Was it beauty killed the beast or was it readjustments? If you have not passed out during our review of the respective roles of adjustable rates and subprime lending in the housing debt bust, here's some detail from Edward Pinto.
Pinto, a mortgage-finance industry consultant and chief credit officer at Fannie Mae in the late 1980s, explains that default risk on an original loan increases geometrically the closer you get to no money down. A default propensity of 1 on a property bought with 80 percent financing increases to 2 at 90 percent financing, 4 at 95 percent, and 8 at 100 percent.
It's the rapid growth of these low/no-money-down purchases that Pinto has tracked back to the passage of the Federal Housing Enterprises Financial Safety and Soundness Act of 1992. Most of that growth took place not through the Federal Housing Authority but through private lenders guaranteed by the government-sponsored entities Fannie Mae and Freddie Mac.
Why did the GSEs count so many high loan-to-value mortgages as prime on their books? Because they were relying on shady measures of loan quality. "Fannie and Freddie consider subprime a certain classification: If it's a lender that traditionally does subprime, or has a division that does subprime, they'd count it as that." So as loan originators passed along low-down-payment loans as prime, Fannie and Freddie (which were required by the 1992 act to provide more support for low-income and underserved purchasers) played along with the fiction. The GSEs finally revealed their exposure to alt-A, subprime, negative amortization loans and other junk in Fannie Mae's 10Q from the third quarter of 2008. (If this link [pdf] doesn't work, the relevant disclosure from page 182-183 is at right.
"Through the end of 2003, self-denominated subprime, impaired credit, as a percentage of the overall mortgage market, didn't change," Pinto says. "The rapid growth in the loan portfolio came from low down-payment debt and other mortgages based on flexible underwriting. There were more flexible definitions of income, such as including an energy tax credit as income. There were people proving their credit history by rental receipts."
In Pinto's view, self-described subprime lending, which increased rapidly in 2004, was the "third stage." First D.C. used relaxed underwriting to nudge up the national rate of homeownership (which grew more than 5 percent between 1993 and 2005, when it peaked at 69.2 percent and began dropping quickly). Next came a period of equity extraction, as owners of overpriced houses began refinancing at artificially depressed interest rates and getting green money from their home equity. ("Cash-out refi appraisals tend to be inflated," Pinto notes.) Actual subprime lending was kind of the last gasp, demonstrating Robert Shiller's point that the bubble caused subprime, not the other way around.
What does this mean for the original point of this discussion: that adjustable-rate mortgages across all loan-quality classes correlate with defaults more reliably than does any particular class? It's true that ARMs continue to lead the default curve, and one of the weird wrinkles of the last decade is that the popularity of ARMs mushroomed during a time of relatively low interest rates. But Pinto's larger view is that de facto subprime lending has been going on for so long, with Fannie and Freddie's connivance, that by the time the default wave hit, already it was impossible to say which was which.