Aims that help inflation policy land on target

It is considered dangerous nowadays to fly without an auto-pilot, whether you are an airline or a central bank. Not that the central banks ever really had a reliable auto-pilot, but they thought they did when monetary aggregates were in vogue in the 1970s and 1980s. These days the use of monetary targets has all but been abandoned as an automatic signalling mechanism for interest rate policy, but there is still a restless feeling that some form of fixed rule would be desirable to help frame monetary decisions.

In the recent past, there has been increasing interest in a rule suggested by John Taylor of Stanford University, which links interest rates in a mechanistic way to the inflation rate and the amount of spare capacity in the economy. This is a rule which we are destined to hear a lot more about. It is under intense scrutiny at the moment in both the Treasury and the Bank of England, and has been the subject of very favourable comment from Alan Blinder, outgoing vice chairman of the Federal Reserve in Washington. Although no central bank would admit to following the rule blindly, one of its main attractions lies in the fact that it appears to mimic the actual behaviour of the central banks with remarkable accuracy, so it can be used for forecasting purposes, as well as for guiding policy makers.

The rule explicitly acknowledges that central banks should have two separate objectives - the long-run control of price inflation, and the short-run stabilisation of output around its long-run trend. (Note that the second objective is only to reduce fluctuations in output and employment, not to change their average levels in the long run; the sole long-run objective relates to stable prices.)

These twin objectives may sound unfamiliar to British ears, since the current strategic objective given to the Bank of England by the Treasury involves an inflation target and nothing else. But no one in official circles would deny that there is, in practice, a trade-off between price stability and the stabilisation of output, and that both objectives should have some role in the setting of monetary policy.

This is probably why the Treasury has set a 1-4 per cent target range for inflation, as well as saying that in the long term the objective is to hold inflation to below 2.5 per cent. Although somewhat shrouded in the mists of deliberate obscurity, I take this combination to mean that inflation might be allowed to fluctuate around the 2.5 per cent central objective if this should be considered necessary to stabilise output and employment.

The danger with trying to follow a twin objective, however, is that it can seduce policy makers into accepting a rise in inflation pressure for far too long, on the grounds that they are trying to "stabilise" output when in fact output and employment are already well above their sustainable levels. The private sector knows that this temptation exists, so they build into their inflation expectations a permanent risk premium on the grounds that it may one day happen. This risk premium makes inflation harder to eliminate than it need otherwise be, even if the authorities always "behave themselves".

One way of overcoming this problem is to tie the hands of the central banks into a formal policy rule, such as the inflation target operated in the UK. But it is known that rigid rules of this sort are sub-optimal, because they do not allow any specific role for output stabilisation. The idea of the Taylor rule is that it specifies exactly how the central bank should mix the twin objectives of price stability and output stabilisation. By specifying the exact mix in advance, it avoids the risk that the central bank will be tempted by "special circumstances" to deviate from the straight and narrow.

To operate the Talyor rule, the authorities first decide on a "neutral" level for real short-term interest rates, possibly by looking at the average level which has been attained in previous economic cycles. For the UK, I reckon that the neutral real short rate is somewhere around 3.5 per cent. If inflation is at the target rate (2.5 per cent), and if output is at its trend level, then the authorities should set the real short rate at the neutral level of 3.5 per cent. Adding back the inflation rate, this suggests that the base rate under such circumstances should be about 6 per cent.

The rule then allows base rates to deviate from this level for two reasons. First, if inflation is above the 2.5 per cent target, base rates are increased by half the excess of inflation over its target. Second, if there is an output gap in the economy (ie output is below trend), base rates are reduced by half the extent of the output gap. Hence there is a simple trade-off between increasing interest rates if need be to hit the inflation target, while reducing them if necessary to stabilise output - and vice versa.

In practice, the application of any such rule will obviously lead to many complications, and no central bankers worth their salt (or their salaries) would ever dream of reducing the huge complexities of monetary policy to such a simple technique. Or that, anyway, is what they routinely say when asked about the Taylor rule. But what does their actual behaviour betray about the way they take decisions?

At Goldman Sachs, we have been seeing whether the Taylor rule is capable of tracking the actual interest rate paths set by central banks over the past 10 years. The evidence, as shown in the graphs, is very surprising - it implies that the key central banks have spent much of the past decade setting policy as if they have been following the Taylor rule. This applies as much to the UK as it does to other economies.

This rather startling observation has two clear implications. First, when it comes to forecasting central bank policy - which is the first crucial step towards understanding the behaviour of financial markets - it is very useful to see what the Taylor rule is implying about the future.

At present, as the graphs show, the rule implies that the trend in short- term rates in the Group of Seven countries will be gradually upwards in the next 12 to 18 months, but not by as much as the markets presently predict.

Second, it is possible to use the rule to gauge what UK policy setting should be in place right now. On this, with inflation slightly above target, but the output gap somewhat negative, it suggests that base rates should be almost exactly equal to their neutral level of 6 per cent - which happens also to be exactly where the Chancellor has put them. Furthermore, on the Goldman Sachs forecasts for inflation and GDP growth in the next 12 months, the Taylor rule reckons that the current level of rates will stay about right until after the election.

It would be a mistake to push such a simple mechanism too far. But it is comforting to note that, according to the Taylor rule, the stance of British monetary policy remains about right, even in a pre-election period. Let us hope it stays that way.