How the market can control the banks

Barack Obama has decided to side with a state solution, if not yet a well-thought-out one, to preventing bank failures bringing down the world economy. But there is a market alternative: fix the banks so the bondholders keep bank risk-taking under control – and bear the costs if they fail in that task.

This important debate is not being framed as a state vs market discussion, but it should be. Remember state control has a dismal record in general, and in the finance sector in particular. Regulators entirely missed the bubble, missed the banks’ reliance on short-term financing and missed the fact that so much regulatory arbitrage was going on. Don’t expect things to be much different in 20 years if a state solution is accepted.

The problem is the banks (and potentially non-banks) being too big to fail, creating perverse incentives to take risks and leaving the taxpayer paying for mistakes. The state solution accepts banks are too big, but tries to control that through regulations, restrictions on what they can do (no prop trading or hedge funds), and potentially a cap on size.

A working market solution would obviously be better, as it would be pretty much immune to the high likelihood of regulatory capture – but can a market solution be made to work? I think so.

There are two major problems with a free market in banks:

1. Political unwillingness to let large-ish banks fail, even if they have little or no knock-on effects (think Northern Rock, or for an example of pure politicised nonsense, Dunfermline Building Society).

2. The time taken by bankruptcy procedures, and the associated uncertainty among both bank counterparties and bondholders about the value of their positions. This leads to panic, hurting other banks and creating the potential for a big bank failure to bring down the entire financial system.

The second of these is being addressed through the creation of “living wills” and revamped systems to shut down banks quickly, if not overnight, resolving the uncertainty. I’m sceptical about the likelihood of new systems being able to close down a huge complex bank quickly, but even assuming the new bankruptcy procedures are quick, the collapse of a big bank could still prompt panicking investors to pull out of other banks if they did not expect it – and, of course, bank collapses are like the Spanish Inquisition: no one ever expects them.

What is needed is to convince investors that they always need to factor in the risk of a collapse, by regularly reminding them that their bonds are at risk of turning into the one financial instrument everyone knows is risky: equity. This can be done, using contingently convertible bonds – “CoCos” – which turn into shares when capital falls below a specified level. The legal documents backing CoCos explicitly lay out these risks, but that would not be enough. What is needed is the regular use of CoCos.

I would suggest that banks are forced – yes, I know, regulation – to raise all their debt through CoCos, and to have a ladder of capital ratios at which debt converts. If core capital ends up being fixed at, say, 10%, with a minimum of 5%, then the lowest tier of bonds would convert at 10%, the next at 9%, the next at 8% and so on down to the best-graded bonds, the equivalent of AAA-rated: bonds which convert at 5%, when the bank actually breaches it capital levels and is otherwise going to need a state rescue.

If a bank goes from 10% to 0% overnight – entirely possible – then all the bonds convert at once.

One advantage of this is that investors have to carefully consider which safety level of debt to buy; the safety, or lack of it, is far clearer than in the current grading of bonds and hybrids.

But the real beauty is that conversions of the riskiest debt should be reasonably frequent, as bank capital always fluctuates. This will serve to remind all bondholders that they could lose their money.

By letting investors lose money without system-wide panic, you also solve much of the political problem: if people were warned they could lose money, there is little political reason to rescue them. Complete political capture by Wall Street or the City of London cannot be ruled out, of course: but the fact that Lehman was allowed to fail, along with hundreds of small US banks, suggests politicians are not totally in the pocket of the bankers.

The outcome would be a higher cost of borrowing for the banks, differentiated according to which banks were seen to be taking the biggest risks. This would reduce the likelihood of loony lending decisions, as well as reducing overall borrowing by the banks. It would also be “fair”: if the bank fails, the people who invested in it and gave it the capital to make the bad decisions would be the ones who lose money. Finally, big banks would be able to fail just as easily as a small bank.

There are problems. In a super-boom, you might again have investors ignoring all the warnings and treating senior bank debt as though it were guaranteed, particularly for big banks. A surprise failure of a bank could still prompt a run on similarly-structured banks.

And, the biggest problem of all, banks could still raise short-term financing through the repo market and other off-balance sheet funding systems. The repo market cannot be banned without bringing the entire finance system to a halt, and it seems a sure bet that banks will move to raise more money through structures treated as “counterparties” if the cost of raising debt through bonds is hiked. The answer here should be more transparency, particularly on 364-day debt: it is insane that debt which does not carry over the quarter-end or year-end can be closed out and then repeated on January 1 without disclosure of the risks being taken by the bank.

But this is a problem which would have to be closely monitored, for possible future action against banks which chose to switch their funding to repo markets and derivatives counterparties. If it was needed, a drastic option would be fixed penalties in case of bank failure – automatic “haircuts” on assets in the repo market, for example, as part of the bank failure regime (this would only come in after all bonds had converted, of course, so we’re talking about a bank with 30-40% capital at this point!).

(Before attacking me, note two things. First, this is not an argument for a rampant free market: it is a debate about how best to control the banks, by regulation or by financial measures “set” by people with money at risk. Second, there are loads of other changes that are still needed: for example, charging proper fees for deposit insurance to reflect the residual risk of a bail-out for the deposits; applying competition policy far more aggressively to break up some of the monster banks which are distorting markets today; and some minimal regulation – such as a cap on size – of hedge funds and other sources of instability to which risk capital will migrate if the banks are properly brought under control.)

FT dot comment

FT dot comment is no longer updated but it remains open as an archive.

Politics, economics, high finance and morality – this blog addresses the issues being considered by the FT’s comment team, and their thoughts.

FT dot comment: a guide

Christopher Cook is an FT editorial writer. Before joining the FT in 2008 as a Peter Martin Fellow, he worked for three years for the Conservative party.

Lorien Kite is deputy comment editor, a post he took up in 2009 after four years as a commissioning editor on the analysis page. He joined the FT in 2000.

Ian Holdsworth became assistant features editor in 2009 and was previously chief production journalist for the features pages.