenge none of these cases attracted a deluge of criticism so to me that tells me that tells me two things the agency is being active and they're not making egregious mistakes so if the basic basic legal framework seems to be functioning fairly well are there nonetheless some sorts of areas that are voids in the current enforcement regime and another way to put that question is to say what's prompting all the calls that we're seeing for stronger antitrust enforcement in digital market to me there is a big gaping void in the current framework and that has to do with cereal price markets so none of those cases I just mentioned involved zero price markets despite the near ubiquity of that model at least with consumer facing platforms so none of these case involves a zero price market a lot of people have responded to that void and enforcement by urging a focus on data extraction or Big Data to me that's misguided for a bunch of reasons the biggest of which though is that data extraction is just a really messy thing from a consumer perspective so a lot of data extraction when it's used internally by a firm to improve the product is good for consumers and the concerning usage that is you know extracting data and then selling it to third parties who then target consumers with advertisements or using it to target advertisements that really reflects derived demand so the demand for data here is derived from the demand for advertising right the demand for attention so to me the the core area of concern and where there seems to be a void in current enforcement structures and efforts lies around attention markets attention competition so my proposal is that where we see acquisitions of direct rivals even if they're nascent we should be looking harder to protect attention rivalry cases in the past that seemed to in retrospect maybe represent false negatives what include mergers like Facebook Instagram Google Waze Zillow Trulia cases where attention rivalry is at the core of the merger so that that immediately runs into I think a laundry list of objections a couple of them were able addressed by Lina a couple more objection number one I'll see if I can respond to this so all digital firms compete for our eyeballs so they all operate from a tension perspective in one big market that's really unconcentrated I think that totally misses the mark so attention is the current see in these markets so the objection is essentially the exact equivalent of saying all firms compete for money so there must be one big market that's really unconcentrated just laughable second objection we already look at harm to innovation on the user side the zero price side the attention rivalry side of a lot of platforms this is true or seems to be true at least if you look at something like Zillow Trulia the agency reported that it looked for harm to innovation on the user side of the platform is that enough though is that sufficient to allay all concerns or find all concerns I don't think so unfortunately we don't have as we've heard today a really strong economic theory or it econometrics tools that we can use to assess innovation effects in a merger context and so worsen the problem the qualitative evidence that we would usually rely on in the absence of an icicle economic theory is not likely going to be there so if we think about what we'd be looking forward to maybe a board presentation right where a CEO is trying to sell an acquisition to the board by saying oh after the merger we're not going to innovate anymore never gonna happen so the types of qualitative evidence we would need isn't going to be there I think in a lot of cases so focusing solely on harm to innovation in these contexts it's not going to be enough to catch all the potentially on a competitive effects finally I think and this is maybe the most salient or the most trenchant criticism of the idea of regulating attention in markets in a more interventionist way is that market definition market effects are going to be really hard to measure here we lack prices and that's our favorite tool to use here I think there actually will be useful qualitative evidence that we could use so if you go back and look at investor statements from zillow trulia to use one example post acquisition you've got the CEO of Zillow group saying in an investor call publicly available call that now we've got seventy plus percent of the market for online real estate portals that's pretty compelling it's qualitative right but it's pretty compelling stuff and if that's the kind of stuff that's available publicly one can only imagine what's being said in turn second quantitative evidence and here's where I'd like to sort of urge the FTC to look harder in a specific way with digital firms in particular there is often a great deal of a B testing that goes on in the marketplace it's really easy to do that in a digital market to the extent that the FTC could ask for the results of a B testing when firms are changing advertising loads and looking at the competitors also that I think that would be a fantastic idea and would help give us an idea of what happens in a digital market when a dominant firm increases advertising loads to users where does that substitution go - it could be a really useful tool for us to use okay thank thank you John to Bill Rogers great well thank you very much for having me here today and thanks for it's such a great job organizing this very timely conference my understanding is the first panel is supposed to focus on two questions first what framework should we use to evaluate mergers between incumbents and nascent competitors in high-tech industry and second is the current law and the current version of the merger guidelines consistent with using the correct framework or with the law or the guidelines have to be changed or altered in order to use the correct framework I'm going to focus more on the first question of what the correct framework should be because I think this isn't a question a question that economics can shed the most light on and I'm an economist however let me very briefly say regarding the second question that I agree with a lot of what has been said earlier that I view the law and the merger guidelines as relatively general statements flexible general statements of general principles that are pretty sensible and my own non-lawyer view of it is is that one could interpret existing law the existing merger guidelines as being completely consistent with a sensible and correct welfare analysis of the problem so I don't think there's any need to change the law or even to change the merger guidelines I think all the right principles are in place if I was on the next panel that's supposed to address should we get a little tougher on any mergers or have we been looking at these mergers as carefully as we ought to using the correct framework I probably say given the events I've seen in the last five years looking back on them that possibly regulators should be a little tougher or look more closely at a number of mergers and perhaps there have been a number of mergers that have been approved that at least in retrospect with hindsight one wonders why they were approved or whether they really should have been or whether there was enough evidence at the time to decide that perhaps that you shouldn't go forward with but I don't think there's really any need for a change in the law and I think the the basic elements of the correct framework are in place what I'm going to do today is focus on one specific issue related to what the right approach is and what I'm going to do is describe two different social welfare problems that you could we could attempt to solve when we were choosing a policy on how to evaluate these types of mergers okay the first policy is going to be not quite correct and be extremely complicated the second policy is going to be exactly correct and fantastically more complicated than the extremely complicated policy and one message I want to leave you with today is is that I wonder whether or not we should be considering using the not quite correct policy because it's a type of question that type of wealth or evaluation although it's extremely complicated I could imagine evidence being brought to it and evaluated whereas the second perfectly correct question I'm not so sure this is the case okay so let me start just as a benchmark with describing what I think the standard competition problem is when to mature firms come to the DOJ or the FTC and say we'd like to merge okay because they're you know and of course we really just look at the competitor I and assess the competitive effects of the merger see to what extent it would cause competition would be reduced and price would go up quality would go down or perhaps even look at innovation effects on the one hand and then on the other hand ask if there are compensating efficiencies and try and do some sensible analysis as best we can to predict whether or not consumer welfare would be higher or lower with the merger okay now what changes when one mature firm and a nascent competitor come to the agencies and say we'd like to merge so we have a big incumbent who wants to merge with a start-up that has created a new idea so it's already been in the market for a while thinking doing innovation it's created the idea for a new product perhaps it's already begun to introduce it to the market but there's still a lot of uncertainty maybe the product is still going to evolve it isn't quite clear how consumers will use it or how many consumers will want to use it or whether they'll view it as a substitute for the incumbents product or not or to what extent they'll view it as a substitute so there's a huge amount of uncertainty about exactly how this nascent product will fit into the market well I think there are two different welfare problems you could consider when you are asking should we approve this merger okay the first one I'm going to call the ex post problem and the reason I call it the ex post problem is I'm going to say let's take as given that we have that startup today who's already done all of the start-up innovate innovate that he did he's already come up with a new idea tried it out a little bit in practice and he's ready to go or you know he's nascent but he's ready you know he's done a lot of work already he's already done this preliminary innovation okay well at that point taking that as given the competition problem we're looking at with the mature firm and the nascent firm is very similar to the problem we look at with them two mature firms only there's a lot more uncertainty about exactly what the competitive effects are and what the nature of the efficiencies are right but but in principle it's the same type of problem we've got two firms we want to assess the competitive effects it's going to be harder to do this because it's more forward-looking relying on more more predictions we're not quite sure if the start you know what number one would the startup succeed if it was by itself or really has it only come up with an idea that would be useful for an incumbent firm and perhaps it was never even really trying to think of a standalone idea really perhaps the efficient way to organize innovation in this industry was simply that little startups think of new ideas that existing incumbents are better at implementing so that could be what's going on or or in the other hand it could well be an idea that could grow into a new product that really would challenge the rival okay but it's uncertain so if we went ahead and did this analysis we still try and determine whether consumer surplus will be higher or lower with or without the merger but it's good there's going to be more uncertainty so it's going to be a much harder problem and maybe harder to draw the conclusion that the merger will be bad on the on the one hand if we allow the merger the incumbent firm will get this technology and use it in some sense although we don't even really know how the incumbent firm will use it okay if we disallow the merger it's possible the startup will just go down the drain or it's possible the startup survives the product still won't be nearly as good as if the talented incumbent who knows how to implement a new product had taken it over on the other hand the startup might succeed beautifully and we not only have the product available to consumers but we'd have more competition and everyone would be better off so there's a lot of uncertainties but nonetheless it's fundamentally the same problem two firms have walked up to you and you have to say we'll consumer surplus be higher if I allow the merger or I don't allow the merger okay so that's what I'm calling the expose problem it's a slightly more complicated version of the to mature firms problem okay but the point I want to leave you with or the point I want to stress here is that expose problem isn't necessarily the right the theoretically right problem to be considering hey why is that what is the ex post problem ignoring well the ex post problem is a problem is a insane I already have this startup now should I let him be bought or not he's already done his innovation and now what should I do with them okay if you were really setting a merger policy in the real world surely you'd like to take or potentially you'd want to take into account the fact that when I choose a merger policy I'm going to make startup innovation more or less profitable for startups in particular if I make it easier for incumbents to buy startups start doing startup innovation will be more profitable and there's likely to be more startup innovation and in fact if it turns out that this is one of those industries where the efficient way to organize innovation is to let little guys think of new ideas and then big guys implement them I might be getting in the way of just the efficient way of doing R&D and innovation in this industry if I started walking a lot of these mergers because they wouldn't do them and the startups wouldn't do them in the first place if they couldn't if they couldn't sell their product to the incumbent well I could consider that problem too and I could call that the Exxon a problem right looking at the effects of the policy before startups have done their innovation now I still want to answer the first question of given this merger with this startup will consumer surplus go up or down if I allow the merger but that might not be the end right if I want if and I think although I haven't yet worked this out in a formal model if we wrote down the formal model in a well behaved model we would predict that the fully optimal solution might be to be a little more lenient on mergers than the one that just implemented perfectly efficient expose policies and that's because innovation is generally good for consumers so it might be desirable at least in theory to commit to a policy where you purposely commit to approving some mergers that are inefficient ex-post in order to create better innovation incentives Exxon 10 now the second problem the Exxon Tay problem is the perfectly correct problem but it's a complete order of magnitude harder than the ex-post problem which is already a complete order of magnitude harder than the standard problem okay usually when I hear experts on existing antitrust laws kind of get into the details of how they would analyze an actual merger between an actual incumbent and an actual startup they do some version of what I would call describing the ex-post problem that is I think they try and ask the question this is hard to do but I'm going to ask would consumers be better off if I allowed this merger and they really don't consider the issue of what effect will this have on innovation and senta's of startups going forward I think that might be the right idea I'm not sure but I want to submit to you that this might not be a bad idea for two different reasons I think that it's possible that this expose problem is a problem that's simple enough that courts could actually evaluate it and if you ask courts to evaluate problems where there's just going to be completely no factual basis for arriving at any sort of possible reasonable conclusion you're just inviting them to give their own opinion and so it might be best to restrict us to a fairly good question that they really can potentially answer that's going to rely on some objective facts that can be presented before the court the second thing is I'm not sure the two problems would necessarily yield that different announcer anyhow I think if you've got a correct ex-post policy you're still going to allow plenty of mergers where the mergers would have never had any hope really of watching a separate firm they're really just little ideas that the incumbent would have used anyhow and if you apply that policy sensibly I think you're going to allow a lot of merger a lot of these mergers and there still will be good merger incentives and secondly there's a countervailing effect if you loosen up your merger policy to get the startups to invest more the incumbent is going to invest less and that's going to be bad so there's a countervailing effect if you try and loosen up policy away from the efficient policy trying to get more mergers to have more innovation it just may be that the incumbent will frustrate you by investing less so I'm Dallas I'm not sure that the slightly easier problem isn't that bad a problem in any event and at a minimum I think it's a problem that courts could potentially address maybe the way this really would work out in practice often in real cases people always talk about what's the probability that we have to show that the startup would succeed how high does that probability have to be how certainly do we have to be that the murder the firm would survive by itself and actually be a good competitor right maybe in a in a theoretical world you could think of well there'd be a level to set that probability off that produced efficient decisions exposed sufficient decisions and maybe you want to move it around a tiny bit if you were trying to do this fully optimal problem even though no one knows how to do that okay and I might imagine that the real problem that courts and and enforcers will always think of themselves as solving is they take that probability is given whatever it you know it seems to be given the case law and then they just try and investigate whether the merger is ex-post efficient or not overtime through some mysterious process that lawyers know about courts maybe end up doing the right thing even though it's hard to calculate what that is I have no idea but I would submit that it might make sense for us to focus on this slightly simpler incorrect problem even though it is incorrect think thank you Bill I think this is the second panel I've moderated and I think what people may discover is I'm not a very good disciplinarian I I've envisioned this panel as sort of setting up the the next two panels so I have allowed people to go a little bit over allocated time but one thing I do want to do is is after Steve and and will give Paul and Susan a chance to respond so just keep that in mind as you as you do your comments so we'll turn it over to Steve and then thank you below and and thanks to the FTC for having this I'm also grateful that I'm actually speaking because the way things were going that wasn't certain like bill I do believe that the current welfare analysis is the preferred tool not surprisingly I'm an economist but I think these tools need to be carefully applied because one thing that isn't said enough is that not all platforms are created equally they don't all have the same business model they don't all have the same barriers to entry so there's a lot of discussion that people seem to lump together I heard this earlier oh the Google Facebook Amazon Apple right these are all different companies different business models and the Devils in the details and and and I think that's something that is not said enough which is why I wanted to start with that now now going to a theoretical a broad theoretical perspective there's no question that nascent competition is something we appreciate because of two main influences it has one it keeps companies a check in terms of pricing and second it's a great source of innovation now the first is non-controversial you know without competition dominant firms might take advantage of the market price higher produce less quantity that's an easy one the innovation thing is not as easy so going back to 1962 Ken arrow a very celebrated economist suggested that competitive markets are really what is needed for innovation the reason being that in a competitive market if you managed to innovate get a slight cost advantage or a slight quality quality advantage then you're going to gain a lot of market share from the rest of the competitors and that gives you a tremendous amount of incentives to innovate 20 years earlier Joseph Schumpeter and other celebrated economists said something very different that in the long run competitive Aconitum markets do not provide true returns or super normal returns and hence there's not much reason to innovate you actually need market power in order to get the gains from innovation and to have those incentives in place and there have been a lot of studies not enough maybe to try to tease this apart a recent very celebrated article in the quarterly quarterly Journal of economics which is one of the leading journals in economics suggested that there's an inverse u-shape relationship between competition innovation namely some kind of Goldilocks story too little is not good too much is not good some healthy middle peers now I was very happy that Paul mentioned we should encourage our graduate students to work on this topic as it so happens one of my graduate students who is on the job market this year has written a beautiful paper where he did something akin to retrospective analysis which is not easy to do spent a year and a half gathering data where he took data from the deal J's a cartel break up history which clearly was an exogenous shock to a competition because one of the things we worry about and the reason empirical work is so difficult here is that competition and innovation are both determined not only by the relationship between them but by what we call latent or lurking variables and it's really hard to tease that causation correlation story so what he did he went back 30 years collected data of breakups in different markets define the market carefully treatment control obviously like any study there are some assumptions but what he showed is actually that more competition creates less innovation measured by patent investment filings by patent breath and by R&D investments so I think the verdict is out about what is the right amount of competition and by extension nascent competition in order to get innovation from the theory side now let me turn a bit to practice because in theory there's no difference between theory and practice but in practice there is and and I was inspired by Susan and and a handful of other economists who actually spend time in industry I spent two years at eBay building and leading a team of economists I spent a year at Amazon also leading a team of economists and and kind of you know seeing how how things actually work and their relationship with startups and and innovation more broadly now startups are the source of this nascent competition that we're talking about here and the reason startups are created is because the founders and the people who invest in them believe that they will get returns in the future no returns no investment that's kind of straightforward and startups are uncertain you know again echoing Senator Bill said there's a lot of uncertainty in these innovative endeavors so we need to focus on what differential success from failures when we think about startups who engage in investment so first of all it could be bad products now by and large we believe that venture capital financing and other financing are gonna be a pretty strong gateway you know stupid product I'm not going to give you money of course think there are no sand you might think differently but by and large that is one mechanism in place you might have a good product you start investing you need to acquire customers that's something that Susan spoke about at length if you don't have enough money to engage in marketing to get your customers or the marketing costs are a lot more than you thought they would be you might burn all your investment capital and then die not because you have a bad product but because you didn't manage to get that early start the so-called Chicot chicken-and-egg problem last but not least is poor execution and I can't stress enough how many companies fail because of poor execution something that as an academic I never appreciate it it's like oh here's the model in theory it works what's the big deal well again in practice things are very different and this is precisely where acquisition exits have a tremendous amount of value again I'm echoing something that bill said because execution is so difficult it is those large companies who succeeded time and again who have put together the apparatus that helps with execution and that is complimentary to the success of many of these startups so if we go to that question kind of like that ex-post idea that bill promoted we have to ask ourselves if we allow this merger to happen what will probabilistically happen in the future and are there real barriers to entry for future competition that might be a foreclosed if we allow certain mergers to happen and in platform markets which is really what we're talking about here those barriers to entry are primarily about indirect network effects and those are going to be a barrier if one multihoming is costly and again Susan talked about multihoming a lot and to acquiring new customers is difficult which again Susan mentioned and these are tightly connected because if multihoming is easy acquiring customers is easy so my observations from my experience on the tech sector more broadly but especially retail marketplaces since everybody else ignored the times up sign hi I might do that too I know then then first of all for many products and services multihoming requires two or three clicks and two queries so whenever I choose to buy something within less than two minutes I could compare Amazon eBay and Walmart and I often do if it's a more expensive product and truth be told if it's a twelve dollar product I'm not going to bother because if I'm screwed by 40 cents that's okay I could live with that and again every time I do those comparisons I find the prices to be extremely similar second every time I take a ride I have uber and lyft on my phone right next to each other in 45 seconds I will have that price comparison and time comparison sometimes I'm in more of a hurry and I will pay a dollar more to get there faster sometimes I'll rather save that dollar and wait another few minutes then the second thing is that early stage entry has become extremely cheap and very easy to do precisely because of a lot of platforms that came up like cloud computing services what used to be a capital expenditure buying millions of dollars of servers is now pay-as-you-go computing and storage and that makes the early stages of entry very very easy and and here I'm echoing something that Lena said barriers to entry are really critical to look out look at and for startups in the tech sector barriers to entry in early stages have declined dramatically last but not least VC funding is really thirsty for potential entrants one example is jet comm four years ago the company was founded they raised quark close to a billion dollars in venture funding and shortly after that they were bought by walmart.com and now they are driving Walmart's most of Walmart's a platform so so going to another point that Lina mentioned and and she quoted how singer from the previous panel who mentioned this decline in decline in VC funding well there is a study by Oliver Wayman albeit funded by Facebook so full disclosure that's what I read but it shows that VC funding is at record high what has changed is where the VC funding is coming in rather than coming in at early stages it is now coming in at later stages of investment well if you think about the reduction in barriers to entry to start a start-up that makes complete sense that is a market reaction easy to enter hard to execute so VC is coming into that later execution phase so I'm going to conclude with with very quick quickly with three points so first I am convinced that the current tools guided by solid economic thinking and and and and those of guide empirical analysis are adequate to deal with the topic that we're talking about today second as we move forward I think we really have to be careful to do things on a case-by-case basis there is no one-size-fits-all tool or application and and and and in that respect and and this is again something that Lina mentioned I am a huge fan of retrospective analysis and I wish we did enough we did more of it I don't think we're doing enough of that retrospective analysis government agencies that have amazing data when they evaluate mergers could be a wonderful source for that kind of analysis last but not least I think that evidence suggests that in these so-called platform markets entry barriers are low multihoming is easy nascent competition is not under thread by these acquisitions I think on the contrary acquisitions help spur execution which then lead to more opportunities for innovation thank you thank you thank you Steve so let's start to will thanks but I'll I'm gonna confine myself to two categories of comments one on the nature of the tools and the second on the management of the jewels so on the nature of the tools I fully agree with Steven that the tools are adequate I think the agencies have a very full toolbox in fact I think the toolbox is kind of overflowing and therein lies the risk so I think the technical term for what has happened to the toolbox over the last couple of decades as the tools have evolved in response to changing circumstances is that the tools have become squishier so take for example the temporal issues that Paul identified in terms of the one year to year kind of thing and the much more qualitative kinds of measures that found their way into the 2010 guidelines I think that was an effort to avoid some of the inaccuracies of kind of progress see and Med of of bright-line rules about time periods but you see the analog in a lot of areas besides time periods for example market share so market share and vertical theories and I think that the first time this really became clear to me was in the time warner turner merger that the FTC handled more than a decade ago and the notion that that you see in the analysis to a public comment is that the share of foreclosure is a function of what is actually needed by the complementor and so you know there the theory was the barriers that would be posed by new programming entities by control of more of the conduit the notion that you'll see in the analysis there was to launch a significant new program or programming network you needed to be able to reach about 60 percent of the subscribers nationwide well yeah the the the inverse of sixty percent is forty percent if you can foreclose forty percent that's enough you don't need the traditional monopoly share or anything like that and so instead of a bright line you know suddenly it was a measurement that depended on the circumstances in the case and the competitive theory involved okay so as these tools have gotten squishier the risk of misunderstanding has increased and one example near and dear to my heart is innovation markets which the term which the 2017 IP guidelines has finally gotten rid of and and probably a good thing even though there was nothing wrong with the underlying concept it's just that as interpreted and applied and nobody seemed to understand what that underlying concept was the concept came about as a direct result of the GM's @f merger in which you had two companies that barely competed at all in the downstream goods markets just as a geographic matter but it happened that to innovate in this product market you've needed a massive amount of manufacturing facilities there was an iterative iterative process between the manufacturing and the innovation and these were the only two companies that had it so even though they didn't compete much downstream they competed heavily in innovation and benefits of that leapfrogging competition was felt worldwide even in markets in which they did not compete in the goods markets and so the focus of the innovation market theory was on specialized assets so if you've had a type of innovation that anybody in his garage could come up with you really didn't worry too much about the restraint on innovation that came about from a merger because you didn't know where the next breakthrough was gonna come from and so it didn't make sense for the antitrust enforces to worry about that whole case hinged on the specialized nature of the assets needed to innovate and the fact that very very few players had it so that that's the one set of issues that have that poses an example of of the danger of the risk of misunderstanding as these concepts become more elastic I think you know currently this notion of acquisitions of data may be another area that we really ought to think hard about before we leap at latest and shiniest theory I don't think these are really zero price markets I mean you know for the most part when you're looking at murders of two companies with significant cases of data you're really talking about data acquisitions as input purchases they're getting data those data are useful to them in their business and helpful to them as they can be downstream and they are paying a price to get that data it's not a monetary price they're paying in the form of offering consumers services so you know the CEO joke that if you're paying a zero price you're not the consumer you're the product and and I think that that has a lot of truth in how we ought to think about these murders okay so another point about the nature of the tools is that as these tools become squishy or if you will the complexity of the trade-offs has increased and you know the asymmetries that Paul pointed to in terms of when we look at potential entry as a potential anti-competitive harm and when we look at a potential entry as curing a harm from the merger they're not always symmetric and I think in some sense that they don't need to be what you're in a broader sense what you're looking at is a decision theory framework right what are the what are the consequences of being wrong and the probabilities of being wrong and the administrative costs of getting it writer in in both dimensions and you make those trade-offs against each other and the probabilities don't necessarily have to balance if the consequences don't balance and so I think it's very much a case-by-case and you know maybe maybe a great example of how granular those case-by-case determinations might be you know I I was always an admirer of chairman nurses handling of the Genzyme case in which you really drill down into you know what are the what are the incentives affecting the behavior of the CEO of the acquiring company and you know sometimes I think you have to do that there are harms that we have to be cognized cognizing of in in these kinds of areas where we're looking at nascent competition and to reach to reach for a somewhat old and old technology example Lille separate court where you know again more than a dozen years ago and Commissioner Anthony made some species about this one you had a branded pharmaceutical company acquiring a an isomer or a company that had an isomer of its product it appeared that the isomer was a product improvement some argued against a merger on theory that separate core was the greatest potential competitive threat to the Lille the problem was that the the this nascent competitive threat you know by law would not be able to market this nascent competitive threat for four years because it clearly infringed Lily's basic patent and so you know I think there was a case where both the probabilities and the potential consequences weighed clearly in favor of letting the merger go through because you know the probabilities were you know weighing the near-term benefit of the greater execution and the removal of the blocking position against this somewhat theoretical benefit of the future competition or four years from now and the consequences of that trade-off was benefits to patients now health benefits versus you know some theoretical of benefit of price competitions on the years gone the road so I think you need to make those trade-offs carefully okay let me just turn very quickly to a couple of words about management when I was at the Commission there was a commissioner who was a very fond of behavioral economics and was an advocate of applying it and at every possible turn and I could understand how it applied to our consumer protection mission and I never could quite get how it applied to our competition mission until it hit me one day that if you turn the telescope around and traded on the enforcers behavioral economics would tell you quite a bit about the cognitive biases of the people within the building who are doing these investigations and making the case recommendations and voting on the case recommendations when it struck me it seemed like a brilliant insight until just a couple of days ago and I was searching the web for something else and I stumble across a really nice article written by Bill kvass Ike and Jim Cooper that made exactly at that point so there's nothing new Under the Sun but we should pay attention to the fact that narratives are powerful they're especially attractive when they seem like novel insights you know my feeling about or my belief that gee I had this brilliant idea about behavioral economics might be might be an example of that where you weigh much more heavily things that are novel and exciting and fun and you stick to them even if a sober analysis of the potential consequences might lead you in the in the other direction so that's point one on management a second point is adding attitudinal approaches former acting chairman ole Hassan has spoken a lot about regulatory humility and that's that that I think scratches the surface of what enforcers might think about as they approach some of these new complexities I also think we should think about some process approaches to managing these tools one thing that always struck me was there there has been an unwritten rule that's grown up at the FTC about the number of meetings you get with the Bureau director and it's colloquially referred to as one meeting rule and I think it makes a lot of sense in traditional horizontal mergers where the volume is such that if you allowed more than one meeting you'd never do anything else and the analytical paths are straightforward enough that additional meetings wouldn't help very much but in novel and cutting-edge areas I I would really like to see that unwritten rule abolish I think there's nothing worse than having the staff go down a particular route or you know pursue a particular hobby horse maybe driven by some of these cognitive biases I was talking about and work for months and months or or years or the matter alway to when management finally pays attention and find that there are whole dimensions of the issue that they've overlooked so one meeting rule devil's advocates I think can be very useful and maybe I ought to be formalized people have thought about retrospectives and I heartily endorse that although they're highly resource intensive and so you know all these things I think become ever more important as these tools will become just a little more ephemeral and difficult to to follow clear and well-traveled roads Thanks okay well I think we are actually at the time so I'm not gonna eat into the next panels time so we're gonna take a 10-minute break and we'll be back here with the second panel on some competition [Applause]
Up Next
Two-Sided Markets Theory: Pricing Strategies and Network Effects
@geostadt578
2.3K views•2020-09-02
Mundell-Fleming Model: Negative Goods Market Shock Explained
@Inlecture
831 views•2020-05-07
Behavioral Economics Explained: Rationality, Nudges, and Risk
@crashcourse
1.1M views•2016-03-12
The Age of Easy Money: Fed & Inflation | Full Documentary
@frontline
21.2M views•2023-03-15
Related Study Plans & Knowledge Roadmaps
Structured learning paths in Economics