This problem resurfaced again last week, when the Central Statistical Office boldly announced that the recession had not been as bad as it thought and that the recovery has so far been better than it looked. The CSO cut its estimate of the fall in national output between its 1990 peak and its 1992 trough from 3.9 to 3.6 per cent. And it raised its estimate of growth in the year to the second quarter from 3.3 to 3.7 per cent, the strongest rate since the days of the Lawson boom.
These revelations are unlikely in themselves to conjure up the fabled 'feel-good' factor for which Conservative backbenchers are waiting so anxiously. Despite the rapid pace of economic growth, the impact on voters' pay packets has been limited. Big tax increases, meagre pay rises and weak job creation meant that the economy's total wage and salary bill rose by less in the second quarter of this year than in any equivalent period since the spring of 1967.
But the figures still suggest the economy is growing unexpectedly strongly and may well be eating into industry's spare capacity too quickly for comfort. This means that Kenneth Clarke and Eddie George may have to raise interest rates sooner rather than later to be confident of achieving their long- term target for underlying inflation of less than 2.5 per cent in early 1997. (Whether that target is sensible is another matter.)
The CSO's decision to revise up its estimates of growth in each of the last three quarters is not as surprising as it looks. When statisticians are short of hard information on output in particular sectors of the economy, they simply extrapolate past trends to come up with plausible predictions. But this process naturally tends to underestimate the true level - and growth - of output when the economy is accelerating out of recession.
With about pounds 170bn worth of output to measure in each quarter, it is hardly surprising that it takes the statisticians a good while to find it all. Four out of every five revisions to national output figures raise the original estimates. The CSO announced last week, for example, that it had managed to unearth another pounds 1.4bn of output that was produced in 1992 and pounds 1.1bn of output that was produced in 1991. So the level of output in the second quarter of this year is now 0.6 per cent higher than the CSO had previously assumed.
Growth estimates for the first half of this year have been revised up mainly because service sector output has turned out to be larger than early calculations suggested. Higher output of North Sea oil and gas also contributed.
Services production is now thought to have grown at a rate equivalent to 3.2 per cent a year between January and June, compared with the previous estimate of 2.4 per cent. Transport and communication are growing most strongly, with financial and business services, distribution, hotels and catering also doing well.
Statisticians always find it more difficult to pin down service sector output than factory production - it is easier to identify a car plant's output than that of a management consultant. Surveys of the service sector have increased in frequency and more businesses are being forced to participate - but the difficulties cannot be resolved entirely.
In theory, it should be possible to verify calculations of the economy's output, because gross domestic product can be measured in two other ways. The value of goods and services produced is, by definition, equal to the amount of money spent on them. And all this expenditure becomes someone else's income, either a worker's wages or a shareholder's profits.
Unfortunately, the output, income and expenditure measures of gross domestic product can differ dramatically. This makes it hard for the policymaker to know how much the economy is producing, and how quickly it is growing.
This helps explain why the Government acted too late to stop the Lawson boom getting out of control in the late 1980s. Between July and September 1988, the expenditure measure of gross domestic product was 4.5 per cent lower than the income measure and almost 6 per cent lower than the output measure. The output measure was growing twice as quickly as the expenditure measure, which was regarded as the more reliable. The authorities could not be certain which measure was right - but unfortunately chose not to err on the side of caution.
Mr Clarke may risk falling into a similar trap as Lord Lawson, although the economy is obviously growing less quickly now than it was at the height of the last boom.
The CSO now regards the output measure as the best indicator of short-term movements in gross domestic product, so it massages the expenditure and income measures to bring them into line. The figures for stocks of unsold goods and company profits are altered so that the quarter-on-quarter growth rates of expenditure and income match those for output. The graphic shows what happens if you undo this massaging process.
Although the CSO has revised up its estimate of growth in the economy over the past year from 3.3 to 3.7 per cent, this turns out to be the lowest of the three official measures of growth. National expenditure in the second quarter of the year was 4.8 per cent higher than a year earlier, a rate unlikely to be sustainable for long without inflationary pressures building up and the trade gap widening. Growth on this measure has been running at rates equivalent to between 4 and 5.7 per cent a year in each of the last four quarters, well above the 2 to 2.5 per cent that the economy has managed on average over recent decades.
The massaging process only affects the estimates of quarter-on- quarter growth rates through the year. Chastened by the experience of the late 1980s, the CSO puts much more effort during the process of compilation into ensuring that the three measures of gross domestic product are consistent. So discrepancies of the size seen in 1988 will not emerge in the published totals again - but a good deal of behind-the scenes pummelling and squeezing may be needed to keep the figures in line.
Darren Winder, economist at Warburg Securities, argues in his latest UK Economic Briefing that the output measure could be underestimating gross domestic product by as much as a full percentage point in the first two quarters of the year. The expenditure measure could always be overstating growth, by flattering the trade position. But hard evidence has been impossible to come by.
'Even if it is just the level of gross domestic product that is being understated, rather than the growth rate, this still implies that the amount of spare capacity in the economy may not be as great as some commentators suggest,' he argues. Coincidentally, David Miles at Merrill Lynch has cut his estimate of the economy's spare capacity to less than 2.5 per cent on the basis of last week's figures.
The Treasury and the Bank of England will no doubt be concerned that the gross domestic product figures may be sending the same confused messages as in 1988. But they will surely be reassured by the make-up of recovery.
In output terms, economic growth is being boosted by the buoyant North Sea oil and gas sector. This may be relatively non-inflationary, because higher oil output helps the trade position and supports the pound. But it is important to remember that the impact of the oil industry on the real economy cannot be ignored entirely, as the changing fortunes of Aberdeen over the last two decades demonstrate only too well.
On the expenditure side, growth is 'rebalancing' healthily. Consumer spending is slowing under the influence of tax increases and the weakness of the housing market. Investment is picking up slowly but surely, although any surge in companies' capital spending could result in a worrying short-term surge in imports before it has time to expand the economy's productive potential.
This rebalancing is mirrored in the income measure of gross domestic product, where growth is being dominated by rising profits, while wages and salaries remain subdued. A rise in profits as a proportion of national income bodes well for investment and the sustainability of recovery. It also gives companies plenty of scope to absorb pressure on their costs without passing it on to consumers in the form of higher prices.
Mr Clarke will return from his holiday to find a recovery that is nicely structured but could be growing too quickly. The implications for inflation are far from certain, but the lesson of the 1980s is that it is safest to assume the worst.Reuse content