While prime, Alt-A and even some subprime mortgage loans look like they'd be able to weather as much as a 30% home price decline on the coasts and a 10% decline in the middle of the country - the swaps used in this year's vintages to hedge interest rate risk may end up significantly hindering some of those transactions in such a scenario, according to a simulated housing market decline scenario recently conducted by Standard & Poor's. In an environment of such rapidly declining home prices, analysts assumed falling short-term interest rates, positive for adjustable-rate borrowers and others in need of refinancing, but negative in terms of payouts from trusts to swap counterparties.
Similar to the scenario that contributed to later-vintage high-yield CDOs suffering numerous downgrades, overhedging for rising rates can be enough to push an already defaulting portfolio to eat through credit enhancement.
The majority of the 2005 vintage subprime mortgage transactions contain swap agreements tied to one-month Libor, whereas few of the 2004 and 2003 vintages include them. 2005 vintage subprime mortgage transactions with swaps initially carrying a BBB-' rating fell to BB' by 2008 under the scenario, while BB' rated bonds fell to a D' rating, assuming a five-year swap with a 4.75% strike rate.
A scenario of incrementally increasing home prices nationwide would certainly support a positive outlook for RMBS performance, as borrowers would be able to refinance out of hybrid products and capitalize on appreciation; and a soft landing scenario, or stabilization in home prices, coupled with rising short-term interest rates, would bode well for swaps on 2005 vintages.
Overall, the rating agency found that only the lowest-rated RMBS tranches experienced defaults under the scenario. For example, in order to correspond with hypothetical losses associated with market-value declines, using an AA' loss severity test on coastal areas and a BBB' loss severity in middle America, the rating agency found that 2003 vintage subprime loan losses from 2Q05 through 2Q08 would constitute 2.77% of pools, while the 2005 vintage would experience a 5.82% loss. Alt-A ARMs, on the other hand, would perform substantially better, with 2003 vintages experiencing 0.54% losses, and the 2005 vintage with 1.08% losses.
A nationwide home price decline has not occurred since the Great Depression, according to the rating agency, but S&P Chief Economist David Wyss said, "This is unthinkable. It can't go on forever - but the question is, when is it going to end, and when it does end, what is it going to do to the economic markets?'" Wyss added that historically, when home prices have accelerated too quickly, they've simply stabilized for a couple years, which is the most likely scenario in this environment.
Historical home prices are roughly 2.6 times greater than average household earnings, according to Wyss, and the national average is currently 3.2 times earnings. That would mean home prices would need to fall somewhere in the neighborhood of 20% to become normalized. In areas such as San Diego, where a home costs about 10 times household earnings, and in New York, where prices are about nine times earnings, the fall could be much more dramatic, he said.
S&P simulated a 20% decline in home prices over the next two years, including a 30% decline on both the East and West coasts, as well as a 10% decline in the middle of the country. As a result of the home price depreciation, the test assumed a decline in gross domestic product from about 3% to 1.2% in 2006, and 2.3% in 2007. The test, used for 2003 through 2005 vintages, did not assume a recession, but did factor in spikes in unemployment, to 5.8% and 6.5%.
S&P combined that scenario with a regression analysis performed by UBS Securities in a May 18, 2004, Mortgage Strategist article and found defaults would increase by up to 90% on the coasts and 60% in the middle of the country. Under the scenario, from 2005 to 2008, borrowers on the coasts would default at a rate equal to S&P's BB' foreclosure frequency, while others would follow a B+' frequency. Prime jumbo borrowers, who typically have fewer defaults, would experience a B+' and B' foreclosure frequency, respectively, on the coasts and in the middle of the country. Investor properties, dependent on continued appreciation, would have a BBB' and BB' foreclosure frequency, whereas subprime borrowers would perform at a rate equal to S&P's BB+' and BB-' foreclosure frequencies, on the coasts and in the middle of the country, respectively.
(c) 2005 Asset Securitization Report and SourceMedia, Inc. All Rights Reserved.