Tony Sio – Balance new-and-exciting technologies with the tried-and-tested

CSDs have lost their fear of blockchain and are now pioneering its application to the securities industry, say Andreas Lundell and Tony Sio of Nasdaq, which provides post-trade technologies to CSDs, include a corporate actions notifications and instructions service that is attracting a lot of interest. Lundell and Sio urge CSDs to reduce the risks of adopting new technologies by collaboration and controlled experimentation – and posit the surprising notion that surveillance techniques should be used at the post-trade as well as the trading level.

“Many central securities depositories (CSDs) are looking to evolve their offering to the markets, and many are also poised to shift from technologies that are 20-to-25 years old,” says Andreas Lundell, head of product management, Nasdaq CSD solutions, at software vendor Nasdaq. “We want to be their partner, supporting them from a technology and knowledge perspective. There is a risk of looking at technology as an isolated phenomenon. CSDs need to see technology as a way of enhancing existing services or adding new services. There are many new technologies out there that can help them do that.”

One of those new technologies is blockchain. For a period between 2015 and 2017, CSDs were fearful that distributed ledgers would put them out of business. But Andreas Lundell reckons CSDs are more confident of their future now. “Even at the 2017 World Forum of CSDs in Hong Kong, we noticed a shift in the attitude of CSDs,” he says. “They realised there is still a need for a strong, centralised entity to govern the markets and, rather than be scared of the technology, they began to consider how it could be utilised for their benefit.”

CSDs are leading efforts to identify useful applications of blockchain technologies

In fact, Lundell reckons CSDs are now in the forefront of efforts to understand how distributed ledger technology (DLT) can be deployed to improve existing services, and how they can support crypto-assets or what Nasdaq calls tokenised assets. Nasdaq is itself a leader in the field, testing DLT in grey market trading of pre-IPO allocations. The firm is also working on DLT for collateral management, mutual fund settlement in the Nordic markets, proxy voting and the settlement and safekeeping of tokenised assets issued on to DLT networks.

Lundell sees tokenised asset settlement and custody in particular as an interesting opportunity for CSDs that is hampered by the uncertain legal and regulatory treatment of the asset class. “CSDs exist to handle regulated instruments,” he says. “So far, CSDs are looking at how existing infrastructure can be used to support tokenised assets rather than investing in new technology.” An obvious instance of this is the idea that CSDs can become governors of – or gatekeepers to – permissioned DLT networks.

Undeterred, Nasdaq is working with market infrastructures on DLT projects. It delivered a DLT-based electronic voting service to STRATE, the South African CSD. The firm was part of Project Ubin, a venture led by the Singapore Exchange (SGX) and the Monetary Authority of Singapore (MAS) to prove that transactions in tokenised assets issued on to multiple DLT networks can be settled securely by delivery against payment (DvP). In addition, Nasdaq formed a joint venture with Citi to put cash payments on blockchain. Lastly, Nasdaq Ventures has invested in Symbiont, which is exploring institutional applications capable of servicing the whole digital asset lifecycle, using blockchain and smart contracts technology.

Collaboration reduces risk when testing new technologies

Nasdaq demonstrably likes to work with other firms, and Lundell thinks CSDs should follow their example. “The best way for CSDs to approach this new technology is through collaboration,” he argues. Lundell applauds the work of the CSD Working Group on DLT convened by the International Securities Service Association (ISSA).

Tony Sio, head of marketplace regulatory technology at Nasdaq, agrees on the value of collaboration in the use of new technologies. “We should be sharing our experience of implementations of these new technologies, so that we can learn from each other,” he says. “Everyone is going to make mistakes as we embrace these new technologies and the more that we can learn from each others’ mistakes the more successful we are going to be.”

Sio adds that in any new technology project it is important to fully engage employees (who might be fearful of losing their jobs) and customers (who might perceive that they will incur additional costs) and regulators (which might change the rules). And Nasdaq should know. It runs numerous exchanges and central counterparty clearing houses (CCPs) and CSDs around the world. Since they run on technology supplied by its own market technology division, Nasdaq cannot be accused of selling untested software.

The group sells matching engines, clearing platforms and surveillance services to marketplaces as well as settlement and custody services. All these services run on conventional technology. “We will not see the core services provided by an existing CSD run on DLT in the short or medium term,” predicts Lundell. “We need to understand more of the benefits by looking at specific services or isolated asset classes. There needs to be a period of exploration and experimentation.”

The modular design of Nasdaq technology allows CSDs to pick and choose

In its own core technology, Nasdaq is pursuing a strategy which sounds old-fashioned by comparison with DLT: modularisation. All software built by Nasdaq is designed on single Nasdaq Financial Framework platform for assembly on a modular basis, so buyers with different budgets as well as different needs can purchase any of the pieces but also only the pieces they need.

The Nasdaq CSD platform, for example, which was revamped in 2014-15, is built on the Nasdaq Financial Framework. It offers segregated and omnibus account opening and closing for issuers and investors, settlement of transactions, corporate actions processing, instrument management, including reference data such as ISIN codes, and three varieties of securities financing: securities lending, repo and pledges.

Nasdaq corporate actions module is attracting interest from CSDs

Andreas Lundell expects the corporate actions component in particular to be sold and implemented on a stand-alone basis, and predicts that it will appeal to the banks and brokers that use CSDs as well as CSDs themselves. Certainly, the securities markets ought to be interested in an application that can make corporate actions processing more efficient. After all, corporate actions have acquired legendary status as a source of high levels of manual processing and excessive levels of cost and risk. “There is huge interest in the market for this technology,” confirms Lundell. “The future looks very bright for this product.”

The Nasdaq corporate actions module allows users to key in the details of a corporate action manually or capture the details from an internal system (provided the event is couched in an ISO 15022 or 20022 message format). This functionality is aimed primarily at issuer CSDs rather than issuers or data vendors, including any foreign securities that originate with a foreign issuer CSD. “We do not gather data from multiple sources and nor do we do any data cleansing,” explains Lundell. “Because we are aiming our product at CSDs. We typically see them as the source of the data.”

Once the details of the corporate action are entered or captured, the module then matches the issuer of the event with the eligible holdings in the Nasdaq CSD custody system. Next, it creates and distributes notification messages to the investors, including an estimate of any tax that must be withheld. As investors respond, their instructions are also captured and entitlements settled automatically in the form of cash (for, say, a dividend) or securities (for, say, a rights issue).

Corporate actions do of course come in dozens of guises, and in both national and regional variants. But the Nasdaq approach of breaking corporate actions into components that can be combined to match any corporate action issuers and their advisers can devise solves this problem effectively for CSDs charged with automating the processing of corporate actions.

“It is a challenge,” admits Lundell. “We have the standard corporate actions pre-configured in the system but you can also configure events based on different components. The smallest component in our corporate actions module is the movement of cash or securities. We have a number of different movements, including the movement of fractional entitlements. You can combine those movements in any way you like to cater for your specific type of corporate event.”

This ability to break a transaction into components which can then be recombined is what makes the Nasdaq CSD platform so flexible in general. It is not surprising, for example, that its securities financing module can settle stock loan and borrow and repo and reverse repo transactions, and pledges of securities. What is unusual is that the platform can also service the lifecycle events in a stock loan or repo, such as dividend payments or recalls or collateral substitutions, precisely because transactions can be supported by piecing together different combinations of separate components.

The component-based approach  – also known as a micro-services approach – has one further advantage: degrees of affordability. The principal cause of price inflation in a technology project is customisation. Nasdaq likes to reserve this for larger clients and offer smaller CSDs not a standard product but an affordable product with the components most valuable to them included.

Nasdaq also helps smaller CSDs cut hosting, maintenance and upgrade costs by providing Cloud-based services. However, even smaller CSDs have to adapt the technologies they use to the idiosyncrasies of their own market. Paradoxically, says Lundell, this can mean their purchases require more bespoking than the larger CSDs.

Surveillance technology is useful at the post-trade level as well the trading level

The component-led approach has also enabled Nasdaq to offer CSDs something even more unusual than securities financing lifecycle management tools: surveillance services. On the face of it, surveillance technology is of use primarily to on-market trading platforms rather than post-trade services. In fact, “bad actors” looking to launder money or manipulate markets or insider deal also make use of settlement and custody services.

Tony Sio argues that this offers CSDs the opportunity to enrich the picture captured by surveillance technologies at the trading level. An unusual position in a particular stock, for example, will be visible at the CSD as an increase in assets in custody. That might indicate a takeover bid is in the offing. Similarly, positions being run down might indicate short-sales. An increase in the volume of transactions being settled on behalf of an investor, or of assets being moved between accounts, might signify money is being laundered.

“If we have account-opening and closing, settlement and custody information as well as trade execution information, it fills out our picture of market behaviour considerably,” says Sio. “If you look at trading information only, you may not be aware of the particular actor, or of his positions.” He adds that CSDs are already using the Nasdaq surveillance technology to spot “bad actors,” such as money launderers, in the post-trade arena.

Though this capability is of obvious appeal to vertically integrated financial market infrastructures that combine trading, clearing and settlement, the availability of components specific to each of these activities means that even stand-alone CSDs are potential buyers of market surveillance technology.

After all, a CSD is in a good position to monitor the theft of securities disguised as a repo transaction or a stock loan, for example. Indeed, this work can be seen as part of an effective cyber-security policy. CSDs are also well-placed to spot so-called “passive mistakes” made by increasingly autonomous trading technologies. Software glitches or rogue algorithms are now much more likely to make an expensive mistake than the fat finger of a trader or a settlement clerk.

By capturing this data, the Nasdaq surveillance technology informs decision-making. “We suck up all the data that is passing through your internal systems, and augment that with additional data such as balances, positions, bank account details and social media postings, to help people understand how their market is evolving and where their business is going within it,” says Sio. “We are, essentially, using tools we originally developed to identify bad actors to identify people who are unaware that they are being bad actors, and so help clients become more pro-active about changes they need to make in their market rules.”

AI and machine learning reduce costs by ordering priorities

Market surveillance provides a perfect use-case for artificial intelligence (AI) and machine learning. “The problem I deal with is to find particular patterns of repetitive behaviour in extremely large data sets,” says Sio. “Those patterns of behaviour not only differ slightly from case to case but change slightly over time.” Nasdaq has multiple AI and machine learning initiatives in train, but it has one application live already in the Nordic market. It uses AI to select which patterns to investigate first, based on the past behaviour of analysts, with the aim of reducing the volume of false positives – and the costs incurred in investigating them.

He says there is a wider lesson about new technologies to be extracted from the experience: explore new technologies, but in a controlled and experimental fashion, and absorb the lessons. It was obvious that machine learning was applicable to market surveillance, for example, but not exactly how. In addition to using machine learning to rank issues in order of priority, Nasdaq also ran other experiments to use machine learning to detect anomalous behaviour and identify groups of bad actors. “Your first approach may not be right,” says Sio. “You have to be prepared to pivot, based on what you learn.”

Sio also emphasises that technology is a means, not an end. “CSDs need to be careful not to chase every shiny new toy,” he explains. “They need to assess whether the new technology is appropriate to the business change they wish to make. If a new technology does not change your business model, or the experience of your client, either you are looking at it incorrectly or it is not suitable.”

One reason Nasdaq is focusing on the mutual fund market in the Nordics as a potential use case for DLT, for example, is the fact that the modus operandi is so obviously in need of change: mutual fund settlement is characterised by non-digital communication and manual processing of purchases, redemptions and switches.

This message may be a familiar one. But the phrase so often used about blockchain technologies – a solution in search of a problem – is a reminder of how seductive a new technology can be. Nasdaq, which provides tested services based on established technologies while experimenting with new technologies such as blockchain and AI is a reminder that progress depends not on cynicism but on conjecture and experiment. For systemically important financial market infrastructures, Nasdaq provides an object lesson in how to strike the right balance between the new-and-exciting and the tried-and-tested.

Show open popup early birds