Basel Committee’s CVA shocker ignores trade-offs

Ditching own models for CVA risk is too binary and eliminates possibility of further dialogue

Never have banking supervisors made their distrust of internal models more obvious than in the past year.

In March alone, the Basel Committee on Banking Supervision unveiled proposals to scrap the advanced measurement approach, the own-models approach to operational risk capital. That was swiftly followed by proposals to sharply curb the use of internal models for credit risk and scrap them for credit valuation adjustment (CVA) risk.

Where internal models are still permitted, it isn't exactly a picnic either. The Fundamental review of the trading book, finalised in January, contains a key test for obtaining model approval that major banks find almost impossible to pass – causing many to question whether it is even worth the effort.

When making big decisions such as these, supervisors must consider the attendant trade-offs. What is being given up for an additional benefit elsewhere, perhaps in the larger interest of things? In the case of their decision to ditch internal models for CVA risk, there are two sides to the argument.

One US regulator argues the proposal to remove internal models was due to the lack of comparability between different approaches used by banks. A second national regulator warns that the complexity of CVA risk modelling carries dangers, and could lead to significant model risk. The Basel Committee has argued that increases in central clearing and bilateral margining for over-the-counter derivatives should temper the overall level of CVA risk over time.

On the other side, banks complain that forcing them to use a less advanced approach could encourage them to engage in ‘regulatory hedging', instead of hedging actual risks that aren't recognised by the capital rules. In addition, they point out they have invested heavily in the past few years to find better ways of modelling, risk managing and optimising derivatives valuation adjustments such as CVA. If incentives to use sophisticated modelling did not exist, those efforts might dwindle, curbing risk management and innovation at banks. And those efforts are crucial when it comes to understanding complex risks such as CVA.

Competition and rivalry have added another dimension to the issue - with some regional banks rejoicing at the idea that bigger, richer banks may no longer be able to gain an unfair advantage by using advanced models. "The regulators have a point," says one risk manager at a European regional bank. "At this stage, I would agree we don't need this type of additional complexity."

The nature of trade-offs means things aren't black and white and there must be some give-and-take. In this case, the trade-off to consider is that the simpler standardised approaches don't reflect risk very well, but an own-models approach is more complex and allows for differences between banks.

With its CVA risk proposals, the Basel Committee pitched its tent at one extreme end of this spectrum. Complicated internal models were discarded and the idea of simplicity emerged victorious.

The answer is a marked contrast to the committee's last set of proposals, released just one year ago, and it isn't clear whether the all-or-nothing approach is particularly helpful. If it had chosen to retain some degree of modelling while increasing scrutiny of the model approval process, then regulators and banks might have been able to work together and respond to each other's concerns. In the case of CVA risk, at least, that no longer looks like an option.