Has the Collins Amendment reached its endgame?

Scott Bessent wants to end the dual capital stack. How that would work in practice remains unclear

Key points Summary Q&A

• US Treasury Secretary Scott Bessent suggested regulators should consider eliminating the dual-stack requirement for bank capital calculations.
• The dual stack is a product of the Collins Amendment, and requires banks to assess capital using both standardised and internal models, applying the higher of the two.
• Bessent argued that the dual requirement lacked a principled methodology and simply inflated capital requirements.
• Ending the dual requirement could reduce compliance costs, but it would not on its own lessen the capital demands imposed by Basel III regulations.

These summaries are generated by an LLM, but edited by humans.

It is an unwritten rule in Washington that laws named after sitting senators cannot be amended or repealed without their consent.

US Treasury secretary Scott Bessent apparently didn’t get the memo.

In a speech in July, he urged prudential regulators to “consider abandoning” the so-called dual-stack requirement for calculating banks’ capital requirements.

The dual-stack stems from a provision of the Dodd-Frank Act introduced by Republican senator Susan Collins. The Collins Amendment requires US banks to calculate their capital adequacy using regulators’ standardised approaches as well as their own internal models, and to apply the higher of the two outputs.

Ending the dual requirement may lower compliance costs, but it won’t lower the capital hit from the Basel III endgame

Bessent questioned the rationale behind the amendment. “This dual-requirement structure did not derive from a principled calibration methodology,” he said in July. “It was motivated simply to reverse-engineer higher and higher capital aggregates.”

Collins may take issue with that framing.

Dodd-Frank made two crucial changes to the existing capital regime. First, it added capital requirements for operational risk and credit valuation adjustments (CVAs). Second, US banks were allowed to calculate capital using internal models for the first time.

The problem Collins saw was that few banks outside the top tier had the resources to deploy internal models, which tend to produce lower capital requirements than the regulator-set standardised approaches that characterised the original Basel I framework introduced in 1988.

Related

The Collins flaw: backstop turned binding constraint 

US legislative tweak was meant to prevent banks from using their own capital models too liberally. It’s now something different

The Collins Amendment was therefore designed to ensure the new rules did not lead to a reduction in capital requirements for the largest banks or give them an undue advantage over smaller lenders.

To turn the spirit of the law into a detailed rule, prudential regulators asked banks to calculate two sets of capital metrics and comply with whichever was higher. The so-called advanced approach included CVA and credit, market and operational risk exposures, and allowed banks to use internal models. The other, often known as the Collins floor, was essentially the old methodology, incorporating only credit and market risk and using only the standardised approach. The dual capital stack was born.

Within five years, the Collins floor had become the binding constraint for most large banks. And if it wasn’t enough already, the introduction of the new stress capital buffer (SCB) in 2020 made it so.

Previously, both capital stacks were topped off with a 2.5% capital conservation buffer. From March 2020, the Collins stack was instead augmented with a dynamic buffer derived from the maximum capital depletion in the severely adverse scenario for the Federal Reserve’s annual supervisory stress tests. The SCB generally works out a good deal higher than 2.5% – typically around 4–5% each year, prior to the gentler 2025 round of stress tests. That made the Collins floor even more binding compared with the advanced approach plus capital conservation buffer.

Careful what you wish for

By the time the SCB was introduced, US banks were already thinking about the next challenge on the horizon: the implementation of the third iteration of the Basel capital framework. Rumours that US regulators would diverge significantly from global standards by abolishing the internal ratings-based (IRB) approach for credit risk were weighing heavily on bankers’ minds.

The implications were potentially dramatic. Credit is by far the largest component of risk-weighted assets for most banks – typically accounting for more than 70% of total exposures. The industry settled on a lobbying strategy to try to mitigate the impact. The plan was to preserve the double stack, even though part of the rationale for the Collins Amendment would become redundant if internal models were being scrapped for credit risk.

The banks behind this plan were mostly those with large SCBs. They reasoned that, even if credit risk was no longer calculated using internal models, overall capital requirements under the advanced approach would still be lower than the Collins stack due to the size of their SCBs. In short, the implementation of Basel III would lead to no change in the binding capital constraint.

Related

CCAR at a turning point, but which way is forward? 

Banks sniff an opportunity to push the Fed for more openness over stress test models – and seize capital benefits

Then, one man threw a spanner in the works. In July 2023, when Michael Barr, then the Fed’s vice-chair for supervision, unveiled his proposal for implementing Basel III, he followed the lobbyists’ suggestion of retaining the dual requirement, but with one crucial deviation. The SCB would be applied to both stacks, replacing the 2.5% capital conservation buffer for banks relying on the advanced approach. With a stroke of the pen, Barr had turned the whole lobbying strategy upside down.

Market risk was the only internally modelled component in the new advanced approach stack proposed by Barr – which the Fed called the enhanced risk-based approach (ERBA) – and therefore the only element that might yield lower capital requirements than the Collins floor. The ERBA also contained operational risk and CVA, which are not part of the Collins floor. Unless those op risk and CVA numbers were tiny, the ERBA plus SCB was almost certain to be a new (and tougher) binding constraint for banks, leaving the Collins stack largely irrelevant.

That irrelevance most likely explains why Bessent called on regulators to dump the double stack in July. But it’s not clear how to solve the underlying problem. Axing the ERBA stack looks like a non-starter. Doing so would remove operational risk and CVA from the US capital framework altogether, but these are real balance-sheet risks that banks have to manage.

Perhaps senator Collins could be persuaded that her amendment is no longer necessary. But this wouldn’t prevent the big increase in capital requirements from Basel III, especially from the loss of internal models for credit risk. And if the Collins floor is eliminated, then the SCB would presumably still have to be added to the ERBA as the only remaining stack.

Put simply, ending the dual requirement may lower compliance costs, but it won’t lower the capital hit from the Basel III endgame. That would require fundamental changes to the ERBA.

Shortly before the November 2024 presidential election that brought Barr’s term as vice-chair to an end, he proposed a series of tweaks to the ERBA that would have eased its impact, especially for cleared derivatives. But these minor changes wouldn’t do much to lower capital requirements for the largest US banks. The most effective way to moderate the Basel III endgame would be to preserve the IRB for credit risk.

Strangely, this seems to be the one piece of deregulation for which there is bipartisan opposition. Barr’s successor, Michelle Bowman, has given no indication that she is preparing to save internal models for credit risk, and bankers speaking to Risk.net in recent weeks seem to have given up the ghost. For them, Bessent’s call to end the double stack is little more than a consolation prize.

Editing by Kris Devasabai