21. Post Trade Clearing, Settlement & Processing

The following content is provided under a Creative Commons license Your support will help MIT OpenCourseWare continue to offer high quality educational resources for free To make a donation or to view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at ocw.mit.edu GARY GENSLER: Today’s a little bit of a discussion about clearing and settlement systems Maybe you feel you’ve heard everything you needed from Jeff Sprecher when he was asked a question when he was with us Jeff runs some of the largest clearing houses in the world at Intercontinental Exchange not only because he started in the energy business, but he has big credit default swap clearing houses and energy clearing houses and the like And he said, well, I don’t know about blockchain technology for clearing and settling But I’m going to try this Bitcoin project that you all know about and he talked about But today we’re going to try to dig into, well, maybe there’s something there, because there are projects going on And why are there projects? They’re mostly permissioned and not permissionless systems But I think there’s something there This will be one of the more optimistic– if you’re judging how often I get to the maximal side, I think there is probably something there, at least for permissioned blockchain projects in the clearing and settling I’m going to– we’ll, of course, go through the readings And these are the sort of topics we’ll go through Clearing and settling services– I’m going to try to set this stage as to what this is Frankly, when I was on Wall Street for 18 years, I couldn’t really define what clearing and settling Sorry, Ross You thought I could define it when I was on Wall Street? What’s that? AUDIENCE: I just strictly know someone else can’t either GARY GENSLER: Yeah Yeah Well, when you’re in the merger and acquisition area, you usually don’t think about what it means to clear something and to settle it And frankly, even usually when you’re on the trading floor, that’s something somebody else takes care of Just the nature of– how many people have worked on trading floors here? All right So I’m going to have seven or eight of you start And you’re on notice now In about eight to 10 minutes, you’re going to help You might prove what I just said to Ross, that you can’t define what clearing and settling is because somebody else took care of it But we’ll see We’ll see And then we’ll talk about blockchain technology applicability, some of the projects that are going on, and then move over to this There was a reading from the International Swaps and Derivatives Association, ISDA, what they’re doing around smart contracts in this space But what I want to do first is a little bit of talking about final projects, because we’re at that time of the year And this is to help Then I’m going to go through eight or nine questions If you haven’t already thought through your projects, these are eight or nine questions that I would think you’d want to think about, which is another way of saying it will probably enhance your performance What else can I say? And number one– and it’s a little, small print I’m sorry Or no, it’s not too bad– is what’s the value creation proposition, or pain point, if you wish So are you creating value? Are you solving a pain point? Or as Jeff Sprecher would say, what are you doing that’s going to make something cheaper, faster, better? What’s the value proposition? Whatever space There’s two or three groups that are doing trade finance There’s one or two doing payments There’s some group doing commercial– I mean, consumer credit rating or so forth There’s a wide variety There’s one or two groups that have come to see me that are doing things outside of finance But regardless, what’s the value proposition? What pain points are you addressing? Or as Jeff would say, what are you going to do better, faster, cheaper? If you can’t answer that threshold question, shake it loose Kick it around Or move on to a different use case, frankly Secondly, [INAUDIBLE] you get into, like, what is this blockchain technology? What transactions or data is going to be recorded? Because this is fundamentally, ultimately, a database technology, even the permissionless systems And you’ll see a later question is why this versus

traditional database system? But what transactions, what data– whether it’s health care records, whether it’s consumer credit data, supply chain, moving a piece of property along a supply chain, but what is the actual data and transactions that you’re recording? Questions? What multiple stakeholders need to write and read access to the ledgers? And I underline the word multiple Because really, when you think about the value proposition of blockchain technology, it’s about immutable records, or so-called immutable records That’s verifiable, immutable records amongst multiple parties And it cost extra money Everything we’ve been studying about in this semester is about there’s more complexity There’s more cost, in some way, of this decentralized, even permissioned systems, decentralized system But you get something for it You get some verification of transactions, of data, of mutability But you also have to have some logic that there’s multiple parties that need to both write and read A distributed database that just does read only– I can read my Bank of America bank account I can see what’s in that sort of ledger It doesn’t have to be on a blockchain So multiple parties writing, why is that important? Forming consensus, in this circumstance Which specific cost of verification or networking are you lowering? We talked about a lot of the verification costs, whether it’s privacy, censorship resistance, whether it’s some direct cost Maybe you’re going right at something that’s got a bunch of economic rents in it So there was a series of verification costs we talked about What networking costs are you trying to reduce? There’s probably been something in the order of 4,000 or 5,000 white papers, many that you can find You can’t find all of them But if you’re going at something that you think a white paper has been written about or a blockchain company has been started on, figure out what the competition is doing And for the two groups, and it may be more, that are doing something around trade, finance, and supply chain, you know there’s a bunch of stuff you can read We’ll be studying it next Tuesday Alin? AUDIENCE: Can you give us an example of a networking cost? I’m not sure what that is GARY GENSLER: So networking cost is often why people are using a native token But networking costs are how to jumpstart a network Uber jumpstarted a network, in a sense, without blockchain technology But an example of networking costs are if you walked out of the airport and you didn’t think any drivers are using this application, then you wouldn’t be terribly interested in using it So– AUDIENCE: But who incurs the cost? GARY GENSLER: Often the cost of building a network can be a tremendous startup cost So how to start a network often is you go to a bunch of venture capitalists You raise money You advertise And you try to get participants And one of the challenges of health care records– we don’t have our health care experts in here today I was looking around We have two students from Harvard’s public policy school doing graduate work and sometimes join us But health care records, for instance, there’s some really interesting blockchain technology solutions But the networking challenge is how do you get a bunch of medical providers, hospitals, doctors, and users to actually join the network So that would be an example of network cost is basically adoption often Does that help? AUDIENCE: Yeah GARY GENSLER: What are the competitors doing? I’m not looking for plagiarism But if you’re in a space, payments or supply chain or consumer credit or some others, where there’s three or six or 10 white papers out there and maybe somebody has even raised 20 million or a couple million in an ICO, then it would be interesting You should feel free to look at those And maybe you’ll be informed by it And you’ll say, yeah, I’m doing it better Or they’ve got it poorly here Or their project’s in Estonia And I’m going to do a digital ID project in the US And our regulatory system is different than Estonia But I wouldn’t be bashful about looking at the competition In fact, I would say that’s– we’re a business school

You always look at the competition And you see if you can do either what they’re doing better or what they’re doing in a different geography or in a little bit different market Back to how Alin likes to define things, I’ve not used the term blockchain technology yet on this slide, have I? But why are append-only logs at multiple party consensus the best solution? Which kind of goes back to why not just a traditional database But why append-only, timestamped, verifiable logs? They’re connected by cryptographic hash functions But why have these logs? And why have multiple parties in essence form a consensus about the state of the ledger? Alin? AUDIENCE: Yes I have a question about the timestamped, apend-only logs Is that only relevant in blockchain, that traditional databases do not have these kinds of features, timestamp and append-only logs? GARY GENSLER: They can have time– timestamps in essence In essence, it’s innovation of the early ’90s And then Satoshi Nakamoto grabbed it It was the concept that you take a block of data and you connect it to the next block with a cryptographic hash function And once you’ve done that, then it’s committed You can’t amend anything that’s before because it’s a unique hash And traditional databases did not have similar commitment schemes They do have other commitments schemes, but not this commitment scheme of taking a block of data, having the hash function, which then means never shall you amend what came before That’s how I understand it Did you have a different view on traditional databases? AUDIENCE: Well, I guess this gets to the root of the question Is there something new about blockchain technology? In my opinion, there isn’t, actually 20 years ago, we knew how to make a traditional database and make it full of power and make it a census protocol, or what people nowadays like to call a permissioned blockchain So we knew how to do that And that would give you an append-only log It’s nothing exceptional GARY GENSLER: Yeah So there are debates as to where is the border and boundary between a permissioned blockchain and a traditional database But I’ve chosen to, at least for this class, a hyperledger fabric, we’ll call that a blockchain project But recognize there are some who would say, well, that’s– unless you have a native token, it’s not really blockchain Eric AUDIENCE: –one could call point recent days that the trust is the key attribute that’s not the factor in this equation, because you can get a multiple– I mean, multiple party consensus doesn’t have value on its own It’s actually multiple party consensus is– it’s actually something you had to come up as a necessity that stems from the fact that you’re dealing with a decentralized arrangement in which you’re not trusting any of the parties that intervening in the exchange of information Right? GARY GENSLER: Well, I agree with the first half, that you wouldn’t even want to use Hyperledger Fabric unless you really need multiple parties writing to the ledger And I think that does separate it a little bit You could take an Oracle database and allow it to be multiple parties, too But I think that separates it You might be more the purist than I am, Eric I don’t think it has to be completely open, where you don’t know all the participants But even in a financial– we’ll talk about the Australian Stock Exchange in a half an hour or so And when we talk about that, that’s a limited closed loop But they believe that they are actually taking a lot of cost out of the verification side They think they’re getting a lot– they’re lowering verification costs And they would contend they’re lowering networking costs, because they don’t have the same reconciliations between multiple databases Now, you might call that verification cost But I think they would call it both verification and networking cost There’s multiple back offices no longer reconciling If your final project has a permissionless system with a native token, why do you need a native token? Is it to jumpstart the ICO that you want to raise money for? I’m not averse to that

But at least fess up and say that’s how we’re crowdsourcing this thing If it’s about that you really believe there’s some benefit to the token economics, that it’s an incentive system to run your new competition to Uber or an incentive system for health care records or an incentive system for supply chain, talk about what that incentive system is And are you trying to motivate jumpstarting the network or maintaining the network later? And it could be both It could be either I’m just saying it would be good to have a discussion Is the native token just about crowdfunding? Is the native token also about jumpstarting this network and then an incentive to maintain the network amongst multiple parties? What trade-offs of scalability, performance, privacy, security, coordination? I’m not asking you to solve that And in fact, you can even jump ahead three and five and seven years and assume that some of these things will be taken care of So recognize that in the permissioned blockchains, they seem to have a higher performance, a lot higher performance, than the native token permissionless systems But if you’re proposing a native token permissionless system, I’d just like a short discussion of how much these scalability issues matter to you, particularly when most of the native token permissionless systems can’t handle much more than 1,000 or a handful 1,000 of transactions a day It is true that Bitcoin handles quite a bit more than that But five or 10 transactions a second, is that going to work for whatever your solution is? In fact, I’d like to see some groups use native tokens I want to have a variety of things I read in December rather than reading all permission systems And then again, can a permissioned blockchain or traditional database adequately address whatever use case if you do have a permissionless system? Thinking about that, how can broad adoption be realized? Are you going to use a native token to equip broad adoption? How are you going to get a bunch of hospitals signed on? How are you going to get a bunch of banks signed on? You’re going to do like Jeff Sprecher did and give 50% of his equity away There was that conversation he talked about he had with Goldman Sachs, that he realized that was the only way to get his business started in the late 1990s But just how are you going to get that adoption? And to the extent, this might be a paragraph, but what type of customer interface or how might it be better than other customer UIs? So these are just some of the things, if you were thinking about a real business And I think that the market is moving away from that big, frothy bull market, where there was as many as 400 or 500 ICOs a month And it was raising anywhere from 2 to 5 billion a month October was about 200 ICOs And it was about 800 million And I think it’s going to continue to decline, partly because they don’t all work But partly because people are now getting more rigorous as to what’s really going on If you are using a permissioned blockchain, you don’t have to tell me whether it’s Hyperledger or Corda But if somebody does and you have something thoughtful to say there, I’m intrigued to find out why this one versus that one It’s really just to show your critical analysis of this subject And if you want to prove yourself on why it’s Corda versus Hyperledger Fabric, that’ll be interesting if somebody has a point of view But I’m not looking for that level of either computer science knowledge This is from a business perspective Think of me as a venture capitalist If I was investing, how you pull it together So any questions on that? So we had six or seven people that actually worked in banking Who did we have at this table? Ah So what’s clearing? AUDIENCE: If I recall correctly, you need to prove the identity of the other party in order to make sure that you’re sending the asset to the right– GARY GENSLER: All right So you have to prove the identity So it’s about validating and authenticating an identity Anything else in clearing? Who else worked in trading? I had somebody back at the back table worked in– ah AUDIENCE: Is it that they have the assets to be traded? GARY GENSLER: They have the assets to be traded Well, most equity markets trade on transaction date

plus two days So somewhere in that two days, you have to confirm they have the asset That is correct You don’t have to do it on t plus zero, normally by t plus 1 AUDIENCE: I think– I could be wrong But I think you had to do as well would be the stock exchange knowing what is the net exposure that you have at the [INAUDIBLE] GARY GENSLER: Net exposure Net exposure Very key thing So there’s authentication At some point between execution and transaction, you have to actually know you have the security– and then net exposure, which we’re going to look at on another page This simple chart of two people cutting the lawns, but a buyer and seller meet But a buyer and seller meet in an exchange in traditional securities markets through brokers That’s called an inner mediated access In the crypto exchange markets, those two brokers on the left-hand side and the right-hand side, the sellers and buyers brokers, don’t exist In the crypto exchange markets, basically you have direct access to the market But in the traditional stock exchanges, whether it’s London, Tokyo, New York, Beijing– where’s the exchange in India? AUDIENCE: Mumbai GARY GENSLER: Oh, it’s at Mumbai now? I forgot 30 years ago, I visited it and walked on the floor It was called the Bombay Stock Exchange back then But regardless of where the stock exchange is, there’s some clearing house separate from the stock exchange doing these functions that we just talked about So three steps– execution What’s it mean to execute a trade? Alpha, do you know what it means to execute a trade without reading? Should I go back? AUDIENCE: Yeah Because I’ve traded before GARY GENSLER: Yeah That’s why I called on you AUDIENCE: We did a great job with a separate execution team That helped us with that But signing the docs, making sure that the account information is all there, if it’s been done through NSDA GARY GENSLER: All right So this sounds like it’s about derivatives, executing a derivatives contract But what if you’re just buying 100 shares of Apple stock? What’s it mean to execute a trade? AUDIENCE: To be on the phone or on the screen and agree on a price and a quantity GARY GENSLER: There you go It’s just you bid a price And somebody hits the bid or lifts the offer Those are the terms when you’re on the phone I left your offer I hit your bid Or at least it was when I did it Still? Or no, maybe not What’s that? Still All right Question? AUDIENCE: Just not on the phone anymore [LAUGHTER] GARY GENSLER: Ah, yeah That’s it What’s that? AUDIENCE: You just hit a button GARY GENSLER: Just hit a button AUDIENCE: I worked on the [INAUDIBLE] desk at Goldman So we didn’t talk to anyone GARY GENSLER: You didn’t talk to anybody? Can you go months without talking to anybody? AUDIENCE: No, we talked a little GARY GENSLER: What’s that? AUDIENCE: We talked a little bit [? But it was always ?] [? kinda ?] weird GARY GENSLER: All right I guess I’m a dinosaur Yeah Yeah The screens that I used to look at at Goldman Sachs, it was green Yeah, yeah, yeah You know You know the screens Kelly is thinking, what is he talking about? Right? Look it up on the internet [? You’re ?] [? fine. ?] [? You’re ?] [? fine. ?] There wasn’t any color on those screens So execution is basically I buy, you sell, we agree upon a price, a price and a volume That’s really what execution is But when you execute, you still need to do a few other things And it’s called clearing and settling When I was on a trading floor, I couldn’t have told you what the difference was between clearing and settling That was something somebody else did, frankly I’ll admit it They were really terrific professionals But they had their job I had a different job Clearing is performing all of these functions Authenticating somebody is there, making sure you actually have those securities, and importantly this thing called netting How about settlement? Anybody want to say? I know it’s written up there But anybody want to just say what– we’ve talked about ledgers for seven or 10 weeks now What do you think a settlement really means in the context of a ledger? Did I see a hand up here? Yes, I did I did see a hand up What’s settlement? AUDIENCE: Running the entries on the ledger What’s in, what’s out GARY GENSLER: Right It’s making a final recording on a ledger, finally moving something on a ledger

You could be moving the ownership of an asset or the ownership of cash And when you do both at the same time, it’s called delivery versus payment, or DVP It’s just a final change of data record on a ledger In the old days– I’m talking about before I was doing it But 40 years to hundreds of years ago, it was physically delivering a physical piece of paper which was a equity security or a bond security But it all got dematerialized starting in the 1970s And by the 1990s, it was pretty much– in this country and the UK– it was pretty much dematerialized securities, that everything’s digitized now in the equity markets in nearly every country that you’re all from and maybe with some small exceptions So execution is I buy, you sell, volume price We’ve legally agreed to something verbally And billions of dollars can trade on the phone or in the computer Legally, it’s executed Clearing is trying to make sure that you actually have the securities, and I actually have the cash We do some netting We authenticate And settlement is actually moving the data on the ledgers Three steps This is the benefit of netting It was just one chart that I could find somewhere It’s actually the Australian Stock Exchange had it on their website But if you have multi-parties in the transaction, which was on the left hand side– and these are just arrows of all the different transactions that could occur in one little example– and instead you interpose a central intermediary, you can do a lot of netting And look at participant C. Participant C only has an exposure of selling 35 of whatever they’re selling, some stock, versus all these other transactions And this is a very simple case But if you have 20 or a hundred clearing members that are all big brokerage houses and they’re doing thousands or even hundreds of thousands of trades a day, you have an awful lot of transactions And the gross transactions versus the net is quite a difference That netting can bring down numbers 90 plus percent or maybe even 99% or more So central intermediaries– I know this goes against Satoshi Nakamoto’s whole concept– but central intermediaries came about in the securities market for a bunch of reasons But one of the biggest ones was netting Brotish? And it’s a really important concept in the back office of Wall Street AUDIENCE: I’m not familiar with the concept So [INAUDIBLE] GARY GENSLER: You’re saying you’re familiar? AUDIENCE: I’m not familiar with this concept, so a clarification question GARY GENSLER: Please AUDIENCE: When we do netting, it seems that we are losing a lot of information here from the gross transactions So is it– GARY GENSLER: That’s right AUDIENCE: –just the transfer of the asset is done through the netting, but the records of all the gross transactions are still there with the central intermediaries? GARY GENSLER: Very good question So are we losing information? Or are we just lessening the lines of counterparty risk in the transactions? And the answer is, if done properly, you still keep the information of all of these And in the real world, this is just eight transactions It’s netted down to four Right? So this is eight transactions netted down to four Usually this is millions of transactions netted down to tens of thousands or maybe millions of transactions literally netted down to, if there was 100 clearing members, it would be 100 transactions, because they’d each have one net transaction per security, per security And so you need to keep the information of the millions of transactions to properly record for your customers, for your taxes, for all sorts of other reasons But you’ve netted down for counterparty reasons Does that help? Question over here Sean AUDIENCE: I was just curious It’s very understandable for equities But for principal commodities, [? the full ?] commodity derivatives, for instance in the clearing space, does it actually involve the physical delivery of the netted commodity [INAUDIBLE]?? Because there’s a different type for delivery, and how that–

GARY GENSLER: OK So Sean’s asked the question, what if you’re not dealing with digitized assets, but physical assets? I think You used the word commodity But I was presuming you meant physical So some of the innovations of the 19th century relate here when they started to have standardized contracts for commodities that were held in warehouses In essence, to facilitate some of this netting was also why central warehouses, this clearing house in the middle in physical commodities, could be a warehouse that held the gold, held the silver, held the corn or wheat And if they’re holding that corn or wheat, they started to have specifications as to what standards there were In gold and silver, it was percent It was called the, let’s see, assay– I’m trying to remember the word– but what percent gold or silver it is But in corn and wheat, there was a more detailed what’s going into the warehouse to commoditize it so that it was a standard contract If it’s truly physical and not commodity, it’s much harder Then you can’t really net down my house versus Shimon’s house They’re two different houses Does that help? But literally by the late 19th century, even in the commodities space running out of Chicago, they figured out central clearing houses with a lot of counterparty risk So there’s an economic reason we have centralized clearing houses And the biggest economic reason is netting And netting lowers something called counterparty risk Instead of having literally millions of transactions, you’ll have thousands And you’ll have less counterparty risk It also lowers one other thing So it lowers counterparty risk And what is the other thing you think it lowers? Anybody? AUDIENCE: Transactions cost GARY GENSLER: Transactions cost because it’s efficient We used to call it you take a lot of lines off the page So in the old human days, you’d have to have fewer humans actually pushing paper Hugo AUDIENCE: Another clarification question So from the participants side, so from a hedge fund side or whatever or a trading desk side, do they just go about their day trading everything? And then the next day the clearing house tells them, OK, you made all these transactions yesterday But we really only need 35 [INAUDIBLE] Or how does it [INAUDIBLE]? GARY GENSLER: In that case, looking back to this, you are either the buyer or seller And you’re dealing with a broker So if you’re a hedge fund, you often enter into an arrangement with a broker And it’s called a prime brokerage arrangement And you might have a prime brokerage arrangement with more than one Wall Street firm But you rarely would have it with more than three So even if I’m doing a trade through Merrill Lynch, if my prime brokerage arrangement is at Goldman Sachs, Merrill is going to get a commission But they’re going to give up my prime brokerage arrangements to Goldman Sachs And that helps me, me the hedge fund, get more netting as well And that’s a development in the last 20 years, this concept of prime brokerage But to answer your question, yes I go about my business I trade all day And unless I get a phone call or an electronic communication from my prime brokerage saying you’ve hit your lines, you do have credit limits and position limits in essence But your prime brokerage– and I might say, well, I need something moved over I need some more credit today But yes You go about your business until your prime brokerage says slow down Shimon? AUDIENCE: But [INAUDIBLE] question says that if you’re a lot like LTCM– GARY GENSLER: Long-Term Capital Management AUDIENCE: –at some point, they were running a balance sheet of a trillion dollars And part of the reason why, when bankrupt is they didn’t want to use one prime broker, because they don’t want to reveal all their trades But then their prime brokers could not net out their positions, because I may hold one leg with you and another leg with you The net may be zero But there’s no way for you to actually reconcile that or reconcile that risk So it’s a real problem GARY GENSLER: Long-Term Capital Management was a hedge fund operating in Connecticut It was founded in the early 1990s by a man named John Meriwether, who is a very famous trader from Salomon Brothers

and very successful And then he started it with some remarkable finance faculty from here, at MIT, and elsewhere– Bob Merton, who is a Nobel laureate Brilliant, brilliant team of folks But they ran into a little challenge four years later They had about $125 billion balance sheet, $4 billion in capital, and 1.2 trillion in swaps Why do I know the numbers? I got the phone call with my one-year-old sitting on my lap And I picked up the phone And I said, hello? And the voice on the other line said, Treasury operator I’ve got the Treasury secretary for you And I get on the phone I go, yes, Bob? It was Bob Rubin It was the Treasury secretary Bob, what’s up? It was a Saturday Isabel was on my lap He says, do you know anything about Long-Term Capital Management? He goes through some of the figures I’ll never forget He says, I hear from Alan– that would be Alan Greenspan, who was the head of the Federal Reserve– I hear from Alan that they have over a trillion dollars of derivatives That’s Shimon’s number He says, is that a big number? Well, you have to understand, Bob’s sense of humor was quite– at that moment, he knew it was a big number So I flew up the next day And with somebody from the Federal Reserve Bank in New York, Peter Fisher, we sat around the conference table looking at these facts and figures And then I got back on the plane and flew back to Washington to be with my family The next day was actually Rosh Hashanah, the Jewish holiday And I wasn’t going to work And I had to call Bob and his deputy, Larry Summers, up and give them my feedback And I said, this thing’s going down Bear Stearns was their prime brokerage And Bear Stearns, who 20 years later failed themselves, but Bear Stearns said, we’re pulling the prime brokerage account if you can’t post another half a billion dollars by Wednesday Well, there was no way they could post another half a billion dollars, because their $4 billion of capital we could best figure on that Sunday was somewhere covering around 400 or 500 million at that point in time Well, when you have a lot of leverage, you have $1.2 trillion of derivatives which are all sort of leveraged in one way or another and you’ve got $125 billion balance sheet, even with the best minds, the best finance faculty– and they really are terrific individuals And Meriwether was quite a trader Sometimes markets go against you Correlation risk, liquidity risk, and so forth So they pulled the plug the next day, basically That’s the Long-Term Capital Management story I remember But it’s just because I kind of had to live it But to answer your question, you can trade all day until your prime broker says, Hugo, slow down, because they’re extending credit to you– the prime broker is– intraday And one of the big risks, one of the big risks in the financial crisis– and certainly Ben Bernanke, who is chairing the US Federal Reserve and others had to deal with it in the middle of 2008– is most commercial banks, so they don’t have a broker They’re banks They’re the brokers themselves– were extending significant amount of intraday credit to each other that, in 2008, wasn’t terribly well measured in the repo markets and in other markets And that’s changed subsequently But you wouldn’t have been a big bank You would have probably been shut down Sorry Or maybe not So netting is the biggest benefit of clearing But it’s also efficiency and taking down a lot of operating lines of code This is a chart I’m not going to spend a lot of time on it But this is right off the Australian Stock Exchange website And I thought, well, why not take that? They have 2 million investors, the market participants or brokers in a much smaller market than the US And Australia has 77 brokers, registered, regulated brokers And then they have the clearing house called CHESS, C-H-E-S-S, which I can’t quite remember what it all stands for But SS is Settlement System Clearing house– AUDIENCE: [INAUDIBLE] GARY GENSLER: What’s that? What’s it? AUDIENCE: Clearing House Subregistered System GARY GENSLER: Subregistered System Thank you Is it Equity for E? AUDIENCE: There is none GARY GENSLER: There is none But they have the Australian Stock Exchange Clear and Australia Stock Exchange Settlement, because they actually put it in two different places, two different things The clearinghouse is doing what we talked about–

mostly netting, lowering the lines of code, and then settlement And basically, because we’re in a trade day plus two days, on the first day is all the clearing and the netting And you better show up with your securities And on the second day is settlement Electronics and technology could make this simultaneous But most market participants like to have some ability to do gross trading against net’s positions So there’s a lot of economics and market practice and structure behind keeping some delayed settlement And the markets have grown up around that Ross AUDIENCE: If it’s not delayed, if it was instantaneous, wouldn’t that get rid of the benefits of net? And you’d have to have every trade go right through So it has to be that you wait until the end of the day or whatever period [INTERPOSING VOICES] GARY GENSLER: So Ross just raised the number one reason, economic reason, why simultaneous clearing and settling is not that likely to happen The markets have built themselves around the benefits of netting and also stock loan, the ability to borrow that security somewhere But you’re absolutely right If there were simultaneous execution and settlement, which is what the Bitcoin network in essence is supposed to be and blockchain networks can be, then you, in essence, have to prefund Every time Hugo trades, he’s going to have to provide his security to trade But then Tom would get the security faster It would change the entire prime brokerage model It doesn’t mean markets, Ross, couldn’t adapt and be structured that way But it would be fundamentally a different structure Is this is a hand up? AUDIENCE: So the [INAUDIBLE] is really big with derivatives I don’t see it I don’t see what we use– I don’t see it being a primary issue with equities So if you were going to apply this to equities, trading, any cash instrument basically, I don’t see the huge economic value for netting GARY GENSLER: The huge economic benefit for netting currently is a lot of operational efficiency AUDIENCE: Sure But if you’re going to place that with some– GARY GENSLER: No And then also, because there is two day counterparty risk, you are lowering that two day whatever measure equity volatility, two day equity volatility risk, usually not collateralized AUDIENCE: I’m just commenting on whether this– you said the price that you pay for replacing this system with the blockchain based one would be that you cannot net up GARY GENSLER: I should have said something in addition And the market participants would have to then prefund their trades I think the larger thing– so if I speak in too much financial technology, let me say current markets are you can sell or buy a security without owning the security or having the cash You can buy something and not have the cash You have to deliver the cash in two days And usually by that evening or the next day, you have to show you have the cash You can sell a security without owning the security And you only have to really kind of deliver it, generally speaking, the next day, on t plus 1 If you went to a true simultaneous clearing settling and execution market, you’d have to have the cash and the security before you execute And I think that’s the larger piece of it, much larger than netting AUDIENCE: I understand the cash But you need the security, right? Because if you’re going to buy it– if you sell it and you don’t have it, then you know it’s going to take two days to get one to cover your first GARY GENSLER: So through your prime brokerage arrangement– Hugo, remember, has a couple He’s a big hedge fund And he has what’s called a prime brokerage account Not at Bear Stearns any longer, but maybe it’s at Goldman Sachs He’s kind of fancy He’s getting that Then his prime brokerage account, in his legal contractual arrangements with them, they will find that security for him, maybe from some other customer

Maybe that security is held in somebody else’s what’s called a margin account And they will enter into what’s called a stock borrow But when he executes on the phone, he doesn’t technically already have that security He could be shorting it And his prime brokerage account could then go find it from the Goldman Sachs network AUDIENCE: But it’s not because you have plus two But it’s because you have a broker GARY GENSLER: Yes That’s right, combined with t plus 2 Let’s take this And then I’m going to move on AUDIENCE: Can you just take that one step further and explain why that’s beneficial broadly versus just the system we’ve grown accustomed to? GARY GENSLER: It’s a really good question Is this just the system we’ve grown accustomed to? Or is it beneficial? It’s hard to tell It is definitely the system we’ve grown accustomed to over a couple of centuries And as we’ve gone from transaction date plus five, which most markets were I’m talking about 50 and 100 years ago, maybe it was even t plus 10 at some point But t plus five to t plus three to now t plus two, we’ve started to challenge that But in the US, the Depository Trust Corporation provides same day settlement And very few accounts ask to do same day settlement And I think it might be just they’ve grown accustomed to it But I think there is some benefit in this stock, loan stock borrow side But I stand with what I said earlier I think the market could work the other way But it would be a big adjustment And there are some benefits, because then Hugo, this hedge fund, is being extended credit by his prime brokerage account And the prime broker is making some money on it But also Hugo’s taking risk in that arrangement Prefunding it would be a different model Some argue it would have less liquidity in the marketplace It would make shorting stocks a little costlier I’m not sure they’re right But their argument is that I’d have to borrow the securities always, 100%, and halve it off before I sell, even if I’m a multi-billion dollar hedge fund that has some credit worthiness Kelly AUDIENCE: So the paper mentions that their new system fails to basically recognize these challenges, such as the fact that it’d be advantaging the capital and operational efficiencies But it goes on to say that it would be a step backwards for the world’s most liquid market So what would necessarily be implications for illiquid markets? Would it basically be that people can make trades without having to interface with extensions of credit? GARY GENSLER: Are you talking about illiquid markets that are centrally cleared or illiquid markets that have no central clearing? AUDIENCE: I guess maybe both I’m just trying to understand what the counter would be, that saying that there would be advantages, efficiencies in liquid markets But what with the opposite be? Would there not be implications for the [INAUDIBLE] markets? GARY GENSLER: So Kelly’s asking– I don’t remember the comment, because it was in the Australian Stock Exchange one So in non-cleared markets, things that are just so illiquid– my house It’s not cleared It’s not centrally cleared Certain parts of the derivatives markets aren’t centrally clear They’re not, quote, “liquid” enough So I don’t know I don’t know the comment Maybe see me in class And I’ll look at the comment and try to nail it for you Is that all right? So I wanted to put this chart up, because it was from 2011 I remembered it when I was at the CFTC But somebody at the Fed of Chicago made this chart about the US clearing systems Yellow are all clearing houses Red, the SEC, the Federal Reserve, New York– AUDIENCE: SBD? GARY GENSLER: Yeah SEC, CFTC, and the Federal Reserve But they might have– New York I don’t even know who New York SB– AUDIENCE: New York State Board of Derivatives GARY GENSLER: Yeah Somebody from the Chicago Fed was trying to make the case that our regulatory system in the US was a little bit complicated And I think this was in the midst of the debate about Dodd-Frank, that they thought all oversight of clearing houses should move to the Federal Reserve But I remembered the chart and found it And I wanted to share it with you But the yellow are all the clearing houses There are a lot of clearing houses in the US There’s one for bonds There’s one for equities There’s multiple ones for derivatives, both from cash and interest rate derivatives to energy derivatives to credit default derivatives and so forth There’s ones for options

So I just wanted to lay it out Finance is pretty complex This is all digitized, by the way But in this complexity, that’s also where I think there’s some opportunity, whether it’s permissioned blockchain and not fully permissioned list with tokens, and inspired by– the green are exchanges And I don’t even think they put a color up for brokers, did they? Because then you’d have hundreds There are multiple parties sharing ledgers and reconciling their ledgers Reconciliation means I keep a ledger Tom keeps a ledger It’s both on our database systems And we keep testing and ensuring that we have the same records This is a system that, by its complexity and by its copies of ledgers and reconciliation, has a lot of opportunity for greater efficiency And the question is whether blockchain technology, maybe permissioned, can bring some of that efficiency That’s the thought And that’s what the chart was just trying– by the way, the Congress didn’t agree with the Fed They left the authorities at the SEC and CFTC And those are for stories other times I personally felt it could be a shared authority And we ended up with some shared authorities But I thought that the CFTC and SEC have domain knowledge And it really important to keep the SEC involved in securities clearing and the CFTC involved in derivatives clearing for that domain knowledge, but that the Federal Reserve also came in under something called Title VIII of Dodd-Frank And they also got authorities So we made it more complicated, because now there’s multiple dotted lines for this The Federal Reserve basically has an account with everybody Or they have some oversight of everybody So then the question on blockchain technology, could you take the clearing house out of the middle and do the thing on the right? And my question for you all is, do you think the economics– I’m not talking about the technology I’m not talking about whether it’s performance If you wish, you can go five or 10 years down the road and say it’s got all this scalability and performance Do you think the technology, without something in the middle, what’s the economics about that? Anybody? Eilon? AUDIENCE: I still want to support the left side for liquid markets because of the netting GARY GENSLER: Because of netting AUDIENCE: Yes I can’t see a blockchain solution working well when you need netting to enjoy the efficiencies of selling and shorting, shorting basically GARY GENSLER: Yeah But the question really is, can you somehow, on the right with shared ledger, still get some benefit of netting? Right? You’re saying you’re inclined here, because you want to get the netting for economic reasons So as a ledger structure, can you get the netting here? Shimon? AUDIENCE: You don’t need the netting if it happens simultaneously Netting is it it is basically– again, with cash instrument I’m not talking about derivatives Derivatives is a different issue With cash instruments, the only reason why you have counterparty risk is because it’s not clear– it was my understanding Correct me– is because you don’t have instantaneous settlement If you have a [INAUDIBLE] settlement, then you don’t care about netting AUDIENCE: You care because you face so many transaction fees AUDIENCE: No But that’s going to be a lot more efficient So why– again, do you think the [? real ?] transaction fees are a big economic issue here? AUDIENCE: Yes That’s what I think GARY GENSLER: And I’m sorry that you don’t AUDIENCE: So now it’s fine I don’t know I mean, I don’t know For really intense help, how much do you pay per transaction with the clearing house and without? Maybe the net will give you only 10% of the transaction But the cost of having the clearing house is more than 10 times AUDIENCE: Maybe AUDIENCE: So the economics depends on those numbers, I think GARY GENSLER: Is that a question? Hugo? AUDIENCE: Yeah So you still can get some of the netting Let’s just talk about blockchain transactions on Bitcoin You can batch transactions So you can have multiple inputs and multiple outputs And that’s kind of like netting Then you have to have a second system on top of that that keeps track of what’s going on throughout the day And then it just does the transactions on a daily basis or an hourly basis or whatever That takes care of that netting Right? GARY GENSLER: So Hugo contends maybe

you can get this economic thing called netting and still have a shared distributed ledger Or you could do it off chain I think, Shimon to your question, this is a big economic thing either because we’ve just grown accustomed to it over a couple centuries or it’s fundamental to how markets trade, that we don’t prefund It’s that economics that blockchain technology will challenge But we might stay with a central clearing house if the economics are that value Well, the economics of, in essence, not prefunding, and thus, some delayed settlement Whereas in the past, we always had delayed settlement at first because there was so much paperwork We had to have delayed settlement We went digitized And now, really in 2018, we don’t need delayed settlement or batch processing So now you’ve exposed, well, wait What are the economics? Tom and then Alin? AUDIENCE: Could you still have a delayed settlement with a blockchain system just by using a smart contract? GARY GENSLER: You could Absolutely AUDIENCE: I mean, you still keep this system that people are accustomed to, but do it without the economic red for the clearing house’s sides or [INAUDIBLE] GARY GENSLER: Yes And it’s what the Australian Stock Exchange says they are using with the digital asset company, Blythe Masters’ company that’s their outside vendor using the HyperLedger Fabric And they’re using the code They call a DAML, or Digital Asset Machine Language? I don’t know what the M is But they call it DAML Alin? AUDIENCE: Yes We’ve said you cannot prefund on the right Is that correct? GARY GENSLER: I think you could through smart contracts I think you could find this But the challenge with just the picture, the picture suggests there’s no central counterparty to do all the netting And so the question is, can you have this ledger structure and still find a way to net down all the positions? AUDIENCE: Well, like Shimon said, there is no need for that You get that implicitly by recording everything and executing GARY GENSLER: But if you’re recording– all right I’m agreeing with that, too But that’s where I get to if you need to do Shimon’s approach, which is simultaneous execution and settlement, then you need to prefund, even if that prefunding is by nanoseconds But you need to prefund AUDIENCE: Not necessarily You could, sort of optimistically, you could post a transaction that says, I want to buy a stock And I’m not going to prove to you that I have money to buy it You record it, if you were saying optimistically And later on– GARY GENSLER: It depends The word record– let me use a different word, settle Settlement means final legal rights have moved from one party to another So the word settlement has a meaning that I just want to come back to It means I’m recording on a ledger that I no longer have legal rights So this alpha has legal rights to it And I think you would have to prefund if you were doing true settlement ledgers AUDIENCE: Sure, sure I guess I’m saying it seems like you could simulate the clearing house In the right side of this picture, the clearing house just distributed using the consensus protocol GARY GENSLER: Right I think you’re right that you could But if you’re going all the way to where Shimon is, like Bitcoin, and you’re doing final settlement, final settlement recordation, then I think you have to prefund A couple of questions over here Eric, Ross AUDIENCE: The reason that the prefunding was [INAUDIBLE] performed in the validation of the block process, because– GARY GENSLER: It could I said it could even be done during the nanoseconds But it has to be done before you move the settlement And that’s what’s called clearing So it’s a question of whether you have simultaneous execution and settlement, or you still have execution, clearing, settlement, even if separated by seconds, but still literally get granular I apologize This stuff is that granular Back here And then I’m going to move on AUDIENCE: Why do you consider prefunding– maybe you don’t But it seems simply you consider prefunding before you’re allowed to transact a negative economically GARY GENSLER: I shouldn’t sound that way It’s just that for a couple years, we’ve had delayed settlement And so I’m saying if you are a proponent of a system that

has simultaneous execution and settlement, you just have to deal with it You’re changing the economic models that have been so common in the capital markets And when you take that on, you just have to think through what are the economics of that shift, from in essence a delayed settlement, thus modest but real counterparty risk that Hugo’s got a prime broker and he’s got that, and moving to this new model in capital markets So I shouldn’t come across as negative But I’m saying at least it’s a hurdle that you’ve got an adoption curve that’s going to be a challenge, I think, at a minimum DTCC, when I first spoke to them– and I’ve had a lively, ongoing dialogue with the folks there on an occasional basis This was the biggest issue they raised They said they looked at blockchain technology starting several years ago They talked to their customer base And they couldn’t find that their customer base really wanted to go to simultaneous execution and settlement They have same day settlement But it’s end of day settlement So there’s still several hours But they said, we don’t have a lot of customers that really want simultaneous execution and settlement Now, that was commenting on the demand side It’s another question whether you could adopt this lower cost, do something, and change your capital market So this is another look at it This is also from the Australian Stock Exchange, I think, website if I remember But maybe it was another website The question is, could you take out these middle parts? This is like blockchain architecture for clearing central counterparty and the central securities depository Because there’s also somebody who’s got the central registry, and that central registry are the two things Could you take this out completely? That was the question, in a sense So here are the projects that are actually live or being explored And we’re going to talk a little bit about Australian Stock Exchange in the few minutes we have left But this Australian stock Exchange Project started in 2016 A bunch of requests for proposals, like higher digital asset holding, they’re actually actively working on it It’s a 25-year-old system And Jeff Sprecher said it, too He said, you know, they had a legacy system they needed to replace They decided to go with the new technology of the day But Jeff wasn’t convinced they had to But they feel that it was the right thing to do, to go with that new technology of the day They believe– and they’ve put out reports as recently as several weeks ago They believe they’re on the right path and that 15 or 20 years from now, every clearing house will be using what they would consider a permissioned blockchain solution, because it will so significantly lower the back office costs for their market participants who are now holding records Each bank, was it 77 members from an earlier– yeah Each bank has to hold records It could be held in a shared ledger And they would lower the reconciliation cost And they believe the second advantage is using smart contracts Now, are they exactly the smart contracts like on the Ethereum network? No, because they don’t have a native token They don’t have a DAP and a native token But it’s still automating certain processes, automating certain I’ll call it back office processes So they’re very committed to it And they’re going to roll it out in 2020 And then, for 12 months, run it simultaneous with their legacy system So their cut over for assuring that it’s resilient is at 12 month test Most major market infrastructures that roll things out do not roll simultaneously for 12 months But they all roll simultaneous Whether it’s a month, whether it’s 90 days, you can’t cut over the back office of the New York Stock Exchange and just test it for one day Or you can It’s just kind of choppy and a big risk AUDIENCE: Quick question Are they going to use a stable coin to settle the transaction? How are they going to settle the transaction? Because I can see how you issue securities digitally But you cannot issue cash GARY GENSLER: So this is very good It’s very good So all of these clearing houses, they are the final custodian for the security

Securities are dematerialized, meaning they’re all digital And so those digital assets are held on a ledger and what’s called the golden record This is the words used in the securities business The gold record or the golden record, meaning it is the legal, verifiable thing you can go to in a court of law And it says, this is who owns this Alpha owns it Not Hugo Sorry You’re working with Goldman Sachs, though, right? That golden record is there for securities Who has the golden record for cash? AUDIENCE: The central bank? GARY GENSLER: The central bank or the fractional banking system, their commercial bank subledger Clearing houses around the globe would like to be able to have direct access to central bank money And they currently do not in any of the major jurisdictions that I know I think So what Australia does now, this is what they do now They net everybody down They get to t plus two They’re on that second day Everything’s netted down And they’re about to do the whole closing out One second before they do the security settlement, they do the cash settlement They call this Delivery Versus Payment, DVP But they’re a clearing house And they do not want to take risk with the central bank They take all the money and settle all the cash with the Royal Bank of Australia, RBA I guess Do I have the right name? Or is it a different– is RBA one of the banks in Australia? But they settle all the cash And just within a second, they settle We think of it as DVP But it’s actually technically not Inside their clearing systems, they have a form of Australian dollars that is effectively digitally, in their system, digital money In essence, what they’re doing is clearing their internal digital money versus Royal Bank money AUDIENCE: Thank you GARY GENSLER: All right Sorry So these other projects– Estonia, the Tallinn Stock Exchange, is exploring proxy voting and registration What’s interesting– who runs the Estonian Stock Exchange? NASDAQ NASDAQ So NASDAQ runs it And they’re doing this project NASDAQ announced and went live in 2015 on an illiquid So Kelly, you asked me about illiquids earlier NASDAQ started a blockchain technology platform in 2015 on private securities, which were not– they certainly weren’t liquid I don’t know enough about what they did there as to how illiquid they were But they felt, all right, this is a place we could do this And it’s still gone live It’s got very little use But it’s still there And they say they’re going to try to do it for mutual funds, when you and I can just by interest in mutual funds And NASDAQ’s going to offer that to mutual funds, partly because mutual funds aren’t in DTCC from what I understand The others are just really some tests Japan Exchange Group for the Tokyo Stock Exchange has done some tests They’ve written papers They’re sort of interesting papers The last one was earlier this year They’re really trying it just on post trade matching One of the securities firms in Japan, I can’t remember, is really pressing the Tokyo Stock Exchange to do this So there’s some local commercial politics that are going on I can’t remember which exchange is– not exchange One of the brokerage houses is pushing them to do it All permissioned systems No, to my knowledge, permissionless systems Let’s talk about ISDA for a few minutes, derivatives common domain model One of my former commissioner’s colleagues, Scott O’Malia, runs ISDA That’s a international trade organization It’s largely, largely dominated by the 16 big banks around the globe that probably transact 98% or 99% of derivatives around the globe And it was started in the 1980s to do all the forms, all the forms to enter into the legal contractual obligations called swaps And over the years, it’s transitioned It’s also an advocacy group

And it lobbies various regulators in Europe and the US and Asia But they’ve, a year or so ago, said, wait a minute We could take a lot of the back office of swaps, particularly swaps that are not going into clearing houses So it’s called the non-cleared or uncleared swaps And maybe we can create smart contracts to automate a bunch of contractual terms So that’s the second bullet point And they contracted out They put out a request for proposal, basically, for somebody to come up with machine readable standard representation Call it code, a scripting language I can’t remember if it’s– oh, it’s written in JavaScript So written in Java and JavaScript, a standard machine readable code of various events If it’s a 20-year swap that’s going to pay every quarter 80 quarterly payments, each of those quarterly payments, fixed versus floating payments, is somehow put into machine readable code And Scott and ISDA’s view was, if we could get a broad consensus and publish that machine readable code that everybody agrees on, maybe that machine readable code could take the place of legal contracts ISDA, for 30 plus years, one of their lines of business is creating those legal contracts Maybe we can get a step ahead of this and get consensus amongst these 16 big banks There’s about 100,000 users of swaps when you say all the customers It’s not measured in the tens of millions I doubt anybody in this room personally has a 20-year or 10-year swap But I don’t want to embarrass Maybe somebody does But it’s an institutional product, largely That’s their product And it covers new transactions, rate resets, partial terminations, as I listed up there Questions? And it’s definitely very much inspired by Ethereum and smart contracts and DAPs But in essence, it’s not really blockchain technology It’s inspired by it, in a sense AUDIENCE: They’re just rewriting the contracts instead of– they think they can have more specificity and certainty of about what the contract means by writing it in contract GARY GENSLER: And importantly, importantly, efficiency– to automate that which humans are still doing, even if it’s in written contract So it’s also automation that’s inspiring them But you’re absolutely right It’s taking legal contract, getting into machine readable JavaScript code, getting consensus on what that code is, getting some courts to recognize it And then you won’t even need to write some of this, a rate reset or a partial termination, into the contract You’d be referencing this The longer term objective is to say maybe we can put even more of it right into code AUDIENCE: People would have to pool their money somewhere It’s like we were talking about, right? GARY GENSLER: That’s correct AUDIENCE: It’s going to execute itself, which is going to have to be a pool of money somewhere GARY GENSLER: That you might have to some prefund or self fund or something Let me go back here, and then up AUDIENCE: If they’re using blockchain on it, it can’t be– GARY GENSLER: They call it a blockchain project But it’s a– I think I’ve tried to capture the essence of what I think they have here It’s definitely blockchain inspired I’m just not sure it’s really stored on a blockchain AUDIENCE: Yeah Because if you look at credit default swap, you have to reference a credit rating somewhere digitally And wherever that’s referenced digitally, if someone makes a mistake, does a transaction period that’s based on that mistake and it’s fixed, someone’s going to have to be able to reverse that transaction GARY GENSLER: So what’s raised is what about if it’s referencing some oracle, referencing some price source, and there’s an error I believe, as I understand it, they’re writing some of that into the code as well So if it’s referencing a price off the New York Stock Exchange, how it does that, what open API it references, how it pulls it And they’re trying to say, let’s make sure that it’s what’s really verifiable There’s one more slide on this that will help This is their steps This is their publicly disclosed five steps This was in an October paper It wasn’t part of your reading But select the parts of derivatives contracts first Automation would be worthwhile They’re effective It’s efficient Let’s find something we can automate Ross, that’s their key driver

Then try to express it the legal terms in a more formalized way Get some standardization about the law Get that be represented as functions, meaning computer code So where can we be efficient and automate? Where can we take the legal words and get agreement, put it into code, combine that into templates? I think some of this is referencing oracles And then validate that the templates actually work and, ultimately, even get some courts of law to accept them That’s their game plan Shimon? AUDIENCE: I was thinking that there is potentially huge implications beyond efficiency, right? Because if you’re at Treasury or Goldman Sachs signs a contract with someone at Orange County, their ability to assess what is actually written in the 30-page contract was minimal or will be minimal But it’s a lot easier to do risk management or assessment if you could just use it as an API And you have an outside company’s basically providing you with the risk analysis Right? GARY GENSLER: I agree with that We’re still very much at the early days But I think this is an interesting, real, live thing There’s a lot of money behind this The markets are hundreds of trillions of dollars If my former colleague and ISDA pulls this off, they’ll probably drive some efficiency automation They might lower some counterparty risk and event risk or crisis risk I think you could embed these into true blockchain technology But I’m not sure it’s dependent upon it The automation of a smart contract Next Tuesday, next Tuesday there’s a couple extra readings We swapped out some readings It’s trade finance Thank you [APPLAUSE]