Separating risky casino banks from High Street lenders is a safety-first approach. But a more robust sector may mean weak lending for many years. What chance an economic revival?

Policymakers trying to grapple with the aftermath of the financial crisis have to resolve a tension that puts them in an acute quandary. They need to take action that helps to lift the economy out of the mire but also try to make the banking system more robust.

These ideas frequently pull in opposite directions. The aim of aiding the recovery argues in favour of boosting bank lending—either keeping, or even raising, the ratio of bank lending to bank capital. Yet the goal of more robustness calls for banks doing less lending per unit of capital. How are these ideas squared? And if they're not reconciled, how should they be prioritised?

The pat answer is that banks have to carry more capital. It isn't a matter of them doing less lending but rather of them carrying more capital per unit of lending. In practice, though, cash is in short supply. It is still not that easy to secure capital on acceptable terms from the private sector, and governments are not exactly anxious to dish out more money after the squillions that they have already "invested". Meanwhile, the banks have regulators, the press, the public and governments breathing down their necks. The result is likely to be weak lending, probably for years to come.

Making the system stronger involves much more than just capital holdings. On the agenda is a radical, structural overhaul of the industry. One suggestion is to separate "the utility from the casino"—or keep the provision of basic money transfer and savings deposits in separate institutions. Tight controls would be placed on what assets they could hold and the type of business they could do.

These bodies would be supported by the state, and riskier financial activities left outside this net. The idea is that institutions in this category could be left to their own devices and may be allowed to fail, without the hope of government support. In this way, the complex regulatory infrastructure, including the capital requirements that govern the banking system as a whole, could supposedly be dismantled. This idea of "narrow banking" is both intellectually appealing and seductive. But I doubt whether it will work well in practice.

Drawing a distinction between the activities within and outside the net could prove troublesome. Would governments forbid institutions outside it from carrying out basic banking activities? If not, then surely such bodies would come to provide apparently more attractive products and services than the slower-moving, "safer" insider banks.

This would lead to a large shadow banking system outside the protected state sector. The state would not be able to walk away from such institutions if they failed. In practice, I doubt there is a neat answer. It is probably a good idea to separate the utility from the casino, but I have doubts that even once this is done, governments will be able to keep out of regulating the casino as well.

Roger Bootle is managing director of Capital Economics and economic adviser to Deloitte His new book, The Trouble with Markets (Nicholas Brealey) is out now.