ISDA Common Domain Model – more interesting than it sounds
(Last updated: )
A persistent background of abnormally low rates and flattened yield curves combined with increased regulation and capital requirements, has long mandated that the easiest way for banks to “make” money has been for them to save it. The industry has generally failed in this regard. Despite large-scale reductions in head count, physical relocation to north- near- and offshore, across the board costs remain stubbornly fixed at 70-75% of turnover. Enterprise-wide cost reductions have been subsumed by increases in compliance and reporting costs, even without factoring in the fines for failure in these regards. With the onset of MiFID 2, a typical (cleared) trade may be reported by and to the following entities: Dealer> Execution Venue> Middleware> Dealer> CCP> Trade Repository> Regulator. The same trade will have to be reported to the various actors at a number of stages pre- and post: setup, pricing, execution, capture, allocation, confirmation, regulatory reporting, clearing, settlement, risk calculation, ensuing life-cycle events. At each intersection of entity and event, a different descriptor may be required to denote the same trade, each reconciled with the other and to the underlying reality. The result is a Tower of Babel, built from the crumbling bricks of legacy systems – absurdly complex, inefficient, opaque, wasteful and expensive.
The modern wrecking ball to clear this old labyrinth already exists – distributed ledger technologies (DLT) allow for a univocal “golden” record that can be simultaneously shared amongst all relevant market participants. Unlike traditional trade taxonomies, the DLT record can also contain life-cycle events as well as risk and value data. The potential benefits are blindingly obvious- transforming data integrity while slashing reconciliation and reporting costs. A number of banks and consortia have accordingly invested heavily in the development and testing of DLT technologies. However, absent the equivalent of Google Translate, unilateral disparate implementation runs the risk of reproducing the fragmented, info-silo’ed situation that currently persists. Perhaps larger fragments, with faster internal data flows, but effectively the same old mess. Network effect benefits can only accrue if the network’s growth is unimpeded, this requires common standards, whether these be a telephone system, a shared internet protocol or a trade description methodology.
To mix semi-mythical metaphors, the sword to cleave this Gordian knot is the ISDA Common Domain Model (CDM). If readers are not already aware of the CDM, we recommend the white paper outlining the concept published by ISDA on 17 October. Designed to facilitate the instantiation of trades on a DLT, adoption of the CDM is intended to be the first vital step to the promised land of self-executing smart contracts. The CDM takes a radical approach, upending traditional taxonomic hierarchy. Rather than beginning with the product eg. semi-semi GBP 5yr IRS, the CDM begins at the “event” stage. An event represents an action that may be applicable to any trade eg. initiating the trade, cancelling, amending, novating etc. The few first stage events identify actions which are independent of agreed contractual terms. The second layer of the picture denotes “Dependent Events” defined by contract eg. floating rate observation, record of daily value, option exercise. A simple FX trade is therefore described as an event- the exchange of cash between two parties, price is captured by the two amounts exchanged. A bond sale is denoted by distinct records- the cashflows that represent the bond “definition”- principal out, principal in, coupon- the transfer records of the bond as defined and the cash purchase amount. Beginning with events rather than products or elements thereof, one can quickly build a record of more complex transactions- simple derivatives are built from independent and dependent events, complex derivatives are built from previously-defined simpler derivatives. The process scales up to whole portfolios, allowing for the automatic operation of portfolio events such as collateral calls, capital and risk calculations.
As with many “obvious” inventions (wheels on suitcases springs to mind), one’s first reaction is to ask why it wasn’t done this way long before. The CDM represents an inherently future-proof and infinitely-scalable solution to a problem which has only grown since the derivative industry’s inception. In addition, it represents the key to unlocking the true potential of nascent technologies that have the ability to revolutionise the industry. Irrespective of its clever design, the CDM will stand and fall with the extent of its adoption. The industry stands at a critical juncture, fundamental technological change is coming; whether market participants decide to adopt a common standard or go their own separate ways, it behoves everyone to make the decision on a fully-informed basis. The CDM is currently being evaluated for different use-cases, the working group welcomes contributions.