WEBVTT - How a DOJ Economist Approaches Antitrust in America 0:00:03.120 --> 0:00:17.960 Bloomberg Audio Studios, Podcasts, Radio News. 0:00:19.880 --> 0:00:23.120 Hello and welcome to another episode of the Odd Thoughts Podcast. 0:00:23.280 --> 0:00:24.640 I'm Tracy Allaway. 0:00:24.360 --> 0:00:25.480 And I'm Joe Wisenthal. 0:00:26.000 --> 0:00:29.440 Joe, we've done a little bit at this point on antitrust. 0:00:29.520 --> 0:00:32.960 I feel like, kind of interestingly, the first time it 0:00:33.080 --> 0:00:38.400 ever really came up substantively for us was I think 0:00:38.440 --> 0:00:41.440 it was like way back in twenty twenty or maybe 0:00:41.440 --> 0:00:44.480 twenty twenty one, where we were talking about the potential 0:00:44.520 --> 0:00:48.000 for antitrust enforcement to bring down inflation. I don't know 0:00:48.040 --> 0:00:49.360 if you remember that totally. 0:00:49.479 --> 0:00:51.200 No, I do remember it. I think we were talking 0:00:51.200 --> 0:00:54.319 about it actually in the context of shipping at the time. Yeah, 0:00:54.360 --> 0:00:56.960 of course, the major source of inflationary pressure was the 0:00:57.000 --> 0:01:00.440 stress on supply chains. But then that brought away, I 0:01:00.480 --> 0:01:03.120 would say. And then it's since you know people paying 0:01:03.120 --> 0:01:05.600 attention to it more, which is that you know how 0:01:05.600 --> 0:01:09.759 to bottlenecks emerge? How do inflationary bottlenecks merge? One theory 0:01:09.920 --> 0:01:14.240 one source of these inflationary bottlenecks is parts of the 0:01:14.440 --> 0:01:16.520 economy where there's not very good competition. 0:01:17.080 --> 0:01:17.399 Yeah. 0:01:17.480 --> 0:01:21.160 Absolutely. And the other interesting thing that's been happening recently 0:01:21.440 --> 0:01:25.440 is it feels like definitions of anti trust and anti 0:01:25.480 --> 0:01:29.600 competitive practices have been expanding, right. You hear this term 0:01:29.720 --> 0:01:33.639 hipster antitrust all the time, this idea that maybe anti 0:01:33.640 --> 0:01:38.480 competitive behavior doesn't just materialize in things like prices, but 0:01:38.640 --> 0:01:40.720 also in things like labor practices. 0:01:40.800 --> 0:01:44.720 Totally. I think the rise of the huge tech giants 0:01:44.840 --> 0:01:48.760 has contributed to this because you do have this situation 0:01:49.160 --> 0:01:52.840 in which it feels like there are entities. Look, if 0:01:52.840 --> 0:01:56.280 you have like a gigantic steel monopoly or whatever, how 0:01:56.360 --> 0:01:59.080 is that going to manifest itself as being bad for 0:01:59.120 --> 0:02:02.640 the economy, Probably by really high steel prices, right. I 0:02:02.680 --> 0:02:06.040 think in some of these other business models, you sense 0:02:06.160 --> 0:02:12.080 that there is concern about corporate power, corporate concentration, extremely 0:02:12.160 --> 0:02:16.600 powerful companies, but it doesn't necessarily manifest as higher prices. 0:02:16.639 --> 0:02:18.440 In fact, it maybe lower prices, and it may be 0:02:18.720 --> 0:02:20.919 or maybe there's something not priced at all, like right, 0:02:21.000 --> 0:02:23.360 none of us pay Facebook directly or whatever it is 0:02:23.360 --> 0:02:25.520 some of these big social media giants, And so I 0:02:25.560 --> 0:02:28.760 think that has probably contributed to some extent for like, 0:02:29.080 --> 0:02:32.880 how do we address market power through outside the lens 0:02:32.880 --> 0:02:34.440 strictly of consumer prices? 0:02:34.680 --> 0:02:37.560 Absolutely, And the other thing just going back to shipping. 0:02:37.760 --> 0:02:40.400 It feels like anti trust is also potentially getting more 0:02:40.560 --> 0:02:44.520 attention in the context of supply chain resiliency. Right, So, yes, 0:02:44.720 --> 0:02:47.480 you can have high prices because of a monopoly that 0:02:47.520 --> 0:02:50.280 dominates a particular industry, but you could also have a 0:02:50.320 --> 0:02:54.240 situation where if something happens, if there's a choke point 0:02:54.400 --> 0:02:57.119 on that business, as we've seen a number of times 0:02:57.480 --> 0:03:00.799 in recent years, that it creates a vulnerable for the 0:03:00.960 --> 0:03:04.320 entire economy. So all of this got me thinking, you know, 0:03:04.480 --> 0:03:09.639 clearly there is an impact from antitrust on the overall 0:03:10.520 --> 0:03:13.800 shape and functioning of the economy, and you see that 0:03:13.919 --> 0:03:16.600 kind of burst into the policy debate every once in 0:03:16.680 --> 0:03:19.600 a while. But I really wanted to dive into how 0:03:19.639 --> 0:03:22.480 do you start to judge the effects of corporate concentration 0:03:23.040 --> 0:03:26.400 on the overall economy? And I have the perfect guest 0:03:26.440 --> 0:03:28.600 for you, Joe Great, We're going to be speaking to 0:03:28.720 --> 0:03:33.080 someone whose day job is basically exactly this. We're going 0:03:33.160 --> 0:03:35.960 to be speaking with Joanna Marinesque. She is the principal 0:03:36.000 --> 0:03:40.000 economist at the Department of Justice's Antitrust Division, also a 0:03:40.040 --> 0:03:43.520 professor at the University of Pennsylvania. Joanna, thank you so 0:03:43.600 --> 0:03:44.960 much for coming on all thoughts. 0:03:45.160 --> 0:03:46.360 Thank you so much for having me. 0:03:46.840 --> 0:03:50.520 What does a principal economist at the DOJ actually do 0:03:50.720 --> 0:03:52.480 walk us through your sort of day to. 0:03:52.520 --> 0:03:55.840 Day absolutely so you know, in my day job, what 0:03:55.960 --> 0:04:00.120 we do is we investigate different companies for potential violation 0:04:00.320 --> 0:04:03.080 of the anti trust laws and so you know all 0:04:03.120 --> 0:04:06.520 these competition issues that you just talked about. And so 0:04:06.760 --> 0:04:09.840 as an economist, what I do there is I oversee 0:04:09.880 --> 0:04:13.360 the work of my whole team of PhD economists where 0:04:13.920 --> 0:04:17.599 they go and do data analysis in order to uncover 0:04:18.200 --> 0:04:21.800 the effects of anti competitive practices on all sorts of 0:04:21.839 --> 0:04:26.440 outcomes like prices and wages and the quality of products 0:04:26.880 --> 0:04:29.200 and so on and so forth. And we use both 0:04:29.320 --> 0:04:33.520 quantitative analysis as well as interviews and documents that we 0:04:33.600 --> 0:04:36.719 are able to get from the companies like emails, to 0:04:36.839 --> 0:04:40.880 better understand how competition works in that particular case and 0:04:41.440 --> 0:04:45.080 whether there is in fact a sign of anti competitive behavior. 0:04:45.279 --> 0:04:48.000 So we talked about in the intro the sort of 0:04:48.040 --> 0:04:51.560 antitrust thinking, maybe widening the aperture of where you would 0:04:51.640 --> 0:04:56.760 look to find evidence of uncompetitive behavior beyond perhaps just 0:04:56.880 --> 0:04:59.760 consumer prices. What type of data are you looking when 0:04:59.760 --> 0:05:01.920 you talk about your going shifting through all this data 0:05:01.960 --> 0:05:04.680 to look for evident what are the types of data 0:05:04.720 --> 0:05:07.000 that you might look at and what might be I guess, 0:05:07.360 --> 0:05:10.800 the signatures of a company or an industry in which 0:05:10.839 --> 0:05:13.240 anti competitive actions are a problem. 0:05:14.000 --> 0:05:16.240 So there are two main types of data that I 0:05:16.279 --> 0:05:19.640 would point to. The first one is on market shares, 0:05:19.839 --> 0:05:24.080 so you look at the share of the companies in 0:05:24.120 --> 0:05:26.480 their market. So this could be a labor market or 0:05:26.520 --> 0:05:29.760 a product market. So you know, if a company has 0:05:29.800 --> 0:05:32.280 a high share and merges with another company that also 0:05:32.320 --> 0:05:34.960 has a high share, then that's likely to raise anti 0:05:34.960 --> 0:05:40.360 competitive concerns in the sense that prices or wages might 0:05:40.440 --> 0:05:44.440 be affected in a bad direction. So that's one way 0:05:44.560 --> 0:05:48.000 that we look at things again, calculating market shares, and 0:05:48.080 --> 0:05:52.000 so that's for one approach to looking at potential for 0:05:52.120 --> 0:05:56.160 anti competitive harm. And we can also talk about dominant firms, 0:05:56.160 --> 0:05:59.919 so firms that have monopoly power. That's also one of 0:06:00.120 --> 0:06:03.000 ways to measure it is through market shares. And then 0:06:03.040 --> 0:06:07.160 the second way is looking at what we called head 0:06:07.200 --> 0:06:10.120 to head competition, So how much do firms directly compete 0:06:10.160 --> 0:06:14.040 with each other? And this would be found in things 0:06:14.120 --> 0:06:16.240 like they are called win loss data. So when you 0:06:16.320 --> 0:06:19.880 look at sales data, you're looking at what company you 0:06:20.080 --> 0:06:22.599 lost your business too. And that would show if you 0:06:22.680 --> 0:06:26.080 often lose business to another company that now you're trying 0:06:26.120 --> 0:06:29.240 to merge with. Well, once you're merging, there's no longer 0:06:29.279 --> 0:06:33.000 going to be that competition, the intense competition between the two. 0:06:33.360 --> 0:06:37.000 And so that's the second like broad category of data. 0:06:37.040 --> 0:06:39.719 And there's similar data for workers. If you constantly lose 0:06:39.760 --> 0:06:43.760 your workers to your competitor and now you're merging, that 0:06:43.800 --> 0:06:45.080 competition is going to be lost. 0:06:45.200 --> 0:06:46.320 Oh that's really interesting. 0:06:46.400 --> 0:06:49.640 So something like that win loss data, how would you 0:06:49.720 --> 0:06:51.960 actually access that? Where does that come from? 0:06:52.080 --> 0:06:56.000 So that comes directly from the company's data. So we 0:06:56.480 --> 0:06:58.840 because you know we're the Department of Justice, if you're 0:06:58.880 --> 0:07:02.919 under investigation with the companies must send us their data, 0:07:03.640 --> 0:07:06.240 not any data, but of course the data that's needed 0:07:06.279 --> 0:07:10.200 for the investigation that concerns the specific concern we have 0:07:10.640 --> 0:07:14.600 with potential adverse effects on competition. So companies send us 0:07:15.040 --> 0:07:19.320 the data, whether it's again on the product for labor markets, 0:07:19.360 --> 0:07:22.440 it would be data from things like an applicant tracking 0:07:22.480 --> 0:07:25.640 system where they see where people are coming from. So 0:07:25.680 --> 0:07:28.120 that can also allow us to know where they're hiring 0:07:28.160 --> 0:07:32.840 employees from and therefore trace back the competition for labor. 0:07:33.120 --> 0:07:35.920 So that sort of data you would get that once 0:07:36.080 --> 0:07:39.800 an investigation had already been started. Talk to us about 0:07:40.240 --> 0:07:45.920 identifying potential investigations and what sort of information and data 0:07:46.080 --> 0:07:48.600 you would be looking at there, or maybe you're looking 0:07:48.680 --> 0:07:52.480 at something like the overall impact on the labor market 0:07:52.600 --> 0:07:55.080 or prices, which we've already discussed. 0:07:55.200 --> 0:07:58.600 Right, So for mergers, that's sort of the bread and butter. 0:07:59.120 --> 0:08:03.960 There is requirement for companies to file so called HSR 0:08:04.080 --> 0:08:07.600 forms for any transaction that's above some threshold, and so 0:08:07.720 --> 0:08:11.760 in those forms they give us some basic information about 0:08:11.960 --> 0:08:14.760 why they're seeking to merge, you know, what markets this 0:08:15.160 --> 0:08:17.680 might affect, and so on and so forth. And actually 0:08:17.680 --> 0:08:22.360 we now have new proposed additional information on labor so 0:08:22.400 --> 0:08:25.560 that it's easier to look at potential labor market effects. 0:08:25.720 --> 0:08:28.760 So based on that, there's an initial screening to see 0:08:28.800 --> 0:08:32.160 if we think that there might be potentially anti competitive 0:08:32.160 --> 0:08:35.240 effects from the merger. Then you know, there's an initial 0:08:35.320 --> 0:08:38.480 investigation phase and basically it goes from step to step 0:08:39.120 --> 0:08:41.920 until we are satisfied that it's likely not going to 0:08:42.000 --> 0:08:45.439 substantially lessen competition or it could end up in a 0:08:45.520 --> 0:08:48.920 lawsuit to try to block the merger if after having 0:08:48.960 --> 0:08:51.880 done a lot of analysis and data and interviews, we 0:08:52.000 --> 0:08:56.520 conclude that there's a substantial risk that the transaction, the merger, 0:08:56.559 --> 0:08:58.079 would lessen competition. 0:08:58.840 --> 0:09:02.200 Not necessarily a specific merger or an example, but just 0:09:02.360 --> 0:09:05.480 what would be a style type of something. You know, 0:09:05.559 --> 0:09:09.760 imagine company widget company A wants to merge with widget 0:09:09.800 --> 0:09:14.400 company B. They fill out this hs R form for you, 0:09:14.920 --> 0:09:18.360 what might be an example of something that shows up 0:09:18.880 --> 0:09:21.640 that crosses some threshold and in fact, is there a 0:09:21.679 --> 0:09:24.360 threshold like is there a level of X or Y 0:09:24.520 --> 0:09:26.760 in which like the red flag goes off, or is 0:09:26.800 --> 0:09:30.559 there some level of subjectivity walk us through like looking 0:09:30.559 --> 0:09:33.480 at that data when the data flashes some alarm to 0:09:33.520 --> 0:09:35.520 take this is not a good deal. 0:09:36.000 --> 0:09:39.400 So the main there are many, of course considerations that'd 0:09:39.400 --> 0:09:41.520 be hard to go through here, but one of the 0:09:41.600 --> 0:09:44.960 main things is the market shares. Okay, so if we 0:09:45.120 --> 0:09:49.360 have some evidence that these companies are competing in some 0:09:49.520 --> 0:09:52.640 market for some widgets, some specific widgets, so that's a 0:09:52.640 --> 0:09:55.880 specific product market, let's say, and if both of those 0:09:55.920 --> 0:09:59.760 companies have a high enough market share, that raises alarm 0:09:59.760 --> 0:10:01.960 and in the merger guidelines. So we have guidelines where 0:10:02.000 --> 0:10:04.520 we tell everybody this is public. You can go see them. 0:10:04.559 --> 0:10:06.679 They are brand new. We just put them out twenty 0:10:06.679 --> 0:10:09.520 twenty three Merger Guidelines. It tells you exactly what that 0:10:09.640 --> 0:10:13.679 threshold is. So what must your share be for us 0:10:13.720 --> 0:10:18.040 to conclude that there is a risk that competition might 0:10:18.080 --> 0:10:20.960 be hurt in this merger. And again, the general principle 0:10:21.080 --> 0:10:25.319 is that the two companies must have a high enough 0:10:25.360 --> 0:10:27.520 share in the market. So basically the market we call 0:10:27.559 --> 0:10:30.880 it concentrated, so there's few companies, each one has a 0:10:31.080 --> 0:10:34.920 large share. And further, we require that after the merger, 0:10:35.480 --> 0:10:39.440 the share of the two companies combined increases enough that 0:10:39.520 --> 0:10:44.040 it threatens competition. So that's really a key screening tool 0:10:44.280 --> 0:10:48.000 that we use, and that's going to play an important role, 0:10:48.559 --> 0:10:51.640 you know, as we go on with the investigation, and 0:10:51.800 --> 0:10:54.880 why are shared is important. They are important as an 0:10:54.920 --> 0:10:59.640 indicator of how anti competitive effects could happen. And this 0:10:59.760 --> 0:11:02.840 could be either through a reduction in what we call 0:11:02.880 --> 0:11:07.640 again competition direct competition for customers or for workers, or 0:11:07.960 --> 0:11:11.439 it could also be through potentially an increased risk of 0:11:11.600 --> 0:11:16.640 collision that whether explicit or implicit. The few actors they 0:11:16.679 --> 0:11:19.360 are in the market, and the easier it is for 0:11:19.400 --> 0:11:23.800 them to agree, let's say to you know, increased prices 0:11:23.800 --> 0:11:26.120 a little bit or let's say keep wages down a 0:11:26.160 --> 0:11:27.839 little bit more, something like that. 0:11:43.000 --> 0:11:46.920 What happens in a non competitive economy? Walk us through 0:11:46.920 --> 0:11:50.760 the sort of side effects of having high corporate concentration. 0:11:51.480 --> 0:11:55.760 Yeah, so there are many adverse effects, but broadly, some 0:11:55.840 --> 0:12:01.640 of these effects are higher prices, lower wages is less innovation, 0:12:02.000 --> 0:12:05.520 so you know, just broadly, and lower product quality, lower 0:12:05.520 --> 0:12:09.120 wage quality. So, for example, for the labor market, if 0:12:09.120 --> 0:12:11.599 we just take that as an example, because of a 0:12:11.679 --> 0:12:14.840 lack of competition, so if you're an employee and you 0:12:14.920 --> 0:12:16.920 don't have anywhere else to go, that would be a 0:12:16.920 --> 0:12:19.400 good firm that you would like. Similarly, just as well 0:12:19.440 --> 0:12:22.280 as your current employer. That means that the employer can 0:12:22.320 --> 0:12:24.160 pay you less because where are you going to go? 0:12:24.679 --> 0:12:27.439 And so you know, I have a paper, for example, 0:12:27.440 --> 0:12:31.200 and others have written related academic papers, and you can 0:12:31.240 --> 0:12:35.079 show that if we got rid of this lack of competition, 0:12:35.240 --> 0:12:39.120 wages may increase by as much as twenty percent on average. 0:12:39.200 --> 0:12:41.080 You know, in the extreme case that we could, you know, 0:12:41.120 --> 0:12:44.320 without all issues with competition in the labor market. Now, 0:12:44.559 --> 0:12:47.440 if you look at the product market, so you know 0:12:47.520 --> 0:12:50.840 there's a lack of competition in various markets for various widgets. 0:12:51.400 --> 0:12:56.160 Not only that will increase prices, reduce quality, and reduce 0:12:56.200 --> 0:13:00.480 consumer choice, reduce innovation, but interestingly, that will also have 0:13:00.559 --> 0:13:04.120 an adverse effect on workers, which is an indirect effect 0:13:04.160 --> 0:13:08.040 because when you have a monopoly, a monopoly makes less stuff. 0:13:08.080 --> 0:13:12.360 They produce less widgets, fewer widgets, and in order to 0:13:12.679 --> 0:13:16.360 produce fuel widgets, well, that also means fuel widgets means 0:13:16.360 --> 0:13:19.120 you need fewer workers to produce the wigets. So there's 0:13:19.200 --> 0:13:22.760 a less demand for labor, less demand for workers. So 0:13:22.800 --> 0:13:26.840 therefore that reduces employment opportunities. And there's really cool new 0:13:26.880 --> 0:13:30.439 academic work that's still work in progress by Tanya Babina 0:13:30.440 --> 0:13:33.959 and co authors suggesting that when the anti trust authorities 0:13:34.040 --> 0:13:36.920 go in and enforce the anti trust law and stimulate 0:13:37.400 --> 0:13:41.000 competition the product market, you see an increase in employment 0:13:41.600 --> 0:13:45.760 as business flourishes, and they also need more workers to 0:13:45.840 --> 0:13:47.680 produce whatever it is that they're producing. 0:13:48.240 --> 0:13:52.160 One of the arguments that you know, one thing that 0:13:52.200 --> 0:13:55.000 can theoretically happen. I don't know if it happens in practice. 0:13:55.080 --> 0:13:58.160 But one thing that can theoretically happen is that two 0:13:58.240 --> 0:14:02.439 companies can combine and become more efficient, and that means 0:14:02.440 --> 0:14:04.480 that they have economies of scale, and maybe they can 0:14:04.559 --> 0:14:09.559 lower prices. Maybe that scale brings consumer benefits with the 0:14:09.600 --> 0:14:13.480 lower prices, et cetera. From an economist perspective, I'm sure 0:14:13.520 --> 0:14:16.440 there are many instances where companies say that that's what 0:14:16.520 --> 0:14:20.120 will happen, and it seems at least possible. Is it 0:14:20.160 --> 0:14:23.480 possible in your view that concentration can lead to better 0:14:23.560 --> 0:14:27.840 consumer outcomes, either a from a price perspective or from 0:14:27.880 --> 0:14:30.200 a here's a very powerful company that then becomes a 0:14:30.240 --> 0:14:33.000 wage ceter or you just sort of would you be 0:14:33.080 --> 0:14:34.880 per se skeptical of those claims. 0:14:35.480 --> 0:14:39.320 So, you know, in general, as with every claim where 0:14:39.520 --> 0:14:42.040 you know the companies have an interest in making that claim, 0:14:42.240 --> 0:14:45.520 we need to examine it, you know, with fairness and 0:14:45.560 --> 0:14:47.720 with evidence. And you know, if you look at the 0:14:47.760 --> 0:14:51.520 merger Guidelands, there's a whole section on efficiencies, recognizing that 0:14:51.520 --> 0:14:55.160 that's a possibility, but we also ask that those efficiencies, 0:14:55.200 --> 0:15:00.400 if claimed, be well documented, you know, clearly articulated, verified 0:15:00.440 --> 0:15:04.080 by outside experts, so that we can start to potentially 0:15:04.120 --> 0:15:07.360 consider them, and we find that oftentimes that's just not 0:15:07.480 --> 0:15:10.360 the case. And so you know, again it's all a 0:15:10.440 --> 0:15:13.400 fact based investigation and you have to look at it. 0:15:13.680 --> 0:15:16.840 But that's certainly something that we consider in the holistic 0:15:16.880 --> 0:15:20.840 assessment of what's going to happen. Since it's a predictive exercise, 0:15:20.880 --> 0:15:23.360 we have to assess the risk. We take that into 0:15:23.360 --> 0:15:24.480 account as one element. 0:15:25.280 --> 0:15:29.240 Just to press you on this part further, as an 0:15:29.240 --> 0:15:31.680 economist coming to the table and there are other people, 0:15:31.760 --> 0:15:34.280 there are lawyers or others at the table, what are 0:15:34.320 --> 0:15:37.840 some of the tools in your arsenal that you might use, 0:15:37.960 --> 0:15:40.280 or tools in your toolbox per se, that you might 0:15:40.520 --> 0:15:44.280 use to make such a prediction to evaluate claims of 0:15:44.400 --> 0:15:46.440 economies of scale and such things. 0:15:47.040 --> 0:15:51.800 So what we essentially, it's important to note that it's 0:15:51.840 --> 0:15:54.880 not up to us the Department of Justice, to make 0:15:54.920 --> 0:15:58.480 that point. It's rather that the companies must be able 0:15:58.560 --> 0:16:02.960 to prove with some evidence that these efficiencies are likely 0:16:03.000 --> 0:16:06.120 to result and be not only that there are efficiency 0:16:06.480 --> 0:16:09.000 at all, which might well be the case, but that 0:16:09.040 --> 0:16:13.280 they are sufficiently substantial that they will negate the anti 0:16:13.280 --> 0:16:16.520 competitive effects that we fear, and so and of course, 0:16:16.600 --> 0:16:20.840 the specifics against specific But what's really important is not 0:16:20.960 --> 0:16:24.320 just to show that there are some efficiencies, but that 0:16:24.440 --> 0:16:29.240 they are again substantive enough that the initial concern that 0:16:29.280 --> 0:16:32.280 we had for all the adverse anti competitive effects is 0:16:32.440 --> 0:16:37.080 likely not to materialize because you know, those efficiencies make 0:16:37.120 --> 0:16:39.840 the market so much more competitive. But again the burden 0:16:39.920 --> 0:16:42.680 of proof in terms of who must show that that's 0:16:42.720 --> 0:16:44.400 the economists of the other side. 0:16:45.960 --> 0:16:49.240 So one of the more interesting cases that the DOJ 0:16:49.560 --> 0:16:52.960 took on over the past because it was a few 0:16:53.040 --> 0:16:58.360 years since it started, but the Penguins acquisition of Simon 0:16:58.400 --> 0:17:02.440 and Schuster, and part of this is media naval gazing 0:17:02.520 --> 0:17:05.680 because it involves the publishing industry, and if you read 0:17:05.760 --> 0:17:08.240 some of the complaints and the other documents around this, 0:17:08.480 --> 0:17:12.639 there is just oodles and oodles of information on like 0:17:12.760 --> 0:17:16.239 the actual numbers behind publishing and how they come up 0:17:16.280 --> 0:17:20.720 with things like author advances, how much they spend on marketing. 0:17:20.640 --> 0:17:21.440 Stuff like that. 0:17:21.840 --> 0:17:25.760 But walk us through the economic lessons of a case 0:17:26.119 --> 0:17:29.120 like that, because there seem to have been some trade 0:17:29.119 --> 0:17:30.720 offs in that particular action. 0:17:31.640 --> 0:17:34.080 So you know, I was really excited to see that case. 0:17:34.160 --> 0:17:37.239 This was ongoing when I arrived in my job at 0:17:37.280 --> 0:17:40.280 the division and just to table sat a little bit. 0:17:40.320 --> 0:17:44.199 That case was a case where Penguin Random House was 0:17:44.200 --> 0:17:47.320 seeking to merge with another big publisher, Simon and Schuster, 0:17:47.680 --> 0:17:51.040 and we always first look at competition in this industry, 0:17:51.080 --> 0:17:55.000 and in this case were interesting competition for authors. So 0:17:55.040 --> 0:17:58.119 you know this big publisher publishing houses, they are trying 0:17:58.160 --> 0:18:02.480 to buy the rights to author's books, and so for 0:18:03.080 --> 0:18:07.240 big name authors, they're on that many publishing houses. In fact, 0:18:07.240 --> 0:18:10.399 there are only five so called big five publishing companies 0:18:10.800 --> 0:18:15.560 that vuy typically for books by famous authors. And so 0:18:15.600 --> 0:18:20.359 we focus specifically on what we called anticipated top selling books, 0:18:20.359 --> 0:18:22.600 which were books that would get an advance of two 0:18:22.680 --> 0:18:25.399 hundred and fifty thousand dollars or more. So we're talking 0:18:25.440 --> 0:18:29.600 here about you know, likely big selling books. And so 0:18:29.760 --> 0:18:33.600 in that market again for those big big books, Penguin 0:18:33.640 --> 0:18:36.800 Random House had the market share there. We got market share, 0:18:36.840 --> 0:18:39.919 remember we talked about that before, of thirty seven percent 0:18:39.960 --> 0:18:42.840 and Simon and Schuster twelve percent, and so you can 0:18:42.920 --> 0:18:46.680 see that together after the merger, if it had happened, 0:18:46.680 --> 0:18:50.720 they would essentially have half of that market. And so 0:18:50.840 --> 0:18:55.399 that's like a really significant change in market shares. That 0:18:55.480 --> 0:18:58.760 would give these publishing houses significant market power. And what's 0:18:58.800 --> 0:19:03.600 the market power here, it's the market power over authors. 0:19:03.760 --> 0:19:06.960 When authors are shopping around the manuscript to find a 0:19:07.040 --> 0:19:10.840 publisher that they like that gives them good conditions. Again, 0:19:10.880 --> 0:19:12.600 that's what they do. They shop around and they make 0:19:12.640 --> 0:19:16.920 competition play who's giving me the best terms? And once 0:19:17.040 --> 0:19:20.720 these publishing houses would have merged, that would remove an 0:19:20.760 --> 0:19:24.919 option for the authors to get a better deal for 0:19:25.040 --> 0:19:28.160 their books. And so therefore we predicted that they would 0:19:28.240 --> 0:19:33.320 get lower advances, and that would therefore damage competition in 0:19:33.440 --> 0:19:36.560 that market for the author's work. And this is the 0:19:36.680 --> 0:19:39.000 judge agreed with us. And so this is the first 0:19:39.080 --> 0:19:43.640 merger that was blocked primarily on a theory of labor 0:19:44.200 --> 0:19:50.359 market power, where those publishing houses would exercise power over 0:19:50.520 --> 0:19:54.560 the labor of the authors in this specific case, is. 0:19:54.520 --> 0:19:57.639 There a you know you mentioned okay, a company with 0:19:57.720 --> 0:20:00.399 thirty seven percent market share want to merge with twelve 0:20:00.400 --> 0:20:03.719 percent market share almost fifty percent? Is there a number 0:20:04.119 --> 0:20:07.199 in which this is too much? Does it vary by industry? Like, 0:20:07.280 --> 0:20:11.439 where is this threshold exist such that you could say, okay, 0:20:11.480 --> 0:20:14.240 this is an okay level of market power or this 0:20:14.440 --> 0:20:15.560 is too much. 0:20:15.920 --> 0:20:19.240 Right, So that's in the merger guidelines. We have a 0:20:19.840 --> 0:20:22.359 famous you know, this is nerdy nerdy, but we have 0:20:22.440 --> 0:20:25.720 an index try The HHI is actually the Herd fineral 0:20:25.840 --> 0:20:29.960 Hershman index that essentially boils down to market shares. So 0:20:30.400 --> 0:20:33.880 if the market shares in the market are generally high enough, 0:20:34.000 --> 0:20:36.080 you're going to have a market that's already we call 0:20:36.119 --> 0:20:39.840 it highly concentrated. And then from there, if the merger 0:20:39.920 --> 0:20:43.040 is going to increase it sufficiently in this concentration, making 0:20:43.080 --> 0:20:47.800 the market even more concentrated with fewer players, then we 0:20:48.040 --> 0:20:50.560 think that it, you know, presumptively, is going to cause 0:20:50.560 --> 0:20:54.040 an anti competitive effect. And this, you know, this is 0:20:54.080 --> 0:20:57.120 this is something that you can calculate, and the thresholds 0:20:57.160 --> 0:20:58.719 are there in the merger guidelines. 0:20:58.880 --> 0:21:02.840 Wait, so, speaking of calculations and nerding out, when it 0:21:02.920 --> 0:21:07.240 comes time to try to predict the effect of this 0:21:07.320 --> 0:21:12.280 type of consolidation on advances for authors and their proposed books, 0:21:12.560 --> 0:21:16.480 presumably you go about actually modeling that in one way 0:21:16.600 --> 0:21:19.919 or another. How do you do something like that for publishing, 0:21:19.960 --> 0:21:24.560 which is kind of a notoriously opaque industry in many ways. 0:21:24.320 --> 0:21:27.400 That's right, So, you know, as I was saying, it's 0:21:27.440 --> 0:21:30.760 always very fact specific, and actually one of the coolest 0:21:30.800 --> 0:21:33.199 things that I've learned by being at the division is 0:21:33.320 --> 0:21:36.840 just the very many different details of each industry that 0:21:36.880 --> 0:21:40.119 you know, as an academic, you don't access this level 0:21:40.160 --> 0:21:43.320 of detail, you know, or manage as a journalist. If 0:21:43.359 --> 0:21:46.800 you could get all the you know, various emails and 0:21:46.880 --> 0:21:49.920 all these documents about what's really going on, it's really juicy. 0:21:50.400 --> 0:21:54.760 So but essentially here in this case, there's a process 0:21:54.760 --> 0:21:58.919 of bargaining and shopping around the manuscripts, and there's various 0:21:58.920 --> 0:22:01.960 ways how this could happen, but typically it's some sort 0:22:01.960 --> 0:22:07.119 of auction like format, and so therefore our expert in 0:22:07.160 --> 0:22:11.160 this case we work with an outside economic expert who 0:22:11.240 --> 0:22:14.200 does some of the analysis, and so he modeled an 0:22:14.240 --> 0:22:19.240 auction style situation where those firms, now instead of each 0:22:19.320 --> 0:22:24.199 bidding independently, would be together and calculated. You know, the 0:22:24.280 --> 0:22:28.280 fact that it would likely result in a significant decrease 0:22:28.359 --> 0:22:30.280 in what the authors were going to be able to 0:22:30.320 --> 0:22:32.880 get for their books. 0:22:33.400 --> 0:22:36.200 You know, one thing in staying with the books example, 0:22:36.359 --> 0:22:39.600 but maybe it's not the best example. Another element of 0:22:39.640 --> 0:22:43.640 these fights is fights over the definition of what constitutes 0:22:43.680 --> 0:22:46.480 a market, and maybe in books, it's kind of clear. 0:22:47.040 --> 0:22:50.560 We're talking about books and there's five big publishers, and so, okay, 0:22:50.800 --> 0:22:55.199 there's an advance and everything. But one could theoretically make 0:22:55.240 --> 0:22:58.280 the argument that the market is more than just five 0:22:58.280 --> 0:23:01.320 book publishers. And we live in an age of social 0:23:01.359 --> 0:23:04.720 media and writers maybe not doing a book, maybe doing 0:23:04.800 --> 0:23:08.480 a deal with Netflix and turning it into a documentary 0:23:08.920 --> 0:23:12.840 or other vehicles that are not precisely booked, but that 0:23:12.960 --> 0:23:16.600 may be a vehicle for an author or researcher or 0:23:16.720 --> 0:23:19.880 writer to get there out. Where does the fight over 0:23:19.960 --> 0:23:23.680 what constitutes the market itself occur or when does that occur? 0:23:24.240 --> 0:23:27.439 Well, you're really pointing the you know, putting your finger here. 0:23:27.480 --> 0:23:30.440 You're putting your finger on a fundamental question in those 0:23:30.720 --> 0:23:34.679 anti trust cases, which is what's the market. And importantly 0:23:34.720 --> 0:23:37.480 it's it's not like there is the market, it's we 0:23:37.560 --> 0:23:41.439 have to show a relevant market based on how competition works. 0:23:41.480 --> 0:23:45.919