 Good morning. So welcome to the second day of the annual research conference of the CB. So I'd like to start by thanking the organizers for inviting me to chair this session. So we'll start with German Gutierrez from New York University, who will present a paper on how EU markets become more competitive than the US markets. So it's some good news, at least for the European side. So, German. Can everyone hear me? Well, first off, thank you very much for inviting me. It's a pleasure to be here. This is joint work with Thomas Philippon with both at NYU. Let me start with a little bit of history. The US, in many ways, was a pioneer of free markets. In 1890, it passed the Sherman Act. So it was sort of the first major antitrust law. And then in the 1970s, it went through a major deregulation wave along with the UK. This meant that in the 1990s, the US was, in many ways, a clear leader when it came to free markets. And this is shown in many discussions. For instance, Alessina and Giovazzi in the early 2000s said, if Europe is to arrest its client, it needs to adopt something closer to the African-American free market model. That's basically what we want to do in this paper. Did this happen? And in particular, we want to ask if this happened by contrasting the evolution of markets in the US versus Europe. But before diving deep, let's think about some examples in some particular markets. So we can think about internet access, for example, and just compare prices and competition across regions. This is difficult because, of course, it depends on the size of demographics, the concentration of population. That affects pricing. But the industry participants have done a lot of comparisons. For instance, the public integrity group compared five cities in the US to five French cities that were common in as many dimensions as they could. And what they find is that consumers in France had a choice between seven providers, on average. Whereas in the US, they had at most two and often just one. So this tells us about consumer choice, the level of concentration. We can then look at prices. And you find very similar dynamics. So international comparison, a bunch of different countries. What you find is that European countries are doing fairly well, at least when it comes to internet prices, where consumers can get internet at monthly cost of $35. The US is essentially twice as much at $70, $66. You can do the same for cell phone services. The OECD has similar indices. You find similar results. We can look at other markets. For instance, the airlines market. So this is a market in which the concentration in the US increased very rapidly in recent years. It went from about seven major providers to just four through a major wave starting in 2008, which means that today, the concentration in the US is in excess of 80%. Whereas in the EU, it's less than 40%. You had also a lot of entry of low-cost providers. Again, looking at prices and profits, what you find is that profits in the US are essentially twice as large as they are in Europe today. Now, this we think are interesting examples, because these are two industries that haven't undergone major technological changes in recent years, yet they have very different trends across the two regions. Nonetheless, we want to think about this more broadly. And in particular, if we compare across all of the different industries, we find fairly similar results. So what this graph shows you is the net profit rate, essentially operating surplus over sales for the US and Europe. So the solid green line is the US. And what you see is that they both start at about the same level in 2000. But since then, the profit rate in the US started to increase from about 13% to 15%. The trend in Europe is very different. So the raw series, the red one, is just the operating surplus over production. And you see that that's been pretty stable in Europe. Now, you may think this is because of industry mix. The US is more weighted in tech. And profits have gone up in tech. But in fact, if we control for industries, you get the blue line. And if anything, the trend is even more decreasing in Europe compared to the US. Similarly, we can look at concentration. And we find a similar picture. So the solid line, again, shows you the weighted average herfindo. So compute herfindo for every industry in the US. And then take the weighted average across the industries. And you see that this started to increase around 2000. This is based on compute stat. You can do the same with the sensors. Find very similar results. The red line and blue line are for Europe. Now, you need to think about what's the relevant market. We show two lines here. This is, of course, fairly aggregated. But the red line shows you treating every country as a single market. And computing herfindo for the industries within that country. The blue line treats the EU as a single market. And in reality, it's somewhere in between. You can think of transitioning from the red line towards the blue line as the single market gets implemented. Nonetheless, the picture is very different. What you see is that concentration has declined at the country level. It's been relatively stable in the aggregate. But certainly, it has not increased. Not as much as it increased in the US. There's some debate about, are you using the right denominator? Would you use different sources? We've shown in the paper a bunch of different robustness tests using the ECB's component data, using Clem's data sources, as well as the census for the US. And the picture is pretty consistent. Not only that, it's pretty consistent across sectors. So if you look at the different industries and sectors, you find something very similar. So this is fairly surprising. If you ask someone in the 1980s or 1990s, what do you expect the evolution to be of competition across the two regions, you probably wouldn't have gotten the answer that you would expect Europe to become more competitive. So what we try and do here is to sort of propose an explanation for this. And there's going to be two components to this. There's going to be a theory. So we develop a model of political support to help us think through one of the dynamics when a region establishes a supranational regulator for competition. And what the model is going to tell us is that the supranational regulator is going to be more independent and more pro-competition than the national ones that it's replacing. And I'll talk about that in a little bit. The model is going to give us three key predictions, which we're then going to test in the data. Now, these are predictions about antitrust, about product market regulation. These are sort of different things to test. So we're also going to look at prices. In particular, we're going to try and get to consumer prices by using PPP data and compare the evolution of markups in the US and EU. And last, I will conclude by discussing a little bit of implications for competition policy. So with that, I don't have time to get into the details of the model. But essentially, the model is a model of political support where politicians are going to design a regulator. And they're going to endow the regulator with a certain amount of independence. The regulator will then enforce the rules, depending on the independence it has. But it might be captured by one industry. And when it's captured, it's going to allow higher profits. The politicians will get some benefits from that. We can think about that as the politicians design a regulator for their particular country. Or different countries getting together to design a supranational regulator. What the model tells us is that when they establish a supranational regulator, the politicians are going to choose a higher degree of independence. And the markets that result are going to be more competitive, meaning markups will be lower. The intuition for this is essentially that you may benefit from your country or an industry capturing the regulator. But you worry more about the regulator being captured by the industry of a different country or the markets of a different country. So that tells us we think that the supranational regulator is going to be more independent. We can learn a little bit more from the model. In particular, we can think about what are going to be the benefits for the different countries that enter the supranational regulator. And then also think about indigenous lobbying. So if we introduce lobbying into the model, what the model tells us is that countries with more independent regulators, we should see less lobbying. And this is interesting not just because of the amount of lobbying itself, but because it introduces dynamics to sort of the model. We can think of politicians in the EU designing the regulator in the 1990s. But then the elasticity to lobbying or the amount of lobbying that happens may change over time. For instance, if globalization increases and the cost-benefit trade-off to lobbying changes. So this introduces unintended consequences. If lobbying becomes more important, because firms start to lobby more, the fact that you established a more independent regulator ex ante is going to mean that your markets are going to be more competitive relative to the US. And so this is what Tomah likes to call the European forward combo. So initially you had some intended amount of independence and competition. But then if lobbying becomes more important, that means it becomes more important. And so your markets become even more competitive than initially intended. So with that, we're going to test the first two predictions. This is about antitrust and regulation. Then we're going to take a step back and look at prices, as I mentioned, and then come back to the last proposition. So first, for the first one, it tells us that when establishing the supernatural regulator, it's going to be made more independent. We can test whether this is the case, both by looking at the laws and policies on the books. There's been a lot of work in the law literature to compare this. So we're going to use a lot of those indicators. But then we can also, that doesn't necessarily mean that enforcement will be tougher. So we're also going to look at the amounts of enforcement and the trends in enforcement over time. So what we show here is the indicators of competition law and policy from Hilton and Deng. So they've essentially ranked the laws and policies for 100 countries by type of infringement, so dominance, restrictive trade, and merger. And what's interesting, what we show here is the red line is for European national competition authorities, the green line is for DigiComp, and then blue is for North America. And what's interesting here is you can see that DigiComp is indeed tougher than European national competition agencies. And we don't show it here, but it's tougher than every single country, according to this measure. And interestingly, it's also tougher than the US. So this is sort of very consistent with the model. Now this is just one indicator. We can look at another indicator from the OECD. This is a competition law and policy indicators, and find very similar results. So in this case, a lower score means it's tougher regulator. And you have four different dimensions, and you can see that DigiComp scores the best possible score for scope of action, policy on anti-competitiveness, and probability of investigation. Now the different European countries, you can see them along the line. They have reasonably good scores, but again, DigiComp is tougher than every single country. The US is the country all the way at the right, where you can see it gets the best score on policy on anti-competitiveness, but substantially worse for scope of action and probability of investigation. Now these three really map to the model fairly well. This is sort of what are the capabilities that you endow the regulator to have. The last one about advocacy. So this is how much the competition regulator can advocate for changes being made, say about product market regulation. This is not necessarily something that maps directly to the model, but nonetheless, you see that DigiComp scores pretty good and better than the US. So these are two indicators of our law and policy. We can now move on to the actual enforcement. Now this is very difficult. Of course, enforcement is highly endogenous. It also depends on the level of punishments that are made, but we're still gonna sort of try and do the best with what we can. Starting with merger, this is an area that there's been a lot of research, and here we can have a pretty reasonable comparison. In particular, Bergman et al. asks the conceptually right question for what we're thinking about here. In particular, they gather granular data about particular mergers, and then they ask the question of if the EU were to examine the US as merger, what would be the predicted challenge? And what they find is that the predicted challenge rate would be 12 percentage points higher in the EU than in the US. This is based on cases before 2004. So it does suggest that indeed the challenge rate for mergers is tougher in Europe, and it is also consistent with some of the discussion in the US. In particular, this is based on work from Tuoca, which shows you what is the likelihood of a merger being challenged, depending on the number of remaining competitors. So all the way at the top is if there's just gonna be one remaining competitor. In 1996, there was a nearly 100% chance of the merger being challenged, and over time you see that that's been pretty consistent. But the more interesting part is sort of moving down a little bit towards the middle. So that's when there were five, six, or more competitors remaining. In 1996, you would see challenge rates in those mergers, right? About half of those were challenged, but somewhere between 25% and a half. But over time, this has sort of declined substantially, so that today, basically the FTC has stopped enforcing mergers that leave five or more competitors. So again, we sort of find a decline in the US. Looking at different types of enforcement, so abuse of dominant or monopolization, you find similar trends. Now, this is difficult. The data that's available varies, and the definitions of different agencies also vary. But what we've done here is we've tried to get the longest time series that we can, with sort of whatever definition is available, so that we can focus on trends rather than levels. In particular, here we show the number of cases brought forth by the DOJ, and we have data since 1970, and we're gonna contrast that with the number of formal decisions issued by DG Comp on this type of cases. Now, formal decisions is tougher than cases, so only some cases get decided via formal decisions, so you can't quite compare the levels, but if you look at the trends, they're very different. The DOJ used to bring somewhere between five and 15 cases in the 70s, then between two to five from the 80s to 2000, when sort of the Chicago School played a bigger role. And since 2000, the DOJ has brought a single monopolization case. By contrast, if you look at the formal decisions in the EU, those have remained fairly stable. And you can think of particular cases, and across many different types of cases, we see continuous activity in the EU. So you can think of dominant IT platforms with Google and Android recently, standard-setting organizations with Samsung and Motorola, Amazon, and so on. So we find a lot of activity in Europe. So that's sort of what we have for antitrust in the paper with more detail about different types of cases, but let's move on to the second prediction of the model, in particular on these cross-sectional effects. And also, let's start thinking a little bit about regulation. This is a dimension that doesn't necessarily map to the model directly in the sense that you establish a regulator, and the regulation gets implemented by every country. So in particular, the way it's worked in Europe is in cooperation with European institutions, particular countries have to implement regulatory policies. Nonetheless, we find similar results. So to start, let's just make sure that product market regulation actually predicts profits. So this shows you just the relationship between product market regulation in 1998 and the profit rate in 1998. And you see that the relationship is fairly strong. Again, we see that the US was sort of a leader in regulation, so it has a second best score, next only to the UK. And most European countries are somewhere in the middle. What happened since is pretty striking. So you can see that here, every dot is a European country, and the line is the US. So in 1998, most European countries were scored worse than the US. Over time, they've improved a lot. So this was an explicit goal of the Lisbon agenda. The Lisbon agenda didn't succeed in some dimensions, but in this case, it was very successful. A lot of reforms were implemented so that today, most European countries score better than the US, with just a few scoring a little bit worse. There's a lot of work looking at the particular reforms that were implemented to achieve this. And they sort of show that this led to lower profits, this led to lower concentration, and more competition in Europe. Now with this, we can then ask the question of, is it true that countries that had weaker institutions benefited more from delegating to European institutions? Now, to start, we should acknowledge that part of what happened, part of this decline that you see in this graph, is just convergence globally, as different countries adopted some of the US policies on regulation. So across all countries, what this shows is the starting level of the product market regulation and then the change. That decline, and you see a strong, sort of negative relationship for all of them, but the relationship is stronger for European countries. So there was a bigger decline in Europe in product market regulation than there was in the rest of the world. And more interestingly, this decline was largely concentrating countries that had weaker initial institutions. So what we show here is measuring the quality of institutions in 1996 based on the World Bank's corruption control. You can look at different indices like the government quality, government effectiveness, quality of government, and so on, and you find very similar results. Meaning countries that had weaker institutions essentially benefited more from sort of the European policies of deregulation and they declined their product market regulation more. So we talked about antitrust. We now talked about regulation. We see that there have been big changes and on this policies has this actually affected the prices. And in particular, here we're interested in consumer welfare. So we want to think about the prices that consumers are paying when they go purchase products. So what we're gonna do is we're gonna use PPP data to try and derive a markup and study the evolution of that markup of Europe relative to the US. We know from Balasa Samuelson that prices increase with wages, right? We confirm that this relationship remains pretty strong today. In fact, it's sort of one to one of wages to prices. Given that we can define a markup in Europe, a given country, I relative to the US. So that the markup is essentially the difference in the relative prices controlling for the relative wages. And think of this as the residual in a Balasa Samuelson regression essentially. We wanna think about how this residual change from 2000 onwards, right? So we compute this cumulative change in the markup for a given country, by from time 2000 up to a given point in time. And we compare that to the cumulative changes in concentration again, relative to the US. So this is what it looks like in the time series. So looking aggregating sort of 10 of the major European economies. The red line is the markup. The green dotted line is concentration. So as you can see, the markup was sort of relatively stable starting in the mid 2000. It started to decline in Europe relative to the US. And it declined by about 6%. This is a big decline if you think that this is markup. This is telling us that prices controlling for wages declined by about 6% in Europe relative to the US. And if you look at concentration, this is sort of very consistent. There's of course a lag as you would expect. But as concentration decline in European countries, we see that the markup also declined relative to the US. Now this is just the time series. We can do the same in the cross section. So in particular, the first column uses the cumulative change in HHI. Column three looks at the concentration ratio. And we find that this relationship is indeed pretty strong. Now this is different data, but nonetheless we can add a country fixed effect. And we can add a year fixed effect so that essentially we're identifying on the differences in the declines in concentration across European countries. And we find that the relationship remains quite strong. Looking at the last column, this sort of tells us that if the concentration ratio declines by 10% relative to the US, in reality declined about 7%, the markup is going to decline by about 7%. So it's pretty consistent. It's a robust relationship across the different countries. And we can also look at the evolution of the markups for particular countries. So this is so more interesting, the country all the way at the top is Italy. This is a country where we know productivity growth has been relatively weak. And you don't see a substantial improvement or any improvement in the markup at all. Moving down, you see Germany and France almost at the bottom, Spain as well. So these are countries in which we see a bigger decline in prices relative to the US. So this is at the country level, we're working to expand it to the industry and more granular levels. So we talked about antitrust, we talked about regulation and prices. The last prediction of the model was that we should expect more lobbying in the US if institutions are less independent. So we're gonna look at that next. And to give a little bit of background, there is a lot of discussion in the US. This comes from a quote or a letter that Jerry Paul wrote to the FTC when they started the investigation of Google. Now Google is a case that the FTC considered and decided not to bring, whereas recently DigiComp ruled on it. The letter basically says, I believe that application of antitrust against Google would be a misguided step that could lead to congressional action reducing the ability of the FTC to do its job. And this is not an isolated incident. Just for Google, there were 13 Congressmen who sent letters to the FTC. So somewhat common, it's suggestive that there is potential for some effect of lobbying. So with that, we can actually compare the expenditures. So this shows you what are the total lobbying expenditures to the federal institutions in the US compared to lobbying in the EU, to EU institutions. So as you can see, lobbying in the US is about three times higher as it is in Europe. And interestingly, there's a much larger percentage of lobbying that is done by business in the US relative to Europe. These are aggregates, perhaps more interestingly, if you then match this to firm-level data, the elasticity of lobbying to size are about 0.15 in Europe compared to 0.62 in the US. So you see that the larger firms in the US lobby a lot more and they increase the lobbying much faster. We can then look at outcomes. There's been, again, work in political science looking at this that basically shows that in the US there's a much higher likelihood of corporations succeeding in their goals, both in absolute terms and relative to institutions that sort of fight for the public goods. So 89% of corporations succeeded in their goals compared to 40% or 37% of foundations and city-side groups. In Europe, you still see some success of lobbying firms, but then citizen groups and foundations have about the same rate of success as corporations. The explanation that's been put forth, at least by Mahoney, was that this was driven by campaign contributions. So we can look at campaign contributions next. And if we think that the difference is lobbying or striking, this is even more. So this shows you total campaign expenditures for federal elections as a percentage point of GDP. And what you see is that campaign expenditures in the US are anywhere from about three to 20 times higher than in European countries. And here we just have a sample of countries for which we could get data. So again, consistent with the model, we're seeing a lot more lobbying, a lot more campaign expenditures in the US relative to Europe. Now, we've talked a lot about regulatory activity so far. Could it be that there's an overreach? You see that regulators are doing too much. And essentially, this is the whole point of the Chicago School, they're reducing the productivity. So what we do to test this is look at the industries in which cases were held. So this is an industry panel. And we look at what is the effect of enforcement activity in the EU on first concentration and profit rates. And as you can see, indeed more cases lead to lower concentration, lower profit rates. But they don't seem to have a negative effect on productivity. In fact, this is noisy and it depends on the specification that you use. But if anything, the coefficient is positive. So, that's all. Thank you. Thank you very much. So we have the discussion, Shabnam Kalemni Oskom from University of Maryland. So, Shabnam. Thank you so much for inviting me to discuss this paper. It's a very interesting paper as you just listened. And if I can have my slides, I can start. Slides. All right, so let me start by a little bit talking about the state of the literature. Because if you are not working directly in this literature, actually, it can get quite confusing with two sets of, you know, kind of researchers looking at two sets of facts that are related, but at the same time also distinct, which are facts on markups and facts on concentration. So, the literature, okay, my title got cut off, so that's not good because I have a lot of stuff there. Is there a way we can adjust this for the title? Yeah, okay. Let's just do this. So, literature says concentration increase in the US. So, this is actually a very, very large literature focusing on the US. In fact, the first paper that started focusing other countries is by Herman and Thomas. So, that literature, European literature, still at infancy, and the literature in the US puts two interpretation on this, the higher concentration in the US. So, the first interpretation is a more, you know, less benign, kind of more pessimistic, which is about market power. This interpretation goes in a way that these two sets of facts, facts on concentration and facts on markups go together. So, concentration, sectoral concentration increase in US, but markups also increase in US. In fact, Herman and Thomas, they have another paper. The first paper on this was by Delocre and Eckhout. You know, these guys says, you know, this is alarming because of course, this is going to imply low investment, low productivity. So, we don't like it, right? So, this is like lower competition. Now, there is a more benign, more optimist interpretation of these facts in the US and that's the technology interpretation. So, the work by Alter et al. and Eberle and Coulter says, well, look, the way the technology evolves actually allows some really large firms, some superstar firms, get a higher share of the market. Okay, and that's, you know, combined with the intangible capital these firms use, you know, Amazon, Google's of the world, it makes it hard to decide what is going on with investment. If anything, this is, this can lead to higher productivity and higher investment. So, these are very, very different interpretations. I mean, obviously super different policy implications. So, it's important to understand what is going on. Now, here comes this paper and this paper says, well, can we learn more about these issues if you look at Europe? They argue forcefully actually, you know, in a series of work, this paper and the previous paper, it can be the technology interpretation. Why? Because clearly European countries and US, these are both at the technological frontier, you know, how different technology can be among these countries, but maybe there is an institutional explanation to these differences in concentration increasing in the US and decreasing in the US. So, the authors convincingly argue that it can't be technology, but the differences in institutions in US and Europe maybe have a bite to explain these differences in concentration, and that is about anti-trust law enforcement. So, they are going to argue that the way US and the European Union enforces anti-trust law can explain the differences in concentration. They provide a model and they provide evidence supporting their argument. Okay, so my general impression is this is an excellent paper. It is very provocative. I think it's super useful to look at other countries and get out of the US if you want to know more about this issue. And this paper really makes you think outside the box. They're still working on it, so I think there are a lot to do to strengthen the interpretation. So, my comment is going to be exactly on that. So, I'm not going to argue with their general conclusion. I really like a lot this line of work. I think if they take care of some measurement issues and little bit think more on how to link the model and the evidence, I think that's going to really help their interpretation and they can further support their argument. Let me start with the measurement issues. So, the first measurement issues is about definition of the market. So, how people do the concentration. Let me talk about concentration. I'm going to mention markups a little bit, but this paper is really about concentration. And the way you look at concentration, and markups are very different. You saw from the Hermann's presentation, they start looking at markups too, but let me just focus on concentration right now. That's their main thing. Concentration is measured by a Herfindal index. And that is about, you know, you go and take the market shares of the firms and then you square it. And of course, the first question here in terms of the measurement is how you define relevant market. You ask the IO economists, they would tell you, well, you know, the market definition should be based on the consumption's substitutability. You want to have a high cross-press elasticity of demand. This is exactly why IO people go and look at very narrowly defined industries. Clearly, in a context of a more macro study, this is hard. So, you know, the authors look at the sectors, SIC sectors. You know, that might be a little tricky because if you talk to anti-trust practitioners, SIC-based industries may not be the relevant market. So you might have low concentration based on SIC codes, industries, but then there might still have monopolist lurking inside that industry within a smaller market. So it's important to link a little bit of what do we mean with the market to what anti-trust practitioners mean with the market. And they may not coincide. In European cases, it's important because as you all know, when we think about single European countries versus EU as a single market, there is going to be a lot of national interest, political factors, certain sectors are going to be defined under national interest. So there's going to be a lot of issues that relates to role of global competition and foreign ownership, how, you know, Europe is going to be very sensitive in terms of mergers and acquisitions, especially cross-border mergers and acquisitions when it comes to certain sectors. And I'm going to talk more about that. All this makes us a little bit maybe cautious in terms of putting that arrow there, direct link from observing a low concentration at the SIC industry level to what you want to go, which is European firms are competing more. That's really what anti-trust is about. The message you want to give in this paper is European firms are competing more for customers than the American firms. But they can definitely improve on that because now they are going more granular. This is something very hard to do with the sector level data. They are, you know, you can do more with the firm level data. You can, you know, define the markets, you know, checking with the anti-trust definition and all that. So clearly, and that's exactly the way the authors are now going forward. The second measurement issue is about firm selection. This is more important when you start doing markups. I'm not going to say much about this because as you see, they just started working on prices and markups, and in the paper, they didn't have much, but this is again the direction now the authors are taking. What I'm going to show here is the markup calculation from two different papers. So the first one, the one with the red lines, is from Deloquent and Account, their global market power paper. So you see, like, you know, Europe, North America, South America, you know, all these, like, super high markups. The numbers you see there is like, markups are increasing like 60%, 70%. It's like, it's gigantic. I mean, this is like, you know, really, really big increases. They use listed firms. All those red lines, and you can just focus on the first little window saying Europe there, that's an increase of over 60%, that is based on listed firms. This markup calculation is based on a production function estimation. Herman and Tomah is not doing it this way, but this is still important because the huge literature, including the IO economists, do it this way. So, and when you do it this way and you use on the listed firms, you get these gigantic numbers. Now, the bottom figure with the blue line, that's only Europe, that's an order of magnitude smaller. So you go from an 80% markup increase in Europe to just an 8% markup increase. Actually, not even eight, it's seven. That's from a paper by Diaz Fan and Villegas Sanchez that's using all firms' data in Europe from Adios Orbis, which is actually, the data set also now Herman and Tomah is using. So I would be very interested to see this also for concentration. Like, you know, if concentration will also change a lot, if we do these things. But this is just to say that, this is not to say that I'm not believing their first markup calculation that show prices are also decreasing in Europe. I'm just like showing this so that this is very sensitive to what type of firms you are using. The minute you move from sector data to firm data, it's going to be important to look at the firm selection in your sample. Okay, so in terms of my comments in linking the model and evidence, they have a super nice, elegant political model. The key in the model is these institutional differences between the European Union and the United States. So these two countries, continents, U.S. and Europe, they start with different initial conditions. U.S. starts better. But over time, U.S. deteriorates and European Union improves. It's shown very nicely. It's backed up with this very nice low-being and political capture in the model. In the data, a lot of these things shown in static differences. Remember the original figures Herman showed you, like which one is higher, which one is lower. Later on the antitrust cases and enforcement, Herman showed time series data. But I think we should be doing more with the time series data here. Because it's really about deterioration in the U.S. and an improvement in the Europe. And of course this is very hard as Herman said, because different indicators can give you different stories. I'm just going to plot three governance indicators from World Bank, just very standard measures. That shows you exactly what Herman's talking about in terms of initial differences. But then the over time deterioration is not that different. So that's why it is important which indicator you are looking at. This is government effectiveness. I have it highlighted what it captures. It is about independence from political pressures. So related to their political capture idea. And the EU is the red line, the U.S. is the blue line. And there's deterioration in both of them over time. Deterioration is not that different. If you look at regulatory quality, this is about policies and regulations that helps private sector development. Again there's less on the level differences. And EU is doing much better here because EU is nothing going on over time. But there's deterioration in regulatory quality in the U.S. So that goes more with their story. But then we also don't see any improvement in the EU. And the third one, the corruption, this is actually one of the things they do look in a static environment. Here it is dynamic. And as you see in the yellow highlight, this is really about capturing of the state by elites and private interests. So this is really the idea they are after. And here there is definitely a deterioration in the U.S. but there's also a deterioration in the Europe. I mean not that big a difference. I believe there's some more work to do here in terms of maybe certain indicators linked to their model better. And then maybe we can look at both static and dynamic differences in those indicators. Another important issue in this model evidence link is this trust, right? The key assumption in the model is that French doesn't trust German, they trust more to ECB, okay? So the idea is I trust the pan-European institution more than I trust my fellow Europeans. This is actually a really neat idea. And this relationship can be static, can be varying over time, can be varying geographically by countries and regions. So there's a lot actually they can do with this relationship. And it's super important for their model and the argument. In fact, there is actually a literature that shows a very strong correlation between trust and economic integration, both financial integration and trading integration. People like Luigi Gizio work a lot on this. I have a paper on this that shows how the laws on the paper that allows you easier cross-border mergers and acquisitions say the laws make it easier for German firms buy out French firms but you don't do it because you don't trust Germans. So that's actually really important. So to kind of go back to the work I did in 2007 actually for an ECB conference and then kind of re-look at that data, this is plotting exactly that relationship. So on the x-axis I have Germans trusting more or less to French. So the idea is you trust more or less to other nations. That's the x-axis. And the y-axis is your confidence and trust in EU-wide institutions. There's a negative relationship. It's not very strong but there's a negative relationship that says I trust less to other countries, I trust more to EU-wide institutions. I think this can be shown much stronger if the authors use European World Value Survey. I use data from World Value Survey that's more restricted on the coverage of European regions but there seems to be something there. But again of course you also want to show it over time because remember their model is about there are these differences between US and Europe and then these differences intensify the other way in the favor of Europe over time. So if you look at this over time I'm going to do this in map so you can see visually. So what you want to see here this is kind of European regions and countries and darker colors is more trust. So this is going to show 2005 and then when I click I'm going to go to future year. The idea is if the color gets darker there is more trust to others. So French is now trusting more to Germans. If the colors get lighter there is less trust. And obviously you see there's a lot of variation within Europe across European regions and countries. So this is 2005 survey then the colors get lighter in 2010 that means now European nations trust even less to each other. Okay so that goes their way. What about trust in the EU because you want the other one goes to go darker so you want I trust less and less you know as a German, well I'm not German but for the sake of argument as a German I trust less and less to French but I trust more and more to ECB European Union and all these pan-European institutions. And that seems this also gets lighter. Okay so this is a trust in the EU institutions in 2005 and 2010. Of course there is the crisis in the middle of this sample you know so this can be related to that but this is something important I think to pin down because this is central to their model. Okay and my final point is on the interpretation. So it is very important for the authors that it is political capture right. So the idea is you know there's a lot of lobbying in the US you know this you know politicians capture regulators capture by the elite private interests and that's why you know US is going down and in Europe this doesn't happen. Well this is actually extremely hard and Herman all that said that this is not very straightforward because think about you want to link anti-trust enforcement you know over time to political capture and to concentration in data. So in the terms of the example Herman is giving you that letter written to the congressman right I mean in for Google so you want Google paying those people and then that links to the concentration is exactly Google operation. So it's just it's not a very straightforward link to make with the data they have. You definitely need to go to more granular data and you really have to establish robust evidence on that. Lobbying is a very good idea to look at in terms of establishing this relationship and they do actually look at that. The model is going to show you that larger firms lobby more and get larger. Very nice but also we know that from the literature in the US productive firms also lobby more and we know from the firm dynamic literature productive firms of course get larger. So it's not just that the causality goes both ways you know there's also this omitted variable bias here in the relationship between size, market concentration and lobbying. So one idea here they can do and this data do exist for US. I'm not sure if it exists for Europe but this geographical dimension cross state lobbying cross state concentration. So it's a state sector variation that maybe you can do more in testing these lobbying. And the genetic is still going to be hard but at least you're going to have more variation. Now the interpretation also gets tricky when you think about how DGCOMB operates relative to DOJ Department of Justice and FTC Federal Trade Commission. Well you know because when you look at the law on the paper it's not that different but in practice it is possible that DGCOMB defines the market wider. Maybe they are more speculative like the Google case again an example to block these mergers because they don't have to prove their case in court as tightly as DOJ and FTC. In fact when you talk to antitrust practitioners, IO people that's the first thing they think is this called political capture? Maybe it is political capture but clearly more work needed to link that to the political capture. And as I mentioned before in the European case this cross border mergers is going to a very important role because Europeans of course famous in blocking these cross border mergers and not because it is tough for antitrust enforcement but because of this political national interest. So I'm actually going to show you some data on this. So this is from recent work I'm working with Karol Nevillega Sánchez where we look at the concentration measures. So very similar actually to Herman and Tomá using a Herfindahl index. This is Herfindahl index based on top 50 firms market shares. So just you know there are a lot of sectors here and there's variation in sectors and we plot two years what happens to concentration in 99 and 2012. The point I'm trying to make if you look at Tobacco just the first item you see that the light blue line is the you know Herfindahl based on top 50 firms market shares that says look 80% of this market is goes to the top 50 firms. That's the light blue line but the dark blue line are the foreigners but most of that is the foreigners. So the sectors where foreigners have a big market share actually the ones that have the bigger drop in concentration. You can do it more systematically where you look at the foreign concentration in the sectors in the initial year and then look at the drop in the concentration and it's going to be, and this is Europe from 1999 to 2012 the higher the foreigners market share in a sector the bigger the drop in concentration in that sector. Again this is now suggestive it has to be of course done much better with many robustness but at least there might be something there. Well you might think well maybe we should look at FTI restrictiveness because instead of like a you know endogenous outcome like foreign concentration maybe we should look at the laws on the paper which is what they are doing with product market regulation. So this figure plots you the concentration Herfindahl index against the FTI restrictiveness by sector. So there is a weak negative relationship that says you know the more the sector restricts foreigners coming into that sector the lower the concentration that sector but this is very weak it's not significant then you look at the change again change from 1999 to 2012 kind of during their sample period it is indeed the case that if you restrict the sector more for foreign entry for FTI that is the sector that is going to have a bigger drop in concentration. Just to summarize I believe this is a very elegant very thought provoking paper. You know they are definitely onto something and I do believe there is definitely meat in the story relates to the regulatory environment and institutional differences in US and the European Union but I believe we need more evidence to connect these three right you want to connect low concentration in Europe to low market power and that should be due to better anti-trust enforcement both over time and also in cross-section and I believe there is actually an important potential role for cross-border mergers and acquisitions in keeping concentration low in the European Union because you know pretty much in the last 10 years each time when you know Germans trying to buy a firm French Screams so this is actually an important issue for Europe. Thank you. So let us collect some questions so Luke. Yeah thank you. It's rare to have a paper claiming that Europe is doing something better than the US. It makes us immediately skeptical right? But so I like very much the way Shepnam you started your discussion because there is two very important strands of the literature emerging so is it really competition or technology? There may be a way to link these two because what I found actually the most insightful parts of this literature is the rise of star firms. So you mentioned Google but this is across industries in the US and this is definitely not happening in Europe. And so even if it's that technology moves across borders, let's assume that. I mean I think that can be contested as well and this is then the big difference and so what I find quite alarming is that you have this growth of star firms in the US but death rates have also collapsed. So there's a lot of entry and then super high growth. Obviously if then death rates collapse as well then concentration ratios will move. So maybe there's a way to link this technology literature the rise of star firms to concentration and some of the things you look at in your paper to kind of link and hopefully separate these two stories. I wanted to have your reaction since you're much better informed both of you on this literature. Okay, have the fourth row. Thank you. One comment and one question. The comment, I wonder actually whether you can look at it from the other way, the other side, sorry, talking about populism. Why is the EU so unpopular? We know of course there's a demand and a supply side to this populist wave and a supply side and I can look at this mostly from the side of the UK and the leave campaign has been often supported by people who try to protect their market share. So this kind of maybe the realization that you can't lobby the EU as much as you can lobby and capture national governments has its parts in there and just to give you a quote with that, Robert Murdoch was asked why he hates the European Union and he answers, well if I go to Downing Street they listen to me, if I go to Brussels they don't even receive me. So that might go in there. Then I have a question for Shepnam. On the governance indicators, I wonder whether this is the right way to look at it because for the EU it is a country average, right? And of course it moves over time and I think it's that weighted by GDP but of course we also know that certain countries grew much faster than others partly due to convergence like in Central and Eastern Europe. So I wonder whether this declining trend for example in control of corruption is actually driven really by deterioration or is it just driven that the countries that are more corrupt actually were also growing faster not because they were more corrupt but because of the convergence process. Thanks. Yes, over there. Over there first and then Bernard. Sorry, and then over there. Yeah. No, on the back. Okay. Sorry. On the back. I'm trying to follow the order. Okay, thank you. It's maybe more of a remark than a question. I think one element that I've been missing which I think is an important fact of EU law is the whistleblower rule in cases of price fixing. There have been an enormous number of price fixings in the EU that have gotten basically looked at and the whistleblower rule is this, is that the first company that comes forward that says we've been doing price fixing doesn't get the huge fines that the EU gives and the other firms suffer from that obviously. And there have been enormous number of firms coming forward saying trying to be the first, this has happened in the car industry, this has happened in the washing powder industry, this has happened in a lot of industries. So I think that also keeps markups low essentially by having this. And I wonder, I don't know the US, but I don't think this whistleblower rule is there in the US, but I think that's an institutional feature that's underestimated the power of this one. Bernard? Does this work? Yeah. Great work. Since we're at the ECB, I want to ask a bit about monetary policy implications or relevance. The first point is the increase in competition coincides more or less with the introduction of the euro. So is there, could there be an effect from the single currency on top of the single market or could you try to distinguish what is due to the single market in the early 90s from maybe some amplifying effect from price transparency through the single currency or is this all second order? Second point, you pointed to minus 6% in the markup since the early 2000s. That's a big thing. Does it affect the optimal rate of inflation? So should there be a lower inflation target? Since this is probably something that is not in the control of the monetary policy makers. Third point, like look, what is the time profile of the effect on inflation? If we have these superstar firms like Amazon and now Tesla, they have a lot of losses initially invested market shares and then they profit from the monopoly position later on. So you have very low inflation in the initial phase but then you have much higher inflation later on. Is that something that banks should worry about? And the final suggestion is for the next paper. I suppose your work is mainly on non-financial corporates, could you actually do the same thing for the financial sector? There's people in this building and elsewhere saying we need to concentrate and have big banks to be more profitable, which is a different angle than the Brussels Digi competition angle. It would be very helpful to have a work extended to that kind of question in the financial sector. Thank you. Okay, second rule. So I think the nationality of the star firm is very important because for instance, in corporate governance, we know that it is relatively easier to enforce the law against foreigners. So I am wondering to what extent the stronger level of enforcement of the competition authority in Europe is driven by the fact that a lot of the formal actions are against foreign firms and the star firms are from the US in particular. Would be nice to see those numbers without the foreign cases. Okay, one last question. I really enjoyed this a lot. It was really a fascinating paper. I had two comments. One of which was that, and I'm speaking primarily from a US perspective, was that there's a huge amount of the US economy that's regulated primarily at the hyper-local level rather than at the national level. And those have very different dynamics. So a third of consumer expenditures and housing. And there is some extent to which mergers of large construction companies are regulated by the DOJ, but that is trivial relative to what hyper-local communities do in terms of regulating things. And those dynamics are quite different. And indeed the usual story is in fact, local capture by builders and banks actually promotes growth and promotes supply rather than the opposite. That in fact it ends up being that the counter to that is when homeowners actually capture the locality and then they block all new construction. But there are other areas like retail for example where regulation is often done at the local level as well. So I would encourage you a lot to think about not just the EU US differences at the macro regulatory level, but also at the micro regulatory level. Particularly in housing, but in other sectors as well. And then the larger sort of very macro point that I wanted to make is that when the AEA was founded 140 odd years ago or so, it's clear that monopoly was the biggest thing on Francis and Mazel Walker's mind, that this was the social loss that was driving things. And that was true throughout most of the next 60 or 70 years, but I would say for most of the 20 odd years in which I started in this profession, we sort of thought monopoly had become relatively unimportant and that we worried much more about externalities like carbon emissions. Have we reached a point in which monopoly should again be seen as being the primary or the second form of market failure that we worry about? And if so, would you be comfortable putting numbers down in terms of social losses that would make the case that relative to the pollution congestion, whatever we think of the large, in classic Pogovian externalities that are out there, monopoly is actually a greater social harm. Thanks very much. Hermann. Lots of questions, I'll touch on one again. So first, thank you very much. I think I agree that the point of contrasting technology and trying to learn by contrasting the different regions about whether technology is what's driving the strengths is very important. I should have emphasized more that in terms of markups, yeah, so we recently got firm level data. Part of what the reason, part of why we look at aggregates is that we think that that sort of solves some of the issues of whether you do that at the firm level. So it solves some of the measurement issues. The fact that we don't see profits rising in Europe to us is fairly convincing that you don't see it at the aggregate. But nonetheless, I think we need to do a lot of work at the micro level. We need to look at the World Bank and think more about the evolution of indicators over time, so I agree with that. In terms of granular political culture, we're looking at that. So we're in the process of sort of merging lobbying data at the US and EU level to the firm level so that we can actually ask some of those more granular questions, but that takes some time. In terms of sort of cross-border antitrust and M&A activity, so there is work that not for M&A, but for broader antitrust activity that shows that, in fact, there doesn't seem to be a bias against foreign corporations. Once you look at the data, once you look at the fines and so on, there doesn't appear like DGCOM has been penalizing foreign corporations more. For M&A, I think that's a different story and it's a broader sort of question of capture at different levels than just DGCOM itself. But I agree that that's an important point. Okay. In terms of sort of linking competition and technology, I think that's part of what we're trying to understand, whether this is sort of a good or a bad trend. Of course, if we have the superstar firms, we should expect technology to go up. Death rapes would rise at the beginning and then eventually fall if there's no entry, if sort of firms cannot compete. There's the whole literature in sort of frontier firms getting further away that then could potentially depress entry and it could in theory lead to rising concentration. I think what's important is understanding whether this is sort of beneficial or not and what is really limiting entry, whether it's the fact that these firms are super productive and they continue to invest a lot and get more productive over time or if they sort of gotten to the point where they're far enough away and they've moved on to sort of protect their rent and erect more barriers to entry. So I think a lot of the work that we have and what we looked at before was about investment. We find that in fact, you don't see these leaders invest a lot, you don't see these leaders sort of getting more productive. In fact, a lot of the relationships that you find promote or about concentration rising with DFB, they hold up until 2000 but then they break thereafter. So we do have a bit more of pessimistic view of recent years that yes, at some point this was because more productive firms were very bigger but that may be going away now and we may be at a point that things are not so good. To the point of populism, I agree I think that's sort of part of the point that delegating to the EU is a way in which you can sort of mitigate this influence from local governments and local elites. Let's see, monetary policy, I think there's a lot more work to be done on that. I think we haven't thought enough. Last in terms of local regulation, yeah, absolutely agree. I think occupational licensing, all those subjects are very important. We have another project trying to look at state level regulation and study its effect on entry, which is sort of to the point of what does it mean for barriers to entry and is it that incumbents that love me more than the press, the level of entry into their industries. So yeah, a lot to do. Just to answer the question, I totally agree. They have to do it by country by country too. This is what exactly I work on right now, but their existing figures done the GDP rates, so I just plotted different indicators also doing the GDP base, but that's important. So let me say one thing, I forgot to say this and I didn't have time to put in my slides, but when Ed asked this question, I just remember that I saw this paper this week. There's a new paper by Esteban Rosse Hansberg and co-authors on local versus national concentration in the US, and it's very interesting actually they show this increased concentration in the US as the other papers as worked by Hormann and Tomah, but they also show local level it is decreasing and it promotes actually growth and investment. So local defined more like in the antitrust IOA, so it is actually I think very important. And the multiple and finance question, we actually look at this finance sector, finance is the only sector in Europe that concentration increased actually from 1999 to 2012. But this literature is really focused on non-financial, but the finance would be on the other side. Okay, thank you very much. So we stop now for a coffee break and we're back at 20 past 10. Thank you.