Politicians still think they can fine-tune the economy. This is folly

Setting policy is hampered not only by the need to rely on forecasts - having to steer using the rear-view mirror - but also by uncertain data

Some time during the course of today when the Bank of England conducts its money market operations it should become clear whether or not Kenneth Clarke has agreed to an increase in interest rates this month. If he has decided against it, one of the reasons is likely to have been the need to wait for the first estimate of gross domestic product in the final quarter of 1996, a figure due to be published at the end of the month.

If it shows the economy grew by as much as or more than in the third quarter, when GDP rose by 0.7 per cent, it could well trigger an increase in the cost of borrowing next month. For the economy's trend rate of growth is thought to be about 2.25-2.5 per cent a year, or about 0.6 per cent a quarter. Faster growth is likely to fuel inflationary pressure, requiring a rise in interest rates to take the froth off the economy. If the quarterly change in GDP is a fraction of a percentage point too high, we are likely to be paying more for our mortgages soon.

An essay in the latest issue of Economic Trends, one of the monthly publications of the Office for National Statistics, sheds an intriguing light on the use of economic statistics in policy decisions. The author, Henry Neuberger, explains how the national accounts were developed precisely for the purposes of policy-making. The first attempt to construct national accounts with policy in mind was made by economist JM Keynes in his paper "How to Pay for the War" in 1940, in which he tried to assess the economy's taxable capacity.

As the following year's Budget White Paper noted: "During 1940 the resources devoted to personal consumption and to the demands of central government and local authorities together exceed the resources available from the net national income."

Keynesian economists during the subsequent decades came to regard the national statistics as the tools that enabled government to read off the required levels of its tax and spending plans. One standard text of the 1960s said: "We should approach the economic system as an engineer approaches a complicated piece of machinery." However, the habit of fine-tuning policy generated by this approach was subsequently discredited among academic economists. The economy is just not that mechanistic. There are unexpected shocks, people's behaviour changes over time. And what's more, the statistics are sometimes wrong.

One of the most notorious cases was the under-recording of exports in the 1960s and 1970s. This error had a profound impact on economic policy, for it led a generation of economists to believe that the balance of payments was a serious constraint on British growth. The government could not allow too much expansion without running into the trade buffers with imports running too far ahead of exports. The high hopes for the management of the economy crumbled into despair because of the stop-go cycle that resulted.

Even though the error was uncovered during the 1970s, the balance of payments remained the bugbear of the 1974-79 Labour government. A balance of payments crisis turned Denis Healey back from Heathrow Airport to meet the International Monetary Fund 20 years ago. The IMF prescribed tough public spending cuts as a condition of the emergency loan, and the winter of discontent followed two years into the cuts.

But the monstrous balance of payments deficit that triggered the crisis was later revised away by the statisticians. Today's estimate of the 1975 deficit is pounds 1.5bn, or about 1.5 per cent of GDP. The scale of the deficit relative to the size of the economy was bigger during the early 1990s.

The late 1980s provide another example. There are three ways of measuring GDP: add up output, add up incomes or add up expenditure. They ought to be the same, but never are, and in the late Eighties the gap between the measures grew significantly. When the scale of the Lawson boom became clear, the Treasury blamed over-loose policy on the unreliability of the GDP measures. It had been impossible to tell how close the economy was to its capacity ceiling, according to an internal inquest into the episode.

The Treasury report in 1989 concluded that one problem had been reductions in spending on gathering statistics. Extra effort and resources put into collecting national accounts data since then mean that the size of revisions to GDP and balance of payments figures is dramatically less than before. As the chart shows, a gap between the three measures of GDP has reopened in recent quarters, but it is nothing like as big as it was in the late 1980s.

Even so, the fact that there are any revisions at all presents a difficulty in the current framework of policy, which involves making a judgement about the precise state of the economy month by month. The income measure of GDP fell in the third quarter of 1996, whereas the output measure jumped. The ONS focuses on the output measure as the most reliable short-term indicator but even so jiggles the published number, which does not add up to the sum of its components.

And the revisions, small as they are, point to different interest rate decisions. For example, when Mr Clarke decided to cut interest rates last June, GDP growth in the first quarter of 1996 was estimated to be 0.4 per cent. The latest figures put it at 0.6 per cent. When he increased base rates in December, the published third-quarter change in GDP was 0.8 per cent - now revised down a little to 0.7 per cent.

Martin Weale, director of the National Institute of Economic and Social Research, and chairman of a statistics users group, is researching the question of whether or not it would be better for the Chancellor and Governor of the Bank of England to meet quarterly rather than monthly. But he would also like the ONS to put health warnings on different categories of statistics. "If you knew there was a margin of uncertainty, you would not respond so much to the most recent data," he says.

That, at least, might be how a rational academic would react to knowing that setting policy is not only hampered by the need to rely on forecasts - having to steer using the rear-view mirror - but also by uncertain data - using the mirror to peer through a misted rear windscreen.

However, apart from the brief flirtation with pure monetarism in the early 1980s when the only thing that determined interest rates was how fast the (fairly accurately measurable) money supply was growing, policy makers have preferred to make policy as often as they can. With the current monetary arrangements, fine-tuning is back with a vengeance. Mr Clarke decides to move interest rates a quarter point because of a margin of 0.2-0.3 per cent in quarterly GDP growth to hit an inflation target two years hence.

The folly is not that statistics get revised. That is inevitable, and the UK's statisticians are better than most. It is the fact that politicians still think they can handle the economy with the precision of a mechanic following a blueprint.

Interest rates should go up this week. If they do not, they should go up next month instead. This is because most of the data over the past several months have pointed to growth well above trend. The GDP figure published between now and the next monetary meeting will not make any difference.