CSDs, like other financial institutions, create and store data every time they settle a transaction or move an asset between accounts. Retrieving the details of a transaction is akin to the proverbial search for a needle-in-a-haystack. That did not matter much until regulators started demanding better data management and control as part of their efforts to avert a repeat of the global financial crisis of 2007-08. Fortunately, the crisis has spawned data management solutions as well as problems. One of them is Belgium-based INTIX.
Every financial institution is riddled with data silos. Central securities depositories (CSDs) are no exception. Like banks, their data storage and management systems reflect decades of ad hoc solutions to pressing problems, accentuated by different syntaxes and formats, including domestic and proprietary ones. Unlike banks, the data stores of CSDs are rarely complicated by a history of corporate mergers. But they nevertheless face stiff challenges in the management of data.
CSDs are scarcely unusual in this respect. If data as a whole was in one place more often than not, and stored in a single structured format, search engines would have a lot less work to do. As it happens, it was the example of the search engine that inspired the foundation eight years ago of a new business dedicated to solving the data management problem in the financial services industry. It is called INTIX, and it is based in Mechelen, midway between Brussels and Antwerp.
INTIX supplies a search engine for the financial services industry
“The question we asked ourselves was, `Can we become a Google for the financial services industry?’” asks Marc Braet, CEO of the company he conceived in 2010 and co-founded in 2011. “Financial institutions have all the data they need, but it is in complex structures and different systems, so they lack a convenient means of getting access to it, and do not always even know which system it is in. We developed a solution which leaves all the data in the underlying systems but enables the users to retrieve what they need by entering criteria that correspond to the data they are looking for. They can comb through the results to find the exact transaction they want to look at.”
The key to the solution is indexing the data. Any data set that is couched in an industry standard format, such as SWIFT, FIX, Edifact or ISO 20022 can be made accessible via dictionaries of the message types. Proprietary formats necessitate the creation of a directory from samples of the data, though they tend to contain much the same components as the mainstream industry standards. A so-called “sniffer” detects the format in which the data is held, and then “reads” it using the right dictionaries to understand whether a data field is, say, a beneficiary or a value date or an amount.
“The results of the sniffers and the readers is stored in our system, but the original transaction remains where it is,” explains Braet. “When end-users want to consult the data, our system makes a call to the location of the original transaction and says, `Give me the byte steam of that transaction.’ We get the raw byte stream of the transaction, which is then mapped against the dictionary in our system, so it can actually be read by the user.” Importantly, the system can report across data formats, and piece together the elements of a single transaction even if they are stored in different formats.
The system can work with unstructured as well as structured data
Predictably, the system works best with structured data, because the fields can be identified more easily. But the system can also search, Google-style, for transactions which mention particular names or currencies or amounts between certain dates. So the system can work with unstructured data as well, such as contracts, service agreements and invoices in PDF formats, but without the refinement possible with structured data. Even then, the addition of more criteria does make it possible to narrow the field in the same way as any advanced search function.
Regulators expect financial institutions to have control of their transactional data
The initial driver was compliance. Since the financial crisis of 2007-08, regulators have sought to improve market oversight and reduce market abuse by insisting financial institutions develop the ability to tell them what transactions they have carried out, in which markets, at what price, in what quantity, and with whom. The penalties for failing to comply are not trivial. On 19 March this year UBS was fined £27.6 million by the Financial Conduct Authority (FCA) for making a total of 135.8 million errors in transaction reports over the previous nine-and-a-half years.
Although it would be unusual for a central securities depository (CSD) to be fined by a regulator, they do work under the burden of exceptionally high expectations from regulators. The confidence that securities transactions will settle on time, and that securities will be kept safely, is crucial to investor confidence in the securities markets. “CSDs are a domain we are looking at seriously, and where we want to extend our business,” says Braet. “A large financial market infrastructure is not that different from a large bank. They both process transactions and store details of transactions.”
But he also finds that CSDs are different from banks in one crucial respect: their first priority is prevention rather than rectification. Although the original INTIX product, which gives users the means to look up transactions, is of interest to CSDs, they are much more interested in the value-added services the company is developing.
Control of transactional data improves decision-making and risk management
These are designed to provide business intelligence (“How fast is our business with China growing?”), transaction monitoring (“Alert me when a payment of $1 million made”), risk management (“What is the value of the transactions at risk of failing to settle?”) and transaction tracking (“Follow the entire life-cycle of my transactions from execution to settlement”) services.
Securities transactions do of course pass outside the ambit of a single institution, as details are matched with the counterpart, or cleared at a central counterparty clearing house (CCP) or settled at a CSD. INTIX is not a transaction routing network, which is why the company is now working with what Braet calls a “large infrastructure in the payments space” to enable clients that use the network provided by that infrastructure to see the current status of their inbound and outbound payments as they pass from initiation to settlement.
“As long as the data is accessible, there is no limit to the reach of our system,” says Braet. This is the principal appeal of the INTIX technology to CSDs. As systemically important institutions dedicated to minimising settlement failure, they value early warning of transactions at risk of failing to settle and tools that can help them find the highest value transactions to repair first. The risk management and transaction tracking technologies offer CSDs those benefits.
“CSDs are attracted by the idea of end-to-end monitoring of all their flows and the assurance of having a watchdog sitting over their platform,” says Braet. “In a number of organisations, our technology is feeding the dashboards in their crisis management rooms.” In fact, risk managers rival chief compliance officers and heads of post-trade operations as the readiest buyers of INTIX technology. “INTIX gets picked up by other stakeholders in the organisation,” adds Braet. “So our product gets distributed quite easily across our clients’ organisations.”
INTIX offers a standard technology solution not a bespoke data service
Conscious that internal IT departments can be a barrier to internal proliferation of this kind – the most direct competitor INTIX faces is the in-house build – Braet emphasises that his company sells a product customers can run themselves, not a service. This suits customers, which tend to resist putting transactional data in the Cloud but are happy to install INTIX technology in their data centres and give internal users access to it. Braet also emphasises that INTIX is a standard product, updated and upgraded automatically for everyone by the familiar techniques of remote version control, though he adds that it can also be configured to the particular needs of the individual client.
The configuration work is done by a team of in-house Java developers, the leaders of which have worked at INTIX since the company was set up. “We need developers, not data scientists, because what we deliver is technology, not data,” explains Braet. “I often have to explain to clients that we do not even want to see their data because it puts us at risk of breaching confidentiality and data protection requirements. Our technology is like buying Microsoft Excel. You can use it to make calculations and draw charts and tables, but you have to put the data into the spreadsheet yourself.”
But if the INTIX technology is a generic product, it is not a general one. “We work only with transactional data,” says Braet. “And we work only in the financial services industry.” This is easily explained. Marc Braet, his co-founder Wouter Van Santvliet and his chief marketing officer André Casterman all worked in financial services. It would be surprising if their technology did not find a ready audience among CSDs. After all, one – Maroclear – has already bought it.