We have Mark Thoma to thank for pointing out this story on research by Nancy Wallace at the Haas School (UC Berkeley). She found serious deficiencies in the risk models used by banks to evaluate residential mortgages and predicted widespread problems a year before the subprime meltdown. She and her colleagues have just completed a detailed study of commercial mortgages which concluded they similarly are being mispriced. She finds that investment grade bonds are at risk, and it won’t take much of a correction to produce defaults at a higher-than-projected rate.
I wonder how long it will take the media to pick up on this story….
Earlier this year a flurry of bad news in the subprime mortgage market spooked Wall Street and generated more debate about the volatility of the residential real estate market. But the rapid meltdown in the subprime mortgage market was not as surprising to Haas School Professor Nancy Wallace, whose research has raised serious concerns about risk models widely used in the residential mortgage industry. Now Wallace predicts that limitations with risk-management methods in the securitized commercial mortgage market will lead to turmoil in that corner of real estate.
“This is a wake-up call,” Wallace says of the subprime mortgage collapse. “Banks became overly sanguine about the risks of these mortgages based on short-term performance measures without appropriate controls for the longer-run cycles and price volatility of real estate markets.”
Wallace, a professor at the Haas School since 1986 and chair of the Real Estate Group, has spent 18 years studying real estate price dynamics. In the past 12 years, she has expanded her research into options markets, particularly mortgage- and asset-backed securities. She has recently found that Wall Street banks and bond-rating agencies underestimate the risk of many of these new securities.
In her latest research paper, Wallace examined the $478 billion commercial mortgage-backed security market, which has grown an average 18% per year since 1997. Mortgage-backed securities are bond-like assets that are created by bundling residential or commercial loans and selling them on the secondary market.
Building A New Model
Working with Haas Associate Professor Richard Stanton and Rice University Assistant Professor Christopher Downing, Wallace developed a new way to calculate the implied volatility of the return on properties underlying commercial mortgage-backed securities. The trio was the first to provide an empirical test of such a model, using a sample of 14,000 properties in 206 deals between 1996 and 2005. They found that the return volatilities by property types are substantially larger than those calculated by either rating agencies or investment banks.
“Our estimates suggest that some investment-grade (as opposed to speculative) bonds in commercial mortgage-backed securities are likely to be at greater risk of default than current ratings suggest,” Wallace says. “We are concerned that defaults will be higher than expected if there are even relatively modest commercial property market corrections.”
Wallace believes the models used by banks, Wall Street, and bond-rating agencies to estimate default risk and to price mortgage-backed bonds rely too heavily on aggregate indexes based on short-term performance without accounting for long-run real estate cycles.
“The banks and rating agencies have been lulled into a sense of complacency about how volatile real estate returns are,” she says.
Nailing Down Mortgage Risk
Wallace proved prescient in making a similar argument about the residential real estate market more than a year before defaults on the riskiest home loans soared to a four-year high in March. Wallace, Stanton, and Downing showed as early as 2005 that banks were incorrectly estimating default risk and prepayment in the residential market because their commonly used models did not properly account for the important role of housing price dynamics on borrower decisions.
In another recent study, Wallace, Downing and Haas School Professor Dwight Jaffee were the first to document that banks were selecting the riskiest pools of home mortgages – the lemons – to sell into the securitized bond market. They found that those pools were more likely to contain mortgages in which borrowers prepaid or defaulted on their loans. This paper has drawn intense scrutiny from Freddie Mac and has helped to pressure the quasi-governmental entity into disclosing more information about underlying mortgages.
Although loan cherry-picking by banks and Freddie Mac may sound like a bad practice, sound economics suggests that it increases liquidity to banks for lower quality loans. The transfer of termination risks into a variety of bonds allows a broad spectrum of suitable investors to participate in this market, Wallace explains.
“It means that the market has innovated and created new capital market structures that really do solve a significant problem and provide much needed liquidity to lower-valued mortgage assets,” Wallace says. The challenge, as Wallace has shown, is to provide accurate and transparent models that investors can use to price and correctly calculate their true risk exposure.