Advocacy for a reference data format in crypto-financial industry
The crypto assets market industry is at the dawn of its standardisation and institutionalisation. But when we speak of the “standardisation” or “regulation” of this sector, we often focus our mind on exchanges operators, virtual asset service providers or token issuers, rarely on data resulting from transactions. Data is one of the links of the entire transactional chain. Making the crypto industry and more secured not only involves making market operations AML-compliant, but also ensuring that they are expressed in an intelligible and well-known format to facilitate their post-trade processing and recording. Then, dealing with the crypto-market and blockchain-transactions data can be risky if they are not well described. With this article, we wanted to show what lessons we can get from the traditional finance industry to improve crypto data format within a global referential.
“Reference data”: what does it mean in the traditional finance industry?
“Language is a source of misunderstanding” (1).
A transaction is basically expressed with a bunch of related information: an object, a price, a date. This kind of information is called “market data” and refers to a trade that occurred in a specific place (an exchange). But each market or exchange can have its own data format, and each financial product can have its own appellation. Sometimes, a same asset or category of assets can be expressed differently from an exchange to another. Like the language, an appellation, symbol, or acronym can mean many things — and the opposite is often true. With the 2007 financial crisis, regulators pointed out the fact that a lack of cleaned and standardized data could have led to major settlement issues between market participants and then participated to financial risks increasing. Following this, data standards so-called “reference data” have emerged in traditional finance.
In a nutshell the role of the reference data is to standardize the description of the transactions within a unique policy in order to drive efficiency and accuracy in the transactions processing. It completes the market data with additional items such as financial classification, closing and settlement prices, transaction participants, calendar information, counterparty, purchase/sale information, exchange name, currencies (etc.), and then express each of these components under standardized and common-shared identifiers or formats. These identifiers are either global (for instance: International Securities Identification Numbers (ISIN), Market Identifier Codes, Legal Entity Identifier, …) or local (by sector, region, or others). Reference data is mostly supposed to help the front, middle and back-office operators for settlement reporting, portfolio management, risk calculation and compliance, and the accountant in its cashflow tracking and balance sheet establishment missions. Partly pulsed by the Basel Committee on Banking Supervision, their implementation seeks to mitigate risks by avoiding failed, missing or incorrect transactions in their processes — a misidentified product or value can naturally lead to transaction failure and money loss. But it seems not being the case in the crypto finance yet…
Reference data in the crypto-financial industry
On the contrary, even though we observed that crypto exchanges are making a significant effort to match their API structure with financial habits, there’s still work to carry out. The crypto market industry is still in its early stage and does not benefit from a well-defined, unified, and global framework for data formatting and messaging standardization. Applying reference data to business and internal procedures would mostly rely on manual executions due to the different market data sources and related format. Indeed, it involves matching crypto market data and internal transactions reporting with a corresponding semantic manually. This can be even more complex as underlying data, product terminology and market events may change and occur in a high frequency: new tokens listed almost daily, emergence of similar tokens due to fork events, are all kinds of events that can disturb the analysis of data resulting from trading markets. As a result, operators spend time and money in updating processes and analysing data to put them in existing boxes — for instance, a Capgemini study (2) showed that the global data integration efforts of the traditional finance cost almost $2 billion in the nearly 2010… –. You can guess it is even worse in the case of DEXs (decentralized exchanges): these platforms are mostly built by developers (for developers) — with their own language — who are more animated by the decentralization philosophy than the standardization of the market. The only manner to assure that the token is the right one is auditing the contract address and analyse the blockchain…
Finally, operators dealing with huge quantity of crypto-related data seem to seek data management systems which would help them to implement straight-through processing and mitigate risks between disparate financial systems. But this can be made possible only if raw data is constantly reshaped on the basis of global, accurate and worldwide crypto data standards.
Can financial standards be applied to the crypto industry?
First of all, we need to define what type of asset is cryptocurrency to know what standards, what ISOs must be met.
If we consider that cryptos are currencies then they must respect the ISO 4217 which defines the three-letters symbol of each currency and precious metal (e.g. USD, JPY, CAD…). How is it constructed? The first two letters represent the country and the third one is the name of the currency (D for dollar, Y for yen etc…). For precious metals that are not linked to any country, we use the X as the first letter to represent the internationality (e.g. XAU: gold). Cryptos must therefore each have their code starting with X and then two letters naming them. The ISO 4217 recognizes already the XBT denomination for Bitcoin. Fun fact: there are only 676 possible combinations with the X in the first place while there are about 8000 cryptos listed on CoinGecko. This standard is therefore difficult to apply to the crypto world if not amended.
We can also consider crypto assets as financial instruments. In fact, the Commodities Futures Trading Commission (CFTC) classified cryptocurrencies as commodities since the publication of the Digital Commodity Exchange Act of 2020. If so, they should then follow the guidelines of ISO 10962 and be classified by a CFI (Classification of Financial Instruments). This CFI is a six-letters code that classify assets by class, where each letter represents a sub-category. If we follow the recommendations, the code for crypto assets would be ITMXXX (I for Spots markets, T for Commodities and M for Others, X are used when there are no additional details). There may even be another appropriate code for futures contracts or associated indices. So there seems no real problem with the application of this standard.
In contrast, the fact that whether a cryptocurrency is a security is more ambiguous. Its definition varies a lot between jurisdictions. In the United States, the definition is broader: the case between the Securities Exchanges Commission (SEC) and Ripple Labs assumes that XRP is a security. Starting from this assumption, they have to attribute an ISIN code to each cryptocurrency (ISO 6166). But how can we reasonably attribute an ISIN code to an asset which hasn’t a country of affiliation? In the traditional finance, when it’s undefined, they use the prefix XA, XB, XC or XD. Then, regulators could create an easy-to-use ISIN format or equivalent for cryptos and appoint a committee dedicated to the attribution.
All of these problems lead to that the only way to bypass them in traditional finance is to sell trackers’ shares.
The case of marketplaces
The above-mentioned problems concern the identification of an asset, not a transaction related to it. On traditional markets, the standardization of transactions is achieved through the Market Identifier Code (MIC). It is applied to global markets and sub-markets. For instance, the code for NASDAQ is XNAS and for the NASDAQ Options Market is XNDQ according to ISO 10383. To transfer it to the world of cryptos, we must also consider the fact that everything is “swappable” for everything and even more on DEXs. Perhaps it is more appropriate to give a code for all trading pairs that would take into account the two individual ISINs of the assets that constitute the pair?
In conclusion, the standardization of the crypto financial industry from an accounting and legal point of view remains complicated. In fact, international legal standards are still unclear and make it difficult to harmonize. These rules are however essential for interoperability between institutions and all international actors thanks to a well-defined format of the transactions which will allow a clean recording and an essential transparency. That’s why we actively support the development of the creation of an identifier that would characterize trading pairs individually within an exchange.
[Updated July 12th, 2021] But most of all, we’re now proud to be part of the DTI Foundation’s PAC to build the Digital Token Identifier (DTI) under ISO 24165 with major actors of the place 🔥 Follow the news here → https://dtif.org/
Written by Benoit Chambon & Gautier Humbert — June 16th, 2021.
(1) Antoine de Saint-Exupéry, in Le Petit Prince.
(2) Capgemini, “Reference Data and its Role in Operational Risk Management”, Risk & Compliance, 2012.
