Mortgage credit availability in the United States has remained subdued over the entire post-financial crisis period. The Urban Institute’s Housing Credit Availability Index remains close to its historically lowest value (Exhibit 1). Evidence of persistently tight mortgage credit standards is also seen in other popular mortgage credit availability indexes, such as the American Enterprise Institute’s Mortgage Risk Index, or National Mortgage Default Rate. This index was recently extended further back historically by Davis, Larson, Oliner, and Smith (2021) and shows that mortgage credit conditions have also been tight relative to those that prevailed in the mid-1990s. Arguably, important contributing factors include lack of product innovation in the mortgage market, as well as anemic competition in the secondary mortgage market due to insufficient private capital. Private capital is important to ensure access to responsible mortgage credit for communities that are historically and currently underserved by the mortgage market, as access to wealth through homeownership remains persistently unequal in the United States.
In this post, we describe how the U.S. banking agencies—the Federal Reserve, OCC and FDIC—could expand availability of mortgage credit for low- and moderate-income households. This approach would enable a more efficient shift of mortgage credit risk from banks to the private sector and revitalize the private securitization market without relaxing bank capital requirements or inducing a deterioration of underwriting standards. More specifically, the U.S. regulatory capital rules need to be amended to encourage the transfer of credit risk to the private market, subject to rigorous rules. This would enable banks to apply securitization capital treatment to transactions where assets that remain on the U.S. GAAP balance sheet that result in (1) a bona-fide transfer of credit risk on those assets; and (2) where those assets have been put presumptively beyond the reach of the originating bank’s creditors even in the event of bankruptcy or receivership (for example through a transfer that meets the FDIC Safe Harbor). These are so called Consolidated Securitization Risk Transfer structures (CSRTs).
The application of securitization capital treatment to those assets would also bring existing U.S. regulatory capital rules in line with the international Basel capital framework. This would significantly expand the ability of banks of all sizes to distribute mortgage credit risk (and other retail credit risk) to private investors. It would also foster more competition in the mortgage secondary market and reduce reliance on Fannie Mae and Freddie Mac (collectively the government-sponsored enterprises, or GSEs) for mortgage funding. Importantly, the ability to transfer risk while maintaining economic and servicing interest in the loans ensures alignment of interest (i.e., having skin in the game). It also gives banks the opportunity to test innovative product features aligned to first-time home buyers and affordable housing initiatives beyond product offerings currently available through government-backed programs (i.e., available through the GSEs, FHA, VA and USDA).
A CSRTs transaction allows the bank originating a mortgage loan to transfer credit risk on a pool of mortgage loans to investors in capital markets through issuance of fully funded asset-backed securities (ABS). Exhibit 2 illustrates the structure of a typical CSRTs deal.
The originating bank legally transfers a pool of mortgages (or other credit exposures) to a funding vehicle (securitization SPE) in compliance with an existing FDIC rule on what types of entities are bankruptcy-remote for purposes of assessing a bank’s risk. This is known as the Safe Harbor rule.[1] The securitization SPE issues ABS tranches whose cash flows depend on the performance of the underlying mortgage. The originating bank retains the senior tranches of the ABS, while the lower-rated tranches, including the first loss, are typically sold to third-party investors. The originating bank also retains the servicing of the mortgages so that it maintains its client relationships.
A policy issue arises because the bank would typically be required to consolidate the securitization SPE under U.S. GAAP, given its control over the mortgage assets as servicer and its retained economic interest in the mortgages through the senior ABS tranches. Because the transferred mortgages remain on the originating bank’s consolidated U.S. GAAP balance sheet, the transaction does not qualify for securitization capital treatment under the current U.S. capital rules, despite the bona fide transfer of credit risk.
Banking regulators generally prefer that the regulatory capital treatment follow the accounting treatment, even when the goals of the accounting rules do not coincide with those of the capital rules. However, following the accounting rules in this case means overstating the risk of losses on mortgage loans to the bank. This discourages banks from holding mortgages on balance sheets, potentially curtailing availability of credit to low- and moderate-income borrowers who are not currently served by the traditional products from the existing government-backed programs (GSEs, FHA, VA, and USDA). Following the accounting rules also puts the U.S. rules out of conformance with the international standard, which does consider actual risk in the calculation of risk-weighted assets. A further consequence of not recognizing the transfer of credit risk in CSRTs transactions is to continue to concentrate mortgage credit risk on the balance sheet of the GSEs and the U.S. government, leaving taxpayers exposed to the systemic risk posed by these entities in situations of market stress.
If capital requirements reflected actual risk, banks would be incentivized to originate and hold mortgages that benefit from credit risk transfers through CSRTs, instead of originating them strictly for sale to the GSEs. The risk weight under the securitization capital approach would then be comparable to that for holding GSE securities. This would make mortgage credit more readily available because investors are willing to take on marginally more risk. Creditworthy applicants not currently served by government-sponsored programs would also benefit from improved efficiencies in the mortgage origination process. Enabling capacity for banks to hold more mortgages in portfolio would facilitate needed changes to recalibrate GSE regulation. These include capital and liquidity requirements that are likely to result in an increase in GSE guarantee fees, mitigating any potential impacts of such changes on overall mortgage credit availability.
Incentivizing banks to retain mortgages is especially important for creditworthy borrowers where the GSEs require a more intensive manual underwriting process, involving the collection and review of additional information for the credit decision.[2] Typically, such applications are from households making smaller down payments, because they have more limited savings and assets, or from those with below-prime credit scores or higher payment-to-income ratios. These households often have lower incomes or are people from minority groups. The manual GSE underwriting process is more likely to end up with the borrower paying a higher rate or with the application denied. In those cases, it would be more efficient for banks to deploy internal underwriting and pricing processes, rather than engage in a lengthy back-and-forth with the GSEs while facing an uncertain outcome. However, without a viable option for retaining the loans in portfolio and distributing the associated credit risk, banks are generally unwilling to take on a direct origination and servicing role.
It is also important to note that, unlike existing products in the markets such as credit-linked notes (CLNs), CSRTs can be used by both small and large banks, because investors are not exposed to the counterparty credit risk of the originating bank.[3] This allows more banks to underwrite and hold mortgage loans, which increases the availability of mortgage credit, without undue loosening of underwriting standards or an excessive rise in mortgage default risk.[4] For example, a creditworthy borrower rejected by the GSE models may be approved by another model that does not rely on the same information set. More broadly, greater diversity in models or ways of evaluating credit quality would create incentives for financial institutions to innovate. For example, this might include developing new mortgage products tailored to community needs and adopting alternative data to enable improved risk identification and measurement, particularly within local or regional contexts.
Moreover, by establishing alternative channels for underserved applicants, banks would be better positioned to provide alternatives to borrowers in addition to those offered in the government-insured mortgage loan market (FHA, VA, and USDA), reducing the dependence of low- and moderate-income neighborhoods on the support of the federal government. In addition, CSRTs could play an important role in supporting financial inclusion by bringing more private-sector capital to support lending to historically underserved populations. CSRTs could thus help promote access to wealth through homeownership which remains persistently unequal in the United States—often in communities of color.
Finally, as is generally recognized about GSE issuance of credit risk transfers, “CRT programs have created a new financial market for pricing and trading mortgage credit risk, which has grown in size and liquidity over time.”[5] In the context of bank issuance of CSRTs, market-implied pricing of such structures will give useful feedback on bank internal processes for measuring and managing mortgage credit risk. More efficient pricing of mortgage credit risk as promoted by CSRTs can be expected to benefit both lenders and borrowers in the long run.
Conclusion
The current securitization capital rules overstate a bank’s risk of holding a pool of mortgages for which there has been a bona fide transfer of credit risk through a CSRTs transaction. The result is inconsistent with international capital standards and a drag on multilateral efforts to increase credit availability to low- and moderate-income households and reduce the concentration of credit risk in taxpayer-backed GSEs.
[1] 12 CFR 360.6. This rule ensures that the transferred assets are out of the reach of the originating bank’s creditors, even in the event of its bankruptcy or receivership administered by the FDIC.
[2] This more intensive process determines whether the prospective borrower’s risk of default (or the risk associated with the value of the property as collateral) is within the range deemed acceptable to the GSEs.
[3] CLNs have several practical limitations for issuers and investors. First, due to their exposure to counterparty credit risk, they are limited to large issuers with favorable credit ratings. Second, CLNs have a limited investor base and are therefore difficult to scale. CLNs also create an accounting asymmetry between assets and liabilities, since the underlying loans are subject to accrual accounting while issued CLNs are carried at fair value.
[4] For various reasons, CSRTs are unlikely to promote the kind of excessive risk-taking that occurred during the housing bubble period, when private mortgage-backed securities proliferated. The fact that the originating bank retains the senior tranches under the strengthened post-crisis capital regime disincentivizes excessive loosening of credit standards. CFPB rules intended to protect consumers from harmful mortgage products, such as Ability-to-Repay rules, also help curtail excessive risk taking. Finally, as argued in Susan M. Wachter, “Credit Risk Transfer, Informed Markets, and Securitization,” Federal Reserve Bank of New York Economic Policy Review, Vol. 24, No. 3, December 2018, pp. 117–137, three features of the CRT market, as currently constituted, make excessive risk-taking unlikely: (1) transparency, through the full provision of information on the mortgages underlying the CRTs; (2) open pricing in liquid markets; and (3) no counterparty risk.
[5] See David Finkelstein, Andreas Strzodka, and James Vickery, “Credit Risk Transfer and De Facto GSE Reform,” Federal Reserve Bank of New York Economic Policy Review, Vol. 24, No. 3, December 2018, pp. 88–116. For further discussion of the informational role of CRTs, see here.
