Loss Severity Leverage
I’ve been analyzing on some surplus notes from a mutual life insurer which is part of a group of mutual insurers. Mutual insurers are opaque, because there are no large private interests external to the company with a concentrated interest in the well-being of the company. This company lost significant money in 2008 and the first quarter of 2009. Most of it was writedowns on non-GSE (not Fannie, nor Freddie) residential mortgage bonds, where they bought mezzanine or subordinated bonds.
Wait. Some definitions:
Senior bonds: those rated AAA, representing the group that gets paid first in a securitization. In a securitization where the loss experience is so bad that the senior bonds take losses, typically all of the senior bonds get paid pro-rata.
Subordinate bonds: Bonds that receive high yields in a securitzation (rated BBB and below), but take losses first as losses emerge. High risk, high return (maybe).
Mezzanine bonds: If the subordinate bonds get wiped out, the Mezzanine bonds (rated AA and A) are next. Not much extra yield, but less chance of loss.
When a securitization goes bad, the juniormost bonds get losses allocated to them until they wiped out, then it goes to the next most junior class. This highlights a difference between the loss severities of corporate bonds and structured bonds. With corporate bonds, there is some recovery of principal — not all of the principal, of course, but 40% on average. With structured bonds, typically you get all of your principal paid, or none of your principal paid (excluding early amortization).
I realized this for the first time when I analyzed the Criimi Mae Securitizations back in 1999. The securitization trusts contained mostly junk-rated CMBS tranches, and junk-rated securitizations of other CMBS deals, and they sold off investment grade participations in the securitizations. This was the messiest set of deals that I ever analyzed. It reinforced one idea to me, that if you are not senior in a deal, you may lose it all. With structured finance, there is loss severity leverage in mezzanine and subordinate bonds.
Going back to the firm I was analyzing, When I looked at their performance last year, I thought, “Stable company. Bread and butter life company. As my old boss said, ‘It takes more than incompetence to kill a mutual life insurer, it takes malice.’ ”
Today I am not so sure. I once was a mortgage bond manager, and I bought senior securities because the yield spreads on the lower-rated securities were so small. This company, like Principal Financial, bought mezzanine and subordinate bonds in significant measure, though they did slow down after 2005.
I have estimated the likely losses on the bonds that the company owns, and it is unlikely but not impossible that the company dies in the next few years. The parent mutual company has ponied up money recently, and will probably do so in the future. Who can tell?
My estimates of loss are less than those that the bond market is estimating. If we marked the assets of the company to market, it would be significantly insolvent. Insurance policies are high credit quality obligations, they don’t vary as much as bonds that are risky. Now, if I had the Actuarial Opinion and Memorandum, perhaps I could know the truth. That group of documents analyzes the long-term ability of a life insurance company to survive. That document is sadly not public. If it suggested that the company in question needed additional capital, that would be reflected in the public filings, and there is no such reference for the company in question.
Can You Afford To Lose It All?
Anytime one buys a mezzanine or subordinated security, or buys a surplus note, or trust-preferred security, or other bit of junior debt or preferred stock, one must ask, “Can I afford to lose it all?” When there are senior investors, investors that are more junior can get whacked. Senior investors ordinarily must be paid in full before junior investors get paid.
This applies even to AA and A-rated structured securities. They aren’t senior, so they can lose all their principal. And that’s what this life insurance company bought a lot of, picking up a princely few extra tenths of a percent in interest over the AAA bonds for a lot more risk.
How Did These Securities Get Such High Ratings?
Uh, seemed like a good idea at the time? 
As I’ve said before, it is impossible to set regulatory capital levels or subordination levels for assets that have not been through a failure cycle. New asset classes have no track record of failure, and as such, estimating the likely expected present value of losses, and how variable losses could be is just an educated guess. Sometimes they would apply models of related asset classes and tweak them. That’s still a guess, though.
Also, using statistics for assets held on balance sheets, and applying them to securitizations, where the originator has little skin in the game, made the assumptions made for losses on residential mortgage lending too low.
The rating agencies did have competitive pressures to get business, but my view is that they did not have sufficient relevant loss experience to guide them.
What Should Have Been Done?
Regulators abdicated their duties by merely relying on the rating agencies. Ratings are fine in theory, and someone has to make judgments of relative risk, but the macro decision of what classes of investments are permitted, and what capital level to hold against them belongs to the regulators. The regulators haven’t done well in setting up rating agencies of their own, so let the rating agencies continue on, but let the regulators be smarter in how they set the capital levels (higher for structured products than for corporates and munis, because of loss severity leverage), and what they allow regulated entities to invest in.
My view is that any new asset class has to go through a failure cycle before regulators allow regulated entities to invest in them. Investments in such new assets should be treated as a deduction from equity.
Wait, We Need to Invest in Those Assets to Stay Competitive!
That scream came the regulated entities. Sorry guys, the risks of untried asset classes belongs to unregulated entities that don’t have deposit insurance, state guarantee funds, etc., where there is no systemic risk. If they go under, no one in the public domain should care.
In principle, it shouldn’t matter within a group of regulated companies what the rules are, so long as they are enforced similarly by the regulators.
Anything else?
One final note — there is one medium-sized mutual insurer flying under the radar that I think its state regulator is unaware of the risks that it faces, and the rating agencies as well… it carries high ratings from all the main insurance raters. Given the project that I am currently working on, I can’t reveal the name, but the guaranty funds would be more than capable of handling the failure.
June 11th, 200912:38 pm at
yes, this is something i’ve been wondering about and which hasn’t been covered much.
what was/will be the extent of miscalculations of the loss rates of mezzanine tranches of securitized products, and who bought and held these products?
the structured credit models, even if they *might* do a good job on supersenior or equity tranches are pretty poorly equipped to handle mezzanine tranche pricing — and as you said the additional premium for taking this risk was quite small.