The Banking Industry Should Learn from the Insurance Industry

I can’t comment on everything, at least above the degree of quality that I try to impose on myself.  (I know, the standards could be raised. 😉 )  But I did want to comment on a paper on banking capital regulations that came out of the Jackson Hole conference.  Odd Numbers and Naked Capitalism commented on it, and I thought both had good things to say.  I have my own twist to share, having been a risk manager inside two insurance companies.

The basic idea of the paper is that risk levels have to be reduced at banks, but banks want to stay highly levered so that they can earn high returns on equity, so asking them to reduce debt levels or internal leverage is not feasible.  Instead, why not have them buy insurance policies that pay out during banking crises?  Then they will have the capital when it is needed, and they can continue to lend in all environments.

(SIgh.)  I have oversimplified their arguments, but I have done it to help make some points, which are:

1) The cost of the insurance policy will get factored into the equity calculation for return on equity, at least at far as a prudent bank manager would view it.  The insurance policy is illiquid, and its cost should be reckoned as a part of the surplus it replaces, which may allow for a reduction in overall surplus levels carrying the business.  (Note to regulators: anytime you allow a financial entity a reduction in required surplus from a risk transfer agreement, you should analyze the alternative of using the premium(s) paid to add to surplus, and ask, which looks better.  Also, these are collateralized agreements, but in uncollateralized agreements, analyze the counterparties, and deny surplus credit frequently.)

2) Do the authors realize how expensive these agreements should be?  Consider:

  • The insurer is asked to post the collateral, which takes money out of its surplus.
  • The insurer is asked to be ready to lose an asset at a very bad point in the credit cycle.
  • The monies are invested in Treasury securities, so there is no possiblity for the insurer to make money from investing the premium more aggressively, but still safely.
  • Large banking crises happen about once every 20 years or so, with smaller ones more frequent.  With a long enough agreement, the loss of the Treasury collateral is almost certain, making the cost high.
  • If these were common, a sort of moral hazard would develop, similar to what has happened with the financial guarantors.  Banks would conduct business aggressively, realizing that they have the capital backstop.  Initial results would look good, until the crisis. Then, double surprise! The insurers figure out that they didn’t charge enough for the insurance, and the banks find out that their losses were larger, because of their aggressive behavior.  Wound banks be willing to pay premiums around 5-15% of the face amount insured, depending upon where the risk trigger kicks in?

3) Beyond that, there would probably be a scarcity of providers.  Few want to dedicate a large portion of their capital bases to the events that are entirely a process of human action.

Take a lesson from the reinsurance industry.  Ideally, you would want an agreement that took the risks directly off of your books, such that the capital would come when you specifically had losses above a threshold.  That’s been done in the insurance industry for reinsuring companies as a whole, and the reinsurers have usually come off the worse for it.  The insurers almost always know their risks better than the outsiders.  Reinsurers prefer to reinsure specific risks that they can underwrite, not companies as a whole.

But, lest I merely seem to be a critic, let me offer three suggestions for how to try to make this work.

1) Call Ajit Jain at Berkshire Hathaway.  They have the capital.  Give him a detailed proposal of what you want, and let him give you the quote that makes your jaw drop, or, watch him decline the business, unless you put your bank into a straitjacket of terms and limitations of coverage.

2) Try setting this up through an Industry Loss Warranty.  You would get paid capital during bad times if the industry has suffered losses past a threshold, and you have suffered losses in excess of a threshold as well.

3) Or, try setting this up as a catastrophe bond.  Borrow money through the bond at a high rate of interest.  Junk bond buyers will fund you.  During a crisis, if the banking industry losses exceed a threshold, the principal of the notes gets written down, and voila!  You have capital when you need it.  Note that the junk bond buyers should require more of a premium here, because bank losses tend to be correlated with other junk bond losses — no big benefit from diversification here.

I will leave aside the idea of setting up captive reinsurance sidecars, because those are just regulatory arbitrage.

My main point here is that I don’t think that this type of insurance will work.  Even for those willing to contemplate the structure, the true price will be too high for the banks to gain any benefit.  Perhaps this could be done on a limited basis for one more turn of the credit cycle, but I think for those that offer the insurance, the banks that buy it, and the regulators, they will be less than happy with the results.

In my opinion, we need to bring down leverage ratios for the banks, slowly but inexorably.  If that hurts their ROEs, well, I’m sorry.  If we are going to do a leveraged fiat money system, the leverage must be considerably lower than where we are now, and all synthetic exposures (derivatives) must be brought on balance sheet as if they were cash transactions.  It is that lack of transparency and increase in leverage that has made our financial system so much more risky, as this other paper from the Jackson Hole conference states.  I did not feel that the discussants really understood what they were talking about, because they could see the micro level risk reductions from derivatives, but miss the added leverage, lack of transparency, and concentration of risk in the hands of parties that were greedy for yield, and may not be able to make good on all agreements in a crisis.

Much complexity and leverage will need to be unwound before this credit crisis is over.  The era of high ROEs for banks should be over for some time, that is, until the regulators fall asleep again during the next boom phase.  Some things rarely change.