• Log in
• Register free
• Subscribe
• Help
• Tour & Trial IC Advantage

• Share Tips
• Trading
• Companies
• Markets Data
• Columnists
  • Bear Bull
  • Chris Dillow
  • No Free Lunch
  • Simon Thompson
  • The Trader
  • Your Opinion
• Funds Data
• Covered Warrants
• Research Tools
• Investment guides
• Q and A
• Blog
• Video

No right way to regulate banks

Created:

26 October 2009

Written by:

Chris Dillow

It would be a great shame - though a high probability - if Mervyn King's call for a radical restructuring of the banking industry were to degenerate into a mere Bank-Treasury spat. In truth, Mr King raises a profound point about the nature of risk when he says that "The belief that appropriate regulation can ensure that speculative activities do not result in [bank] failures is a delusion."

The problem, says Mr King, is that it's impossible to know for sure what level of capital ratio would make a bank safe. Mr King doesn't say so explicitly, but the reason for this is that the risks confronting banks are unquantifiable. They are what Donald Rumsfeld famously called "unknown unknowns". The small probability of catastrophic defaults cannot be quantified precisely. Nor can the small probabilities of a collapse in liquidity - which is what caused our recent difficulties. Banks exist in the realm of Knightian uncertainty. Which means a regulator cannot say with any confidence to a bank: "you are taking too much risk". We just don't know how much is too much.

Of course, this problem could be solved by setting capital ratios very high indeed. But this carries enormous economic costs; it would choke off lending and economic growth.

Mr King's speech is, then, a hugely valuable - and I think correct - counterweight to the dominant (if now fragile!) ideology of our time, which says it is possible to quantify and manage everything precisely.

However, although Mr King has the diagnosis right, I'm not sure he has the remedy. Splitting banks into tightly regulated utilities on the one hand and "small enough to fail" speculative operations on the other carries its own problems.

1. Can we draw a bright line between "utility" and "casino" banking? Martin Wolf thinks not. I sympathize. For example, it's not obvious that a plain vanilla loan which sits on a utility bank's books is necessarily safer than securitized loans bundled together by a "casino." As Alasdair Milne points out, the AAA tranches of mortgage-backed securities really do carry tiny default risk. Their problem is that their liquidity risk varied.

2. Limiting the size of "casino banks" would in effect limit competition. Imagine a casino bank found a way of lending cheaply to firms, say by securitizing risk cleverly. It would then grow quickly. But once it hit the size ceiling, it couldn't do so any more. Having no incentive to expand, further, it would merely raise the cost of lending. Firms would then be starved of cheap finance, unless there was vigorous entry into the market.

3. Would the government really be tough enough to enforce consistently a "small enough to fail policy"? In good times, when confidence and liquidity are high, it might be willing to do so. But banks usually fail in bad times, not good ones. And wouldn't the government then be tempted to rescue banks, in order to preserve "confidence"? There is, I fear, a time inconsistency problem here. Saying "we'll let casino banks fail" is one thing. Actually doing it when the economy is fragile is quite another.

I say this not to glibly dismiss the idea. Instead, I do so to raise a disturbing question. Could it be that there is no perfect way of regulating banks? Fundamentally, we have a trade-off. Either we make banking very low-risk indeed, in which case we choke off lots of economic growth and entrepreneurship. Or we risk banks failing and getting tax-payer bail-outs, but at least enjoy some lending growth in between the busts. Mightn't this trade-off be more acute than we think?