Is the Treasury's Foreclosure Prevention Program a Failure?

During the Whitewater investigation, President Bill Clinton was famously quoted quibbling over the meaning of the word "is," saying, "It depends upon what the meaning of the word 'is' is." Perhaps the Treasury could take a similar approach when asked whether its foreclosure prevention program (HAMP) has been a failure. It depends upon what the meaning of the word "failure" is. Although it has failed in many ways, perhaps it has not in every way.

Where It Has Failed

The Treasury released its report (.pdf) for September today. The program is still struggling to find new applicants and to convert trials to permanent modifications. Here's the latest chart showing its progress, or lack thereof:

HAMP cht1 2010-09.png

You can see that new trials have held pretty steady in the 20,000 to 25,000 range for several months now. The number of new permanent modifications has declined back to this level as well, with the fewest in September so far this year. Meanwhile, the total cancellations are still higher than either of these figures at 46,455.

Cancellations have plagued the HAMP program. Through September, 699,924 trial modifications had been canceled. That's more than the 495,898 modifications that had been made permanent. And 29,190 of those had been canceled as well. That means, of its 1.37 million trials started, 53.2% have been canceled, while 34.1% have been made and remain permanent.

It's hard to spin this in a positive way, especially considering that the Obama administration initially announced that the program would help between 7 and 9 million borrowers. Even assuming the low end of that range, the program is only 6.7% of the way there more than 18 months later. Its current course implies that it won't come close. There are only 173,592 trials still active, with only around 25,000 new trials being added each month. Its only hope now is that new principal reduction programs succeed. They may begin to have an impact this quarter.

Where It Hasn't Failed

It's probably a stretch to say that this program has succeeded in any way, but it hasn't failed terribly in one way: re-defaults are relatively low. According to the September report:

At nine months, nearly 90% of homeowners remain in a permanent modification, with 11% defaulted. Fewer than 16% of permanent modifications are 60+ days delinquent.

This chart it includes shows just how few of the permanent modifications have so far re-defaulted (click for bigger version):

HAMP Re-Default cht 2010-09.png

If you know much about mortgage modifications, then you know many are destined to fail. Yet, those which make it all the way permanent through HAMP have been doing fairly well. Even a 25% default rate is a decent result, and that's the statistic for 60+ days delinquent for the permanent modifications that are a year old. Some of the younger vintages are trending even better.

Of course, HAMP's ultimate fate is not yet known. Over the next several years, more of these loans can, and almost certainly will, re-default. But at this time, the numbers look relatively low. But it also probably matters how you define re-default. If you include all of the trial modifications that failed, then the numbers would look much, much worse. But considering that these mortgages are only officially modified once they achieve permanent status, it's probably fair to consider the statistics above as legit.

Then, there's just the question of cost. Although $75 billion was allocated to the HAMP program from the bank bailout fund, it's unclear how much of that money has been or will be spent. If half has been spent, then that would come out to around to a cost of around $80,000 per successful permanent modification. That would be pretty expensive. Unfortunately, the Treasury does not provide this level of detail in its monthly reports.

So has HAMP been a failure? It depends on your point-of-view. It certainly hasn't lived up to its expectations, but the Treasury has managed to create a rather selective program that has so far experienced relatively few re-defaults. That might not be much, but it's something.