Why the Bank should opt for policy overkill

Monetary policy is a matter of assessing the balance of risks. In an uncertain world, is it better, at any given time, to err on the side of policy overkill, or policy underkill? This column will argue that, in present circumstances, the incentives on the Bank are clearly in the direction of risking overkill. For, if it makes the opposite mistake, the mopping-up operation could last the entire Parliament.

These results follow from a standard model which links the degree of spare capacity in the economy - the output gap - with the change in the rate of inflation. This was first estimated by David Walton of Goldman Sachs a few years ago, and while it is extremely simple, it captures the main features of the inflation process, and has operated with a tolerable degree of accuracy for some years.

Essentially, the model suggests that for every 1 percentage point of output gap, the inflation rate will change by 0.15 percentage points per quarter. The model is symmetrical between positive and negative output gaps, and interestingly finds almost no independent role for the exchange rate to influence inflation over the long term.

The elasticity between the output gap and inflation may not appear to be very large, but of course inflation continues to rise indefinitely for as long as output is above trend, so the cumulative effects of a policy mistake on inflation become very sizeable and persistent over time.

For example, if output is 2 per cent above trend for two years, then inflation will have risen by 2.4 percentage points from its starting rate by the end of the period. In order subsequently to get inflation back down by the same amount, output needs to be 2 per cent below its trend level for two years. When we allow for the fact that it takes some time, perhaps a couple of years, for output to shift from 2 per cent above trend to 2 per cent below trend, it becomes plain why it takes so long to regain control of inflation once it has been let out of the bag.

Fortunately, inflation has not yet been let out of the bag, so it is still possible that we can avoid all of these unpleasant consequences altogether. But time is getting short, and the likelihood of being lucky enough to judge the stance of policy just right is undoubtedly diminishing with every month that passes.

After last week's strong GDP statistics, Goldman Sachs reckons that the true level of output in mid-1997 - after making appropriate adjustments for the likely future upward revisions to GDP statistics - is already about 1 per cent above trend, and the positive output gap is rising all the time. The adverse impact of this situation on inflation is temporarily being disguised by the rise in the exchange rate, but this is unlikely to persist indefinitely. Hence the urgent need to bring output growth back to its trend rate of 2.5 per cent or so.

The good news is that, on the consensus view of prospects for next year, this is exactly what will happen. The stance of fiscal policy is tightening sharply because of the Chancellor's tough public spending plans, and overall monetary conditions have similarly tightened with the rise in base rates and sterling.

With today's policy settings, output growth should slow to 2.3 per cent in 1998, the positive output gap will never become too large, and inflation will be curtailed at roughly 3 per cent in 1999. This is a very good outcome, but it is a knife-edge solution which can only occur if everything works out perfectly.

The graphs show what would happen if, instead, the economy does not respond exactly as expected to the Bank's tug on the policy reins in the next 18 months. The first graph shows what would happen in the case of "policy underkill" - a situation in which the Bank does not tighten policy enough, and in which GDP growth persists at about 4 per cent next year. By the end of 1998, the die is cast. Output is 3 per cent above trend, and inflation is rising by almost 0.5 per cent a quarter. From then on, all choices are unpleasant, and there is no happy exit from the mess by the end of the Parliament.

The graph shows two possible choices from then on. In scenario one, the Bank seeks to avoid a hard landing, and slowly squeezes GDP growth down to zero over a three-year period. The trouble with this scenario is that output remains above trend for some time, so inflation continues to rise until the end of the year 2000, by which time it has reached 7 per cent. This seems politically infeasible, since it would mean the total abandonment of the government's 2.5 per cent inflation target.

Accordingly, scenario two shows what would happen if the Bank opted to get inflation back on target by the end of the Parliament. In this case, the output gap would need to be corrected quickly, and a hard landing would be necessary, with GDP growth going negative in 1999, and remaining below 2 per cent up to 2002. Inflation would slowly come back on track, but only at the expense of what would feel like a long and deep recession.

Now let us turn to the second graph, which shows what would happen if the Bank makes a mistake in the direction of "policy overkill" in the near future. Instead of output getting out of hand next year, GDP growth averages 1.2 per cent in 1998, and briefly dips to zero by the end of the year. This is sufficient to eliminate the positive output gap, and indeed to leave output below trend by the end of next year. Inflation therefore falls to well within its target band from 1999 onwards, and GDP growth can be allowed to rebound to 3 per cent or more for the last three years of the Parliament.

There is absolutely no doubt that, from the point of view of the Bank's policy committee, the second graph is far more appealing than the first. In other words, policy overkill today produces a far preferable out-turn for the next five years than policy underkill. (It has to be admitted that the level of GDP, and of employment, is higher for most of the Parliament in the latter case; but this is more than offset by the high path for inflation, which is disastrous for the newly independent Bank.)

The Bank's incentive is therefore to lean towards overkill - ie risking a situation in which the second graph, rather than the first, comes to pass. This probably means another 1 percentage point on base rates by the year end.

How could this conclusion be wrong? Only if we are significantly underestimating the impact of the Thatcherite policy reforms of the past 15 years, which might have increased potential output by more than we are assuming. If so, output growth can be higher for longer than our model is permitting, without any rise in inflation. This is indeed possible, and it is a risk in the opposite direction, pointing to the advisability of policy underkill this year. But it would seem unwise to bet the life of the Government, and of the independent Bank, on such uncertain foundations.