distributed Financial Market Infrastructure (dFMI) and the Disintermediation of Digital Assets
Can Blockchain really make a difference in regulated financial markets?
A major appeal of Blockchain, more generically described as “Distributed Ledger Technology” (DLT), from the perspective of financial market participants has been the promise of simpler and less risky operational processes. But this is only a part of its full potential and misses an even bigger story: how DLT will enable a big disruptive shift in the shape of financial market structure.
Financial Markets are currently intermediated by Financial Market Infrastructures (FMIs) or Financial Market Utilities (FMUs). These infrastructures exist with the express purpose of providing shared processes that are explicitly designed to reduce cost and various types of operational complexity and risk.
This is the first of a series of posts where we outline a vision of a “Distributed FMI” (dFMI), a vision that goes beyond the narrow, technology-centric goals of operational efficiency improvements that are found in many DLT initiatives.
dFMI is a multi-disciplinary construct that combines distributed protocols and cryptography with economic mechanism design to re-imagine how the functions performed by FMIs today can be replaced with infrastructure that is not concentrated in FMI intermediaries. Rather, the power is with the participating members/users of the infrastructure, effectively returning to the mutualized ownership structures that were prevalent as recently as 30 years ago. dFMI will help contribute to a financial system that is more resilient, competitive and anti-fragile.
What are FMI’s?
According to the US Federal Reserve, FMIs are
“..multilateral systems among participating financial institutions, including the system operator, used for the purposes of clearing, settling, or recording payments, securities, derivatives, or other financial transactions.”
These systems can include:
1. Payment Systems
2. Central Securities Depositories
3. Securities Settlement Systems
4. Central Counterparties
5. Trade Repositories
The purpose of FMIs has been to facilitate the efficient interaction of financial market participants. They enable the settlement of contractual obligations (trades) of multiple counterparties via a single entity rather than many bilateral relationships. In classic economic terms, they reduce transaction costs between participants.
Why do Financial Markets Need FMIs?
The current banking system was built during the 18th Century industrial revolution to facilitate commerce at long distance. Back then, banks used an account-based methodology, based on double-entry bookkeeping, to facilitate this activity. They were trusted by their clients and each other to support payments for business and retail clients. All of these bank relationships were bilateral.
As Capital Markets have developed, the processes used for industrial commerce were adapted to facilitate the clearing and settling of financial products. In effect, the account based system for commercial banking was adapted to intermediate financial assets.
However, the development of a set of processes for finance based on a system designed for trade naturally creates weaknesses. And as gaps in the banking system appeared (either in the form of a newly required service, a hitherto unmitigated risk, or a means to mutualise a shared cost) FMIs have grown to fill them. The first such examples were Clearing Houses that were first formed in the 19th Century to reduce the cost, settlement risk and operational risk of financial transactions.
Starting in the 1960’s a regime shift came in to play driven by the advent of computing power. The overall process was called “dematerialisation” and involved developing a legal and operational framework that took advantage of the efficiencies offered by the client server architecture available at that time. This process involved not just the digitalisation of double entry bookkeeping, but also legal innovations, such as the “intermediated security” where a security is held in the name of an intermediary on behalf of the beneficial owner. It created a new market structure and ensured that account based intermediation was built into the technological infrastructure. It was a digital plan that made sense back then. And it cemented the role of the centralised FMI.
Outside the world of finance, modern digitalisation is not about automation of existing processes as it had been in the 80s, 90s and early 00s. Digitalisation, as practised in Silicon Valley today, is about taking “Digital Assets” and developing business models based on those assets that could not have existed in the analogue world. These new business models improve the customer experience by delivering the service where and when the consumer needs it, effectively shrinking the distance between consumer and producer by eliminating high cost low value-add steps. In the world of Finance, as Chris Skinner notes in many of his blogs on Digital Banking:
“Digital is a complete transformation of the fabric and foundations of the (banking) organisation.”
New financial assets, such as cryptocurrencies, are being created in native digital form; they have no paper legacy and they allow peer-to-peer transactions at distance. Given these innovations, the relevant question today is what form should a modern FMI take? We can be sure we wouldn’t create account based intermediaries if we were digitising paper-based proprietary claims today as these intermediaries would not be required to solve the dematerialisation problem of the late 1960s.
Understanding Financial Intermediation
Banks, brokers and fund managers are all intermediaries. Individuals might hold cash or investments via any of these directly. Many of us have indirect relationships too; our pension fund is an intermediary, holding these assets on our behalf. In either case, there is an account with that intermediary and our asset is recorded on the ledger of that institution.
When it comes to settling our financial transactions we need to trust these institutions to carry out that transfer of value. And they bilaterally reconcile their accounts to prevent double spend. The asset becomes an “intermediated asset” where our intermediaries exchange those assets in their own names on our behalf.
To transfer the value, behind your account is an asset custody and settlement process. This will normally require an FMI to intermediate the relationships between multiple banks, brokers and so on. The value transfer happens via debits/credits across the books of the FMI: account based intermediation. If you make a payment, in the UK that might go through the Faster Payments infrastructure or CHAPS, the Real-Time Gross Settlement System (see image below). If you are a customer of Goldman Sachs in the US, your Apple, Facebook and Amazon shares will be held by Goldman at the DTCC.
The use of FMIs has often, though not always, resulted in efficient processes which support very high transaction volumes and the exchange of large transaction values. Achieving those economies of scale has meant that there are few rather than multiple FMIs. In one country, there might be one or two payments mechanisms and one Central Securities Depository. Numbers vary, but the fingers on one hand are always enough to keep track.
There is an asymmetry to the overall picture; many banks, many brokers, many fund managers trade freely with each other in a peer-to-peer or bilateral pattern. But they settle those trades via a few FMIs; a highly concentrated hub-and-spoke picture, which creates problems of its own. The BIS-CPSS, which defines the high-level Principles for Financial Market Infrastructures to which national financial regulators conform explains:
“..while safe and efficient FMIs contribute to maintaining and promoting financial stability and economic growth, FMIs also concentrate risk. If not properly managed, FMIs can be sources of financial shocks, such as liquidity dislocations and credit losses, or a major channel through which these shocks are transmitted across domestic and international financial markets.”
Thus, intermediated assets have the side-effects of: (1) creating a pattern of intermediation that is highly tiered and concentrated; and (2) creating inter-intermediary exposures that can amplify systemic risk in the event of an intermediary’s failure.
In simpler language, this is conducive to monopoly economics and concentration of risk. Both of those effects are mitigated through more regulation, capital requirements, and other legal and (often manual) operational devices that increase complexity and cost in what is, in effect, a very humble function: the legal transfer of an asset from Alice to Bob.
Distributed Ledger Technology and Financial Intermediation
Although existing FMIs do drive operational efficiency, their ability to radically overhaul business processes has been limited. With FMIs so critical to financial stability, they are mandated to be risk averse. And this risk averse mindset has driven a particular architectural thinking: automating the individual steps in pre- existing manual processes rather than refactoring processes with digital methods that would not have previously been possible.
As discussed earlier, Financial Market Infrastructure was based on a client server architecture where each intermediary effectively has its own database. The problem created by this architecture is the need to agree the state of each of these databases at all the points where the intermediaries “hand over” to each other. DLT is based on the idea that separate copies of a database that are synchronised to always eventually agree the same state: a shared ledger. In this model, processes can share the same data and be guaranteed that the state of the database is agreed. This eliminates the need for multiple copies of the same information to be stored in individual silos leading to a massive reduction in the need for reconciliation processes.
This view of DLT focuses on the complexities that arise from the inevitable reconciliation between all those different intermediary accounts (and their mirror accounts in the systems of the asset owners). But this is a misdiagnosis of the problem; as if account intermediated digital assets are foundational and we just need to upgrade the technology behind the construct. It ignores the potential of the distributed technology to disrupt the market structure: the business models, processes, and legal form of digital assets.
Lets go back to the physical processes we are trying to mimic. Alice transfers value to Bob. Ideally, Alice would transfer to Bob peer-to-peer. Today’s processes require Alice to instruct her intermediary to pay Bob. Alice’s intermediary will find Bob’s intermediary and potentially use another intermediary, an FMI, to transfer the value.
And why was intermediation created? To make more efficient value transfer across distance and time in pre-computer times and to take advantage of technology when computing first became available. In the shared ledger view, “the ledger” would replace the ledger of account intermediary. So that hub-and-spoke graph we spoke of earlier remains, with all the monopoly and systemic risk problems that go with it. In effect, we have created a logically centralised database and given all the participants a real-time view of it. We still have not enabled Alice to directly transfer to Bob.
To borrow Vitalik Buterin’s taxonomy, it’s a strategy that pursues technological decentralisation with political (ie legal/commercial) centralisation. Vitalik describes this as “architectural decentralisation”. (In a future post, we question whether DLT can even achieve full technological decentralisation. The practical compromises needed to integrate with legacy account intermediaries have the tendency to introduce Single Points Of Failure (SPOFs) and Single Points Of Trust (SPOTs) that remove most of the benefits of technological distribution.)
So how can we get rid of the hub and spoke model? In the Silicon Valley jargon, what is the new digital business model? As information technology moves away from client-server to peer-to-peer and consensus-based networks, it’s time to question whether digital assets need to be “intermediated assets” at all. The answer is to take inspiration from the world of cryptocurrencies and use the tokenized forms of financial assets on DLT. In this new model, the graph of settlement relationships starts to look identical to that of trade relationships.
The Distributed Financial Market Infrastructure
In the Silicon Valley vision of digitisation, analogue processes are reimagined in new ways from previously unavailable digital assets with the purpose of shrinking time and space between service and the customer. In the financial markets, time and space had already been shrunk through the physical process of account intermediation and this process was automated in the second half of the 20th century via the commoditization of information technology. DLT allows us to return to the original tokenized form of financial assets removing the need for inefficient, somewhat risky and expensive settlement intermediaries.
Thus, in our vision, dFMI is more than just a narrow exercise in creating financial market operational efficiency by creating a global ledger for these financial intermediaries. While this is a worthy goal, it does not either need the DLT solution nor utilise the full power of it.
We believe that the dFMI construct requires an approach that adapts the multi-disciplinary blockchain philosophy of “economic mechanism design” to the FMI problem domain and offers a powerful alternative to the incrementalist DLT proposition.
Our starting point in this post is a recognition that dFMI should be based on digital assets in tokenized form, where no intermediation between the sender and receiver of the asset is required. The consequence? dFMI has no balance sheet containing customer value transfer transactions. Future posts explore the issues around DLT interoperation, technological decentralisation, privacy, network economics and governance for the dFMI context.
Rhomaios Ram, Head of Services & USC Consortium Executive, Clearmatics
Tweet us @Clearmatics