During each working day, thousands of large payments are made electronically between banks, on behalf of their customers, using the Chaps system. Some individual payments exceed pounds 100m. But although the customer may get the money, say, at mid-morning, the banks never settle with each other until late afternoon. They do so by netting off the total amounts they owe each other, so there is a single payment or receipt at each of the banks in the clearing system.
The result is that for hours every day, up to the moment the clearing takes place, banks build huge exposures to other banks. A bank's exposure may easily grow in a few hours to more than its capital. So if one bank collapses or for some other reason fails to pay at the end of the day, it may bring down others in its wake. Add in an international payments dimension and the risks escalate.
The unspoken assumption of the present system, with its large exposures to risk during the day, is that central banks will intervene if there is a failure. This could be expensive, but no central bank would stand idly by while its banking system seized up. From the central bankers' point of view, this amounts to a hidden subsidy, with the authorities bearing a risk but not being paid for it.
Under the new proposal for a Real Time Gross Settlement System, banks will clear every individual Chaps transaction between themselves, as it happens. They will do this through accounts at the Bank of England. There will be no build-up of exposure during the day.
This will cost some money for equipment, but much more for the extra funds the clearing banks will have to keep available at any one time to allow immediate settlement - the reason why the Bank has offered them a credit facility to increase the liquidity of the new system.
But those will be minor penalties compared with the benefits of reducing the risks at a time when some international banks have taken to living perilously close to the brink.
The collapse of Herstatt in Germany is still remembered because of the payment system risks. Given that Herstatt was nearly 20 years ago, it comes as a shock to hear from Mr Leigh-Pemberton that the issues raised by Herstatt have still not been satisfactorily addressed.