Why Isins are not fit-for-purpose OTC derivatives identifiers

SocGen’s Bill Stenning says draft Mifid II regulatory technical standards need fit-for-purpose symbology

European regulators are hoping to make the over-the-counter derivatives market more transparent – a key goal and a sensible one – but doing so requires a reliable means of categorising and identifying an instrument, and officials currently propose borrowing identifiers from other markets to achieve it. That poses problems.

The identifier proposed is an International Securities Identification Number (Isin), already used very effectively to tag individual bonds, stocks and other securities. Swaps are not securities, however, and there are good reasons why Isins – and other common identifiers – are not already being used in the OTC markets.

Later in this article, I will outline a simple draft solution to this problem, intended to prompt discussion and provoke the generation of improved proposals. Using an interest rate swap as an example, the proposed identifier would capture the standard terms of a contract in a process that could be set in advance. It could then have attached to it trade-specific information detailing the notional, currency, trade date, maturity date, fixed rate, spread and central counterparty (CCP).

But, first, it is helpful to scope out the challenge to be addressed.

The push to apply Isins to OTC derivatives comes from draft regulatory technical standards (RTS) issued by the European Securities and Markets Authority (Esma), fleshing out the base text of Europe's revised Markets in Financial Instruments Directive and its associated regulation (Mifid II and Mifir). The draft RTS, published in September 2015, calls for an Isin to be attached to every trade executed on a trading venue or by a systematic internaliser (SI) – an SI is one of the new categories of market participant introduced by Mifid II and is expected to apply to most large dealers.

OTC derivatives are not securities

As we consider instrument identifiers in the context of OTC derivatives, a fundamental difference between OTC derivatives and securities is central to this discussion, namely the means of creating inventory, or ‘stock' if you prefer.

In the world of securities, stock is created via issuance. In this way, bonds and equities come into existence through an extended process. Subsequently, ownership of this finite supply of a given stock is transferred through trading.

For OTC derivatives, however, there is no issuance process and inventory is created through the act of trading. As such, a) the supply of inventory in OTC derivatives is not finite, and b) the act of trading does not transfer ownership of existing inventory – rather it develops new supply.

This fundamental difference exposes further divergence between the characteristics of securities and OTC derivatives:

• Exposure to issuer, as opposed to counterparty

In securities, the buyer is exposed to the risk of the issuer. Post-settlement (typically a few days) the identity of the initial counterparty – or even the CCP guaranteeing settlement – becomes irrelevant since the owner faces the risk of the issuer.

In OTC derivatives, when two parties agree a contract they are exposed to the credit risk of each other for the lifetime of the contract. Even where such a contract is subsequently cleared, both parties now face the intervening CCP – there is no concept of an issuer in this construct.

• Fungibility (or a lack thereof)

In securities, if a firm buys an amount of a given bond (say €10 million of a specific Bund) and subsequently sells an amount of that same bond (say €7 million), then the firm is left with the net position in that issue (€3 million).

OTC derivatives function such that when two firms execute a trade a new and unique contract is created. Even if the same two firms subsequently execute a trade with identical terms as the original deal, a second unique contact is brought into being. The implications of this are quite profound. Where a firm executes a trade with one counterparty and an offsetting identical contract with another counterparty, these contracts do not net and neither does the second contract in any way impact the first one – ownership is not transferred.

• Ex ante knowledge of trade terms

In securities, the economic (and other) terms of a given security are known before a trade is executed. In other words, when a buyer seeks to acquire a given bond, he or she agrees the price for the transfer of those previously known economic rights with the seller. While it may seem obvious, when a client calls to enquire about a securities trade, there is a defined range of possible securities that can be traded – or, put more succinctly, a client cannot buy or sell a security that has not previously been issued.

In the world of OTC derivatives, meanwhile, the contrast is considerable. Each contract is different and the range of combinations that could be requested by a client is a) almost limitless and b) not restricted to contracts that have previously been created. While there is a degree of standardisation across about 80% of OTC derivatives markets – around day count conventions and holiday calendars, for example – there is still a wide range of possible permutations of contracts.

A simple example illustrates this well.

Take an enquiry from a client to pay the fixed rate on a five-year euro interest rate swap (IRS) versus the floating Euribor rate. The client will typically, but not always, ask for the prevailing market rate for a swap maturing on spot plus five years, with an assumption that the fixed cashflow is paid annually, and the floating leg resets and pays every six months. Thus, ex ante the dealer has no certainty of the precise terms to be priced. It is easy to see how, even for a product as simple as a vanilla IRS, the level of complexity multiplies quickly.

To illustrate a subset of possibilities:

• The client could also ask for a specific date (such as slightly longer than exactly five years), which would imply a different price than the exact five-year price, as well as a decision as to whether the additional days form an additional period (known as the 'stub period') at the start or end of the swap, and whether that stub period only contains those additional days or whether they are added to the length of the first or last period (short or long stub period);
• The client could ask for a spread of a certain number of basis points to be added to, or subtracted from, the Euribor rate at each reset date with a commensurate adjustment to the prevailing fixed rate;
• The client could ask for a different payment frequency on the fixed leg, or a different frequency for the reset and payments on the floating leg;
• The client could ask for the trade as part of a package (such as a spread, butterfly or other grouping of instruments) that is quoted as a single price; and
• The client might ask for a forward start date.

As the example is extended into more complex asset classes, such as interest rate swaptions, the number of parameters required to describe a product becomes much higher and thus the number of possible permutations can become even more challenging.

Why can't a securities identifier be allocated to an OTC derivative?

Of course, it is mechanically possible to assign a unique identifier to each specific contract. There are a finite number of OTC derivatives that have ever been traded and trading volumes are such that any reasonable numbering system should be able to cope with future contract volumes. However, allocating an identifier to each separate contract is not the hard part. The challenge is classifying ‘like' trades in a way that provides meaningful information to market participants. How does a contract identifier convey information about the type of derivative to which it refers – for classification purposes – and how can it be allocated in a timely enough manner?

To illustrate this, building upon the previous example of a five-year euro IRS serves well:

• The client asks for a quote for the fixed rate for a five-year euro IRS, and this will be applicable to a product that is effective from a specific date (in this case two target business days after the trade date) and matures five years later (for the sake of argument this trade takes place on December 15, 2015). If the dealer happens to have an identifier for this fixed rate and maturity date (December 17, 2020) then everything can proceed. But what if the dealer does not have such an identifier? Indeed, what if the identifier for this specific quote has not previously been created? How long will it take for the identifier to be created?

Currently, a typical creation process for a securities identifier involves the submission of forms (often via a website) and, at best, something in the order of a 24- to 48-hour turnaround. As part of a multi-week/month issuance process for securities this is not a problem; in fact it works very well and the identifier can easily be created in time for issuance.

However, in the OTC derivatives space where the identifier for a specific quote does not already exist, suggesting your client calls you back in 24 hours for a quote is clearly absurd. Furthermore, due to the nature of OTC trading, when your client does call you back in 24 hours for a quote on a five-year euro IRS, the five-year date will have moved on a day (to December 18, 2020), which, in turn, could require another identifier to be created, necessitating a further 24-hour delay.

This highlights an underlying feature of OTC derivatives markets: a five-year trade executed today does not have the same economic terms as a five-year trade executed tomorrow, and in fact it is not the same contract.

Surely there must be an easy answer?

There have been two possible solutions suggested to this conundrum but, sadly, neither stands up to scrutiny.

• Solution 1: speed up the issuance process

If it is absurd to tell your client to call back in 24 hours, is there a time sufficiently short that is not absurd – 24 minutes, 24 seconds or 24 milliseconds?

In the world of electronic execution, timeframes have become increasingly short and the ecosystem involved in the trading process has become highly complex – involving order-management systems, execution-management systems, executing brokers, clearing brokers, clients, prime brokers and execution venues – plus the new Mifid II concepts in OTC derivatives of approved publication arrangements and both pre- and post-trade public trade reporting.

Even if an identifier could be issued in real time by a central authority within a few microseconds, it would have to be made publicly available in that same instant and propagated around the whole ecosystem, validated and updated into the infrastructures of market participants prior to any further activity taking place, because all of the participants in a market for a given product must have the same understanding of the precise definition of a specific product to avoid confusion and mispricing.

Notwithstanding the huge challenges of speed and scalability this presents, it is also profoundly unsafe to have parts of the trade execution chain acting on instruments it does not know – risk checks would have to be set aside and suitability could not be checked.

Clearly, this is unworkable even in theory; the fact that none of this infrastructure exists today makes the challenge that much greater.

• Solution 2: pre-allocate identifiers

This proposal suggests each firm or trading venue buys a number of identifiers ahead of time, and each time a quote is made on an instrument the firm has not previously quoted, one of these pre-allocated ‘spare' identifiers is attached to it.

Consider, by way of example, a fairly typical situation where a client calls around half a dozen dealers for a quote on a given instrument - in this case one that has not previously been quoted or traded. All six dealers will, presumably, issue one of their own identifiers against this quote.

Two insurmountable issues quickly arise.

Firstly, as quotes and post-trade reports need to be made public under Mifid II, how does the consumer of this public quote/price data locate sufficient information to fully understand the economic terms of the quote and thus benefit from this transparency? A quote/price/trade report that is ambiguous is, at best, of no value and, at worse, actively misleading. In other words, where does the database of the full economic terms referenced by that identifier reside? How quickly can a consumer of this transparency access the reference data? A post-trade price publication made within 15 minutes of execution, but where the economic terms describing the identifier are not available until, say, the next day, fails to achieve post-trade transparency.

Secondly, it now appears there may be up to six identifiers in issue for the same instrument. This seems to fall foul of the basic fungibility requirements of any securities identifier standard. Having the same instrument represented by multiple identifiers is clearly problematic from a data management perspective.

Of course, a different approach to pre-allocation could be to issue an identifier in advance for all instruments that could possibly trade. However, this approach would require an almost infinite, and certainly impractical, number of identifiers to be created.

What is the price?

To illustrate the complexity of OTC derivatives, consider a seemingly simple concept such as the ‘price' of an interest rate swap. Typically, and as implied in my earlier example, this is considered by many to be the fixed rate. But, in the case of an asset swap it is common to set the fixed rate and coupon dates to be the same as the underlying asset and, subsequently, to negotiate the spread over/under the floating rate index – in such circumstances the price negotiated is the spread as well as the fixed rate.

Or, in the context of an unwind or compression, the parties to a trade agree to set the fixed rate and spread to the same values as the original trade that is to be unwound/compressed, and to align the coupon dates after which the parties negotiate a fee to represent the close-out value of the original trade.

Factoring such subtleties into a product symbology was never considered, or indeed was ever necessary, in the development of existing securities identifiers.

How can we move forward? Are these challenges not true for all identifiers?

The key issue boils down to how much information is included in the instrument identifier and how much in the contract-specific data. At the right level, an identifier can be developed that avoids the pitfalls of approaches where every contract has its own identifier.

In fact, the consultative report Harmonisation of the Unique Product Identifier, published by the Committee on Payments and Market Infrastructures (CPMI) and the International Organization of Securities Commissions (Iosco) on December 17, 2015, centres on this approach, breaking the data describing an individual trade into groups of elements comprising instrument type, product, contract and transaction.

However, the CPMI-Iosco consultation focuses on the goal of facilitating global aggregation of data from the perspective of regulators. While this is an important task – and the CPMI-Iosco approach does recognise many of the key differences between securities and OTC derivatives – it should be noted that the transparency requirements of regulators are very different to those of market participants. For example, inclusion of the names of participants to a trade would not be suitable for market-wide public transparency, while the identification of whether a trade is bilateral or cleared (if cleared, by which CCP) would be highly relevant – in fact, the report would be meaningless without it.

By way of a straw man for market transparency, take, for example, an instrument identifier at the level of euro IRS annual fixed versus semi-annual Euribor, such that an identifier (say, ZZ1234567890) represents this and is defined thus: it is a euro IRS, fixed payments are calculated on an 30/360-day count basis, floating payments on an actual/360 basis, target holiday calendars apply for payments with no settlement lag, the calculation period end dates roll on a modified-following basis in case the market is closed, and the floating rate resets two target days before the start of the calculation period. This level of detail is both useful and necessary to obtain a full economic representation of the contract, and it can be set up in advance in a process much more akin to a ‘securities-style' issuance process.

Attached to this would then be a block of trade-level data specifying the effective date, the maturity date, the amount, the fixed rate, the spread versus Euribor, the fee and the CCP (if applicable). Thus a quote/report could look like:

ZZ1234567890 17Dec2015 17Dec2020 EUR100,000,000 0.2900% 0.00bps EUR135,000 LCH

The consumer of such a quote/price would be able to quickly and easily understand the exact economic terms represented by this information.

Conclusion

The straw-man approach set out above should lend itself to other OTC derivatives asset classes, such as interest rate swaps and credit derivatives. However, more work is required to complete such an analysis and also to consider how this might align with concepts such as Esma's liquidity assessment process under Mifid II. Other approaches could be equally valid and need investigating.

What is abundantly clear is that to be fit-for-purpose as a trade identifier, a symbol needs to carry a minimum amount of trade-specific pricing information. This, at present, cannot be a securities-style instrument identifier, at least not in the way these identifiers currently work. While the CPMI-Iosco consultation on harmonisation of the Unique Product Identifier acknowledges this, the specification by Esma of Isin (in the form prescribed by International Organization for Standardization standard 6166) in the draft Mifid II RTS does not and should be amended prior to the adoption of those standards across the OTC derivatives markets in Europe to allow time for a fit-for-purpose symbology to be designed and adopted.

Bill Stenning is managing director of clearing, regulatory and strategic affairs at Societe Generale Corporate Investment Banking.