There have been many methods mooted for telling the future before the fact: water divination, the stars, legible tea leaves and numerology. None offers any foresight but they are no less blind than our modern soothsayers, the economic forecasters. In the Budget, the Treasury forecast that borrowing for this year would be £43 billion. Then, in the Pre-Budget Report, it said £78 billion. Alastair Darling, the Chancellor, has already given warning that net borrowing next year might hit £118 billion. Retail sales this November rose 1.5 per cent when economists had been forecasting a fall of 0.4 per cent.

This complete failure to see round corners is exacerbated by the tendency of governments to use statistics not as performance measures but as political mantras. Last week, 10 Downing Street and the Home Office were upbraided by Sir Michael Scholar, the head of the UK Statistics Authority, for releasing data on knife crime that was, as he put it, “premature, irregular and selective”. The Home Office had already been criticised in August by the head of the Royal Statistical Society for playing fast and loose with the release of immigration figures. These incidents cast doubt on all the data in the public domain. No sooner does the Government claim an improvement to a service than the integrity of its statistics is impugned. The Institute of Directors recently decried the data on maths and English performance by 11-year-olds. On all the available data, crime is falling but hardly anyone believes the numbers. It is widely assumed that the Government is lying about the number of un- employed people, the numbers on incapcity benefit, the number of immigrants.

A survey by the Office for National Statistics (ONS) found that fewer than one in five Britons believe ministers do not meddle with official figures. Improving public trust was one of the issues that Gordon Brown said would define his premiership. And, indeed, a reform Bill was introduced in April this year. It created the UK Statistics Authority, to operate purportedly at arm's length from government. The new regime introduced an independent governing board for the ONS, a regular report to Parliament and a session before the Treasury Select Committee. It was all very sensible-sounding and none of it has made any great difference at all.

The current system will never inspire confidence. The basic problem is that the Government is still involved at every stage. Government departments actually compile four fifths of all data that are released. Initial drafts of National Audit Office reports go through a negotiated clearance with departments during which contentious findings are usually removed.

The basic principle to follow is that the people who gather interesting facts should themselves be disinterested. Full independence to the UK Statistics Authority is the only way to increase public confidence. That move should be coupled with the removal of the right of politicians to see data before release. The current rules permit a select group of ministers to see market-sensitive information 40 hours before it is made available to the public. France allows just an hour and the United States no more than 30 minutes.

It would also make sense if the range of advice available to politicians were widened. No institution has a monopoly of wisdom about the present, let alone the future, and the Treasury has no claim to either title. Twenty days into Mr Brown's chancellorship in May 1997, a little-noticed Treasury press release noted that competitive tendering for its forecasting had been rejected. It is time that a wider range of numer- ologists was brought in. A richer conversation and an independent body is the only way to make us believe the statistics are not just damned lies.