Bank Supervision: Bye Bye Basel?

“Less is more”: under this motto, at the recent banking symposium in Jackson Hole[1], Andrew Haldane has openly advocated to dismantle the regulatory skyscraper of Basel 3 (the international accord on bank supervision) to replace it with a much narrower set of simple rules. This caused uproar in the financial industry[2], as Haldane is executive director for banking stability at the Bank of England: on the face of it, Heathrow’s watch-tower is proposing to replace jumbo jets with gliders.

Haldane’s argument is brilliant, compelling, even seductive: complexity spawns disasters, simplicity can prevent them. A dog needs no complex algorithms to catch a frisbee; conversely, safety systems at nuclear plants are vulnerable because they are too complicated. Also, finance is not only about risk (we know that the raffle box holds 90 numbers, but we ignore what the next draw will be), but also about uncertainty (we might as well be in the US, where bingo is played with 75 numbers), hence intricate probabilistic models may prove pointlessly expensive . Finally, credit rating models and market risk systems, used by banks to comply with Basel rules, are calibrated on past data and accordingly cannot provide a reliable forecast of the future. Therefore, supervisors would better replace them with simpler tools and look at a straight ratio of capital to assets (the so-called “plain leverage ratio”).

Haldane is right to warn against the hypertrophy of an accord which, born in 1988 with thirty pages, has reached 600 upon its latest revision. But he may stumble on a few logical slips, so his iconoclast vein should be received with caution.

Take nuclear plants first. Complexity in safety systems is the consequence, not the cause, of disaster risk. The danger originates from splitting atoms, not from using computers and sensors to check the temperature of reactors; people could otherwise cut risk by replacing control devices with a Maremma sheepdog. Financial intermediation works the same way: a mighty, dangerous lever. It turns huge passive masses into energy, but radiations can quickly pollute a whole continent. Accordingly, those who want to go back to a less complex supervision should first defuse (and scale back) the business of supervised entities, facing the costs that would follow from lesser scale economies and a possible drop in market liquidity. This amounts to leaving nuclear for coal, which may produce less energy and pollute more, but would not spawn new Fukushimas or Northern Rocks. One cannot, however, cut supervisory firepower unless banks and markets undergo a parallel disarmament. It would be unwise, e.g., to leave OTC trades intact while repealing the rules that require banks to report on potential exposures and credit risk.

Let us now turn to the “plain leverage ratio” (not so plain, actually, given that there is still a number of uncertainties on how to account for off-balance sheet items). Switching to un-weighted accounting data means that a possibly flawed set of risk-weights (the ones in Basel) would be replaced by a definitely flawed set of coefficients, implying that risk is the constant across portfolios and counterparties. “Plain” ratios may hence prove dangerously distracting[3], unless of course you think that withdrawing funds from the central bank to lend them to junk borrowers does not alter a bank’s risk profile. Haldane provides some compelling statistical evidence supporting the plain leverage ratio, showing that it would have pointed at many shaky institutions ahead of the 2007 financial crisis. But this reminds of his critique to bank models, which rely on past data to forecast the future (driving towards the next crash by looking in the rear-view mirror).

As for the difference between risk and uncertainty, it is already embedded in best regulatory practices, as most supervisors now ask banks to stress-test for the effects of highly unlikely events, not just to estimate the losses generated by the “most reasonable” probability space. Costs attached to complex risk models, finally, become widely acceptable when the latter are used to improve and streamline bank management, not just as a tool to feed the supervisors’ statistical bulimia.

So why this call to “reset” Basel? The desire for a new, simpler, bullet-proof regulation may just a way to make up for the weaknesses of supervision, which were the true trigger behind the latest financial crisis (as shown by the fact that EU member states – having the same rules in place –  were very unevenly affected). Supervision is where the focus should be steadily kept, as national dwarves still struggle to harness systemic risk and multinational banking giants. On this account, the recent decision to assign new powers to the ECB for Eurozone banks[4] is not necessarily a step forward. On one hand, it weakens the European Banking Authority (EBA) and provides fresh support to Britain’s regulatory self-sufficiency. On the other hand, it ties the possibility to provide European funds to struggling banks to the takeoff delays of a new body, whose structure and competencies are still hard to imagine. It might take a few years before the dog catches its first frisbee. In the meantime, it may be unwise to clean the regulatory blackboard to make room for a simple leverage ratio.

Basel is desperately in need of harsh criticism like Haldane’s, only a bit more circumstantial. The latest revision of the accord (2010) has led to some over-conservative rules, bound to be equally disregarded by decent banks and rascals. With 200 pages added to the 2004 version, nothing was cut by regulators (who did not look overly inclined to self-criticism). But wait: Basel also includes some strong calls to common sense, e.g. where it mandates that no dividends or bonuses be paid by weakly capitalized banks (throwing its weight behind a tendency that has proved very unwelcome to British banks). You need time and patience to prune it, but that is the only way ahead.


[1] Haldane A.G, Madouros V:, “The dog and the frisbee”, paper given at the Federal Reserve Bank of Kansas City’s 36th economic policy symposium,“The Changing Policy Landscape”, Jackson Hole, Wyoming, 31 August 2012 (http://www.kc.frb.org/publicat/sympos/2012/ah.pdf).

[2] Mallaby S. “Regulators should keep it simple”, Financial Times, September 4, 2012 (http://www.ft.com/intl/cms/s/0/697e5ad4-f5ec-11e1-a6c2-00144feabdc0.html#axzz25aKSPp66).

[3]D’Hulster K.“The Leverage Ratio – A New Binding Limit on Banks”, Crisis Response Policy Brief, 11, 2009, The World Bank  (http://siteresources.worldbank.org/EXTFINANCIALSECTOR/Resources/282884-1303327122200/Note11.pdf).

2 Responses to "Bank Supervision: Bye Bye Basel?"

  1. riverTrader   September 17, 2012 at 9:48 am

    Anyone who would say a dog needs no complex algorithm to catch a frisbee has obviously never worked in artificial intelligence.

  2. damiano guadalupi   September 26, 2012 at 2:47 am

    All people who criticise Basel II framework (first of all, Andrew Haldane) forget to (avoid to) mention that banks are allowed to use internal rating systems for regulatory purposes only after such systems have been validated by national supervisors. Tens of pages of the Basel Accord are dedicated to validation and to the role of supervisors in granting that models are reliable. Apparently the job performed by some European national supervisors is very poor if different PDs are assigned to the same rating grades by banks operating in the same jurisdiction.
    Before putting the blame on the complexity of the regulatory framework, we should question the ability of supervisors to challenge banks in developing reliable models to measure risks.