Earlier this week, the Wall Street Journal called a formidable paper by Andrew Haldane and Vasileios Madouros the “speech of the year.” Given my fears about what I call “complexity risk,” the speech’s central focus on simplicity struck quite a chord. But, the more I read, the more worrisome it seemed. Buttressed by analytics at least as complex as the rules they decry, the British regulators call for two planks on which financial regulation is to rest: a whole lot of capital and loads more supervisory judgment calls . The first criterion – capital – might make sense if banks failed solely due to it, which they don’t. Capital also only counts if it’s applied consistently, which the monumental cracks in global implementation of Basel III show clearly that it can’t. The second criterion – supervisory judgment – is nice, not to mention needed, but valuable only if supervisors can actually do something once they wish to pronounce sentence – which all too often they can’t.
Much in the Haldane-Madouros paper is premised on economic analytics, historical assessments of capital models and some of the formative texts in organization theory — the last an unusual and useful excursion for regulators. However, the paper takes a selective view, focusing principally on the vital work of Herbert Simon and not on theories derived from it. Most important of these in this regard are those that warn against “synoptic” solutions – that is, one-size-fits-all cures. Very high, very simple capital ratios are the Haldane-Madouros cure, but this works only if capital is a perfect cure – penicillin to the infection, to use a medical metaphor along lines frequently adopted in the U.K. paper. But, if the disease isn’t an infection and, instead, a more complex genetic syndrome, you can dope the patient up all you like, but little good will it do.
Capital counts, of course, but so do many other risk factors – liquidity, operational risk, governance lapses due to incentive misalignment and management or customer fraud to name a few major primroses on the path to destruction. Thus, a single cause like capital inadequacy is not in fact the proven cause of bank failure, let alone that of nonbanks priming the shadow system for the next round of systemic derring-do.
Engineering – the premises of which underlie much in organization theory – is based among other things on redundancy. Systems are built on systems and, then, on back-ups because engineers know that one system – say a really big jet engine that nominally can fly a very large plane – can suffer catastrophic failure even when every model says it can’t. So, the more we rely only on capital, the more we ignore other catastrophe drivers, building colossal engines meant to power giant banks, even though banks are still subject to violent weather, pilot error and other dangerous safety problems.
But, what if capital really were the cure? Even then, this paper’s recommendations falter because the ability of global regulators to do anything so simple as a stringent leverage rule is, at best, uncertain. Yesterday, the chair of the Basel Committee argued that the Basel III rules are simple and, thus, meet the Haldane standard, because they include a leverage standard in addition to all the revised risk weightings Basel III hopes will cure defects in Basel II. Basel III’s global standards do indeed include a three percent leverage rule, but it’s hard to find it anywhere other than in the codified global rules. Some nations are moving towards a leverage rule and some like the U.S. have long had one in place for banking organizations. The European Union is, though, a long way from leverage and may well never agree to a simple capital ratio even if set as low as Basel did in hopes of getting concurrence across the globe.
Supervisory discretion is just as tricky as capital when one considers it across borders. In some countries, bank supervisors have formidable powers, including the ability to shut down a bank, should they choose to use them. Mostly, as in the U.S., they don’t. Still, though, supervisors can do something about risk should they spot it. In other nations, supervisors have limited access to banks and can’t tell management what to do no matter how frightened they are. Elsewhere, supervisors aren’t frightened because they are in cahoots with banks or joined with them to do the government’s bidding, risk be damned.
In our 2011 complexity-risk paper, we argue for supervisory discretion, but only when it’s based on clear standards for boards of directors and senior management in critical areas like risk tolerances and only when discretion can be readily turned into discipline. Now, mostly, it can’t so – despite hope – judgment calls may well not matter.
So, what does? This is where what organization theory calls an incremental solution seems the best. In it, you take various findings – e.g., that banks fail for lots of reasons, not just one – and build decision trees based on them. You don’t look for a philosopher’s stone to turn risk into reward; instead, you undertake a holistic approach based on numerous simple indicators of risk that can be combined into a balanced, measurable remedy that, then, can be tested on clear goals to see if it works, refined, and tested again. The complexity-risk paper and, now, our review of Basel III’s travails and recommendations on ways to fix it are incremental – they lay out problems based on an assessment of the sum total cost of pending rules and, then, suggest a set of modest solutions that combine into a very rigorous regulatory framework for even the largest banks. Some of this might not work, but all of it can’t fail all at once – the risk of a simple capital cure that can never be meaningfully implemented backstopped by stringent supervision that can never happen.