The OCC is bravely trying to spin the horrorshow of its botched foreclosure reviews as some sort of positive outcome. It is apparently now trying to present its dereliction of duty in figuring out how to compensate borrowers as no big deal.
But the most amazing finesse, which a New York Times article tacitly accepts, is the OCC’s pretense that it got a good deal for homeowners in the settlement.
Just focus on the cash portion, which is $3.3 billion across the ten servicers in the settlement. The other forms of relief, paralleling the state attorney general/Federal settlement, either aren’t worth much (writing off deficiency judgments) or are for things the banks were inclined to do anyhow.
495,000 complaint letters were filed. The estimates of serious harm from the whistleblowers at the Bank of America site in Tampa Bay ranged from 10% to 80%. The average was 33%, and the estimates also clustered around 30% to 40%. So we’ll use 30%.
To make the math simpler, we’ll use 500,000 x 30% x the maximum award, which was $125,000, which would seem to be warranted with “serious harm” (the people on the modification test all described cases where people who were in mods of various sorts and were paying as the bank stipulated but were foreclosed on, so they seemed to have an adequately stringent notion of what “serious harm” amounted to. Basically, to get the maximum award, the bank had to have eaten your home while you were in a mod or it had to be a Servicemembers Civil Relief Act violation).
You get $18.75 billion. Let’s say maybe the temps were too generous and their estimate is 1/3 too high. You still get $12.5 billion, nearly four times the amount for the banks to divvy up. And you’d have some less large payouts for the people not seriously harmed. If you would have qualified for a mod but the bank never processed your application, or were denied a mod incorrectly, that’s a $15,000 award. The folks who were processing mod complaints say they saw another 30% to 40% instances of less serious harm. So if you assume a $15,000 payout for another 20% (that’s conservative, the real number is probably closer to 30%), you get another $1.5 billion.
$14 billion, which is a conservative estimate of what the banks should have been required to cough up, is more than four times the only part that the OCC got that really matters, hard dollar payments to borrowers that suffered real losses. . No wonder the banks are perfectly happy to pay out $2 billion to consultants who made a mess of things. Those madcap consultant were as clever as foxes.
So the OCC got a turkey of a deal.
And the New York Times is taking up other OCC/bank messaging in its piece. It’s merely making the problem sound as if the issue is that the reviews were supposed to be independent, and now the banks are in charge of handing out the money to homeowners. The conflict issue is theoretically a problem, but it isn’t as big a problem given that we are talking about a fixed, if way too small, pot at each bank. The Times underplays the other major problem, that the banks don’t have enough of a grasp of who was really harmed to dole the money out fairly:
Washington is seeking help from an unlikely group in its effort to distribute billions of dollars to struggling homeowners in foreclosure: the same banks accused of abusing homeowners with shoddy foreclosure practices.
In doing so, the regulators are trying to speed the process after a flawed, independent foreclosure review delayed relief for millions of borrowers, according to people briefed on the matter. But housing advocates worry that the banks, eager to end the costly process, could take shortcuts as they comb through loan files for errors, potentially diverting aid from the neediest homeowners.
Regulators say they will check the work. And banks have already agreed to pay a fixed amount to troubled homeowners, creating another backstop.
Check the work!?! The work was not done and not close to being done.
Do you get the circularity of this reasoning? The reviews were shut down because the consultants couldn’t figure out an efficient process for doing the reviews. The whistleblowers at Bank of American said they were told it would take two years to complete the reviews using the process finally put in place (the most complex test was not put into production until the end of August). Promonotory, Bank of America’s “independent” consultant, put in an “optimization” the week before the project was shut down that was supposed to speed the process, but that was never road tested.
So how many loans were officially completed, as in the temps did all their work and Promontory reviewed them? I’m told a grand total of 4,800 as of December. That is compared to a universe of at least 121,000 letters, the total as of around December 10. If that 4,800 was as of the end of December, it would be relative to a universe of more like 140,000 to 150,000 loans. So the completed loan reviews were 3.2% to 3.9% of the total. And the OCC really believes Bank of America can make informed decisions on who really deserves the funds?
Now there might be places where the work could be completed. But temps were told to leave the premises as soon as the settlement was inked. From what I can infer, the temps look to have completed the tests on 40% of the requested reviews, but one well informed reviewer was under the impression that only 20% were considered to be done from their end, suggesting that half of that total had not gotten through the quality assurance/rebuttal process*. So even this bit of intelligence confirms that the Bank of America were in even worse shape than figure previously disclosed, that of the banks having finished roughly 1/3 of the reviews by the time the process was shut down.
The New York Times article runs some other remarkable PR:
Under the plan outlined last week, the banks will pore over loan files like Ms. Lee’s to identify the worst possible errors. Military personnel illegally foreclosed on, for example, will rank highest on the list. Borrowers who might be current on their loan payments — and therefore did not warrant a foreclosure — will be next.
Regulators will then decide how much money to pay each category of borrower. The worse the errors, the bigger the payout.
This sounds fine if you have no idea of what was going on. What does “might be current on their loan payment” mean? That was monstrously ambiguous at Bank of America for anyone who had a modification. Was someone who was on a trial mod and paying on time current? By common sense, they would be, but Bank of America had only one category that was tagged as current, that of people who did not need special servicing. Anyone who had even missed one payment was put in the “foreclosure” category. Modifications were another “not current” category, even if they had gotten a permanent mod. A temp told how someone who had a husband who had cancer, and was 15 days late on one payment on her permanent mod was foreclosed on two months later.
And even if the OCC takes the view that people who were current on trial mods should be treated as current, the BofA records often don’t reflect that properly. Contractors also report that trial mod payments were put in suspense accounts rather than credited to the mortgage. If a loan got through the G test, the bank and the OCC might pick that up, but as we indicated, it’s pretty much certain that well under half the borrowers that sent letters complaining about mods had a first pass file review on that issue.
In addition, it isn’t hard to imagine that Bank of America would (at most) only want to use the tests that had been completed by the temps, but also gone through the official pushback process known as “quality assurance.” So a much smaller portion of the work completed by the temps would be deemed as usable. This is the only thing we get in the way of an indication of serious problems, and it comes a full 15 paragraphs in, after we get a touching story of a homeowner who really needs the money pronto.
The strategy, though, presents potential conflicts of interests. The banks, in haste to meet tight deadlines, could fail to provide an accurate portrayal of what went wrong. The loan files are also in disarray, officials say, complicating the task for banks.
“The whole process has been a slap in the face to homeowners and a slap on the wrist to banks,” said Isaac Simon Hodes, an organizer with the community group Lynn United for Change. “The latest development shows how there has been no accountability.”
Too bad the Times doesn’t clue readers as to the amounts at issue. $3.3 billion divided by 500,000 people is $6,600 a person. That’s not nothing, but if you lost your home and have good reason to think the bank was responsible, it’s an insult.
Not only is the Times far too forgiving of the banks and the OCC on their framework for doling money out, but it also takes up consultant party line on what needed to be done in the reviews. Notice the contradiction in the section before and after the ellipsis:
Their efforts were stymied, in part, because regulators urged consultants to first scrutinize a random sample of the four million foreclosures before digging into specific homeowner complaints, the people involved said. The decision, the people said, may have undercut the scope of the settlement and potentially deprived homeowners of additional relief…
Some consultants say they sounded repeated alarms about the process. Last spring, a group of consulting firm executives met with comptroller officials in Washington to voice concerns that the reviews were too narrow, according to people with direct knowledge of the meetings.
Did you get that? The consultants weren’t able to do the two workstreams the OCC wanted completed, a review of a large sample of loans, plus a review of the the borrower complaints, yet they were champing to make the tests more elaborate.
As we’ve made clear, the foreclosure review process has been a complete fiasco. It’s been disappointing to see the mainstream media only give a timid and shallow account of how bad it was.
*Around December 10 that there had been 121,000 letters submitted through the IFR process, with a good guesstimate for the final number was 140,000 to 150,000.
A, O, S, B, D, and C tests had been completed for all the requests sent in as of the beginning of August, so, only small teams were being kept on those tests to process new letters coming in. Everyone who could be shifted onto E and F test (fees) were being moved to that. G test (mods) was the other bottleneck, but G test took less time than E and F tests and had nearly as big a labor pool.
An estimate how much of E and F test might have been done:
The first reviewer to go live on E and F test was on August 29. Pretty much everyone except new hires were certified by the end of September. For simplicity, assume an average start date for the total deployed E and F test workforce of mid-September. Allowing for various holidays, that means 15 work weeks to project finish.
I backed out a certain number of staffers of the available pool for E and F (Level II reviewers) at each center to stay on tests B and D. I was told 40 to 50 in Tampa Bay, which is roughly 3 teams, or 42 people. I used similar proportions in Westlake Village and Newark on our assumed staffing #s. That gets you to 477 people working on E and F.
I assumed 3 completed files a week, which was the expected production level.
That give you 21,465 completed files. A source said there were 42,000 loans to be tested for E and F before that final influx. If you add another 25% (high guesstimate for last minute submissions) you get a bit over 50,000. So that would mean more like nearly 40% done for E and F. I’d assume (maybe incorrectly) that E and F was THE bottleneck, since management seemed most fixed on it and it was the last set of tests to go live. But the flip side is even though G test could be done much faster (on average more than one a day), the pool of letters that needed to be reviewed for G test could easily have been more than twice as large as for E and F test.