The IMD Wrap: As crypto rises, yes, reference data is VERY important

Specificity is important. Imagine, if you will, a portfolio manager and a trader hosting a Christmas party. The portfolio manager is making their special dessert that requires a 16-ounce bottle of Bailey’s Irish Cream liqueur and sends the trader out to buy some. The trader goes to the store and finds the Bailey’s. Next to it on the shelf, though, is a 24-ounce bottle of Barry’s Irish Cream for $1 less. The trader buys it, and takes it home, pleased with themself for getting more value for money. At home, the portfolio manager throws a fit and demands that the trader go back out and get exactly what they asked for. (And if you’ve ever been in the “trader’s” shoes, you can come commiserate with me over what’s left of my Barry’s Irish Cream.)

When trading, specificity is critical, especially if you’re trading or managing money on behalf of clients with specified goals and investment criteria. And the foundation that provides the specifics of every tradable asset is its reference data. This information identifies a specific asset and marks it as separate from all other assets, such as the market information code for where a stock or derivative trades, its Legal Entity Identifier, Isin number or, in the case of bonds, its Cusip code, and other attributes, such as corporate actions data or client information.

For many years, reference data was looked down upon, and seen as mere record-keeping or a career only for those who had ambitions to become librarians. Now, however, reference data is finally getting its share of the limelight. Most notably (until a certain US regulator’s decision last week…more on that in a bit), there’s a class action lawsuit challenging the alleged monopoly (and very real fees) charged by Cusip Global Services (for years part of S&P Global, which was forced to divest the cash-cow business as part of its acquisition of IHS Markit, and which is now owned by FactSet).

This shift to prioritize reference data began in the early 2000s, when a combination of factors such as rising trade failure rates and regulations covering Know Your Customer and Anti-Money Laundering sparked the long-overdue realization that—Ding! (cue cartoon light bulb illuminating above our heads)—this reference data stuff isn’t just actually rather important; it’s VERY important.

Why so important? Because it tells you what you’re trading, and what you’re holding. Trades were failing because counterparties may agree to a trade believing they’re trading the same asset, only to find they were talking about similar assets—even issued by the same corporation—but not the same one. Think Bailey’s versus Barry’s—similar, but not the same. But a bottle’s barcode is the definitive identifier for a product, just as an asset’s reference data includes its specific identifier. Once a firm’s middle or back office finds that the identifiers don’t match, everything halts.

Introducing new exchange-traded assets, comprised of other traded assets—and the sheer amount of reference data generated—increases, and also introduces, complexity for data professionals managing the data on these assets and their markets for the first time

This is especially true in the bilateral, over-the-counter markets, where transactions are conducted directly between two parties, who (at least in the past) may have classified names and codes differently, and didn’t use any shared utilities to ensure consistency. This was the impetus for the founding of Mark-It Partners (subsequently renamed Markit and now part of S&P), and has driven the business of data management platforms such as GoldenSource and Alveo (originally FTI and Asset Control, respectively), and led directly to the creation of SmartStream’s Reference Data Utility (RDU).

But even the centrally-cleared, exchange-traded markets can also be confusing. For example, let’s say you wanted to buy shares in Ford Motor Company but instead purchased shares in Forward Industries, which has the ticker symbol FORD. Or, perhaps you wanted to buy shares in the Coca-Cola Company (KO), but instead bought shares in Coca-Cola Consolidated (COKE), which is actually the bottling company that produces and sells the soda.

The individual alphanumeric codes assigned to stocks, bonds, and derivatives ensure consistency throughout your organization and with your counterparties. This info makes sure you’re buying or selling the correct asset, or that you’re not over-exposed to a particular asset. 

Maintaining accurate security masters of reference data can be laborious and tedious—and, keep in mind, that’s now, after years of development, collaboration, and automation. It’s a uniquely complicated area, and it’s about to get more complicated.

The past year has seen the introduction of new types of codes and identifiers covering new asset classes. For example, the Derivatives Service Bureau had a busy year, releasing the Unique Product Identifier for OTC derivatives, then—with the Association of National Numbering Agencies—releasing Isin numbers for digital assets.

And just last week, as reported by my colleague Theo Normanton, Bloomberg extended its Financial Instrument Global Identifier to cover nearly 8,000 crypto assets, which is being done in partnership with crypto data provider Kaiko. Bloomberg issued the first Figis for crypto assets in 2021, and the vendors have now ramped up efforts in response to increasing volumes.  As Theo notes, this makes Figi—which has slowly and strategically gained ground after initially struggling to gain approvals against incumbent identifiers in the past—the most widely used open identifier in the rapidly evolving crypto space.

Theo’s story notes that “some market participants are concerned that if the Figi gained a dominant position in the identifiers space, Bloomberg could use the vast web of data attached to the Figi to make its data more attractive, pulling more paying clients into its ecosystem.” To be sure, some may have concerns about Bloomberg gaining a dominant position. But I would argue (1) Bloomberg already has a dominant position among the firms likely to be trading these assets and using Figis, and (2) the whole point of identifiers is not only to be able to identify an asset, but to be able to use a standard identifier to map between assets and those related to it, and in this case between crypto assets and “traditional” traded assets, such as stocks, options, or exchange-traded funds.

Enter Spot Bitcoin ETFs

The US Securities and Exchange Commission’s seemingly grudging approval of spot Bitcoin ETFs last week will finally see widespread trading of crypto assets on traditional and regulated exchanges, albeit as a bundle of assets into a fund, shares of which are then traded on-exchange. The decision included public dissents by commissioners Hester Peirce and Caroline Crenshaw, and an agreement but with concerns by commissioner Mark Uyeda.

What it will also do is create—and force firms that offer trading in these initial 11 ETFs, no doubt with myriad more to come in the future—a volume of new contracts, each with their own identifiers, along with the identifiers of each underlying Bitcoin contract on their respective crypto marketplace. This also means that anyone seeking to analyze the fundamentals behind each ETF will need to scrutinize the underlying fundamentals in greater detail.

Over the years, the launch of new equities and options exchanges, and the transition of the US securities and options markets in 2001, have been step-change events that resulted in massive increases in data volumes. For the most part, these moves impacted the levels of real-time market data—more potential strike prices based on decimal points rather than fractions means more price levels being hit, and mode price updates being distributed. 

In those cases, the amount of reference data needing to be databased and managed didn’t really change. However, introducing new exchange-traded assets, comprised of other traded assets—and the sheer amount of reference data generated—increases, and also introduces, complexity for data professionals managing the data on these assets and their markets for the first time. This is especially true if the underlying crypto assets have different nuances, are valued differently, or if they trade differently from traditional assets.

Many still struggle with the challenges of managing reference data. I can only see those challenges increasing as more crypto-based assets begin trading on public exchanges. I’m not sure yet if it’s a good or bad thing for investors, but I think it involves a little of both for data professionals. On the one hand, it’s certainly a good thing that the data on these assets will be subjected to the same stringent management and controls as other datasets; but on the other hand, it will undoubtedly mean more work for data professionals and the providers serving this space.

Is your firm ready for a crypto data tsunami? Let me know at max.bowie@infopro-digital.com.