Garbage in, garbage out. It is hugely frustrating when your state-of-the-art risk, pricing or optimisation system is let down by dodgy data, the artful algorithms rendered impotent by gaps and inaccuracies in the information they depend on.
Having spent a number of years tackling the root cause of this problem we've got a lengthy list of gotchas - in the interest of a catchy title we've limited it to the seven we consider to be the most pernicious!
- Unwitting ‘standardisation’ – the temptation to knock a few edges off a square peg and squeeze it into the round hole provided in the data capture platform, to avoid yet another agreement that’s ‘stuck’, waiting for an upgrade to the application.
Of course it will be the provisions where transparency is key that will be swept under the carpet in this way.
- Cross-Template Oblivion – the disregard of a provision that relates to something that would normally be captured under a different template. For example, conveniently ignoring Additional Events of Default that are specific to a CSA, because that’s
something you’d normally capture from the Master.
- Gold-plated Data Dustbin – okay, you are capturing some of the data - but there’s still a high-risk multi-year cross-divisional development programme required before your key business users see much value from it, and who knows what’ll have happened
- Fat-Tail Fatigue – effort is focussed on the high volumes of ISDA Master and CSA agreements, leaving an inconvenient gap in data across a multitude of relatively low volume agreements (non-ISDA, repo and securities lending, et cetera). Is there still
an appetite to invest in all the work required to capture this information properly?
- Automation Fixation – any decent solution will apply technology to the problem, no doubt claiming impressive accuracy stats with esoteric references to machine learning and artificial intelligence. Face it, there’s no silver bullet. Focus on what
you are getting out of the platform and the gap to your requirements, and don’t get distracted by shiny toys!
- Silo Mentality – it may make sense to tackle OTC Derivative collateral annexes first (just as an example), but eventually you’ll want to interrogate your data across
all transaction types. Make sure that support for this isn’t a hand-waving afterthought.
- Exceptions are the Rule – at long last you’ve got a platform that lets you capture your collateral agreements. Except for some tricky independent amounts. And some even trickier collateral definitions. Oh, and support for concentration limits is,
well, limited. Ah, and we can’t handle those complex structured finance agreements with ratings agency criteria. Sure, the support for fund agreements could do with some more work… Sounds familiar?