Chancellor should not snatch at base rate cuts

At the monthly monetary meeting in January, the Chancellor made a comment which gives a valuable insight into the way his mind may be working on base rate policy. He pointed out to the Governor that he had acted promptly and pre-emptively to raise interest rates when inflation pressures were rising in 1994. This, he said, has earned him the right to act similarly decisively on the way down. With the Bank of England now predicting that inflation will be just below the 2.5 per cent target next year (see graph), the financial markets expect another base rate cut to be announced this week.

The authorities have given a lot of thought lately to the speed of adjustment of monetary policy in response to new economic information. There is a feeling in official circles that the UK, like other countries, has had a tendency to react too slowly to changes in the economic climate, and then to eke out interest rate changes for far too long. This pattern of "too little too late" is obviously less than ideal, and it has probably contributed to the instability of the economy in the recent past.

Why do central banks act like this? A superb new study of central bank behaviour by Charles Goodhart at the LSE provides some of the answers. He starts by making a fascinating point, which is quite hard to grasp at first, but becomes more telling the more you think about it. Central banks, he believes, are all trying to do the same thing, no matter how they choose to formulate the precise targets which officially guide them. They are basically trying to stabilise prices in the period ahead, which translated into the real world means holding the officially measured inflation rate to 2.5 per cent or less a year or more into the future. In other words, they are all actually doing exactly what the Bank of England now claims in public to be doing.

Policy at any given moment should therefore be set such that the forecast for inflation a year or so ahead is 2.5 per cent. If this is the case, then changes in policy should be triggered only by a shock which alters the inflation forecast over the relevant horizon. These shocks should be random if the system for forecasting inflation is efficient - in other words, shocks which raise inflation relative to the target should be just as likely as shocks in the opposite direction. And here comes the main point. If the shocks are random, then so too should be the policy response - that is, the pattern of base rate changes should itself be random. Therefore, when we come to examine the behaviour of the authorities over time, we should see a series of higgledy piggledy moves in interest rates which do not follow a neat orderly pattern. A good central bank, on this definition, might be one which raises rates by a point in one month, cuts them by a point and a half the next, and then raises them again a couple of months later.

But instead of this haphazard pattern, which would be optimal, we see precisely the reverse - a series of smallish changes coming at frequent intervals which have a phenomenal tendency to be all in the same direction. When there is a change in base rates in the UK, for example, it is four times as likely to be in the same direction as the last change as it is to be in the opposite direction. Furthermore, there is a tendency for these changes to be in reaction not to the future behaviour of inflation, but to its behaviour in the recent past. Central banks appear to be driving while looking in the rear view mirror. Their broad rule, according to Goodhart, is that they react to a 1 per cent rise in reported inflation by increasing interest rates by about 0.15 per cent per quarter for four or five successive quarters.

Goodhart suggests that central banks behave like this because they need to be able to point to a worsening in actual reported data in order to sustain public support for interest rate changes. Forecasts are not enough. And some people have argued that this backward looking behaviour might not be a bad thing - that monetary policy should indeed be based on published data for historic inflation or nominal income. Forecasting has a bad name, and even Eddie George has allowed himself to argue that it is inherently too uncertain for it to be central to the policy process. Therefore why not wait until published information is available, or (in a different formulation of a similar point) why not act only when explicit "lead indicators", such as monetary growth, say you should?

These arguments may seem beguiling, but on close inspection they become totally incomprehensible. The fact that forecasting is imperfect does not mean that it is entirely useless. There is a very large gap between a view of the future which is held with perfect certainty, and one which is held with no knowledge whatsoever. Clearly, the present state of economic forecasting comes somewhere in between - it is far from perfect, but it is much better than saying you know nothing at all about the future. Since everyone would agree that monetary policy affects inflation only with a lag of about 12-18 months, it cannot make sense to throw away whatever knowledge we can glean about the future when base rate decisions are made.

The present rate of inflation conveys almost no knowledge about the future, so it should be discarded as a policy tool without further thought. And growth in the monetary aggregates scarcely does any better. It does admittedly convey some information about the future, but many other variables can add to this information, so they should be used as well. To use historic inflation data to guide policy is to tie both the Chancellor's hands behind his back; to use the monetary aggregates is to tie one hand. Only by using all the available information in a coherent forecasting system does the Chancellor have both hands free.

Now back to the question of what the Chancellor should do next. On the argument just outlined, the authorities should be moving over time to act more decisively and earlier in response to changes in the economic climate than they have typically done in the past. This might indicate that base rates should drop by (say) another half point quite quickly. But unfortunately, as the Governor has been arguing, the authorities need to worry about building their credibility over time, as well as about the need to act quickly and decisively.

This means that, for a while at least, they should exhibit a clear tendency to move interest rates more quickly on the way up than on the way down. This is especially the case when the markets doubt whether the Government would be willing to reverse the trend again, should it prove necessary, just before the election. The Governor always looks like a spoilsport when he argues this, and he is often accused of being an "inflation nutter". But actually he is quite right - credibility matters. Like virginity, it is easier lost than regained.