Bloomberg Experts: Semantic Tech to Improve Financial Operations
Today, executing financial operations often requires the use of identifiers to clearly specify the different financial instruments used - such as equities, bonds, options and so on. But these IDs may be different in various organizations and countries, which can lead to mistakes and slowing down financial operations. At the recent Digital Transformation & Global Society Conference hosted by ITMO University, the Bloomberg experts Richard Beatch and Inessa Collier, as well as Dr. Yuri Lipuntsov, Associate Professor at the Moscow State University, spoke about new standards of identification of financial instruments and how they can help solve such problems.
The conference discussion topics were sorted into five categories: eSociety, eCity, eHealth, eKnowledge и eScience. The latter included reports related to high-performance computations and Big Data. Richard Beatch, Inessa Collier and Yuri Lipuntsov presented the results of their work: how using the Financial Instrument Global Identifier, or FIGI, can be combined with an informational model of interagency interaction. The new system aims to simplify the information exchange in the global financial sector. But why is it relevant and who will benefit from it?
First, let’s define what these identifiers are. When two market players enter a trade (for instance, one bank sells stock to another), they use certain codification for these securities. They need this in order to be totally sure what kind of deal it is and what it’s about. For international financial instruments, there are ISIN codes assigned in compliance with rules set by the International Standards Organization. The standard regulates the procedures used for assigning these codes, as well as the list of organizations that have the right to do that.
Surely, that may seem simple: here are the codes, just use them. However, there are also national identifiers, as well as internal identifiers (created within a financial firm for its own database), which often leads to confusion. Let’s say two banks want to buy or sell some stock traded on MOEX. One bank is located in France, the other - in Great Britain. When conducting the trade, their trading systems send requests to the MOEX system using the stock symbol assigned by MOEX. But their own databases rely on their internal identifiers, i.e. codes assigned by these banks. After the trade is done, they have to execute the settlement via the central securities depository, which requires using national identifiers. Since this is an international deal, it has to be settled via the international central securities depository, which requires ISIN codes. So, we see that each exchange or trading venue, bank or any other financial organization has its own identifiers.
The situation becomes even more complicated because ISIN codes and many national codes can change due to corporate actions, e.g., merges and acquisitions, name or domicile change. If one of the deal participants has the updated identifiers, and the other does not, the settlement will not be executed in time. Thus, although the existing identification systems play an important role in the international settlement, they are not sufficient for automated processes. As the amount of financial data grows, financial professionals start to understand that the existing data storage and processing systems are no longer effective.
“Big financial companies have large engineering departments that provide support to the data departments. Most of them still work with relational or structured databases. Semantic technologies bring about a different approach to their work: they are used for storing and processing information in a completely different way. Sure, everything that can be done with semantic technologies, can also be achieved using databases. But the problem is that databases don’t have the same potential. What is more, when working with databases, processing speed is slower, and it requires a lot more resources,” explained Richard Beatch.
He added that today the governments of different countries, including the US, European and Asian states, require greater transparency in financial transactions. Such level of detailed reporting amounts to the use of Big Data. The problem is that such data changes very quickly, and in order to ensure the accuracy of its reporting, companies must increase their resources for processing data. And this is unprofitable. Therefore, financial institutions are gradually introducing semantic technologies.
So, what’s so different about semantic technologies? The main difference is that they don’t use mathematical methods to describe data and establish connections to it. Instead, they use metadata, i.e. data that describes other data. This semantic data represents interlinked information that is recorded as the so-called triple: subject-predicate-object. Subject is linked to the object via some semantic relationship expressed via the predicate. For instance, the phrase “the funds are transferred to the account” has a subject ‘funds’ and an object ‘account’ interrelated via a particular semantic relationship ‘transferred’. Semantic relationships are created with the help of ontologies - documents or files that assign the relationship between terms (concepts), i.e. the subjects of semantic relations. In addition, some relations are created automatically based on the existing definitions. However, it is important to note that the meaning of the terms must remain unchanged and be assigned by the same person who created the semantic network.
How can this be applied in the financial sector? Using semantic technologies, the market players can understand semantic relationships between different financial terms: financial instruments and how they are linked to each other, as well to their issuers. The Financial Industry Business Ontology, or FIBO, is one of such ontologies. It is a sort of dictionary of semantic relationships between objects in financial industry. FIBO was developed jointly by the Enterprise Data Management Council and the Object Management Group, an international not-for-profit technology standards consortium.
The Object Management Group also developed FIGI, a unique system of identifiers that is connected to FIBO. The FIGI standard is a 12-character semantically meaningless alphanumerical code with a check digit at the end. These identifiers do not change as a result of corporate actions, which makes them persistent. This aspect is very important for supporting databases and automated processes.
“FIGI is not just an identifier; it allows to understand how things are identified. The goal of its development was not to replace other codes, but to explain their interrelations. For instance, it helps to specify a given trading venue where a financial instrument is traded. We can do that by using the combination ISIN+MIC (sometimes ISIN+MIC+CUR). However, from the perspective of data management the use of a single code will result in better quality of data and will reduce possible mistakes. Surely, one can use stock symbols instead, but that would call for a system of cross-references, which would cause difficulties when some stock symbol changes. Instead, the ticker, ISIN, MIC, and CFI become metadata within FIGIs, which provides the structured relationships,” commented Inessa Collier.
FIGI can also be used in conjunction with traditional methods to increase the efficiency of data analysis. For example, it can improve, risk analysis. At the conference, the experts from Bloomberg and Moscow State University presented a paper that describes interagency information exchanging using FIGI identifiers. As of now, there are already more than 400 million FIGI codes assigned for different financial assets, and their number continues to grow.
“We also wanted to show how FIGI works, and provide the government, the academic society and the financial industry with some roadmap on creating a system for exchanging financial data,” mentioned Inessa Collier.
The introduction of a common system can also assist with the development of Blockchain technology which implies fast processing speed for all transactions, which is not possible when different players on the market use different identifiers.
“Before FIGI, there was no standard that acted as a common permanent identifier of the same financial instrument in different banks. Thus, when banks wanted to enter a deal, they had to map their identifiers to external ones, like stock symbols and ISIN. This mapping is also necessary for regulatory reporting.. FIGI provides a common identifier for the whole community, and that is a key benefit,” noted Richard Beatch.
Andrei Chugunov, Yuri Kabanov, Inessa Collier, Yuri Lipuntsov and Richard Beatch. Credit: econ.msu.ru
The report by Richard Beatch, Inessa Collier and Yuri Lipuntsov was named best in its category at the Digital Transformation & Global Society Conference. The Bloomberg representatives also noted that they are considering collaboration with ITMO University, including possible plans for conducting a round table as part of DTGS-2018.