Why Banks Are Dangerous

To people in the west, questioning banks may seem heretical. Yet, if they believe their banks are robust, it is only because memories are short. They forget the developing country debt crisis of the 1980s, the US savings and loans debacle, the property lending crises of the late 1990 and early 1990s in the UK and US, the Scandinavian banking crisis and the Japanese bubble. The list of recent disasters is breathtaking.

Now, again, financial rescues are being orchestrated by the International Monetary Fund on behalf of the Group of Seven leading industrial countries. The motivation for these packages has as much to do with saving G7 financial institutions (particularly Japanese banks) as helping borrowers. Moreover, if western banks are, for the moment, in passable shape, the same cannot be said of those in Asia. In many of these countries the banking system may now be insolvent.

Why are banks so fragile? The answer is that their underlying structure is not just susceptible to disaster, but a cause of it. Governments respond to crises by assuming much of the risk, thereby encouraging banks to behave still more dangerously.

Commercial banks support a mountain of risk on a slender capital base. The bulk of their liabilities is redeemable at par and on demand, with depositors regarding their money as perfectly safe. Yet bank assets are subject to market risk (changes in asset values), credit risk (insolvency of debtors) and settlement risk (breakdown in the payment chain).

With international lending, there is foreign exchange risk (changes in exchange rates) and transfer risk (inability to obtain foreign exchange).

Banks are also susceptible to political risk and risk of fraud. They are honeypots for governments promoting favoured investments and for businesses trying to do the same. People rob banks, because thats where the money is. Better connected people want to control them for much the same reason.

Banks also exacerbate macro economic instability. In good times, credit expansion raises asset prices and enhances economic buoyancy. As property prices are bid up, the quality of collateral looks excellent and it seems safe to lend yet more. In a downturn, the engine goes into reverse: the value of collateral collapses, bad debt increases and credit contracts. A bank failure threatens the banks that have lent to it. Customers may flee to safety. At worst, many argue, much of the financial system disappears, the money supply collapses and the economy slumps.

Over time policy innovations were introduced to deal with the various threats. Central banks provided liquidity by becoming lenders of last resort; deposits were insured; and monetary policy was aimed at economic stability. Unfortunately, the government support increases the incentive for banks to take risks. An inevitable consequence was greater regulation.

Unfortunately, these remedies are made considerably less effective with international lending. A central bank cannot be a lender of last resort in a foreign currency. Neither can its government insure foreign currency deposits. Moreover, as banks lend abroad, the job of the regulator becomes still more difficult, precisely as its charges assume significant new risks.

The east Asian crisis has displayed virtually all the possible calamities. But if banks are so dangerous, what can we do to make the financial system safer?

The days when banks were the only way to make payments or provide liquidity in a form more useful than cash are gone. Other payment systems exist. An everwider range of financial services can be supplied electronically. As for their once valuable job of collecting peoples savings for investment, that function can be, and increasingly is, performed by securities markets.

So how could we replace banks, or at least avoid having to underpin them with explicit and implicit government guarantees? One radical suggestion is to separate narrow from broad banks. A narrow bank would supply liquid assets and manage the payment system. It would invest in safe short-term securities, such as treasury bills. It would earn some money from its assets, but would probably have to charge its customers for services they have hitherto received free. The liabilities of narrow banks would be safe. Governments could even guarantee them, but would not need to do so.

Then there would be broad banks. These would look much like they do today, with one crucial difference: they would operate without government support. Ideally, there would be no lender of last resort.

This distinction would cut the Gordian knot of bank subsidisation and regulation. But, like many neat and attractive solutions, it leads to some tricky problems. In particular, broad banks could readily recreate the instability we know today.

Most of the time, broad banks could offer customers more attractive terms than narrow banks, because they would be more profitable. When the economy was buoyant people would put more of their money into broad banks, which could then create more credit. When (and if) panic came broad banks would suffer, and create, the disasters known form the era before modern central banking stabilised the financial system.

But is it credible that a government might watch the disappearance of much of its banking system? Not very. The less credible this is, the more likely broad banks are to behave in ways requiring official assistance. Broad banks could then end up as cossetted (and controlled) as todays commercial banks.

The more probable government intervention is, the more important it is to insist on the soundness of broad banks. This brings back the need for regulation. This would be much the same as that required to strengthen todays commercial banks. If the pure narrow bank solution is little help, what are the other possible approaches? At least five further possibilities can be envisaged:

Replacing explicit or implicit government deposit insurance with cross-guarantees, under which banks insure one another.

Demanding greater transparency, along with market-based regulation.

Imposing higher ratios of equity and subordinated debt to overall liabilities.

Matching the maturity of bank liabilities to that of assets.

Matching the value of liabilities to that of assets.

The first of these proposals has been advanced by Bert Ely, a banking consultant in Washington. Its big attraction is that banks would only guarantee institutions they believed were sound.

A question, however, is whether banks would be able to meet their liabilities in a panic and, if not, whether government would be able to let them fall. Sound banks might also be brought down by efforts to rescue failing ones. The proposal would itself create some moral hazard and could further cartelise commercial banking.

The second idea is the market based route to regulation being followed by the Reserve Bank of New Zealand. If depositors can obtain information now available to regulators, they can do the monitoring. But if governments were not prepared to adopt a hands-off policy, market-based regulation would be ineffective. It might also be ineffective if depositors tried to protect themselves by diversifying their deposits, having failed to perceive the interconnections among banks.

Under the third proposal, banks would be closed before becoming insolvent. George Benston of Emory University has suggested higher capital requirements, valuation of capital at market prices and structured early intervention and resolution. By this he means scrupulous monitoring of banks with capital (equity, plus uninsured subordinated debt) of less than 10 per cent of assets and mandatory recapitalisation, or liquidation, of banks with capital below 3 per cent of assets.

The fourth proposal attempts to combine the virtues of narrow and broad banks within the same institution. Under the fifth proposal, bank deposits would no longer be valued at par, but might reflect the value of underlying securities. Higher returns would be obtained on longer-term deposits, at the price of risk to the value of these deposits.

These are not mutually exclusive solutions. The starting point, however, would be recognition that the delivery of vital day-to-day economic functions in institutions that transform long-term assets into cash is unnecessary and risky.

Customers may want safety, convenience and high returns. But they can have this only at the price of waste and economic instability, as risks are shifted to taxpayers and credit generates economic fluctuations.

Narrow banking, to clear payments and provide liquidity, along with broad banking, to provide higher returns to depositors and finance longer-term investment, seems attractive.

But these broad banks would need tough rules on capital, marking of assets to market and transparency. Alternatively, the risk inherent in the broad banks could be eliminated by making them more like mutual funds, matching returns and maturities for both assets and liabilities.

East Asians, note. The days of banks that offered everything to everyone should end. The price they impose is not worth paying.

Martin Wolf argues that rules governing commercial banks need radical change to strengthen the global financial system