But it has earned enough spurs to merit retention in the reassessment of the macro-economic framework which is now under way in the Treasury. And Labour might even think of retaining it in the next Parliament.
An inflation target assumes that, in the long run, macro-economic policy cannot influence output and employment, so policy might as well be aimed solely at inflation. But even if true, the real question is whether this applies in the shorter run as random shocks buffet the economy. This depends on what sort of shocks are the most likely - demand shocks are friendly to inflation targets, supply shocks are not.
The Treasury believes that inflation is largely driven by demand - in fact, determined by the gap between output and its trend level. Once inflation is in the middle of its target band (say 2 per cent), it can be kept there indefinitely if output remains anchored to the trend growth path. So the problem becomes one of stabilising output, with inflation then looking after itself, with the usual lags.
Paradoxically, therefore, the government is committed to an inflation target, but ends up trying to stabilise output, since this is the means by which inflation is controlled. Now consider what happens if an unexpected demand shock hits the system - for example an sudden improvement in consumer confidence. In the first instance, output rises above trend, and a year or two later this will increase the inflation rate above the mid-point of the target band.
The appropriate response is therefore to tighten monetary policy as soon as possible so that output can be rapidly brought back down to its trend level, thus minimising the deviation in inflation above its target. There is therefore no conflict between growth and inflation objectives - both point to the same policy response.
Now consider the very different case of an adverse supply shock hitting the system - for example, a temporary increase in world commodity prices, causing a reduction in the level of UK output which is consistent with stable inflation. The government now faces a conflict between inflation and output objectives. If it seeks to keep inflation stable, it needs to accept that output must fall for as long as the adverse supply shock is in place. Alternatively, if it seeks to avoid a fall in output, then it must accept that inflation will rise. (Incidentally, most economists think inflation will rise permanently if the government refuses to accept the output loss.)
Operating an inflation target when the economy is hit by a supply shock means the government must accept a temporary loss of output to keep inflation close to the centre of the target range. This might generally seem sensible, but it is not axiomatic that the output cost is worth paying for the inflation gain - this depends on the magnitudes involved, and the relative weights the government chooses to apply to output and inflation as policy objectives.
Depending on these magnitudes and weights, it is possible to argue that some other target, such as a nominal GDP target, might be preferable to an inflation target in circumstances where supply shocks hit the system. This would implicitly give some weight to inflation, and some to output, in the government's policy response.
In the real world, however, demand shocks probably dominate supply shocks. Furthermore, if a major supply shock should come along, it would be possible to suspend temporarily the inflation target, reinstating it as soon as the initial adverse effects on inflation have been absorbed. No single target can be optimal in the face of all circumstances, but this sort of approach makes an inflation target quite tolerable, and possibly as good as, or superior to, any simple alternative.
The next question is what the inflation target should actually be. The present formulation is "the bottom half of a 1-4 per cent band", which of course implies, in effect, a band of 1-2.5 per cent, with a central objective of 1.75 per cent. This is far too narrow a band to be an operative target for any length of time.
The accompanying graph shows the extent of inflation fluctuations within economic cycles since the late 1950s. It is obvious that even the full 3 per cent width of the present 1-4 per cent range is not sufficient to cover the fluctuations in inflation that have occurred in the past.
Although this comparison may be distorted by the fact that governments were not actively trying to stabilise inflation in past cycles, a band width of four percentage points seems the minimum required. With a central objective of 2 per cent inflation (the same as the Bundesbank's target, and enough to allow some flexibility in the relative prices of goods and services without forcing some individual prices actually to fall), this would imply a range of 0-4 per cent.
But why have a range at all, instead of just a central objective? The existence of a 4 per cent ceiling for the range suggests that there is something particularly significant about this level, and that the government's response to an inflation rate of 4.1 per cent would be qualitatively different from their response to 3.9 per cent. But what actually happens is that the government becomes increasingly dissatisfied as inflation rises above the 2 per cent central objective, with no step change in the policy response as the 4 per cent barrier is passed.
This suggests that the existence of a range is arbitrary, and that no particular significance should be attached to the boundaries themselves. However, a breach of the target band, even if temporary, would be very costly to the credibility of policy. So the existence of band limits could backfire as well as mislead.
One way round this is to abandon the concept of a range, and simply to announce a permanent inflation objective of 2 per cent - which is what Germany and France both do. This would have the advantage that policy credibility would not become artificially attached to arbitrary band limits.
Finally, there is the issue of how the target should be enforced. At present, the authorities produce their best estimate of inflation pressures two years ahead, and change policy accordingly. Recently, there have been suggestions that the policy change should be linked to actual inflation, instead of to future inflation, and should be made automatic - for example, a quarter point increase in base rates for every 1 per cent increase in inflation above the central target.
This type of automaticity, however, has two fatal drawbacks. First, because of the time lags between monetary policy changes and their effect on inflation, a rule of this sort would always be two years behind the game. Second, although such a rule may appear to work on some econometric models, this seems to be a feature of the particular model chosen, and has disastrous effects when tried in other models. No Chancellor should (or would) ever risk experimenting in this way with the real economy.
A 2 per cent inflation target, with no target bands, and no change in the enforcement mechanism, seems the best way forward.Reuse content