Opposition to the Basel Endgame capital proposals has focused on populist red herrings, like the supposed adverse impact on credit availability to first time home buyers. These narratives obscure a bigger story. A closer look at the numbers indicates that most of the largest banks have an effective capital charge of zero for operational risk under current rules. The Basel Endgame would change that and it’s a big deal.
Standardized and Advanced Approaches
Larger banks must meet both leverage and risk-based capital requirements. The leverage ratio equals capital divided by total assets. Risk-based requirements adjust the denominator based on the risk of the underlying assets. Two risk-based capital approaches apply to the largest and most complex banks (LISCC [1] banks). The Standardized Approach (SA) assigns a set of risk weights based on broad categories of assets. For example, U.S. Treasuries are risk weighted at zero, most residential mortgages at 50%, and commercial loans at 100%. The Advanced Approach (AA) relies more on the banks’ internal models, but also includes some elements, such as operational risk, not present in the SA.
The Collins Amendment requires banks subject to the AA to also calculate risk-weighted assets under the SA and be subject to the stricter of the two standards. The rule arose from a concern that capital requirements for the most complex and systemically important banks were more lenient than those for everyone else. Collins addressed that specific concern but only for the top of the house requirement and not its individual components. These individual components are fungible, with low credit risk weights under the AA offsetting the operational risk capital charge. Big banks wind up with very low risk weights for credit risk (AA) or an operational risk capital charge of zero under the SA.
The Basel Endgame proposal rests on three key observations. First, modeling approaches to estimate capital charges for credit risk and operational risk don’t work. Second, the response to large bank failures in early 2023 indicates that the current regulatory framework defines “systemically important” banks too narrowly. Third, a modeling approach for market risk can still work, but the approach requires substantial revisions. I’ll focus on the first two and leave the third for another time.
Problems with Model-Based Approaches
As I noted in an earlier post, capital risk weights for residential mortgages under the AA are extemely generous, with most falling in the 10-15% range.[1] Taken by itself, the AA would allow residential mortgages to be leveraged by about 80 to 1. For reference, the risk weight for a “prudently underwritten” residential mortgage is 50% under the SA. The SA risk weight for Fannie/Freddie MBS, which are de facto full faith and credit obligations, is 20%.
Modeling for operational risk is likewise problematic. While operational risk losses are common, the biggest loss events tend to be idiosyncratic and unlikely to follow anything resembling a normal distribution. Regulators recognized these limitations in capital stress testing years ago.
Basel Endgame Proposal
The Basel Endgame proposal would ditch the AA’s models-based approach to credit and operational risk and replace it with an amped-up version of the SA. Moving credit risk weightings under the AA closer to what already exists under the SA would result in a real capital charge for operational risk. The proposal also expands the AA from eight US-domiciled banks to about 35. Making the AA less reliant on models would likely ease that transition.
The table below summarizes risk-weighted assets under Standardized and Advanced Approaches. SA RWA is higher for six of the eight BHCs.
The difference is even more dramatic if we only look at credit RWA under the two approaches. Every BHC has lower credit RWA under the Advanced Approach. On average, credit RWA under the AA is only 66% of credit RWA under the SA. In aggregate, the difference totals $2.3 trillion.
The lower risk weights assigned to credit risk under the AA usually makes the SA the stricter risk-based requirement. But it’s not just that. Lack of an operational risk component under the SA means capital charges for operational risk are effectively zero for most large banks. Even those banks that have higher capital charges under the AA receive a significant offset to their capital requirements due to low credit risk weights. The impact is huge. As shown below, aggregate RWA attributed to ops risk for the eight large banks was nearly $1.8 trillion, or more than a quarter of total RWA.
The Mortgage Red Herring
Large banks launched an astroturf campaign against the proposal emphasizing how it might hurt lower income and first-time home buyers. Large bank lobbyists and CEOs laid it on thick with the Main Street routine. The largest banks don’t devote a large share of their balance sheets to mortgages. Real estate loans are less than 10% of total assets for most LISCC BHCs,[1] compared to a median 40.76% for all BHCs over $10 billion. One CEO also noted the proposal’s impact on loans to farmers, which were 0.03% of that BHC’s total assets. The mortgage proposal ties risk weights to loan-to-value (LTV) ratios, but in substance it doesn’t differ much from the current SA. Risk weights for high LTV mortgages come out a little lower under the Basel Endgame compared to the SA. The impact of the Basel Endgame on operational risk capital charges is both larger and less ambiguous.
[1] A median of 5.60% and a mean of 7.68%.
Regulators have long struggled identify the “correct” capital charge for operational risk. It remains to be seen whether the current models-based approach, the Basel Endgame formula, or some other method is the best way to measure operational risk. But it shouldn’t be zero, as it is for most banks under the current framework. According to the OCC’s Semiannual Risk Perspective, 42% of outstanding matters requiring attention relate to operational risk.
Large banks have complained about capital requirements that they deem as excessive. It’s worth noting, however, that the largest banks are considerably more leveraged than most other banks. Median and average leverage ratios for the current LISCC holding companies as of September 2023 were 7.09% and 6.98%, respectively. In contrast, the median leverage ratio for all banks over $10 billion was 9.47%, or almost 250 basis points higher. Seven of the eight banks fell below the tenth percentile in this measure. Supplemental leverage ratios, which consider off-balance sheet derivatives exposure, are even lower. Why should the largest, most complex, too big to fail banks leverage their balance sheets more than most other banks?
[1] Goldman Sachs is higher at 41.71% while Morgan Stanley is lower at only 5.14%.
[1] Large Institution Supervision Coordinating Committee, the Fed’s shorthand for systemically important bank holding companies (BHCs).