An Errata on a Recent Bloomberg Editorial on SRTs

A recent Bloomberg editorial incorrectly draws parallels between current synthetic risk transfers (SRTs) and pre-2008 private label mortgage-backed securities that contributed to the Global Financial Crisis. This response will address the editorial's key mischaracterizations. Readers interested in learning more about SRTs should read our recent primer.

SRTs are well-established in Europe and Canada but were virtually nonexistent in the U.S. until recently. The higher equity costs following Silicon Valley Bank's failure created incentives for U.S. banks to engage in SRT transactions. These instruments allow banks to shed credit risk and thereby receive a lower risk weight for the exposure, and thus a lower capital charge. According to a recent report, U.S. banks executed 15 SRT transactions during 2023.

In an SRT, a bank either sells the junior credit tranche of an exposure or hedges by entering into a credit derivative transaction. In the U.S., SRTs are typically structured as fully funded transactions, where banks receive the entire protection amount in cash at inception, often through the form of a credit-linked note (CLN) or other debt issuance. The bank's repayment to the counterparty depends on the performance of the pool of loans. Crucially, this upfront funding eliminates counterparty credit risk: the bank holds the cash and can retain it in the event the counterparty defaults.^[1]

Therefore, the editorial's assertion that "Unlike equity, it doesn't absorb any and all losses. It applies only to the designated assets, and it could prove worthless in a crisis if the counterparty can't pay," is factually incorrect. Even in partially funded transactions, which are less common, banks employ margin requirements to mitigate counterparty risk.

Second, by drawing parallels to the pre-2008 subprime mortgage boom, the editorial incorrectly suggests that banks are transferring risk from their least creditworthy loans. Most SRT instruments in the U.S. are short-term and reference a pool of high-quality assets, such as prime auto loans. The reference assets are chosen not because they are high-risk, but rather because regulatory capital charges overstate the risk of low-risk assets. In addition, since banks need to compensate investors for taking on credit risk, spreads are much higher for leveraged finance than auto loans, particularly when banks retain the first-loss positions. These higher spreads often make SRTs uneconomical from the bank's perspective.

Third, regarding concerns about banks lending to the same nonbank entities that are acquiring credit risk through SRTs, the editorial presents little evidence that this practice is widespread. Even so, when such lending occurs, bank loans to nonbank institutions are secured and overcollateralized, significantly limiting potential bank losses. Moreover, these nonbank entities are primarily owned by long-term institutional investors (such as pension funds and insurance companies) that have no incentive for a rapid withdrawal of their funding; they stand in contrast to the shadow banks (MMFs, CP investors, repo market participants, etc.) that ran during the GFC. Lastly, the closed-end funds owned by long-term investors have very little risk of insolvency under stress scenarios because these entities are generally very well capitalized.

Finally, beyond the mischaracterizations of SRTs, the editorial appears to advocate for a capital regime where the capital requirements that bind are leverage ratios. We have pointed out many times in the past (one example) that a binding leverage ratio encourages banks to take on more risk, not less risk, and ultimately leads to suboptimal economic outcomes.

[1] Because cash is risk-weighted at 100%, banks' leverage ratios are about unchanged after they engage in an SRT.