A persistent theme of this blog has been the crucial role that a revised capital standard for Fannie and Freddie will play in determining how effectively and efficiently the companies will be able to carry out their traditional function of profitably providing large volumes of affordable mortgage financing to a wide range of borrower types in a post-conservatorship world.

Since the conservatorships, opponents and critics of the companies have insisted that any future version of them as shareholder-owned companies include the requirement that they hold “bank-like” amounts of capital. This demand has been couched in terms of safety and soundness (“more capital is better”), but in fact the goal of advocates of this approach is to burden Fannie and Freddie with capital requirements that are far out of proportion to the risks of the mortgages they finance, forcing them to raise their guaranty fees to levels that make their secondary market activities overpriced and less competitive, to the benefit of primary market lenders. As I’ve often noted, there is no economic or regulatory reason why Fannie and Freddie’s capital should be equal or anywhere close to that of commercial banks. Fannie and Freddie are not multi-product and multinational lenders; they are mono-line insurance companies, limited to a single asset type—residential mortgages—whose historical credit loss performance has been dramatically better than banks.

Last June, FHFA put a Proposed Rule for Enterprise Capital out for comment. I was among many who submitted one. While I didn’t say this in my comment letter, it was clear to me that in its specification of the risk-based standard mandated by the Housing and Economic Development Act (HERA), FHFA had used a combination of unrealistic structural elements and conservative assumptions to push Fannie and Freddie’s calculated capital requirement substantially higher, to 3.24 percent as of September 30, 2017. My comment addressed the elements and assumptions I thought were unjustified—including not counting guaranty fees as offsets to credit losses—and I suggested ways that FHFA could and should improve the quality and accuracy of its risk-based capital specification to align it much more closely with the risks of the mortgages the companies are guaranteeing and the way their business works. A number of other commenters made similar observations.

FHFA’s revised capital proposal is due to be released shortly, so it was with considerable concern that I read the recent spate of comments by the new FHFA Director, Mark Calabria, repeating the general (and insupportable) assertion that “Fannie and Freddie ought to operate under essentially the same capital rules as other large financial institutions.” These statements raised the possibility that he had chosen to side with the companies’ opponents on the capital issue, but I also thought he might not know the facts about the differences in risk between Fannie and Freddie’s business and that of commercial banks. So last week I wrote a two-page note for the Director and sent it to FHFA General Counsel Alfred Pollard, asking him to submit it both for the record and to Calabria. Here is what the note said:

“A recent interview with Fox Business News quoted you as saying, ‘I think our objective over time is that you have capital levels at Fannie and Freddie that are comparable to other large financial institutions,’ and that 4.5 percent capital was ‘kind of in the neighborhood of where we’re looking at.’

While I agree that Fannie and Freddie’s capital should be comparable to that of other large financial institutions, comparable is not the same as equal. Capital must be related to risk, and if FHFA engineers its risk-based capital standard for Fannie and Freddie’s credit guarantees to produce an average capital percentage equal or even close to banks’ Basel III risk-weighted capital percentage for residential mortgages of 4.5 percent, the companies’ capital standard would be far more stringent than banks’, because: (a) the delinquency and default rates of single-family mortgages owned or guaranteed by Fannie and Freddie have been less than one-third those of residential mortgages held by banks, and (b) bank capital also must cover the interest rate risk of funding mortgages in portfolio, whereas Fannie and Freddie’s capital standards treat credit and interest rate risks separately, and additively.

Historical data on single-family residential mortgage delinquencies and defaults for Fannie and Freddie versus commercial banks are readily available from the companies themselves and from the FDIC and Federal Reserve, and they tell a clear and consistent story.

During the time I was Fannie’s CFO from 1990 through 2004, the company’s rate of single-family serious delinquencies (90 days or more past due) averaged about 55 basis points, while over the same period the average serious delinquency rate for residential mortgages held on the balance sheets of commercial banks was 2.3 percent. Following the financial crisis both delinquency rates spiked, peaking in the first quarter of 2010—Fannie’s at 5.53 percent and banks’ at 11.54 percent. Both have since declined substantially, but banks’ 2.67 percent serious delinquency rate on residential mortgages in the first quarter of this year still was three and a half times Fannie’s 75 basis-point delinquency rate. In addition, Fannie now publishes its serious delinquency rate on the single-family loans it has guaranteed or purchased since the end of 2008, when it tightened its underwriting standards. These loans now make up 93 percent of Fannie’s total book, and their rate of serious delinquency in the first quarter of 2019 was only 33 basis points.

Post-crisis default rates tell a similar story. In the amicus curiae brief I submitted on July 6, 2015 for a court case involving Fannie and Freddie I noted, ‘Solid data now exist, and they show that, from 2008 through 2014, the average loss rate on residential mortgages owned or guaranteed by Fannie and Freddie was 0.45 percent per year. The loss rate for residential mortgages owned by commercial banks during this same period was 1.43 percent per year—more than three times as high.’

Most people are surprised to learn that single-family mortgage delinquency and default rates at commercial banks consistently exceed the comparable rates recorded at Fannie and Freddie, by a factor of three. I believe the primary reason is that banks finance a much greater percentage of home equity loans. But whatever the reason, banks’ much higher delinquency and default rates mean that Fannie and Freddie can provide an equal degree of protection against mortgage credit risk with one-third the capital that banks require.

A second critical consideration for determining what constitutes a level playing field on capital for banks and Fannie and Freddie is the fact that banks’ 4.5 percent Basel III capital requirement for single-family mortgages has to cover not just credit risk but also the risk of financing 30-year fixed-rate and capped adjustable-rate mortgages with consumer deposits and short-term purchased funds. Capital standards for Fannie and Freddie treat these risks separately; currently they have a 45-basis point minimum for guaranteed mortgages (now universally viewed as too little), and a 250-basis point minimum for mortgages financed in portfolio. The companies thus have to hold 205 basis points in interest rate risk-related capital for their on-balance sheet mortgages, even though they can and do match their durations with a combination of long-term debt and derivatives, and constantly ‘rebalance’ to keep those durations matched as interest rates change. Since banks have more interest rate risk in their mortgage portfolios than Fannie and Freddie, the capital required for the interest rate risk they bear should be higher as well.

A true level playing field on capital between banks’ on-balance sheet mortgage holdings and Fannie and Freddie’s single-family credit guaranty business, therefore, is nowhere near 4.5 percent. Banks’ 4.5 percent capital ratio properly applies to the companies’ portfolio holdings, making their credit risk capital no more than 2.5 percent (4.5 percent less at least 2.0 percent for interest rate risk), and the huge differential between banks’ and Fannie and Freddie’s historical credit performance should further reduce the companies’ required capital for their credit guaranty business—to a maximum of 1.5 percent, and arguably less.

For the above reasons, a very good case can be made for FHFA to set 1.5 percent as the minimum capital requirement for Fannie and Freddie’s credit guaranty business, and 3.5 percent as the minimum for their portfolio mortgages. More importantly, it would be a serious mistake for FHFA to employ layered conservative assumptions to push the capital percentage required by its risk-based capital stress test artificially high, in order to make it look more ‘bank-like.’ Doing so is unjustified, and would decouple the capital and pricing for the mortgages the companies guarantee from their true risks, making their services more expensive and less efficient for no good reason, to the disadvantage of the low- and moderate-income homebuyers Fannie and Freddie were chartered to serve.”

Once these facts about relative mortgage risks are known and acknowledged, it becomes obvious that it is incorrect to impose a bank-like capital standard on Fannie and Freddie. On top of that, as Fannie (and I) discovered more than thirty years ago, the correct way to specify a risk-based capital standard is nearly as obvious.

Prior to the early 1980s Fannie had done all of its business as portfolio purchases, funded with debt. Not long after I joined the company it began issuing mortgage-backed securities (MBS) using pools of newly-originated mortgages for which it bore the risk of loss, rather than having recourse to the originating lender. Fannie’s charter had been tailored for an entity that financed all of its mortgages on-balance sheet, so its new MBS, which for Fannie as well as Freddie were considered contingent liabilities and thus accounted for off-balance sheet, had no required regulatory capital. We had to determine how to capitalize, and then price, these credit guarantees ourselves.

The stress-test based approach we adopted was virtually identical to what FHFA should be doing to specify the risk-based capital requirement for Fannie and Freddie mandated by HERA. First, you use historical data to pick a defined degree of credit stress you want to be able to survive. (We initially picked a credit loss rate we deemed to have less than a one percent chance of occurring, later tightening that to a one-half of one percent chance; under HERA, FHFA gets to pick whatever loss scenario it wishes.) Next, you take your existing book of business with its fee rates, and break it into “buckets” of product types and risk combinations. Assuming no replacement business, you then simulate the performance of each of these buckets using different amounts of initial capital until you determine the exact amount of capital that, when combined with the guaranty fee income from the loans you project to remain outstanding during the stress period and allowing for projected administrative expenses, just allows that bucket of loans to survive the loss scenario you’ve chosen. Finally, you add a reasonable, clearly identified cushion of conservatism, and you have your required capital percentages. It’s that simple.

Unlike banks’ ratio-based standards, Fannie and Freddie’s stress-based capital calculations adjust automatically as the riskiness of their business changes, because they’re done at the risk and the product level. If FHFA wants to be even more conservative, it should pick a higher threshold of stress for the companies to meet (and defend that choice), rather than use unrealistic or unjustified elements to force the capital requirements higher. The latter break the link between capital and risk, and also affect Fannie and Freddie’s ability to attract a broad and profitable a range of business. FHFA’s decision in last year’s proposal not to count guaranty fee income in the stress test illustrates this problem. A properly designed risk-based standard permits the companies to trade off initial capital, guaranty fees and return targets to manage their mix of business, while remaining in capital compliance. For certain loan types it may make sense to charge a lower guaranty fee and compensate with higher initial capital (and thus accept a lower expected return); for other business a higher guaranty fee and lower initial capital may be the right choice. If guaranty fees are not counted in the stress test, however, this flexibility is lost. As a general rule, any significant distortions in the specification of the risk-based capital standard will constrain how the companies can do their business, sometimes in unpredictable ways.

There truly is a right and a wrong way to specify a risk-based capital standard, and the difference between the two is evident. When FHFA releases its revised capital proposal, therefore, there will be little doubt as to whether it has made a straightforward attempt to produce an accurate risk-based standard, engineered the exercise to produce a particular predetermined outcome, or tried to land somewhere in the middle.

Potential investors in a recapitalization of Fannie and Freddie should pay close attention to where FHFA comes out on this proposal, because it will be a window into the agency’s likely regulatory posture with respect to the companies post-conservatorship. At the one extreme, if FHFA makes only minor changes to its June 2018 capital proposal, this would strongly indicate that it intends to favor ideology and politics over economics in its future regulation, which should be a bright red light for any investor with a functioning memory. At the other extreme, a clean risk-based capital standard—which would result in the least amount of unnecessary capital, the lowest average guaranty fees, and the most flexibility to use cross-subsidization to broaden the companies’ product mix and support affordable housing—would be a green light for investment.

A middle-ground outcome on the capital standard, however, seems to be the most likely, and it will be the hardest for investors to evaluate. In this case, they will be well advised to closely examine the details of the proposed standard, assess the potential for any aspects of it to be used to their disadvantage, and insist that anything objectionable be remedied to their satisfaction before they agree to put new capital into the companies. Investors never will have as much influence in determining how the next version of Fannie and Freddie will operate and be regulated as they will prior to a first round of equity raising.