GAO Report on Foreclosure Reviews Misses How Regulators Conspired with Banks Against Homeowners

I suppose one has to be grateful for any official pushback against failed regulatory initiatives, such as the just-released GAO report criticizing the Independent Foreclosure Reviews. Of course, in this instance, I am charitably assuming that these reviews were a failure. They have certainly proven to be an embarrassment to the lead actor, the OCC, which has tried to maintain as low a profile as possible on this topic rather than offer any defenses.

But “failure” assumes that the OCC and the Fed did not achieve their real objective, which was to protect the banks. That hardly appears to be the case. The short story of the reviews is that to dampen down criticism of the many foreclosure horrors revealed in the media and in courtrooms all over the country, borrowers who were foreclosed on or had foreclosure actions underway in 2009 and 2010 were promised an independent review and compensation if they were found to have suffered financial harm. And even though the abrupt termination of the reviews has left the regulators with a lot of egg on their face, the result is that the banks paid a lot less than if the reviews had lived up to their billing.

Maxine Waters, to her credit, tried to put a little heat under the OCC and Fed by requesting that the GAO look into the matter. But as Dave Dayen stressed in his commentary on the GAO report, the overseer was blinkered in its approach:

Furthermore, its narrow scope – GAO only looked at the regulators’ design and ovesight of the foreclosure reviews, rather than what the independent reviewers did, and in fact they used the bank consultant reviewers as primary sources – tends to give a very circumscribed picture of the reviews. You could even say that this report will help get the bank consultants off the hook by putting the blame on OCC and the Fed.

The biggest problem, though, is the GAO was tasked only to do a very high level review of process, which meant it looked for how procedural weaknesses led to bad outcomes. It did not question the intent of the review, nor did it examine at how the reviews operated in practice, as opposed to theory (remember, the OCC relied on interviews of the consultants and questionnaires to them; there was no checking of the consultants’ processes or guides with independent experts, and the odds are very high that the only personnel that the GAO met from the consultants were individuals who would be attuned to and protective of their firms’ interest).

Some of the gaps in the report are simply stunning. For instance, the GAO points out at several junctures the intent of the exercise:

According to regulators, the goals of the foreclosure review were for consultants to identify as many harmed borrowers as possible, to treat similarly situated borrowers across all 14 servicers similarly, and to help restore public confidence in the mortgage market.

The GAO never considered that these goals are in conflict. If the reviews had indeed exposed the full extent of the rot in servicing, it would have undermined, not increased confidence in the mortgage market. It is obvious that the OCC either believed the bank PR that borrowers were deadbeats and complaints, for the most part, were simply the creation of clever foreclosure defense lawyers, or they had an inkling that there were serious failings, and so Potemkin reviews were necessary to shield the banks from liability. If they could claim to have made a meaningful investigation and found little amiss, that would undermine borrowers’ efforts to get a hearing in courts and to press for tougher curbs on servicers

Another conflict the GAO never considered was conflicts of interest; in fact, the word “conflict” does not appear once in the entire document. Yet as numerous independent parties pointed out from the very outset, the structure of having miscreant banks select and pay directly for “independent” reviews turned them into “bought” reviews. Sheila Bair described what happened when she was pressed for management changes at Citigroup in the wake of a bailout in the form of guarantees on $306 billion of toxic assets. A consultant was brought in to shield CEO Vikram Pandit:

When the “independent consultant” report came back in the fall, it compared Pandit to small European bank CEOs and gave him glowing marks. As for its review of the rest of Citi’s management, it gave high grades to Pandit loyalists while criticizing those who were not viewed as part of the Pandit team…

That was my first and last experience in asking bank consultants to assist regulators in reviewing bank operations. They are hopelessly conflicted, given their desire to secure future consulting work at those big banks. The consultants clearly considered their primary client to be Vikram Pandit. Indeed, they reported to him regularly on their review and sought his input until we found out about it and objected…..But Citi’s primary regulators, the OCC and NY Fed, didn’t seem to mind one bit.

While the GAO took its signals from the OCC and Fed and didn’t question the true interests of the consultants, Waters is not taking the matter lying down and is introducing legislation this week to curb this form of putting the foxes in charge of the henhouse.

A third major gap in the report is its failure to look at the borrower-requested reviews in any meaningful way. The report focuses mainly on failings that made the reviews overly labor-intensive and inconsistent. For instance, it criticizes the regulators for failing to talk to community groups, housing counselors, or other stakeholders. It also spends a great deal of time on problems with the sampling methodology for the loans that were to be examined in addition to the ones where a review was requested. It also discusses how the reviewers reinvented the wheel in terms of the state law elements of the reviews, where different servicers came up with different approaches and answers. For instance, at least one didn’t look at state law rules on fees at all but merely relied on “investor guidelines” meaning Fannie/Freddie/FHA/VA requirements. But the OCC consent orders specifically required that the consultants examine compliance with state law. It’s astonishing that the GAO blandly reports this “investor only” review as an inconsistency, rather than a blatant failure to adhere to the consent order requirements.

The GAO similarly mentions that the reviewers expected from 0 errors on certain types of loans and the highest level of errors anticipated was 10%. Huh? These are astonishingly low assumptions. The GAO does discuss how overly low assumptions can affect sampling. But it failed to consider how these assumptions served as “anchoring,” a well known cognitive bias. An assumption of low error rates would lead the consultants, even if they really had been independent, to have trouble with results that depart significantly from your assumptions. If you expect a 0 error rate and you find 15% of the sample to have problems, you’ll assume the problem is your sample or your error definition.

But what I found most striking was the way the GAO managed not to talk at all about how the borrower reviews were conducted. Yes, they did talk about the problems with borrower outreach, and also discussed in some detail how the regulators haven’t said much about how borrowers who asked for a review will be compensated, and they don’t look to have come up with a way to assure that similarly-situated borrowers at different servicers will be treated the same. But there was absolutely no consideration of the issues exposed by our whistleblowers: that the consultants and the servicers were working hard to suppress any findings of harm, when the evidence of harm was widespread.

And this gets back to a basic question: why was the sampling being done at all? Remember, in the engagement letters, the effort devoted to the sampling was roughly the same as that expended on the borrower letters. Why did that make any sense? If you had adequate borrower outreach and education as to what types of harm would be eligible for compensation, why would you need the sampling, or at least sampling of that scale? Absent any explanation, it’s hard not to imagine the sampling was intended to come up with low error rates (as in confirm the low expected error rates) which would then be used to justify low findings of harm in the borrower letters.

But we are past that point, so the GAO pulls the veil and focuses on the mess that is left in the wake of the abruptly-shuttered reviews, which is also pretty ugly. Consider this section of that discussion:

In most cases, servicers, with regulators’ approval, have engaged the third-party consultants to review borrowers’ files in two categories (SCRA and foreclosed borrowers who were not in default) to determine whether borrowers experienced those specific types of harm. According to one third-party consultant, at the time of the agreements that led to the amended consent orders, consultants were waiting on additional guidance from regulators to complete aspects of these reviews.

Now you probably missed the “gotcha” in that section” “borrowers who were not in default”. Shouldn’t that include borrowers who had gotten modifications and were complying with the terms of the mod, yet were foreclosed upon? There are tons of cases like that, where payments were misapplied or where the bank simply started refusing to accept payment even though a mod had been executed. There are also instances of banks refusing payments from borrowers who were current and proceeding to foreclosure (I’ve just heard of a new case like that and I may be writing it up). Yet at Bank of America, Promontory deemed borrower who were having their checks returned by the bank not to be making payments! So with this sort of Catch 22 going on (presumably “if you were foreclosed on, you must have been in default”) you can rest assured that “borrowers who were not in default and were foreclosed upon” will be as scarce as unicorns.

The OCC’s excuse is that this new approach will be less inconsistent than the one that had been in place:

As discussed earlier, regulators had limited and unsystematic centralized control mechanisms to monitor consistency among the foreclosure review processes and did not have the information to assess the implications of any differences. According to regulators, achieving consistent results for borrowers, so that similarly situated borrowers receive similar payment amounts, is a goal of the amended consent orders, as it was of the foreclosure review process. OCC staff stated that the direct payments provided under the amended consent orders will likely be more consistent than what would have occurred under the foreclosure review because servicers are using a standard framework and objective criteria to categorize borrowers and all borrowers in a particular category will receive the same payment amount.

But even then, the GAO isn’t convinced:

Without using mechanisms to centrally monitor the consistency of servicers’ activities to categorize borrowers, regulators may risk delays in providing direct payments to borrowers and inconsistent results.

So even where the GAO does look, what it finds is pretty ugly, but you have to read through bureaucrat-speak to discern the implications. I hope Sherrod Brown is in the mood in his hearings later this week to put the consultants and regulators on the spot. It will take some determined probing. The obfuscation is almost always thicker when there is more to hide.