Federal Reserve Vice-Chairman Michael Barr recently previewed “broad and material changes” to the Basel Endgame capital proposal. Summarizing changes to a 1,000-page proposal in a 24-minute speech will leave out a lot of details. However, we can get a good sense of the proposed changes based on what we already know about the old proposal and the details that Barr has supplied.
Tiering
A handful of the largest and most complex banks are subject to additional capital requirements related to credit risk modeling, trading activity, operational risk, and the inclusion of unrealized losses in regulatory capital. The original Notice of Proposed Rulemaking (NPR) would make a much wider group of banks subject to these additional requirements. The revised proposal excludes banks between $250-750 billion from revised market risk (trading) and credit valuation adjustments (CVA). They would still be subject to the new requirements for credit risk, operational risk, and unrealized losses.
Banks between $100-250 billion would only be subject to the unrealized losses requirement. The rules only apply to losses on available-for-sale (AFS) securities and not those designated as held-to-maturity (HTM). Requiring more banks to reflect unrealized losses in capital represents the most direct response to the Silicon Valley Bank failure. This change could incent banks to designate more securities as HTM, providing them with less flexibility to address interest rate risk. That’s exactly what happened when the largest banks became subject to the requirement.
However, that is also likely a long-term rather than a short-term concern since most bond portfolios are already underwater. If a bank transfers its AFS securities to HTM, the unrealized loss goes into Accumulated Other Comprehensive Income (AOCI). Banks could still take a substantial AOCI hit. For example, Charles Schwab Bank reported (June 2024) an unrealized loss on AFS of $7.7 billion and a total unrealized loss of $20.6 billion. The bank reported AOCI of -$14.8 billion, which roughly reflects the total tax-effected loss.
Mortgages
The proposed treatment of residential mortgages represented the most controversial element of the NPR. The NPR would ditch a model-based approach for the largest banks with one that assigns risk weights using loan-to-value (LTV) ratios. The revised proposal appears to follow a similar approach, but with more generous risk buckets. The original proposal assigned more conservative risk weights than the Basel Committee recommends for international banks. Regulators plan to make the risk weights “in line with the calibration developed in the Basel process.” The table below shows risk weights by LTV range under the NPR, under the Basel standard, and under the Standardized Approach that applies to small and midsize banks.
Barr cited competitive equality with domestic banks as a rationale for the initial set of risk weights. The revised approach gives a clear advantage to larger banks. Foreign banks don’t generally compete with U.S. banks in the residential mortgage market. It’s a little hard to see why smaller banks would like to provide their larger peers with a competitive advantage. Barr’s speech implicitly assumes that the Standardized Approach applies a 50% risk weight under all LTV buckets. A closer look at the capital regulation suggests otherwise but the regulations are also maddeningly vague in this respect.
Barr left unaddressed two other controversial aspects of the original proposal. Current regulations allow high LTV loans to qualify for a higher risk weight if they have private mortgage insurance (PMI). Regulators expressed concern with the enormous systemic and wrong-way risk that reliance on PMI can pose. A PMI company may have deeper pockets than a community bank, but their capacity pales in comparison to the mega banks to which this rule would apply. The NPR also excluded a bank’s model-based estimates in calculating mortgage risk weights under the Advanced Approach. These models failed miserably during the Global Financial Crisis. Current estimates do not appear realistic, either. For example, the models-based approach results in an effective risk weight of 11.05% for residential mortgages at the nation’s largest bank. Fannie and Freddie MBS, which have virtually no credit risk, have a risk weight of 20% under the Standardized Approach.
Securities Financing Transactions (SFTs)
Current risk-based capital rules are unusually generous towards SFTs, such are repos and margin loans. The presence of collateral can not only reduce the risk weight of these transactions but can directly reduce the balance subject to risk weighting. A loan to a risky borrower secured by risky collateral can still have an effective risk weight of zero. RWA reductions aren’t dollar for dollar – regulators reduce the collateral value through a “haircut” tied to its market volatility. Even with these haircuts and daily marking of the collateral to market, banks can and do lose money on SFTs. That’s especially true in cases that may involve financial fraud, such as Steinhoff and Archegos. The NPR proposed minimum haircuts (shown below) when the counterparty is unregulated. Firms subject to prudential regulation (e.g., banks) are less likely to commit massive fraud, or so the theory goes.
The revised proposal will remove this minimum haircut floor. Barr acknowledged that this feature of the NPR was consistent with the Basel Standard but noted that “several other major jurisdictions” have not adopted this approach. U.S. regulators are electing to kick the can down the road, likely secure in the belief that they can toss front-line supervisors under the bus if we see more SFT meltdowns.
Tax credit equity funding
Regulatory capital rules generally assign a 400% risk weight to equity exposures that are not publicly traded. However, current rules assign a risk weight of only 100% to “non-significant equity exposures.” The NPR proposed to eliminate this carve-out, which would largely affect tax equity investments in clean energy projects. This element of the proposal generated considerable opposition, especially from more progressive members of Congress that might otherwise support higher capital standards for banks. The proposed revision will restore the 100% risk weight for these exposures.
Operational Risk
As I noted in an earlier post, the expressed concerns with the Basel Endgame’s approach to mortgages obscured a bigger story. Capital charges for operational risk only apply under the Advanced Approach. However, since the model-based risk weights are so low, the Standardized Approach is typically the more binding constraint. That means the effective capital charge for operational risk is zero for most banks. The revised proposal tweaks the operational risk charge. Barr did not quantify the impact of these changes in his speech.
One expected change deserves further scrutiny. The revised proposal will “no longer adjust a firm’s operational risk charge based on its operational loss history,” to “reduce fluctuations in a bank’s operational risk capital requirements over time.” It’s hard to see how a bank’s actual experience isn’t relevant, especially since it can reflect its operating model and control environment. Bank regulators like stable and predictable capital ratios because predictability makes their jobs easier.
Trading Activity
Changes in the treatment of trading activity have been in the works for more than a decade and reflect the lessons from the Global Financial Crisis. I don’t want to get too much into the weeds here but want to focus on one proposed change from the NPR. The re-proposal will introduce a multiyear implementation period for the profit and loss attribution (PLA) tests. PLA tries to align a bank’s risk models with its actual experience. Say you have a Rates Trading desk and the risk model includes assumptions regarding duration, convexity, volatility, and correlation. You can assess how well that model works by comparing the theoretical P&L based on the risk model and market moves with the actual P&L.
PLA is a Trading 101 concept and large trading operations do this as a matter of course. There is, however, a range of practices of how banks implement PLA, and the Basel Endgame will set some minimum requirements and capital penalties for falling short. I’m sympathetic to the idea that making these adjustments will be a major undertaking for some (maybe most) banks. However, since these proposals have been in the works for years, banks have also had a lot of time to get their risk models and systems up to speed.
Impact
Regulators estimated that the initial Basel Endgame proposal would increase capital requirements by roughly 16% in aggregate. The largest and most complex banks would see the largest increase. Barr estimates that the revised proposal would raise capital requirements by roughly 9% for the largest and most complex (G-SIB) banks. Incorporating unrealized losses in capital for the non-GSIB banks would add to capital requirements by roughly 3-4%. The other requirements would increase capital requirements by 0.5%. While indications of the proposal’s aggregate impact are helpful, a further breakdown might be especially useful. Such a breakdown could compare capital requirements under existing rules, the original Endgame proposal, and the current proposal for each major area of the proposal. Capital requirements are fungible and can interact with one another, but regulators can at least provide ranges. That will help the public better understand the proposal.