• The U.S. Basel III Endgame (“B3E”) proposal related to securitization prescribes the most restrictive approach to set capital requirements for banks’ securitization exposures in the developed world. At the same time, securitization markets are central to the majority of credit being extended in the U.S. economy, including mortgages, credit cards, auto loans, equipment and other small business financing, and other retail consumer assets. Punitive changes in the securitization capital framework will impact the cost of credit for virtually every consumer and business in the U.S.
• As proposed, the U.S. rules will, in many cases, result in significantly more capital for securitized assets than what is required under the current rules. This contradicts the proposed changes to the risk weighting for retail exposures which reduce the amount of capital in most cases, that banks will have to hold against retail loans. Additionally, because the rules are not risk sensitive (i.e., they do not take into consideration the expected performance or riskiness of the underlying loans), in many cases, more capital is required for securitizations of loans that are expected to experience relatively lower losses than for loans expected to experience higher losses.
• Taken together, this will result in more capital (read “higher cost”) for banks to finance securitized assets. Consequently, the U.S. B3E proposal could have severe detrimental impacts on the ability of banks to finance consumer, business, and other credit, and to make markets in securitization bonds, increasing interest rates and reducing the availability of credit, thereby harming main street as well as U.S. financial markets’ global competitiveness.

Background

Put most simply, securitization is a means of providing cost-effective funding to originators of consumer and business credit whereby those originators use their loans as collateral for borrowing or the issuance of securities.

Securitization allows banks and other lenders to provide more credit to consumers and businesses, and at a lower cost, than would be possible if they instead held the loans on their balance sheets. The loans can be placed in a securitization, where investors exchange cash for the bonds that are created. This same structure can be used where a bank provides a loan which, given protections in the form of excess collateral, should require less capital and thus, can be provided at a lower cost than the bank providing the consumer loans directly. In other words, securitization allows for a more efficient cycling of lending capital through the financial system. Securitization products are also used by banks and others to manage or hedge risk. Investors in securitization include mutual funds, pension funds, insurance companies, banks, hedge funds, corporate treasuries, sovereign wealth funds, and other foreign governmental entities.

According to SIFMA data, in 2022 over $1.5 trillion in mortgage-backed securities were issued, and over $200 billion in asset-backed securities were issued. Assets that are commonly securitized include residential and commercial mortgages (“mortgage-backed securities”), student loans, auto loans and leases, credit cards receivables, equipment, solar and cell phone tower lease cash flows, and other types of receivables (non-mortgage-backed securities are referred to as “asset-backed securities”). While it varies year to year, recently 70% or more of residential mortgage loans in the US have been funded by securitization, and studies have quantified how securitization has lowered the cost of obtaining a mortgage.^[1] This stands in contrast to Europe, where bank lending is a far greater component of consumer and commercial lending than securitization, and hence, the cost of consumer credit is generally higher.

Capital requirements play a key role in the ability of banks to participate in securitizations to fund lending. Higher capital requirements would force banks to hold less inventory leading to lower ABS liquidity and higher spreads which in turn raises costs for consumers and businesses.

How are capital requirements for securitization exposures calculated?

In 2016, the Basel Committee released the international standards for securitization capital framework, ^[2] which constitutes a part of the B3E standards published in 2017. The B3E offers four approaches to calculate capital requirements for securitization exposures – internal ratings-based approach (“SEC-IRBA”), external ratings-based approach (“SEC-ERBA”), internal assessment approach (“IAA”), and standardized approach (“SEC-SA”). The SEC-IRBA is the most risk-sensitive whereas the SEC-SA is the most conservative and least risk-sensitive approach.

In addition, in response to the global financial crisis the Basel Committee and the International Organization of Securities Commissions (“IOSCO”) published “Criteria for identifying simple, transparent and comparable securitisations” in July 2015. The goal of Simplicity, Transparency, and Comparability (“STC”) was to help stakeholders (e.g., originators and investors) evaluate the risks and returns of a particular securitization exposure, thereby “lower the hurdles of assessing securitisations” ^[3] and incentivize healthy growth of securitization markets. Securitization exposures meeting the STC criteria would enjoy preferential treatment (i.e., lower overall capital charges) under the SEC-IRBA, the SEC-ERBA (external ratings-based approach) and the SEC-SA under the Basel standards. The STC framework and the four capital treatment approaches have been adopted by all other major jurisdictions.

In July 2023, the U.S. banking agencies released their proposal implementing the B3E in the U.S. (“U.S. proposal”). In contrast to other major jurisdictions, the U.S. proposal removes the internal ratings-based approach for credit risk and does not adopt the STC framework. ^[4] The Dodd-Frank Act prohibits the use of external credit ratings for bank capital requirements. Consequently, the U.S. proposal does not implement the SEC-ERBA and adopts only the SEC-SA for capital charges on large banks’ securitization exposures. The SEC-SA determines the applicable risk weight based only on standardized parameters reflecting the broad category of underlying pool of assets (e.g., mortgage, corporate etc.), the seniority of the securitization exposure, and an important multiplicative adjustment called the “p-factor.” It completely ignores the expected performance of the underlying pool of loans in assessing capital requirements. As a result, the SEC-SA in the U.S. proposal could double or even triple the capital required on certain securitization exposures relative to the current U.S. capital rules and the securitization framework under the EU and Canada’s B3E implementation, which Federal Reserve Board Governor Michelle Bowman worries could bring “potential harm to U.S. bank competitiveness in the global economy.” ^[5]

As an illustrative example, banks lend to prime auto loan originators at an advance rate of c. 88% (i.e., the bank lends $88 collateralized by $100 of prime auto loans; losses on the loans would therefore need to exceed $12 for the bank to suffer any impairment on its loan). This lending would generally be rated AAA by the rating agencies which is commensurate with the over-collateralization (i.e., the $12 in the prior example) being sufficient to cover 4-5x historical losses. The table below compares the risk weight for this lending under the current U.S. capital rules, the U.S. B3E proposal and the approaches available to banks in other jurisdictions for the same lending.

For unsecuritized assets (e.g., loans), under the B3E proposal the broad category of the underlying pool of assets determines the capital requirements for holding them, but for securitized assets, the p-factor is an important additional parameter. The p-factor plays 2 main roles in the SEC-SA framework: (1) it controls the degree of capital penalty for securitization (i.e., it causes the aggregate capital required for holding all tranches of a securitization to exceed that for holding the underlying pool of assets alone, thereby disincentivizing the use of securitization as a credit risk transfer tool by banks) – often referred to as “securitization capital non-neutrality”, and (2) it controls the allocation of capital across different tranches of a securitization. ^[6]

Figure 1 below illustrates the impacts of raising p-factor to 1 from 0.5 on securitization capital non-neutrality and the re-allocation of capital requirements across securitization tranches. It is clear the securitization is capital non-neutral since the securitization capital non-neutrality far exceeds 100%, a larger p-factor only exaggerates the securitization capital non-neutrality (from 351% to 455%) making it more expensive to securitize assets.

Figure 1. The impacts of p-factor on securitization capital non-neutrality and risk weight across securitization tranches. Assuming tranche thickness equals to 2% with attachment point ranges from 0%-42% and detachment point ranges from 2%-44% respectively, 

How does the U.S. proposal create perverse incentives for securitization exposures?

The U.S. proposal could create perverse incentives for securitization exposures due to (1) revisions to the p-factor and the risk weighting of underlying pool of assets, and (2) the lack of appropriate risk-sensitivity of the SEC-SA framework.

The SEC-SA calculates tranche risk weight base on four inputs – attach and detach points, the ratio of delinquent underlying exposures to total underlying exposures in the securitization pool (i.e., W), and the standardized capital requirements for the securitization pool (i.e., K_g), in addition to the p-factor. K_g is adjusted to account for the impacts of W via parameter K_A (defined as ). That is, K_A effectively ascribes a 625% risk weight to delinquent exposures underlying securitization in contrast to a 150% risk weight held directly on balance sheet. This is further compounded by a higher p-factor.

The U.S. proposal would double the p-factor to 1 from 0.5 under the current U.S. capital rules leading to higher capital requirements for securitization exposures. The proposal would ascribe a lower risk weight to certain retail loans held on a bank’s balance sheet (e.g., 85% for prime auto loan exposures) than the current U.S. capital rules (e.g., 100% for prime auto loan exposures) which ought to result in lower capital requirements for securitization exposures. The combined impacts of both changes however, tend to raise capital requirements for senior tranches while lowering capital requirements for the junior-most tranches as shown in Table 1 below.

Table 1. The SEC-SA risk weight applicable to prime auto loan (classified as regulatory retail exposures, i.e., assigned 85% risk weight under the U.S. proposal, but 100% risk weight under the current U.S. capital rules) securitization exposures with the p-factor=1 as proposed in the U.S. proposal.

Table 1 shows that the risk-weighted asset (“RWA”) and, as a result, the required capital for the retaining Class A, AAA rated notes in the structure essentially doubles. ^[8] Take the A2A tranche for example, the risk weight would go up from 30.9% under the current U.S. capital rules to 61.1% under the U.S. proposal – an increase of 98%. However, the capital increase bears no relation to the actual risks inherent in the underlying pool of prime auto loans. In fact, relative to the current capital rules the U.S. proposal would ascribe a lower risk weight to prime auto loans.

The overall securitization capital surcharge (i.e., non-neutrality) for this securitization would rise from 153% under the current capital rules to 172% under the proposal – a 2000 basis points increase, even though the underlying pool of assets are considered less credit risky (i.e., assigned a lower credit risk weight under the proposal). A capital surcharge of 172% indicates that the capital required for holding the securitization would have been over 1.7x of that for holding the underlying pool of prime auto loans. As a result, the U.S. proposal could make it more expensive for banks to transfer credit risk via securitization. In addition, banks could be incentivized to shift out of senior tranches in exchange for more junior and riskier tranches – a perverse incentive that counters the principle and practice of sound risk management.

Additionally, unlike the SEC-IRBA whereby the p-factor is dependent on the expected performance of the underlying securitization pool (including the probability of default and loss given default), the SEC-SA fixed the p-factor at 1 ignoring the expected performance of the underlying securitization pool. Consequently, the SEC-SA under the U.S. proposal could require lower capital on a senior tranche backed by a subprime pool than a senior tranche backed by a prime pool despite the fact that expected losses on the subprime pool will erode more of the collateral balance than the expected losses on the prime pool. Table 2 shows that capital requirements for retaining most senior tranches backed by a pool of subprime auto loans would decrease under the U.S. proposal. This is in stark contrast to the capital increase for senior tranches backed by prime auto loans reported in Table 1. All else equal, this could result in banks either having to charge substantially more on the securitized loans to the prime auto lender than to the subprime auto lender or to seek to lend more in subprime than prime.

Table 2. The SEC-SA risk weight applicable to subprime auto loan (classified as regulatory retail exposures, i.e., assigned 85% risk weight under the U.S. proposal, but 100% risk weight under the current U.S. capital rules) securitization exposures with the p-factor=1 as proposed in the U.S. proposal.

The U.S. proposal would subject a large bank’s trading and market-making activities to the Fundamental Review of the Trading Book framework (“FRTB”). The FRTB consists of two capital components – general market risk capital and issuer default risk capital (“DRC”). For the purpose of the DRC, the risk weight applicable to certain securitization exposures would be calculated using the SEC-SA framework. ^[9] Thus, these same perverse incentives would carry over even if the bank holds these tranches for the purpose of trading or market making.

How to mitigate the perverse incentives the U.S. proposal creates for securitization exposures?

The root cause of these perverse incentives is the lack of appropriate risk-sensitivity of the SEC-SA framework which could have detrimental effects on the functioning of the U.S. securitization markets. A few actions could be taken to mitigate such perverse incentives: (1) at the very least, revert the p-factor to 0.5 as in the current U.S. capital rules instead of doubling it as in the U.S. proposal, (2) adopt the SEC-IRBA, and (3) implement the STC framework.

The U.S. proposal should revert the p-factor to 0.5 from 1 to avoid the perverse incentives created by a higher p-factor and reduce the degree of securitization capital surcharge. Our concerns with the excessive securitization capital non-neutrality are shared by several major jurisdictions where mitigation actions are being taken. For example, considering that the “[risk-weighted amount] resulting from the application of the SEC-SA is not commensurate with the risks posed to the institution or to financial stability”, the UK Prudential Regulatory Authority published a discussion paper on “adjustments to the Pillar 1 framework for determining capital requirements for securitisation exposures”. ^[10]

In addition, the U.S. proposal should adopt the SEC-IRBA. As explained earlier, the SEC-IRBA takes into account the expected performance of the underlying pool of assets in setting capital requirements for securitization exposures. The SEC-SA, however, ignores the expected performance of the underlying pool. As a result, it is the least risk-sensitive and most conservative securitization framework offered by the Basel standards. To ensure capital requirements that are commensurate with risks arising from securitization exposures, the SEC-IRBA should be adopted.

Finally, the U.S. proposal should implement the STC framework. The Basel standards set out the STC criteria to help mitigate the uncertainty related to asset risk, structural risk, governance, and operational risk associated with securitization. Less uncertainty and more confidence in the performance of STC transactions would justify a reduced degree of conservatism being built into the securitization capital frameworks through capital non-neutrality. The Basel Committee states explicitly that “[a]ll other things being equal, a securitization with lower structural risk needs a lower capital surcharge than a securitization with higher structural risk; and a securitization with less risky underlying assets requires a lower capital surcharge than a securitization with riskier underlying assets.” ^[11] The STC framework would help lower the hurdles of assessing securitization exposures and incentivize healthy and responsible growth of the U.S. securitization markets.

Conclusion

The U.S. proposal requires large banks to set capital requirements for securitization exposures using the SEC-SA approach. This framework is the least risk-sensitive and most conservative amongst the four approaches offered in the Basel standards and adopted by other major jurisdictions. Additionally, the proposal does not implement the STC framework which was designed by the Basel Committee to support healthy and responsible growth of securitization markets. As a result, the proposal would result in capital requirements that are not commensurate with risks of the securitization and could create perverse incentives for banks’ involvement in securitization markets. Securitization markets have been a cornerstone of the U.S. capital markets and a key source of funding for the broader U.S. economy. Without appropriate mitigative actions, the U.S. proposal could have detrimental effects on the securitization markets and the broader economy

Authors

Dr. Guowei Zhang is Managing Director and Head of Capital Policy for SIFMA

Chris Killian is Managing Director, Securitization and Corporate Credit for SIFMA

Footnotes

^[1] See, e.g., “TBA Trading and Liquidity in the MBS Market”, James Vickery and Joshua Wright (2013), available here: https://www.newyorkfed.org/medialibrary/media/research/epr/2013/1212vick.pdf

^[2] See https://www.bis.org/bcbs/publ/d374.pdf

^[3] See https://www.bis.org/bcbs/publ/d332.pdf

^[4] See https://www.sifma.org/resources/news/the-federal-reserve-should-remove-gold-plating-in-the-basel-3-endgame/

^[5] See https://www.federalreserve.gov/newsevents/speech/bowman20231109a.htm.

^[6] In the SEC-SA and the SEC-IRBA, “the p-factor is calculated on a tranche basis with the senior tranche typically having a lower p-factor compared to subordinated tranches. All things being equal, a higher p-factor for the mezzanine tranche relative to the senior tranche would result in higher capital requirements for the mezzanine tranche.” (https://www.bankofengland.co.uk/prudential-regulation/publication/2023/october/securitisation-capital-requirements#footnote-6)

^[7] Attachment point for a securitization tranche represents the threshold at which credit losses will first be allocated to the tranche. And detachment point represents the corresponding threshold at which credit losses of principal allocated to the position would result in a total loss of principal.

^[8] The capital reduction for holding the most senior tranche is because of the lower risk weight floor (i.e., 15%) relative to the current rule (i.e., 20%), but in this case, relates only to a Money Market tranche which is structurally senior in order to qualify under Rule 2a-7 but is a short, relatively small part of the senior capital stack.

^[9] FRTB classifies securitization exposures into two groups – correlation trading vs non-correlation trading. The prime auto loan securitization transaction presented in Table 1 would be considered non-correlation trading. The risk weight applicable to non-correlation trading securitization exposures for the purpose of DRC is calculated using the SEC-SA.

^[10] https://www.bankofengland.co.uk/prudential-regulation/publication/2023/october/securitisation-capital-requirements

^[11] See https://www.bis.org/bcbs/publ/d374.pdf

Related Resources

• Pennsylvania + Wall

  Blog Series on US Bank Capital Requirements

• SIFMA Research

  US Asset Backed Securities Statistics