Martin Sexton of London Market Systems Limited examines the benefits of the SKOS network for product classifications.
Being top of the game is the aim of any trade association and FIX members are not scared of putting their heads above the trenches when it comes to identifying a clean and simple solution to an industry issue. Not only has a proposal been put forward that meets the OTC derivative products reporting requirements associated with Dodd-Frank, but it is also working with others such as ISITC and X9 to create a viable solution to meet the need for a globally acceptable classification scheme under the ISO banner. There is an overlap between these two initiatives in that financial products need to be appropriately classified; and this is where semantics has a role to play.
As soon as ontology is mentioned a concern is always raised as to the potential impact on a project due to the risk of adding latency into the deliverable time line. There is the impression that any semantic analysis is complex and yields little in tangible benefits. So what is the missing link? Could it as simple as identifying a framework capable of linking a semantic model onto a physical model?
Simple Knowledge Organisation System (SKOS) is just that framework; it allows the user to provide meaning and understanding to financial product classifications and, where appropriate, a visual representation can be used to help convey its intended use and ensure the industry exploits its capabilities and achieve maximum potential.
If we take the recent FIX Protocol regulatory reporting proposal, SKOS allows us to link the FIX terms with those outlined in the final rule (17 CFR Part 43 – Real-Time Public Reporting of Swap Transaction Data). ISDA for their part has developed a set of taxonomies. Based on four key terms the strict hierarchy under each asset class comprises three related terms, namely Base product, sub-Product and Transaction type. A further extension to this scheme is provided for commodities to support the Settlement type (this can have the value “cash” or “physical delivery”). The FIX proposal covers all asset classes and base products whilst the ISDA proposal only focuses on OTC derivative products.
The CFTC Rule identifies two key facets, namely asset class and contract type, each of which can be broken down further by sub-asset class and contract sub-type respectively. SKOS provides the ability to manage the relationship between all the terms and most important link them to the associated FIX equivalent tags.
At a high level the relationships between the CFTC’s final rule, the FIX and the ISDA proposals can be represented in a diagrammatic form.
There is also an awareness that the product taxonomies capabilities has limited capability of classifying risk, therefore the FIX proposal also includes a taxonomy specifically design to quantify the risk associated with the position.
In addition to capturing the relationship between terms, it is also possible to render details about each term (or Concept in SKOS). The figure that follows is a snippet of the FIX taxonomy used as part of trade lifecycle, principally for order routing. The taxonomy comprises the FIX tag 460 Product which can be further broken down by SecurityType (Tag 167). The example includes the associated labels for each tag, supplementary information such as a definition of the term and any external references:
Other industry standards lend themselves to being represented in a SKOS framework, the most obvious of these that would benefit from this framework is ISO 10383 the Market Identification Code (or MIC). Apart from the concept of a market itself, it also includes information about the locale of the market (country and city), its market status (eg Active), the live date and the industry acronym for the market. The tabular nature of the ISO MIC means that it can be easily transposed into SKOS.
Some users of SKOS have indicated how easy it is to construct a classification, though it somehow doesn’t feel right. This makes sense, if one does not apply some form of management framework around the data. For example, if one examines the ISO MIC, the market itself is normally governed by the laws of the jurisdiction in which it resides. That relationship needs to be captured and appropriately documented. It may also reside in a city where other markets exist. Thus the exchange has both a relationship to the city as well as the city to all the markets. SKOS provides sufficient documentation constructs to allow one to define context and provide examples.
So what are the advantages of SKOS? Apart from the reduced time to market, given the ease at which one can construct a classification, the main advantage is that it allows one to identify explicit relationships between terms within a scheme, in a human readable form. Once “skosified” and with the help of an appropriate engine, the user can then undertake complex searches that can allow one to identify similarities, refine the searches and ultimately improve the design of the classification itself.
Efforts continue to be put into managing the deluge of regulations and FIX protocol continues to respond to the industry’s requirements in a prompt and adaptive nature. The introduction of semantics can provide FIX members with the latest weaponry to remain at the top of the game.
[1] TopBraid Composer Graphical Screenshot
