The Federal Housing Finance Agency (FHFA) released its notice of proposed rulemaking on “Enterprise [Fannie Mae and Freddie Mac] Capital Requirements” on June 12. After the comment period closed on November 16, there were 77 entries on FHFA’s comment log. Half (38) were from individuals who wrote short comments ranging from one sentence to a few paragraphs expressing their personal points of view, and there were eight administrative-type entries (correspondence to and information or meeting summaries from FHFA).
The remaining 31 entries broke out this way: One each from Fannie and Freddie; 5 from private individuals—Myself, the team of Ed Golding, Laurie Goodman and Jun Zhu (who work at the Urban Institute but submitted their piece as individuals), and Andrew Davidson (who submitted three separate comments); 1 from an affordable housing advocate—The Center for Responsible Lending; 3 from private mortgage insurance entities—Genworth, Arch Capital, and U.S. Mortgage Insurers; 4 from Wall Street interests—Moelis & Company, Pershing Square Capital Management, Katz Capital, and The Structured Finance Industry Group; 5 from community lending groups— The Independent Community Bankers of America, The Community Home Lenders Association, The Credit Union National Association, The National Association of Federally-Insured Credit Unions, and Heartland Credit Union Association; 4 from large trade associations—The American Bankers Association, The Mortgage Bankers Association, The National Association of Realtors, and The National Association of Homebuilders; three from Washington think tanks—The American Enterprise Institute, The Housing Policy Council, and The American Action Forum), and four from multifamily interests—The DUS Advisory Council, Walker and Dunlop, NorthMarq Capital, and the National Multifamily Housing Council and National Apartment Association, commenting jointly.
For those whose interest is in multifamily lending I must offer an apology: from the start of the capital exercise I have concentrated almost exclusively on the single-family side, since that is where my expertise (and most of the controversy) lies. But multifamily lenders serve a critical segment of the housing community, and getting Fannie and Freddie’s capital right for these borrowers is just as important as it is in single-family. As soon as I can, I intend to dig into the multifamily comments, try to get up to speed on the issues there, and see if there is anything I can add to the dialogue that might be helpful.
Commenters on the single-family standard, as a rough generalization, fell into two groups: those whose primary concern was how the proposed capital rule could (or needed to) be improved in order to strike the desired balance between taxpayer protection and the provision of an ample flow of mortgage credit to as wide a range of potential borrowers as possible, and those whose comments were colored by a different objective—whether it be ideological (reducing the role of “government” in housing), competitive (engineering the capital rule so as to move business away from Fannie and Freddie to a preferred financing source), or something else.
The focus of this post is on comments intended to improve the FHFA standard. (If those who wish to use capital to constrain the companies end up carrying the day, the issues the former group is trying to get right won’t matter as much.) From this perspective, I found the most useful and constructive comments to have come from Fannie Mae, the team of Golding, Goodman and Zhu from the Urban Institute, The Center for Responsible Lending, and the three mortgage insurance entities (who had valuable insights into how the capital dynamics of insurers differ from those of banks). And even including the comments from the “we’re going to wind this thing [Fannie and Freddie] down, we’re going to kill it, we’re going to drive a stake through its heart, and we’re going to salt the earth so it can never grow back” crowd, the submissions as a whole revealed a number of areas of legitimate concern:
- The lack of transparency of the FHFA proposal, which makes it difficult to understand it in sufficient detail to offer constructive suggestions for how to improve it (or to prevent unintended consequences);
- Near-universal agreement that using current loan-to-value ratios in the standard creates severe problems of procyclicality, which need to be addressed;
- The adverse effect on mortgage rates of layering conservative assumptions one on top of the other;
- The need for FHFA to acknowledge the role of guaranty fees in absorbing credit losses;
- A significant degree of overcapitalization of higher-risk mortgages relative to their historical performance;
- Confusion as to how the minimum capital ratio, loss reserves, the going concern capital buffer, any systemic risk buffer (proposed by some) and excess capital relate to one another in the standard, and
- Reservations about the proposed equity capital relief for securitized credit risk transfers.
Below are brief elaborations of each of these points:
Lack of transparency. U.S. Mortgage Insurers spoke for many commenters when it said, “The NPR does not provide sufficient factual detail to permit interested parties to ‘comment meaningfully’ on the proposal.” USMI said that, “FHFA has chosen to abandon the existing capital rule in its entirety…using instead as a foundation for this NPR the CCF [FHFA’s Capital Conservation Framework], which has not been publicly released.” USMI added, “The remedy is to treat this proposal as an ANPR [Advance Notice of Proposed Rulemaking], solicit further comments on the appropriate capital regulations for the Enterprises, and, in particular, solicit comments on whether an insurance company capital model is more appropriate in light of the fact that these entities are not banks.”
Pershing Square Capital and The Center Responsible Lending (CRL) believe FHFA’s current standard may be too complex to fix, and for that reason are prepared to support a simple 2.5 percent leverage ratio for all of Fannie and Freddie’s business.
Use of current loan-to-value ratios in stress test. This feature was almost universally cited as problematic, because of its strong procyclical effect. The Urban Institute’s Golding, Goodman and Zhu noted with understatement that, “It will be difficult for a GSE to manage to a requirement that can double in two years.” Several commenters (including Fannie, the Urban Institute team, Pershing Square, the ICBA and the ABA) supported moving back to the original LTVs used in the 1992 standard. Yet nearly an equal number (including Genworth, Arch Capital, the Housing Policy Council and the MBA) proposed sticking with mark-to-market LTVs but having FHFA adjust them for “fundamental housing values,” which would somewhat increase required capital in strong housing markets and somewhat reduce it in weak ones.
I was surprised that trying to offset current LTV volatility by attempting to estimate when homes are overvalued or undervalued had as much support as it did. Most obviously, such estimates of fundamental value always will be subjective. Second, this approach still would have the companies holding too little capital near market peaks and too much capital at troughs. And third, it leaves unsolved the problem of how to set guaranty fees when required capital is not known at the time of pricing. But given the number of comments in favor of this alternative, FHFA will need to examine it seriously.
Layered conservatism. FHFA says it “considers it prudent to have risk-based capital requirements that include components for credit risk, operational risk, market risk, and a risk-invariant going-concern buffer; that require full life-of-loan capital for each loan acquisition; that are calculated to cover losses in a severe stress event comparable to the recent financial crisis, but with house price recoveries that are somewhat more conservative than experienced following that crisis; and that do not count future Enterprise revenue toward capital.” Numerous commenters, including the community lending groups, wanted FHFA to quantify the effect this layered conservatism had on guaranty fees and mortgage rates. As the NAR noted, again with understatement, “the significant caution taken in creating a conservative rule mutes or may hinder the Enterprises’ ability to support their public mission.”
The Urban Institute group, CRL, Arch Capital, USMI, the NAR and the NAHB all called attention to another significant, but hidden, form of conservatism in the standard: since the 2008 financial crisis, the adoption of the qualified mortgage and ability to repay standards, improved appraisals, changes in underwriting and stronger capitalization of mortgage insurance companies all have reduced both the likelihood and severity of a similar crisis in the future, but none of these favorable changes are reflected in the stress loss percentages in FHFA’s look-up tables or risk multipliers.
In addition, embedded in Fannie’s comment is a subtle but important fact: FHFA has built all of these levels of conservatism into a stress standard that uses current value LTVs which, given the recent run-up in home prices, are near record lows. Were FHFA to switch to an original LTV-based standard—as it should—Fannie said its required capital under the FHFA standard would rise “by nearly $25 billion.” This would make Fannie’s capital requirement for its September 30, 2017 book not the 343 basis points estimated by FHFA, but well over four percent—much higher than almost all of the commenters on the standard are expecting it to be.
Exclusion of guaranty fees in stress test. No commenter defended FHFA’s decision not to count guaranty fees in determining required stress capital (although a few suggested that some type of discount or “haircut” might be appropriate to account for the uncertainty in estimating the magnitude of those fees). USMI had perhaps the most cogent comment on this topic: “It would be highly ironic (if not irrational) to require higher fees based on capital needs, but not count those fees when determining capital requirements.”
Excessive capital requirements for higher-risk mortgages. Both the Urban Institute team and CRL did extensive analyses of the FHFA proposal by loan attribute. The Urban Institute team wrote, “Our conclusion: the lower FICO higher LTV mortgages require more capital than is necessary relative to their less risky brethren.” CRL was more specific, saying, “The model would disproportionately place the cost of this systemic failure capital [capital required to pass FHFA’s stress test, plus cushions] on lower wealth borrowers, by a factor of as much as ten to one,” and it made the added point that “not counting revenues adds a further bias to high LTV low FICO QM loans” because these loans prepay more slowly during stress periods, and thus can cover a higher percentage of their losses with guaranty fees. CRL concluded its analysis by saying, “The proposal has greatly excessive capital demands. This will harm the housing market and overall economy, and it will place the heaviest unnecessary burden on working families of modest means.”
Interaction of capital components and loss reserves. Several commenters are concerned that FHFA has not thought through how the different components of its proposed capital standard relate to one other, or what each one’s purpose is, resulting in what USMI calls “a ‘mash-up’ of different capital concepts.” Genworth questions “the need for both a going concern buffer and a minimum leverage requirement,” while the criticism of USMI is more general: “It is not clear that the proposal takes into account the full extent of, and interactions between, loss reserves, the minimum leverage standard, the incentives provided by prompt corrective action for the Enterprises to hold excess capital and the market pressure for the Enterprises to hold even more capital.”
Fannie raised an important question about how the FHFA capital rule would be affected by implementation of the Current Expected Credit Loss (CECL) accounting method scheduled to take effect on January 1, 2020. Under CECL, a company will set its loss reserves based on the “expected lifetime losses” of its loans, rather than its “incurred losses,” as currently. The FHFA capital proposal states that, “The credit risk capital requirements in the proposed rule are based on unexpected losses (stress losses minus expected losses) over the lifetime of mortgage assets.” I had interpreted this to mean that as Fannie and Freddie’s loss reserves rose under CECL during a stress period (along with their expected losses), their required equity capital would fall by a comparable amount, since the total stress standard will not have changed. Fannie’s comment suggests it does not believe that will be the case. FHFA must clarify this relationship. For an insurer, loss reserves and capital serve the same purpose: to absorb credit losses. If an accounting change (CECL) can push up the companies’ total required capital, that is a flaw in the FHFA proposal that needs to be fixed.
Too favorable treatment of credit risk transfer securities. Finally, numerous commenters questioned the cost-effectiveness of securitized CRTs, with most (including Fannie) focusing on their likely non-availability during times of stress, when they would be most needed. The Urban Institute team, CRL, Arch Capital and Genworth all said that the cost of Fannie and Freddie’s CRT securities should be deducted from any capital relief given to the companies when CRTs are used.
* * *
As should be evident from the above, enough questions about and serious objections to FHFA’s proposal were raised by those who commented on it that the agency almost certainly will need to rework and reissue it. This has obvious implications for the timing of any administrative mortgage reform that preserves the role of Fannie and Freddie in the secondary market. The companies cannot be released from conservatorship until they are recapitalized, and they cannot be recapitalized until they have binding regulatory capital requirements. Fannie and Freddie’s required capital will determine not only the dollar amount of new equity they would have to raise but also how they would have to price their business, which in turn will affect the volume and mix of business they could do (and, through these volume and loan quality impacts, their investment values as going concerns). A workable final FHFA capital regulation is at the same time an essential step on the path to administrative reform and a challenge that must be met to get there.
The Mortgage Bankers Association—which is solidly on record in favor of legislative reform that creates competitors to Fannie and Freddie rather than administrative reform that recapitalizes them—clearly sees the potential for protracted discussions over the risk-based capital standard to delay administrative action indefinitely. Its comment letter is not subtle about this. In addition to advising FHFA that a lengthy, “multi-round set of notice-and-comment periods would give FHFA the opportunity to fine-tune the proposed rule, benefiting from outside points of view, improving it in stages,” MBA told FHFA, “In fact, as can be seen from the structure of our comment letter, we find ourselves with more questions than answers with respect to many aspects of the proposed rule. We would ask that FHFA seriously consider the full set of questions posed when developing additional data and information to release prior to further rounds of notice and comment.” I counted all of the MBA’s questions and there were 86 of them—enough, should FHFA elect to address them all in detail, to keep the agency tied up for months.
FHFA, instead, should “get on with it.” By now it undoubtedly will have reviewed and categorized all the comments it has received. After analyzing them thoroughly, it should take the initiative to meet with the individuals or groups whom it believes have the best insights into or ideas for constructive changes to the standard, with the objective of creating a new less complex, more consistent, cohesive, transparent and defensible proposal that it can issue as a new notice of proposed rulemaking, with advance buy-in from key stakeholders. Supporters of administrative reform also need to engage on this. They have the most reason to get Fannie and Freddie’s capital right, and to get it right as quickly as possible.