By Mark Somers, Technical Director at 4most Europe (

William Coen, the Basel Committee Secretary General said in a recent interview in the Financial Times, that he believes that the avalanche of new risk regulation is coming to an end. Risk functions across financial services will no doubt be relieved. Certainly banks across the globe have been kept busy in recent years; increasing the capital supporting their operations by nearly $0.5 trillion since the credit crisis, implementing a new Capital Accord (Basel III), adopting forward looking credit loss accounting rules (IFRS9) and the introduction of annual stress tests to name a few. However, to assume this implies no further change is likely, would be wrong.  The standards are set, the fires are largely out, now the work of implementation is in full swing and as usual the devil is in the detail.

Increasing use of mathematical models in risk

During the early phases of the economic crisis there were respected voices (for example Andrew Haldane at the BoE and Thomas Hoenig at the US Federal Deposit Insurance Corp) calling time on the complex mathematical models that in some instances, failed spectacularly and have often been viewed as being a key ingredient of the volatile and opaque nature of pre-2008 finance.  Whilst backstops in the form of leverage ratios (a simple to calculate ratio of capital to the balance sheet) have been implemented as part of Basel III, the desire to go further – to put into effect a Luddite reversion to the days of manually-led risk decisions – has not caught on in banks and consequently regulators can’t ignore their complex models that guide important decisions either.

It remains true that banks that can use models effectively to measure risks and thereby provide better quantification of them to investors and regulators, will have a competitive advantage and will dominate their markets.  Whilst BIS and other prudential regulators may have put less reliance on models for capital, the current introduction of IFRS9 into accounting principles will demand across the board, from the simplest mono-line mutual to more sophisticated international institutions, perhaps the most significant increase in risk model complexity of retail and commercial banks since the introduction of the Basel II internal ratings (IRB) system nearly 10 years ago.

Mark Somers
Mark Somers

Enhanced approaches to understand and manage model risk

With more reliance on risk models (not less),becoming inevitable then a more mature response to model complexity is to consider and formalise assessments of model risk. This is an area that is likely to mature considerably with the use of internal model “triangulation” and external benchmarking – using not just one but diverse independent models that have different assumptions to assess the uncertainty of potential outcomes. This is important to assess both the systematic model risk inherent in the regulatory RWA calculation within the Basel capital formulae and to understand model risks within internal bank models used in impairment,stress testing, pricing or credit decisions.

Another key element to manage model risk is to ensure a robust governance structure. This requires capturing expert judgement and insights from monitoring and back-testing to systematically review qualitative factors about models and aggregate these inputs to ensure models are working and used appropriately. Model governance provides the framework that should hold model builders to account – importantly, emphasis needs to shift from trying to simply guarantee that the single “best” model is used by the business to ensuring a range of model assumptions are considered and disagreement in model outputs are understood in the organisation rather than blindly following the number spat out of a process.

Increased sophistication in trying to understand stress scenarios; the interplay between economy, politics and future regulation

While economics remains a contentious discipline, economic history tells us unequivocally that economists are spectacularly bad at forecasting the timing of turning points of economic cycles. This is likely to be because at some level these phenomena exhibit a form of self-organised criticality – like the heap of sand in the bottom of an egg timer, the steepness of the slope is well defined but the timing and location of the next “sand slip” is not predictable. The same phenomena is present in many natural processes involving imperfect dynamic equilibria and economics is full of examples of such processes too.

Given the trigger of the next crisis may be unknowable then dealing with extreme events needs to focus on scenario analysis not backward looking historical models. To make this approach meaningful however many tens or hundreds of diverse scenarios need to be considered, their probabilities and severities assessed. This compares to the handful of relatively banal scenarios that regulators currently assess. To do this effectively will require new tools and a fresh approach.

About the author

Mark Somers is Technical Director at specialist analytics consultancy 4most Europe, based in London.  The company provides a range of products and services across credit risk, fraud and marketing, working with blue chip clients predominantly in the banking, retail and mobile sectors.

About 4most Europe (

4most Europe Ltd is a specialist credit risk analytics consultancy with offices in London and Edinburgh. The company provides a range of products and services across credit risk, fraud and pricing, working with blue chip clients predominantly in the retail banking and mobile sectors. The company offers a flexible, competitive model, either working with clients to manage regulatory change or delivering and implementing business critical solutions.

Comments are closed