In the wake of the 1,089-page capital proposal, debate is waging on well-trod battlegrounds such as whether the new approach will dry up credit and thus stifle growth. I’ve my own view on this, but my initial read of the proposal points to a still more fundamental issue: some of it makes absolutely no sense even if one agrees with the agencies’ goals. Here, I lay out two bedrock assumptions that contradict the rule’s express intent and will have adverse unintended consequences to boot. God knows what lurks in the details.
The first “say what” in the sweeping rules results from the new “higher-of” construct. Credit and operational -risk models are entirely gone and market-risk models are largely eviscerated. Instead, big banks must hold the higher of the old, “general” standardized approach (SA) or the new, “expanded” SA. Each of these requirements is set by the agencies – models mostly never allowed. Further, a new “output floor” – different from Basel’s approach to this model’s constraint – is also mandated as yet another safety net preventing anyone gaining any advantage from any possible regulatory-capital arbitrage.
Why then not just demand that big banks use a standardized weighting the agencies think suffices? Must banks be put through the burden of calculating two ratios when they have no ability to arbitrage requisite capital weights? Do the agencies not even trust themselves to set capital standards that are now sometimes higher, sometimes lower as God gives them to know probability of default and loss given default after ten years working with the current rules? After all, even if these core assumptions are too lenient and the output floor crumbles, there’s always the stress capital buffer to back them up – now a bit restructured to make it another, still more formidable safety net preventing the standardized risk weightings the agencies – not banks – set.
Do the agencies so mistrust their ability to set standardized models and output floors as well as the stress capital buffer? If so, one must despair of the new rule ever making banking any safer even as it surely makes the financial system still more shadowy.
And what of supervision? The agencies reserve their right to add more capital if they don’t like the look of a bank’s risk. Do they also mistrust themselves here? Perhaps with reason, but then the onus is on them, not the banks.
And, it’s clear the new, double-barreled approach has still other unintended consequences. We’re only ankle deep in the agencies’ Big Muddy, but we have read the residential-mortgage finance section with care. As our in-depth analysis detailed, it’s simply impossible to know how much regulatory capital must be held against a low-risk mortgage because of the interplay of the old and new SAs with the proposal’s decision to end capital recognition of mortgage insurance. Does it still count under the old SA despite the risks the agencies now think MIs run because that requires a higher capital weighting even if the loan itself is low risk and MI thus adds only cost? Who knows, but it seems likely that banks will continue not to make these loans as long as the capital costs relative to risk and return prove prohibitive.
The reason the agencies want big banks to hold unnecessary capital makes even less sense than the underlying higher-of construct. The rule’s rationale is to ensure that big banks have no edge over smaller banks when the agencies’ own SA setting is lower because due deliberation has now concluded that risk is lower than the current SA requires. This foundational rationale lends credence to big-bank accusations that the new rules are founded more on vengeful political ideology than solid analytics.
If a current risk weight is too high, then why not lower it for all banks rather than come up with a complex system of multitudinous capital charges designed to do the worst for big banks only because doing better for them – not to mention for an analytically-credible rule – would require revising the rules for small banks? Can’t the regulators do the proverbial walk and still chew gum?
Apparently not.