Decentralized Derivatives Clearing Part 2
Note: (This is part 2 in a series exploring Autonity’s Decentralized Derivatives Clearing model. Autonity is a public, EVM based, proof-of-stake blockchain for decentralized clearing of smart derivatives contracts. Find out more)
Contract Specs and Execution Modalities
In OTC markets, people often want to trade contracts where certain parameters (e.g., start date, end date, etc.) can be customised. So the OTC market standardises (the body that defines these standards is the International Swaps and Derivatives Association “ISDA”) contract specifications, but these specifications contain parameters that are customised by the trading counterparts at trade execution time.
With OTC derivatives, trades are negotiated over message or Request-for-Quote (“RFQ”) platforms¹, where liquidity takers can request quotes from liquidity providers (“LP’s”) over a platform-specific protocol. Customised parameters are (optionally) included in the quote request and sent to one or more LPs, who reply with a quote. The taker has a few seconds to hit the bid, lift the offer or pass. This provides for a sort-of-but-not-really “electronic trading” experience. Alternatively, trading can even be done over the humble telephone.²
In contrast, the parameters of an XTD (exchange traded derivative) derivatives contract spec must be identical. This is because the execution modality of an XTD is a Central Limit Order Book (“CLOB”). A CLOB is a multi-unit double auction and as such assumes that all of the units bid and offered are homogeneous. Therefore, contract specifications are not only standardised, their parameters are also fixed by the exchange.³
The OTC market’s one-to-many modality of trade execution contrasts sharply with the many-to-many modality of a CLOB, but on a bit of reflection, it is clear that contract customisation demands an execution modality that the exchange model cannot provide.
But this is not the only key difference.
Decoupling trading from clearing
The key observation to make about this one-to-many execution modality is not the execution modality per-se, but the decentralized interface for trade submission to the clearing process. In OTC derivatives there is no single trade execution venue⁴ tethered to the post-trade CCP or trade compression service, and there need not be a venue at all. What matters is that Alice and Bob agreed to a trade and have submitted it for clearing. How and where they traded is irrelevant.
The contrast with XTDs here is profound. With XTDs all trades are executed through the exchange’s CLOB. Trading is of course architecturally centralised⁵ but it is also logically centralised in that there is one and only one trading venue. XTD’s exist within systems that vertically integrate the trading venue with the clearing mechanism. To take a random example, the CME’s E-mini S&P 500 contract for September 2023 settlement (code ESU3) . Alice and Bob can only trade this contract on the CME’s ESU3 CLOB, nowhere else.
As one can see below, in DeFi most so-called “decentralized derivatives” protocols also logically centralise their trade execution, emulating the vertical integration of trade execution and clearing that is found in XTDs. It would seem that these projects view the non-custodial and permissionless access to their system as a defining feature of “decentralisation”. For Autonity, these are merely necessary conditions. The Autonity project works from the premise that a derivatives market is only decentralized if Alice and Bob are free to choose how they discover each other and trade.
Logically Decentralized Derivatives Trading
Logically decentralized derivatives trading means that Alice and Bob are free to trade their contract wherever, however they want and subsequently submit their trade execution for clearing. So why is it the case that XTDs do not work like this? To use the example above, why isn’t the market structure one where multiple trading venues list ESU3 and submit their trade executions to the CME for clearing? These alternative venues could be CLOBs with different features, or entirely different execution modalities.
Trading wants to decentralise
There is nothing intrinsic to the contract specifications of XTDs that results in this monopolisation of liquidity in a single exchange’s order book. In fact, contract standardisation lends itself to venue competition and heterogeneity. This is the phenomena we see with most asset markets, equities and crypto being the most prominent examples. If you do not want to trade AAPL on NASDAQ, you can go to any number of alternative lit or dark trading venues. And of course BTC or ETH can be traded on hundreds of different crypto exchanges around the world.
Trading wants to decentralise for a variety of (often inter-related) reasons. One reason concerns venue reach. Each trading venue can only directly reach a subset of the participants who want to trade the thing. There are also differing preferences of market participants over execution modalities and features. Traders of large positions are often concerned about the toxic order flow of High Frequency Trading (“HFT”) in lit venues and prefer certain dark pools of liquidity to minimise their slippage, for example.
The business model of trading venues is matching user orders according to some execution modality and collecting fees on transactions. An ecosystem of competing, differentiating venues for trading the same thing is a stable equilibrium and a natural state of affairs in free and competitive markets.
The Autonity project works from the premise that a derivatives market is only decentralized if Alice and Bob are free to choose how they discover each other and trade.
Centralised clearing wants to integrate
But the business model of CCP derivatives clearing is a natural monopoly. A contract specification cleared via CCP A is by definition a different product from the same contract specification cleared via CCP B. Competition and economic efficiency will drive all clearing to a single CCP because participants want to trade the contract that has the largest netting set.⁶
It is not difficult to see that the enterprise value of a derivatives CCP is greater if it vertically integrates a CCP monopoly with a monopoly trading venue. If derivatives are implemented via legal contracts instead of smart contracts on a public blockchain, clearing eligibility is necessarily a permissioned arrangement. So the CCP’s power to permission trade submission to itself is used to achieve vertical integration: the CCP simply refuses to clear any contracts that are traded on a venue not owned and operated by itself.
This is indeed what can be observed in practice with XTDs. Most derivatives CCPs are owned by the same group that owns the trading venue.⁷ For example, CME Clearing is the CCP⁸ that clears every contract on all the XTD trading venues operated by CME, CBOT, NYMEX, and COMEX⁹. All five entities are owned by CME Group.
Disruptive startups in the TradFi market have tried to break this vertical integration and bring about logically decentralized trading of XTDs. For example in the 2010s ELX Futures adopted an “open access” model where trades executed on its platform could be cleared at CME Clearing. They supported trading on popular bond and interest rate products offered by CME and introduced new spread execution modalities, different fee structure etc. This was done through an arrangement known as “Exchange of Futures for Futures” (EFF). The EFF arrangement would have allowed market participants to transfer their positions from CME to ELX and vice versa, while still clearing those positions through CME Clearing.
But the CME Group was predictably unhappy with the arrangement and used its control over clearing eligibility to prohibit the clearing of ELX trade executions.¹⁰ 12 years on, EFF is still not a feature of XTD market structure. Such is the power of the TradFi CCP monopoly.
(This is the end of Part 2, to read part 3)
Footnotes:
¹ API’s to these platforms are typically limited in functionality (if they exist at all) as the platform operator is in persistent fear that price-discovery will get commoditised by standard protocols and order aggregation platforms.
² A typical exchange might go like this. Alice “I’d like an all-or-none quote for XYZ with features A, B and C, in 100m size”; Bob “For that i’m 55 bid at 57”; Alice “Done. I pay 57 for 100m”. Bob (and maybe Alice) will have a valuation model that prices the contract on the inputs A, B, and C. These models are calibrated to the underlying’s market data (and benchmark derivatives quotes too, if dealers stream those).
³ There is also a philosophical difference between XTD and OTC contract specifications. XTD specs are basically a proprietary financial product offered by the exchange. OTC specs are basically open standards and protocols that market participants are free to adopt and adapt.
⁴ A “trading venue” refers to a system where financial things are bought and sold. It’s a regulatory concept and there are different types of regulated trading venues: Exchanges, ATSs, ECNs, etc. Historically, RFQ systems were not considered trading venues and fell outside the regulatory perimeter. But it looks like this is changing. See Clare (2023).
⁵ The exchange’s CLOB is client-server, there is no zk-rollup here!
⁶ You can see this dynamic at play in the few cases where competing futures exchanges compete for market liquidity with identical contract specifications in the same underlying. For example, the 3M Euribor interest rate futures on ICE vs the 3M Euribor future on Eurex (Bloomberg tickers: ER1 Comdty and FP1 Comdty, respectively). The ICE contract won an effective monopoly over 3M Euribor liquidity with an open interest thousands of times greater than the thinly traded Eurex contract that had an identical specification.
⁷ And where there isn’t group ownership of the two, there is typically an exclusivity arrangement between the CCP and the venue.
⁸ The regulatory designation in the US for a derivatives CCP is called a Designated Clearing Organization (“DCO”).
⁹ The regulatory designation in the US for an XTD derivatives trading venue is called a Designated Contract Market (“DCM”).
¹⁰ One can only assume that CME’s motivation was competitive, although the CME initially cited regulatory compliance as the motivation for the prohibition. But this reason was invalidated when the CFTC later issued a statement saying that ELX’s EFF model was indeed compliant with regulation.
