How ironic and what a strange turn of Fortune’s Wheel that the most financially sophisticated Americans are likely to suffer the greatest pain from loans made to the least financially sophisticated Americans! The loans in question are subprime mortgages, which are defined as loans made to individuals with credit score of 620 or below. A subprime mortgage borrower has, in addition to a low credit score, high debt to income and loan to value ratios, on average 42% and 87% in 2006.
Last year marked a watershed in subprime lending. Interest rates were rising, which normally restricts mortgage affordability especially for marginal borrowers. But a huge, loosely regulated infrastructure of mortgage brokering and securitizing had evolved in the fat years of 2003-05. The machine had taken on a life of its own, and it reacted to changing conditions by relaxing its already none-too-tight underwriting standards. The result was an enormous volume of adjustable rate mortgages to first-time homebuyers who put no money down and often were not even required to document their incomes. This so-called "affordability product," masked under the socially desirable rubric of bringing homeownership to those previously shut out, is at the heart of the subprime problem.
Approximately 20% ($600 billion of $3 trillion) of mortgages originated in 2006 are considered subprime. Of these, nearly three-quarters are adjustable rate mortgages or ARMs. As of March 31, 2007, 15.75% of all adjustable-rate subprime mortgages were delinquent; 10.1% were seriously delinquent; 6.5% were in foreclosure. The year-over-year increase in the delinquency rate of subprime ARMs was 373 basis points; the rate of seriously delinquent loans jumped 211 basis points.
A brief look into asset securitization will help clarify how low-income ARMs became a headache for high-income investors. Wall Street can and has turned any predictable stream of cash flows into securities, the key word being "predictable". Mortgages, with their low delinquency rates and steady monthly payments, have proved to be ideal for securitization since the first Ginnie Mae pools were issued in 1970, but many other asset classes will do. Today, a typical securitized mortgage or asset-backed securitization (not Agency-guaranteed) looks like this:
Once loans are securitized, defaults and losses are absorbed from the bottom up, first in the unrated or equity piece, then in the B, BB, BBB tranches and so on up the credit ladder. These ratings are bestowed by Moody’s, Standard & Poor’s, and/or Fitch. Their input and analysis is critical, and the relationship between the rocket scientists who build the more exotic derivative investment products and the rating services is symbiotic. Securitized mortgage bonds are built to achieve desired levels of ratings for their various tranches. The ratings, in their turn, are based on models of historic performance of the underlying mortgages.
Obviously, the more complicated the instrument, the more critical the need for good historic data. That’s why the earliest Collateralized Debt Obligations (CDOs), which are repackages of existing securities, were built from corporate bonds, where decades of historical performance were available. CDOs differed from MBS in that the risk of ebbs and flows of cash flow streams was exchanged for the credit risk of ultimate repayment, but the same structure of tranching was retained. In the last few years, more and more securitized subprime mortgages and home equity lines (HEL) found their way into CDOs. Citigroup estimates that in 2006, the entire $50 billion of BBB-rated securitized HEL debt was absorbed into CDOs.
Despite the rating services’ disclaimers and warnings, investors often rely so heavily on credit ratings that their investment policies require them to sell bonds that fall below certain levels. Of course, savvy investment managers have been trying to sell subprime MBS and CDOs for months, but the rating service downgrades could well force a fire sale. Moreover, the downgrades are a tacit admission that historical data was insufficient to model the risk associated with these securities, and this means that nobody really knows where the bottom is. The equity pieces, BB, and B tranches of many subprime CDOs are already worthless. We were reliably informed of a BBB tranche, purchased last year at 99, which sold recently at five cents on the dollar. It’s probably not an aberration.
These dislocations have caused "flight to quality" Treasury buying, reduced appetite for credit risk in corporate bonds, and wider spreads in some callable debt. Broadly, investors are re-valuating their appetite for risk, which seemed all-but-unlimited just a few months ago. We expect prepayments of all mortgages to slow, causing the average lives of MBS to lengthen, at least temporarily. The greater risk is if investors became nervous about Fannie Mae and Freddie Mac; that is, if rising delinquencies and foreclosures popped up in the prime loans these Agencies guarantee.
How could this happen? As noted, subprime mortgages are securitized into tranches with varying degrees of credit risk, and by 2006, CDOs had become virtually the only buyers of the riskier tranches. Following the trail backwards, if CDOs are out of the market, few subprime mortgages can be securitized, which means that fewer subprime mortgages will be originated. The whole bottom level of the mortgage market reverts to the rental market. Home values, particularly in the lower tiers, fall because potential buyers can’t get financing, and sellers can’t sell. Refinancings drop off because there is less equity to draw on. Defaults increase as homeowners stop paying mortgages whose balances are greater than the value of their homes. And then the other shoe drops.
It’s an axiom of the mortgage business that, if people have jobs, they will pay their mortgages. So far in this economic cycle, jobs have been plentiful. The unemployment rate has hovered around 4.50% for close to a year. If the subprime mess contaminates the economy and leads to job losses—in construction, in businesses catering to home improvements, in banking and finance, wherever—a vicious cycle could develop. This is the nightmare scenario and the basis for those economists predicting a recession this year or next. It’s very much a minority view. A more likely endgame is that the most significant loan problems—as contrasted to financial problems—will remain concentrated in the bottom tiers.
It’s still far too early for certainty. The mortgage world moves slowly. Payments are only made once a month, and as long as borrowers are willing to cooperate with servicers, all sorts of accommodations can be arranged. Even when foreclosure is initiated, the final disposition can take months or years. Odds are, we won’t know the full impact of the 2006 subprime debacle until 2008 at the earliest.
http://shopyield.com/blog/2007/07/18/historical-data-was-insufficient-to-model-the-risk/
Posted by: Cate | July 18, 2007 at 09:47 AM
An excellent article on the sub-prime problem. Easy to understand yet thorough.
Amer Khan
University of Sydney
Australia
Posted by: Amer Khan | August 03, 2007 at 12:53 AM