Economics: Fresh ideas please, and spare the job juggling

UNEMPLOYMENT has fallen by more than 400,000 since the end of 1992, when it peaked at just under 3 million. This should be a cause for celebration, but nothing is ever that straightforward with the British economy. Look more closely at the statistics and a baffling and contradictory picture emerges.

According to the Department of Employment, the number of people without work and claiming benefit began falling just a few months after the recession ended in 1992. Since then, it has dropped on average by nearly 20,000 a month. This is in marked contrast to the mid-1980s, when unemployment was still rising five years after the recovery in output of goods and services got under way.

Most economists believe this is because the Government's anti- union legislation in the 1980s made it easier and cheaper to hire and fire. But can we really trust the statistics? The Employment Department measures both the jobless total and the number of people in employment in two ways - and each tells a different story.

The headline (or 'claimant') unemployment total counts the number of people without work and receiving benefits. An alternative measure is derived from the quarterly Labour Force Survey, which questions people in 60,000 households. Someone is unemployed on the LFS definition if they say they are without work, are available to start work in the next fortnight and have looked for a job in the last four weeks.

On the face of it, the two measures of unemployment appear to be telling the same story. Both the LFS measure and the claimant count show a similar drop in joblessness since 1992. But, paradoxically, one would expect these measures to diverge at this stage of an economic recovery.

Why? Although the LFS and claimant counts both show unemployment totalling around 2.6 million, only 1.6 million people are unemployed on both definitions. One million people are claiming benefits but tell the LFS they are not looking for work, while another million say they are looking for work but do not claim benefits.

The latter group, who are unemployed only on the LFS definition, normally find work relatively early in the recovery. But people receiving unemployment benefit but not looking for work usually continue claiming benefits until later in the upturn. This means the LFS measure of unemployment normally falls earlier and faster than the claimant count once the economy is growing. But not during this recovery.

Leo Doyle, at Kleinwort Benson, argues that the claimant count may have been falling relatively quickly this time because people have claimed invalidity or sickness benefit instead of unemployment benefit. There has also been a big rise in the number of people telling the LFS that they are not looking for work because they are sick or disabled. Mr Doyle does not dispute that there are around 2.6 million people unemployed today, but he believes that both the LFS and the claimant count have exaggerated the size of the recent fall.

The confused signals from the alternative measures of the jobless total are as nothing against the differences between the two official measures of employment. Employers report that they have shed 101,000 jobs since unemployment peaked, but 294,000 people claim that they have found new jobs in the same period, according to the LFS. Officials at the Employment Department are so baffled by this discrepancy that they have launched an investigation.

The LFS is a sample rather than an exhaustive survey of the entire population, so perhaps the 60,000 households questioned are giving untypical answers?

Possible, but unlikely. Statistical theory suggests that there is less than a one-in-20 chance that a survey as large as the LFS could measure employment incorrectly by more than 150,000 either way. Even if this were the case, a substantial discrepancy would remain.

Part of the explanation is that the recovery is creating what Michael Portillo calls 'the sort of work employers do not care to declare'. These could be jobs for which the pay is so low or the hours so short that the employer does not have to pay national insurance contributions. Or companies may be concealing jobs from the Inland Revenue illegally.

The 'back to work bonus' unveiled by Peter Lilley in his speech last week to the blue-rinsed hordes in Bournemouth is partly intended to smoke out some of these hidden part-time jobs, which ministers believe are escaping the scrutiny of the tax and benefit systems.

There is little point in benefit recipients taking part-time work - or at least declaring it - because a lot of what they earn is offset by cuts in the payments they receive from the DSS. The bonus scheme will allow benefit claimants who declare that they are working part-time to build up a credit of up to pounds 1,000 which they can cash in when they get a full- time job. This should encourage people to remain in touch with the world of work.

The scheme is unlikely to make a big dent in the jobless total, but it is an imaginative attempt to promote work incentives. Kenneth Clarke has hinted that there is more of this to come in the Budget. He is well aware that the interaction between trends in the labour market and the operation of the welfare state remains an urgent policy problem. It is the main explanation for the rise in unemployment in western Europe over the last 30 years.

Advances in technology and competition from low-wage producers overseas have created pressure for greater inequality of earnings in industrial countries. Employers are increasingly prepared to pay more for educated workers but less for the unskilled. As benefits move roughly in line with changes in average earnings, this growth in inequality has left more people at the bottom of the scale in the position where remaining on benefits is more attractive than taking low-paid work.

The macho response to this problem is to cut the level of benefits. The consequence - as the US demonstrates - is appalling inequality and rising crime rates.

The touchy-feely answer is to spend more on education and retraining, so that low earners are better equipped for more highly paid jobs. Mr Clarke has paid lip service to this solution, which is also the policy prescription implied by Gordon Brown's much- mocked 'post neo-classical endogenous growth theory'.

Poncendogrot, as a fellow columnist on this newspaper has dubbed the concept, argues that countries which spend most on education tend to grow most quickly. Unfortunately, this conclusion is much stronger for developing countries than for industrial nations, which already spend a lot on education and training.

As Paul Krugman, the US economist, told a recent symposium organised by the Federal Reserve Bank of Kansas City: 'It is hard to escape the feeling that those who place their faith in education and training as the major solution to the problems of jobs and wages are engaging in wishful thinking, driven by an unwillingness to face up to the harshness of the trade-offs involved.'

The problems are difficult, but not intractable. Wage subsidies, loans for retraining and reforms of Family Credit and Housing Benefit are all fruitful territory. Dubious statistical evidence that unemployment is falling sharply is no excuse for losing interest.