Even in the imprecise world of forecasting you can expect a downturn after a prolonged upturn, says David Smith, and the order book usually gives the signal.
A few years ago, around the 50th anniversary of VE Day, the Central Statistical Office put out a fascinating publication on how statistics had contributed to the war effort. When the whole country, from factory worker to civil service clerk, was engaged on the home front (and the British economy was centrally planned to a degree never seen before) accurate information was vital. And it seems the statisticians did a pretty good job.
In peacetime, however, official statistics have rarely merited an honourable mention. Harold Wilson blamed his 1970 election defeat on a set of rogue trade figures published just before polling day. Denis Healey says the Labour government would not have been forced in 1976 to call in the International Monetary Fund had it not been for official miscalculations of the state of the public finances.
Last year was the annus horribilis of UK official statisticians, now known as the Office for National Statistics. In the spring, ONS figures suggested average earnings were racing away, providing the Bank of England's monetary policy committee with a reason to raise interest rates when it should not have done so. In the autumn, the ONS suggested the picture was even worse than it had estimated. Two weeks later, its figures showed it was in fact dramatically better. Confused? The Bank of England was, and angry too. So was the chancellor, who ordered an official inquiry.
This was not the ONS's only problem. While every business survey suggested that manufacturing was under the cosh, its figures showed a remarkably benign picture. While exporters complained of the worst conditions in two or three decades, the trade statistics were slow to reflect it. Meanwhile, Britain's biggest rebasing of its national accounts for 50 years left everybody uncertain about how fast the economy had been growing, and how much it was slowing.
Some say the statistician's job was made intolerably difficult by cuts in resources and an attempt in the Thatcher years to relieve firms of the burden of official forms, thus affecting the quality of the raw material on which all good sets of data are built.
There are more fundamental problems. Consider two key economic variables, the inflation rate and unemployment level. Both should be subject to the greatest possible measurement certainty. Neither is. The retail prices index is good at what it measures: a basket of goods and services. But few would claim it is the only representative measure of inflation. In common with inflation indices throughout the industrial countries, it suffers from upward bias, mainly because of the difficulty in incorporating quality improvements and technological change.
This bias is important. In countries such as Germany, it means the difference between a low positive rate of inflation and falling prices or deflation.
As for unemployment - the official measure of which has also changed to a higher level under the present Government - there are those who argue that Britain was not far from achieving full employment during last year; others say the true jobless total was four million or more.
Paul Ormerod, a former model-based economic forecaster, argues in his books The Death of Economics and Butterfly Economics that the economy is too complex a mechanism to lend itself to conventional forecasting.
You can lead a horse to water and sometimes it will drink. But sometimes it won't, and other times it may decide to go for a swim. Multiply such uncertainty of response by the millions of decisions that affect an economy every day and you have some idea of the problem.
Another forecasting problem is the herd instinct. Most forecasters would prefer to be wrong in company than risk being right or wrong on their own.
If statistics are suspect and forecasting even more so, what is a businessman to do when faced with a decision on whether to expand or draw in his horns?
The most important statistical indicator for any business is the performance of that business, and the most important economic forecast is provided by the order book. Remember that official statistics, even business surveys, are simply taking that kind of information and filtering it back a couple of months later. For any business, the great advantage is that you know what is going on before any statistician can. Most businessmen are far more attuned to cyclical warnings than any economist. This does not mean economy-wide statistics or forecasts have no value, but that value should be to reinforce your own view. As a check, there should be plenty of opportunity to see how your peers are doing. The time to get worried is when you are struggling but everybody else is thriving. Recent difficulties at Britain's best-known retailer began as what appeared to be an exclusively Marks & Spencer problem but, as it soon became clear, they were within the context of a struggling sector.
Another essential rule is to remember the cycle. In every prolonged upturn, there are those who speculate that the business cycle has been abolished. It happened in Britain in the 1980s and in the US in the 1990s, where the claim was that the exploitation of information technology created something like a perpetual motion machine for the economy. But the business cycle is made of sterner stuff. When growth has been reasonably strong for a prolonged period, learn to expect a downturn. Even without all the adverse factors of the past 18 months, on cyclical grounds alone, I would have expected a 1998-1999 downturn.
The people who are best at guessing where the economy is going also use informal indicators. If taxis are easier to come by or popular restaurants can be booked with ease, suspect that economic activity is weakening. Look at the appointments sections in newspapers. One that works for me is to count the skips in my street to measure the strength of the housing market.
Above all, never fall into the trap of thinking the statisticians and economists know best.
David Smith is economics editor of the Sunday Times.