Data. In the past few years, the concept has become central to almost every industry thanks, in no small part, to the EU’s landmark GDPR legislation.
Concepts like data protection and data scientists are, by now, familiar concepts to most. More esoteric ideas like “data hygiene” and “the data exhaust” may become so as industries develop ways to monetise the data they produce.
The capital markets industry is no exception. The primary bond market has raised almost $5.5tr in 2019 for over 4,000 borrowers from around the world, from huge sovereign multi-billion-dollar benchmarks all the way down to private placements of a few million. The processes that power this industry have been incrementally improved for years, but in many cases, are due to be rebuilt from the ground up to the latest technological standards.
Robotic syndicate bankers
Artificial intelligence or, as it is more properly called in its current form, “machine learning” is on track to turn dumb capital markets deal records into smart, monetisable data.
By feeding in 20 years of primary bond market data, machine learning models could soon make recommendations of the ideal timing, price, size and maturity for a particular issuer during any given market conditions.
“With data-driven models, we can predict with round 80%-90% accuracy, the make-up of an order book,” says Michael Spitz, head of Main Incubator, Commerzbank’s research and development branch.
Obviously, the system works best for frequent borrowers, on whom the most data are available.
For Spitz, the system is about levelling the playing field for banks such as his. “The fact is that the biggest banks have the most data and the most information on which to base the recommendations they make to issuers,” says Spitz. “How do smaller banks with less information compete? They do it with better analytical tools.”
Structuring data
When it comes to streamlining the execution process, the appropriate buzz term is “data harmonisation” with a side helping of “data ownership”.
Electronification of capital markets, even in the primary space, is by now an outmoded goal. Very few transactions require a physical paper trail anymore. The German government in September did away with its archaic requirement for securities transactions to exist in physical form.
But while emailing PDF files and Word documents to each other is already standard practice and allows the almost instant and purely digital transmission of information, the process is still manual and requires human intervention at every stage.
Lawyers for the banks produce bond documentation — a process which itself may require several iterations of re-entering the same data. Then the bonds are issued via an agent and custody bank, sent to a central securities depository, cleared and settled and allocated to investors. Each stage may involve the co-ordination of several different parties, all of which are using the same data. Or they should be, but by this point they may have been separately encoded a dozen or so times.
Avtar Sehra, founder of a UK-based primary bond market focused start-up called Nivaura, says: “As the terms of a new security pass to paying agents, securities depositories and custodians, the data is entered into multiple systems, often in a very manual way. Our solution addresses these inefficiencies by allowing the existing players to process structured data in digital form.”
Nivaura’s system is a platform to manage the formation of a bond, its settlement, and post-trade lifecycle. Each stage of the process integrates seamlessly with market participants’ existing systems, meaning that it can provide efficiency savings and automation for any user, without relying on everyone in the chain using the same product.
At every stage, humans must manually take data from one source to another. Every time that happens, the potential for error is introduced. Removing that stage of the process removes the possibility of mistakes but, perhaps more importantly, frees bankers of a simple and manual task, and allows them to spend their time on processes that can add value to their clients.
That is always a more attractive spin than “allows banks to pay fewer people to do the same job”, but either way, the result should be a reduction in costs, and an increase in efficiency.
This is a problem with more than one solution. Blockchain, perhaps the most hyped technological development of the past 10 years, is finally flexing its muscles and demonstrating some of the promised values. In a sense, data harmonisation is an ideal project for blockchain. It gives each participant access to the “golden source”, the shared ledger that keeps track of ownership of assets.
R3, with its private blockchain system Corda, is among the largest and best-known service providers. Within capital markets, Agora, part of the R3 family, is aiming to build its own blockchain-based bond issuance platform.
Corda is likely to end up as the dominant private blockchain protocol in financial markets. While bitcoin runs on a public blockchain, where anyone can access it and see what transactions have taken place, such a system was swiftly deemed unsuitable for privacy-obsessed financial markets. Corda marries the immutable record and distributed ledger technology of public blockchains with the privacy and supervisory control necessary for bond market operations.
Commerzbank’s Main Incubator has been working hard on blockchain issuance protocols. The concept requires the creation of a “token”, which is tradable on a distributed ledger that represents the legal ownership of a security. That can be done for the issuance of a new bond, or simply creating a digital version of an existing security.
This cuts down on the risk of mistakes simply by ensuring that only one copy need exist, shared between every participant.
Nivaura, on the other hand, while it is has helped to issue bonds on blockchain, has built its platform to function just as well with traditional settlement infrastructure. The core of its product is the automation of the onerous aspects of documentation and connecting different counterparties across the market.
What numerous fintechs, Nivaura included, are finding is that properly structured data, rather than a blockchain, are key to efficiency improvements.
However, Nivaura is certainly aware of the potential benefits blockchain can provide in settlement and has presided over the issuance of several blockchain bonds. Its platform has been developed to function just as effectively with a settlement process based on blockchain as with the traditional means.
In some respects, the question of whether or not a distributed system or a centralised system should win out as the default for primary bond markets rests on a philosophical question of data ownership: should the single data source be centralised or distributed. A matter of preference on which, as yet, no consensus has emerged.
Will banks prefer the surety of owning their own data, or will private blockchain data security standards come to be perceived as an industry standard?
Avtar Sehra, Nivaura’s founder, says: “Market participants want to fully control their data — who has access to it, who can use it. To meet those expectations, we’ve built a multi-layered security infrastructure and an innovative set of user orchestration tools.”
Next steps
Although a blockchain enabled network may not initially appear to be a superior system for primary bond markets, some believe that if it becomes widely adopted, it could open up new opportunities to create value.
Bond.One is a US-based start-up building a blockchain-based network for primary bond issuance. John Mizzi, chief strategy officer, says: “What we’ve built today could work just as well on a centralised database, but the network effects once it’s adopted at scale are fantastic. Everyone using the same source of data means a much more secure and efficient process.”
These network effects will require new pieces of financial infrastructure to come into being. Commerzbank is also working on developing a cash solution for distributed ledgers that can facilitate the cash leg of securities transactions. With a bond on blockchain, as well as a digital cash substitute, settlement can be achieved instantly simply because the distributed ledger removes any need for clearing and settlement, since all participants’ assets are known to the ledger.
DVP (delivery versus payment) protocols are at the heart of the developments. Spitz calls cash on ledger “programmable currency”. In addition to instant settlement, programmable currency would allow the use of smart contracts to do more than just automate or instantly settle payments.
For example, a security could be structured wherein the coupon payment is affected by the fulfilment of certain conditions — Spitz imagines an ESG product wherein coupon payments are affected by carbon footprint reduction. The calculation could be programmed upfront and processed automatically on the payment date.
Of course, a bank’s proprietary system, be it Commerzbank’s programmable currency or JP Morgan coin, is unlikely to attain mass adoption. “The idea is to get pure exposure to the credit risk of the security you’re purchasing,” says Spitz. “People don’t want to take credit risk to a bank on somebody else’s security.”
The next stage will be for this system to become a piece of market infrastructure provided by a central bank, or other public entity.
Where is innovation needed?
The innovative potential of new technologies often comes from re-examining parts of capital markets infrastructure — clearing and settlement or custody, for example — that are often forgotten as less glamourous and lucrative activities. The way they work is often accepted as a bond market’s law of nature.
It took new technologies to expose not only the costs and inefficiencies of these systems but their systemic risks as well.
“If you have a €5bn redemption approaching, and you pay that to a paying agent, and it’s settled in two days, there’s a credit risk,” says Spitz. “With instant delivery versus payment, we can reduce that risk, processing payments without taking any exposure to the paying agent’s credit.”
But instant settlement will not be appropriate in every circumstance. Making a two-way market becomes much more difficult when settlement takes place instantly.
Short selling becomes much more difficult. If you have lent out a security that you wish to sell, the short position must be unwound. A two or three-day settlement process allows a grace period in which the trade can be executed before the shorted security is returned.
Without such a grace period, securities lending will be more unpopular and expensive, and liquidity will be adversely affected.