Comment Rescue – The Regulation Edition

Alternatively titled  “If you want regulations, make sure the rules are crystal clear”

zic gets the honor of today’s comment rescue:

Pointing out the good side of regulation — how it protects small people (despite the concern trolling that always happens about how it’s the little guy screwed) is another.

And a third thing I rarely hear liberals discuss is, you know, actually repealing bad regulation and lightening the regulatory burden were appropriate. We should be looking for those opportunities.

Look no further.  I have just the opportunity she’s looking for: loan buybacks as they apply to the government sponsored entities (“GSEs”), specifically Fannie Mae and Freddie Mac. Some background is in order.

Subject to specific underwriting guidelines, GSEs source their mortgages by buying them from banks.  Given the sheer volume of loans that the GSEs buy, it is functionally impossible and very cost prohibitive for the GSEs to conduct loan level due diligence on each and every loan that it buys prior to purchase.  In lieu of due diligence, since the GSEs want to protect their downside risk against banks selling bad loans, the GSEs require lenders to provide representations and warranties with respect to the quality of those loans.

A loan buyback is the remedy that the GSEs have as recourse against banks that sell loans to the GSEs that are later to not be as they were represented for whatever reason (defaults, fraud, etc.).  In these instances, the GSEs have the right to demand that the banks repurchase bad loans at full value, and the banks must comply.

The loan buyback provision typically expires 36 months after the purchase of a loan; however, to further protect the GSEs against certain kinds of bad acts, in light of what happened in 2008, the regulatory framework includes “life of loan” exclusions that extend the loan buyback provision throughout the life of the loan whether the loan is two years old or 25 years old.  Life of loan exclusions fall under six categories: 1) misrepresentations, misstatements and omissions; 2) data inaccuracies; 3) charter compliance issues; 4) first lien and title matters; 5) legal compliance violations and 6) unacceptable mortgage products.

Let’s go back to zic’s comment.  Setting aside whether or not the federal government should involve itself in the mortgage business (**), the mission of both Fannie Mae and Freddie Mac is to provide liquidity, stability and affordability in the housing markets.  Banks lend through their Fannie Mae and Freddie Mac programs, and in many cases are able to provide loans that are more affordable than through their own balance sheet programs.  The regulatory framework that forces the banks to buy back bad loans helps to provide the banks and the GSEs with a level playing field, protects the GSEs core mission and allows banks to transact in a reasonable environment.   Perhaps indirectly, it protects “the little guy” from predatory lending practices because the banks are not able to pawn off bad loans.  Yes, this a good regulatory framework.  I think it’s very reasonable.

In theory, this is great; however, the banks are seeing problems with they way this has been put into practice: :

Despite these positive steps, life of loan exclusions undermine the existing sunset through ambiguous terms that can be used to force a repurchase even after the sunset. The lack of an independent dispute resolution process further undermines the sunset framework because the GSEs retain their status as sole decision-maker and arbiter of disputes, despite their financial interest in the outcome. Further reforms are necessary to clarity these problematic provisions and reduce the credit overlays that have restricted access to credit.Continued aggressive use of repurchase requests since the financial crisis has left lenders uncertain of their ultimate exposure once a loan has been sold to a GSE. This uncertainty has led to credit overlays as lenders seek to avoid having to repurchase loans for “foot fault” violations.

The banks aren’t complaining about the framework, but rather the ambiguity of the rules that have been developed out of the framework.  Rules require clarity so that those subject to them have clear expectations of what to expect when they transact in the markets.  Consider the following:

1.  Should a bank have the right to cure misrepresentations, misstatements and omissions or data inaccuracies if such defects are considered immaterial in nature and have no bearing whatsoever on the underlying quality of the loan (i.e. a missing page, a mistyped address)?

2.  If in the event there is a good faith dispute over a loan buyback, how should those disputes be handled?  Should a neutral third party oversee the dispute and issue a ruling binding to both parties?

3. Should the banks and the GSEs establish a clear set of rules dictating specific remedies given specific defects (i.e. Given Defect A, Remedy B is required, etc. etc.)?

Given that these issues were not addressed up until recently and the GSEs have been hyper-aggressive in their approach to loan buybacks through interpreting the rules in the broadest manner possible (as is their right – buybacks totaled $81.2 billion between 2011 and 2013 (source), this has placed the banks in a difficult situation.  Yes, the Fannie and Freddie loan programs are profitable; however, banks also have to manage their risk exposure.  Part of that risk exposure is being able to as accurately as possible forecast their exposure to loan buybacks.  When the banks are at risk of being forced to buyback a loan based on a misrepresentation that has no bearing on the underlying quality of the loan, as the ambiguity of the life of loan exclusions, provides, as generally risk averse institutions (***), they will pull back on their loan underwriting, employ credit overlays and/or provide loans to only the very best borrowers within their underwriting parameters.

Once again, I return to zic’s comment. There is a clear opportunity here.  While there’s no need to repeal bad regulation in this case (the regulations are fine IMO), there is a regulatory burden on the banks that should be lightened in a way to still protect the GSEs against bad loans while giving banks more comfort that the GSEs won’t force buybacks in situations that do not warrant it, even if loan documents contain the occasional de minimis error.

Here, a regulatory framework has produced a set of rules where one party needs more clarification in order to be willing to further transact.  Perhaps some of the more anti-corporate or anti-bank among the commentariat here would rather see the private sector say “how high” every time the regulators “say jump”, but that’s not how it works.  The banks aren’t forced to sell to the GSEs and the GSEs rely heavily on the private sector in order to help fulfill their respective missions.  Banks are certainly not going to transact with the GSEs if it leads to undue risk to their own balance sheets through the inability to appropriate manage buyback risk.

The Federal Housing and Finance Agency has taken notice.  A few weeks ago, Melvin L. Watt, Director of the Federal Housing Finance Agency (“FHFA”), addressed certain issues related to loan buybacks in a speech at the annual Mortgage Bankers Association in Las Vegas:

…The current life-of-loan exclusions are open-ended and make it difficult for a lender to predict when, or if, Fannie Mae or Freddie Mac will apply one of them.  

…we are more clearly defining the life-of-loan exclusions, so lenders will know what they are and when they apply to loans that have otherwise obtained repurchase relief… 

…The Enterprises will provide details about the updated definitions for each life-of-loan exclusion in the coming weeks, but let me spend a minute highlighting some aspects of the refined definitions for the first two categories – misrepresentations and data inaccuracies…    

…In defining both of these categories, we are setting a minimum number of loans that must be identified with misrepresentations or data inaccuracies to trigger the life-of-loan exclusion.  This approach allows the Enterprises to act when there is a pattern of misrepresentations or data inaccuracies that warrant an exclusion, but not to revoke repurchase relief they have already granted if they subsequently discover that a lender incorrectly calculated the debt-to-income ratio or loan-to-value ratio on a single loan…  

We are also adding a “significance” requirement to the misrepresentation and data inaccuracy definitions.  In order to require repurchase of a loan under the misrepresentation or data inaccuracy categories, the “significance” test requires the Enterprises to determine – based on their automated underwriting systems – that the loan would have been ineligible for purchase initially if the loan information had been accurately reported…  

Under the revised and modified Framework, the Enterprises will retain their ability to conduct quality control reviews at any time, of course, because this is essential to their risk management practices and is essential to their ongoing safety and soundness…

After FHFA and the Enterprises release the details shortly on these life-of-loan exclusions, there still remains more work to be done on our Representation and Warranty Framework.  On the origination side, FHFA is already focused on developing an independent dispute resolution process.  We are also identifying cure mechanisms and alternative remedies for lower-severity loan defects… 

While nothing has been finalized, this is a positive step in the right direction, and it addresses the issues I raised in my list of three questions above.  With respect to my first question, immaterial errors will not automatically trigger the loan buyback request until a certain number of loans are found containing errors.   This eliminates issues with single loans yet protects the GSE against shoddy underwriting being done on a more systemic basis, a far greater concern than isolated yet inconsequential data errors.  Also, there is no change to the GSEs right to force a loan buyback in the event of material errors.  Based on the above, the issues I raised in questions two and three are currently being addressed.  Again, nothing has been finalized as far as I can tell, but at the very least, the GSEs can stop forcing buybacks based on immaterial errors soon if not immediately.

I think these are reasonable solutions to legitimate issues.  Not everyone agrees.  David Dayen believes that what the banking industry is doing is effectively holding the government hostage in an attempt to roll back mortgage regulations.  While it’s standard fare for any industry group to oppose a regulatory framework that prevents members within that industry from making money, I don’t think Dayen makes a convincing case.  Nothing that Director Watt mentioned in his speech had anything to do with the regulatory framework and under no conditions are the GSEs exposed to additional risk of fraud. The threat of buybacks is still there, with the only difference being that it will be applied only to those situations where it is most appropriate (a blank line where a name is supposed to go is not the kind of inaccuracy that is appropriate grounds for a buyback).  One last point:  Dayen;s claim that the banks are seeking to expand risky lending fail because, with the exception of the down payment reduction addressed by Vikram Bath, which I personally don’t think is that big of a deal ****, is incorrect as the GSEs have not altered their underwriting standards.

Here, I see valid solutions proposed in order to address valid concerns.  I don’t see the banker boogeymen nor do I see anyone – liberals, conservatives or libertarians – talking that much about this.  Zic is right, but it doesn’t only apply to liberals.  Today can be one of those days where we can all discuss it.  I’m up for a conversation.

** – a conversation that I know will not take place in the combox because, well, oh never mind.  Who am I kidding?

*** – don’t choke. I’m not talking about the i-banks.  Commercial banks and other balance sheet lenders (i.e. life companies) approach mortgage lending very conservatively.

**** – I’m less concerned about the recent change allowing Fannie and Freddie to buy loans with as little as 3% because I’m not as bearish about lower down payments given the current mortgage regulations (i.e. ability to pay) and Fannie and Freddie’s willingness to pursue loan buybacks.  

Picture courtesy of Wikipedia Commons

Update – Thanks to a good point made by nevermoor, I need to make a clarification.  I wrote:

a blank line where a name is supposed to go is not the kind of inaccuracy that is appropriate grounds for a buyback

This holds true so long as the exclusion/error is either insignificant or can be remedied by the lender by providing the correct information.  I did not mean to imply that the GSEs can or should ignore significant errors.