Now that the Congressional Oversight Panel has wrapped up operations, it's not clear whether there will be regular government analysis of foreclosure mitigation efforts. SIGTARP and the GAO can produce reports on the topic, but it is uncertain whether they will. For the time being, at least, Treasury seems to be continuing to release HAMP performance data. Let's hope that they continue with this. It took the Oversight Panel considerable wrangling to get Treasury to concede to even this minimal degree of transparency.

Looking at the most recent data release, one can see why. HAMP mods are not holding up well (as I and many others predicted). Through March 2011, there are 670,186 permanent HAMP modifications commenced. Of these, 90,536 are from the previous three months. 83,270 of the permanent modifications have been cancelled, almost all because of defaults (1,126 mortgages got paid off).

Let's figure that HAMP mods aren't cancelled until they run 90 days delinquent. So we're looking at a cumulative default rate of about 14% (=(83,270-1,126)/(670,186-90536)).  Now this isn't the ideal statistic for gauging HAMP mods–ideally Treasury would release the cumulative default rate by mod vintage.  I'm guessing that this would show considerably higher default rates for older mods. The Congressional Oversight Panel was able to undertake this sort of analysis in its December 2010 report (see pp. 95-96) and found that although the cumulative default rate was only 6.9%, the redefault rate for modifications that were 12 months old (as measured by 90 days delinquent) was 21%.  So roughly triple the cumulative rate.  I am not confident that we can extrapolate from 14% to 42%, but I'm pretty sure that the redefault rate at 1 year is over 20% on HAMP mods and probably doesn't plateau at 1 year. I really wish Treasury would release data showing redefault rates by vintage of HAMP permanent modification; OCC/OTS statistics do so. 

To the extent there is good news, it is that HAMP is converting trial mods to permanent at a rate that surpasses cancellation of permanent mods–several months ago, the trend lines indicated that cancellations would surpass conversions by the end of 2010.

Now if HAMP were meant to provide real, long-term stability in the housing market, it's going to come up way short (and indeed, that was always clear to anyone who looked at the program objectively). But looking at the program now, I'm fairly convinced that it was always designed to be a kick-the-can-down-the-road endeavor. It wasn't meant to fix the foreclosure crisis, just to stabilize it and spread losses out over time, including postponing them until the financial system was healthier. And to the extent that was in fact the goal, HAMP might have been partially successful. But it's easy to be successful when you define success downwards.

[Update 1:  Treasury just posted a new HAMP report. I've updated the numbers here.]

[Update 2:  I just got a good query about what's so bad about 14% defaults–afterall, many have been predicting 40-50% default rates. If the only 14% of mods defaulted within 1 year of conversion to permanent, that wouldn't be so terrible.  But again, that's not what the 14% number is showing.  It is just the ratio of the total number of defaults on permanent mods to the total number of permanent mods that are at least 3 months old.  That's not a meaningful number because there isn't a straight line rate of redefault on mods.  The only meaningful statistic for an on-going program is to measure redefaults by vintage.]