Skip to main content

Search form

Comments on Enterprise Capital Requirements FHFA RIN 2590-AA95

November 20, 2018
Mortgage Lending
Comment Letter

Download the full text of the comment.

In determining the capital standards for the GSEs, it is first critical to remember the primary drivers of the 2008 financial crisis and how those conditions have changed, affecting both the likelihood and severity of a future crisis. Next, the assumptions and mechanics of setting the capital regime must be closely examined in order to both set aside sufficient capital and enable the GSEs to provide their essential support for the housing market. Since the cost of holding capital to protect against a future crisis comprises the bulk of the total g-fees charged by the GSEs, care must also be taken to ensure the structure does not impose unnecessary burdens and obstacles to credit worthy borrowers seeking the opportunities of home ownership. Finally, the capital regime must recognize the legacy and continuing presence of discrimination in the housing market and must not perpetuate that discrimination.

Reviewing the proposed FHFA capital structure in light of the above standards shows the following:

  • The proposal contains useful provisions that acknowledge the important changes in the market since the last crisis, and it provides some measures to more accurately set and distribute capital than under the current GSE capital practices. However, substantial modifications are needed to establish an accurate, sufficient and productive capital model.
  • The model only partially recognizes the profound changes in the mortgage market and in the GSEs since the 2008 crisis. The ability to repay/qualified mortgage standards dramatically reduce the likelihood and depth of a crisis, and the proposed model does not fully take this into account. Similarly, the elimination of excessive GSE portfolios and the large risks they imposed and that are now removed are not considered in the model. As a result, the depth of a future crisis and the capital needed to withstand it are substantially overestimated.
  • The proposed model also undercounts by a large amount the resources available to the GSEs to absorb losses in a future crisis, and this further inflates the amount of needed credit risk capital. Omitted and undercounted resources include the proposed additional buffers for market risk, deferred tax assets and going concern buffers, as well as the complete exclusion of ongoing GSE premium revenue, which was proven to a reliable and substantial resource in the last crisis.
  • The proposed distribution of the crisis capital inaccurately and unfairly burdens working families. This capital is a buffer for a future systemic market failure above and beyond the credit risk capital that covers losses that occur throughout the business cycle. The model would disproportionately place the cost of this systemic failure capital on lower wealth borrowers, by a factor of as much as ten to one. These hardworking families are not the cause of these market failures, and they already shoulder much of the economic burden of these systemic failures. The capital model should not further aggravate this unfair distribution of costs of an overall market failure. Lower wealth borrowers are also overcharged by the failure to credit them for their lower and less volatile prepayment speeds that make their loans less costly to guarantee, and by the proposed risk multiplier surcharges that further inflate their capital charges.
  • Looking at the level of proposed capital in the model, for the above reasons, it is too high, and too much of it is placed on working families. Analyses set out in the sections below show that a significantly lower amount of capital would fully absorb the losses that a repeat of the 2008 would produce, even under conservative assumptions that generate a high estimate of those losses.
  • The proposed model properly acknowledges the positive impact of loan seasoning and that this should be considered in capital levels and pricing. However, an alternative model using a multi-vintage approach would better measure the risks across the multi-year portfolio of loans that will exist in a future crisis. This approach would produce a more accurate, less costly and less pro-cyclical capital structure.
  • Finally, the model must achieve safety and soundness, but it must do so in an accurate way that furthers the statutory purpose of the GSEs to ensure a broad, affordable and inclusive housing market. This is particularly important since in the evolving housing market most credit worthy new borrowers will be people of color. Moreover, the model must not perpetuate the discrimination that is a fundamental part of the past and current housing market. The current proposal needs the above changes to achieve these goals.