On August 31, the Urban Institute posted an article by Laurie Goodman titled, “Squeaky-clean loans lead to near-zero borrower defaults—and that’s not a good thing.”
The article begins, “There’s something interesting and important going on in the mortgage market today: borrowers who took out mortgages in the past five years have rarely defaulted, making them better at paying their mortgages than any other group of mortgage borrowers in history.” After giving details on the credit quality and performance of the loans Fannie Mae and Freddie Mac have either purchased or guaranteed from 2011 through the first half of 2015—which she notes “are hardly defaulting at all”—Goodman states, “The performance of mortgages originated over the past few years…suggests that there is plenty of room to safely expand the credit box…Put simply, it’s time to lend again to borrowers with less-than-perfect credit.”
The data highlighted by Goodman on the credit quality and performance Fannie and Freddie’s recent books of business are no mystery; I’ve cited them frequently, most recently in explaining the dramatic improvement in the companies’ projected credit losses in the Dodd-Frank Severely Adverse stress test over the last two years. And I agree with Goodman’s recommendation that because of this credit performance, what she and others call the “credit box”—the product and borrower characteristics deemed to present acceptable risks to lenders or credit guarantors—can and should be opened up, to expand the range of borrowers served by the mortgage finance system. But there is a further critically important observation to be made about these credit data that is not in Goodman’s piece: their striking incongruity with the overwhelming consensus among mortgage reformers—most recently documented in the Urban Institute’s own series of essays submitted for its “Housing Finance Reform Incubator” series—that for the mortgage finance system to operate safely, Fannie and Freddie must be replaced. What Goodman’s article makes clear, without saying it directly, is that today’s mortgage market, with Fannie and Freddie at its center, is in fact too safe. The obvious question this raises is, “What, therefore, are we supposed to be reforming, and why?”
There is considerable irony here. To make the point about how “squeaky-clean” Fannie and Freddie’s 2011-2015 loans are, Goodman notes that the default rate on these loans “is tracking well below [the companies’] mortgages originated from 1999 to 2003, a period of reasonable underwriting standards and fairly low default rates.” So, what happened in 2004? Oh, yes; that was when private-label securities (PLS) surpassed Fannie, Freddie and Ginnie Mae securitization combined as the primary source of residential mortgage financing. PLS placed virtually no limit on the riskiness of the loans they would accept, and as a consequence underwriting standards collapsed throughout the industry. The 2005-2008 books of business of all lenders performed far worse than any years since the Great Depression, and Fannie and Freddie’s were no exception (although we now know that the loss rates on the companies’ mortgages from this period were one-third those of banks, and less than one-tenth those in PLS).
Even as the 2008 financial crisis raged, opponents of Fannie and Freddie ignored the role played by PLS and asserted that the companies were its cause. Treasury forced them into conservatorship in September 2008, and since then there have been constant calls for them to be wound down and replaced. Non-cash expenses booked by their conservator, FHFA, ballooned Fannie and Freddie’s losses and caused them to have to take $187 billion in senior preferred stock they didn’t need and weren’t allowed to repay. The only aspect of Fannie and Freddie’s business that has remained within their control following the conservatorship is the quality of the loans they purchase or guarantee. With the false claim that their lax underwriting caused the financial crisis, calls for them to be put out of business, and the huge financial losses unrelated to their business risks imposed upon them by their conservator, it’s no wonder that the companies tightened their credit standards so dramatically. What else could they have done to be responsive to the criticisms being made of them?
Fannie and Freddie also now have little control over the guaranty fees they charge, since those are based on notional capital guidelines set by FHFA. While FHFA does not disclose what its capital guidelines are or how they are determined, published data on the companies’ target guaranty fees by risk category are consistent with an average capital requirement of about 3.5 percent. Fannie and Freddie’s implied capital and guaranty fee pricing have remained relatively constant since 2013, when FHFA Acting Director Ed DeMarco required them to raise their guaranty fees by 10 basis points not because of the riskiness of their loans but to “encourage more private sector participation” and to “reduce [their] market share.” Mel Watt has not changed FHFA’s capital guidelines since he became Director in January 2014, and having to set guaranty fees consistent with an average capital requirement of 3.5 percent has caused Fannie and Freddie to price many lower-credit score and other higher-risk borrowers out of the market, exacerbating the effect of restrictive underwriting on the size of the credit box.
Ensuring the smooth workings of the mortgage finance system for as many potential homebuyers as possible should be the primary goal of mortgage reform. Yet that is frequently not the case. For a sizable majority of interested parties, since 2008 the term “mortgage reform” has been a code phrase for replacing Fannie and Freddie with some favored alternative. While pitched to the public as “the last major piece of unfinished business of the financial crisis” (to quote Senator Corker), this agenda-driven approach to “reform” has from day one been a solution in search of a problem. And as more time passes, that approach becomes harder to defend. Today, after eight years in which to assess and understand what actually did happen to cause the mortgage crisis, and nearly as long a period of unprecedented restrictiveness in credit conditions, it no longer is credible to claim that replacing Fannie and Freddie with an untested alternative, forcing them to do mandatory risk sharing (of what little credit risk there currently is) or requiring them to hold “bank-like” levels of capital could possibly fix what now ails the U.S. residential mortgage market.
As Goodman points out in her piece, the real problem with that market is the very small size of the credit box. Underwriting certainly has played a role in this, but more important, I believe, is the deliberate lack of guidance from regulators as to how to assess, capitalize and price mortgage credit risk. Prior to the crisis, Fannie and Freddie had defined capital standards—both risk-based and minimum—along with full control of their underwriting processes. With certainty about their required capital they were able to quote lenders a single guaranty fee that covered a diverse range of products, borrower types and risk characteristics, as determined by the lender. As long as lenders met Fannie and Freddie’s underwriting requirements—which could be confirmed using their automated underwriting tools—those loans would be accepted. The credit box was wide, full, and affordably priced.
The financial crisis revealed that all mortgage lenders were undercapitalized (and that private-label securitization, which relied on assessments of credit risk by rating agencies that had a financial incentive to underestimate it, was untenable). Banks have been given updated capital requirements by their regulators, but Fannie and Freddie have not. Their capital requirements are being held hostage to the politicized version of the “reform” debate. Initially, opponents of Fannie and Freddie insisted that they hold an absurdly high (10 percent) amount of capital to back any future credit guarantees they did. More recently, Treasury has used the companies’ absence of capital (almost all of which it has taken in the net worth sweep) as an excuse to avoid addressing the issue of what their capital should be. And FHFA, for its part, has played along with this charade by adding massive amounts of non-cash expenses to their Dodd-Frank Severely Adverse stress tests (as discussed in the last post) to create the impression that the companies would need tremendous amounts of capital to do even the “squeaky-clean” business they’re doing now.
This gamesmanship over Fannie and Freddie’s capital has had real-world consequences. The size of the credit box and risk-based capital are very closely linked. Until we can get the capital right, we won’t get credit box right. And the key to enlarging the credit box at Fannie and Freddie—really the only sure way to do it—is to give the companies true risk-based capital requirements, designed to strike a careful and deliberate balance (as determined by policymakers) between the cost and breadth of access to mortgages on the one hand and taxpayer protection on the other.
FHFA is in a perfect position to take the lead on this initiative, by running an honest stress test of the companies’ business by risk category. But to date Treasury has not permitted FHFA to do that, I believe because it knows that if Fannie and Freddie were given an updated and legitimate risk-based capital standard, Treasury and others would lose the ability to pretend that the companies need so much capital to operate safely that it is in everyone’s best interest to replace them.
This issue is so important to opponents of Fannie and Freddie that new attempts to muddle their capital situation continue to crop up. Most recent is the claim that because banks with more than $50 billion in assets must hold capital equal to 5 percent of their risk-weighted assets (about 3.5 percent of their total assets) after having absorbed losses from their Dodd-Frank stress test, Fannie and Freddie should be required to hold an equally large capital cushion after their stress tests have been run. But this assertion ignores a fundamental difference in the business models of banks versus Fannie and Freddie, as well as the experience of both sets of parties during the financial crisis.
What caused the near-failure of the banking system in 2008 was not that banks’ credit losses exceeded the amount of capital they had; it was the fact that at some banks—particularly the larger ones—losses rose so much so quickly that borrowers withdrew non-insured deposits and many investors refused to roll over short-term funds placed in those banks. Without the trillions of dollars made available by the Fed and the Treasury (on very favorable terms), the affected banks would have been forced to sell large amounts of assets at depressed prices, and that would have wiped out their capital. Banks are highly leveraged institutions, and in addition to non-insured deposits a typical bank has between 10 and 15 percent of its assets in short-term purchased funds. Both can flee quickly in times of stress. It is precisely to retain the confidence of non-insured depositors and investors who provide “hot” money that the Fed requires the banks it subjects to the Dodd-Frank stress tests to pass them with such a considerable margin of safety.
Credit guarantors such as Fannie or Freddie are in a completely different position. They have neither deposits nor hot money on their balance sheets subject to runs. (The short-term debt Fannie and Freddie have is associated with their portfolio businesses, which have had and should continue to have capital requirements different and separate from the credit guaranty business). And when large numbers of their mortgages do fail, it happens relatively slowly. Unlike a bank, if a credit guarantor has enough capital to cover its losses during a stress environment, it can stay in business. We saw that during the financial crisis, when Fannie and Freddie were able to remain profitable on an operating basis (that is, excluding the non-cash expenses added by FHFA) even after the spikes in their credit losses.
In contrast to the arguments for excessive Fannie and Freddie capital requirements, the problems with the credit box are real. They are based on economics, and their resolution will depend on politics. Advocates for replacing Fannie and Freddie with mechanisms they favor will continue to cite stress tests padded with non-cash expenses, and liquidity cushions with no rationale in practice, as the basis for capital requirements that make the companies’ credit guarantees noncompetitive. Yet if policymakers fall prey to these ploys—and either replace Fannie and Freddie with a less efficient secondary market mechanism or force them to hold capital unrelated to the risks of mortgages they guarantee—we won’t solve our credit box problem. Instead, we’ll end up with a “reformed” mortgage system in which some financial institutions are more profitable, but large segments of the home-buying population remain underserved, and all segments pay more for their loans than they should. It’s a simple choice: we can solve the real mortgage problem, or one that’s been made up.