The risk fallacy, page 3
First, because there had been only limited subprime securitization in the past, the models were mostly constructed on data from prime mortgages, which perform differently.
Second, the limited data that existed for subprime securitizations was misconstrued. The models yielded optimistic predictions in part because subprime defaults dropped from 10% of loans in 2001 to 5% in 2005. That created the impression that subprime borrowers were paying with alacrity. They were - but the models didn't take into account the fact that many borrowers were financing their payments with new loans. The borrowers' debts were growing, not shrinking, but the calculations didn't reflect that.
Finally, the models neither perceived that real estate was in a bubble nor grasped the interconnectedness of modern finance: They assumed that such defaults as would inevitably occur wouldn't affect the broader real estate market. "There has never been a nationwide real estate recession" went the mantra, and the models assumed that a slowdown in Miami real estate, say, wouldn't be accompanied by slowdowns elsewhere. That, needless to say, turned out to be wrong.
But none of that mattered at the time. Like a bunch of 8-year-old Little Leaguers who each get a trophy, vast quantities of subprime-related CDO securities were bestowed a AAA rating.
"Optimism about how subprime mortgages would perform led to more than 90% of securitized subprime loans being turned into securities with the top rating of AAA," says IMF economist Randall Dodd.
That AAA seal of approval enticed even cautious pension funds, many of which are barred from low-rated investments, to buy in. That's one reason that annual CDO issuance, tiny in 2003, had jumped to half a trillion dollars per year by 2007. During those same years, the total amount of packaged assets in the market shot from $5 trillion to $13.8 trillion.
Issuers kept right on whipping up more new confections, with synthetic CDOs (backed by credit derivatives linked to ABS, not even the ABS themselves), and CDOs squared (collections of CDOs), and others. These securitizations provided cash to issue more and more mortgages.
And with each new investment product, the number of intermediaries between the security and the underlying investments increased. It was a toxic combination: less and less clarity in greater and greater quantities.
That wave of CDOs, ABS, and MBS couldn't have happened without the national frenzy of borrowing. Alan Greenspan chopped rates dramatically to bolster the economy after the stock market plunge in 2001-02. Money became almost free, and lenders had seemingly endless funds to dispense.
Many of the potential customers with good credit already had loans, so the banks sought less reliable borrowers, to whom they extended credit at higher rates. If some loans didn't go so well, it wouldn't matter. Lenders bet that they could either sell the underlying home for a higher price, which would more than cover the defaulted loan, or persuade the borrower to take out another loan, backed by the home's rising equity, to help make payments on the original debt.
As a result, the share of the mortgage market composed of subprime loans shot from 5% to 15%.
Not only were banks lending to borrowers with shakier credit histories, but they encouraged those shaky customers to load up more borrowing on the same collateral. One house could now back two loans instead of one, so if its value fell and the borrower couldn't pay up, both loans were toast.
"Borrowers were getting [multiple] mortgages with 5% or no money down," says John Geanakoplos, a Yale economics professor. Most alarming, financial companies used their CDOs, ABS, and MBS as collateral to borrow even more money. As Geanakoplos puts it, "Leverage of the economy as a whole increased because there was more borrowing on previously unusable collateral."