A friend of mine last week commented that she was a size 2 in high school, but somehow has become a size 8. If she were a bank, she’d still be forced into her size 2 jeans even if she could only pull them up over one leg and her ability to appear in public was, to say the least, impaired.
Current “tailoring” rules take no account of inflation or, even worse, much that matters when it comes to risk. In 2020, the banking agencies issued tailoring standards categorizing banks via a series of size and “complexity” thresholds that determine applicable prudential standards and supervisory vigilance, or so it was said at the time. The final rule also reserved the regulators’ right to alter a bank’s category –presumably to a tougher one – based on whatever worried them. In practice, the standards have been implemented almost exclusively by reference to a bank’s size and nothing – not even all of the risks evident at some mid-sized banks ahead of the 2023 crisis – led supervisors to look harder.
A bank below $250 billion was deemed simple and safe; any above that threshold, almost certainly not. Further, any bank above $700 billion turned into a pumpkin – that is, a GSIB – even if it is neither global nor systemically-important. Big equals bad even though bad is remarkably indifferent to asset size when there is rapid growth, ill-begotten incentives, lax supervision, or negligent risk management.
The banking agencies rightly plan to revise tailoring, and the FDIC has even begun to do so. However, the focus so far seems principally to be on relief for the smallest banking companies along with inflation indexing. That’s all fine, but it’s far from enough.
Asset size, inflation-adjusted or not, is a poor risk indicator, and the current approach to measuring complexity was flawed from day one – again see SVB, Signature, First Republic, and all too many others. The current complexity standard leaves out critical factors – rapid growth, massive unrealized losses, asset concentrations, and disproportionate reliance on uninsured deposits as just a few clear cases in point. The rule also assumes that nonbanking activities or non-credit exposures are risk-free because they are outside the scope of FDIC resolutions, resolutions which the final rule assumes occur without a wrinkle in weekend purchase-and-assumption transactions.
Been there, done that, it doesn’t happen. Maybe it will, but the FDIC’s new plans to allow greater bank/private-equity integration makes this even less likely and complexity still more dangerous and, under current rules, impenetrable.
Instead of fiddling with size categories, the banking agencies can and should use an improved GSIB-scoring methodology to spot higher-risk banking organizations regardless of size or putative resolvability. A sliding scale of scores up to the scimitar forcing GSIB designation would factor in size, but also consider far better measures of complexity. For example, the scoring methodology captures “substitutability,” – i.e., whether the economy or financial system could live without a bank if it ran aground. Some of the most systemic activities about to be launched in the banking industry relate to digital-asset infrastructure or issuance, activities with small asset footprints, but large concentrated-risk positions and operational vulnerabilities. The GSIB score sort-of catches this now and can be quickly improved in concert with more effective tailoring criteria.
Further, super-regional banks may well be big, but they’re unlikely to be bad when judged in a more sophisticated way because their exposures are capitalized, generally traditional, and community-focused. The super-regional problem is defending the business model from dominant GSIBs, not taking the risks GSIBs absorb thanks to their huge size and resulting economies of scope and scale. There is a deep gap between super-regional complexity and that of the GSIBs that a new scoring methodology can and should quickly capture so that tailoring rules appropriately reflect real risk, not just size.
Is the GSIB-scoring methodology perfect? Of course not. Its measurement of cross-jurisdictional exposures is particularly in need of correction as is the score’s limited consideration of short-term funding sources, indirect off-balance sheet exposures, and the other proven risk indicators noted above. These flaws can and should be quickly fixed and then thresholds can and should be set not just for when a bank is a GSIB or when a small bank poses big risks, but also for when a bank is so not even close that complex, super-stringent rules are nothing but over-kill ensuring greater industry consolidation at still more risk.