FHFA’s Final GSE Capital Rule: Little Credibility and a Short Shelf Life
The Federal Housing Finance Agency (FHFA) received overwhelmingly negative feedback when its GSE capital rule was proposed earlier this year. The main criticisms were that the required level of capital was too high (an 80 percent increase over its predecessor proposal), that it was poorly constructed (with many arbitrary floors and buffers, and an incorrectly designed leverage ratio), and was bizarrely biased against credit risk transfer (CRT). Nevertheless, FHFA Director Mark Calabria just approved a final rule that effectively doubles down on the much-criticized proposal.
I have four reactions to it that are important to understand:
1. The total capital requirement of $283 billion is even higher than in the proposal.
Despite most comments that were submitted on the proposed rule finding the required capital too high, the final rule actually increases risk-based required capital by another 8 percent, or $20 billion (all figures herein as of June 30, 2020). The comments from industry sources lacked some credibility, of course, in that all beneficiaries of the mortgage system, in their own self-interest, prefer low capital requirements in order to keep mortgage credit as inexpensive as possible. However, more technical commenters (for instance, Andrew Davidson, who submitted a 17-page comment letter) also found the required capital too high. I fully concur — it is absolutely well above what is needed to support a policy of having the same capital required by other systemically important financial institutions for the same risk (which I dub “SIFI-consistency”). It is true that determining what level of capital is SIFI-consistent is difficult, given the dramatically different business models of the GSEs compared to banks. (For example, the risk to a SIFI bank from holding a mortgage loan is far higher than if a GSE held the same loan given that, unlike the bank, the GSE issues both a pass-through mortgage-backed security on it, almost wholly eliminating liquidity and interest rate risk, and a CRT, which eliminates most of its credit risk. Therefore, much less capital is required of a GSE than a SIFI bank when holding the same mortgage loan, while still retaining SIFI-consistency.) But that’s no excuse for the capital rule being so far off, and the final rule increasing it further is hard to fathom. In aggregate, the right number for required SIFI-consistent capital is, I estimate, in the $175 billion range (or about $100 billion less) on today’s combined $6.3 trillion GSE balance sheet.
2. The final rule is still significantly convoluted and distorted.
As mentioned above, the proposed rule was criticized for its heavy use of non-risk-based features: arbitrary floors and buffers, plus the incorrectly designed leverage ratio requirement (which was intended by the FHFA to be, and was at the time, cyclically binding, rather than correctly functioning only as a backstop). Almost none of this is changed in the final rule, except that the leverage ratio is no longer binding; however, this result is not because the leverage ratio requirement was reduced (as many commenters suggested), but instead because the risk-based requirement was increased. (It is unclear whether and how often the leverage ratio might still be cyclically binding, as it is only just under 7 percent lower than the risk-based calculation.) Therefore, the final rule will still, if not quite as much as the proposed one, drive distorted transaction-level decision-making, and will result in the GSEs taking actions that are economically bad but falsely look good according to the capital rule. The history of bank capital rules, going back to before Basel I, is that such distorted decision-making has led to significant problems, and the final GSE rule appears to be heading in the same direction.
3. The inexplicable and damaging antipathy toward CRT continues with no consequential change.
The credit risk transfer program, supported overwhelmingly by the industry, elected officials from both parties, and all but a few extreme GSE-skeptic policy specialists, would have, under the proposed capital rule, dramatically declined if not outright ended. This is because it would become falsely uneconomic for the GSEs to do most, if not all, such transactions. (Note that Fannie Mae has already stopped its CRT program, pending finalization of the capital rule.) The FHFA says the final rule is more generous in its allowance for CRT to offset credit risk, but the numbers indicate otherwise: CRT goes from providing a capital reduction of 18.0 percent under the proposed rule to slightly less at 17.1 percent under the final one. Among other consequences, the GSEs are now strongly incented to reconcentrate extraordinarily large amounts of mortgage credit risk back onto their balance sheets, which would likely be the single most backward and damaging systemic risk action since the 2008 financial crisis.
4. The capital rule works directly against the FHFA’s desire for the GSEs to be recapitalized.
The final capital rule will kill, at least for several years, the ability of the two companies to issue any sizeable amount of new common equity. This is unchanged from the proposed rule because their financial results will still show, based upon the capital they would need to raise to meet the regulation, a return-on-equity far below levels demanded by investors. (It will take at least several years for new loans with higher guarantee fees to accumulate before the average return approaches needed levels.) So the FHFA, which says pursuing the end of conservatorship is a top priority, is in fact going in two opposite and conflicting policy directions at the same time, which is not a recipe for either to be successful.
Unfortunately, this final capital rule, like the proposed one before it, seems to me to reflect far too much antipathy to the GSEs and the politics of anti-GSE advocates who have long wanted to dramatically shrink (if not eliminate) the role the two companies play in mortgage markets. While the FHFA is, not surprisingly, claiming that the capital rule strictly reflects professional policy considerations, I believe that this is not really true and that anti-GSE ideology is too much in the driver’s seat. The final capital rule reflects no attempt to create something with broad credibility or acceptability to the industry, stakeholders more generally, or the analytical experts involved in housing finance. In fact, the feedback received during the legally-required comment period indicating just how minimal was the credibility of the proposed rule may, if anything, worsen with the final rule’s publication.
In the end, it may not matter, but could be very problematic in the meantime. I predict that when a President Biden-appointed FHFA Director (or acting Director) takes office, which is likely to be mid-2021 (although it could be as far away as early 2024, depending on the outcome of a case currently before the Supreme Court), the capital rule will quickly be replaced and revised with what is hopefully a more mainstream, properly-designed one that earns broad credibility. As the GSEs will very likely not be out of conservatorship and fully capitalized before the date of that replacement, the new rule might then never actually become binding on the two companies. One could therefore argue that the GSEs should, in the meantime, simply ignore it. However, the GSE management teams can’t realistically be expected to do that with Director Calabria still in office at the FHFA, and so will need to generally adhere to the rule, which will in turn create confusion, wasted effort, and bad decisions. The rule will also generate immediate pressure for a major increase in guarantee fees, which are currently based on a much lower capital requirement. Such an increase, if pursued by the FHFA, would generate incredible noise and friction in the industry and in Washington — it is not obvious how that would all play out, except to increase the desire of the Biden administration to replace Director Calabria as soon as it can.