Streetwise Professor

October 12, 2016

A Pitch Perfect Illustration of Blockchain Hype

Filed under: Clearing,Commodities,Derivatives,Economics,Regulation — The Professor @ 7:31 pm

If you’ve been paying the slightest attention to financial markets lately, you’ll know that blockchain is The New Big Thing. Entrepreneurs and incumbent financial behemoths alike are claiming it will transform every aspect of financial markets.

The techno-utopianism makes me extremely skeptical. I will lay out the broader case for my skepticism in a forthcoming post. For now, I will discuss a specific example that illustrates odd combination of cluelessness and hype that characterizes many blockchain initiatives.

Titled “Blockchain startup aims to replace clearinghouses,” the article breathlessly states:

Founded by two former traders at Societe Generale, SynSwap is a post-trade start-up based on hyperledger technology designed to disintermediate central counterparties (CCPs) from the clearing process, effectively removing their role in key areas.

“For now we are focusing on interest rate swaps and credit default swaps, and will further develop the platform for other asset classes,” says Sophia Grami, co-founder of SynSwap.

Grami explains that once a trade is captured, SynSwap automatically processes the whole post-trade workflow on its blockchain platform. Through smart contracts, it can perform key post-trade functions such as matching and affirmation, generation of the confirmation, netting, collateral management, compression, default management and settlement.

“CCPs have been created to reduce systemic risk and remove counterparty risk through central clearing. While clearing is key to mitigate risks, the blockchain technology allows us to disintermediate CCPs while providing the same risk mitigation techniques,” Grami adds.

“Central clearing is turned into distributed clearing. There is no central counterparty anymore and no entity is in the middle of a trade anymore.”

The potential disruptive force blockchain technology could have for derivatives clearing could bring back banks that have pulled away from the business due to heightened regulatory costs.

I have often noted that CCPs offer a bundle of many services, and it is possible to considering unbundling some of them. But there are certain core functions of CCP clearing that this blockchain proposal does not offer. Most importantly, CCPs mutualize default risk: this is truly one of the core features of a CCP. This proposal does not, meaning that it provides a fundamentally different service than a CCP. Further, CCPs hedge and manage defaulted positions and port customer positions from a defaulted intermediary to a solvent one: this proposal does not. CCPs also manage liquidity risk. For instance, a defaulter’s collateral may not be immediately convertible into cash to pay winning counterparties, but the CCP maintains liquidity reserves and lines that it can use to intermediate liquidity in these circumstances. The proposal does not. The proposal mentions netting, but I seriously doubt that the blockchain–hyperledger, excuse me–can perform multilateral netting like a CCP.

There are other issues. Who sets the margin levels? Who sets the daily (or intraday) marks which determine variation margin flows and margin calls to top up IM? CCPs do that. Who does it for the hyper ledger?

So the proposal does some of the same things as a CCP, but not all of them, and in fact omits the most important bits that make central clearing central clearing. To the extent that these other CCP services add value–or regulation compels market participants to utilize a CCP that offers these services–market participants will choose to use a CCP, rather than this service. It is not a perfect substitute for central clearing, and will not disintermediate central clearing in cases where the services it does not offer and the functions it does not perform are demanded by market participants, or by regulators.

The co-founder says “[c]entral clearing is turned into distributed clearing.” Er, “distributed clearing”–AKA “bilateral OTC market.” What is being proposed here is not something really new: it is an application of a new technology to a very old, and very common, way of transacting. And by its nature, such a distributed, bilateral system cannot perform some functions that inherently require multilateral cooperation and centralization.

This illustrates one of my general gripes about blockchain hype: blockchain evangelists often claim to offer something new and revolutionary but what they actually describe often involves re-inventing the wheel. Maybe this wheel has advantages over existing wheels, but it’s still a wheel.

Furthermore, I would point out that this wheel may have some serious disadvantages as compared to existing wheels, namely, the bilateral OTC market as we know it. In some respects, it introduces one of the most dangerous features of central clearing into the bilateral market. (H/T Izabella Kaminska for pointing this out.) Specifically, as I’ve been going on about for about 8 years now, the rigid variation margining mechanism inherent in central clearing creates a tight coupling that can lead to catastrophic failure. Operational or financial delays that prevent timely payment of variation margin can force the CCP into default, or force it or its members to take extraordinary measures to access liquidity during times when liquidity is tight. Everything in a cleared system has to perform like clockwork, or an entire CCP can fail. Even slight delays in receiving payments during periods of market stress (when large variation margin flows occur) can bring down a CCP.

In contrast, there is more play in traditional bilateral contracting. It is not nearly so tightly coupled. One party not making a margin call at the precise time does not threaten to bring down the entire system. Furthermore, in the bilateral world, the “FU Option” is often quite systemically stabilizing. During the lead up to the crisis, arguments over marks could stretch on for days and sometimes weeks, giving some breathing room to stump up the cash to meet margin calls, and to negotiate down the size of the calls.

The “smart contracts” aspect of the blockchain proposal jettisons that. Everything is written in the code, the code is the last word, and will be self-executing. This will almost certainly create tight coupling: The Market has moved by X; contract says that means party A has to pay Party B Y by 0800 tomorrow or A is in default. (One could imagine writing really, really smart contracts that embed various conditions that mimic the flexibility and play in face-to-face bilateral markets, but color me skeptical–and this conditionality will create other issues, as I’ll discuss in the future post.)

When I think of these “smart contracts” one image that comes to mind is the magic broomsticks in The Sorcerer’s Apprentice. They do EXACTLY what they are commanded to do by the apprentice (coder?): they tote water, and end up toting so much water that a flood ensues. There is no feedback mechanism to get them to stop when the water gets too high. Again, perhaps it is possible to create really, really smart contracts that embed such feedback mechanisms.

But then one has to consider the potential interactions among a dense network of such really, really smart contracts. How do the feedbacks feed back on one another? Simple agent models show that agents operating subject to pre-programmed rules can generate complex, emergent orders when they interact. Sometimes these orders can be quite efficient. Sometimes they can crash and collapse.

In sum, the proposal for “distributed clearing to disintermediate CCPs” illustrates some of the defects of the blockchain movement. It overhypes what it does. It claims to be something new, when really it is a somewhat new way of doing something quite common. It does not necessarily perform these familiar functions better. It does not consider the systemic implications of what it does.

So why is there so much hype? Well, why was Pets.com a thing? More seriously, I think that there is an interesting sociological dynamic here. All the cool kids are talking about blockchain, and nobody wants to admit to not being cool. Further, when a critical mass of supposed thought leaders are doing something, others imitate for fear of being left behind: if you join and it turns out to be flop, well, you don’t stand out–everybody, including the smartest people, screwed up. You’re in good company! But if you don’t join and it becomes a hit, you look like a Luddite idiot and get left behind. So there is a bias towards joining the fad/jumping on the bandwagon.

I think there will be a role for blockchain. But I also believe that it will not be nearly as revolutionary as its most ardent proponents claim. And I am damn certain that it is not going to disintermediate central clearing, both because central clearing does some things “decentralized clearing” doesn’t (duh!), and because regulators like those things and are forcing their use.

Print Friendly

13 Comments »

  1. Excellent article.

    There is also an excellent Bank of England paper from 2007 all about “alternative” CCP models which is well worth a read. They model something called “ring clearing” which, from the limited information out there, sounds similar to what both SynSwap and the (currently mothballed) NetOTC solutions were envisioning. It’s well worth a read (no association with the authors, I’m just surprised it’s not had more air time!):

    http://www.bankofengland.co.uk/research/Documents/workingpapers/2007/WP321.pdf

    Comment by Chris Barnes — October 13, 2016 @ 4:06 am

  2. @Chris. Thanks for the kind words, and for reminding me of the paper. I read it when it came out. Hard to believe it’s been 9 years.

    Ring clearing was used at the CBT prior to the adoption of central clearing in 1925. It was an entirely voluntary arrangement. Every FCM had “ring clerks” who would get together and try to find trades that would “ring out”, e.g., A had sold to B who had sold to C who had sold to D who had sold to A. When a ring was found, the parties would negotiate to see if everyone wanted to eliminate the trades in exchange for payments of differences (i.e., payment of mark-to-market values). The arrangement was voluntary–if one party in the ring didn’t agree, the clerks would scurry around to try to form another one.

    Ring clearing is analogous to the multilateral compression service operated by TriOptima.

    I’m skeptical that SynSwap will do this, because it would require smart contracts that (a) reference other smart contracts, (b) give permission ex ante for SynSwap to tear up contracts among an unknown set of participants. This would essentially be granting SynSwap the right of offset, which would be interesting but (a) I doubt it does, and (b) that wasn’t mentioned specifically in the article.

    The ProfessorComment by The Professor — October 13, 2016 @ 10:24 am

  3. Indeed. The problem with smart contracts here are as you describe:
    – near instantenous propagation of shock waves
    – absolute trust in their correctness..

    The first means no firebreaks for any shocks, which, in something where failure may be truly catastrophic, is really bad.

    The second is dubious, because you can either have only very simple contract (where you can prove the correctness, so would have to be sub-Turing), or you can have complex contracts, but ones where it may be impossible to prove they are what you wanted (because they are fully Turing, and thus suffer from Goedel Theorem trap).
    Human languages, which is what contracts are written right now are fully GT – but everyone knows it (it’s the reason we have lawyers), so it’s not a problem.

    TBH, if I was a regulator, I’d be more worried about smart contracts than welcoming them.

    Comment by vlade — October 14, 2016 @ 2:26 am

  4. The question is whether disintermediation is needed or not. If it is for good and needed, blockchain is the ideal tech to address this need. Blockchain is yet to or just coming off the age and it will mature as the demand arise. Having said, blockchain in its current stage is very much ideal for immutable transaction capturing from multiple parties in a trustless environment. For now it is ideal to adopt blockchain as transaction capture layer and integrate with proven and time tested business solutions to achieve best of both.

    Comment by ravi jagannathan — October 15, 2016 @ 6:36 am

  5. […] article was previously published on the author’s blog and has been republished here with permission. Some minor edits have been […]

    Pingback by Cryptostyle | NEWS: A Pitch Perfect Illustration of Blockchain Hype — October 16, 2016 @ 8:25 am

  6. […] article was previously published on the author’s blog and has been republished here with permission. Some minor edits have been […]

    Pingback by Virtual Mining Bitcoin News » A Pitch Perfect Illustration of Blockchain Hype — October 16, 2016 @ 8:45 am

  7. You sound like the person saying postage is fine the way it is, e-mail is trying to reinventing the wheel.

    Comment by D.F — October 16, 2016 @ 9:13 pm

  8. I agree. The fact that a clearinghouse reduces counter-party risk is reason enough to continue using them. The people making this proposal do not appear to understand why a central clearinghouse exists.

    Comment by Dave D'Rave — October 17, 2016 @ 12:07 pm

  9. There’s a 2pm panel at the FIA expo on Wednesday discussing this very issue.

    Comment by Abe Froman — October 17, 2016 @ 3:08 pm

  10. @Abe-And I’ll be there! I’ll be chairing one panel, and participating in another. I’ll attend the BC panel.

    The ProfessorComment by The Professor — October 17, 2016 @ 6:50 pm

  11. @SWP…I have been meaning to report that I have finished “Fragile By Design”. What an excellent treatise on banking and banking crises. Calomiris and Haber detail the relationship of banks to government, and they use as examples of the Great Bank Bargain the historical development of banking in several different countries to show how the principles they describe apply.
    There have been books on the recent banking crisis, and there will be more to follow, but this is required reading for anyone who wants to begin to understand what happened about eight years ago.

    Comment by Richard Whitney — October 17, 2016 @ 7:11 pm

  12. SWP- When Accenture is pushing blockchain, you know it is peak hype- and I share your skepticism than blockchains can solve all business issues out there. I am sure Accenture was pushing ecommerce when Pets.com went public. Nonetheless, I think you are throwing out the bathwater in a domain that you are too knowledgable to excuse it. I will try to answer some of your detailed queries.

    Centralized vs. Decentralized. Since most traditional CCPs are mutualized, the risk warehouse (or capital buffer) is not centralized. Instead the default risk is borne by its eligible members after margin is consumed, so effectively the risk is decentralized to trusted and bonded participants. The CCP is not full decentralized like Bitcoin, but is decentralized amongst the biggest players. More legitimate blockchain startups in the clearing space such as Clearmatics and R3 have this model in place. The SPAN margining algorithms are centralized- or better yet standardized as a protocol. It is the margining arrangement that is centralized in a CCP, rather than the risk per se. The blockchain startups are the same way.

    Bi-lateral swaps vs. Cleared Swaps. There are many financial markets that do not have CCPs- albeit minor ones. The FX and IR markets being the largest global markets, the number of cleared trades is tiny compared to the non-cleared OTC market. Since most of these contracts are delivered financially- and easily cash settled, the dealers themselves or Prime brokers play the role of netting and calculating margin. High liquidity and real-time calculation make this suitable for these markets, and a traditional CCP just adds cost and regulatory uncertainty. CCPs instead are highly utilized in markets where the delivery is often not cash settled, such as equity and commodity markets. CCPs provide enormous benefits in these markets, particularly credit support for hedge funds and non-rated entities. But CCPs come with a cost, regulatory burden and lack of innovation. If decentralized blockchains can create an immutable ledger to safely trade (from a credit perspective) stocks or commodities 24/7, in any geographical area, then liquidity will gravitate towards that solution rather than the old man’s trading house- also known as the CCP. The profit that ICE and CME generate from clearing far and away exceeds profit they generate from trade matching or other pre-to-post trade services. The blockchain will disintermediate the most profitable part of exchanges- the clearing house.

    Today, clearing house must apply for clearing approval for each particular contract.For example, if they have a derivative contract that settles heating oil futures in NY Harbor, they must reapply for another Heating Oil contract that settles at Houston Ship Channel, and then rules around each settlment and delivery date. Each month CME and ICE add dozens of new contracts for clearing after they get regulatory approval- a huge bureacratic queue. Instead- a very simple algorithm could describe all heating oil price risk, and a participant could define the delivery location, a specific settlement date, ticket size and specs. This is how the OTC market works already for bespoke derivatives, and there is nothing preventing a clearing operation from providing this simple service- other than the regulatory hurdles of being an offical CCP. You don’t need the blockchain to do this, but CCPs are generally dinosaurs in risk measurement, and the demand for clearing complex derivatives far exceeds supply at the given time.

    Also you did not touch on this, the blockchain can provide enormous cost savings for post-trade services. Tabb and others estimate that the error rate in derivative trading exceeds 10% of tickets- and settlement often takes 2-10 days. Mostly trading firms now have back-offices with a headcount 2-3x their front office staff. Blockchain can eliminate those staff and the errors, and create enormous savings for trading firms.

    The FU option in the OTC market is not one to be underestimated- and I think we will see some CCPs go bust in the next financial crisis. After Lehman’s collapse the OTC market continued to operate normally Smart Contracts and real-time clearing and settlement on the blockchain do not offer this flexibility (FU Options), but if they are de-centralized clearing firms, hopefully the losses will be more widely spread. You might be interested in reading about the DAO on Ethereum blockchain this past summer, for a drama filled Sorcerer’s Apprentice.

    In sum, I think a blockchain clearing startup can potentialy provide clearing services for 1) less than 1/100th of current cost and will less errors; 2) offer clearing 24/7 for all types of financial players- like the decentralized FX market that exists today; 3) can provide clearing for thousands of derivatives that are not eligible today; 4) provide a decentralized capital buffer that makes systematic risk much less than a traditional CCP.

    Don’t let the blockchain hype chafe you. It is revolutionary technology like the TCP/IP protocol of the internet. It will not fix most things in the business world- and Accenture will provide zero value add- but IMO centralized derivative clearing has numbered days, like the trading pits of yore.

    Comment by Cypriot — October 19, 2016 @ 7:24 am

  13. @Cypriot FX is not decentralized. CLS Bank and CLS members are playing the role of a CCP wrt netting.
    The financial world will always be regulated and regulation is not compatible with decentralized immutability.

    Comment by GlmWgr — October 20, 2016 @ 12:27 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress