by Tanta on 12/06/2007 01:39:00 PM

While we're all eagerly crowding around the teevee waiting for the Great Loan Modification Speech (can it get any nerdier than that? Can it? Sheesh) I want to comment on this little statistic, that is getting thrown around a lot:

Modified loans frequently re-default. Joshua Rosner at Graham-Fisher & Co. says 40% to 60% of subprime and Alt-A borrowers who have their loans modified end up defaulting anyway within the next two years. Fitch Ratings puts the recidivism rate at a slightly lower 35% to 40% for good modification programs.

Let us bear in mind that such statistics have to be based on loans that were modified no more recently than 2005 (newer mods would not have a 24-month post-modification history). It is quite possible, indeed it is likely, that modifications done in 2005 and earlier (when there were many more refi opportunities and most borrowers could sell their homes for at least the loan amount) were done for borrowers with problems like job loss or illness that either simply recurred or that created other (non-mortgage) debt problems down the road.

This is not to argue that modifications done now for loans originated in 2005 and after would perform better. Or worse. Or the same. It is to say that we are probably in uncharted territory and that "past history" was a lousy guide when we made the loans and might be a lousy guide when we have to work them out.

Furthermore, 40-60% is a very big range. In the absence of other information, I would certainly guess that most if not all of that variation is due to servicer quality, not borrower quality. If that is true, then performance of modified loans can be improved by something that is well within the control of the industry.

Now you can go back to waiting around the teevee . . .