There was a time -- just a few months ago -- when insuring bonds was commonplace. But then the top-notch credit ratings of insurers (which are what make insurance attractive) were suddenly at risk. And just as suddenly, state and local government issuers that have used bond insurance in the past now are electing not to do so, creating havoc for many players in the municipal bond market. In January, the credit rating of AMBAC insurance was dropped two notches, lowering the ratings and market value of the 138,000 municipal bonds (worth $500 billion) that it insures. Downgrades of other insurers followed.
The cause of the crisis has nothing to do with the bond market per se. The insurance companies -- the MBIAs, FGICs and AMBACs of this world -- underwrote huge amounts of residential mortgage-backed deals, and those deals are now tottering into default. As they sour, the insurers have to pay claims -- and that is draining their capital.
This isn't how things are supposed to go. When the bond insurance business first started in the 1970s, it was based on a simple concept. Municipal bonds, especially those that rely on taxes and traditional user charges for repayment, have zero risk in terms of default. A government borrower might have a fiscal hiccup, but it never dies. For the insurers, it made good business sense to write a policy that guarantees a government's debt-service payments.
The companies needed to show state regulatory commissions and the credit-rating agencies that they had the financial wherewithal to back up the insurance. If they did, the credit agencies bestowed on them ratings of up to AAA. In the last analysis, the rating agencies became the final arbiters of the viability of the insurance business, since buying insurance essentially amounted to a government entity "renting a rating." Bond investors, in turn, rely on the insurers to decide if a deal is OK, resting assured by the insurer's superior credit rating that they will get paid in any event. Over the years, the use of bond insurance grew. To date, half of the $2.6 trillion outstanding municipal bonds carry insurance.
But it's been an increasingly competitive business. As profit margins on insuring municipal debt narrowed, the bond insurers stepped into other markets. These steps were taken in conjunction with the rating agencies, which adopted new models for measuring risk. The rating agencies found themselves with a growing army of new types of debt to rate -- vehicles that had no track record. The insurers were close behind, insuring various newfangled instruments, largely relying on the credit rater's assessments to heft up coverage. An incredible and indecipherable maze ofinterdependencies was created. Novel and complex deals were rated by the rating agencies and then insured by insurance companies that, in turn, were rated by the agencies on the basis of the deals they insured. But 2007 brought an end to the latest bubble with the implosion of the subprime mortgage market.
One need not be a wizard to see that the first half of this decade was an unprecedented time in the residential housing markets. (Could anyone of right mind believe that housing prices would grow by double digits annually and interest rates go down forever?) Yet, this peculiar period became embedded in the computer models used to estimate risk in residential mortgage obligations. Those risk models then became "inapplicable" as housing prices went into reverse, interest rates went up and mortgage defaults exploded.
The insurance companies, it turned out, had insured flakey deals that carried credit ratings that proved unjustified. Now, the rating agencies are saying "I'm sorry" by downgrading the insurance companies. Downgrades can potentially put the insurers out of business and are very damaging to investors that hold insured bonds. Something is woefully out of control and unaccountable in the whole process.
There is no easy cure. We celebrate the "smartest guys in the room" and applaud financial innovation, since that is where the big money is made. But there are risks. Financial engineering assumes there are laws of rational behavior that are captured in the mathematical models that drive analysis. But these constructs are based on historical relationships and, in many cases, the history used is both short term and unique to a certain time. Maybe one should restrict the highest credit ratings to those borrowers and deal structures that have a time-tested track record and are conservative both in terms of the underlying creditworthiness and the structure of the transaction. But absent ways to curb competitive pressures, the use and abuse of insurance and the ratings upon which it relies remain up to the markets to decide.
You may use or reference this story with attribution and a link to