The Independent Banking Commission has proposed that conglomerate banks in the UK should ring fence their retail banking operations. The objective is to reduce the subsidy to large investment banks and the related risk exposures to taxpayers which arise from the belief that such banks are ‘too big to fail’
The effect of the Commission’s proposals is that the taxpayer would still be expected to protect depositors, and perhaps other bank creditors, from the consequences of banks’ bad lending, but would be spared the losses from the banks’ inept gambling. This would be reassuring for those who bailed out RBS and UBS, through little consolation for those who bailed out HBOS and Anglo-Irish Bank. The issue is how to make the ring fence effective.
What activities are appropriate to a retail bank? . A retail bank takes deposits from its customers and lends to businesses and households, and maintains liquidity to meet withdrawals. In short, it performs the traditional functions of a bank.
But even the most boring bank needs a treasury operation to manage day to day funding requirements. Still, there is a large difference between a treasury which services the work of retail bankers and a treasury which hopes to derive a stream of profits from speculative trading in global capital markets. The transition from the former model to the latter created UK banks whose balance sheets are many times larger than their loans or deposits.
As HSBC has pointed out in its submissions, international accounting standards already require a distinction between banking activities and trading activities. In principle this focuses on the right issue. But banks have demonstrated how flexible a distinction based on inference of motive can be: bad assets, which can be held at book value, are said to be intended to be held to maturity while profits from good assets are marked to market.
The core problem is that banks have no intention of abiding by the spirit, rather than the letter, of any regulatory rules. Indeed they have developed profitable business in regulatory arbitrage – selling instruments which avoid regulatory burdens by changing the form of the transaction but not the substance. Effective rules therefore have to be direct and simple.
A suitable requirement might be that a high proportion – 90% or more – of a retail bank’s assets be in residential mortgages, government stock or loans to non-financial businesses. The latter really means SMEs, since the treasury operations of many large corporations are effectively internal banks. Such a rule also helps solve a further difficulty: how to stop the ring fence being as permeable as, for example, a Chinese wall.
In the absence of restriction the retail bank could simply lend all its funds to the investment banking arm of the group; when the investment bank fails the retail bank then has no assets. Treating such intra group loans as third party exposures, as the Commission proposes, would block that stratagem; but strict limitation on acceptable retail bank assets is needed to prohibit more complex transactions with similar practical effect.
And limits on the retail bank’s dealings in derivatives are also necessary, since modern financial innovation allows almost any desired exposure to be written as a derivative contract. It is hard to see why a retail bank needs to trade any derivatives at all on its own account other than interest rate swaps.
These measures might be reinforced by a statutory duty on the directors of the retail bank to protect its deposits. The purpose is to use the threat of personal liability to force them to face the conflicts of interest inherent in a complex holding company structure. If it proved difficult to find directors willing to take on such a responsibility, that would tell us the separation was not working.
If the ring fencing of retail banking were effective, it would be a big step towards creating a British financial system more resilient to future crises. The sanguine reaction of banks and markets to the Vickers’ proposal suggests that they do not believe the proposed measures will really make much difference. It is up to the Commission to prove them wrong.