It appears that after many dark years, the housing market is finally turning the corner. Home values are up in most metro areas and serious delinquencies (defined as payments being at least 90 days late) are down. According to the most recent Mortgage Bankers Association survey, even the share of risky subprime loans in serious delinquency has fallen by 18% since the beginning of 2011.
The one exception to this trend of good housing news?  Federal Housing Administration (FHA)-insured mortgages. Their serious delinquency rate has risen by 6% over the same time period. This is only part of the sobering news from FHA, which recently announced that its main single-family mortgage insurance fund is insolvent, with a negative net worth of -$13.5 billion.
FHA’s failure to accurately account for unemployment risk is a key factor behind these troubles. Recent research I conducted with Joseph Tracy of the Federal Reserve Bank of New York shows that FHA has systematically underestimated the true default risk in its insured portfolio. Its forecast error is so large that even much improved housing market fundamentals have not been able turn the fund right-side up.
The agency’s empirical analysis finds that unemployment risk is not an economically important factor in explaining individual borrower default—or even a significant contributor to the risk that taxpayers would have to bail out FHA’s insurance fund.
This contravenes what common sense and sound theories of default tell us about unemployment risk and the decision to default. If a household becomes unemployed and does not obtain another job quickly, default is in fact very likely. An unfortunate hallmark of our recent downturn — very long average length of unemployment — has only made that likelihood greater. Once a household exhausts its liquid resources, it literally cannot continue making the mortgage payment and defaults. Layer these risk factors atop FHA policy that encourages little up-front equity investment and the high default rates and insurance fund losses are hardly surprising.
My research finds that the true impact of a borrower becoming unemployed on its probability of default is about 100 times greater than FHA estimates. This means that becoming unemployed is a more powerful predictor of default by an individual borrower than even their credit score or initial leverage ratio. At the portfolio level, FHA’s current attempt to control for unemployment risk yields no real improvement in the reliability of its default and loss forecasts
Actuarial modeling has a penchant for putting people to sleep. But history tells us that policymakers and taxpayers must be wide awake to the dangers of FHA’s assumptions. One need not look far into the past — to the government bailout of Fannie Mae and Freddie Mac — to recall what happens when government entities evaluate risk through rose-colored glasses.
We can no longer afford that kind of reckless optimism.
Currently, unemployment risk remains a largely unpriced factor in estimating the true cost of the FHA insurance program. Until FHA accurately accounts for that risk, the fund is likely to see high defaults and insurance-fund losses going forward. Further, without an accurate assessment of FHA’s financial position, Congressional leaders will be hindered in their efforts to tailor more fundamental programmatic reforms.
Continued improvement in housing and labor markets will boost the FHA balance sheet, but the unobserved credit risk in the FHA portfolio is so high that we should expect high losses even while those markets strengthen. A taxpayer bailout may well be the price that must be paid when risks are downplayed and losses underestimated.
Gyourko is an adjunct scholar at the American Enterprise institute and the Martin Bucksbaum Professor of Real Estate, Finance and Business & Public Policy at The Wharton School of the University of Pennsylvania. He also serves as director of the Zell/Lurie Real Estate Center at Wharton and is chairman of the Real Estate Department.