Despite the apparent world economic upturn, excessively tight monetary targets could be dangerous if recession starts

The last International Monetary Fund World Economic Outlook contained a section entitled "Can inflation be too low?". The suggested answer was "Yes". Coming from an organisation regularly attacked as a pillar of western capitalist orthodoxy, the conclusion is pretty remarkable. It is also highly topical when so many central banks are pondering when to raise interest rates.

A serious difficulty about zero inflation is that it puts a floor under how far the real rate of interest - that is, the rate corrected for inflation - can fall. In the recession of the mid-1970s, triggered by the first oil price explosion, the world experienced double-digit inflation; and in many countries the real rate of interest was negative. This could well be one reason why the world did not experience more than a pause in the upward movement of output.

If, however, inflation is absent, nominal interest rates cannot fall below zero, where they have more or less been in Japan for several years. Moreover, most personal and many corporate borrowers have to pay a couple of percentage points above official interest rates. So if a near-zero base rate or prime rate is not enough to boost consumption or stimulate investment sufficiently in a recession, the economy will not easily recover.

That is why the IMF warns that "deflation blunts the effectiveness of monetary policy . . . There is a danger that with real interest rates prevented from falling, output recovers only slowly and deflation gathers force ... If the shocks hitting the economy are large enough, this dynamic interaction can lead to a downward spiral that cannot be tackled by short-term interest rate policy alone."

This situation resembles what John Maynard Keynes called a liquidity trap, although the mechanics are somewhat different. The IMF authors suggest pre-emptive measures to prevent depression from gathering force. One way is to "adjust the short-term nominal interest rate aggressively" - as has been done in the US and Britain - before a downward spiral can begin.

They would also like central banks to respond not merely to recession but also to deviations of output below its sustainable trend, as well as to deviations of inflation from its target. As a last resort, they cite unorthodox measures such as direct purchases of assets, including long-term bonds and foreign currency. The most fully elaborated of such proposals involves the Japanese government buying back its short-term bonds from the public and is described by Tim Congdon in the May issue of Central Banking. A still more unorthodox suggestion has been made by Brendan Brown in his book The Yo-Yo Yen (Palgrave) for a currency reform in which, say, 95 new yen are substituted for 100 old yen not spent by a certain date.

Although these ideas have emerged from a discussion of the prolonged Japanese recession, concern has extended much further. For instance, the January 29 minutes of the US Federal Reserve open market committee contain a cryptic paragraph saying: "Members discussed staff background analyses of the implications for the conduct of policy if the economy were to deteriorate substantially in a period when nominal short-term interest rates were already at very low levels. Under such conditions, while unconventional policy measures might be available, their efficacy was uncertain, and it might be impossible to ease monetary policy sufficiently through the usual interest rate process ... The members agreed that the potential for such an economic and policy scenario seemed highly remote, but could not be dismissed altogether." Like the IMF, the Fed favoured taking preemptive easing actions to prevent economic weakness and price declines from reaching a point where the zero interest limitation constrains policy.

It is in this context that the IMF argues in favour of a low but positive rate of inflation. The Fed would surely agree, although it prefers to operate pragmatically, without any explicit target for either output or inflation. The IMF authors argue that it is best "if inflation does not fall below 1-1½ per cent, implying target inflation rates above that level".

An immediate implication is that Gordon Brown, the British chancellor of the exchequer, was right to proclaim a positive 2½ per cent inflation target, with downward deviations being regarded as as bad as upward ones. It also follows that the European Central Bank's aim of "stable prices", which is defined as a measured inflation rate of between zero and 2 per cent, is too strict.

True enough in principle; but dwelling on this point only encourages those who erroneously attribute the unemployment problems of the eurozone to the ECB rather than to the malfunctioning of the area's labour markets. Indeed, it is far from clear that actual ECB policy - as distinct from proclaimed objectives - is more stringent than that of the Bank of England's monetary policy committee. And it makes a difference which index is used - on the harmonised European one, UK inflation is lower than the eurozone rate.

The problem of interest rate floors may seem remote if, as forecasters suggest, the world economy is moving towards a restoration of normal growth rates; but suppose they are wrong? Unfortunately, we cannot rule out the possibility of one or more large-scale terrorist attacks in the US or Europe. Nor can we dismiss the worries of those who believe that the US recovery cannot continue for long on the basis of negative saving and rising consumption. The American consumer has been buttressed by rising real wealth from home ownership, and until recently from equity holdings. Stocks may still be overvalued; and the property bubble could burst at any time. The mere possibility of these events makes one need to look at economic defences against renewed recession, even if it is not the most likely short-term danger.