Is insurance assured?

Featured in "Operational Risk" - November 2003

As if there weren't enough issues surrounding op risk, I'm not sure the insurance question is close to being answered. At first, Basel put insurance in square brackets within the AMA, but at least it gave it a run. Then the EU upped the ante by offering insurance for all Approaches - Basic, Standardised and Advanced.

But joy was short-lived. Now it's back to AMA only. Fine in principle, but there must be a danger that those banks which have most need of insurance to withstand significant shocks will review the costs given, whatever they decide, that they will fall further behind their sophisticated Advanced cousins in the capital cost/benefit stakes.

But is it even definite that insurance will survive for the AMA banks? Regulators have never been entirely happy with insurance. Is the global insurance market well enough capitalised to support the banking industry? Is it sufficiently well regulated - especially reinsurers, on whom the real burden falls? And of course there's always the received wisdom that insurers never pay. [Have these guys had such a rotten experience with their house and car insurances?]

As for capital, two camps have emerged. The first argues that unless insurance can be turned into a capital substitute and effectively look like a guarantee, it should not be allowed as a capital 'mitigant'. The other camp argues that insurance is what it is - not a guarantee, but a contract of indemnity based on fortuity i.e. you have to prove the insured event has happened before the insurer is liable to pay. The debate should therefore be about safeguards and the value of insurance.

The tensions between the two camps come through in the qualifications for allowing insurance as a 'mitigant' and in the mysterious 20% cap, which nobody has explained satisfactorily. Of course, it could be argued that if it's in the AMA, and the AMA's about self-assessment, there's no need to apply any qualifications, but perhaps we'd better row along with the idea for the time being.

The Operational Risk Research Forum recently sent a questionnaire around the insurance market about the criteria. The results were instructive. First, the market wants to stay with 12 month terms and 90 (preferably 60) day cancellation periods - on both sides. It's had enough of multi-year policies and the considerable losses which they've caused. But regulators want a haircut from 12 months down - so all policies would be haircut. It's back to the great debate - should cover be guaranteed for 12 months, or is this really about renewal risk, which is usually slight? Next, the single A criterion. The insurers accepted this, but pointed out that A means very different things to the main rating agencies and claims paying ability is different from long-term financial strength, which is what ratings are usually about. And ratings change, usually down, which is a problem for long-tail liabilities. As regards the exclusions criteria, these run counter to good market practice. A number of 'regulatory' exclusions are already mandatory; and all policies are valid for events occurring up to the time of receivership, when they automatically terminate.

Finally, and curiously, where's the dog that did not bark - speed of payment? Perhaps the 'insurance' camp won that one and persuaded their colleagues that this is primarily a question of liquidity rather than of risk capital. We shall tiptoe quietly past.

I also have a suspicion that there's an assumption that each policy has an intrinsic value which, of course, it does not. Even if two policies had precisely the same coverage, terms and conditions, their value to two different banks will be quite different - a subtlety which, in some ways thankfully, is no longer referred to in the regulatory papers.

I've been nervous about insurance for some time - first, when it seemed it wasn't going to make the cut, and now, when I'm not convinced that all the difficult issues have been grasped. Or, to be more exact, that when they are grasped the anti-insurance camp may win and insurance will be quietly dropped.

This would be a disaster. On the one hand, if regulators have problems it's more than likely insurers and the capital markets can solve them. On the other, there's really no reason to distinguish insurance from other risk control techniques, and therefore allow it under self-assessment within the AMA.

At bottom, I'm not convinced we have a done deal, or that insurance is top of the regulators' worry list. If banks and insurers want to make sure that it moves from the 'too difficult' pile and survives in one form or another, they need to be like the England rugby forwards - and engage. Before the whistle blows, with no extra time.

John Thirlwell

© John Thirlwell 2004-07. All rights reserved.
Any reuse in whole or part requires our consent
Design by