Statistics chief says jobs count is not believed

`New standard' needed for employment figures

The head of the Government Statistical Service has warned that the public will never regain confidence in the integrity of the official unemployment statistics unless the monthly count is changed to an internationally recognised standard.

In an interview with the Independent Bill McLennan, who is also director of the Central Statistical Office, responsible for macro economic statistics and the national accounts, said, "nobody believes" the benefits claimants count.

Instead, he said, the UK should publish monthly figures from the International Labour Organisation's Labour Force Survey which defines employment and unemployment. These are published quarterly in the UK. In the interview after announcing his resignationto return to Australia as chief statistican, Mr McLennan revealed other radical ideas which he has suggested to John Major.

In a memo to the Prime Minister, he has proposed the merger of the Office of Population Censuses and Surveys with the Central Statistical Office to combine the collecting arms of social and economic statistics. This would make better use of available data to build a statistical picture of social problems such as the underclass and crime. He is confident of winning support for a "statistical shop front" in London, open to the public, housing a national library of all published statistics.

He has also suggested to the Prime Minister approval of a UK Statistics Act, which would consolidate the changes and improvements he has introduced. This would incorporate a code of practice for official statistics which he has published and is expected to be finalised before he leaves. But, one of his most controversial proposals, discussed with colleagues but not yet with the Prime Minister, is to abandon the current method of publishing unemployment-related benefit counts which the Government introduced in 1982.

This counts the number of people claiming benefits who are "actively seeking work". Previously the Government measured the number of people claiming benefits who registered for work.

Since then, according to the Unemployment Unit, a trade union-funded body, there have been more than 30 changes to the claimant defintion. Until a year ago the unit's researchers compared the pre-1982 registered for work figures with the claimant count, and they last calculated that the claimant count was about one million less than the registered for work figures would have been.

Because of controversy over the changes, the Royal Statistical Society, the statisticians' professional body, set up a working party comprising the present and two past presidents which is due to report in ther new year.

Mr McLennan said that although the unemployment numbers were "published correctly" the public's perception was that they were not to be trusted.

"A lot more people are now willing to trust the Government's statisticians. But I'm also acutely aware that we haven't removed this problem ... I don't think this problem will be removed until you remove the source of it. The only way we are going to stop this ... is to spend some more money and bring out the International Labour Organisation-based definitions monthly".

The ILO definition draws in some people who are either ineligible for unemployment-related benefits or who choose not to register a claim. The claimant count takes in those with relatively low earnings from part-time work who are also claiming unemployment-related benefits legitimately. These would be classified as being employed rather than unemployed.

In recent years the claimant count and Labour Force Survey figures have become closer. In summer 1994 the LFS figure was 2.64 million - 112,000 higher than the 2.52 million in the claimant count.

Interview, page 3

Tomorrow: The Good News Unit and how ministers use health statistics.

LIES, damn LIES "There are three kinds of lies: lies, damn lies and statistics."

Disraeli's distrust of statistics is shared by many people today. They are used - and misused - by politicians, civil servants, and officialdom in general as a tool of convenience.

Can we trust figures on issues such as crime, unemployment, inflation, spending on the NHS, hospital waiting lists, school league tables, company results? Are government statisticians sufficiently independent of political pressure.

Do you have experience of statistics which hide the truth? Do you work with official figures? The Independent would welcome contributions to the debate on statistics. Write, phone or fax to David Felton, Home Editor, The Independent, 1 Canada Square, Canary Wharf, London E14 5DL. Tel: 0171-293 2000. Fax: 0171-293 2047/2051.