On October 4, the Federal Housing Finance Agency (FHFA) announced that it would be hiring a financial advisory firm to help it “develop a Roadmap to responsibly end the conservatorships” of Fannie Mae and Freddie Mac. As they carry out their mandate, this firm and those to be retained by Treasury, Fannie and Freddie will benefit greatly from incorporating into their work some simple facts about Fannie and Freddie’s competitive situation and capitalization, set forth below.
The bank competitive issue
There is considerable confusion and misinformation about the roles Fannie, Freddie and commercial banks play in the residential mortgage market, and the competition between them. Much of this has stemmed from proponents of legislative mortgage reform insisting that the most important reform objective is to create a “level playing field” in the market, without ever identifying what the playing field is. In fact, today there is no meaningful area in which Fannie, Freddie and commercial banks compete directly, or even are in the same business. There is, therefore, no common playing field that needs to be, or could be, leveled.
Prior to Fannie and Freddie’s conservatorships in 2008 they did compete with commercial banks as holders of single-family mortgages and mortgage-backed securities (MBS). Banks were the largest mortgage holders. At December 31, 2007, banks held $2.98 trillion in single-family mortgages—$2.01 trillion in whole loans and $973 billion in Fannie, Freddie, Ginnie Mae or private-label MBS—on their balance sheets, giving them a 26.5 percent share of the $11.27 trillion single-family market. On the same date Fannie and Freddie together held $1.45 trillion in single-family mortgages and MBS, for a combined 12.9 percent market share. After the companies were placed in conservatorship, however, Treasury required them to shrink their portfolios, first by 10 percent per year and then by 15 percent per year. As a consequence, as of June 30, 2019 Fannie and Freddie’s balance sheet mortgage holdings totaled only $390 billion (3.6 percent of single-family mortgages outstanding), and going forward their portfolios will be limited to purposes incidental to their credit guaranty business.
With Fannie and Freddie having been directed to exit the one business they did have in common with banks—investing in mortgages for profit—their sole remaining business is guaranteeing the credit on residential mortgages. Banks aren’t in the credit guaranty business at all. Banks originate and service mortgages, which Fannie and Freddie do not do, and have never done. Banks also are deposit-based financial intermediaries, and in that capacity incur substantial liquidity risk (the threat of a “run on the bank”), interest rate risk (many of their assets are funded with short-term deposits and purchased funds), and credit risk on multiple asset types (with historical loss rates far higher than those of single-family mortgages). Fannie and Freddie are not intermediaries, and don’t have those risks. Today, the businesses done by Fannie, Freddie and banks have virtually no overlap.
What, then, is the competitive issue? It’s that deposit-based banks compete with capital markets-based investors to own fixed-rate mortgages. Fannie and Freddie are indirectly involved in this competition, because capital markets investors will not invest in mortgages without a credit guaranty from a reliable, high-quality third party. Since the collapse of the private-label securities market in 2007, Fannie and Freddie have been the only trusted sources of conventional (non-government guaranteed) single-family mortgage guarantees. In a very real sense, these two companies form the gateway through which all international capital markets investors channel funds into the U.S. conventional fixed-rate mortgage market, and the guaranty fees the companies charge are the tolls for using this gateway.
At the end of 2007 Fannie and Freddie charged an average of 21 basis points to guarantee a single-family mortgage. In the second quarter of this year they were charging 56 basis points, 35 basis points more, to provide the same credit guaranty. Ten basis points were added by Congress in 2011 with the Temporary Payroll Tax Cut Continuation Act (TCCA), with the proceeds going to Treasury, and another ten by acting FHFA Director Ed DeMarco in 2013 to “reduce [Fannie and Freddie’s] market share,” and “encourage more private sector participation.” Together, these twenty basis points raised the cost of providing a secondary market credit guaranty without adding to its value, and significantly changed the relative economics of deposit-based versus capital markets-based mortgage investing.
One often reads that banks have been losing market share in mortgages. That’s true for origination and servicing, where over the past decade banks have been supplanted by non-bank lenders such as Quicken Loans and LoanDepot as the main sources of these activities. But this has nothing to do with Fannie and Freddie; it instead is a consequence of the actions of Congress and bank regulators, as well as banks’ own business preferences. And the share change one doesn’t read about is the huge jump over the same period in bank mortgage investment. Banks’ $2.98 trillion—and 26.5 percent market share—in single-family mortgage holdings at December 31, 2007 have soared to $4.01 trillion ($2.26 trillion in whole loans and $1.75 trillion in MBS) of the $10.97 trillion in single-family mortgages outstanding at June 30 of this year, for a 36.6 percent market share. Since just before the crisis, banks have added over a trillion dollars of single-family mortgages to their balance sheets, and increased their share of that market by more than ten percentage points.
It’s not surprising that a jump in bank holdings of single-family mortgages and MBS would follow sharply higher Fannie and Freddie guaranty fees. Banks benefit when guaranty fees are too high relative to a loan’s credit risk, in at least three ways. First, non-bank lenders more often will get a better price for a loan by selling it to a correspondent bank as a whole loan rather than securitizing it with Fannie or Freddie as an MBS. Second, since lenders base their mortgage rates on the cost of selling into the secondary market, higher guaranty fees increase the yields, and the profits, on the unsecuritized loans banks retain in portfolio. Third, artificially high guaranty fees force Fannie and Freddie to overprice their guarantees on lower-risk loans and underprice them on higher-risk loans. Banks can take advantage of this to reduce their credit losses by swapping their higher-risk loans for MBS—which they either can sell or keep in portfolio (and benefit from the lower Basel III capital requirement accorded Fannie and Freddie MBS)—while retaining their lower-risk mortgages as whole loans in portfolio (where they incur no Basel III capital penalty for funding them short).
This is why banks and their supporters have been so insistent that Fannie and Freddie hold “bank-like” capital, in spite of the fact that they are nothing like banks. If the companies’ guaranty fees can be kept high or pushed higher through overcapitalization, the capital markets channel they facilitate will be less efficient, mortgage rates will be higher, and the amount of mortgages banks hold in portfolio, the spreads at which they hold them, and their ability to reduce their mortgage credit losses all will benefit. For banks these may be sensible competitive objectives, but they should have no bearing on how FHFA and its financial advisors assess the amount of capital Fannie and Freddie should be required to hold given the risks of their business. Getting the companies’ capital right is essential to their successful release from conservatorship. And fortunately, it’s not that difficult to do.
Fannie and Freddie capital
In announcing the September 30 letter agreement allowing Fannie and Freddie to retain more capital, FHFA Director Calabria said, “The Enterprises are leveraged nearly 1000-to-one, ensuring they would fail during an economic downturn—exposing taxpayers once again.” This statement implies that it is normal for Fannie and Freddie to lose money during a recession. It is not. Prior to the 2008 financial crisis, neither Fannie nor Freddie’s credit losses ever exceeded their net guaranty fees (guaranty fees less administrative expenses) in their roughly forty years of existence, even during recessions. The highest Fannie’s credit losses ever got was 11 basis points as a percent of total loans in portfolio and mortgage-backed securities outstanding (in 1988), and for the fifteen years I was CFO (1990-2004) its average credit loss rate was less than 4 basis points per year. Fannie and Freddie’s average net guaranty fee today, excluding the TCCA fee paid to Treasury, is 33 basis points. In a normal recession, neither company will come close to losing money; each will remain highly profitable.
The outsized losses experienced between 2008 and 2012 by all mortgage holders, not just Fannie and Freddie, were the result of policy and market failures that have since been remedied. In the late 1990s, the Federal Reserve under Alan Greenspan declined to regulate risky lending practices in the newly-emerging subprime market, favoring market regulation instead. Neither the Fed nor Treasury changed their regulatory stances when many of these practices began spreading to the prime mortgage market in 2003, nor when private-label securities (PLS) became the dominant means of secondary market financing for all single-family mortgages in 2004. In the absence of any prudential regulation, near-unlimited access to mortgages for unqualified borrowers through PLS issuance fueled an unsustainable boom in home sales, construction and prices that continued until the fall of 2007, when the PLS market finally collapsed. With PLS financing suddenly gone, and other lenders pulling back in an attempt to protect themselves, housing sales and starts plummeted, and home prices fell by 25 percent peak-to-trough before they could stabilize.
We learned from our mistakes. Post-crisis, the Fed and Treasury (and even Greenspan) admitted their deregulatory posture during the previous decade was an error. Congress in 2010 passed the Dodd-Frank Act, requiring lenders to apply an “ability to repay” rule to mortgage borrowers and, through its qualified mortgage standard, effectively prohibiting the riskiest mortgage products and loan features that proliferated during the PLS bubble. And investors simply abandoned the PLS market due to its inherent conflicts of interest, and have not come back. These reforms and changes make a recurrence of the mortgage market excesses of 2003-2007 extremely unlikely, and without those excesses anything similar to the subsequent 25 percent collapse in home prices that occurred from 2007 through 2011 is equally improbable.
The requirement that Fannie and Freddie hold sufficient capital to withstand another 25 percent nationwide decline in home prices is thus much more a protection against future policy mistakes than against the inherent riskiness of single-family mortgages in a severe downturn. This is an important point to remember when assessing how much additional conservatism needs to be built into the standard. And as to the amount of capital Fannie or Freddie need to survive a 25 percent home price decline, there is no mystery at all about that. We have data from the previous episode to tell us.
I am most familiar with Fannie’s experience. Fannie’s credit losses exceeded its net guaranty fees for five years, from 2008 through 2012 (in 2013 they did not). During that period the company’s single-family credit losses totaled $76.2 billion. It was able to cover a little over a third of those losses with roughly $27 billion in net guaranty fee income ($15 billion from loans on its books at the end of 2007, which paid off at a 20 percent annual rate, and $12 billion from business added through the end of 2012), leaving $49.2 billion (or just under 2.0 percent of Fannie’s single-family loan balance at December 31, 2007) to be covered with capital, assuming no additional cushion. But these total losses, and capital amount, significantly overstate what would happen in the future. Data published by Fannie show that between 40 and 60 percent of its 2008-2012 credit losses came from two loan types—interest-only (I/O) adjustable rate mortgages and no- or low-documentation (“Alt A”) loans—that now are prohibited by Dodd-Frank. Taking the middle of this range, in a repeat of the crisis without the I/O and Alt A loans Fannie’s five-year single-family credit losses would be $38 billion. Fannie would be able to cover more than all of these losses with guaranty fee income, because its current 34 basis-point net guaranty fee rate would generate $45 billion ($26 billion from the existing book and $19 billion from new business) over the five-year stress period. Counting only income from the existing book (i.e., no going concern income), Fannie would need $12 billion, or about 50 basis points of its initial loan balance, to survive the stress—again with no additional cushion.
We can use these “real world” data to assess the results of two stress tests recently run against Fannie’s business: this year’s Dodd-Frank test (which also has a 25 percent home price decline, but over only a 9-quarter period), applied to the company’s December 31, 2018 book, and a stress test run by FHFA based on the parameters of its 2018 capital proposal, applied to Fannie’s September 30, 2017 book (the two books have similar risk profiles). The Dodd-Frank stress test projected Fannie’s credit losses at $7.2 billion, and its net revenues at $17.4 billion. The FHFA stress test produced “net credit risk” (FHFA’s version of credit losses) for Fannie of $70.5 billion—nearly ten times the Dodd-Frank stress loss amount—and included no projection of net revenues, because FHFA did not count revenues as an offset to credit losses.
These are strikingly different results from measuring what purports to be the same thing—credit losses and revenues stemming from a 25 percent drop in home prices—and neither are close to what actually happened during and after the financial crisis. The Dodd-Frank stress test, which is binding on banks but not on Fannie (or Freddie), produces credit losses for Fannie that are far too low relative to its adjusted 2008-2012 experience, and includes revenues as an offset to them (as a stress test should). In contrast, the stress test based on FHFA’s 2018 capital proposal, which is only for Fannie and Freddie and will be binding on them, produces credit losses for Fannie that are far too high relative to 2008-2012, and by giving no credit to revenues substantially inflates its required capital. Returning to an earlier theme, this is nowhere near a “level playing field” for the stress test that is binding on banks and the one that will be binding on Fannie and Freddie, in an area—measuring the systemic risk to our financial markets—where the playing field should be level. FHFA is only responsible for the mechanics of its own stress test, and here it must ensure that its results properly align with readily available benchmarks from historical experience.
Fannie and the FHA
One additional, indirect, way to assess the reasonableness of the capitalization approach proposed by FHFA for Fannie and Freddie is to impose it upon the only other entity limited to the single-family credit guaranty business, the FHA.
FHFA’s 2018 capital proposal features a “single-family credit risk pricing grid for new originations.” Applying this grid to the average loan-to-value (LTV) ratios, credit scores, and refinance percentages of the business done in 2018 by Fannie (77.0 percent, 743 and 35.0 percent, respectively) and the FHA (91.9 percent, 670 and 23.5 percent) can give us a rough capital requirement for each. The required credit loss capital from these grids for Fannie’s 2018 new business is 2.2 percent, and for the FHA’s it’s 7.8 percent. But we’re not finished. Fannie has private mortgage insurance (PMI) on loans with LTVs over 80 percent, and company data show that over the past 20 years PMI has reduced Fannie’s loss severity on high LTV loans by an average of 42 percent. With relatively few exceptions (refinances of conventional loans into FHA), the FHA has only high LTV loans, and no PMI. Adjusting for this, the 2018 proposed FHFA grids that would require 2.2 percent in credit loss capital for Fannie’s 2018 new business would require more than 12.0 percent if applied to the FHA’s. The FHA’s capital at December 31, 2018 was 2.76 percent.
Since the end of 2007, loans financed by the FHA have grown by 350 percent, while loans financed by Fannie and Freddie have grown by 10 percent. Capital and pricing differences are almost certainly the reasons for this enormous growth disparity.
* * *
By far the most critical element of the “Roadmap to responsibly end the conservatorships” is Fannie and Freddie’s capital standard. If FHFA, Treasury and their financial advisors can agree on a standard that both protects the taxpayer from loss and enables the companies to carry out their credit guaranty function on a scale and at a cost that satisfies homebuyers’ needs, Fannie and Freddie will have successful businesses, and investors willingly will supply the new equity required to return them to private ownership.
The simple facts noted above will help to find the right answer here. Those responsible for determining Fannie and Freddie’s required capital can’t allow themselves to be fooled by claims that it must be “bank-like,” or anywhere close to it; the companies are in no way like banks, so there is no economic reason for their capital requirements to be similar. Also, the principals and their advisors need to be aware that the 25 percent home price decline Fannie and Freddie are being asked to protect against is extremely unlikely in the absence of another major and protracted error in regulatory policy, so the stress test replicating this scenario does not need to be made more severe by further conservatism built into it. The specification of the FHFA stress test must be straightforward and understandable. It should produce credit losses and revenues that closely align with Fannie and Freddie’s 2008-2012 experience, adjusted for changes in their books of business, and clearly identify and explain any cushions or additional elements of conservatism. Finally, required “going concern” capital should reflect revenues from new business expected to be done during the stress period, as well as the existence of a catastrophic risk backstop from Treasury, for which the companies will be paying a commitment fee.
The right answer for Fannie and Freddie’s capital will lead to their smooth and successful exit from conservatorship. And the wrong answer will be obvious, because it will be visibly inconsistent with historical experience, and cause Fannie and Freddie’s credit pricing to be notably misaligned with credit pricing elsewhere in the market. The pathway to the right answer for FHFA, Treasury and their financial advisors is clearly marked; they only have to follow the facts to stay on it.