Historically, the link between the state and the banking system has been umbilical. Starting with the first Italian banking houses in the 13th century, banks were financiers of the sovereign. Sovereign need was often greatest following war.
That is just an FYI. As you know, I think that the link between money and militarism goes back even farther. So does Niall Ferguson. There are other historical views, and mine may not be in the mainstream, but neither is it uninformed.
Anyway, the main point of the paper (thanks to Peter Boone and Simon Johnson for the pointer) is to emphasize the time-inconsistency problem in banking regulation. Key quote:
Ex-ante, [bank regulators] may well say "never again" [to bailouts]. But the ex-post costs of crisis mean such a statement lacks credibility. Knowing this, the rational response by market participants is to double their bets. This adds to the cost of future crises. And the larger these costs, the lower the credibility of "never again" announcements. This is a doom loop.
As I have said before, the challenge is to create a credible way to allow financial firms to fail. I have suggested creating a crisis response plan and organization that can keep the ATM's working and checks clearing in the event of a bank failure. Make sure that you can swoop in and keep the system running even while bankruptcy proceedings are taking place. Once again, I say we should try to make the system easy to fix rather than hard to break. Instead, current proposals are an attempt to make large banks "too regulated to fail." There is no such thing.
More excerpts from the paper are below.
As in the Middle Ages, perceived risks from lending to the state are larger than to some corporations. The price of default insurance is higher for some G7 governments than for McDonalds or the Campbell Soup Company. Yet there is one key difference between the situation today and that in the Middle Ages. Then, the biggest risk to the banks was from the sovereign. Today, perhaps the biggest risk to the sovereign comes from the banks.
After reviewing the history of responses to banking crises, the authors sum up:
Taken together, this evidence paints a consistent picture: a progressive rise in banking risk and an accompanying widening and deepening of the state safety net. There is a ratchet. This ratchet is evidence of a policy time-consistency problem.
If protection of depositors is felt to be a public good, these losses instead risk being borne by the state, either in the form of equity injections from the government (capital insurance), payouts to retail depositors (deposit insurance) or liquidity support to wholesale funders (liquidity insurance). The gains risk being privatised and the losses socialised. Evidence suggests this is a repeated historical pattern.
The authors cite five strategies that banks use for taking advantage of the asymmetry in who bears the risk:
--more assets classified as "trading assets" (this appears to me to depend on some accidental features of accounting treatment)
--business line diversification. If banks were limited to narrow lines of business, then they would not be in each others' businesses, so that their risks would be less highly correlated with one another.
--high default assets, which are assets with good returns in normal times but high potential loss from default in bad times
--writing out-of-the-money options (which are a form of high default assets)
The solutions that the author advocate include:
--contingent capital, meaning debt that converts to equity in a crisis.
--reducing concentration in the banking sector. This is a particularly bad problem in Europe (and some of us suspect that the bailouts the U.S. undertook were driven in part by fears about Europe's banking system). Simon Johnson and I would add that a concentrated banking sector creates a political economy problem--it gives the banks more political leverage over their regulators.