Americans back from holiday can catch up on the Solvency II stress test result with this slide (from regulator EIOPA’s pdf presentation):
The regulator created four stresses. The most stressful turned out to be the adverse scenario, with 13 failures. This is kind of a good news/bad news situation. Yes, 10% of groups tested failed. But the shortfall was fairly small, €4.4B from an aggregate capital base of €577B (€427B after stress).
Wall Street Journal summarizes the adverse scenario:
The adverse scenario assumed a widening of government-debt spreads—the difference in yields on debt issued by riskier and more stable borrowers, such as Germany—to examine the sector’s resilience to sovereign risk. For instance, it assumed a 2.55 percentage-point increase in yields on Greek sovereign debt, a 2.58 percentage-point increase on Irish government bonds, and a 2.46 percentage-point surge for Portuguese government bonds. For Spanish sovereign debt, that figure was 1.65 percentage points.
The adverse scenario also tested whether insurers could weather a natural disaster of such great magnitude that it can be expected just once in 200 years, such as the multiple disasters that hit Japan in March.
It also looked at what would happen if inflation in claims increased 2 percentage points faster than currently expected, and whether that would cause a shortfall in reserves for property-and-casualty claims.
But how stressful is that? I’m concerned particularly with the stress on government bond holdings. The stress assumes that German and French bonds remain robust while other sovereign debt deteriorates.
But that’s like fighting the last war. Most companies should already be insulated against a blowup of, say, Greek debt, since the problem is well known. A stress test, on the other hand, would challenge a company’s German or French (or for that matter, U.S.) holdings.
I’m not suggesting the stress test is covering up industry weakness. Most likely, it’s not, since the stress test looks to my crude eyes less stressful than September 2008. And of course failure to meet the S-II capital standard could also imply the standard is too high.
It may help to put this in a U.S. perspective. Even in the messiest of times, I don’t think 10% of U.S. insurers would fail. But in a calamity I would like to know the shortfall of the failed companies totaled less than 1% of the capital base. That means the remaining companies could take over and, combined with state guaranty funds, fulfill the failed insurer’s obligations.