Quantcast

Treasury Issues Correction on Mod Program Default Numbers

Paul Kiel | 8/12/2010, 1:13 p.m.

Last month, the Treasury Department released its latest numbers on the current status of homeowners who'd managed to receive a permanent modification through the government's foreclosure prevention program. The results suggested homeowners were doing miraculously well. But it turns out, the numbers were too good to be true.

On Friday, the Treasury released revised numbers that still show a relatively low level of default, but one significantly higher than the earlier estimate. For instance, Treasury had estimated that less than 3 percent of homeowners who began a permanent modification last summer had fallen more than three payments behind. That number turned out to be about 15 percent.

So how did the government get such an important metric for the program so wrong? All that's clear is that Fannie Mae, which works under a $113 million contract to administer the program for the Treasury Department, made mistakes when crunching the numbers. In a statement, Treasury spokesman Mark Paustenbach said the mistakes had been fixed, and both Fannie and Treasury had hired outside consultants to verify the new results.

The newly published default rate does remain encouraging. Modifications under the program's rules reduce homeowner payments on average far more than modifications offered outside of the program by banks, and as a result, a higher proportion of homeowners are expected to be able to maintain the payments. Early indications support that.

For example, a recent regulatory report on mortgages [3] shows that modifications begun in the second quarter of 2009 (all done outside of the government program) had defaulted in about 33 percent of cases after nine months. That's about twice the default rate the government program has shown so far.

Of course, it's worth keeping in mind that relatively few homeowners have actually emerged with permanent modifications through the program. Though about 1.3 million homeowners have begun the three-month trial period, fewer than 400,000 have emerged with permanent modifications. Far more have either been booted from the program or left waiting more than half a year for a final answer.