Evolving disruption: the near to medium term future of collateral management
Recent advances in technology are beginning to have a major impact across financial services. New technologies are coming on stream and financial institutions are becoming much more efficient at adopting them.
But while applications for new technologies and initiatives are becoming clear, the overall impact and level of disruption they will bring is far from obvious today. In
the second part of this study, we examine three key new concepts and technologies that have the potential not just to increase efficiencies and automation with the collateral environment but to revolutionise processes, participants and operations across the industry.
All-to-all collateral platforms
The first disruptive concept is all-to-all collateral platforms. The concept has been around for some time and is more commonly termed peer-to-peer. However, the more suitable term “all-to-all” is gaining currency as the concept evolves.
The idea to create a platform through which entities across the market can exchange collateral directly with each other was born out of post-crisis fears of a shortage in available collateral to post against the vast swathes of over-the-counter (OTC) positions that regulators were mandating should be cleared or margined.
At the same time, it was becoming clear the capacity of banks to act as intermediaries in the market and to warehouse the risk in trading and collateral was going to be severely impacted by the capital and leverage rules within Basel III.
Several all-to-all collateral platforms were launched to great interest in the market and, while there has been some activity on them, so far they have not lived up to the hype.
Numerous factors are holding all-to-all platforms back. The first, says J.P. Morgan’s Michael Albanese, is that the problem they were designed to address, the withdrawal of banks from the market, hasn’t emerged.
“Banks haven’t disappeared as intermediaries as much as some might have imagined. There are still large volumes of trades in repo, securities lending, etc. that rely on banks to intermediate. Put simply, fears that banks would completely step back have not come to pass.”
There are also operational issues for the buyside to take part. Firms need the expertise, staffing and technology to manage counterparty risk, credit and Know Your Customer checks. Ben Challice notes that outsourcing these activities to an agent would mitigate many of these operational concerns.
Other firms are launching similar models but with less disruptive goals seeking to provide greater mediation between banks and dealers rather than to disintermediate them entirely.
“An all-to-all market is taking shape but there is still a long way to go,” says BNY Mellon’s Mark Higgins.
“Positive market conditions over the past few years have perhaps reduced the urgency to look for alternatives to bank- based financing, although there remains the need to access supplementary forms of liquidity. The peer-to-peer and all-to-all platforms that have emerged in the last couple of years provide additional options and a place to find new counterparties.”
He added that any relaxation of the current constraints on banks’ balance sheets would further reduce the need for all-to-all markets.
Another bank executive interviewed for the study voiced concern over the longer terms implications of the success of the all- to-all market.
“Non-banks operate in an environment where they don’t have to abide by the same rules as us,” he says. “If a bank offers a hedge fund a 180-day margin commitment they need to go out and fund that with somebody else and there is a certain amount of balance sheet leverage capital intensity to that.
“If they don’t fund it they need to hold liquidity against it and that costs them money, or they can fund it through the treasury. The peer-to-peer market is not exposed to that but the more it grows the more the regulators will realise that there is risk building.
“Basel III was not designed to shift the risk elsewhere in the system, it was designed to address the risk. If liquidity leaks into the unregulated market, risk will build up. The implications of that have not played out yet. The largest hedge funds are becoming bank-like but they are not treated like a bank in regulatory perspective.”
However, despite the challenges, as the market becomes more familiar with the concept and some of the operational issues are addressed, the appeal of all-to-all platforms is likely to grow. In addition, moves to standardise the collateral environment and tag securities will play to the strengths of the all-to-all platforms.
“All-to-all is one promising solution the market is looking at to improve liquidity,” says GlobalCollateral’s Grimonpont.
“Some of these platforms are becoming reality and are seeing business flows. Other initiatives around securities finance are also gaining ground.”
Artificial intelligence, machine learning and robotic process automation
Artificial intelligence has been a presence in collateral management for several years, driving the algorithms that run collateral optimisation programmes. However, as data becomes more centralised and the technology advances, there is significant scope for expansion.
While the increasing sophistication of collateral optimisation algorithms has been driven mainly by the ability to take-in more data points and process information faster, machine learning, where algorithms learn from the data and improve themselves, is coming to increased prominence across financial markets.
Collateral efficiency is ripe for disruption from machine learning and associated artificial intelligence technology. Martin Seagroatt, marketing director for securities finance and collateral management at Broadridge, identifies several manual processes that could be disrupted by machine learning.
Legal agreement electronification, regulatory analysis, reconciliations and disputes and client counterparty communications are all processes that currently rely on large elements of manual and human interaction.
He says that collateral pricing, optimisation, liquidity forecasting and counterparty credit risk could all be significantly enhanced using machine learning.
Initiatives are underway across the market to introduce advanced AI into operations and processes. CACEIS uses an AI platform to decide lending fees for its clients in one of the first examples of commercial use of the technology in securities financing.
Other banks have also made senior hires with AI and Robotic Process Automation (RPA) experience with a view to broadening the application of the technology across operations. “We are exploring ways to enhance highly manual, repetitive processes with RPA,” says Rob Scott, head of custody, collateral and clearing at Commerzbank.
“Further advancements in RPA could, for example, enhance not only the monotony and predictability of process but also allow the expert to engage with clients much earlier on, therefore speeding up a process. That would save time, which can then be better used to understand and optimise the existing client experience.”
A consolidated and accurate view of data is central to effective implementation of AI solutions adds Rob Frost, head of product development at Pirum.
“For machine learning you need the data and you need to move towards a real time, standardised data model. From there you can start to automate more processes. All the manual touchpoints today can ultimately be automated.”
Distributed ledger technology and tokenisation
All-to-all platforms and machine learning represent evolutions in existing processes but distributed ledger (DLT) or blockchain technology has the potential for revolution.
A number of initiatives are already underway in the industry and could fundamentally change how collateral is processed but the trend is likely to be evolution before revolution.
DLT offers ways to increase efficiency exponentially through the creation of a single, immutable record of transactions. Already projects are under way to “tokenise” collateral so it can be processed on a blockchain, reducing settlement times and negating the requirement to move collateral.
Tokenisation effectively is the creation of a digital record of ownership. In theory, any asset with a specified value can be tokenised from a Picasso painting to a Treasury bill. So in a tokenised world, ownership of every asset can be digitally tagged according to its owner.
While this doesn’t sound like a huge innovation, when this is combined with a blockchain, this record of digital ownership can be instantly and immutably transferred.
Not only does this mean that, with the correct legal framework, assets can be kept in a custody account and only the token (i.e., record of legal ownership) be transferred, the efficiency of transfer on a blockchain means that any asset can be broken up into multiple parts with different owners ascribed different percentages, potentially vastly increasing the pool of available collateral in the market and the specificity with which collateral can be allocated.
In addition, the ability to tokenise any asset means that collateral-to-collateral trades can be made directly between asset classes that previously would have been difficult if not impossible such as a collateral upgrade trade directly from gold to Gilts. It also means that baskets can be created of almost any assortment of collateral.
HQLA-x, a distributed ledger pioneer is beta-testing its platformthat will enable the tokenisation and trading of DigitalCollateral Receipts, tokenised baskets of collateral, which will trade on R3’s Corda blockchain.
“There is clearly a need for interoperability in the collateral market that will enable firms to move collateral around in a cost effective and instant way,” says Olly Benkert, chairman of HQLA-x.
“The industry has had an aversion to working together to achieve this, which is understandable as they compete with each other. We have come up with a way of tokenising baskets on a distributed ledger and swapping the tokens so that the collateral doesn’t have to move between custodians.”
The firm has created a legal framework that it says enables full legal title transfer and in March, Credit Suisse and ING conducted a test trade on the platform exchanging legal ownership of a basket of securities but without the securities moving between their respective custody accounts.
In August, Deutsche Bourse announced an investment in the firm and said it will work with it to connect Clearstream to the platform.
“What we need to be successful is for the custodians to come on board. We are not asking the banks and buyside to completely overhaul how they operate collateral management for this to work. We need the custodians to feed into our platform to maintain existing infrastructures,” says Benkert.
J.P. Morgan’s Challice is excited by the potential of tokenisation: “It is something we are exploring and clients like the concept.
“We see our role as the custodian/collateral manager as the translation layer between the old world infrastructure of physical settlement, custody banks, agency banks and sub-custodian networks and the new world of that digital representation,” he says.
Evolving existing infrastructure
For now at the very least, DLT is expected to complement existing processes and workflows rather than overhauling the infrastructure.
Challice notes that tokenisation is not that different to the shift to pledge in many respects.
“Pledge is a different legal representation of a collateral obligation from title transfer. But a token could be fully digitised and a much more mobile representation of that,” he says.
BNY’s Mark Higgins also sees DLT as an evolution. “Triparty agents use the client’s custody footprint to make decisions about how to manage that client’s collateral, but in a DLT world do you need to hold the assets within a confined custody environment?
“If under a DLT construct assets don’t need to be moved, do you need to have a custodial network and physical accounts? These are questions that will be addressed over time. Triparty collateral management will continue to exist, but will it look like it does today? Probably not.”
Looking at the collateral cycle, there are clearly areas in which DLT can continue to improve efficiencies. Despite recent advantages, current settlement networks look outdated.
The industry has come a long way from the need for paper certificates for title transfer but it still takes more than 24 hours to settle. Tokenising the collateral on a blockchain could reduce that to seconds.
“Blockchain has the potential to streamline the efficiency and significantly reduce costs and duplicative reconciliation across various organisations,” says Commerzbank’s Rob Scott.
Others believe that progress will come without the need for blockchain solutions. “My view is that DLT is not a critical step. It is one of several possible solutions,” says Albanese.
“Other solutions the industry has attempted in the past are locking down a piece of collateral in a control account and documenting legal terms that make it clear what the custodian needs to do if the collateral provider default.
“DLT has accelerated the conversation and has got some parts of the market excited but it is not a sine qua non of being able to achieve certainty of legal ownership without moving the collateral.”
Grimonpont concludes: “The world is changing, however most of these initiatives are being built on the existing building blocks of the industry. Currently many of those new initiatives are being built alongside existing market infrastructures as they provide the necessary credibility and comfort."
To download the full whitepaper, click here.