Print

There is a better way to stop bank failures

Tuesday 24 June 2008 10:06AM

Stuff happens. People make mistakes, parts go wrong, unexpected things crop up. When disaster ensues, politicians, journalists and savants blame the blunderers, the component manufacturers and the perversity of fate. But good systems, whether in engineering, in management or in regulation, are robust. When those that rely on infallibility break down, the fault lies with the structure – not the individuals, the parts or the gods.

Richard Bookstaber’s fine book on modern financial innovation, A Demon of Our Own Design, distinguishes loosely and tightly coupled systems. The postal service is a complex network, but individual failures are contained and have no consequences for the integrity of the whole. Aircraft are safe because every critical element has a back-up. But the indicators of reactor core temperature at Three Mile Island were designed to give readings only within the planned operating range; neither the control room nor the instrumentation, Bookstaber explains, was designed with emergency operation in mind.

It seems absurd to design a monitoring or regulatory system on the premise of normality, but that is what we routinely do. Our systems are tested on data from periods of stability and our inspections confirm that everything functions as we intended. Northern Rock did not fail because the Financial Services Authority did not maintain adequate minutes, or because the tripartite authorities did not keep each other informed. It failed because the control room had not been designed with emergency operation in mind. The problems could, of course, have been defused at any time by the public authorities writing out a sufficiently large cheque – as they eventually did for Northern Rock, and as the Federal Reserve did much more promptly for Bear Stearns. Many social and economic problems can be ameliorated by the prompt disbursement of $30bn, although it is difficult to see why protecting the counterparty exposures of large financial institutions comes near the top of the list of priorities.

In a column last September, I made three proposals that would have led to the better handling of Northern Rock. Adequate deposit protection would have given an immediate assurance that no small saver would lose money, and ended the run. A special administration scheme would have given adequate power to deal with a failing bank. The government plans to implement both these measures. It should also act on a third proposal: that the Financial Services Compensation scheme becomes a preferred creditor in any liquidation.

Banks traditionally borrowed short from the public and lent long to business. The growth of innovative intermediation severed the link between borrower and lender. The outcome has been to make retail depositors less safe, not more, as their deposits have become collateral for proprietary trading. In an ideal world, retail deposits would be backed by genuinely first-class and liquid assets. Competition in the retail market would then be on the basis of relative effectiveness in the collection and processing of deposits. It is ironic that Northern Rock’s efficient administration is being wound down owing to losses in different activities.

To impose that ideal through regulation would be intrusive – and also unnecessary. The preferential status of retail deposits would ensure that if deposits were not supported by government or similar quality obligations, wholesale lenders would be at risk for any shortfall. Such a structure is probably the only realistic regulatory intervention that would have prevented the Northern Rock fiasco. If both securitised mortgages and the retail deposit base had been ring-fenced, the residual balance sheet would have been insufficient to support the bank’s risky expansion.

This proposal amounts to the privatisation of banking regulation. If public agencies are to supervise seriously the strategies of high street and investment banks, we might as well nationalise them; the proposal is entertained only because everyone knows it is not really serious. Such supervision will continue to be an exercise in box-ticking, analogous to confirming that today the reactor temperature was within the normal range, and equally helpless when, one day, it is not.

 Print