1 00:00:01,095 --> 00:00:03,420 The following content is provided under a Creative 2 00:00:03,420 --> 00:00:04,810 Commons license. 3 00:00:04,810 --> 00:00:07,020 Your support will help MIT OpenCourseWare 4 00:00:07,020 --> 00:00:11,110 continue to offer high quality educational resources for free. 5 00:00:11,110 --> 00:00:13,680 To make a donation or to view additional materials 6 00:00:13,680 --> 00:00:17,640 from hundreds of MIT courses, visit MIT OpenCourseWare 7 00:00:17,640 --> 00:00:18,526 at ocw.mit.edu. 8 00:00:24,070 --> 00:00:25,790 GARY GENSLER: Today's a little bit 9 00:00:25,790 --> 00:00:31,010 of a discussion about clearing and settlement systems. 10 00:00:31,010 --> 00:00:32,720 Maybe you feel you've heard everything 11 00:00:32,720 --> 00:00:35,270 you needed from Jeff Sprecher when he was asked 12 00:00:35,270 --> 00:00:37,340 a question when he was with us. 13 00:00:37,340 --> 00:00:39,180 Jeff runs some of the largest clearing 14 00:00:39,180 --> 00:00:42,860 houses in the world at Intercontinental Exchange 15 00:00:42,860 --> 00:00:47,660 not only because he started in the energy business, 16 00:00:47,660 --> 00:00:51,650 but he has big credit default swap clearing houses and energy 17 00:00:51,650 --> 00:00:53,630 clearing houses and the like. 18 00:00:53,630 --> 00:00:57,740 And he said, well, I don't know about blockchain technology 19 00:00:57,740 --> 00:00:58,910 for clearing and settling. 20 00:00:58,910 --> 00:01:02,180 But I'm going to try this Bitcoin project that you all 21 00:01:02,180 --> 00:01:03,800 know about and he talked about. 22 00:01:03,800 --> 00:01:06,277 But today we're going to try to dig into, 23 00:01:06,277 --> 00:01:08,360 well, maybe there's something there, because there 24 00:01:08,360 --> 00:01:10,010 are projects going on. 25 00:01:10,010 --> 00:01:11,380 And why are there projects? 26 00:01:11,380 --> 00:01:16,580 They're mostly permissioned and not permissionless systems. 27 00:01:16,580 --> 00:01:20,160 But I think there's something there. 28 00:01:20,160 --> 00:01:22,560 This will be one of the more optimistic-- 29 00:01:22,560 --> 00:01:27,120 if you're judging how often I get to the maximal side, 30 00:01:27,120 --> 00:01:29,540 I think there is probably something there, at least 31 00:01:29,540 --> 00:01:32,420 for permissioned blockchain projects in the clearing 32 00:01:32,420 --> 00:01:34,940 and settling. 33 00:01:34,940 --> 00:01:38,270 I'm going to-- we'll, of course, go through the readings. 34 00:01:38,270 --> 00:01:41,060 And these are the sort of topics we'll go through. 35 00:01:41,060 --> 00:01:43,010 Clearing and settling services-- 36 00:01:43,010 --> 00:01:47,330 I'm going to try to set this stage as to what this is. 37 00:01:47,330 --> 00:01:50,210 Frankly, when I was on Wall Street for 18 years, 38 00:01:50,210 --> 00:01:53,210 I couldn't really define what clearing and settling. 39 00:01:53,210 --> 00:01:53,960 Sorry, Ross. 40 00:01:53,960 --> 00:01:56,950 You thought I could define it when I was on Wall Street? 41 00:01:56,950 --> 00:01:57,790 What's that? 42 00:01:57,790 --> 00:01:59,570 AUDIENCE: I just strictly know someone else can't either. 43 00:01:59,570 --> 00:02:00,362 GARY GENSLER: Yeah. 44 00:02:00,362 --> 00:02:00,872 Yeah. 45 00:02:00,872 --> 00:02:03,080 Well, when you're in the merger and acquisition area, 46 00:02:03,080 --> 00:02:04,910 you usually don't think about what 47 00:02:04,910 --> 00:02:07,657 it means to clear something and to settle it. 48 00:02:07,657 --> 00:02:09,199 And frankly, even usually when you're 49 00:02:09,199 --> 00:02:12,110 on the trading floor, that's something somebody else 50 00:02:12,110 --> 00:02:14,840 takes care of. 51 00:02:14,840 --> 00:02:16,580 Just the nature of-- 52 00:02:16,580 --> 00:02:20,470 how many people have worked on trading floors here? 53 00:02:20,470 --> 00:02:20,970 All right. 54 00:02:20,970 --> 00:02:23,850 So I'm going to have seven or eight of you start. 55 00:02:23,850 --> 00:02:26,310 And you're on notice now. 56 00:02:26,310 --> 00:02:30,490 In about eight to 10 minutes, you're going to help. 57 00:02:30,490 --> 00:02:34,147 You might prove what I just said to Ross, that you can't define 58 00:02:34,147 --> 00:02:36,480 what clearing and settling is because somebody else took 59 00:02:36,480 --> 00:02:37,100 care of it. 60 00:02:37,100 --> 00:02:38,400 But we'll see. 61 00:02:38,400 --> 00:02:40,140 We'll see. 62 00:02:40,140 --> 00:02:42,450 And then we'll talk about blockchain technology 63 00:02:42,450 --> 00:02:46,060 applicability, some of the projects that are going on, 64 00:02:46,060 --> 00:02:48,330 and then move over to this. 65 00:02:48,330 --> 00:02:51,135 There was a reading from the International Swaps 66 00:02:51,135 --> 00:02:54,600 and Derivatives Association, ISDA, 67 00:02:54,600 --> 00:03:00,090 what they're doing around smart contracts in this space. 68 00:03:00,090 --> 00:03:02,613 But what I want to do first is a little bit 69 00:03:02,613 --> 00:03:04,530 of talking about final projects, because we're 70 00:03:04,530 --> 00:03:06,120 at that time of the year. 71 00:03:06,120 --> 00:03:09,420 And this is to help. 72 00:03:09,420 --> 00:03:12,990 Then I'm going to go through eight or nine questions. 73 00:03:12,990 --> 00:03:17,280 If you haven't already thought through your projects, 74 00:03:17,280 --> 00:03:18,690 these are eight or nine questions 75 00:03:18,690 --> 00:03:22,740 that I would think you'd want to think about, 76 00:03:22,740 --> 00:03:24,810 which is another way of saying it will probably 77 00:03:24,810 --> 00:03:28,260 enhance your performance. 78 00:03:28,260 --> 00:03:31,530 What else can I say? 79 00:03:31,530 --> 00:03:34,260 And number one-- and it's a little, small print. 80 00:03:34,260 --> 00:03:34,890 I'm sorry. 81 00:03:34,890 --> 00:03:37,110 Or no, it's not too bad-- 82 00:03:37,110 --> 00:03:41,340 is what's the value creation proposition, or pain point, 83 00:03:41,340 --> 00:03:42,100 if you wish. 84 00:03:42,100 --> 00:03:44,070 So are you creating value? 85 00:03:44,070 --> 00:03:45,428 Are you solving a pain point? 86 00:03:45,428 --> 00:03:47,220 Or as Jeff Sprecher would say, what are you 87 00:03:47,220 --> 00:03:51,880 doing that's going to make something cheaper, faster, 88 00:03:51,880 --> 00:03:53,610 better? 89 00:03:53,610 --> 00:03:55,470 What's the value proposition? 90 00:03:55,470 --> 00:03:56,880 Whatever space. 91 00:03:56,880 --> 00:04:00,320 There's two or three groups that are doing trade finance. 92 00:04:00,320 --> 00:04:02,220 There's one or two doing payments. 93 00:04:02,220 --> 00:04:06,990 There's some group doing commercial-- 94 00:04:06,990 --> 00:04:11,470 I mean, consumer credit rating or so forth. 95 00:04:11,470 --> 00:04:12,942 There's a wide variety. 96 00:04:12,942 --> 00:04:14,400 There's one or two groups that have 97 00:04:14,400 --> 00:04:18,209 come to see me that are doing things outside of finance. 98 00:04:18,209 --> 00:04:22,660 But regardless, what's the value proposition? 99 00:04:22,660 --> 00:04:25,170 What pain points are you addressing? 100 00:04:25,170 --> 00:04:26,900 Or as Jeff would say, what are you going 101 00:04:26,900 --> 00:04:30,000 to do better, faster, cheaper? 102 00:04:30,000 --> 00:04:33,960 If you can't answer that threshold question, 103 00:04:33,960 --> 00:04:35,820 shake it loose. 104 00:04:35,820 --> 00:04:37,350 Kick it around. 105 00:04:37,350 --> 00:04:42,270 Or move on to a different use case, frankly. 106 00:04:42,270 --> 00:04:43,950 Secondly, [INAUDIBLE] you get into, 107 00:04:43,950 --> 00:04:46,170 like, what is this blockchain technology? 108 00:04:46,170 --> 00:04:50,610 What transactions or data is going to be recorded? 109 00:04:50,610 --> 00:04:52,650 Because this is fundamentally, ultimately, 110 00:04:52,650 --> 00:04:57,130 a database technology, even the permissionless systems. 111 00:04:57,130 --> 00:05:00,210 And you'll see a later question is why this versus 112 00:05:00,210 --> 00:05:01,860 traditional database system? 113 00:05:01,860 --> 00:05:03,810 But what transactions, what data-- 114 00:05:03,810 --> 00:05:05,880 whether it's health care records, 115 00:05:05,880 --> 00:05:10,260 whether it's consumer credit data, supply chain, 116 00:05:10,260 --> 00:05:13,350 moving a piece of property along a supply chain, 117 00:05:13,350 --> 00:05:16,230 but what is the actual data and transactions that you're 118 00:05:16,230 --> 00:05:17,550 recording? 119 00:05:17,550 --> 00:05:19,820 Questions? 120 00:05:19,820 --> 00:05:23,510 What multiple stakeholders need to write and read access 121 00:05:23,510 --> 00:05:24,590 to the ledgers? 122 00:05:24,590 --> 00:05:26,390 And I underline the word multiple. 123 00:05:26,390 --> 00:05:30,080 Because really, when you think about the value proposition 124 00:05:30,080 --> 00:05:35,950 of blockchain technology, it's about immutable records, 125 00:05:35,950 --> 00:05:38,180 or so-called immutable records. 126 00:05:38,180 --> 00:05:41,630 That's verifiable, immutable records 127 00:05:41,630 --> 00:05:45,090 amongst multiple parties. 128 00:05:45,090 --> 00:05:46,560 And it cost extra money. 129 00:05:46,560 --> 00:05:49,710 Everything we've been studying about in this semester 130 00:05:49,710 --> 00:05:51,330 is about there's more complexity. 131 00:05:51,330 --> 00:05:55,290 There's more cost, in some way, of this decentralized, 132 00:05:55,290 --> 00:06:00,300 even permissioned systems, decentralized system. 133 00:06:00,300 --> 00:06:02,100 But you get something for it. 134 00:06:02,100 --> 00:06:06,100 You get some verification of transactions, of data, 135 00:06:06,100 --> 00:06:08,070 of mutability. 136 00:06:08,070 --> 00:06:10,800 But you also have to have some logic 137 00:06:10,800 --> 00:06:12,420 that there's multiple parties that 138 00:06:12,420 --> 00:06:15,600 need to both write and read. 139 00:06:15,600 --> 00:06:19,390 A distributed database that just does read only-- 140 00:06:19,390 --> 00:06:23,290 I can read my Bank of America bank account. 141 00:06:23,290 --> 00:06:26,230 I can see what's in that sort of ledger. 142 00:06:26,230 --> 00:06:28,180 It doesn't have to be on a blockchain. 143 00:06:28,180 --> 00:06:31,720 So multiple parties writing, why is that important? 144 00:06:31,720 --> 00:06:36,700 Forming consensus, in this circumstance. 145 00:06:36,700 --> 00:06:39,610 Which specific cost of verification or networking 146 00:06:39,610 --> 00:06:40,510 are you lowering? 147 00:06:40,510 --> 00:06:43,780 We talked about a lot of the verification costs, 148 00:06:43,780 --> 00:06:47,350 whether it's privacy, censorship resistance, 149 00:06:47,350 --> 00:06:49,258 whether it's some direct cost. 150 00:06:49,258 --> 00:06:50,800 Maybe you're going right at something 151 00:06:50,800 --> 00:06:53,590 that's got a bunch of economic rents in it. 152 00:06:53,590 --> 00:06:55,750 So there was a series of verification costs 153 00:06:55,750 --> 00:06:57,460 we talked about. 154 00:06:57,460 --> 00:07:00,490 What networking costs are you trying to reduce? 155 00:07:04,480 --> 00:07:06,430 There's probably been something in the order 156 00:07:06,430 --> 00:07:11,470 of 4,000 or 5,000 white papers, many that you can find. 157 00:07:11,470 --> 00:07:13,062 You can't find all of them. 158 00:07:13,062 --> 00:07:14,770 But if you're going at something that you 159 00:07:14,770 --> 00:07:16,660 think a white paper has been written about 160 00:07:16,660 --> 00:07:19,240 or a blockchain company has been started on, 161 00:07:19,240 --> 00:07:22,080 figure out what the competition is doing. 162 00:07:22,080 --> 00:07:23,920 And for the two groups, and it may be more, 163 00:07:23,920 --> 00:07:26,620 that are doing something around trade, finance, and supply 164 00:07:26,620 --> 00:07:30,730 chain, you know there's a bunch of stuff you can read. 165 00:07:30,730 --> 00:07:32,260 We'll be studying it next Tuesday. 166 00:07:32,260 --> 00:07:32,973 Alin? 167 00:07:32,973 --> 00:07:35,390 AUDIENCE: Can you give us an example of a networking cost? 168 00:07:35,390 --> 00:07:36,670 I'm not sure what that is. 169 00:07:36,670 --> 00:07:40,690 GARY GENSLER: So networking cost is often 170 00:07:40,690 --> 00:07:44,170 why people are using a native token. 171 00:07:44,170 --> 00:07:50,730 But networking costs are how to jumpstart a network. 172 00:07:50,730 --> 00:07:54,270 Uber jumpstarted a network, in a sense, 173 00:07:54,270 --> 00:07:55,850 without blockchain technology. 174 00:07:55,850 --> 00:07:57,750 But an example of networking costs 175 00:07:57,750 --> 00:08:00,200 are if you walked out of the airport 176 00:08:00,200 --> 00:08:04,790 and you didn't think any drivers are using this application, 177 00:08:04,790 --> 00:08:08,480 then you wouldn't be terribly interested in using it. 178 00:08:08,480 --> 00:08:09,720 So-- 179 00:08:09,720 --> 00:08:12,860 AUDIENCE: But who incurs the cost? 180 00:08:12,860 --> 00:08:16,760 GARY GENSLER: Often the cost of building a network can 181 00:08:16,760 --> 00:08:19,860 be a tremendous startup cost. 182 00:08:19,860 --> 00:08:22,052 So how to start a network often is 183 00:08:22,052 --> 00:08:23,760 you go to a bunch of venture capitalists. 184 00:08:23,760 --> 00:08:25,010 You raise money. 185 00:08:25,010 --> 00:08:26,600 You advertise. 186 00:08:26,600 --> 00:08:29,430 And you try to get participants. 187 00:08:29,430 --> 00:08:32,900 And one of the challenges of health care records-- 188 00:08:32,900 --> 00:08:35,990 we don't have our health care experts in here today. 189 00:08:39,825 --> 00:08:40,700 I was looking around. 190 00:08:40,700 --> 00:08:44,240 We have two students from Harvard's public policy 191 00:08:44,240 --> 00:08:48,170 school doing graduate work and sometimes join us. 192 00:08:48,170 --> 00:08:50,460 But health care records, for instance, 193 00:08:50,460 --> 00:08:54,650 there's some really interesting blockchain technology 194 00:08:54,650 --> 00:08:56,070 solutions. 195 00:08:56,070 --> 00:08:57,920 But the networking challenge is how do you 196 00:08:57,920 --> 00:09:02,500 get a bunch of medical providers, hospitals, doctors, 197 00:09:02,500 --> 00:09:06,350 and users to actually join the network. 198 00:09:06,350 --> 00:09:09,230 So that would be an example of network cost 199 00:09:09,230 --> 00:09:11,395 is basically adoption often. 200 00:09:11,395 --> 00:09:12,020 Does that help? 201 00:09:12,020 --> 00:09:12,645 AUDIENCE: Yeah. 202 00:09:15,390 --> 00:09:17,730 GARY GENSLER: What are the competitors doing? 203 00:09:17,730 --> 00:09:19,380 I'm not looking for plagiarism. 204 00:09:19,380 --> 00:09:23,580 But if you're in a space, payments or supply chain 205 00:09:23,580 --> 00:09:26,130 or consumer credit or some others, 206 00:09:26,130 --> 00:09:29,010 where there's three or six or 10 white papers out there 207 00:09:29,010 --> 00:09:31,680 and maybe somebody has even raised 20 million 208 00:09:31,680 --> 00:09:37,050 or a couple million in an ICO, then it would be interesting. 209 00:09:37,050 --> 00:09:38,730 You should feel free to look at those. 210 00:09:38,730 --> 00:09:40,470 And maybe you'll be informed by it. 211 00:09:40,470 --> 00:09:42,220 And you'll say, yeah, I'm doing it better. 212 00:09:42,220 --> 00:09:44,550 Or they've got it poorly here. 213 00:09:44,550 --> 00:09:47,070 Or their project's in Estonia. 214 00:09:47,070 --> 00:09:50,730 And I'm going to do a digital ID project in the US. 215 00:09:50,730 --> 00:09:54,240 And our regulatory system is different than Estonia. 216 00:09:54,240 --> 00:09:58,060 But I wouldn't be bashful about looking at the competition. 217 00:09:58,060 --> 00:09:59,400 In fact, I would say that's-- 218 00:09:59,400 --> 00:10:00,400 we're a business school. 219 00:10:00,400 --> 00:10:02,130 You always look at the competition. 220 00:10:02,130 --> 00:10:05,040 And you see if you can do either what they're doing better 221 00:10:05,040 --> 00:10:07,140 or what they're doing in a different geography 222 00:10:07,140 --> 00:10:09,500 or in a little bit different market. 223 00:10:09,500 --> 00:10:11,750 Back to how Alin likes to define things, 224 00:10:11,750 --> 00:10:14,660 I've not used the term blockchain technology yet 225 00:10:14,660 --> 00:10:17,120 on this slide, have I? 226 00:10:17,120 --> 00:10:22,940 But why are append-only logs at multiple party consensus 227 00:10:22,940 --> 00:10:25,780 the best solution? 228 00:10:25,780 --> 00:10:27,730 Which kind of goes back to why not just 229 00:10:27,730 --> 00:10:28,960 a traditional database. 230 00:10:28,960 --> 00:10:34,390 But why append-only, timestamped, verifiable logs? 231 00:10:34,390 --> 00:10:37,420 They're connected by cryptographic hash functions. 232 00:10:37,420 --> 00:10:39,790 But why have these logs? 233 00:10:39,790 --> 00:10:41,770 And why have multiple parties in essence 234 00:10:41,770 --> 00:10:46,665 form a consensus about the state of the ledger? 235 00:10:46,665 --> 00:10:47,307 Alin? 236 00:10:47,307 --> 00:10:47,890 AUDIENCE: Yes. 237 00:10:47,890 --> 00:10:51,590 I have a question about the timestamped, apend-only logs. 238 00:10:51,590 --> 00:10:54,420 Is that only relevant in blockchain, 239 00:10:54,420 --> 00:10:56,210 that traditional databases do not 240 00:10:56,210 --> 00:11:01,740 have these kinds of features, timestamp and append-only logs? 241 00:11:01,740 --> 00:11:04,740 GARY GENSLER: They can have time-- 242 00:11:04,740 --> 00:11:08,690 timestamps in essence. 243 00:11:08,690 --> 00:11:11,130 In essence, it's innovation of the early '90s. 244 00:11:11,130 --> 00:11:13,070 And then Satoshi Nakamoto grabbed it. 245 00:11:13,070 --> 00:11:16,520 It was the concept that you take a block of data 246 00:11:16,520 --> 00:11:20,120 and you connect it to the next block with a cryptographic hash 247 00:11:20,120 --> 00:11:20,862 function. 248 00:11:20,862 --> 00:11:22,820 And once you've done that, then it's committed. 249 00:11:22,820 --> 00:11:26,050 You can't amend anything that's before 250 00:11:26,050 --> 00:11:29,460 because it's a unique hash. 251 00:11:29,460 --> 00:11:33,880 And traditional databases did not have similar commitment 252 00:11:33,880 --> 00:11:34,380 schemes. 253 00:11:34,380 --> 00:11:36,360 They do have other commitments schemes, but not 254 00:11:36,360 --> 00:11:39,720 this commitment scheme of taking a block of data, 255 00:11:39,720 --> 00:11:44,700 having the hash function, which then means never shall 256 00:11:44,700 --> 00:11:48,520 you amend what came before. 257 00:11:48,520 --> 00:11:49,960 That's how I understand it. 258 00:11:49,960 --> 00:11:53,295 Did you have a different view on traditional databases? 259 00:11:53,295 --> 00:11:54,670 AUDIENCE: Well, I guess this gets 260 00:11:54,670 --> 00:11:56,620 to the root of the question. 261 00:11:56,620 --> 00:11:59,510 Is there something new about blockchain technology? 262 00:11:59,510 --> 00:12:02,050 In my opinion, there isn't, actually. 263 00:12:02,050 --> 00:12:06,140 20 years ago, we knew how to make a traditional database 264 00:12:06,140 --> 00:12:09,380 and make it full of power and make it a census protocol, 265 00:12:09,380 --> 00:12:12,497 or what people nowadays like to call a permissioned blockchain. 266 00:12:12,497 --> 00:12:13,580 So we knew how to do that. 267 00:12:13,580 --> 00:12:15,372 And that would give you an append-only log. 268 00:12:15,372 --> 00:12:18,820 It's nothing exceptional. 269 00:12:21,808 --> 00:12:22,600 GARY GENSLER: Yeah. 270 00:12:22,600 --> 00:12:26,800 So there are debates as to where is the border and boundary 271 00:12:26,800 --> 00:12:28,420 between a permissioned blockchain 272 00:12:28,420 --> 00:12:31,190 and a traditional database. 273 00:12:31,190 --> 00:12:35,210 But I've chosen to, at least for this class, 274 00:12:35,210 --> 00:12:39,890 a hyperledger fabric, we'll call that a blockchain project. 275 00:12:39,890 --> 00:12:43,970 But recognize there are some who would say, well, that's-- 276 00:12:43,970 --> 00:12:48,610 unless you have a native token, it's not really blockchain. 277 00:12:48,610 --> 00:12:49,275 Eric. 278 00:12:49,275 --> 00:12:56,330 AUDIENCE: --one could call point recent days that the trust is 279 00:12:56,330 --> 00:13:00,760 the key attribute that's not the factor in this equation, 280 00:13:00,760 --> 00:13:04,290 because you can get a multiple-- 281 00:13:04,290 --> 00:13:08,950 I mean, multiple party consensus doesn't have value on its own. 282 00:13:08,950 --> 00:13:12,760 It's actually multiple party consensus is-- 283 00:13:12,760 --> 00:13:17,140 it's actually something you had to come up as a necessity that 284 00:13:17,140 --> 00:13:19,150 stems from the fact that you're dealing 285 00:13:19,150 --> 00:13:22,960 with a decentralized arrangement in which you're not trusting 286 00:13:22,960 --> 00:13:26,940 any of the parties that intervening 287 00:13:26,940 --> 00:13:28,790 in the exchange of information. 288 00:13:28,790 --> 00:13:29,620 Right? 289 00:13:29,620 --> 00:13:32,020 GARY GENSLER: Well, I agree with the first half, 290 00:13:32,020 --> 00:13:36,130 that you wouldn't even want to use Hyperledger 291 00:13:36,130 --> 00:13:39,880 Fabric unless you really need multiple parties writing 292 00:13:39,880 --> 00:13:40,870 to the ledger. 293 00:13:40,870 --> 00:13:43,990 And I think that does separate it a little bit. 294 00:13:43,990 --> 00:13:45,860 You could take an Oracle database 295 00:13:45,860 --> 00:13:48,340 and allow it to be multiple parties, too. 296 00:13:48,340 --> 00:13:51,130 But I think that separates it. 297 00:13:55,090 --> 00:13:57,220 You might be more the purist than I am, Eric. 298 00:13:57,220 --> 00:14:02,220 I don't think it has to be completely open, where 299 00:14:02,220 --> 00:14:05,710 you don't know all the participants. 300 00:14:05,710 --> 00:14:07,600 But even in a financial-- we'll talk 301 00:14:07,600 --> 00:14:10,090 about the Australian Stock Exchange 302 00:14:10,090 --> 00:14:12,630 in a half an hour or so. 303 00:14:12,630 --> 00:14:17,230 And when we talk about that, that's a limited closed loop. 304 00:14:17,230 --> 00:14:19,060 But they believe that they are actually 305 00:14:19,060 --> 00:14:22,083 taking a lot of cost out of the verification side. 306 00:14:22,083 --> 00:14:23,500 They think they're getting a lot-- 307 00:14:23,500 --> 00:14:25,750 they're lowering verification costs. 308 00:14:25,750 --> 00:14:28,150 And they would contend they're lowering networking costs, 309 00:14:28,150 --> 00:14:30,490 because they don't have the same reconciliations 310 00:14:30,490 --> 00:14:33,550 between multiple databases. 311 00:14:33,550 --> 00:14:35,590 Now, you might call that verification cost. 312 00:14:35,590 --> 00:14:37,630 But I think they would call it both verification 313 00:14:37,630 --> 00:14:40,140 and networking cost. 314 00:14:40,140 --> 00:14:43,390 There's multiple back offices no longer reconciling. 315 00:14:47,770 --> 00:14:51,010 If your final project has a permissionless system 316 00:14:51,010 --> 00:14:54,160 with a native token, why do you need a native token? 317 00:14:54,160 --> 00:14:58,450 Is it to jumpstart the ICO that you want to raise money for? 318 00:14:58,450 --> 00:15:00,520 I'm not averse to that. 319 00:15:00,520 --> 00:15:04,000 But at least fess up and say that's how 320 00:15:04,000 --> 00:15:05,950 we're crowdsourcing this thing. 321 00:15:05,950 --> 00:15:08,250 If it's about that you really believe 322 00:15:08,250 --> 00:15:12,130 there's some benefit to the token economics, 323 00:15:12,130 --> 00:15:16,660 that it's an incentive system to run your new competition 324 00:15:16,660 --> 00:15:20,230 to Uber or an incentive system for health care records 325 00:15:20,230 --> 00:15:23,470 or an incentive system for supply chain, 326 00:15:23,470 --> 00:15:26,810 talk about what that incentive system is. 327 00:15:26,810 --> 00:15:30,280 And are you trying to motivate jumpstarting the network 328 00:15:30,280 --> 00:15:33,210 or maintaining the network later? 329 00:15:33,210 --> 00:15:35,308 And it could be both. 330 00:15:35,308 --> 00:15:36,100 It could be either. 331 00:15:36,100 --> 00:15:39,770 I'm just saying it would be good to have a discussion. 332 00:15:39,770 --> 00:15:43,350 Is the native token just about crowdfunding? 333 00:15:43,350 --> 00:15:45,930 Is the native token also about jumpstarting 334 00:15:45,930 --> 00:15:47,580 this network and then an incentive 335 00:15:47,580 --> 00:15:50,580 to maintain the network amongst multiple parties? 336 00:15:54,570 --> 00:15:57,780 What trade-offs of scalability, performance, privacy, security, 337 00:15:57,780 --> 00:15:58,620 coordination? 338 00:15:58,620 --> 00:16:00,510 I'm not asking you to solve that. 339 00:16:00,510 --> 00:16:03,600 And in fact, you can even jump ahead three and five and seven 340 00:16:03,600 --> 00:16:06,240 years and assume that some of these things 341 00:16:06,240 --> 00:16:07,980 will be taken care of. 342 00:16:07,980 --> 00:16:13,110 So recognize that in the permissioned blockchains, 343 00:16:13,110 --> 00:16:15,420 they seem to have a higher performance, a lot 344 00:16:15,420 --> 00:16:18,700 higher performance, than the native token permissionless 345 00:16:18,700 --> 00:16:19,630 systems. 346 00:16:19,630 --> 00:16:23,940 But if you're proposing a native token permissionless system, 347 00:16:23,940 --> 00:16:27,510 I'd just like a short discussion of how much 348 00:16:27,510 --> 00:16:32,460 these scalability issues matter to you, particularly when 349 00:16:32,460 --> 00:16:35,280 most of the native token permissionless systems 350 00:16:35,280 --> 00:16:37,920 can't handle much more than 1,000 351 00:16:37,920 --> 00:16:40,980 or a handful 1,000 of transactions a day. 352 00:16:40,980 --> 00:16:43,740 It is true that Bitcoin handles quite a bit more than that. 353 00:16:43,740 --> 00:16:49,200 But five or 10 transactions a second, 354 00:16:49,200 --> 00:16:51,720 is that going to work for whatever your solution is? 355 00:16:54,810 --> 00:16:57,980 In fact, I'd like to see some groups use native tokens. 356 00:16:57,980 --> 00:17:01,010 I want to have a variety of things I read in December 357 00:17:01,010 --> 00:17:02,860 rather than reading all permission systems. 358 00:17:05,960 --> 00:17:11,170 And then again, can a permissioned blockchain 359 00:17:11,170 --> 00:17:12,920 or traditional database adequately 360 00:17:12,920 --> 00:17:14,839 address whatever use case if you do 361 00:17:14,839 --> 00:17:17,030 have a permissionless system? 362 00:17:17,030 --> 00:17:22,119 Thinking about that, how can broad adoption be realized? 363 00:17:22,119 --> 00:17:24,619 Are you going to use a native token to equip broad adoption? 364 00:17:24,619 --> 00:17:26,952 How are you going to get a bunch of hospitals signed on? 365 00:17:26,952 --> 00:17:29,200 How are you going to get a bunch of banks signed on? 366 00:17:29,200 --> 00:17:30,908 You're going to do like Jeff Sprecher did 367 00:17:30,908 --> 00:17:33,428 and give 50% of his equity away. 368 00:17:33,428 --> 00:17:34,970 There was that conversation he talked 369 00:17:34,970 --> 00:17:37,280 about he had with Goldman Sachs, that he realized 370 00:17:37,280 --> 00:17:40,070 that was the only way to get his business started 371 00:17:40,070 --> 00:17:43,130 in the late 1990s. 372 00:17:43,130 --> 00:17:48,590 But just how are you going to get that adoption? 373 00:17:48,590 --> 00:17:51,260 And to the extent, this might be a paragraph, 374 00:17:51,260 --> 00:17:54,410 but what type of customer interface or how might it be 375 00:17:54,410 --> 00:17:57,518 better than other customer UIs? 376 00:17:57,518 --> 00:17:59,060 So these are just some of the things, 377 00:17:59,060 --> 00:18:02,060 if you were thinking about a real business. 378 00:18:02,060 --> 00:18:05,510 And I think that the market is moving away 379 00:18:05,510 --> 00:18:08,750 from that big, frothy bull market, 380 00:18:08,750 --> 00:18:12,020 where there was as many as 400 or 500 ICOs a month. 381 00:18:12,020 --> 00:18:15,330 And it was raising anywhere from 2 to 5 billion a month. 382 00:18:19,130 --> 00:18:21,200 October was about 200 ICOs. 383 00:18:21,200 --> 00:18:22,790 And it was about 800 million. 384 00:18:22,790 --> 00:18:25,220 And I think it's going to continue to decline, 385 00:18:25,220 --> 00:18:28,165 partly because they don't all work. 386 00:18:28,165 --> 00:18:29,540 But partly because people are now 387 00:18:29,540 --> 00:18:34,560 getting more rigorous as to what's really going on. 388 00:18:34,560 --> 00:18:36,687 If you are using a permissioned blockchain, 389 00:18:36,687 --> 00:18:38,270 you don't have to tell me whether it's 390 00:18:38,270 --> 00:18:39,388 Hyperledger or Corda. 391 00:18:39,388 --> 00:18:41,180 But if somebody does and you have something 392 00:18:41,180 --> 00:18:44,570 thoughtful to say there, I'm intrigued to find out 393 00:18:44,570 --> 00:18:47,300 why this one versus that one. 394 00:18:47,300 --> 00:18:50,180 It's really just to show your critical analysis 395 00:18:50,180 --> 00:18:51,810 of this subject. 396 00:18:51,810 --> 00:18:53,570 And if you want to prove yourself 397 00:18:53,570 --> 00:18:57,020 on why it's Corda versus Hyperledger Fabric, 398 00:18:57,020 --> 00:18:59,880 that'll be interesting if somebody has a point of view. 399 00:18:59,880 --> 00:19:03,440 But I'm not looking for that level of either computer 400 00:19:03,440 --> 00:19:06,620 science knowledge. 401 00:19:06,620 --> 00:19:08,360 This is from a business perspective. 402 00:19:08,360 --> 00:19:10,170 Think of me as a venture capitalist. 403 00:19:10,170 --> 00:19:14,820 If I was investing, how you pull it together. 404 00:19:14,820 --> 00:19:16,270 So any questions on that? 405 00:19:19,000 --> 00:19:20,890 So we had six or seven people that 406 00:19:20,890 --> 00:19:22,440 actually worked in banking. 407 00:19:22,440 --> 00:19:24,792 Who did we have at this table? 408 00:19:24,792 --> 00:19:26,258 Ah. 409 00:19:26,258 --> 00:19:27,050 So what's clearing? 410 00:19:31,500 --> 00:19:33,110 AUDIENCE: If I recall correctly, you 411 00:19:33,110 --> 00:19:37,230 need to prove the identity of the other party 412 00:19:37,230 --> 00:19:39,640 in order to make sure that you're sending 413 00:19:39,640 --> 00:19:40,640 the asset to the right-- 414 00:19:40,640 --> 00:19:40,920 GARY GENSLER: All right. 415 00:19:40,920 --> 00:19:42,580 So you have to prove the identity. 416 00:19:42,580 --> 00:19:46,140 So it's about validating and authenticating an identity. 417 00:19:46,140 --> 00:19:47,730 Anything else in clearing? 418 00:19:47,730 --> 00:19:49,270 Who else worked in trading? 419 00:19:49,270 --> 00:19:52,110 I had somebody back at the back table worked in-- 420 00:19:52,110 --> 00:19:52,646 ah. 421 00:19:52,646 --> 00:19:55,830 AUDIENCE: Is it that they have the assets to be traded? 422 00:19:55,830 --> 00:19:57,840 GARY GENSLER: They have the assets to be traded. 423 00:19:57,840 --> 00:20:01,770 Well, most equity markets trade on transaction date 424 00:20:01,770 --> 00:20:04,560 plus two days. 425 00:20:04,560 --> 00:20:06,720 So somewhere in that two days, you 426 00:20:06,720 --> 00:20:08,820 have to confirm they have the asset. 427 00:20:08,820 --> 00:20:09,730 That is correct. 428 00:20:09,730 --> 00:20:12,000 You don't have to do it on t plus zero, 429 00:20:12,000 --> 00:20:14,380 normally by t plus 1. 430 00:20:14,380 --> 00:20:15,240 AUDIENCE: I think-- 431 00:20:15,240 --> 00:20:15,990 I could be wrong. 432 00:20:15,990 --> 00:20:19,050 But I think you had to do as well would be the stock 433 00:20:19,050 --> 00:20:21,380 exchange knowing what is the net exposure that you 434 00:20:21,380 --> 00:20:22,380 have at the [INAUDIBLE]. 435 00:20:22,380 --> 00:20:24,240 GARY GENSLER: Net exposure. 436 00:20:24,240 --> 00:20:25,140 Net exposure. 437 00:20:25,140 --> 00:20:26,110 Very key thing. 438 00:20:26,110 --> 00:20:27,480 So there's authentication. 439 00:20:27,480 --> 00:20:32,130 At some point between execution and transaction, 440 00:20:32,130 --> 00:20:35,100 you have to actually know you have the security-- 441 00:20:35,100 --> 00:20:39,260 and then net exposure, which we're going 442 00:20:39,260 --> 00:20:41,850 to look at on another page. 443 00:20:41,850 --> 00:20:45,900 This simple chart of two people cutting the lawns, 444 00:20:45,900 --> 00:20:48,330 but a buyer and seller meet. 445 00:20:48,330 --> 00:20:51,500 But a buyer and seller meet in an exchange 446 00:20:51,500 --> 00:20:55,950 in traditional securities markets through brokers. 447 00:20:55,950 --> 00:20:59,360 That's called an inner mediated access. 448 00:20:59,360 --> 00:21:04,323 In the crypto exchange markets, those two brokers 449 00:21:04,323 --> 00:21:06,240 on the left-hand side and the right-hand side, 450 00:21:06,240 --> 00:21:09,495 the sellers and buyers brokers, don't exist. 451 00:21:12,230 --> 00:21:15,480 In the crypto exchange markets, basically you 452 00:21:15,480 --> 00:21:20,980 have direct access to the market. 453 00:21:20,980 --> 00:21:22,710 But in the traditional stock exchanges, 454 00:21:22,710 --> 00:21:26,025 whether it's London, Tokyo, New York, Beijing-- 455 00:21:31,260 --> 00:21:32,982 where's the exchange in India? 456 00:21:32,982 --> 00:21:34,370 AUDIENCE: Mumbai. 457 00:21:34,370 --> 00:21:36,720 GARY GENSLER: Oh, it's at Mumbai now? 458 00:21:36,720 --> 00:21:37,590 I forgot. 459 00:21:37,590 --> 00:21:40,460 30 years ago, I visited it and walked on the floor. 460 00:21:40,460 --> 00:21:43,990 It was called the Bombay Stock Exchange back then. 461 00:21:43,990 --> 00:21:46,440 But regardless of where the stock exchange is, 462 00:21:46,440 --> 00:21:49,320 there's some clearing house separate from the stock 463 00:21:49,320 --> 00:21:54,490 exchange doing these functions that we just talked about. 464 00:21:54,490 --> 00:22:00,240 So three steps-- execution. 465 00:22:00,240 --> 00:22:01,988 What's it mean to execute a trade? 466 00:22:01,988 --> 00:22:03,780 Alpha, do you know what it means to execute 467 00:22:03,780 --> 00:22:05,652 a trade without reading? 468 00:22:05,652 --> 00:22:06,360 Should I go back? 469 00:22:06,360 --> 00:22:06,985 AUDIENCE: Yeah. 470 00:22:09,270 --> 00:22:10,710 Because I've traded before. 471 00:22:10,710 --> 00:22:11,730 GARY GENSLER: Yeah. 472 00:22:11,730 --> 00:22:12,900 That's why I called on you. 473 00:22:12,900 --> 00:22:15,400 AUDIENCE: We did a great job with a separate execution team. 474 00:22:15,400 --> 00:22:16,670 That helped us with that. 475 00:22:16,670 --> 00:22:21,990 But signing the docs, making sure that the account 476 00:22:21,990 --> 00:22:24,470 information is all there, if it's been done through NSDA. 477 00:22:24,470 --> 00:22:25,470 GARY GENSLER: All right. 478 00:22:25,470 --> 00:22:28,740 So this sounds like it's about derivatives, executing 479 00:22:28,740 --> 00:22:29,730 a derivatives contract. 480 00:22:29,730 --> 00:22:33,610 But what if you're just buying 100 shares of Apple stock? 481 00:22:33,610 --> 00:22:36,435 What's it mean to execute a trade? 482 00:22:36,435 --> 00:22:38,310 AUDIENCE: To be on the phone or on the screen 483 00:22:38,310 --> 00:22:39,830 and agree on a price and a quantity. 484 00:22:39,830 --> 00:22:40,955 GARY GENSLER: There you go. 485 00:22:40,955 --> 00:22:43,970 It's just you bid a price. 486 00:22:43,970 --> 00:22:48,440 And somebody hits the bid or lifts the offer. 487 00:22:48,440 --> 00:22:50,480 Those are the terms when you're on the phone. 488 00:22:50,480 --> 00:22:52,070 I left your offer. 489 00:22:52,070 --> 00:22:53,970 I hit your bid. 490 00:22:53,970 --> 00:22:56,640 Or at least it was when I did it. 491 00:22:56,640 --> 00:22:57,420 Still? 492 00:22:57,420 --> 00:22:59,360 Or no, maybe not. 493 00:22:59,360 --> 00:23:00,590 What's that? 494 00:23:00,590 --> 00:23:01,220 Still. 495 00:23:01,220 --> 00:23:02,820 All right. 496 00:23:02,820 --> 00:23:03,320 Question? 497 00:23:03,320 --> 00:23:05,248 AUDIENCE: Just not on the phone anymore. 498 00:23:05,248 --> 00:23:06,212 [LAUGHTER] 499 00:23:06,212 --> 00:23:07,250 GARY GENSLER: Ah, yeah. 500 00:23:07,250 --> 00:23:08,560 That's it. 501 00:23:08,560 --> 00:23:09,260 What's that? 502 00:23:09,260 --> 00:23:10,640 AUDIENCE: You just hit a button. 503 00:23:10,640 --> 00:23:11,450 GARY GENSLER: Just hit a button. 504 00:23:11,450 --> 00:23:13,658 AUDIENCE: I worked on the [INAUDIBLE] desk at Goldman 505 00:23:13,658 --> 00:23:15,190 So we didn't talk to anyone. 506 00:23:15,190 --> 00:23:17,175 GARY GENSLER: You didn't talk to anybody? 507 00:23:17,175 --> 00:23:19,050 Can you go months without talking to anybody? 508 00:23:19,050 --> 00:23:20,570 AUDIENCE: No, we talked a little. 509 00:23:20,570 --> 00:23:20,990 GARY GENSLER: What's that? 510 00:23:20,990 --> 00:23:21,942 AUDIENCE: We talked a little bit. 511 00:23:21,942 --> 00:23:22,700 [? But it was always ?] [? kinda ?] weird. 512 00:23:22,700 --> 00:23:24,050 GARY GENSLER: All right. 513 00:23:24,050 --> 00:23:26,600 I guess I'm a dinosaur. 514 00:23:26,600 --> 00:23:27,360 Yeah. 515 00:23:27,360 --> 00:23:27,860 Yeah. 516 00:23:27,860 --> 00:23:31,280 The screens that I used to look at at Goldman Sachs, 517 00:23:31,280 --> 00:23:34,120 it was green. 518 00:23:34,120 --> 00:23:35,100 Yeah, yeah, yeah. 519 00:23:35,100 --> 00:23:36,800 You know. 520 00:23:36,800 --> 00:23:38,190 You know the screens. 521 00:23:38,190 --> 00:23:40,250 Kelly is thinking, what is he talking about? 522 00:23:40,250 --> 00:23:40,750 Right? 523 00:23:43,260 --> 00:23:44,610 Look it up on the internet. 524 00:23:44,610 --> 00:23:47,370 [? You're ?] [? fine. ?] [? You're ?] [? fine. ?] There 525 00:23:47,370 --> 00:23:50,280 wasn't any color on those screens. 526 00:23:50,280 --> 00:23:53,100 So execution is basically I buy, you 527 00:23:53,100 --> 00:23:58,070 sell, we agree upon a price, a price and a volume. 528 00:23:58,070 --> 00:24:00,620 That's really what execution is. 529 00:24:00,620 --> 00:24:04,170 But when you execute, you still need to do a few other things. 530 00:24:04,170 --> 00:24:05,892 And it's called clearing and settling. 531 00:24:05,892 --> 00:24:07,600 When I was on a trading floor, I couldn't 532 00:24:07,600 --> 00:24:09,142 have told you what the difference was 533 00:24:09,142 --> 00:24:10,520 between clearing and settling. 534 00:24:10,520 --> 00:24:14,070 That was something somebody else did, frankly. 535 00:24:14,070 --> 00:24:16,770 I'll admit it. 536 00:24:16,770 --> 00:24:18,440 They were really terrific professionals. 537 00:24:18,440 --> 00:24:19,460 But they had their job. 538 00:24:19,460 --> 00:24:21,110 I had a different job. 539 00:24:21,110 --> 00:24:24,380 Clearing is performing all of these functions. 540 00:24:24,380 --> 00:24:26,810 Authenticating somebody is there, 541 00:24:26,810 --> 00:24:28,860 making sure you actually have those securities, 542 00:24:28,860 --> 00:24:31,680 and importantly this thing called netting. 543 00:24:31,680 --> 00:24:32,780 How about settlement? 544 00:24:32,780 --> 00:24:33,620 Anybody want to say? 545 00:24:33,620 --> 00:24:34,828 I know it's written up there. 546 00:24:34,828 --> 00:24:36,680 But anybody want to just say what-- 547 00:24:36,680 --> 00:24:41,390 we've talked about ledgers for seven or 10 weeks now. 548 00:24:41,390 --> 00:24:43,730 What do you think a settlement really means 549 00:24:43,730 --> 00:24:47,190 in the context of a ledger? 550 00:24:47,190 --> 00:24:49,290 Did I see a hand up here? 551 00:24:49,290 --> 00:24:50,190 Yes, I did. 552 00:24:50,190 --> 00:24:51,240 I did see a hand up. 553 00:24:51,240 --> 00:24:51,990 What's settlement? 554 00:24:51,990 --> 00:24:53,190 AUDIENCE: Running the entries on the ledger. 555 00:24:53,190 --> 00:24:54,497 What's in, what's out. 556 00:24:54,497 --> 00:24:55,330 GARY GENSLER: Right. 557 00:24:55,330 --> 00:25:00,000 It's making a final recording on a ledger, 558 00:25:00,000 --> 00:25:02,280 finally moving something on a ledger. 559 00:25:02,280 --> 00:25:06,480 You could be moving the ownership of an asset 560 00:25:06,480 --> 00:25:08,940 or the ownership of cash. 561 00:25:08,940 --> 00:25:10,840 And when you do both at the same time, 562 00:25:10,840 --> 00:25:13,130 it's called delivery versus payment, or DVP. 563 00:25:16,310 --> 00:25:21,490 It's just a final change of data record on a ledger. 564 00:25:21,490 --> 00:25:22,630 In the old days-- 565 00:25:22,630 --> 00:25:25,010 I'm talking about before I was doing it. 566 00:25:25,010 --> 00:25:27,970 But 40 years to hundreds of years 567 00:25:27,970 --> 00:25:30,240 ago, it was physically delivering 568 00:25:30,240 --> 00:25:34,840 a physical piece of paper which was a equity security or a bond 569 00:25:34,840 --> 00:25:36,100 security. 570 00:25:36,100 --> 00:25:39,850 But it all got dematerialized starting in the 1970s. 571 00:25:39,850 --> 00:25:41,950 And by the 1990s, it was pretty much-- 572 00:25:41,950 --> 00:25:44,260 in this country and the UK-- 573 00:25:44,260 --> 00:25:48,400 it was pretty much dematerialized securities, 574 00:25:48,400 --> 00:25:52,960 that everything's digitized now in the equity markets in nearly 575 00:25:52,960 --> 00:25:56,290 every country that you're all from and maybe 576 00:25:56,290 --> 00:25:59,200 with some small exceptions. 577 00:25:59,200 --> 00:26:03,240 So execution is I buy, you sell, volume price. 578 00:26:03,240 --> 00:26:06,760 We've legally agreed to something verbally. 579 00:26:06,760 --> 00:26:10,810 And billions of dollars can trade on the phone 580 00:26:10,810 --> 00:26:13,330 or in the computer. 581 00:26:13,330 --> 00:26:14,650 Legally, it's executed. 582 00:26:14,650 --> 00:26:17,620 Clearing is trying to make sure that you actually 583 00:26:17,620 --> 00:26:20,780 have the securities, and I actually have the cash. 584 00:26:20,780 --> 00:26:21,670 We do some netting. 585 00:26:21,670 --> 00:26:23,320 We authenticate. 586 00:26:23,320 --> 00:26:26,680 And settlement is actually moving the data on the ledgers. 587 00:26:26,680 --> 00:26:29,020 Three steps. 588 00:26:29,020 --> 00:26:30,460 This is the benefit of netting. 589 00:26:30,460 --> 00:26:32,750 It was just one chart that I could find somewhere. 590 00:26:32,750 --> 00:26:34,690 It's actually the Australian Stock Exchange 591 00:26:34,690 --> 00:26:37,540 had it on their website. 592 00:26:37,540 --> 00:26:40,430 But if you have multi-parties in the transaction, 593 00:26:40,430 --> 00:26:42,400 which was on the left hand side-- 594 00:26:42,400 --> 00:26:45,460 and these are just arrows of all the different transactions 595 00:26:45,460 --> 00:26:48,850 that could occur in one little example-- 596 00:26:48,850 --> 00:26:56,310 and instead you interpose a central intermediary, 597 00:26:56,310 --> 00:26:58,840 you can do a lot of netting. 598 00:26:58,840 --> 00:27:01,270 And look at participant C. Participant C 599 00:27:01,270 --> 00:27:05,890 only has an exposure of selling 35 of whatever they're 600 00:27:05,890 --> 00:27:10,930 selling, some stock, versus all these other transactions. 601 00:27:10,930 --> 00:27:14,610 And this is a very simple case. 602 00:27:14,610 --> 00:27:20,940 But if you have 20 or a hundred clearing members that are all 603 00:27:20,940 --> 00:27:24,300 big brokerage houses and they're doing thousands or even 604 00:27:24,300 --> 00:27:28,550 hundreds of thousands of trades a day, 605 00:27:28,550 --> 00:27:32,690 you have an awful lot of transactions. 606 00:27:32,690 --> 00:27:35,750 And the gross transactions versus the net 607 00:27:35,750 --> 00:27:37,560 is quite a difference. 608 00:27:37,560 --> 00:27:41,660 That netting can bring down numbers 90 plus percent 609 00:27:41,660 --> 00:27:45,410 or maybe even 99% or more. 610 00:27:45,410 --> 00:27:47,780 So central intermediaries-- 611 00:27:47,780 --> 00:27:51,060 I know this goes against Satoshi Nakamoto's whole concept-- 612 00:27:51,060 --> 00:27:57,370 but central intermediaries came about in the securities market 613 00:27:57,370 --> 00:27:58,580 for a bunch of reasons. 614 00:27:58,580 --> 00:28:03,890 But one of the biggest ones was netting. 615 00:28:03,890 --> 00:28:06,140 Brotish? 616 00:28:06,140 --> 00:28:07,990 And it's a really important concept 617 00:28:07,990 --> 00:28:09,640 in the back office of Wall Street. 618 00:28:09,640 --> 00:28:11,473 AUDIENCE: I'm not familiar with the concept. 619 00:28:11,473 --> 00:28:12,370 So [INAUDIBLE]. 620 00:28:12,370 --> 00:28:13,540 GARY GENSLER: You're saying you're familiar? 621 00:28:13,540 --> 00:28:15,000 AUDIENCE: I'm not familiar with this concept, 622 00:28:15,000 --> 00:28:16,460 so a clarification question. 623 00:28:16,460 --> 00:28:17,657 GARY GENSLER: Please. 624 00:28:17,657 --> 00:28:19,240 AUDIENCE: When we do netting, it seems 625 00:28:19,240 --> 00:28:21,250 that we are losing a lot of information 626 00:28:21,250 --> 00:28:22,630 here from the gross transactions. 627 00:28:22,630 --> 00:28:23,680 So is it-- 628 00:28:23,680 --> 00:28:24,100 GARY GENSLER: That's right. 629 00:28:24,100 --> 00:28:25,990 AUDIENCE: --just the transfer of the asset 630 00:28:25,990 --> 00:28:27,940 is done through the netting, but the records 631 00:28:27,940 --> 00:28:30,040 of all the gross transactions are still there 632 00:28:30,040 --> 00:28:31,450 with the central intermediaries? 633 00:28:31,450 --> 00:28:32,825 GARY GENSLER: Very good question. 634 00:28:32,825 --> 00:28:35,230 So are we losing information? 635 00:28:35,230 --> 00:28:41,110 Or are we just lessening the lines of counterparty risk 636 00:28:41,110 --> 00:28:42,400 in the transactions? 637 00:28:42,400 --> 00:28:44,290 And the answer is, if done properly, 638 00:28:44,290 --> 00:28:48,880 you still keep the information of all of these. 639 00:28:48,880 --> 00:28:53,880 And in the real world, this is just eight transactions. 640 00:28:53,880 --> 00:28:55,910 It's netted down to four. 641 00:28:55,910 --> 00:28:56,410 Right? 642 00:28:56,410 --> 00:28:58,720 So this is eight transactions netted down to four. 643 00:28:58,720 --> 00:29:02,110 Usually this is millions of transactions netted down 644 00:29:02,110 --> 00:29:06,420 to tens of thousands or maybe millions of transactions 645 00:29:06,420 --> 00:29:11,010 literally netted down to, if there was 100 clearing members, 646 00:29:11,010 --> 00:29:13,200 it would be 100 transactions, because they'd 647 00:29:13,200 --> 00:29:21,750 each have one net transaction per security, per security. 648 00:29:21,750 --> 00:29:24,000 And so you need to keep the information 649 00:29:24,000 --> 00:29:26,430 of the millions of transactions to properly 650 00:29:26,430 --> 00:29:29,880 record for your customers, for your taxes, 651 00:29:29,880 --> 00:29:33,210 for all sorts of other reasons. 652 00:29:33,210 --> 00:29:36,270 But you've netted down for counterparty reasons. 653 00:29:36,270 --> 00:29:39,270 Does that help? 654 00:29:39,270 --> 00:29:40,390 Question over here. 655 00:29:40,390 --> 00:29:40,890 Sean. 656 00:29:40,890 --> 00:29:43,145 AUDIENCE: I was just curious. 657 00:29:43,145 --> 00:29:45,220 It's very understandable for equities. 658 00:29:45,220 --> 00:29:47,796 But for principal commodities, [? the full ?] 659 00:29:47,796 --> 00:29:52,060 commodity derivatives, for instance in the clearing space, 660 00:29:52,060 --> 00:29:54,540 does it actually involve the physical delivery 661 00:29:54,540 --> 00:29:59,700 of the netted commodity [INAUDIBLE]?? 662 00:29:59,700 --> 00:30:02,242 Because there's a different type for delivery, and how that-- 663 00:30:02,242 --> 00:30:02,950 GARY GENSLER: OK. 664 00:30:02,950 --> 00:30:04,590 So Sean's asked the question, what 665 00:30:04,590 --> 00:30:06,660 if you're not dealing with digitized assets, 666 00:30:06,660 --> 00:30:07,990 but physical assets? 667 00:30:07,990 --> 00:30:08,490 I think. 668 00:30:08,490 --> 00:30:09,780 You used the word commodity. 669 00:30:09,780 --> 00:30:11,865 But I was presuming you meant physical. 670 00:30:14,760 --> 00:30:17,370 So some of the innovations of the 19th century 671 00:30:17,370 --> 00:30:19,740 relate here when they started to have 672 00:30:19,740 --> 00:30:22,770 standardized contracts for commodities 673 00:30:22,770 --> 00:30:26,610 that were held in warehouses. 674 00:30:26,610 --> 00:30:31,550 In essence, to facilitate some of this netting 675 00:30:31,550 --> 00:30:34,730 was also why central warehouses, this clearing 676 00:30:34,730 --> 00:30:36,890 house in the middle in physical commodities, 677 00:30:36,890 --> 00:30:40,310 could be a warehouse that held the gold, held the silver, 678 00:30:40,310 --> 00:30:42,830 held the corn or wheat. 679 00:30:42,830 --> 00:30:45,560 And if they're holding that corn or wheat, 680 00:30:45,560 --> 00:30:48,230 they started to have specifications as to what 681 00:30:48,230 --> 00:30:50,390 standards there were. 682 00:30:50,390 --> 00:30:52,850 In gold and silver, it was percent. 683 00:30:52,850 --> 00:30:59,065 It was called the, let's see, assay-- 684 00:30:59,065 --> 00:31:00,440 I'm trying to remember the word-- 685 00:31:00,440 --> 00:31:04,100 but what percent gold or silver it is. 686 00:31:04,100 --> 00:31:07,070 But in corn and wheat, there was a more detailed 687 00:31:07,070 --> 00:31:12,680 what's going into the warehouse to commoditize it so that it 688 00:31:12,680 --> 00:31:14,990 was a standard contract. 689 00:31:14,990 --> 00:31:20,230 If it's truly physical and not commodity, it's much harder. 690 00:31:20,230 --> 00:31:26,290 Then you can't really net down my house versus Shimon's house. 691 00:31:26,290 --> 00:31:28,905 They're two different houses. 692 00:31:28,905 --> 00:31:31,460 Does that help? 693 00:31:31,460 --> 00:31:33,870 But literally by the late 19th century, 694 00:31:33,870 --> 00:31:37,200 even in the commodities space running out of Chicago, 695 00:31:37,200 --> 00:31:39,140 they figured out central clearing houses 696 00:31:39,140 --> 00:31:43,610 with a lot of counterparty risk. 697 00:31:43,610 --> 00:31:45,560 So there's an economic reason we have 698 00:31:45,560 --> 00:31:46,970 centralized clearing houses. 699 00:31:46,970 --> 00:31:51,980 And the biggest economic reason is netting. 700 00:31:51,980 --> 00:31:56,330 And netting lowers something called counterparty risk. 701 00:31:56,330 --> 00:32:00,530 Instead of having literally millions of transactions, 702 00:32:00,530 --> 00:32:02,660 you'll have thousands. 703 00:32:02,660 --> 00:32:04,398 And you'll have less counterparty risk. 704 00:32:04,398 --> 00:32:05,690 It also lowers one other thing. 705 00:32:05,690 --> 00:32:07,130 So it lowers counterparty risk. 706 00:32:07,130 --> 00:32:09,130 And what is the other thing you think it lowers? 707 00:32:09,130 --> 00:32:11,223 Anybody? 708 00:32:11,223 --> 00:32:12,390 AUDIENCE: Transactions cost. 709 00:32:12,390 --> 00:32:14,682 GARY GENSLER: Transactions cost because it's efficient. 710 00:32:17,280 --> 00:32:20,850 We used to call it you take a lot of lines off the page. 711 00:32:20,850 --> 00:32:23,520 So in the old human days, you'd have 712 00:32:23,520 --> 00:32:28,390 to have fewer humans actually pushing paper. 713 00:32:28,390 --> 00:32:29,650 Hugo. 714 00:32:29,650 --> 00:32:32,230 AUDIENCE: Another clarification question. 715 00:32:32,230 --> 00:32:35,650 So from the participants side, so from a hedge fund side 716 00:32:35,650 --> 00:32:39,615 or whatever or a trading desk side, 717 00:32:39,615 --> 00:32:41,740 do they just go about their day trading everything? 718 00:32:41,740 --> 00:32:44,183 And then the next day the clearing house 719 00:32:44,183 --> 00:32:46,600 tells them, OK, you made all these transactions yesterday. 720 00:32:46,600 --> 00:32:49,420 But we really only need 35 [INAUDIBLE].. 721 00:32:49,420 --> 00:32:51,790 Or how does it [INAUDIBLE]? 722 00:32:51,790 --> 00:32:56,110 GARY GENSLER: In that case, looking back to this, 723 00:32:56,110 --> 00:32:57,760 you are either the buyer or seller. 724 00:32:57,760 --> 00:32:59,200 And you're dealing with a broker. 725 00:32:59,200 --> 00:33:03,713 So if you're a hedge fund, you often 726 00:33:03,713 --> 00:33:05,380 enter into an arrangement with a broker. 727 00:33:05,380 --> 00:33:07,640 And it's called a prime brokerage arrangement. 728 00:33:07,640 --> 00:33:09,640 And you might have a prime brokerage arrangement 729 00:33:09,640 --> 00:33:11,680 with more than one Wall Street firm. 730 00:33:11,680 --> 00:33:14,580 But you rarely would have it with more than three. 731 00:33:14,580 --> 00:33:17,130 So even if I'm doing a trade through Merrill Lynch, 732 00:33:17,130 --> 00:33:18,900 if my prime brokerage arrangement is 733 00:33:18,900 --> 00:33:25,420 at Goldman Sachs, Merrill is going to get a commission. 734 00:33:25,420 --> 00:33:27,970 But they're going to give up my prime brokerage arrangements 735 00:33:27,970 --> 00:33:28,870 to Goldman Sachs. 736 00:33:28,870 --> 00:33:32,950 And that helps me, me the hedge fund, get more netting as well. 737 00:33:32,950 --> 00:33:36,100 And that's a development in the last 20 years, 738 00:33:36,100 --> 00:33:39,670 this concept of prime brokerage. 739 00:33:39,670 --> 00:33:41,260 But to answer your question, yes. 740 00:33:41,260 --> 00:33:42,370 I go about my business. 741 00:33:42,370 --> 00:33:43,560 I trade all day. 742 00:33:43,560 --> 00:33:48,130 And unless I get a phone call or an electronic communication 743 00:33:48,130 --> 00:33:51,870 from my prime brokerage saying you've hit your lines, 744 00:33:51,870 --> 00:33:56,880 you do have credit limits and position limits in essence. 745 00:33:56,880 --> 00:33:58,650 But your prime brokerage-- 746 00:33:58,650 --> 00:34:01,650 and I might say, well, I need something moved over. 747 00:34:01,650 --> 00:34:05,220 I need some more credit today. 748 00:34:05,220 --> 00:34:05,862 But yes. 749 00:34:05,862 --> 00:34:08,070 You go about your business until your prime brokerage 750 00:34:08,070 --> 00:34:11,020 says slow down. 751 00:34:11,020 --> 00:34:11,520 Shimon? 752 00:34:11,520 --> 00:34:13,690 AUDIENCE: But [INAUDIBLE] question says that 753 00:34:13,690 --> 00:34:16,580 if you're a lot like LTCM-- 754 00:34:16,580 --> 00:34:18,372 GARY GENSLER: Long-Term Capital Management. 755 00:34:18,372 --> 00:34:20,622 AUDIENCE: --at some point, they were running a balance 756 00:34:20,622 --> 00:34:22,230 sheet of a trillion dollars. 757 00:34:22,230 --> 00:34:24,409 And part of the reason why, when bankrupt is they 758 00:34:24,409 --> 00:34:27,145 didn't want to use one prime broker, 759 00:34:27,145 --> 00:34:29,270 because they don't want to reveal all their trades. 760 00:34:29,270 --> 00:34:31,389 But then their prime brokers could not 761 00:34:31,389 --> 00:34:35,570 net out their positions, because I may hold one leg with you 762 00:34:35,570 --> 00:34:38,739 and another leg with you. 763 00:34:38,739 --> 00:34:41,120 The net may be zero. 764 00:34:41,120 --> 00:34:43,429 But there's no way for you to actually reconcile 765 00:34:43,429 --> 00:34:46,670 that or reconcile that risk. 766 00:34:46,670 --> 00:34:48,610 So it's a real problem. 767 00:34:48,610 --> 00:34:50,360 GARY GENSLER: Long-Term Capital Management 768 00:34:50,360 --> 00:34:55,250 was a hedge fund operating in Connecticut. 769 00:34:55,250 --> 00:34:57,590 It was founded in the early 1990s 770 00:34:57,590 --> 00:34:59,790 by a man named John Meriwether, who 771 00:34:59,790 --> 00:35:03,610 is a very famous trader from Salomon Brothers 772 00:35:03,610 --> 00:35:05,300 and very successful. 773 00:35:05,300 --> 00:35:10,460 And then he started it with some remarkable finance 774 00:35:10,460 --> 00:35:14,060 faculty from here, at MIT, and elsewhere-- 775 00:35:14,060 --> 00:35:17,000 Bob Merton, who is a Nobel laureate. 776 00:35:17,000 --> 00:35:20,220 Brilliant, brilliant team of folks. 777 00:35:20,220 --> 00:35:22,940 But they ran into a little challenge four years later. 778 00:35:22,940 --> 00:35:26,750 They had about $125 billion balance sheet, $4 billion 779 00:35:26,750 --> 00:35:30,800 in capital, and 1.2 trillion in swaps. 780 00:35:30,800 --> 00:35:32,000 Why do I know the numbers? 781 00:35:32,000 --> 00:35:34,880 I got the phone call with my one-year-old sitting on my lap. 782 00:35:34,880 --> 00:35:36,620 And I picked up the phone. 783 00:35:36,620 --> 00:35:38,120 And I said, hello? 784 00:35:38,120 --> 00:35:40,790 And the voice on the other line said, Treasury operator. 785 00:35:40,790 --> 00:35:42,830 I've got the Treasury secretary for you. 786 00:35:42,830 --> 00:35:43,820 And I get on the phone. 787 00:35:43,820 --> 00:35:44,930 I go, yes, Bob? 788 00:35:44,930 --> 00:35:45,960 It was Bob Rubin. 789 00:35:45,960 --> 00:35:47,210 It was the Treasury secretary. 790 00:35:47,210 --> 00:35:47,950 Bob, what's up? 791 00:35:47,950 --> 00:35:48,850 It was a Saturday. 792 00:35:48,850 --> 00:35:50,980 Isabel was on my lap. 793 00:35:50,980 --> 00:35:53,590 He says, do you know anything about Long-Term Capital 794 00:35:53,590 --> 00:35:54,180 Management? 795 00:35:54,180 --> 00:35:56,050 He goes through some of the figures. 796 00:35:56,050 --> 00:35:56,800 I'll never forget. 797 00:35:56,800 --> 00:35:59,020 He says, I hear from Alan-- 798 00:35:59,020 --> 00:36:02,090 that would be Alan Greenspan, who was the head of the Federal 799 00:36:02,090 --> 00:36:02,620 Reserve-- 800 00:36:02,620 --> 00:36:05,005 I hear from Alan that they have over a trillion dollars 801 00:36:05,005 --> 00:36:06,710 of derivatives. 802 00:36:06,710 --> 00:36:08,470 That's Shimon's number. 803 00:36:08,470 --> 00:36:10,357 He says, is that a big number? 804 00:36:10,357 --> 00:36:12,440 Well, you have to understand, Bob's sense of humor 805 00:36:12,440 --> 00:36:17,410 was quite-- at that moment, he knew it was a big number. 806 00:36:17,410 --> 00:36:19,990 So I flew up the next day. 807 00:36:19,990 --> 00:36:22,630 And with somebody from the Federal Reserve 808 00:36:22,630 --> 00:36:25,660 Bank in New York, Peter Fisher, we 809 00:36:25,660 --> 00:36:27,880 sat around the conference table looking 810 00:36:27,880 --> 00:36:30,040 at these facts and figures. 811 00:36:30,040 --> 00:36:33,850 And then I got back on the plane and flew back to Washington 812 00:36:33,850 --> 00:36:34,840 to be with my family. 813 00:36:34,840 --> 00:36:37,510 The next day was actually Rosh Hashanah, the Jewish holiday. 814 00:36:37,510 --> 00:36:40,230 And I wasn't going to work. 815 00:36:40,230 --> 00:36:43,910 And I had to call Bob and his deputy, Larry Summers, 816 00:36:43,910 --> 00:36:46,160 up and give them my feedback. 817 00:36:46,160 --> 00:36:49,040 And I said, this thing's going down. 818 00:36:49,040 --> 00:36:52,100 Bear Stearns was their prime brokerage. 819 00:36:52,100 --> 00:36:55,010 And Bear Stearns, who 20 years later failed themselves, 820 00:36:55,010 --> 00:36:58,730 but Bear Stearns said, we're pulling the prime brokerage 821 00:36:58,730 --> 00:37:01,460 account if you can't post another half 822 00:37:01,460 --> 00:37:04,015 a billion dollars by Wednesday. 823 00:37:04,015 --> 00:37:05,390 Well, there was no way they could 824 00:37:05,390 --> 00:37:07,010 post another half a billion dollars, 825 00:37:07,010 --> 00:37:08,840 because their $4 billion of capital 826 00:37:08,840 --> 00:37:11,210 we could best figure on that Sunday was somewhere 827 00:37:11,210 --> 00:37:16,427 covering around 400 or 500 million at that point in time. 828 00:37:16,427 --> 00:37:18,010 Well, when you have a lot of leverage, 829 00:37:18,010 --> 00:37:20,860 you have $1.2 trillion of derivatives which are all 830 00:37:20,860 --> 00:37:22,690 sort of leveraged in one way or another 831 00:37:22,690 --> 00:37:25,660 and you've got $125 billion balance sheet, even 832 00:37:25,660 --> 00:37:29,870 with the best minds, the best finance faculty-- 833 00:37:29,870 --> 00:37:32,160 and they really are terrific individuals. 834 00:37:32,160 --> 00:37:35,440 And Meriwether was quite a trader. 835 00:37:35,440 --> 00:37:38,390 Sometimes markets go against you. 836 00:37:38,390 --> 00:37:42,210 Correlation risk, liquidity risk, and so forth. 837 00:37:42,210 --> 00:37:47,090 So they pulled the plug the next day, basically. 838 00:37:47,090 --> 00:37:49,850 That's the Long-Term Capital Management story I remember. 839 00:37:49,850 --> 00:37:54,140 But it's just because I kind of had to live it. 840 00:37:54,140 --> 00:37:57,620 But to answer your question, you can trade all day 841 00:37:57,620 --> 00:38:02,550 until your prime broker says, Hugo, slow down, 842 00:38:02,550 --> 00:38:04,980 because they're extending credit to you-- 843 00:38:04,980 --> 00:38:08,250 the prime broker is-- 844 00:38:08,250 --> 00:38:09,510 intraday. 845 00:38:09,510 --> 00:38:11,520 And one of the big risks, one of the big risks 846 00:38:11,520 --> 00:38:13,300 in the financial crisis-- 847 00:38:13,300 --> 00:38:16,710 and certainly Ben Bernanke, who is chairing the US Federal 848 00:38:16,710 --> 00:38:20,190 Reserve and others had to deal with it in the middle of 2008-- 849 00:38:20,190 --> 00:38:23,800 is most commercial banks, so they don't have a broker. 850 00:38:23,800 --> 00:38:24,700 They're banks. 851 00:38:24,700 --> 00:38:27,690 They're the brokers themselves-- were extending 852 00:38:27,690 --> 00:38:30,150 significant amount of intraday credit 853 00:38:30,150 --> 00:38:36,980 to each other that, in 2008, wasn't terribly well measured 854 00:38:36,980 --> 00:38:40,460 in the repo markets and in other markets. 855 00:38:40,460 --> 00:38:44,400 And that's changed subsequently. 856 00:38:44,400 --> 00:38:46,250 But you wouldn't have been a big bank. 857 00:38:46,250 --> 00:38:49,130 You would have probably been shut down. 858 00:38:49,130 --> 00:38:50,360 Sorry. 859 00:38:50,360 --> 00:38:52,160 Or maybe not. 860 00:38:52,160 --> 00:38:57,020 So netting is the biggest benefit of clearing. 861 00:38:57,020 --> 00:39:00,770 But it's also efficiency and taking down 862 00:39:00,770 --> 00:39:04,320 a lot of operating lines of code. 863 00:39:04,320 --> 00:39:05,000 This is a chart. 864 00:39:05,000 --> 00:39:06,420 I'm not going to spend a lot of time on it. 865 00:39:06,420 --> 00:39:08,550 But this is right off the Australian Stock Exchange 866 00:39:08,550 --> 00:39:09,050 website. 867 00:39:09,050 --> 00:39:10,970 And I thought, well, why not take that? 868 00:39:10,970 --> 00:39:15,040 They have 2 million investors, the market participants 869 00:39:15,040 --> 00:39:18,160 or brokers in a much smaller market than the US. 870 00:39:18,160 --> 00:39:24,340 And Australia has 77 brokers, registered, regulated brokers. 871 00:39:24,340 --> 00:39:28,840 And then they have the clearing house called CHESS, C-H-E-S-S, 872 00:39:28,840 --> 00:39:31,090 which I can't quite remember what it all stands for. 873 00:39:31,090 --> 00:39:33,490 But SS is Settlement System. 874 00:39:33,490 --> 00:39:35,170 Clearing house-- 875 00:39:35,170 --> 00:39:36,940 AUDIENCE: [INAUDIBLE] 876 00:39:36,940 --> 00:39:38,900 GARY GENSLER: What's that? 877 00:39:38,900 --> 00:39:39,660 What's it? 878 00:39:39,660 --> 00:39:41,942 AUDIENCE: Clearing House Subregistered System. 879 00:39:41,942 --> 00:39:43,400 GARY GENSLER: Subregistered System. 880 00:39:43,400 --> 00:39:44,300 Thank you. 881 00:39:44,300 --> 00:39:45,890 Is it Equity for E? 882 00:39:45,890 --> 00:39:46,890 AUDIENCE: There is none. 883 00:39:46,890 --> 00:39:48,600 GARY GENSLER: There is none. 884 00:39:48,600 --> 00:39:51,910 But they have the Australian Stock Exchange Clear 885 00:39:51,910 --> 00:39:55,080 and Australia Stock Exchange Settlement, 886 00:39:55,080 --> 00:39:58,050 because they actually put it in two different places, two 887 00:39:58,050 --> 00:39:59,400 different things. 888 00:39:59,400 --> 00:40:02,100 The clearinghouse is doing what we talked about-- 889 00:40:02,100 --> 00:40:06,850 mostly netting, lowering the lines of code, 890 00:40:06,850 --> 00:40:08,320 and then settlement. 891 00:40:08,320 --> 00:40:12,700 And basically, because we're in a trade day plus two days, 892 00:40:12,700 --> 00:40:16,440 on the first day is all the clearing and the netting. 893 00:40:16,440 --> 00:40:18,490 And you better show up with your securities. 894 00:40:18,490 --> 00:40:21,160 And on the second day is settlement. 895 00:40:21,160 --> 00:40:24,820 Electronics and technology could make this simultaneous. 896 00:40:28,140 --> 00:40:30,560 But most market participants like 897 00:40:30,560 --> 00:40:34,050 to have some ability to do gross trading against net's 898 00:40:34,050 --> 00:40:35,640 positions. 899 00:40:35,640 --> 00:40:39,030 So there's a lot of economics and market practice 900 00:40:39,030 --> 00:40:43,510 and structure behind keeping some delayed settlement. 901 00:40:43,510 --> 00:40:45,260 And the markets have grown up around that. 902 00:40:45,260 --> 00:40:46,500 Ross. 903 00:40:46,500 --> 00:40:50,555 AUDIENCE: If it's not delayed, if it was instantaneous, 904 00:40:50,555 --> 00:40:52,430 wouldn't that get rid of the benefits of net? 905 00:40:52,430 --> 00:40:54,940 And you'd have to have every trade go right through. 906 00:40:54,940 --> 00:40:57,240 So it has to be that you wait until the end of the day 907 00:40:57,240 --> 00:40:58,948 or whatever period [INTERPOSING VOICES].. 908 00:40:58,948 --> 00:41:01,860 GARY GENSLER: So Ross just raised the number one reason, 909 00:41:01,860 --> 00:41:08,070 economic reason, why simultaneous clearing 910 00:41:08,070 --> 00:41:13,650 and settling is not that likely to happen. 911 00:41:13,650 --> 00:41:15,660 The markets have built themselves 912 00:41:15,660 --> 00:41:20,400 around the benefits of netting and also 913 00:41:20,400 --> 00:41:25,390 stock loan, the ability to borrow that security somewhere. 914 00:41:25,390 --> 00:41:27,220 But you're absolutely right. 915 00:41:27,220 --> 00:41:31,870 If there were simultaneous execution and settlement, which 916 00:41:31,870 --> 00:41:34,840 is what the Bitcoin network in essence is supposed to be 917 00:41:34,840 --> 00:41:38,973 and blockchain networks can be, then you, in essence, 918 00:41:38,973 --> 00:41:39,640 have to prefund. 919 00:41:42,370 --> 00:41:44,510 Every time Hugo trades, he's going 920 00:41:44,510 --> 00:41:46,630 to have to provide his security to trade. 921 00:41:49,560 --> 00:41:53,790 But then Tom would get the security faster. 922 00:41:53,790 --> 00:41:57,790 It would change the entire prime brokerage model. 923 00:41:57,790 --> 00:42:00,190 It doesn't mean markets, Ross, couldn't adapt 924 00:42:00,190 --> 00:42:02,860 and be structured that way. 925 00:42:02,860 --> 00:42:05,230 But it would be fundamentally a different structure. 926 00:42:05,230 --> 00:42:08,280 Is this is a hand up? 927 00:42:08,280 --> 00:42:11,670 AUDIENCE: So the [INAUDIBLE] is really big with derivatives. 928 00:42:11,670 --> 00:42:13,180 I don't see it. 929 00:42:13,180 --> 00:42:14,290 I don't see what we use-- 930 00:42:14,290 --> 00:42:17,400 I don't see it being a primary issue with equities. 931 00:42:17,400 --> 00:42:20,140 So if you were going to apply this to equities, trading, 932 00:42:20,140 --> 00:42:23,890 any cash instrument basically, I don't 933 00:42:23,890 --> 00:42:27,520 see the huge economic value for netting. 934 00:42:27,520 --> 00:42:32,230 GARY GENSLER: The huge economic benefit for netting currently 935 00:42:32,230 --> 00:42:36,095 is a lot of operational efficiency. 936 00:42:36,095 --> 00:42:36,720 AUDIENCE: Sure. 937 00:42:36,720 --> 00:42:39,470 But if you're going to place that with some-- 938 00:42:39,470 --> 00:42:40,180 GARY GENSLER: No. 939 00:42:40,180 --> 00:42:45,520 And then also, because there is two day counterparty risk, 940 00:42:45,520 --> 00:42:48,880 you are lowering that two day whatever 941 00:42:48,880 --> 00:42:55,460 measure equity volatility, two day equity volatility risk, 942 00:42:55,460 --> 00:42:57,210 usually not collateralized. 943 00:43:01,580 --> 00:43:05,120 AUDIENCE: I'm just commenting on whether this-- 944 00:43:05,120 --> 00:43:12,240 you said the price that you pay for replacing this system with 945 00:43:12,240 --> 00:43:16,088 the blockchain based one would be that you cannot net up. 946 00:43:16,088 --> 00:43:18,380 GARY GENSLER: I should have said something in addition. 947 00:43:18,380 --> 00:43:20,270 And the market participants would have 948 00:43:20,270 --> 00:43:23,090 to then prefund their trades. 949 00:43:23,090 --> 00:43:26,120 I think the larger thing-- 950 00:43:26,120 --> 00:43:29,580 so if I speak in too much financial technology, 951 00:43:29,580 --> 00:43:33,930 let me say current markets are you can sell or buy 952 00:43:33,930 --> 00:43:39,910 a security without owning the security or having the cash. 953 00:43:39,910 --> 00:43:42,420 You can buy something and not have the cash. 954 00:43:42,420 --> 00:43:45,480 You have to deliver the cash in two days. 955 00:43:45,480 --> 00:43:47,700 And usually by that evening or the next day, 956 00:43:47,700 --> 00:43:50,580 you have to show you have the cash. 957 00:43:50,580 --> 00:43:53,520 You can sell a security without owning the security. 958 00:43:53,520 --> 00:43:55,320 And you only have to really kind of deliver 959 00:43:55,320 --> 00:44:00,910 it, generally speaking, the next day, on t plus 1. 960 00:44:00,910 --> 00:44:04,690 If you went to a true simultaneous clearing settling 961 00:44:04,690 --> 00:44:07,270 and execution market, you'd have to have 962 00:44:07,270 --> 00:44:12,140 the cash and the security before you execute. 963 00:44:12,140 --> 00:44:14,380 And I think that's the larger piece of it, 964 00:44:14,380 --> 00:44:17,657 much larger than netting. 965 00:44:17,657 --> 00:44:18,990 AUDIENCE: I understand the cash. 966 00:44:18,990 --> 00:44:20,580 But you need the security, right? 967 00:44:20,580 --> 00:44:22,620 Because if you're going to buy it-- 968 00:44:22,620 --> 00:44:25,380 if you sell it and you don't have it, 969 00:44:25,380 --> 00:44:30,140 then you know it's going to take two days to get one 970 00:44:30,140 --> 00:44:32,262 to cover your first. 971 00:44:32,262 --> 00:44:34,720 GARY GENSLER: So through your prime brokerage arrangement-- 972 00:44:34,720 --> 00:44:36,310 Hugo, remember, has a couple. 973 00:44:36,310 --> 00:44:38,235 He's a big hedge fund. 974 00:44:38,235 --> 00:44:40,360 And he has what's called a prime brokerage account. 975 00:44:40,360 --> 00:44:43,570 Not at Bear Stearns any longer, but maybe 976 00:44:43,570 --> 00:44:44,750 it's at Goldman Sachs. 977 00:44:44,750 --> 00:44:46,420 He's kind of fancy. 978 00:44:46,420 --> 00:44:49,370 He's getting that. 979 00:44:49,370 --> 00:44:51,220 Then his prime brokerage account, 980 00:44:51,220 --> 00:44:53,620 in his legal contractual arrangements with them, 981 00:44:53,620 --> 00:44:58,440 they will find that security for him, maybe 982 00:44:58,440 --> 00:45:00,060 from some other customer. 983 00:45:00,060 --> 00:45:02,550 Maybe that security is held in somebody else's 984 00:45:02,550 --> 00:45:04,410 what's called a margin account. 985 00:45:04,410 --> 00:45:08,520 And they will enter into what's called a stock borrow. 986 00:45:08,520 --> 00:45:11,640 But when he executes on the phone, 987 00:45:11,640 --> 00:45:14,670 he doesn't technically already have that security. 988 00:45:14,670 --> 00:45:16,540 He could be shorting it. 989 00:45:16,540 --> 00:45:20,190 And his prime brokerage account could then go find it 990 00:45:20,190 --> 00:45:23,550 from the Goldman Sachs network. 991 00:45:23,550 --> 00:45:25,820 AUDIENCE: But it's not because you have plus two. 992 00:45:25,820 --> 00:45:27,450 But it's because you have a broker. 993 00:45:27,450 --> 00:45:28,200 GARY GENSLER: Yes. 994 00:45:28,200 --> 00:45:31,133 That's right, combined with t plus 2. 995 00:45:31,133 --> 00:45:31,800 Let's take this. 996 00:45:31,800 --> 00:45:33,060 And then I'm going to move on. 997 00:45:33,060 --> 00:45:35,102 AUDIENCE: Can you just take that one step further 998 00:45:35,102 --> 00:45:37,770 and explain why that's beneficial broadly 999 00:45:37,770 --> 00:45:39,950 versus just the system we've grown accustomed to? 1000 00:45:39,950 --> 00:45:41,700 GARY GENSLER: It's a really good question. 1001 00:45:41,700 --> 00:45:43,783 Is this just the system we've grown accustomed to? 1002 00:45:43,783 --> 00:45:45,120 Or is it beneficial? 1003 00:45:45,120 --> 00:45:46,020 It's hard to tell. 1004 00:45:46,020 --> 00:45:47,550 It is definitely the system we've 1005 00:45:47,550 --> 00:45:50,700 grown accustomed to over a couple of centuries. 1006 00:45:50,700 --> 00:45:55,870 And as we've gone from transaction date plus five, 1007 00:45:55,870 --> 00:45:58,500 which most markets were I'm talking 1008 00:45:58,500 --> 00:46:01,200 about 50 and 100 years ago, maybe it 1009 00:46:01,200 --> 00:46:03,030 was even t plus 10 at some point. 1010 00:46:03,030 --> 00:46:07,850 But t plus five to t plus three to now t plus two, 1011 00:46:07,850 --> 00:46:09,240 we've started to challenge that. 1012 00:46:09,240 --> 00:46:12,090 But in the US, the Depository Trust Corporation 1013 00:46:12,090 --> 00:46:14,460 provides same day settlement. 1014 00:46:14,460 --> 00:46:20,300 And very few accounts ask to do same day settlement. 1015 00:46:20,300 --> 00:46:23,990 And I think it might be just they've grown accustomed to it. 1016 00:46:23,990 --> 00:46:27,320 But I think there is some benefit in this stock, 1017 00:46:27,320 --> 00:46:30,710 loan stock borrow side. 1018 00:46:30,710 --> 00:46:33,750 But I stand with what I said earlier. 1019 00:46:33,750 --> 00:46:37,100 I think the market could work the other way. 1020 00:46:37,100 --> 00:46:40,620 But it would be a big adjustment. 1021 00:46:40,620 --> 00:46:42,510 And there are some benefits, because then 1022 00:46:42,510 --> 00:46:45,780 Hugo, this hedge fund, is being extended credit 1023 00:46:45,780 --> 00:46:48,580 by his prime brokerage account. 1024 00:46:48,580 --> 00:46:50,810 And the prime broker is making some money on it. 1025 00:46:50,810 --> 00:46:55,970 But also Hugo's taking risk in that arrangement. 1026 00:46:55,970 --> 00:46:59,960 Prefunding it would be a different model. 1027 00:46:59,960 --> 00:47:03,380 Some argue it would have less liquidity in the marketplace. 1028 00:47:03,380 --> 00:47:07,595 It would make shorting stocks a little costlier. 1029 00:47:07,595 --> 00:47:08,720 I'm not sure they're right. 1030 00:47:08,720 --> 00:47:10,700 But their argument is that I'd have 1031 00:47:10,700 --> 00:47:15,920 to borrow the securities always, 100%, and halve it off 1032 00:47:15,920 --> 00:47:19,760 before I sell, even if I'm a multi-billion dollar hedge 1033 00:47:19,760 --> 00:47:22,788 fund that has some credit worthiness. 1034 00:47:25,310 --> 00:47:25,810 Kelly. 1035 00:47:25,810 --> 00:47:31,210 AUDIENCE: So the paper mentions that their new system fails 1036 00:47:31,210 --> 00:47:35,170 to basically recognize these challenges, such as the fact 1037 00:47:35,170 --> 00:47:37,000 that it'd be advantaging the capital 1038 00:47:37,000 --> 00:47:38,632 and operational efficiencies. 1039 00:47:38,632 --> 00:47:40,090 But it goes on to say that it would 1040 00:47:40,090 --> 00:47:43,080 be a step backwards for the world's most liquid market. 1041 00:47:43,080 --> 00:47:46,000 So what would necessarily be implications 1042 00:47:46,000 --> 00:47:47,620 for illiquid markets? 1043 00:47:47,620 --> 00:47:50,530 Would it basically be that people can make trades 1044 00:47:50,530 --> 00:47:54,275 without having to interface with extensions of credit? 1045 00:47:54,275 --> 00:47:56,650 GARY GENSLER: Are you talking about illiquid markets that 1046 00:47:56,650 --> 00:47:58,960 are centrally cleared or illiquid markets that 1047 00:47:58,960 --> 00:48:00,692 have no central clearing? 1048 00:48:00,692 --> 00:48:01,900 AUDIENCE: I guess maybe both. 1049 00:48:01,900 --> 00:48:05,380 I'm just trying to understand what the counter would 1050 00:48:05,380 --> 00:48:09,130 be, that saying that there would be advantages, efficiencies 1051 00:48:09,130 --> 00:48:10,520 in liquid markets. 1052 00:48:10,520 --> 00:48:13,750 But what with the opposite be? 1053 00:48:13,750 --> 00:48:17,930 Would there not be implications for the [INAUDIBLE] markets? 1054 00:48:17,930 --> 00:48:19,463 GARY GENSLER: So Kelly's asking-- 1055 00:48:19,463 --> 00:48:21,130 I don't remember the comment, because it 1056 00:48:21,130 --> 00:48:24,680 was in the Australian Stock Exchange one. 1057 00:48:24,680 --> 00:48:30,430 So in non-cleared markets, things 1058 00:48:30,430 --> 00:48:33,400 that are just so illiquid-- my house. 1059 00:48:33,400 --> 00:48:34,240 It's not cleared. 1060 00:48:34,240 --> 00:48:36,400 It's not centrally cleared. 1061 00:48:36,400 --> 00:48:39,400 Certain parts of the derivatives markets aren't centrally clear. 1062 00:48:39,400 --> 00:48:42,360 They're not, quote, "liquid" enough. 1063 00:48:42,360 --> 00:48:43,110 So I don't know. 1064 00:48:43,110 --> 00:48:44,580 I don't know the comment. 1065 00:48:44,580 --> 00:48:45,750 Maybe see me in class. 1066 00:48:45,750 --> 00:48:49,080 And I'll look at the comment and try to nail it for you. 1067 00:48:49,080 --> 00:48:51,610 Is that all right? 1068 00:48:51,610 --> 00:48:54,520 So I wanted to put this chart up, because it was from 2011. 1069 00:48:54,520 --> 00:48:57,070 I remembered it when I was at the CFTC. 1070 00:48:57,070 --> 00:48:59,290 But somebody at the Fed of Chicago 1071 00:48:59,290 --> 00:49:02,890 made this chart about the US clearing systems. 1072 00:49:02,890 --> 00:49:05,700 Yellow are all clearing houses. 1073 00:49:05,700 --> 00:49:07,900 Red, the SEC, the Federal Reserve, New York-- 1074 00:49:11,162 --> 00:49:12,238 AUDIENCE: SBD? 1075 00:49:12,238 --> 00:49:13,030 GARY GENSLER: Yeah. 1076 00:49:13,030 --> 00:49:15,190 SEC, CFTC, and the Federal Reserve. 1077 00:49:15,190 --> 00:49:17,090 But they might have-- 1078 00:49:17,090 --> 00:49:17,590 New York. 1079 00:49:17,590 --> 00:49:19,093 I don't even know who New York SB-- 1080 00:49:19,093 --> 00:49:21,010 AUDIENCE: New York State Board of Derivatives. 1081 00:49:21,010 --> 00:49:23,420 GARY GENSLER: Yeah. 1082 00:49:23,420 --> 00:49:25,340 Somebody from the Chicago Fed was 1083 00:49:25,340 --> 00:49:28,760 trying to make the case that our regulatory system in the US 1084 00:49:28,760 --> 00:49:31,080 was a little bit complicated. 1085 00:49:31,080 --> 00:49:33,350 And I think this was in the midst of the debate 1086 00:49:33,350 --> 00:49:35,000 about Dodd-Frank, that they thought 1087 00:49:35,000 --> 00:49:37,280 all oversight of clearing houses should 1088 00:49:37,280 --> 00:49:39,140 move to the Federal Reserve. 1089 00:49:39,140 --> 00:49:40,940 But I remembered the chart and found it. 1090 00:49:40,940 --> 00:49:42,900 And I wanted to share it with you. 1091 00:49:42,900 --> 00:49:45,590 But the yellow are all the clearing houses. 1092 00:49:45,590 --> 00:49:47,870 There are a lot of clearing houses in the US. 1093 00:49:47,870 --> 00:49:49,410 There's one for bonds. 1094 00:49:49,410 --> 00:49:50,870 There's one for equities. 1095 00:49:50,870 --> 00:49:52,970 There's multiple ones for derivatives, 1096 00:49:52,970 --> 00:49:55,550 both from cash and interest rate derivatives 1097 00:49:55,550 --> 00:49:58,370 to energy derivatives to credit default derivatives 1098 00:49:58,370 --> 00:49:59,480 and so forth. 1099 00:49:59,480 --> 00:50:03,080 There's ones for options. 1100 00:50:03,080 --> 00:50:08,080 So I just wanted to lay it out. 1101 00:50:08,080 --> 00:50:11,300 Finance is pretty complex. 1102 00:50:11,300 --> 00:50:15,200 This is all digitized, by the way. 1103 00:50:15,200 --> 00:50:17,960 But in this complexity, that's also 1104 00:50:17,960 --> 00:50:20,570 where I think there's some opportunity, whether it's 1105 00:50:20,570 --> 00:50:24,200 permissioned blockchain and not fully permissioned 1106 00:50:24,200 --> 00:50:28,190 list with tokens, and inspired by-- 1107 00:50:31,580 --> 00:50:33,800 the green are exchanges. 1108 00:50:33,800 --> 00:50:36,230 And I don't even think they put a color up for brokers, 1109 00:50:36,230 --> 00:50:36,730 did they? 1110 00:50:36,730 --> 00:50:39,890 Because then you'd have hundreds. 1111 00:50:39,890 --> 00:50:44,330 There are multiple parties sharing ledgers and reconciling 1112 00:50:44,330 --> 00:50:45,170 their ledgers. 1113 00:50:45,170 --> 00:50:47,210 Reconciliation means I keep a ledger. 1114 00:50:47,210 --> 00:50:48,510 Tom keeps a ledger. 1115 00:50:48,510 --> 00:50:50,480 It's both on our database systems. 1116 00:50:50,480 --> 00:50:54,245 And we keep testing and ensuring that we have the same records. 1117 00:50:57,030 --> 00:51:00,120 This is a system that, by its complexity 1118 00:51:00,120 --> 00:51:05,960 and by its copies of ledgers and reconciliation, 1119 00:51:05,960 --> 00:51:10,100 has a lot of opportunity for greater efficiency. 1120 00:51:10,100 --> 00:51:12,620 And the question is whether blockchain technology, 1121 00:51:12,620 --> 00:51:17,260 maybe permissioned, can bring some of that efficiency. 1122 00:51:17,260 --> 00:51:18,480 That's the thought. 1123 00:51:18,480 --> 00:51:20,272 And that's what the chart was just trying-- 1124 00:51:22,280 --> 00:51:25,490 by the way, the Congress didn't agree with the Fed. 1125 00:51:25,490 --> 00:51:28,550 They left the authorities at the SEC and CFTC. 1126 00:51:28,550 --> 00:51:31,880 And those are for stories other times. 1127 00:51:31,880 --> 00:51:34,580 I personally felt it could be a shared authority. 1128 00:51:34,580 --> 00:51:36,650 And we ended up with some shared authorities. 1129 00:51:36,650 --> 00:51:41,010 But I thought that the CFTC and SEC have domain knowledge. 1130 00:51:41,010 --> 00:51:44,270 And it really important to keep the SEC involved in securities 1131 00:51:44,270 --> 00:51:48,260 clearing and the CFTC involved in derivatives clearing 1132 00:51:48,260 --> 00:51:52,970 for that domain knowledge, but that the Federal Reserve also 1133 00:51:52,970 --> 00:51:55,700 came in under something called Title VIII of Dodd-Frank. 1134 00:51:55,700 --> 00:51:57,440 And they also got authorities. 1135 00:51:57,440 --> 00:52:00,320 So we made it more complicated, because now there's 1136 00:52:00,320 --> 00:52:04,430 multiple dotted lines for this. 1137 00:52:04,430 --> 00:52:08,060 The Federal Reserve basically has an account with everybody. 1138 00:52:08,060 --> 00:52:12,410 Or they have some oversight of everybody. 1139 00:52:12,410 --> 00:52:14,335 So then the question on blockchain technology, 1140 00:52:14,335 --> 00:52:16,460 could you take the clearing house out of the middle 1141 00:52:16,460 --> 00:52:18,470 and do the thing on the right? 1142 00:52:18,470 --> 00:52:22,500 And my question for you all is, do you think the economics-- 1143 00:52:22,500 --> 00:52:24,570 I'm not talking about the technology. 1144 00:52:24,570 --> 00:52:26,930 I'm not talking about whether it's performance. 1145 00:52:26,930 --> 00:52:29,540 If you wish, you can go five or 10 years down the road 1146 00:52:29,540 --> 00:52:32,660 and say it's got all this scalability and performance. 1147 00:52:32,660 --> 00:52:36,110 Do you think the technology, without something 1148 00:52:36,110 --> 00:52:40,250 in the middle, what's the economics about that? 1149 00:52:40,250 --> 00:52:42,200 Anybody? 1150 00:52:42,200 --> 00:52:42,970 Eilon? 1151 00:52:42,970 --> 00:52:45,880 AUDIENCE: I still want to support the left side 1152 00:52:45,880 --> 00:52:47,645 for liquid markets because of the netting. 1153 00:52:47,645 --> 00:52:49,020 GARY GENSLER: Because of netting. 1154 00:52:49,020 --> 00:52:49,603 AUDIENCE: Yes. 1155 00:52:52,800 --> 00:52:54,180 I can't see a blockchain solution 1156 00:52:54,180 --> 00:52:56,850 working well when you need netting 1157 00:52:56,850 --> 00:53:00,720 to enjoy the efficiencies of selling and shorting, 1158 00:53:00,720 --> 00:53:01,580 shorting basically. 1159 00:53:01,580 --> 00:53:02,372 GARY GENSLER: Yeah. 1160 00:53:02,372 --> 00:53:05,580 But the question really is, can you somehow, 1161 00:53:05,580 --> 00:53:10,090 on the right with shared ledger, still 1162 00:53:10,090 --> 00:53:12,410 get some benefit of netting? 1163 00:53:12,410 --> 00:53:12,910 Right? 1164 00:53:12,910 --> 00:53:14,950 You're saying you're inclined here, 1165 00:53:14,950 --> 00:53:19,820 because you want to get the netting for economic reasons. 1166 00:53:19,820 --> 00:53:22,690 So as a ledger structure, can you get the netting here? 1167 00:53:22,690 --> 00:53:24,620 Shimon? 1168 00:53:24,620 --> 00:53:26,120 AUDIENCE: You don't need the netting 1169 00:53:26,120 --> 00:53:29,090 if it happens simultaneously. 1170 00:53:29,090 --> 00:53:32,273 Netting is it it is basically-- 1171 00:53:32,273 --> 00:53:33,440 again, with cash instrument. 1172 00:53:33,440 --> 00:53:34,857 I'm not talking about derivatives. 1173 00:53:34,857 --> 00:53:36,710 Derivatives is a different issue. 1174 00:53:36,710 --> 00:53:38,330 With cash instruments, the only reason 1175 00:53:38,330 --> 00:53:39,410 why you have counterparty risk is 1176 00:53:39,410 --> 00:53:41,610 because it's not clear-- it was my understanding. 1177 00:53:41,610 --> 00:53:43,640 Correct me-- is because you don't 1178 00:53:43,640 --> 00:53:45,890 have instantaneous settlement. 1179 00:53:45,890 --> 00:53:47,480 If you have a [INAUDIBLE] settlement, 1180 00:53:47,480 --> 00:53:48,897 then you don't care about netting. 1181 00:53:48,897 --> 00:53:51,550 AUDIENCE: You care because you face so many transaction fees. 1182 00:53:51,550 --> 00:53:52,092 AUDIENCE: No. 1183 00:53:52,092 --> 00:53:53,925 But that's going to be a lot more efficient. 1184 00:53:53,925 --> 00:54:00,570 So why-- again, do you think the [? real ?] transaction fees are 1185 00:54:00,570 --> 00:54:02,300 a big economic issue here? 1186 00:54:02,300 --> 00:54:02,910 AUDIENCE: Yes. 1187 00:54:02,910 --> 00:54:03,930 That's what I think. 1188 00:54:03,930 --> 00:54:06,040 GARY GENSLER: And I'm sorry that you don't. 1189 00:54:06,040 --> 00:54:07,220 AUDIENCE: So now it's fine. 1190 00:54:07,220 --> 00:54:07,762 I don't know. 1191 00:54:07,762 --> 00:54:08,726 I mean, I don't know. 1192 00:54:08,726 --> 00:54:11,030 For really intense help, how much 1193 00:54:11,030 --> 00:54:15,050 do you pay per transaction with the clearing house and without? 1194 00:54:15,050 --> 00:54:18,950 Maybe the net will give you only 10% of the transaction. 1195 00:54:18,950 --> 00:54:21,230 But the cost of having the clearing house 1196 00:54:21,230 --> 00:54:22,903 is more than 10 times. 1197 00:54:22,903 --> 00:54:23,570 AUDIENCE: Maybe. 1198 00:54:23,570 --> 00:54:25,760 AUDIENCE: So the economics depends on those numbers, 1199 00:54:25,760 --> 00:54:27,840 I think. 1200 00:54:27,840 --> 00:54:30,570 GARY GENSLER: Is that a question? 1201 00:54:30,570 --> 00:54:31,140 Hugo? 1202 00:54:31,140 --> 00:54:32,130 AUDIENCE: Yeah. 1203 00:54:32,130 --> 00:54:35,220 So you still can get some of the netting. 1204 00:54:35,220 --> 00:54:38,250 Let's just talk about blockchain transactions on Bitcoin. 1205 00:54:38,250 --> 00:54:40,110 You can batch transactions. 1206 00:54:40,110 --> 00:54:42,960 So you can have multiple inputs and multiple outputs. 1207 00:54:42,960 --> 00:54:44,590 And that's kind of like netting. 1208 00:54:44,590 --> 00:54:46,682 Then you have to have a second system 1209 00:54:46,682 --> 00:54:48,390 on top of that that keeps track of what's 1210 00:54:48,390 --> 00:54:49,557 going on throughout the day. 1211 00:54:49,557 --> 00:54:52,430 And then it just does the transactions 1212 00:54:52,430 --> 00:54:56,070 on a daily basis or an hourly basis or whatever. 1213 00:54:56,070 --> 00:54:58,570 That takes care of that netting. 1214 00:54:58,570 --> 00:54:59,160 Right? 1215 00:54:59,160 --> 00:55:00,660 GARY GENSLER: So Hugo contends maybe 1216 00:55:00,660 --> 00:55:03,150 you can get this economic thing called netting 1217 00:55:03,150 --> 00:55:06,570 and still have a shared distributed ledger. 1218 00:55:06,570 --> 00:55:09,630 Or you could do it off chain. 1219 00:55:09,630 --> 00:55:12,900 I think, Shimon to your question, 1220 00:55:12,900 --> 00:55:15,360 this is a big economic thing either 1221 00:55:15,360 --> 00:55:17,220 because we've just grown accustomed to it 1222 00:55:17,220 --> 00:55:21,360 over a couple centuries or it's fundamental to how markets 1223 00:55:21,360 --> 00:55:25,460 trade, that we don't prefund. 1224 00:55:25,460 --> 00:55:29,770 It's that economics that blockchain technology 1225 00:55:29,770 --> 00:55:31,690 will challenge. 1226 00:55:31,690 --> 00:55:34,060 But we might stay with a central clearing house 1227 00:55:34,060 --> 00:55:35,770 if the economics are that value. 1228 00:55:35,770 --> 00:55:40,900 Well, the economics of, in essence, not prefunding, 1229 00:55:40,900 --> 00:55:45,390 and thus, some delayed settlement. 1230 00:55:45,390 --> 00:55:48,210 Whereas in the past, we always had delayed settlement at first 1231 00:55:48,210 --> 00:55:50,100 because there was so much paperwork. 1232 00:55:50,100 --> 00:55:52,080 We had to have delayed settlement. 1233 00:55:52,080 --> 00:55:53,700 We went digitized. 1234 00:55:53,700 --> 00:55:59,480 And now, really in 2018, we don't need delayed settlement 1235 00:55:59,480 --> 00:56:03,130 or batch processing. 1236 00:56:03,130 --> 00:56:05,010 So now you've exposed, well, wait. 1237 00:56:05,010 --> 00:56:07,670 What are the economics? 1238 00:56:07,670 --> 00:56:08,765 Tom and then Alin? 1239 00:56:08,765 --> 00:56:10,890 AUDIENCE: Could you still have a delayed settlement 1240 00:56:10,890 --> 00:56:14,290 with a blockchain system just by using a smart contract? 1241 00:56:14,290 --> 00:56:15,290 GARY GENSLER: You could. 1242 00:56:15,290 --> 00:56:15,920 Absolutely. 1243 00:56:15,920 --> 00:56:19,250 AUDIENCE: I mean, you still keep this system 1244 00:56:19,250 --> 00:56:23,210 that people are accustomed to, but do it 1245 00:56:23,210 --> 00:56:25,910 without the economic red for the clearing house's 1246 00:56:25,910 --> 00:56:27,080 sides or [INAUDIBLE]. 1247 00:56:27,080 --> 00:56:27,830 GARY GENSLER: Yes. 1248 00:56:27,830 --> 00:56:31,010 And it's what the Australian Stock Exchange 1249 00:56:31,010 --> 00:56:36,860 says they are using with the digital asset company, Blythe 1250 00:56:36,860 --> 00:56:39,620 Masters' company that's their outside vendor using 1251 00:56:39,620 --> 00:56:40,910 the HyperLedger Fabric. 1252 00:56:40,910 --> 00:56:42,350 And they're using the code. 1253 00:56:42,350 --> 00:56:47,585 They call a DAML, or Digital Asset Machine Language? 1254 00:56:47,585 --> 00:56:48,710 I don't know what the M is. 1255 00:56:48,710 --> 00:56:51,606 But they call it DAML. 1256 00:56:51,606 --> 00:56:52,797 Alin? 1257 00:56:52,797 --> 00:56:53,380 AUDIENCE: Yes. 1258 00:56:53,380 --> 00:56:56,063 We've said you cannot prefund on the right. 1259 00:56:56,063 --> 00:56:56,730 Is that correct? 1260 00:57:01,927 --> 00:57:04,260 GARY GENSLER: I think you could through smart contracts. 1261 00:57:04,260 --> 00:57:05,790 I think you could find this. 1262 00:57:05,790 --> 00:57:10,860 But the challenge with just the picture, the picture 1263 00:57:10,860 --> 00:57:13,110 suggests there's no central counterparty 1264 00:57:13,110 --> 00:57:15,720 to do all the netting. 1265 00:57:15,720 --> 00:57:19,290 And so the question is, can you have this ledger structure 1266 00:57:19,290 --> 00:57:23,760 and still find a way to net down all the positions? 1267 00:57:23,760 --> 00:57:25,610 AUDIENCE: Well, like Shimon said, 1268 00:57:25,610 --> 00:57:28,140 there is no need for that. 1269 00:57:28,140 --> 00:57:30,450 You get that implicitly by recording everything 1270 00:57:30,450 --> 00:57:31,672 and executing. 1271 00:57:31,672 --> 00:57:33,755 GARY GENSLER: But if you're recording-- all right. 1272 00:57:33,755 --> 00:57:35,150 I'm agreeing with that, too. 1273 00:57:35,150 --> 00:57:36,930 But that's where I get to if you need 1274 00:57:36,930 --> 00:57:42,360 to do Shimon's approach, which is simultaneous execution 1275 00:57:42,360 --> 00:57:45,180 and settlement, then you need to prefund, 1276 00:57:45,180 --> 00:57:47,550 even if that prefunding is by nanoseconds. 1277 00:57:47,550 --> 00:57:49,377 But you need to prefund. 1278 00:57:49,377 --> 00:57:50,460 AUDIENCE: Not necessarily. 1279 00:57:50,460 --> 00:57:52,430 You could, sort of optimistically, you 1280 00:57:52,430 --> 00:57:55,895 could post a transaction that says, I want to buy a stock. 1281 00:57:55,895 --> 00:57:57,270 And I'm not going to prove to you 1282 00:57:57,270 --> 00:57:58,980 that I have money to buy it. 1283 00:57:58,980 --> 00:58:01,900 You record it, if you were saying optimistically. 1284 00:58:01,900 --> 00:58:04,738 And later on-- 1285 00:58:04,738 --> 00:58:05,780 GARY GENSLER: It depends. 1286 00:58:05,780 --> 00:58:10,200 The word record-- let me use a different word, settle. 1287 00:58:10,200 --> 00:58:13,470 Settlement means final legal rights have moved 1288 00:58:13,470 --> 00:58:16,380 from one party to another. 1289 00:58:16,380 --> 00:58:18,770 So the word settlement has a meaning that I just 1290 00:58:18,770 --> 00:58:20,060 want to come back to. 1291 00:58:20,060 --> 00:58:22,730 It means I'm recording on a ledger 1292 00:58:22,730 --> 00:58:25,070 that I no longer have legal rights. 1293 00:58:25,070 --> 00:58:26,890 So this alpha has legal rights to it. 1294 00:58:29,580 --> 00:58:32,080 And I think you would have to prefund if you were 1295 00:58:32,080 --> 00:58:34,480 doing true settlement ledgers. 1296 00:58:34,480 --> 00:58:36,100 AUDIENCE: Sure, sure. 1297 00:58:36,100 --> 00:58:38,860 I guess I'm saying it seems like you could simulate the clearing 1298 00:58:38,860 --> 00:58:40,420 house. 1299 00:58:40,420 --> 00:58:42,442 In the right side of this picture, 1300 00:58:42,442 --> 00:58:43,900 the clearing house just distributed 1301 00:58:43,900 --> 00:58:45,070 using the consensus protocol. 1302 00:58:45,070 --> 00:58:45,940 GARY GENSLER: Right. 1303 00:58:45,940 --> 00:58:48,593 I think you're right that you could. 1304 00:58:48,593 --> 00:58:50,260 But if you're going all the way to where 1305 00:58:50,260 --> 00:58:54,590 Shimon is, like Bitcoin, and you're doing final settlement, 1306 00:58:54,590 --> 00:58:57,170 final settlement recordation, then I 1307 00:58:57,170 --> 00:58:59,257 think you have to prefund. 1308 00:58:59,257 --> 00:59:00,590 A couple of questions over here. 1309 00:59:00,590 --> 00:59:02,700 Eric, Ross. 1310 00:59:02,700 --> 00:59:06,470 AUDIENCE: The reason that the prefunding was [INAUDIBLE] 1311 00:59:06,470 --> 00:59:11,662 performed in the validation of the block process, because-- 1312 00:59:11,662 --> 00:59:12,620 GARY GENSLER: It could. 1313 00:59:12,620 --> 00:59:15,110 I said it could even be done during the nanoseconds. 1314 00:59:15,110 --> 00:59:18,380 But it has to be done before you move the settlement. 1315 00:59:18,380 --> 00:59:21,140 And that's what's called clearing. 1316 00:59:21,140 --> 00:59:24,080 So it's a question of whether you have simultaneous execution 1317 00:59:24,080 --> 00:59:28,130 and settlement, or you still have execution, clearing, 1318 00:59:28,130 --> 00:59:31,890 settlement, even if separated by seconds, 1319 00:59:31,890 --> 00:59:35,250 but still literally get granular. 1320 00:59:35,250 --> 00:59:36,000 I apologize. 1321 00:59:36,000 --> 00:59:39,380 This stuff is that granular. 1322 00:59:39,380 --> 00:59:41,310 Back here. 1323 00:59:41,310 --> 00:59:43,380 And then I'm going to move on. 1324 00:59:43,380 --> 00:59:45,863 AUDIENCE: Why do you consider prefunding-- 1325 00:59:45,863 --> 00:59:46,530 maybe you don't. 1326 00:59:46,530 --> 00:59:49,315 But it seems simply you consider prefunding 1327 00:59:49,315 --> 00:59:50,690 before you're allowed to transact 1328 00:59:50,690 --> 00:59:52,512 a negative economically. 1329 00:59:52,512 --> 00:59:54,220 GARY GENSLER: I shouldn't sound that way. 1330 00:59:54,220 --> 00:59:56,550 It's just that for a couple years, 1331 00:59:56,550 --> 00:59:59,440 we've had delayed settlement. 1332 00:59:59,440 --> 01:00:03,750 And so I'm saying if you are a proponent of a system that 1333 01:00:03,750 --> 01:00:08,390 has simultaneous execution and settlement, 1334 01:00:08,390 --> 01:00:09,810 you just have to deal with it. 1335 01:00:09,810 --> 01:00:11,720 You're changing the economic models 1336 01:00:11,720 --> 01:00:15,720 that have been so common in the capital markets. 1337 01:00:15,720 --> 01:00:17,180 And when you take that on, you just 1338 01:00:17,180 --> 01:00:22,240 have to think through what are the economics of that shift, 1339 01:00:22,240 --> 01:00:25,280 from in essence a delayed settlement, 1340 01:00:25,280 --> 01:00:28,220 thus modest but real counterparty 1341 01:00:28,220 --> 01:00:32,120 risk that Hugo's got a prime broker and he's got that, 1342 01:00:32,120 --> 01:00:36,380 and moving to this new model in capital markets. 1343 01:00:36,380 --> 01:00:38,480 So I shouldn't come across as negative. 1344 01:00:38,480 --> 01:00:42,320 But I'm saying at least it's a hurdle that you've 1345 01:00:42,320 --> 01:00:45,320 got an adoption curve that's going to be a challenge, 1346 01:00:45,320 --> 01:00:47,710 I think, at a minimum. 1347 01:00:51,550 --> 01:00:54,220 DTCC, when I first spoke to them-- 1348 01:00:54,220 --> 01:00:58,480 and I've had a lively, ongoing dialogue with the folks 1349 01:00:58,480 --> 01:01:02,110 there on an occasional basis. 1350 01:01:02,110 --> 01:01:03,970 This was the biggest issue they raised. 1351 01:01:03,970 --> 01:01:06,640 They said they looked at blockchain technology 1352 01:01:06,640 --> 01:01:08,950 starting several years ago. 1353 01:01:08,950 --> 01:01:11,080 They talked to their customer base. 1354 01:01:11,080 --> 01:01:13,330 And they couldn't find that their customer base really 1355 01:01:13,330 --> 01:01:16,990 wanted to go to simultaneous execution and settlement. 1356 01:01:16,990 --> 01:01:18,430 They have same day settlement. 1357 01:01:18,430 --> 01:01:19,990 But it's end of day settlement. 1358 01:01:19,990 --> 01:01:23,253 So there's still several hours. 1359 01:01:23,253 --> 01:01:24,670 But they said, we don't have a lot 1360 01:01:24,670 --> 01:01:28,390 of customers that really want simultaneous execution 1361 01:01:28,390 --> 01:01:29,680 and settlement. 1362 01:01:29,680 --> 01:01:32,645 Now, that was commenting on the demand side. 1363 01:01:32,645 --> 01:01:34,020 It's another question whether you 1364 01:01:34,020 --> 01:01:37,500 could adopt this lower cost, do something, and change 1365 01:01:37,500 --> 01:01:38,420 your capital market. 1366 01:01:43,260 --> 01:01:46,380 So this is another look at it. 1367 01:01:46,380 --> 01:01:49,080 This is also from the Australian Stock Exchange, 1368 01:01:49,080 --> 01:01:50,560 I think, website if I remember. 1369 01:01:50,560 --> 01:01:52,080 But maybe it was another website. 1370 01:01:52,080 --> 01:01:54,870 The question is, could you take out these middle parts? 1371 01:01:54,870 --> 01:01:57,180 This is like blockchain architecture 1372 01:01:57,180 --> 01:02:00,180 for clearing central counterparty 1373 01:02:00,180 --> 01:02:02,490 and the central securities depository. 1374 01:02:02,490 --> 01:02:07,100 Because there's also somebody who's got the central registry, 1375 01:02:07,100 --> 01:02:09,270 and that central registry are the two things. 1376 01:02:09,270 --> 01:02:11,640 Could you take this out completely? 1377 01:02:11,640 --> 01:02:15,510 That was the question, in a sense. 1378 01:02:15,510 --> 01:02:19,320 So here are the projects that are actually live or being 1379 01:02:19,320 --> 01:02:19,830 explored. 1380 01:02:19,830 --> 01:02:21,330 And we're going to talk a little bit 1381 01:02:21,330 --> 01:02:25,330 about Australian Stock Exchange in the few minutes we 1382 01:02:25,330 --> 01:02:25,830 have left. 1383 01:02:25,830 --> 01:02:28,980 But this Australian stock Exchange Project 1384 01:02:28,980 --> 01:02:30,630 started in 2016. 1385 01:02:30,630 --> 01:02:32,340 A bunch of requests for proposals, 1386 01:02:32,340 --> 01:02:34,420 like higher digital asset holding, 1387 01:02:34,420 --> 01:02:36,360 they're actually actively working on it. 1388 01:02:36,360 --> 01:02:38,770 It's a 25-year-old system. 1389 01:02:38,770 --> 01:02:40,200 And Jeff Sprecher said it, too. 1390 01:02:40,200 --> 01:02:42,360 He said, you know, they had a legacy system 1391 01:02:42,360 --> 01:02:43,470 they needed to replace. 1392 01:02:43,470 --> 01:02:46,950 They decided to go with the new technology of the day. 1393 01:02:46,950 --> 01:02:50,180 But Jeff wasn't convinced they had to. 1394 01:02:50,180 --> 01:02:53,240 But they feel that it was the right thing 1395 01:02:53,240 --> 01:02:58,250 to do, to go with that new technology of the day. 1396 01:02:58,250 --> 01:03:01,970 They believe-- and they've put out reports as recently 1397 01:03:01,970 --> 01:03:03,830 as several weeks ago. 1398 01:03:03,830 --> 01:03:05,540 They believe they're on the right path 1399 01:03:05,540 --> 01:03:08,390 and that 15 or 20 years from now, every clearing house 1400 01:03:08,390 --> 01:03:10,070 will be using what they would consider 1401 01:03:10,070 --> 01:03:12,320 a permissioned blockchain solution, 1402 01:03:12,320 --> 01:03:17,030 because it will so significantly lower the back office 1403 01:03:17,030 --> 01:03:20,000 costs for their market participants 1404 01:03:20,000 --> 01:03:22,010 who are now holding records. 1405 01:03:22,010 --> 01:03:27,620 Each bank, was it 77 members from an earlier-- yeah. 1406 01:03:27,620 --> 01:03:29,930 Each bank has to hold records. 1407 01:03:29,930 --> 01:03:32,750 It could be held in a shared ledger. 1408 01:03:32,750 --> 01:03:35,240 And they would lower the reconciliation cost. 1409 01:03:35,240 --> 01:03:36,980 And they believe the second advantage 1410 01:03:36,980 --> 01:03:38,630 is using smart contracts. 1411 01:03:38,630 --> 01:03:41,570 Now, are they exactly the smart contracts 1412 01:03:41,570 --> 01:03:43,170 like on the Ethereum network? 1413 01:03:43,170 --> 01:03:46,100 No, because they don't have a native token. 1414 01:03:46,100 --> 01:03:48,170 They don't have a DAP and a native token. 1415 01:03:48,170 --> 01:03:51,410 But it's still automating certain processes, 1416 01:03:51,410 --> 01:03:56,510 automating certain I'll call it back office processes. 1417 01:03:56,510 --> 01:03:58,130 So they're very committed to it. 1418 01:03:58,130 --> 01:04:00,950 And they're going to roll it out in 2020. 1419 01:04:00,950 --> 01:04:05,210 And then, for 12 months, run it simultaneous 1420 01:04:05,210 --> 01:04:07,590 with their legacy system. 1421 01:04:07,590 --> 01:04:13,980 So their cut over for assuring that it's resilient 1422 01:04:13,980 --> 01:04:16,650 is at 12 month test. 1423 01:04:16,650 --> 01:04:18,510 Most major market infrastructures 1424 01:04:18,510 --> 01:04:21,850 that roll things out do not roll simultaneously for 12 months. 1425 01:04:21,850 --> 01:04:24,210 But they all roll simultaneous. 1426 01:04:24,210 --> 01:04:28,380 Whether it's a month, whether it's 90 days, 1427 01:04:28,380 --> 01:04:31,200 you can't cut over the back office of the New York Stock 1428 01:04:31,200 --> 01:04:35,370 Exchange and just test it for one day. 1429 01:04:35,370 --> 01:04:36,490 Or you can. 1430 01:04:36,490 --> 01:04:40,944 It's just kind of choppy and a big risk. 1431 01:04:40,944 --> 01:04:42,090 AUDIENCE: Quick question. 1432 01:04:42,090 --> 01:04:45,285 Are they going to use a stable coin to settle the transaction? 1433 01:04:45,285 --> 01:04:47,160 How are they going to settle the transaction? 1434 01:04:47,160 --> 01:04:49,368 Because I can see how you issue securities digitally. 1435 01:04:49,368 --> 01:04:50,740 But you cannot issue cash. 1436 01:04:50,740 --> 01:04:51,870 GARY GENSLER: So this is very good. 1437 01:04:51,870 --> 01:04:52,495 It's very good. 1438 01:04:52,495 --> 01:04:56,940 So all of these clearing houses, they 1439 01:04:56,940 --> 01:05:00,840 are the final custodian for the security. 1440 01:05:00,840 --> 01:05:04,900 Securities are dematerialized, meaning they're all digital. 1441 01:05:04,900 --> 01:05:07,890 And so those digital assets are held on a ledger 1442 01:05:07,890 --> 01:05:10,710 and what's called the golden record. 1443 01:05:10,710 --> 01:05:13,320 This is the words used in the securities business. 1444 01:05:13,320 --> 01:05:16,140 The gold record or the golden record, meaning it 1445 01:05:16,140 --> 01:05:19,083 is the legal, verifiable thing you 1446 01:05:19,083 --> 01:05:20,250 can go to in a court of law. 1447 01:05:20,250 --> 01:05:22,020 And it says, this is who owns this. 1448 01:05:22,020 --> 01:05:22,800 Alpha owns it. 1449 01:05:22,800 --> 01:05:25,080 Not Hugo. 1450 01:05:25,080 --> 01:05:26,790 Sorry. 1451 01:05:26,790 --> 01:05:28,950 You're working with Goldman Sachs, though, right? 1452 01:05:31,770 --> 01:05:34,830 That golden record is there for securities. 1453 01:05:34,830 --> 01:05:36,915 Who has the golden record for cash? 1454 01:05:36,915 --> 01:05:38,040 AUDIENCE: The central bank? 1455 01:05:38,040 --> 01:05:41,700 GARY GENSLER: The central bank or the fractional banking 1456 01:05:41,700 --> 01:05:44,400 system, their commercial bank subledger. 1457 01:05:47,370 --> 01:05:49,290 Clearing houses around the globe would 1458 01:05:49,290 --> 01:05:54,280 like to be able to have direct access to central bank money. 1459 01:05:54,280 --> 01:05:58,100 And they currently do not in any of the major jurisdictions 1460 01:05:58,100 --> 01:05:58,670 that I know. 1461 01:05:58,670 --> 01:05:59,950 I think. 1462 01:05:59,950 --> 01:06:04,790 So what Australia does now, this is what they do now. 1463 01:06:04,790 --> 01:06:06,800 They net everybody down. 1464 01:06:06,800 --> 01:06:08,120 They get to t plus two. 1465 01:06:08,120 --> 01:06:09,380 They're on that second day. 1466 01:06:09,380 --> 01:06:10,490 Everything's netted down. 1467 01:06:10,490 --> 01:06:14,350 And they're about to do the whole closing out. 1468 01:06:14,350 --> 01:06:17,650 One second before they do the security settlement, 1469 01:06:17,650 --> 01:06:20,300 they do the cash settlement. 1470 01:06:20,300 --> 01:06:23,382 They call this Delivery Versus Payment, DVP. 1471 01:06:23,382 --> 01:06:24,590 But they're a clearing house. 1472 01:06:24,590 --> 01:06:28,000 And they do not want to take risk with the central bank. 1473 01:06:28,000 --> 01:06:31,420 They take all the money and settle all the cash 1474 01:06:31,420 --> 01:06:35,170 with the Royal Bank of Australia, RBA I guess. 1475 01:06:35,170 --> 01:06:36,280 Do I have the right name? 1476 01:06:36,280 --> 01:06:40,300 Or is it a different-- is RBA one of the banks in Australia? 1477 01:06:40,300 --> 01:06:42,310 But they settle all the cash. 1478 01:06:42,310 --> 01:06:45,790 And just within a second, they settle. 1479 01:06:45,790 --> 01:06:47,230 We think of it as DVP. 1480 01:06:47,230 --> 01:06:49,360 But it's actually technically not. 1481 01:06:49,360 --> 01:06:53,380 Inside their clearing systems, they 1482 01:06:53,380 --> 01:06:55,540 have a form of Australian dollars 1483 01:06:55,540 --> 01:07:02,590 that is effectively digitally, in their system, digital money. 1484 01:07:02,590 --> 01:07:04,090 In essence, what they're doing is 1485 01:07:04,090 --> 01:07:11,497 clearing their internal digital money versus Royal Bank money. 1486 01:07:11,497 --> 01:07:12,330 AUDIENCE: Thank you. 1487 01:07:12,330 --> 01:07:13,520 GARY GENSLER: All right. 1488 01:07:13,520 --> 01:07:14,750 Sorry. 1489 01:07:14,750 --> 01:07:16,160 So these other projects-- 1490 01:07:16,160 --> 01:07:19,310 Estonia, the Tallinn Stock Exchange, 1491 01:07:19,310 --> 01:07:21,690 is exploring proxy voting and registration. 1492 01:07:21,690 --> 01:07:27,080 What's interesting-- who runs the Estonian Stock Exchange? 1493 01:07:27,080 --> 01:07:28,430 NASDAQ. 1494 01:07:28,430 --> 01:07:30,030 NASDAQ. 1495 01:07:30,030 --> 01:07:32,360 So NASDAQ runs it. 1496 01:07:32,360 --> 01:07:33,710 And they're doing this project. 1497 01:07:33,710 --> 01:07:38,450 NASDAQ announced and went live in 2015 on an illiquid. 1498 01:07:38,450 --> 01:07:41,650 So Kelly, you asked me about illiquids earlier. 1499 01:07:41,650 --> 01:07:47,750 NASDAQ started a blockchain technology platform in 2015 1500 01:07:47,750 --> 01:07:51,140 on private securities, which were not-- 1501 01:07:51,140 --> 01:07:53,420 they certainly weren't liquid. 1502 01:07:53,420 --> 01:07:55,790 I don't know enough about what they did there as 1503 01:07:55,790 --> 01:07:58,370 to how illiquid they were. 1504 01:07:58,370 --> 01:08:03,380 But they felt, all right, this is a place we could do this. 1505 01:08:03,380 --> 01:08:06,175 And it's still gone live. 1506 01:08:06,175 --> 01:08:08,120 It's got very little use. 1507 01:08:08,120 --> 01:08:09,565 But it's still there. 1508 01:08:09,565 --> 01:08:10,940 And they say they're going to try 1509 01:08:10,940 --> 01:08:14,720 to do it for mutual funds, when you and I can just 1510 01:08:14,720 --> 01:08:18,020 by interest in mutual funds. 1511 01:08:18,020 --> 01:08:22,310 And NASDAQ's going to offer that to mutual funds, 1512 01:08:22,310 --> 01:08:24,290 partly because mutual funds aren't 1513 01:08:24,290 --> 01:08:28,399 in DTCC from what I understand. 1514 01:08:28,399 --> 01:08:33,720 The others are just really some tests. 1515 01:08:33,720 --> 01:08:36,590 Japan Exchange Group for the Tokyo Stock Exchange 1516 01:08:36,590 --> 01:08:37,850 has done some tests. 1517 01:08:37,850 --> 01:08:38,852 They've written papers. 1518 01:08:38,852 --> 01:08:40,310 They're sort of interesting papers. 1519 01:08:40,310 --> 01:08:43,729 The last one was earlier this year. 1520 01:08:43,729 --> 01:08:47,840 They're really trying it just on post trade matching. 1521 01:08:47,840 --> 01:08:50,810 One of the securities firms in Japan, I can't remember, 1522 01:08:50,810 --> 01:08:54,620 is really pressing the Tokyo Stock Exchange to do this. 1523 01:08:54,620 --> 01:08:58,600 So there's some local commercial politics that are going on. 1524 01:08:58,600 --> 01:09:01,430 I can't remember which exchange is-- 1525 01:09:01,430 --> 01:09:02,149 not exchange. 1526 01:09:02,149 --> 01:09:06,740 One of the brokerage houses is pushing them to do it. 1527 01:09:06,740 --> 01:09:08,310 All permissioned systems. 1528 01:09:08,310 --> 01:09:12,960 No, to my knowledge, permissionless systems. 1529 01:09:12,960 --> 01:09:16,490 Let's talk about ISDA for a few minutes, derivatives 1530 01:09:16,490 --> 01:09:20,750 common domain model. 1531 01:09:20,750 --> 01:09:24,560 One of my former commissioner's colleagues, Scott O'Malia, 1532 01:09:24,560 --> 01:09:25,640 runs ISDA. 1533 01:09:29,689 --> 01:09:32,479 That's a international trade organization. 1534 01:09:32,479 --> 01:09:37,520 It's largely, largely dominated by the 16 big banks 1535 01:09:37,520 --> 01:09:43,100 around the globe that probably transact 98% or 99% 1536 01:09:43,100 --> 01:09:46,550 of derivatives around the globe. 1537 01:09:46,550 --> 01:09:50,359 And it was started in the 1980s to do all the forms, 1538 01:09:50,359 --> 01:09:54,140 all the forms to enter into the legal contractual obligations 1539 01:09:54,140 --> 01:09:55,980 called swaps. 1540 01:09:55,980 --> 01:09:59,760 And over the years, it's transitioned. 1541 01:09:59,760 --> 01:10:01,500 It's also an advocacy group. 1542 01:10:01,500 --> 01:10:04,470 And it lobbies various regulators 1543 01:10:04,470 --> 01:10:06,840 in Europe and the US and Asia. 1544 01:10:06,840 --> 01:10:08,970 But they've, a year or so ago, said, wait a minute. 1545 01:10:08,970 --> 01:10:14,900 We could take a lot of the back office of swaps, particularly 1546 01:10:14,900 --> 01:10:17,930 swaps that are not going into clearing houses. 1547 01:10:17,930 --> 01:10:20,840 So it's called the non-cleared or uncleared swaps. 1548 01:10:20,840 --> 01:10:24,320 And maybe we can create smart contracts 1549 01:10:24,320 --> 01:10:27,110 to automate a bunch of contractual terms. 1550 01:10:27,110 --> 01:10:28,860 So that's the second bullet point. 1551 01:10:28,860 --> 01:10:30,290 And they contracted out. 1552 01:10:30,290 --> 01:10:32,480 They put out a request for proposal, 1553 01:10:32,480 --> 01:10:34,700 basically, for somebody to come up 1554 01:10:34,700 --> 01:10:37,970 with machine readable standard representation. 1555 01:10:37,970 --> 01:10:41,300 Call it code, a scripting language. 1556 01:10:41,300 --> 01:10:43,220 I can't remember if it's-- 1557 01:10:43,220 --> 01:10:45,200 oh, it's written in JavaScript. 1558 01:10:45,200 --> 01:10:47,651 So written in Java and JavaScript, 1559 01:10:47,651 --> 01:10:53,150 a standard machine readable code of various events. 1560 01:10:53,150 --> 01:10:57,830 If it's a 20-year swap that's going to pay every quarter 80 1561 01:10:57,830 --> 01:11:02,660 quarterly payments, each of those quarterly payments, 1562 01:11:02,660 --> 01:11:05,810 fixed versus floating payments, is somehow 1563 01:11:05,810 --> 01:11:08,290 put into machine readable code. 1564 01:11:08,290 --> 01:11:11,570 And Scott and ISDA's view was, if we 1565 01:11:11,570 --> 01:11:16,010 could get a broad consensus and publish that machine readable 1566 01:11:16,010 --> 01:11:19,400 code that everybody agrees on, maybe that machine readable 1567 01:11:19,400 --> 01:11:23,390 code could take the place of legal contracts. 1568 01:11:23,390 --> 01:11:27,150 ISDA, for 30 plus years, one of their lines of business 1569 01:11:27,150 --> 01:11:30,590 is creating those legal contracts. 1570 01:11:30,590 --> 01:11:32,990 Maybe we can get a step ahead of this 1571 01:11:32,990 --> 01:11:38,170 and get consensus amongst these 16 big banks. 1572 01:11:38,170 --> 01:11:41,240 There's about 100,000 users of swaps 1573 01:11:41,240 --> 01:11:42,830 when you say all the customers. 1574 01:11:42,830 --> 01:11:45,870 It's not measured in the tens of millions. 1575 01:11:45,870 --> 01:11:47,990 I doubt anybody in this room personally 1576 01:11:47,990 --> 01:11:50,870 has a 20-year or 10-year swap. 1577 01:11:50,870 --> 01:11:52,430 But I don't want to embarrass. 1578 01:11:52,430 --> 01:11:54,920 Maybe somebody does. 1579 01:11:54,920 --> 01:11:59,440 But it's an institutional product, largely. 1580 01:11:59,440 --> 01:12:00,850 That's their product. 1581 01:12:00,850 --> 01:12:03,880 And it covers new transactions, rate resets, 1582 01:12:03,880 --> 01:12:08,930 partial terminations, as I listed up there. 1583 01:12:08,930 --> 01:12:10,510 Questions? 1584 01:12:10,510 --> 01:12:15,380 And it's definitely very much inspired by Ethereum 1585 01:12:15,380 --> 01:12:19,260 and smart contracts and DAPs. 1586 01:12:19,260 --> 01:12:24,360 But in essence, it's not really blockchain technology. 1587 01:12:24,360 --> 01:12:26,370 It's inspired by it, in a sense. 1588 01:12:26,370 --> 01:12:30,060 AUDIENCE: They're just rewriting the contracts instead of-- 1589 01:12:30,060 --> 01:12:33,410 they think they can have more specificity and certainty 1590 01:12:33,410 --> 01:12:37,020 of about what the contract means by writing it in contract. 1591 01:12:37,020 --> 01:12:40,570 GARY GENSLER: And importantly, importantly, efficiency-- 1592 01:12:40,570 --> 01:12:44,530 to automate that which humans are still doing, even 1593 01:12:44,530 --> 01:12:46,900 if it's in written contract. 1594 01:12:46,900 --> 01:12:49,130 So it's also automation that's inspiring them. 1595 01:12:52,413 --> 01:12:53,580 But you're absolutely right. 1596 01:12:53,580 --> 01:12:57,030 It's taking legal contract, getting into machine readable 1597 01:12:57,030 --> 01:13:01,320 JavaScript code, getting consensus on what that code is, 1598 01:13:01,320 --> 01:13:04,008 getting some courts to recognize it. 1599 01:13:04,008 --> 01:13:05,550 And then you won't even need to write 1600 01:13:05,550 --> 01:13:09,510 some of this, a rate reset or a partial termination, 1601 01:13:09,510 --> 01:13:11,130 into the contract. 1602 01:13:11,130 --> 01:13:15,470 You'd be referencing this. 1603 01:13:15,470 --> 01:13:17,780 The longer term objective is to say maybe 1604 01:13:17,780 --> 01:13:23,623 we can put even more of it right into code. 1605 01:13:23,623 --> 01:13:26,040 AUDIENCE: People would have to pool their money somewhere. 1606 01:13:26,040 --> 01:13:27,290 It's like we were talking about, right? 1607 01:13:27,290 --> 01:13:27,760 GARY GENSLER: That's correct. 1608 01:13:27,760 --> 01:13:29,320 AUDIENCE: It's going to execute itself, 1609 01:13:29,320 --> 01:13:31,612 which is going to have to be a pool of money somewhere. 1610 01:13:31,612 --> 01:13:34,580 GARY GENSLER: That you might have to some prefund or self 1611 01:13:34,580 --> 01:13:36,410 fund or something. 1612 01:13:36,410 --> 01:13:38,650 Let me go back here, and then up. 1613 01:13:38,650 --> 01:13:42,450 AUDIENCE: If they're using blockchain on it, it can't be-- 1614 01:13:42,450 --> 01:13:44,450 GARY GENSLER: They call it a blockchain project. 1615 01:13:44,450 --> 01:13:46,620 But it's a-- 1616 01:13:46,620 --> 01:13:49,430 I think I've tried to capture the essence of what 1617 01:13:49,430 --> 01:13:50,580 I think they have here. 1618 01:13:50,580 --> 01:13:52,250 It's definitely blockchain inspired. 1619 01:13:52,250 --> 01:13:54,520 I'm just not sure it's really stored on a blockchain. 1620 01:13:54,520 --> 01:13:55,145 AUDIENCE: Yeah. 1621 01:13:55,145 --> 01:13:58,030 Because if you look at credit default swap, 1622 01:13:58,030 --> 01:14:00,515 you have to reference a credit rating somewhere digitally. 1623 01:14:00,515 --> 01:14:03,470 And wherever that's referenced digitally, if someone makes 1624 01:14:03,470 --> 01:14:05,990 a mistake, does a transaction period that's 1625 01:14:05,990 --> 01:14:07,945 based on that mistake and it's fixed, 1626 01:14:07,945 --> 01:14:09,820 someone's going to have to be able to reverse 1627 01:14:09,820 --> 01:14:11,245 that transaction. 1628 01:14:11,245 --> 01:14:12,620 GARY GENSLER: So what's raised is 1629 01:14:12,620 --> 01:14:16,760 what about if it's referencing some oracle, referencing 1630 01:14:16,760 --> 01:14:19,640 some price source, and there's an error. 1631 01:14:19,640 --> 01:14:21,440 I believe, as I understand it, they're 1632 01:14:21,440 --> 01:14:24,300 writing some of that into the code as well. 1633 01:14:24,300 --> 01:14:26,630 So if it's referencing a price off the New York Stock 1634 01:14:26,630 --> 01:14:30,080 Exchange, how it does that, what open API 1635 01:14:30,080 --> 01:14:34,190 it references, how it pulls it. 1636 01:14:34,190 --> 01:14:38,240 And they're trying to say, let's make 1637 01:14:38,240 --> 01:14:40,253 sure that it's what's really verifiable. 1638 01:14:40,253 --> 01:14:42,170 There's one more slide on this that will help. 1639 01:14:42,170 --> 01:14:43,220 This is their steps. 1640 01:14:43,220 --> 01:14:47,000 This is their publicly disclosed five steps. 1641 01:14:47,000 --> 01:14:48,560 This was in an October paper. 1642 01:14:48,560 --> 01:14:50,930 It wasn't part of your reading. 1643 01:14:50,930 --> 01:14:53,330 But select the parts of derivatives contracts first. 1644 01:14:53,330 --> 01:14:54,622 Automation would be worthwhile. 1645 01:14:54,622 --> 01:14:55,372 They're effective. 1646 01:14:55,372 --> 01:14:56,000 It's efficient. 1647 01:14:56,000 --> 01:14:58,170 Let's find something we can automate. 1648 01:14:58,170 --> 01:15:00,980 Ross, that's their key driver. 1649 01:15:00,980 --> 01:15:06,680 Then try to express it the legal terms in a more formalized way. 1650 01:15:06,680 --> 01:15:09,870 Get some standardization about the law. 1651 01:15:09,870 --> 01:15:12,370 Get that be represented as functions, 1652 01:15:12,370 --> 01:15:14,860 meaning computer code. 1653 01:15:14,860 --> 01:15:16,620 So where can we be efficient and automate? 1654 01:15:16,620 --> 01:15:18,300 Where can we take the legal words 1655 01:15:18,300 --> 01:15:21,240 and get agreement, put it into code, 1656 01:15:21,240 --> 01:15:23,130 combine that into templates? 1657 01:15:23,130 --> 01:15:26,430 I think some of this is referencing oracles. 1658 01:15:26,430 --> 01:15:29,100 And then validate that the templates actually 1659 01:15:29,100 --> 01:15:32,400 work and, ultimately, even get some courts of law 1660 01:15:32,400 --> 01:15:33,960 to accept them. 1661 01:15:33,960 --> 01:15:35,670 That's their game plan. 1662 01:15:40,829 --> 01:15:41,407 Shimon? 1663 01:15:41,407 --> 01:15:43,490 AUDIENCE: I was thinking that there is potentially 1664 01:15:43,490 --> 01:15:45,800 huge implications beyond efficiency, right? 1665 01:15:45,800 --> 01:15:50,590 Because if you're at Treasury or Goldman Sachs 1666 01:15:50,590 --> 01:15:55,950 signs a contract with someone at Orange County, 1667 01:15:55,950 --> 01:15:57,690 their ability to assess what is actually 1668 01:15:57,690 --> 01:16:03,840 written in the 30-page contract was minimal or will be minimal. 1669 01:16:03,840 --> 01:16:08,430 But it's a lot easier to do risk management or assessment 1670 01:16:08,430 --> 01:16:12,660 if you could just use it as an API. 1671 01:16:12,660 --> 01:16:14,730 And you have an outside company's 1672 01:16:14,730 --> 01:16:16,700 basically providing you with the risk analysis. 1673 01:16:16,700 --> 01:16:17,627 Right? 1674 01:16:17,627 --> 01:16:18,960 GARY GENSLER: I agree with that. 1675 01:16:18,960 --> 01:16:21,190 We're still very much at the early days. 1676 01:16:21,190 --> 01:16:24,210 But I think this is an interesting, real, live thing. 1677 01:16:24,210 --> 01:16:27,100 There's a lot of money behind this. 1678 01:16:27,100 --> 01:16:30,260 The markets are hundreds of trillions of dollars. 1679 01:16:30,260 --> 01:16:33,540 If my former colleague and ISDA pulls this off, 1680 01:16:33,540 --> 01:16:36,000 they'll probably drive some efficiency automation. 1681 01:16:36,000 --> 01:16:38,820 They might lower some counterparty risk and event 1682 01:16:38,820 --> 01:16:41,880 risk or crisis risk. 1683 01:16:41,880 --> 01:16:44,640 I think you could embed these into true blockchain 1684 01:16:44,640 --> 01:16:45,540 technology. 1685 01:16:45,540 --> 01:16:48,750 But I'm not sure it's dependent upon it. 1686 01:16:48,750 --> 01:16:51,840 The automation of a smart contract. 1687 01:16:51,840 --> 01:16:54,330 Next Tuesday, next Tuesday there's 1688 01:16:54,330 --> 01:16:55,560 a couple extra readings. 1689 01:16:55,560 --> 01:16:56,870 We swapped out some readings. 1690 01:16:56,870 --> 01:16:58,140 It's trade finance. 1691 01:16:58,140 --> 01:16:58,980 Thank you. 1692 01:16:58,980 --> 01:17:02,330 [APPLAUSE]