1 00:00:20,460 --> 00:00:29,070 See you here in a event which is outside our normal series because we wanted to create something to take the 2 00:00:29,070 --> 00:00:37,170 opportunity of Professor Jason Furman being in Oxford today to give him an opportunity to share his thoughts with you. 3 00:00:37,170 --> 00:00:49,380 As you know, Jason is not only original thinker, but he's a doer and has been quite instrumental in many domains. 4 00:00:49,380 --> 00:01:00,780 Not pleased with President Obama in the US and more recently in the review of the digital economy for the Secretary of State of Business, 5 00:01:00,780 --> 00:01:11,190 as well as the chancellor in the UK. This combination of interest in ideas being at the forefront of ideas and, uh, 6 00:01:11,190 --> 00:01:15,510 changing things in the real world is, of course, dear to the heart of the Oxford Martin School, 7 00:01:15,510 --> 00:01:21,180 which was created for that purpose and to the programme, which I now direct in the school, 8 00:01:21,180 --> 00:01:25,770 which is on the intersection of technological and economic change. 9 00:01:25,770 --> 00:01:31,980 Jason is the Professor of Practise of Economic Policy at Harvard University at the Kennedy School. 10 00:01:31,980 --> 00:01:37,200 He's a non-resident fellow at the Peterson Institute. And he's also this coming. 11 00:01:37,200 --> 00:01:43,170 You're going to be responsible for introductory economics for Harvard with David Lipson, 12 00:01:43,170 --> 00:01:47,820 who was part of the presidential campaign for President Obama. 13 00:01:47,820 --> 00:01:58,350 And before he became president and then chairman of the Council of Economic Advisers from 2013 to 2017 and for many other duties as well. 14 00:01:58,350 --> 00:02:01,650 Interestingly enough, he not only worked with President Obama, 15 00:02:01,650 --> 00:02:10,560 but he also worked with President Clinton and Aspirin President Gore in his campaign and of course, 16 00:02:10,560 --> 00:02:15,330 with many others that we know well, including Joe Stiglitz at the World Bank. 17 00:02:15,330 --> 00:02:21,600 And then subsequently, he's been director of the Hamilton Project, which is an economic policy research group. 18 00:02:21,600 --> 00:02:27,000 Focussing on the intersection of economic policies and visions of shared growth 19 00:02:27,000 --> 00:02:31,530 has been senior fellow at the Brookings Institution and taught at many, 20 00:02:31,530 --> 00:02:38,040 many universities extremely wide ranging interests, including health economics, tax policy, 21 00:02:38,040 --> 00:02:45,610 technology policy, domestic macro and micro, etc. published many, many articles and books. 22 00:02:45,610 --> 00:02:54,720 And in addition to his Ph.D. from Harvard, has a Masters from LSC and an additional one in public policy from Harvard. 23 00:02:54,720 --> 00:03:02,820 Together with Diane Coyle, who's speaking here on the 66th of June, so in two weeks on Thursday. 24 00:03:02,820 --> 00:03:15,000 And Philip Marsden, who's here? He chaired the review expert panel as it was called on competition in the digital economy and 25 00:03:15,000 --> 00:03:22,470 had a very significant impact that I hope with its 20 recommendations [INAUDIBLE] share with us. 26 00:03:22,470 --> 00:03:29,880 And then I hope it impacts on the UK if those responsible for it are not too distracted by other activities. 27 00:03:29,880 --> 00:03:35,320 Jason, it's a huge pleasure to have you here and welcome you to the stage. 28 00:03:35,320 --> 00:03:43,560 Yeah. This is been live cost of gets a Q&A. 29 00:03:43,560 --> 00:03:47,790 Be aware of that when it happens. I'm grateful. 30 00:03:47,790 --> 00:03:55,890 Ian, thank you for that introduction. Thank you for hosting me here at the Martin School and to have a chance to 31 00:03:55,890 --> 00:04:02,640 really talk through the recommendations that we made for digital competition. 32 00:04:02,640 --> 00:04:07,930 This is something that an expert panel that I chair that includes Philip Marsden, who is here. 33 00:04:07,930 --> 00:04:17,610 I'm Diane Coyle, who is mentioned, also an economist, Amelia Fletcher and a computer scientist named Derek McCauley came up with. 34 00:04:17,610 --> 00:04:29,010 Before I get to our recommendations, I wanted to situate my own personal work in this space with how I came to it. 35 00:04:29,010 --> 00:04:38,310 As Ian said, I have a wide ranging set of interests in economics and in economic policy. 36 00:04:38,310 --> 00:04:49,470 And for me, the single animating question has been why income growth has slowed so much in the advanced economies. 37 00:04:49,470 --> 00:04:56,490 The United States incomes used to grow at three percent a year. Now, incomes grow at about half a percent a year. 38 00:04:56,490 --> 00:05:01,440 That goes from doubling once a generation to taking over a century. 39 00:05:01,440 --> 00:05:16,050 For incomes to double if you look one level further down, you find that this slowdown in income growth is a combination of slower productivity growth, 40 00:05:16,050 --> 00:05:22,900 which started slowing around the 1970s and has been in quite severe slowdown for the last decade. 41 00:05:22,900 --> 00:05:28,660 And also rising and high levels of inequality. 42 00:05:28,660 --> 00:05:40,620 So the combination of the pie growing less quickly and the pie being more unequally divided is what lies behind this. 43 00:05:40,620 --> 00:05:46,500 To me, that was a little bit like a massive crime and put me in almost a paranoid mindset as 44 00:05:46,500 --> 00:05:53,430 I walked around looking for suspects that could possibly have committed that crime. 45 00:05:53,430 --> 00:06:04,170 I eventually gave up looking for any one suspect that committed it, but I came to understand that it was many, many small things that contributed. 46 00:06:04,170 --> 00:06:13,650 And in particular, if you could find anything that potentially played a role in both productivity and in inequality, 47 00:06:13,650 --> 00:06:25,470 that that would be quite important. Joe Stiglitz was celebrating the 50th anniversary of his teaching in two thousand fifteen, 48 00:06:25,470 --> 00:06:34,170 and I decided with a former colleague and friend of mine, Peter Orszag, to do a paper in honour of Joe Stiglitz for that conference. 49 00:06:34,170 --> 00:06:44,010 And what we focussed on was whether there was a link between a lot of the micro issues that Joe was interested in that taught us about that. 50 00:06:44,010 --> 00:06:51,420 I worked some with him in terms of concentration at the micro economic level in the economy. 51 00:06:51,420 --> 00:07:01,560 And then some of the macro themes that had preoccupied Joe like slower growth and especially in his case, rising inequality. 52 00:07:01,560 --> 00:07:08,730 And in this first part of the talk, I'm going to talk about that increased concentration and the analysis. 53 00:07:08,730 --> 00:07:14,790 We had some of the research since then that documents the increase in concentration 54 00:07:14,790 --> 00:07:21,780 throughout the economy and tries to tease out some of the consequences of it. 55 00:07:21,780 --> 00:07:31,920 That'll be a segue way to the expert panel, which is focussed in particular on one sector, the technology sector. 56 00:07:31,920 --> 00:07:39,210 So I'll tell you a little bit about what we were set up to do, go through some of the major questions that we had to answer. 57 00:07:39,210 --> 00:07:45,720 Talk about two sets of our proposals X Anti-regulatory X Plus X Post Merger Enforcement. 58 00:07:45,720 --> 00:07:50,190 And finally, talk about next steps. 59 00:07:50,190 --> 00:08:02,710 So starting with that increased concentration, I'm now going to show you some pretty definitive evidence for it. 60 00:08:02,710 --> 00:08:06,670 Those are the two beer companies we have right now. 61 00:08:06,670 --> 00:08:09,790 One is called InBev. The other is called Heineken. 62 00:08:09,790 --> 00:08:15,340 There's a lot of other, at least in the United States, craft beers that look like they're from small breweries. 63 00:08:15,340 --> 00:08:21,910 But most of them have been purchased by one of these two companies. 64 00:08:21,910 --> 00:08:29,320 This trend of many companies consolidating down in this case for the most part to two companies is something 65 00:08:29,320 --> 00:08:40,930 you see in lots of segments of the economy in that paper for the Stiglitz Fest Shrift in two thousand fifteen. 66 00:08:40,930 --> 00:08:49,690 I put together this chart. It shows something like 13 sectors and in 12 of them, concentration was increasing. 67 00:08:49,690 --> 00:08:54,730 That is a smaller number of firms or firms taking a larger share. 68 00:08:54,730 --> 00:09:03,370 And in one of them, it wasn't. At the time, I was very clear that and in a subsequent report, the Council of Economic Advisers put out, 69 00:09:03,370 --> 00:09:09,400 it's not like you would bring an antitrust case against the transportation and warehousing industry 70 00:09:09,400 --> 00:09:15,940 based on the fact that the 50 largest firms increase their market share by 11 percentage points. 71 00:09:15,940 --> 00:09:23,170 You can have increased market share at this high level of aggregation and more competition in any given area. 72 00:09:23,170 --> 00:09:33,760 Conversely, you could have a decline in concentration measured in this way and actually have, you know, less choices for consumers. 73 00:09:33,760 --> 00:09:39,160 So the question in this isn't, you know, this is definitive proof of anything. 74 00:09:39,160 --> 00:09:45,010 Of course it's not. It's is this suggestive. You know, what does it look like? 75 00:09:45,010 --> 00:09:50,290 Since then, there have been a number of papers which have dug a couple levels deeper. 76 00:09:50,290 --> 00:09:54,400 These are again all using aggregate macroeconomic data. 77 00:09:54,400 --> 00:09:58,870 And they found things like in seventy five percent of U.S. industries, 78 00:09:58,870 --> 00:10:09,520 there's been an increase in concentration level at the three digit industry level that the C.R. for the C.R. 40, 79 00:10:09,520 --> 00:10:13,540 the share revenue share for the top four and 40 firms in the HHI, 80 00:10:13,540 --> 00:10:22,960 which is a measure of how much competition there is in a sector, have trended up for six hundred and seventy six industries. 81 00:10:22,960 --> 00:10:33,630 Another paper did add a more disaggregated basis and found increases in the HHI, et cetera. 82 00:10:33,630 --> 00:10:42,540 These are all macro papers. These wouldn't stand up any better in a court of law, nor should they than the last set of papers I showed you. 83 00:10:42,540 --> 00:10:47,100 The question is not can you poke holes in these types of measures? 84 00:10:47,100 --> 00:10:52,530 And a lot of people have poked holes in these types of measures. It's like anything in economics. 85 00:10:52,530 --> 00:10:55,260 Are these measures useful or do they tell you something? 86 00:10:55,260 --> 00:11:02,460 And you know, these measures tend to be correlated with, you know, industries that are more concentrated, having higher mark-ups, 87 00:11:02,460 --> 00:11:10,920 having higher profit rates, having, you know, lower investment relative to what you would expect them to have. 88 00:11:10,920 --> 00:11:17,640 So it's a reasonably good explanatory variable for some of what we want. 89 00:11:17,640 --> 00:11:21,720 There are a set of papers, and the point of this slide is not for you to read it, 90 00:11:21,720 --> 00:11:28,830 but for you to have the sense that there's a lot of things on it that, you know, have that. 91 00:11:28,830 --> 00:11:34,530 Iowa economists have done the types of studies that are closer to what holds up in a 92 00:11:34,530 --> 00:11:42,060 court of law that looks in a really detailed way at a well-defined market and asks, 93 00:11:42,060 --> 00:11:45,570 you know, is there more concentration in it? 94 00:11:45,570 --> 00:11:51,990 And it's something we've seen in agriculture, appliances, beer, fertiliser, financial services, 95 00:11:51,990 --> 00:11:58,670 hospitals, railroads, wireless, you know, amongst other industries in the economy. 96 00:11:58,670 --> 00:12:08,360 So, you know, trying to look at the economy from the firm level, from the micro level in aggregate, it all up is tricky. 97 00:12:08,360 --> 00:12:15,200 You can either take a really close look at, you know, one piece of the economy, which is what a lot of these papers do. 98 00:12:15,200 --> 00:12:20,540 You or you can take a really blurry look at the economy of a whole. That's what the others did. 99 00:12:20,540 --> 00:12:25,630 Both of those types of resolution seem to be. 100 00:12:25,630 --> 00:12:29,670 Telling you a similar thing. 101 00:12:29,670 --> 00:12:38,940 The question, then, is why has concentration gone up, why in most industries are there fewer firms today than there were in the past? 102 00:12:38,940 --> 00:12:45,040 And they have, you know, by definition, larger shares? 103 00:12:45,040 --> 00:12:56,440 In part, it's for good and natural reasons. And in fact, you can have more competition leads to greater concentration. 104 00:12:56,440 --> 00:13:02,800 You could imagine a sector like retail where there's a lot of small businesses. 105 00:13:02,800 --> 00:13:08,650 They're not very competitive with each other. Consumers don't have a lot of choices. 106 00:13:08,650 --> 00:13:14,350 And then one or two big retail giants come in that are really, really efficient. 107 00:13:14,350 --> 00:13:24,130 That increased competition that give consumers new choices and in the process, knock some of the smaller, inefficient businesses out. 108 00:13:24,130 --> 00:13:34,660 You would observe in that sector an increase in concentration that increasing concentration would be evidence that there was more competition. 109 00:13:34,660 --> 00:13:43,060 And, you know, broadly speaking, would be a good thing. This is something you see under the heading of superstar firms. 110 00:13:43,060 --> 00:13:47,350 David Autor and Company and McKinsey have done analysis of that. 111 00:13:47,350 --> 00:13:53,440 Globalisation can be a force for greater competition that can lead to increased 112 00:13:53,440 --> 00:14:02,140 consolidation and finally increasing returns to scale and network externalities. 113 00:14:02,140 --> 00:14:08,860 There's also a set of bad and unnatural causes for increased concentration. 114 00:14:08,860 --> 00:14:17,310 If you used to stop a company from merging. And now you have some new theory and you allow the merger. 115 00:14:17,310 --> 00:14:21,540 You're going to observe an increase in concentration that increase in 116 00:14:21,540 --> 00:14:26,640 concentration isn't going to reflect any of the superstar firms or globalisation. 117 00:14:26,640 --> 00:14:36,970 It's just going to reflect a policy shift. Similarly, if you have less antitrust enforcement, so firms used to be able to, you know, 118 00:14:36,970 --> 00:14:45,260 raise it used to be stopped when they were trying to raise rivals costs or exclude rivals, and now they no longer are. 119 00:14:45,260 --> 00:14:51,960 You could have more regulations that create more barriers to entry in a sector. 120 00:14:51,960 --> 00:14:57,180 And I put to question Marks on it, because I think the jury is out on this literature, 121 00:14:57,180 --> 00:15:04,860 but there's a large literature on the fact that with the rise of huge asset managers and especially index funds, 122 00:15:04,860 --> 00:15:10,830 that even if there is competition between different companies, those companies all have the same owners. 123 00:15:10,830 --> 00:15:18,690 So the airlines are all owned by BlackRock and Warren Buffett and State Street and BlackRock. 124 00:15:18,690 --> 00:15:25,350 Warren Buffett and State Street don't have an interest in the airlines competing, no matter how many airlines they were. 125 00:15:25,350 --> 00:15:38,740 If there is that number of owners? So there's a set of bad and unnatural causes as well that cause less competition. 126 00:15:38,740 --> 00:15:49,950 These stories are different in different sectors in the US economy, I think retail fits the good causes story quite well. 127 00:15:49,950 --> 00:15:55,920 And hospitals fit the bad causes story quite well. 128 00:15:55,920 --> 00:16:03,570 Tyler Cowen has a new book out called Big Business, and he calls it a love story to an underappreciated institution. 129 00:16:03,570 --> 00:16:13,830 He is much less worried about monopolisation of the economy than I am, but he also uses the example of hospitals as an exception to his thesis. 130 00:16:13,830 --> 00:16:18,340 An example of something where we used to stop mergers. Now we allow them. 131 00:16:18,340 --> 00:16:26,690 There's not really a whole lot of efficiency benefits to them. There are higher prices as a result. 132 00:16:26,690 --> 00:16:36,600 So we're now 15 minutes into this talk, and I haven't mentioned the tech sector, which is the topic of the report. 133 00:16:36,600 --> 00:16:43,170 Well, tech has a combination of good and bad reasons for concentration, and at a high level, 134 00:16:43,170 --> 00:16:50,610 the key to policy is preserving the good reasons for the increased concentration in tech. 135 00:16:50,610 --> 00:16:57,360 You know, the efficiencies, the benefit you have from being on a social network with your friends, 136 00:16:57,360 --> 00:17:06,930 the benefits of the scope that a bunch of products can get bundled in together in a very seamless, convenient way for you. 137 00:17:06,930 --> 00:17:11,940 So those are all good reasons for scale. But then there's a set of bad reasons, 138 00:17:11,940 --> 00:17:23,700 ways in which there have been killer acquisitions or behaviour by the incumbents or others that I'll go into in a moment that we want to get rid of. 139 00:17:23,700 --> 00:17:31,500 And in some sense, if you can design policy to get rid of the bad reasons and preserve the good reasons, 140 00:17:31,500 --> 00:17:35,070 you can be a little bit agnostic as to what the outcome is. 141 00:17:35,070 --> 00:17:41,520 Rather than saying, you know, we know in this industry there's going to be 20 companies or there's going to be two companies know. 142 00:17:41,520 --> 00:17:52,110 We know in this industry we want to do everything we can to give entry and competition a chance and then we'll see how many there are. 143 00:17:52,110 --> 00:18:00,840 There are. And that's part of why, you know, in proceeding on this, I wouldn't start from breaking the companies up and deciding that, 144 00:18:00,840 --> 00:18:06,840 you know, we know the efficient scale that would get rid of some of the good reasons for that scale. 145 00:18:06,840 --> 00:18:12,930 I would address the bad ones and then see what happens. 146 00:18:12,930 --> 00:18:23,910 So, you know, a year ago, I could have given the talk I just gave you right now, and I would have ended right here and said, You know, 147 00:18:23,910 --> 00:18:31,170 I have a good principle for how I think about digital competition, which is get rid of the bad things and keep the good things. 148 00:18:31,170 --> 00:18:38,310 Sort of hard to argue with the idea that you want to get rid of the bad things that also are equally hard to argue with the idea. 149 00:18:38,310 --> 00:18:46,140 You want to keep the good things. The trick is what does that mean and how would you do that in practise? 150 00:18:46,140 --> 00:18:52,200 And so I was excited when the chancellor of the Exchequer asked me if I wanted to chair a digital 151 00:18:52,200 --> 00:19:00,300 panel to answer exactly this question went into that discussion with a strong belief in competition, 152 00:19:00,300 --> 00:19:11,550 a strong belief that left to itself, we wouldn't have competition, but pretty much no strong beliefs at all beyond that in terms of the form. 153 00:19:11,550 --> 00:19:15,540 So I'll tell you a little bit about the expert panel. 154 00:19:15,540 --> 00:19:22,740 Our terms of reference asked us to consider the potential opportunities and challenges as the emerging 155 00:19:22,740 --> 00:19:31,230 digital economy may pose for competition and pro competition policy and to make recommendations. 156 00:19:31,230 --> 00:19:36,330 In areas like online advertising, mergers, takeover and anti-competitive practises. 157 00:19:36,330 --> 00:19:39,590 Opportunities to enhance competition. 158 00:19:39,590 --> 00:19:49,940 And how to assess consumer impacts in services that are you always have to do the air quotes around free for consumers. 159 00:19:49,940 --> 00:19:55,310 The panel have already named everyone, I think this is the only extant photo of the five of us. 160 00:19:55,310 --> 00:20:03,170 But Philip, you can see, as I said, it brought computer science, law, economics together, had an excellent secretariat. 161 00:20:03,170 --> 00:20:10,980 There are no extant photos of them. And we had 11 evidence gathering roundtables in London. 162 00:20:10,980 --> 00:20:18,930 We got more than 60 written pieces of evidence submitted, many of them really quite thoughtful and quite extensive. 163 00:20:18,930 --> 00:20:24,810 We consulted, you know, many other experts. 164 00:20:24,810 --> 00:20:32,940 John Vickers is not responsible for any of the errors amongst them and engaged with governments in the United States, 165 00:20:32,940 --> 00:20:41,610 France, Australia, the OECD and the G7 in the course of it. 166 00:20:41,610 --> 00:20:48,270 For me, it was interesting because this wasn't just one part of the government that cared. 167 00:20:48,270 --> 00:20:57,450 We were reporting into the Treasury and Bays, the Department for Digital, DCMS, digital, culture, media and Sport. 168 00:20:57,450 --> 00:21:03,630 And then three regulators were interested to various degrees CMA, Ofcom and ICAO. 169 00:21:03,630 --> 00:21:11,280 So there was a lot of trying to understand the existing functions and navigate a 170 00:21:11,280 --> 00:21:18,960 complex institutional landscape and a distinction between government and regulators. 171 00:21:18,960 --> 00:21:25,670 And in March, we came out with our report unlocking digital competition. 172 00:21:25,670 --> 00:21:32,170 It's a report that, you know, is it's free, so you can all read it. 173 00:21:32,170 --> 00:21:38,460 OK. You know, in coming up with recommendations, you want to start, 174 00:21:38,460 --> 00:21:43,110 obviously with an analysis of the economy and I think different people who have 175 00:21:43,110 --> 00:21:48,780 different ideas about what should or shouldn't be done in the digital economy. 176 00:21:48,780 --> 00:21:54,810 Answer the following four questions in different ways from each other. 177 00:21:54,810 --> 00:22:02,020 The first question is, is competition in the digital sector beneficial? The second is, is it absent? 178 00:22:02,020 --> 00:22:06,290 The third is, is the lack of competition costly? 179 00:22:06,290 --> 00:22:13,280 Presumes an answer to the second one, and the last is, can there be competition in the digital sector? 180 00:22:13,280 --> 00:22:19,170 Just want to quickly take you through how I would answer each of these questions? 181 00:22:19,170 --> 00:22:24,780 You know, when is competition beneficial? Competition is beneficial. 182 00:22:24,780 --> 00:22:32,130 It lower prices, which give you higher quantities, a greater quality and greater variety. 183 00:22:32,130 --> 00:22:41,310 If you think consumer preferences are good, then competition is a great way to satisfy those consumer preferences. 184 00:22:41,310 --> 00:22:49,320 If you think consumer preferences are bad, then competition won't be helpful, and in fact, competition might be harmful. 185 00:22:49,320 --> 00:22:54,540 And so people use digital platforms for different purposes. 186 00:22:54,540 --> 00:23:00,420 That guy on the left is shopping and that guy on the right is clearly a terrorist. 187 00:23:00,420 --> 00:23:10,770 If what you're worried about online is sharing information about how to commit terrorist attacks or child pornography or bullying or the like, 188 00:23:10,770 --> 00:23:13,890 competition policy isn't going to be the solution for you. 189 00:23:13,890 --> 00:23:21,750 In fact, competition could even make all of those worse by lowering the price, increasing the quantity, increasing the variety. 190 00:23:21,750 --> 00:23:30,030 If what you're worried about is people connecting with friends, people shopping, people trying to find jobs, 191 00:23:30,030 --> 00:23:37,050 people trying to open up their own small businesses, people trying to, you know, rate and establish trust networks. 192 00:23:37,050 --> 00:23:47,850 People who care about privacy, which is something that consumers care about than competition, can help with all of those. 193 00:23:47,850 --> 00:23:54,360 When you go to some of the discussions of all the different things people are worried about in the digital economy. 194 00:23:54,360 --> 00:24:00,600 A lot of them are in the online harms type of sphere, not in the standard. 195 00:24:00,600 --> 00:24:08,130 What consumers care about privacy would be something that's in the overlap of the two of them, because I think consumers do care about privacy. 196 00:24:08,130 --> 00:24:13,020 They often don't have choices, like a friend of mine who, you know, 197 00:24:13,020 --> 00:24:20,430 announced on Facebook he was quitting Facebook because he was sick of all of the terrible things the company was doing. 198 00:24:20,430 --> 00:24:26,340 And from now on, he said, you can follow me on Instagram. 199 00:24:26,340 --> 00:24:30,810 So, you know, there's just not a lot of choices in that regard. 200 00:24:30,810 --> 00:24:37,950 But, you know, sometimes you go to conferences that are competition conferences and you have a lot of people wander around with hammers and to them, 201 00:24:37,950 --> 00:24:48,660 everything looks like a nail. And you know, the answer to this first question of is competition beneficial is in a lot of cases, yes. 202 00:24:48,660 --> 00:24:53,880 In some cases it's irrelevant. In some cases, it may even be counterproductive. 203 00:24:53,880 --> 00:25:01,660 And so nothing I'm going to talk about is a substitute for anything you want to do about, you know. 204 00:25:01,660 --> 00:25:09,400 Cyber terrorism, child pornography and possibly even issues like privacy. 205 00:25:09,400 --> 00:25:17,710 And in fact, if you do more to increase competition, you're probably going to need to do not less but do even more on all of those. 206 00:25:17,710 --> 00:25:23,650 It's a little bit easier to police some of that conduct if you have a single monopolist that has 207 00:25:23,650 --> 00:25:28,660 a lot of rents that they can devote to it and that can have a relationship with the government. 208 00:25:28,660 --> 00:25:34,360 Once you have 12 companies, you know, doing search, doing social networking during messaging, 209 00:25:34,360 --> 00:25:41,200 you might need to do even more of the non-competition regulation as well. 210 00:25:41,200 --> 00:25:43,990 So that's the first question, is competition in the digital sector beneficial? 211 00:25:43,990 --> 00:25:48,970 I think in many cases, yes, but I want to start out that that's not always the case. 212 00:25:48,970 --> 00:25:55,650 The second question is, is competition in the digital sector absent? 213 00:25:55,650 --> 00:26:03,180 There's a number of ways to look at this. I'm going to start out looking at a point in time. 214 00:26:03,180 --> 00:26:12,510 You know, this is the share time spent on popular social media services on Facebook and Instagram are one company, of course. 215 00:26:12,510 --> 00:26:16,140 And that 75 percent snap is 20 percent. 216 00:26:16,140 --> 00:26:21,390 Twitter's even less. And Pinterest is tiny. 217 00:26:21,390 --> 00:26:27,330 This data from our report you can look at a lot of different segments like online search, 218 00:26:27,330 --> 00:26:31,380 digital advertising, mobile operating systems and social media. 219 00:26:31,380 --> 00:26:38,280 And you see, increasingly the market share of the top two companies is approaching zero. 220 00:26:38,280 --> 00:26:46,890 And in all of these segments, the top two companies are drawn pretty much from one of five companies in all of them. 221 00:26:46,890 --> 00:26:55,050 The dominance has been stable now for 15 years, and most of them have had stable now for 15 years or so. 222 00:26:55,050 --> 00:27:03,180 And in many of them, there seem to be characteristics of the market that make it winners, take most or make the market tip, 223 00:27:03,180 --> 00:27:09,930 whether that's the network externality that you benefit from your friends being on the network, 224 00:27:09,930 --> 00:27:15,120 that advertisers benefit from advertising on the thing that has a lot of people on it. 225 00:27:15,120 --> 00:27:20,790 Something like mobile operating systems, Microsoft had a quite good mobile operating system. 226 00:27:20,790 --> 00:27:26,370 They had to pay people to write apps for their system because their market share was, I don't know, 227 00:27:26,370 --> 00:27:34,290 twenty twenty five percent, whereas people were willing to pay Apple to write for iOS. 228 00:27:34,290 --> 00:27:43,230 It seemed in mobile operating that having a third system, even if it was quite good, even if was supported by a major company, wasn't enough. 229 00:27:43,230 --> 00:27:47,460 Google tried to enter social networks, couldn't survive there. 230 00:27:47,460 --> 00:27:53,010 So there's a set of powerful forces that tend to lead to tipping that tend to lead 231 00:27:53,010 --> 00:27:59,880 to market take most everything I'm showing you showed you like social media search, 232 00:27:59,880 --> 00:28:06,210 mobile operating. These are all particular activities of companies at their heart. 233 00:28:06,210 --> 00:28:10,140 Some of the major companies here are actually advertising companies. 234 00:28:10,140 --> 00:28:15,510 They're not search, messaging, social networks and the like. 235 00:28:15,510 --> 00:28:29,490 And in online advertising, you're something very close to a duopoly, with Google dominating and Facebook reasonably large as well. 236 00:28:29,490 --> 00:28:38,490 The tougher question to ask is, you know, I've talked about whether this competition in the market, 237 00:28:38,490 --> 00:28:44,220 you know, right now you have a choice of buying Pepsi, Coca Cola or other. 238 00:28:44,220 --> 00:28:51,600 There is much less of that choice in digital platforms right now, so there's not a lot of competition in the market. 239 00:28:51,600 --> 00:29:02,490 The hopeful story would be that this competition for the market you see here that earlier on MySpace was the dominant social network. 240 00:29:02,490 --> 00:29:09,780 But then MySpace, you know, Facebook emerged and MySpace collapsed down to nothing. 241 00:29:09,780 --> 00:29:13,140 This is the idea with Google. The competition is just a click away. 242 00:29:13,140 --> 00:29:21,780 Maybe 90 percent of searches on Google now, but a good search engine comes tomorrow, and everyone en masse could switch over to that. 243 00:29:21,780 --> 00:29:28,740 Everyone en masse could switch over to a separate social network other than Facebook. 244 00:29:28,740 --> 00:29:33,180 And Instagram, of course, doesn't count tomorrow. 245 00:29:33,180 --> 00:29:43,680 So is that possible? I think the answer is probably not, or probably not, at least given the current policies which I'll come to. 246 00:29:43,680 --> 00:29:52,440 In part, the examples we have of that in the platform space were very early on in the development of the internet. 247 00:29:52,440 --> 00:29:56,460 You know, the fact that Lycos didn't have a lasting monopoly on search, 248 00:29:56,460 --> 00:30:04,620 I don't think is a particularly good reason argument against thinking that Google is pretty entrenched in the search business. 249 00:30:04,620 --> 00:30:15,230 These are companies that have expanded their reach, expanded their scope, have bought potential competitors have. 250 00:30:15,230 --> 00:30:25,850 Lots and lots of data, some of which has increasing returns to scale associated with it and thus serves as a barrier to entry. 251 00:30:25,850 --> 00:30:34,190 And so, you know, we can't be sure. You know, I'm sure if we waited long enough, some of these companies would be unseated in some ways. 252 00:30:34,190 --> 00:30:46,640 But, you know, competition for the market doesn't seem to be a whole lot more present than the competition in the market. 253 00:30:46,640 --> 00:30:56,730 So what? These companies give products that people like, the products are free in air quotes. 254 00:30:56,730 --> 00:31:03,420 Should we even care about this or be bothered by this? 255 00:31:03,420 --> 00:31:10,410 I think the answer to that is yes. And that it is costly and there's a number of different costs. 256 00:31:10,410 --> 00:31:16,380 The first one. Fiona Scott Morton is fond of pointing out that zero on, as she puts it, 257 00:31:16,380 --> 00:31:21,780 is just another no nos to the left of zero and those numbers to the right of zero. 258 00:31:21,780 --> 00:31:26,070 If you observe a price of zero, that doesn't mean the price isn't too high. 259 00:31:26,070 --> 00:31:28,230 Might have been that the price would have been negative. 260 00:31:28,230 --> 00:31:38,040 You would have been paid for the use of your data paid to be exposed to advertisers and because of lack of competition, 261 00:31:38,040 --> 00:31:43,790 you were overpaying at a price of zero. 262 00:31:43,790 --> 00:31:50,510 People aren't just paying relat over necessarily overpaying relative to a hypothetical. 263 00:31:50,510 --> 00:31:58,310 You know, when you have an advertising duopoly, they'll be able to charge more for advertising that is going to be paid out of something somewhere, 264 00:31:58,310 --> 00:32:02,090 maybe built into the mark-ups for those products themselves. 265 00:32:02,090 --> 00:32:06,890 So it might be the things you're buying that you saw in the online ad you're 266 00:32:06,890 --> 00:32:11,990 paying more for because the online ad itself cost more because of the duopoly, 267 00:32:11,990 --> 00:32:23,210 so there might even be some cash. Consumers are paying in terms of data and privacy in ways that they partly understand, partly don't know. 268 00:32:23,210 --> 00:32:30,770 Some of this is a counterfactual. What would the world be like if we had more competition? 269 00:32:30,770 --> 00:32:33,050 Quality and variety might be better. 270 00:32:33,050 --> 00:32:42,110 I certainly think, you know, something like the Facebook feed has basically not improved in the last five or more years. 271 00:32:42,110 --> 00:32:48,530 Google searches have sort of more and more advertisements and have improved in other dimensions, 272 00:32:48,530 --> 00:32:54,780 certainly something like Internet Explorer in the two years it went without competition. 273 00:32:54,780 --> 00:33:02,040 Developed almost none of the tab to browsers and other things that the other search engines did. 274 00:33:02,040 --> 00:33:10,590 And so that lost quality and variety is also very closely and intimately related to lost innovation. 275 00:33:10,590 --> 00:33:20,630 So I think we should care about the lack of competition. 276 00:33:20,630 --> 00:33:26,250 Then the last question is, in many ways, I think the toughest question. 277 00:33:26,250 --> 00:33:32,780 Which is, can there be competition in the digital sector? 278 00:33:32,780 --> 00:33:42,870 I gave you before a whole bunch of really good reasons to expect there to be tipping or a winner take most. 279 00:33:42,870 --> 00:33:51,240 If you take those to the extreme, they would say this is just like a you tilt last mile utility that's delivering water to your house. 280 00:33:51,240 --> 00:33:56,640 And if you had a single water company, it was charging people a lot of money for their water. 281 00:33:56,640 --> 00:34:00,950 The solution wouldn't be to try to get three more water companies and have them compete. 282 00:34:00,950 --> 00:34:11,910 It would be to regulate the price of that water company or if it was a telecom, make it carry other people's stuff in addition to its own. 283 00:34:11,910 --> 00:34:18,180 So this question is, is the digital sector a natural monopoly like water? 284 00:34:18,180 --> 00:34:28,260 Or is it a potentially competitive sector that just hasn't had competition because of inadequate policies? 285 00:34:28,260 --> 00:34:31,770 This is the question I'm probably the least sure of. 286 00:34:31,770 --> 00:34:40,680 But I sort of root for the answer to this to be yes, I think we should put in place policies that take us as far as possible, 287 00:34:40,680 --> 00:34:48,840 assuming the answer is yes, but policies that would also work decently well if the answer to this happened to be no. 288 00:34:48,840 --> 00:34:56,520 And if we discover more definitively that no, we may need to dial or pivot in some ways. 289 00:34:56,520 --> 00:35:02,760 And you know, this question to me comes back to the same point I made in general terms, 290 00:35:02,760 --> 00:35:10,950 not in the tech sector about sort of good and bad, natural and unnatural reasons for scale. 291 00:35:10,950 --> 00:35:19,050 It's a good thing to have a social network everyone's on. It's, you know, some of these companies are very efficient. 292 00:35:19,050 --> 00:35:22,700 No, there's an advantage to having brands, especially in something where, you know, 293 00:35:22,700 --> 00:35:29,390 you're not quite sure if you can trust the person, you can't look them in the eye. So those are all natural things. 294 00:35:29,390 --> 00:35:37,880 But then there's unnatural things to kill killer acquisitions, predatory behaviour by incumbents and then standards and data. 295 00:35:37,880 --> 00:35:48,720 And I really wanted to stress this last one. Now, imagine a world where a single company had invented email. 296 00:35:48,720 --> 00:35:56,180 And their email system could only communicate with other people that used their email system. 297 00:35:56,180 --> 00:36:02,480 They several companies entered at once for some reason, a lock or a little small advantage. 298 00:36:02,480 --> 00:36:06,650 One of those companies ended up with 95 percent of people on it. 299 00:36:06,650 --> 00:36:13,710 And the only way you could send an email to your friend is if they use that very same system. 300 00:36:13,710 --> 00:36:19,890 Somebody might come along and say, this is crazy. We should have lots of e-mail companies. 301 00:36:19,890 --> 00:36:25,800 People should have their own choice about what company, what client and we should have competition. 302 00:36:25,800 --> 00:36:32,160 Somebody else would say, No, you don't understand. There is economies of scale in email. 303 00:36:32,160 --> 00:36:39,180 Do you want to break this email company up? And no one could e-mail their friends again unless they happen to be on the same system 304 00:36:39,180 --> 00:36:43,230 or force people to have eight different emails so that they can reach their friends, 305 00:36:43,230 --> 00:36:52,830 depending on who's on which one. Yeah, we could have been in a world like that, but we weren't because email wasn't developed by a single company, 306 00:36:52,830 --> 00:37:00,000 it was developed in a public process with open standards that were interoperable. 307 00:37:00,000 --> 00:37:08,690 And as a result, no matter who your email provider is, no matter what email client you're using to download, compose, messages, 308 00:37:08,690 --> 00:37:17,520 store and the like, you can communicate with anyone else without even knowing what their system is or what client they're using. 309 00:37:17,520 --> 00:37:24,840 And so some of these economies of scale aren't natural, like the laws of physics that say, 310 00:37:24,840 --> 00:37:31,320 you know, is really costly to dig up hype to your house but reflect technical choices. 311 00:37:31,320 --> 00:37:38,820 And the technical choices are made by companies with a motivation that may not be aligned with 312 00:37:38,820 --> 00:37:48,420 what the social benefit is and may be aligned with protecting themselves at the expense of others. 313 00:37:48,420 --> 00:37:52,740 Data to insofar as it's a barrier to entry or a natural monopoly, 314 00:37:52,740 --> 00:37:58,410 I put it under the unnatural category standards, or Davies could have in some sense put them under either. 315 00:37:58,410 --> 00:38:02,310 But I chose to put them under unnatural because again, 316 00:38:02,310 --> 00:38:13,500 data has substantial spill-overs and it's sort of a choice about what our laws are about how we protect information, 317 00:38:13,500 --> 00:38:20,760 protect access to the exclusivity, who owns it and the like that create the system. 318 00:38:20,760 --> 00:38:26,520 We have nots, you know, natural thing that occurred. 319 00:38:26,520 --> 00:38:34,140 And so our goal is, like I said at the outset as a general matter for competition, but in this sector, 320 00:38:34,140 --> 00:38:42,160 it's to try to get rid of as many of these obstacles to competition as possible and then see if you get more actual entry. 321 00:38:42,160 --> 00:38:48,340 More threats of entry, more people afraid about entry. More companies that want to grow up. 322 00:38:48,340 --> 00:38:58,870 Not to be purchased by Google, but to be the next Google and see how far all of that gets. 323 00:38:58,870 --> 00:39:03,670 So to summarise this part of the discussion, you know, these are the answer. 324 00:39:03,670 --> 00:39:13,970 The questions is competition beneficial? I think for the most part, yes, at least for the purposes that I was most focussed on, is it absent? 325 00:39:13,970 --> 00:39:19,580 And if it is absent, is that absence costly? Tried to argue yes. 326 00:39:19,580 --> 00:39:23,100 If your answer to those questions is no, then you don't really need a policy. 327 00:39:23,100 --> 00:39:28,730 You know, the companies are already competing with each other, so we don't need to do anything or they're not really competing. 328 00:39:28,730 --> 00:39:34,070 But that's because everything's free and great, so you don't need any policy, either. 329 00:39:34,070 --> 00:39:41,870 I don't think the no policy change is crazy. I think there's a lot of versions of policy change that would be worse than no policy change. 330 00:39:41,870 --> 00:39:49,100 So you do have to be very careful here. So I think there's some merit in every answer to every one of these questions. 331 00:39:49,100 --> 00:39:53,430 But I certainly think we can do a whole lot better than what we have now. 332 00:39:53,430 --> 00:39:56,180 The last is, is competition policy effective? 333 00:39:56,180 --> 00:40:02,480 And if you don't think it is, you need something more like utility regulation around the price, common carrier and the like. 334 00:40:02,480 --> 00:40:11,380 If you think it can be effective, then you can have pro competition policy. 335 00:40:11,380 --> 00:40:16,970 So I think, yes, I think different branches on this tree. 336 00:40:16,970 --> 00:40:24,260 Have one way of categorising why different people come to different views on policy in this area. 337 00:40:24,260 --> 00:40:28,550 So where does this leave us in terms of recommendations? 338 00:40:28,550 --> 00:40:36,140 I'm going to go through the recommendations now. These were developed for the UK government in most cases. 339 00:40:36,140 --> 00:40:48,080 These are ones that I think would apply to for any media, more large sized economy in the world and certainly would apply for and be, 340 00:40:48,080 --> 00:40:55,430 I think, a great idea in the context of the United States. 341 00:40:55,430 --> 00:41:06,290 I'm going to talk about ex-ante regulatory proposals first, because, you know, we discussed how far you can get with the traditional competition, 342 00:41:06,290 --> 00:41:15,710 merger, antitrust types of rules and felt you couldn't get nearly far enough even if you stopped all mergers. 343 00:41:15,710 --> 00:41:23,610 I think it would be a big mistake, but even if you did, the companies are already quite a large scale, even if you brought fine after fine. 344 00:41:23,610 --> 00:41:35,230 You know, it's taken the commission a decade to deal with, you know, much of three cases and hasn't changed a whole lot. 345 00:41:35,230 --> 00:41:40,420 There's also something when you have this ex post, 346 00:41:40,420 --> 00:41:48,340 especially an antitrust approach on the something costly to it, it takes a long time can be very uncertain. 347 00:41:48,340 --> 00:41:53,440 You can be fined for behaviour. You weren't quite sure whether you were allowed to do or not allowed to do so. 348 00:41:53,440 --> 00:41:59,020 The fine isn't a complete deterrence to that behaviour in the future. 349 00:41:59,020 --> 00:42:03,550 It doesn't set up a clear precedent for behaviour in the future, 350 00:42:03,550 --> 00:42:09,910 and it's not flexible in a rapidly changing landscape because you might have a behavioural remedy, 351 00:42:09,910 --> 00:42:16,630 but no mechanism to change and update that behavioural remedy as the economy changes. 352 00:42:16,630 --> 00:42:20,110 It's not to say antitrust doesn't play a role and isn't important, 353 00:42:20,110 --> 00:42:30,430 but we felt that a lot of where the enforcement needed to happen was set up clear rules in advance. 354 00:42:30,430 --> 00:42:39,760 Be somewhat more detailed about how they'd function, make those rules focussed on easing entry, 355 00:42:39,760 --> 00:42:47,920 and you know, in some sense, the debate between regulate or not regulate is almost a misnomer. 356 00:42:47,920 --> 00:42:53,440 You're definitely going to enforce rules. The question is, you know, when are you enforcing them? 357 00:42:53,440 --> 00:42:55,930 How are you writing those rules down? 358 00:42:55,930 --> 00:43:02,410 So we proposed something called a digital markets unit, and we were agnostic as to whether this would be a new body, 359 00:43:02,410 --> 00:43:15,280 a subsidiary of an existing body or just an office that existed in the CMA, Ofcom or somewhere else in the United States. 360 00:43:15,280 --> 00:43:21,130 You could debate whether this is a new regulator, as recently proposed by Fiona Scott Morton, 361 00:43:21,130 --> 00:43:26,110 or whether this sits in the Federal Trade Commission and the like. 362 00:43:26,110 --> 00:43:34,300 The important issue, or at least the issue that we felt we had something to add on wasn't, you know, where this thing sits. 363 00:43:34,300 --> 00:43:45,430 But what it does and what it does is something you could call participatory regulation with three main functions. 364 00:43:45,430 --> 00:43:54,220 First is a code of conduct. The second is data mobility and open standards, and the third is data openness. 365 00:43:54,220 --> 00:44:00,220 The code of conduct would only apply to companies with strategic market status. 366 00:44:00,220 --> 00:44:07,780 There would be a process, something like what Ofcom has already to designate companies with significant market power or what 367 00:44:07,780 --> 00:44:14,540 the United States F SoC has to discover designate systemically important financial institutions. 368 00:44:14,540 --> 00:44:22,060 Syfy's The designation would last three to five years and you could renew it and re designate the companies, 369 00:44:22,060 --> 00:44:27,040 those companies or companies that are gateways sometimes called bottlenecks that you basically 370 00:44:27,040 --> 00:44:34,540 need to get through them in order to access everything else in the online platform. 371 00:44:34,540 --> 00:44:42,460 And those companies would have to subscribe to a code of conduct that code of 372 00:44:42,460 --> 00:44:47,590 conduct for the business side of platforms would be based on a set of principles. 373 00:44:47,590 --> 00:44:53,700 It would be fleshed out through the participatory multi-stakeholder process and 374 00:44:53,700 --> 00:44:57,460 would be things like providing access to the designated platforms on a fair, 375 00:44:57,460 --> 00:45:04,420 consistent and transparent basis. Things like prominence ranking in reviews are done the same way. 376 00:45:04,420 --> 00:45:10,380 You couldn't be restricted from or penalised for using alternative platforms. 377 00:45:10,380 --> 00:45:18,420 Just to be clear, if I want to start a small business tomorrow and not let anyone else on my platform or say, 378 00:45:18,420 --> 00:45:26,010 if you want to be on my platform, you can't be on anyone else's. Or have my search engine favour my results over anyone else's. 379 00:45:26,010 --> 00:45:32,460 That's fine. Small businesses, even medium businesses, wouldn't fall under this code of conduct. 380 00:45:32,460 --> 00:45:39,590 This is just for companies that are acting as gateways. 381 00:45:39,590 --> 00:45:46,040 The that would be the first part of what the digital markets unit would do. 382 00:45:46,040 --> 00:45:52,340 This isn't wildly different from what's lying in the back of antitrust law. 383 00:45:52,340 --> 00:46:03,890 It's similar in spirit. The difference is the ongoing monitoring flexibility in its application. 384 00:46:03,890 --> 00:46:12,500 And you know, this isn't self-executing. I think if you have a competent regulator that does a good job of this, it would be a great thing. 385 00:46:12,500 --> 00:46:16,490 If you had an incompetent regulator that did a bad job of this, I think it would be a problem. 386 00:46:16,490 --> 00:46:27,550 And you need to figure out, you know, your judgement about how to weigh the costs of the existing system versus some of the risks. 387 00:46:27,550 --> 00:46:32,780 But on balance, I think we're in some sense already doing this through antitrust. 388 00:46:32,780 --> 00:46:39,180 We're just doing it in a clunkier less predictable manner. 389 00:46:39,180 --> 00:46:47,130 The second part of the digital markets unit is the data mobility and open standards. 390 00:46:47,130 --> 00:46:51,270 We chose the word data mobility because it's not just portability, 391 00:46:51,270 --> 00:46:57,780 portability and the GDPR sense might just be that you can download the data and bring it with you. 392 00:46:57,780 --> 00:47:04,620 Data mobility says that you can import your data from system to system. 393 00:47:04,620 --> 00:47:16,240 Open standards are the type I talked about with email and why we have lots of email companies right now and with data mobility and open standards. 394 00:47:16,240 --> 00:47:25,540 You know, what's clear is this isn't just a law you can pass that says you need data mobility and open standards. 395 00:47:25,540 --> 00:47:33,000 You know, now you go ahead and do it. This is really important and complicated technical issues and how to do this. 396 00:47:33,000 --> 00:47:39,660 There's really important privacy issues and other sets of concerns. 397 00:47:39,660 --> 00:47:46,860 So this is certainly not self-executing. There are private efforts underway right now, 398 00:47:46,860 --> 00:47:55,770 like the digital transfer project that the major companies are a part of to share photos and the like that are doing some of this data, 399 00:47:55,770 --> 00:47:58,140 mobility and open standards. 400 00:47:58,140 --> 00:48:08,970 The problem is that those companies don't have the same incentives that we do as a society to do that to the fullest and make all of that succeed. 401 00:48:08,970 --> 00:48:16,910 And when you had open banking here in the United Kingdom, which was a requirement that nine eight. 402 00:48:16,910 --> 00:48:19,560 Eight, eight, eight, 403 00:48:19,560 --> 00:48:32,160 eight large banks would have to essentially open up their APIs and allow a whitelisted set of fintechs to build applications on top of it, you know, 404 00:48:32,160 --> 00:48:41,280 absent being somewhat prescriptive and having a well-funded implementation entity that goes through the use cases, 405 00:48:41,280 --> 00:48:44,580 figures out the number of screens, figures that how you can do it. 406 00:48:44,580 --> 00:48:53,120 You can end up having data mobility and open standards in theory, but in practise it's completely unusable by consumers. 407 00:48:53,120 --> 00:49:02,610 So this is something where you have to go through, you know, messaging, social networking, trust systems and the like and ask, Can it work? 408 00:49:02,610 --> 00:49:09,080 Can it not work? Do we have the funding? Do we have the people to figure it out? 409 00:49:09,080 --> 00:49:12,980 The third function of the digital market unit is data openness. 410 00:49:12,980 --> 00:49:19,760 I talked a little bit about the sense in which in some cases there's increasing returns to scale to data 411 00:49:19,760 --> 00:49:30,680 in the economy that access to data can effectively exclude and act as a barrier to entry to others. 412 00:49:30,680 --> 00:49:35,950 This is again not one where you can pass a law saying, you know. 413 00:49:35,950 --> 00:49:45,730 Facebook, go ahead. Open your servers up, let everyone look at everything. That would be a terrible thing to do, but can you make data trusts? 414 00:49:45,730 --> 00:49:50,870 Sydenham ized data some types of data. 415 00:49:50,870 --> 00:49:55,540 No geospatial mapping data used for autonomous vehicles. 416 00:49:55,540 --> 00:50:02,260 You're not going to make the competitive advantage of a company that they know there's an obstacle that's going to kill people. 417 00:50:02,260 --> 00:50:07,960 And another company doesn't. Again, you need a you can't just pass a law. 418 00:50:07,960 --> 00:50:17,860 You need a unit to do this. You need a unit with expertise and one that's going to go through the different use cases. 419 00:50:17,860 --> 00:50:23,560 So the digital markets unit is the most important proposal that we made. 420 00:50:23,560 --> 00:50:36,370 It's one that there is authority in the UK already to do a certain amount of all of this, but certainly new authority would be needed to do all of it. 421 00:50:36,370 --> 00:50:43,030 In the United States, there's a debate about whether the FTC has under Section five its own authority to do all sorts of things, 422 00:50:43,030 --> 00:50:51,140 including these types of activities or whether new authority would be needed. 423 00:50:51,140 --> 00:50:56,510 Would not confine, you know, this is a little bit of a belt and suspenders type of approach. 424 00:50:56,510 --> 00:51:04,890 And the traditional competition policy has an important role to play too with, 425 00:51:04,890 --> 00:51:09,560 you know, which depends a little bit on your perspective on competition policy. 426 00:51:09,560 --> 00:51:15,320 I won't spend a lot of time on this, but we can do it in the Q&A. But there's a traditional big is bad approach. 427 00:51:15,320 --> 00:51:23,000 There's Robert Bork big as efficient. These are both grossly caricaturing those two approaches. 428 00:51:23,000 --> 00:51:28,550 We situated ourselves with Robert Bork in terms of believing in economic analysis, 429 00:51:28,550 --> 00:51:35,360 believing you need to sort of work through the efficiency questions and not just assume big is bad, 430 00:51:35,360 --> 00:51:42,260 but then that with modern economics, when you do that, it turns out that. 431 00:51:42,260 --> 00:51:50,450 You know, there's a probably greater range of things to be concerned about using this traditional economic tools than the Chicago school was, 432 00:51:50,450 --> 00:52:06,110 a lot of people are in this approach to competition policy but haven't coalesced around a name so Chicago, plus the Taft School or just being correct. 433 00:52:06,110 --> 00:52:10,850 So I won't spend a lot of time on these recommendations, but we can talk more about them. 434 00:52:10,850 --> 00:52:19,100 One is that merger assessment needs a reset. I spelled this so you all could understand it, prioritising scrutiny of digital murders, 435 00:52:19,100 --> 00:52:24,860 updating the merger assessment guidelines, including emphasising potential competition innovation. 436 00:52:24,860 --> 00:52:30,260 We have an important idea on instead of saying Is there a fifty one percent chance of mergers a problem? 437 00:52:30,260 --> 00:52:33,020 Asking How bad a problem is it if it happens? 438 00:52:33,020 --> 00:52:41,120 You know, if you're going to buy something and you know what have been your competitor and now it can't be anymore, you might have a really big harm. 439 00:52:41,120 --> 00:52:47,420 Even with a smaller chance of that happening, we should be worried about it. 440 00:52:47,420 --> 00:52:55,160 We considered and decided not to have a sort of blanket ban or presumption against the major platforms acquiring companies. 441 00:52:55,160 --> 00:53:02,390 That idea might make more sense the United States than it would make in the UK, and it's worth thinking about more there. 442 00:53:02,390 --> 00:53:08,960 But all of this is designed to get at the fact that we basically haven't challenged any platform acquisitions. 443 00:53:08,960 --> 00:53:14,870 So we've never really made a false positive where we accidentally blocked a merger we shouldn't have. 444 00:53:14,870 --> 00:53:22,970 We probably have had a number of false negatives where we didn't block a merger and maybe we should have antitrust. 445 00:53:22,970 --> 00:53:30,830 These are very no offence to lawyers in the Weeds UK specific things about the 446 00:53:30,830 --> 00:53:37,880 Competition and Markets Authority getting more independent decision making. The Competition Appeals Tribunal reviews, you know, 447 00:53:37,880 --> 00:53:44,660 more limited in certain types of ways you to sort of take a thumb on the scale that's currently 448 00:53:44,660 --> 00:53:50,840 sort of adds extra hoops to jump through to do antitrust and try to streamline that. 449 00:53:50,840 --> 00:53:53,150 And then there's a number of other recommendations, 450 00:53:53,150 --> 00:53:59,960 like a market study of online advertising and monitoring, algorithmic collusion and personalised pricing. 451 00:53:59,960 --> 00:54:08,780 This is where algorithms collude with each other to set higher prices something again we can talk about. 452 00:54:08,780 --> 00:54:19,760 So in terms of next steps, as Ian said, this is not the only issue being debated in the United Kingdom right now as I understand it, 453 00:54:19,760 --> 00:54:28,430 but there is quite active effort on the part, especially of civil servants in the government that are reviewing this set of proposals. 454 00:54:28,430 --> 00:54:36,080 Another one from DCMS on online harms and the third one from Frances Cairncross on the media, 455 00:54:36,080 --> 00:54:41,240 and I'm hopeful that they'll come out with some ideas and recommendations. 456 00:54:41,240 --> 00:54:47,510 CMA has been publicly positive about most of what I just said. 457 00:54:47,510 --> 00:54:56,630 I know the chancellor and others have been very positive as well, and there's also a growing global conversation at the G7. 458 00:54:56,630 --> 00:55:04,250 The Stigler centred Chicago recently came out with a report that looks an awful lot like ours, except they put their antitrust chapter. 459 00:55:04,250 --> 00:55:09,890 Before the regulatory chapter. We did the regulatory chapter. Before the antitrust one. 460 00:55:09,890 --> 00:55:15,200 The European Commission had special advisers. The Australians had to report. 461 00:55:15,200 --> 00:55:20,840 The OECD has done things. The FTC in the United States has been holding hearings. 462 00:55:20,840 --> 00:55:25,160 So there's a big global conversation about this as well. 463 00:55:25,160 --> 00:55:26,660 And you know, well, 464 00:55:26,660 --> 00:55:37,460 I don't think we're ever going to have coordinated global action to have similar actions happening at similar times in similar countries. 465 00:55:37,460 --> 00:55:47,360 I think would be certainly the most efficient and effective way to address what is genuinely a global economic issue. 466 00:55:47,360 --> 00:55:56,020 Thank you. 467 00:55:56,020 --> 00:56:10,870 Thanks very much, Jason, for that remarkably broad, clear exposition, not only of the issues you focussed on in the review, but the wider context. 468 00:56:10,870 --> 00:56:17,080 Your capability of sort of looking at the very big picture and getting down into granularity, 469 00:56:17,080 --> 00:56:22,360 not least from a policy or investment perspective, is quite remarkable. 470 00:56:22,360 --> 00:56:28,510 And it's good to know that not only was the UK government very wise and timely in appointing you, 471 00:56:28,510 --> 00:56:35,890 but that this is part of a force that is bigger than the current government and something which is obviously of extreme significance. 472 00:56:35,890 --> 00:56:40,780 So we have about half an hour for discussion questions and answers. 473 00:56:40,780 --> 00:56:41,950 It is being webcast, 474 00:56:41,950 --> 00:56:51,920 so just be aware of that when you pose your questions and we have clearer with who's the organiser of all these events with a roving mic. 475 00:56:51,920 --> 00:57:00,310 I'd like to go first. And people will identify themselves, right? 476 00:57:00,310 --> 00:57:04,650 Yes. Hello. 477 00:57:04,650 --> 00:57:10,020 Thank you for your talk. My name is Ariana. I study cyber security. I specifically consent and privacy. 478 00:57:10,020 --> 00:57:16,110 So we mentioned that people are paying, consumers are paying in terms of data and privacy. 479 00:57:16,110 --> 00:57:22,950 Would you mind just expanding a little bit on what you mean or what you prioritise and what you focus on when you talk about that, 480 00:57:22,950 --> 00:57:26,280 how consumers are paying in terms of data and privacy? 481 00:57:26,280 --> 00:57:36,810 Yeah, I mean, I stopped at a cafe this morning and tried to get online and asked them what the Wi-Fi was, 482 00:57:36,810 --> 00:57:43,440 and they told me it was O2 and that took me to some screen where I had to enter my mobile number. 483 00:57:43,440 --> 00:57:48,690 And then it sent me a text to, you know, you've all probably done this a lot. 484 00:57:48,690 --> 00:57:52,110 And then it took me to the screen and there was a box. Can we share your information? 485 00:57:52,110 --> 00:57:59,190 And it was with that cafe in particular that you just gave, and I didn't check the box and wouldn't give me the Wi-Fi. 486 00:57:59,190 --> 00:58:06,510 No, that's not unreasonable. I mean, there's no reason, you know, the cafe wouldn't give me, you know, a coffee without me paying either. 487 00:58:06,510 --> 00:58:17,670 So there are free business and they can offer something in exchange for something else, and you have a choice of whether or not to do that. 488 00:58:17,670 --> 00:58:25,320 I just think you then can't call it free. And so that's so I'm not objecting to charging in various forms. 489 00:58:25,320 --> 00:58:32,790 You know, I find the advertisements in magazines and newspapers pretty annoying, and I wish I had a thinner magazine with no advertisements. 490 00:58:32,790 --> 00:58:36,690 But, you know, I don't want to pay four times the price for that. 491 00:58:36,690 --> 00:58:46,590 So we have to pay for things somehow, and we don't love any of the ways of paying for things, including parting with cash. 492 00:58:46,590 --> 00:58:50,520 So my two main point, my main point there was just that. 493 00:58:50,520 --> 00:58:58,080 You know, the defence that these are free so you shouldn't be worried isn't a very good defence when you're paying in some manner. 494 00:58:58,080 --> 00:59:02,400 And then, yes, I think you worry. Maybe people don't quite know what they pay. 495 00:59:02,400 --> 00:59:07,290 If you pay one pound, you know you paid a pound if you check the box on a form. 496 00:59:07,290 --> 00:59:11,730 What did I just give them? How much are they going to take from me? 497 00:59:11,730 --> 00:59:17,440 You might need some additional consumer protections of a type you don't really need around the price of coffee. 498 00:59:17,440 --> 00:59:19,830 Someone wants to charge you 20 pounds for coffee, 499 00:59:19,830 --> 00:59:26,790 and they're probably not going to sell a lot if they want to charge like massive data violation in exchange for something, 500 00:59:26,790 --> 00:59:30,510 you know, maybe you didn't realise you paid massive instead of just a little. 501 00:59:30,510 --> 00:59:38,560 And so some consumer protection may be needed, you know, because you don't quite know the price you're paying. 502 00:59:38,560 --> 00:59:42,910 Is anyone on the that terrorism doesn't all the other way? 503 00:59:42,910 --> 00:59:47,750 Thank you, I thank you. I'm Silas. I'm at the Oxford Internet Institute. 504 00:59:47,750 --> 00:59:53,620 And you mentioned that policy that fosters competition might also require more regulation, 505 00:59:53,620 --> 00:59:58,420 and the costs of that regulation would go back to the consumer, presumably through taxation. 506 00:59:58,420 --> 01:00:07,310 So do the do the sort of benefits of the cost benefits in terms of increased competition versus the costs of higher taxes. 507 01:00:07,310 --> 01:00:12,200 Do they count each other or is there an optimum point that's reached right? 508 01:00:12,200 --> 01:00:21,950 So there's two types of costs. One is like literally the cost of paying the people that work in the regulatory unit. 509 01:00:21,950 --> 01:00:32,400 I think that would be relatively trivial compared to the size of these markets and so that if it improved consumer welfare buying. 510 01:00:32,400 --> 01:00:42,690 You know, 0.01 percent that would weigh more than offset the cost of those people, so that's not a trivial or an irrelevant thing. 511 01:00:42,690 --> 01:00:47,400 But it's not what I'm most worried about. We didn't make this recommendation in the report, 512 01:00:47,400 --> 01:00:52,470 but certainly one model you could have is just some form of fee on the activity 513 01:00:52,470 --> 01:00:56,790 of the type that we have in a number of other sectors to fund the regulator. 514 01:00:56,790 --> 01:01:03,990 So there's that cost, which from your question sounded like that's when you're talking about the second cost of if the regulation becomes a 515 01:01:03,990 --> 01:01:12,660 barrier to entry so someone else can enter or constricts your social network to adopt a certain form that's interoperable. 516 01:01:12,660 --> 01:01:19,050 And so you don't get to invent a new type of social network with some new type of features. 517 01:01:19,050 --> 01:01:29,520 And, you know, I think that's a risk, but I also think it's a benefit if you are creating something so a new social network enters. 518 01:01:29,520 --> 01:01:34,140 And if I thought Facebook was innovating at a really rapid pace and its feed, 519 01:01:34,140 --> 01:01:39,540 I'd be more worried about getting in the way of it when that seems like it's not innovating that much. 520 01:01:39,540 --> 01:01:41,850 And maybe we could have other companies come in. 521 01:01:41,850 --> 01:01:50,280 And if we do, that would either force Facebook to innovate more or lead the other ones to another case where you have that is, you know, 522 01:01:50,280 --> 01:01:58,920 venture capital like venture capitalist exit strategies selling to a big company because that can increase the demand for innovation. 523 01:01:58,920 --> 01:02:01,950 But it can also increase the demand for certain types of innovation that don't 524 01:02:01,950 --> 01:02:06,480 threaten certain business models or that complement them in certain ways. 525 01:02:06,480 --> 01:02:12,420 And so maybe you get less of a demand for certain type of innovation if you make mergers a little bit harder, 526 01:02:12,420 --> 01:02:22,020 but more of a demand for, you know, another type of innovation. If you make killer acquisitions harder, so I think these things are balanced. 527 01:02:22,020 --> 01:02:28,260 But I wouldn't worry too much about the group of people doing this. 528 01:02:28,260 --> 01:02:32,760 Hi, my name is Danny. I'm a doctoral student in the politics department. 529 01:02:32,760 --> 01:02:36,750 And I also recently enjoyed your recent interview with Margaret best there. 530 01:02:36,750 --> 01:02:42,180 I thought that was really interesting. Margaret Forster, you had an event that I was really good. 531 01:02:42,180 --> 01:02:52,160 I have a question about what you said last about your. Prospects of coordinated global regulation or coordinated global action, 532 01:02:52,160 --> 01:02:58,490 do you anticipate that we might look at a kind of a soft law approach with a like a Basel type standards? 533 01:02:58,490 --> 01:03:04,670 Were there a set of voluntary standards, but there is no global regulator for something like this? 534 01:03:04,670 --> 01:03:08,640 And if so, what? What are the standards that you see developing? 535 01:03:08,640 --> 01:03:15,320 Obviously, you have your own report here, but are there common links in the in the kind of programmes that you put up in there last night? 536 01:03:15,320 --> 01:03:22,490 Thanks very much. I don't know the answer to your question with any remote degree of confidence, 537 01:03:22,490 --> 01:03:29,360 but one the really big difference between the sector and the banking sector is that every country has banks. 538 01:03:29,360 --> 01:03:35,750 And the major digital platforms are all in one country and that everyone is going 539 01:03:35,750 --> 01:03:40,370 to give up a little bit in terms of their ability to set rules for their own banks 540 01:03:40,370 --> 01:03:45,230 in exchange for everyone else having rules on their bank so that their banks can 541 01:03:45,230 --> 01:03:51,380 blow up the global system is a little bit of the bargain that's built into Basel. 542 01:03:51,380 --> 01:03:59,450 And I don't see how at the current moment you could reproduce a bargain like that here. 543 01:03:59,450 --> 01:04:06,890 I think there's some places where, you know, there's common interest, and I found this in the U.S. government when I was in the government like, 544 01:04:06,890 --> 01:04:14,300 let's say, on taxes, we could put in our budget a proposal that was quite tough on multinational overseas tax avoidance. 545 01:04:14,300 --> 01:04:21,770 If our budget proposal had been passed, the result would have been that that multinational would have paid more taxes to the United States. 546 01:04:21,770 --> 01:04:29,630 But it also wouldn't have been able to hide its UK subsidiary income in the Cayman Islands, so it would have paid more taxes to the UK. 547 01:04:29,630 --> 01:04:31,310 We put that in our budget. 548 01:04:31,310 --> 01:04:39,890 But then, you know, the company would come to the Treasury on a particular issue that happened in a particular country and say, 549 01:04:39,890 --> 01:04:46,880 like, you know, look like the UK is doing something and they're taking more taxes from us than we're an American company. 550 01:04:46,880 --> 01:04:55,190 Why would you allow that? And we tend to side with that company. So we'd have a sort of social welfare improving policy in one space and then 551 01:04:55,190 --> 01:05:00,680 a bit of a rent seeking captured by American company policy and other space. 552 01:05:00,680 --> 01:05:04,650 And those two would co-exist. I worry this could be a similar space. 553 01:05:04,650 --> 01:05:10,040 You know, I think open standards and data mobility. 554 01:05:10,040 --> 01:05:15,110 No, that's not like a way to hurt the United States and help everyone else. 555 01:05:15,110 --> 01:05:19,340 That's a way to help people in the United States have more choices, 556 01:05:19,340 --> 01:05:29,130 help new American companies enter the market and succeed, and also help new UK companies, French companies and the like. 557 01:05:29,130 --> 01:05:33,350 So I don't think in reality, this is a zero sum thing. 558 01:05:33,350 --> 01:05:41,180 In fact, it's much less zero-sum than taxes, where like, sometimes it is the case that this country collects it, that country doesn't. 559 01:05:41,180 --> 01:05:53,210 So I think it's really positive. Some I just don't trust that the political system would always see it as such. 560 01:05:53,210 --> 01:05:58,040 OK, thank you. My name is Greg Taylor, I'm an economist at the Oxford Internet Institute. 561 01:05:58,040 --> 01:06:04,370 So you mentioned the example of mandating the sharing of geospatial data for autonomous vehicles, 562 01:06:04,370 --> 01:06:07,810 and that has very obvious static efficiency benefits. 563 01:06:07,810 --> 01:06:14,240 And but one might worry that if we do that, then the incentive to generate those sorts of data sets in the first place would disappear. 564 01:06:14,240 --> 01:06:19,630 And I wondered if the panel spent much time thinking about those sorts of dynamic inefficiencies. 565 01:06:19,630 --> 01:06:27,890 Efficiencies could be avoided, right? Yeah, I, I think spend time and sitting on it. 566 01:06:27,890 --> 01:06:33,980 I don't think spend time Philip can correct me coming up with a great answer. 567 01:06:33,980 --> 01:06:37,280 In some cases, where you have those types of external benefits. 568 01:06:37,280 --> 01:06:47,360 Yes, you do, you know, assign property rights and give somebody a greater return to them or you have more public investment in the activity. 569 01:06:47,360 --> 01:06:57,410 I think there are, you know, and and I find, you know, part of why our answer was what it was, wasn't that, 570 01:06:57,410 --> 01:07:04,660 like all the merit is on one side and all the merits on the break the companies offer you do nothing that a lot of these are sort of good things. 571 01:07:04,660 --> 01:07:12,020 And it's bad that, you know, like the acquisitions, like some of that's a really good, efficient thing to take a company and acquire them. 572 01:07:12,020 --> 01:07:17,210 Some of it's an inefficient thing because you're foreclosing. And you know, 573 01:07:17,210 --> 01:07:22,880 that means a certain amount of judgement and all of this and asking questions like 574 01:07:22,880 --> 01:07:27,680 you did and coming up with better answers than the one I just didn't give you. 575 01:07:27,680 --> 01:07:32,370 Do you want had anything? It's kind of high. 576 01:07:32,370 --> 01:07:39,580 I'm and tired, I'm a researcher at the Martin School, so my question was on your last point on personalised pricing. 577 01:07:39,580 --> 01:07:44,860 And I wanted to know what your stance on regulating that is because on one hand, you know, 578 01:07:44,860 --> 01:07:51,670 producers benefit from price discrimination, but on the other hand, consumers lose out on their consumer surplus. 579 01:07:51,670 --> 01:07:56,590 So what's the right way to actually regulate personalised pricing? 580 01:07:56,590 --> 01:08:05,080 I should confess my views on personalised pricing are sort of more traditional. 581 01:08:05,080 --> 01:08:11,890 Economist views of this is efficient because it allows the quantity to be greater. 582 01:08:11,890 --> 01:08:16,060 You know, if some people are willing to pay a little bit for something, some people are willing to pay a lot. 583 01:08:16,060 --> 01:08:21,460 People are willing to pay a little bit, you know, get a low price, they buy it. 584 01:08:21,460 --> 01:08:26,170 You only will pay a lot. They get high price, they buy it so you increase quantity. 585 01:08:26,170 --> 01:08:34,660 And that's what we're ultimately concerned with. And I think it often has features of equity and that the people are willing to pay more often 586 01:08:34,660 --> 01:08:40,660 our higher income people that Uber might charge a higher price to somebody who gets picked 587 01:08:40,660 --> 01:08:45,160 up right in front of Goldman Sachs rather than a half a block away because it's doing a 588 01:08:45,160 --> 01:08:50,470 personalised pricing and figuring out who the person is is of all the injustices in the world, 589 01:08:50,470 --> 01:08:54,700 probably not like the one that bothers people the most. 590 01:08:54,700 --> 01:08:58,210 So I confess I'm a little bit less bothered. 591 01:08:58,210 --> 01:09:04,720 That's why a lot of our recommendations are in this sort of monitoring and get to is the personalised pricing. 592 01:09:04,720 --> 01:09:11,830 Sort of abusive or unfair or against protected groups and the like, you know, 593 01:09:11,830 --> 01:09:18,010 sort of who is it directed at and how not, you know, the existence of it and let's say is, 594 01:09:18,010 --> 01:09:25,390 I think the puzzle is that there's so little of this right now in Amazon for does a lot of different prices at different times of day. 595 01:09:25,390 --> 01:09:29,350 It does not charge different people at the same moment, different prices. 596 01:09:29,350 --> 01:09:36,400 And you know, there's a, I think, a huge social obstacle to there doing that. 597 01:09:36,400 --> 01:09:40,390 But one day they'll overcome that and it'll be just like an aeroplane where the person sitting 598 01:09:40,390 --> 01:09:46,610 next to you had a completely different price for the same product that you just bought. 599 01:09:46,610 --> 01:09:52,640 If Johnny may not be gay partners, I'm looking at the US over here, 10 year time span. 600 01:09:52,640 --> 01:10:00,800 Which of your recommendations do you think are most likely to be implemented? All of them. 601 01:10:00,800 --> 01:10:03,170 I think we're getting the picture right. Right, right, yeah. 602 01:10:03,170 --> 01:10:10,130 So first of all, the FCC, the FTC, which is one of our two, we have two competition enforcers in the United States, 603 01:10:10,130 --> 01:10:18,650 the FTC, the Department of Justice and the FTC has had a set of hearings on the questions we've been talking about. 604 01:10:18,650 --> 01:10:24,510 The Department of Justice invited me in to meet with them and talk to them about this report next week. 605 01:10:24,510 --> 01:10:29,750 So even within this government there, the current government, there is interest. 606 01:10:29,750 --> 01:10:33,890 There's definitely a lot more of interest in these questions on the Democratic side, 607 01:10:33,890 --> 01:10:42,620 and I've gotten more questions and engagement on that side, I think on regulation. 608 01:10:42,620 --> 01:10:50,190 Mark Zuckerberg's out there calling for Facebook to be regulated less about the competition side than a set of behaviours. 609 01:10:50,190 --> 01:10:54,320 So I think on that in a decade, there's going to be more regulation. 610 01:10:54,320 --> 01:10:59,000 It's just a question of what form does it take and as it sort of. 611 01:10:59,000 --> 01:11:05,900 Well, designed and conceived, or is it something other than what we proposed and on mergers, 612 01:11:05,900 --> 01:11:14,130 I think the dial is going to turn in the United States, not just in the tech sector, but more generally. 613 01:11:14,130 --> 01:11:21,360 Making it more scrutiny of mergers. The issue there is that absent a legal change, it's the same judges. 614 01:11:21,360 --> 01:11:30,510 And so we've had hospital mergers that very conservative Chicago School FTC Commissioner Josh Wright have voted against, 615 01:11:30,510 --> 01:11:36,810 allowing the hospitals to merge and then end up losing in court, where the court says, oh, like the executive said, it was a good idea. 616 01:11:36,810 --> 01:11:43,960 So we're going to overturn it and allow it. So I think mergers will get tougher, but there's a constraint on that. 617 01:11:43,960 --> 01:11:50,540 And I don't predict any real antitrust action, like breaking companies off. 618 01:11:50,540 --> 01:11:51,620 If we've got the money. 619 01:11:51,620 --> 01:12:01,190 Hello, I'm student fund geography, and because I heard the news yesterday that Google would not provide any services to Chinese company Huawei. 620 01:12:01,190 --> 01:12:08,330 So in that sense, the digital competition is kind of a competition between countries rather than between companies. 621 01:12:08,330 --> 01:12:15,170 So in this context, do you still think the regulation is to be put on two countries? 622 01:12:15,170 --> 01:12:20,920 And if so, how can you persuade the two countries to obey the regulations that you recommend? 623 01:12:20,920 --> 01:12:35,730 Right? I think the US decision on Huawei was genuinely motivated by a set of national security concerns and no. 624 01:12:35,730 --> 01:12:44,430 I don't have current up to date information on those national security concerns, but I think there's some merit to them, 625 01:12:44,430 --> 01:12:50,520 so I wouldn't look at that through the prism of the competition policy type of issues. 626 01:12:50,520 --> 01:12:58,020 I was talking about, I think if the United States had complete confidence in, you know, 627 01:12:58,020 --> 01:13:04,860 why way not engaging in espionage and the like, then Google would be free to sell it. 628 01:13:04,860 --> 01:13:09,270 So I don't think that was. I think Google's business decision would have been to continue selling to walk away. 629 01:13:09,270 --> 01:13:16,650 That's what Google would have liked to have done. I think if it was just up to US economic policy, that would have happened. 630 01:13:16,650 --> 01:13:29,090 I think there a national security issue that, you know, rightly or wrongly, but certainly not utterly spuriously, was what that was based on. 631 01:13:29,090 --> 01:13:34,190 I'm John Hofmeyr from Oxford's Centre for Mutual, an employee owned business. 632 01:13:34,190 --> 01:13:39,660 I have a question about your slide that had to do with obstacles to competition. 633 01:13:39,660 --> 01:13:49,700 And this was under the unnatural category and I was hoping that you might feel comfortable just giving an example or two of predatory behaviour. 634 01:13:49,700 --> 01:13:59,450 Yeah, this is if you give priority to, you know, your own products when someone's searching. 635 01:13:59,450 --> 01:14:03,200 And again, if you're one of a million companies, that's fine. 636 01:14:03,200 --> 01:14:13,340 But if you're a major one, it is, you know, in your search results putting, you know, your sales first instead of others, 637 01:14:13,340 --> 01:14:21,050 making a condition that if people want to sell with you, they can't sell somewhere else. 638 01:14:21,050 --> 01:14:29,210 You know, all of these types of things, and we have a high degree of tolerance for in the economy as a whole, probably rightly so. 639 01:14:29,210 --> 01:14:37,670 In a higher degree of tolerance now than, you know, the doctrines of antitrust would have said we should have had 30 years ago, 640 01:14:37,670 --> 01:14:44,250 but that, you know, in this space, we need to be, I think, more concerned about. 641 01:14:44,250 --> 01:14:51,500 With the companies that have a strategic market position. Hello, my name is Beatrice. 642 01:14:51,500 --> 01:14:54,470 I am a researcher at the Blavatnik School of Government. 643 01:14:54,470 --> 01:15:00,860 And my question is about the definition of the market for your assessment because the market definition seems 644 01:15:00,860 --> 01:15:08,360 to be very relevant both for the regulatory approach and for the ex-post competition policy perspective. 645 01:15:08,360 --> 01:15:14,030 And I was wondering when you were discussing whether or not the digital market is concentrated, 646 01:15:14,030 --> 01:15:21,110 you showed some practise to assess the relevant market is advertising revenue or 647 01:15:21,110 --> 01:15:25,940 numbers of the time that the hours that users spend online and things like that. 648 01:15:25,940 --> 01:15:35,900 I think we have the tools and analytical framework to assess the relevant markets for many of the online platforms that we have nowadays, 649 01:15:35,900 --> 01:15:45,380 considering that they are more two sided markets and many of them have this kind of ecosystem environment where they influence other markets as well. 650 01:15:45,380 --> 01:15:50,630 Yeah. So on the Australian Consumer and Competition Commission, 651 01:15:50,630 --> 01:15:57,560 triple see whatever there that stands for their report on digital platforms said We were assigned the price, 652 01:15:57,560 --> 01:16:01,230 you know, project of digital platforms or digital markets. We set out to do it. 653 01:16:01,230 --> 01:16:06,200 We've done all this research and we realised that this entire report is about Facebook and Google. 654 01:16:06,200 --> 01:16:13,520 And so for the next 400 pages, we're just going to refer to the two of them and not say digital markets. 655 01:16:13,520 --> 01:16:18,560 We say digital markets and digital platforms throughout. 656 01:16:18,560 --> 01:16:25,370 You know, certainly a disproportionate amount of that is, you know, certain companies and certain platforms. 657 01:16:25,370 --> 01:16:29,750 Part of it is that the digital markets unit would need to figure out through 658 01:16:29,750 --> 01:16:36,710 this strategic market position and what other parts of the economy it applied. 659 01:16:36,710 --> 01:16:44,730 Certainly. And I completely glossed over this in my speaking today, but I think we didn't gloss over this in our thinking. 660 01:16:44,730 --> 01:16:50,600 You know, these platforms are very, very different from each other in terms of their business models and the like. 661 01:16:50,600 --> 01:16:57,350 And so that's again, why there isn't like an answer, but types of answers. 662 01:16:57,350 --> 01:17:01,940 And then you figure out how to apply the types of answers to the different markets. 663 01:17:01,940 --> 01:17:07,410 But all of them have in common that, you know. 664 01:17:07,410 --> 01:17:14,910 There are some, you know, gateway entry point bottleneck that you need to go through them to access other 665 01:17:14,910 --> 01:17:23,370 things in a broader economy like a road that standards and data matter quite a lot. 666 01:17:23,370 --> 01:17:33,330 So those would be some of the common features and thus the common policy levers we'd have in the different platforms. 667 01:17:33,330 --> 01:17:38,850 Thanks, Nikita. Well, I'm a researcher, the Faculty of Law and the Oxford Internet Institute. 668 01:17:38,850 --> 01:17:43,620 I was at a question about the first function that you described as the digital services unit, 669 01:17:43,620 --> 01:17:49,020 which was, as I understood, designate designating systemically important companies. 670 01:17:49,020 --> 01:17:55,020 And you drew an analogy with the Sufis and the sort of post-crisis paradigm of the FSB and so on. 671 01:17:55,020 --> 01:18:03,480 So I'm I'm curious to know how you imagine systems being defined in the technical technological context or in the tech sector where, you know, 672 01:18:03,480 --> 01:18:12,330 sort of the parallel of freezing deposits or access to sort of systemically important financial functions doesn't quite translate in the same way. 673 01:18:12,330 --> 01:18:19,530 Right now, if you have a small business and you want people to find you, you need to be, 674 01:18:19,530 --> 01:18:25,230 you know, findable on Google search engine because that's the way people find you. 675 01:18:25,230 --> 01:18:34,200 If you're selling products, you know you may need to be able to sell them on Amazon because that's where I think half of online shopping is. 676 01:18:34,200 --> 01:18:40,980 If you know to Harvard for its newly admitted students, set up a way for them to communicate with each other. 677 01:18:40,980 --> 01:18:47,550 It uses Facebook and the students need to go on Facebook if they don't want to be excluded from the ability to, 678 01:18:47,550 --> 01:18:56,490 like, get to know and see if they know any of their classmates. So the characteristics that these have is, you know, 679 01:18:56,490 --> 01:19:05,070 that word I keep using gateway or bottleneck that it's something to access other things, you know, like a road, 680 01:19:05,070 --> 01:19:13,110 like a fibre optic cable, like a water main that you know, you sort of need to travel over, 681 01:19:13,110 --> 01:19:21,490 get through for a broader set of things, not just for accessing that company's products. 682 01:19:21,490 --> 01:19:30,190 OK. I'm Martin Chalmers. I have a corporate background, but I'm currently an independent adviser in digital business transformation, 683 01:19:30,190 --> 01:19:37,360 and I'm interested in how the success of the proposed unit and policies will be measured. 684 01:19:37,360 --> 01:19:44,110 So will it be on an input basis? How much of the increase in competition or will there be outcome measures of success? 685 01:19:44,110 --> 01:19:52,430 Are we making society healthier and given the complex trade offs involved, is thinking about that possibly part of that consultation process? 686 01:19:52,430 --> 01:19:58,090 So of course, there could be advantages to managing the change with industry making that part of that process. 687 01:19:58,090 --> 01:20:09,400 Yeah, we had an epilogue or conclusion to the report that tried to sketch out what it might look like for consumers to see 688 01:20:09,400 --> 01:20:19,840 greater ability to interoperate between different products to multi home or to switch than you're currently able to. 689 01:20:19,840 --> 01:20:24,610 For, you know, businesses, you'd see more start-ups flourishing, you know, 690 01:20:24,610 --> 01:20:32,920 building their products on top of others or building independent rivals to them. 691 01:20:32,920 --> 01:20:41,580 And, you know, in the economy as a whole, that would sort of look like, you know, more innovation and and more choice. 692 01:20:41,580 --> 01:20:48,250 So are those detailed benchmarks such that, you know, like GDP? 693 01:20:48,250 --> 01:20:51,850 I could say if it goes up this amount, we succeeded in that amount, we failed. 694 01:20:51,850 --> 01:20:58,990 Certainly not and actually can think a lot of economic policies that you could benchmark by GDP, either. 695 01:20:58,990 --> 01:21:08,390 But I think you two want. You know, yes, you want things like the cell phone number portability where I think it happened here, 696 01:21:08,390 --> 01:21:11,950 it's certainly half of the United States where you take your cell phone number from company to company, 697 01:21:11,950 --> 01:21:17,320 and all of a sudden people have been locked in. We're able to switch and that increased competition prices came down. 698 01:21:17,320 --> 01:21:21,880 I think a lot of people knew that that happened, and even if they didn't, 699 01:21:21,880 --> 01:21:30,630 they certainly knew they were seeing a lot more advertisements and enticements to switch and did switch more. 700 01:21:30,630 --> 01:21:36,690 Thank you. My name is Scott Gisele's, and I am not the economist. 701 01:21:36,690 --> 01:21:44,510 I came from the educational field. And my question is you mentioned a few and a few times there. 702 01:21:44,510 --> 01:21:54,870 The them fill a hole in free and competition market will decide what is a fail 703 01:21:54,870 --> 01:22:03,480 fail and how and what is their theory about what is fair competition and market. 704 01:22:03,480 --> 01:22:11,790 You also mentioned the businesses the necessity to achieve a thrust. 705 01:22:11,790 --> 01:22:15,500 You can achieve class in global markets and city. 706 01:22:15,500 --> 01:22:19,800 You have a lot of enemies around you. 707 01:22:19,800 --> 01:22:26,250 And you also mentioned that we should move to balance the. 708 01:22:26,250 --> 01:22:34,160 What is good balance of who will decide and decide what is a good balance of the? 709 01:22:34,160 --> 01:22:39,740 There's a lot of different questions, I think. You know, again, if it's sort of if you have competition, 710 01:22:39,740 --> 01:22:45,950 then you don't worry that much about fairness or company cafe can charge 80 pounds for coffee. 711 01:22:45,950 --> 01:22:52,190 That's not unfair. It's just a stupid thing for them to do because they're not going to sell a lot of coffee. 712 01:22:52,190 --> 01:22:56,390 And the remedy for that is people not buying from that cafe. 713 01:22:56,390 --> 01:23:04,940 When you have, you know, one or one or two companies, then I think it's sort of a little bit harder not to use words like fairness, 714 01:23:04,940 --> 01:23:11,210 even though I think underlying those there hopefully is some notion of efficiency. 715 01:23:11,210 --> 01:23:15,860 That one's getting out for something like balance of harms that would be, you know, 716 01:23:15,860 --> 01:23:19,970 that would be the competition and markets are they bring merger cases now. 717 01:23:19,970 --> 01:23:29,870 They already review a merger and they say, you know, trying to determine whether this merger is in the interests of consumers or not. 718 01:23:29,870 --> 01:23:34,340 The shorthand for that is will it result in a substantial lessening of competition? 719 01:23:34,340 --> 01:23:41,310 And the answer is probably, you know, won't be good. And if it doesn't do that, then, you know, great, be more efficient. 720 01:23:41,310 --> 01:23:44,000 You know, we certainly wouldn't want to stop it. 721 01:23:44,000 --> 01:23:50,480 And so they have a set of judgements they make right now under economic procedures and they'll do modelling. 722 01:23:50,480 --> 01:23:51,710 And but they also look at, you know, 723 01:23:51,710 --> 01:23:59,490 emails people send to each other and they'll like and they'll render a judgement and then the courts get to review that judgement. 724 01:23:59,490 --> 01:24:02,240 And so that's exactly how it would be here. 725 01:24:02,240 --> 01:24:10,690 If it was about a merger, the balance of harms would be CMA making the judgement and then the courts could review that judgement. 726 01:24:10,690 --> 01:24:17,770 Thank you. Yeah. Final question. My name is David, and I work for the UK Economic Statistics Centre of Excellence in London. 727 01:24:17,770 --> 01:24:23,380 I have a question about you mentioned the role of data and algorithms and intangible assets in general, 728 01:24:23,380 --> 01:24:26,440 and we know that these are highly mobile across borders, obviously. 729 01:24:26,440 --> 01:24:33,520 And I think going back to your last slide, I mean, I wanted to what degree than we actually need a kind of multi lateral approach to these things. 730 01:24:33,520 --> 01:24:40,990 And he would like to know your thoughts on kind of reasoned proposals on trying to tax revenues or using networks in that regard, 731 01:24:40,990 --> 01:24:45,860 rather than profits of companies, which are tricky to allocate, obviously. Thanks. 732 01:24:45,860 --> 01:24:56,860 Oh, I mean, if we could have a really a fish and expert global body to do this, that would be better than lots of national ones. 733 01:24:56,860 --> 01:24:59,560 But I just don't think that would happen. 734 01:24:59,560 --> 01:25:08,590 I don't think we'd have a global body, and if we did, it would look more like the entire national telecommunications in the ITU than it does. 735 01:25:08,590 --> 01:25:15,430 You know, the digital markets unit, which at least to date has not done a single thing anyone can complain about. 736 01:25:15,430 --> 01:25:22,000 So I just think this is a matter of being pragmatic that yes, have G20 discussions, 737 01:25:22,000 --> 01:25:27,400 G7 discussions or OECD discussions you have when you have this in mergers now, 738 01:25:27,400 --> 01:25:34,840 I mean, there is relatively similar rules under which what mergers get allowed and don't get allowed in the United States era. 739 01:25:34,840 --> 01:25:41,890 You know, Australia, New Zealand, et cetera. Companies can reasonably predict that if something's not going to be allowed, 740 01:25:41,890 --> 01:25:48,340 it's probably not going to by multiple of them or it will be allowed by multiple of them, so they use similar rules. 741 01:25:48,340 --> 01:25:55,210 Conferences have lots of people flying back and forth and a lot of the same economic experts. 742 01:25:55,210 --> 01:26:02,680 But at the end of the day, it's DG Comp or the CMA or the FTC or the DOJ making the decision. 743 01:26:02,680 --> 01:26:08,260 They may even communicate with each other if it's an international, but it's not a group decision. 744 01:26:08,260 --> 01:26:18,890 So I think that's the best we could aspire to is a similar set of rules, but each being applied and enforced separately. 745 01:26:18,890 --> 01:26:19,250 Great. 746 01:26:19,250 --> 01:26:29,360 Well, thank you for that very rich set of questions, many, many questions, and thanks, Jason, for being so concise and effective in answering them. 747 01:26:29,360 --> 01:26:38,300 We got through many of them. And I think I'm certainly amazed by the breadth and scale of what you are grappling with. 748 01:26:38,300 --> 01:26:44,330 Thanks for coming and spending the evening with us. And thanks to you and all of you for coming. 749 01:26:44,330 --> 01:26:54,342 Thank you.