House under construction

Housing Perspectives

Research, trends, and perspective from the Harvard Joint Center for Housing Studies

Newly-Proposed Changes to the GSE Capital Rule Will Eliminate Harmful Distortions

On September 15, the FHFA issued a formal proposal to amend the regulatory minimum capital requirement that applies to the two government-sponsored enterprises (GSEs) of Freddie Mac and Fannie Mae, known as the “enterprise capital rule” (ECR). My analysis is that these changes are not designed to significantly reduce the total capital required by the GSEs at this time, but rather to eliminate harmful distortions in the current rule that create incentives for the two companies to seek out and hold high-risk assets, which is counter to good public policy. I am therefore very supportive of the proposal.

The ECR has a history that reflects the well-known politicization of housing finance policy in this country. Post-conservatorship, the first ECR proposal was published in 2018, when the FHFA director was Mel Watt, a former Democratic congressman appointed by President Obama. The Watt ECR proposal called for a total capital requirement (based upon September 2017 GSE balance sheets, but revised to reflect year-end 2017 tax rate changes) of $165 billion. It was generally well-regarded, seemingly large enough to cover stressed-market losses with a significant margin, and was constructed such that the GSEs had the proper economic incentives to manage and price their risks well. The most important concern expressed in public comments was that it was too pro-cyclical, meaning the GSEs would be required, in good times, to have a level of capital that would prove insufficient to weather the bad times that could follow. Director Watt’s term expired in January 2019 without the FHFA making revisions and issuing a final rule.

The next director of the FHFA, appointed by President Trump, was Mark Calabria, who had a history of writings and public comments that were strongly biased towards shrinking, if not eliminating, the role of the GSEs in mortgage markets. Under his guidance, the FHFA issued a revised ECR proposal in May 2020, received public comments, and issued a final rule six months later. The Calabria ECR was and still is controversial.

Previously, I wrote articles on both the Calabria ECR proposal and the final rule. Both were critical, primarily with respect to three problems:

  1. The total amount of capital required, based upon the higher of a complex risk-based calculation and a simple leverage requirement (the latter being a set percentage of assets, regardless of the riskiness of those assets), was far too high at $283 billion, in particular being inconsistent with the results of the annual stress tests. This was due to the inclusion of several large, judgmentally-determined buffers (i.e. they were calculated not as a result of quantitative analysis but by the somewhat arbitrary judgment of FHFA officials). As I previously stated, based upon GSE 2019 balance sheets, I estimate the right number was in the $150-175 billion range.
  2. The specifics of the risk-based capital requirement, especially with respect to credit risk transfer (CRT), were significantly distorted so they were at odds with the GSEs making good, economically-driven risk-reward decisions. This was because impactful judgmentally-determined overlays were placed on top of the underlying cost-benefit calculations, which had the result of rendering CRT transactions falsely uneconomic. That is, they were economic to undertake based on the quantitative analysis but not after applying those judgmentally-determined overlays. I note that this result, perhaps not surprisingly, is in line with Director Calabria’s expressed view that “CRT doesn’t work,” a view not broadly shared in the industry or policy community. Thus, the risk-based capital system gives the GSEs strong incentives to hold onto their risks, which makes the US taxpayer more exposed to the GSEs and also increases the systemic concentration of risk in the two companies. This is not good for the financial system broadly nor the housing finance system specifically.
  3. The leverage capital requirement (again, just a set percentage of assets, regardless of the riskiness of those assets) was so high it would almost always be higher than the risk-based calculation (referred to as “binding”), making the lower risk-based capital requirements nearly irrelevant. This is in direct contravention to well-developed global financial institution regulatory practice, where the purpose of a leverage requirement is to be a credible backstop to a risk-based capital system, should the latter fail somehow to capture significant risks. It is well known that, when a leverage capital requirement is binding in this fashion, it strongly incents seeking out and holding high-risk assets, which again would leave the US taxpayer more exposed and the financial system more accident-prone.

My criticisms were not unique, and were shared by many of the commenters on the Calabria ECR proposal. But the FHFA essentially ignored the feedback and only made superficial revisions to produce the final Calabria ECR. It was therefore obvious the rule would likely have a short shelf life once the Calabria era at the FHFA ended. Indeed, as President Biden appointed Sandra Thompson as Acting Director in late June, the shelf life has proven to be quite short—less than three months!

The Thompson revisions, which can be found on page 6 of the FHFA proposal, mainly consist of three changes that are primarily technical in their wording, so it is not obvious to the lay reader what the impacts will be. In fact, there are three important policy implications that will flow directly from those changes:

  1. The amount of capital required will be lower—not by much right now, but avoiding unduly large increases in the future. While the leverage requirement for capital would decline by $74 billion (as of March 2021 balance sheets), the actual capital required (which is, as stated above, the higher of the leverage requirement and a risk-based calculation) would not decline by anywhere near as much, since the risk-based requirement (changed only modestly by the revisions) would become the higher of the two, and thus binding, instead.  The net result would be a small capital reduction required at Freddie Mac, and none at Fannie Mae. (The FHFA’s proposal, curiously, doesn’t specify the reduction in dollar terms.) Importantly, prior to the proposed revisions, as the leverage ratio capital requirement is projected in the future to become much higher than the risk-based one, this unnecessary future increase can be avoided. In addition, there will be a significant improvement in the capital relief given on the existing book of CRT transactions, although the FHFA doesn’t specify the dollar amount, generating that modest reduction in the risk-based calculation. My back-of-the-envelope estimate is that, combined, these two changes would reduce the $283 billion requirement in the final Calabria ECR by only a small percentage, leaving it at no less than $250 billion, which is still far too high in my view.
  2. Decision-making at the GSEs can become less distorted, and more based on true economics and risk, resulting in better-run and more safe-and-sound companies. As stated above, the dominance of capital requirements by a leverage ratio is bad: risk does not play a role in decisions at the transaction level, and in aggregate it creates incentives to simply buy and hold high-risk assets, which is the opposite of what good policy should encourage. This will now be avoided as the leverage ratio will no longer be mostly binding (i.e. higher); it will instead become a credible backstop to a risk-based capital system, in line with best practice in financial institutions regulation.
  3. CRT transactions that are economically efficient can proceed full speed ahead. The anti-CRT bias of the Calabria ECR has mostly been eliminated. Such transactions can now be done when justified by the economics, comparing without distortion the cost of a particular CRT transaction with its benefits. The biggest such benefit, of course, is the reduction in capital required since CRT transfers risk to outside investors, and this can now be reflected in transaction economic calculations. Interestingly, since the Calabria ECR proposal was made early last year, Freddie Mac has continued to do CRT transactions, while Fannie Mae essentially stopped doing so; presumably the proposed Thompson changes will allow Fannie Mae to resume CRT transactions, thereby ceasing the old-school buy-and-hold strategy that it adopted when the Calabria ECR came on the scene.

Viewed in their entirety, these changes are not about reducing the total capital required today by the GSEsthat is reduced by a relatively small amount, which the FHFA has not precisely disclosed. They are instead focused on eliminating critical distortions so that the GSEs are not incented by a poorly constructed capital rule to make bad risk-reward decisions that result in a large concentration of high-risk assets on their books. Not fixing the distortions might cause an ugly outcome, with a systemic concentration of high-risk mortgage assets on the books of the GSEs that could, in some future stress event, cause them to fail again.

I am highly supportive of these changes. In fact, the core of the FHFA proposal document should be required reading among financial institution regulators and other interested parties for its excellent, well-reasoned explanation of why a leverage ratio-based capital requirement should only be a credible backstop to a risk-based one, rather than primarily binding. I have found too many policy specialists and elected officials, and even a few regulators, who do not understand that fundamental of regulation, and instead focus on a simple leverage ratio, which was commonly used by regulators in the 1980s and earlier.

But this begs the question: are these the only changes that should be made to the Calabria ECR? It still produces a high level of total capital required, which I estimate is at least $75 billion too much, and is increasingly inconsistent with the annual stress test results. That has implications for how difficult an exit from conservatorship would be, and what level of guarantee fees would be needed upon exit. As a result, there are members of the housing finance policy community who would like to throw the whole thing out and start again.

That question, or any additional significant changes, however, will have to wait for another day, perhaps when a confirmed director can examine policy concerns that will be relevant in the future, after the GSEs have retained considerably more capital and an exit from conservatorship becomes a more immediate issue. In the meantime, Acting Director Thompson has gotten to the heart of the Calabria ECR’s damaging distortions with this targeted proposal, and the GSEs—and the entire financial system—will be better for it being quickly finalized and implemented.