EMPLOYMENT STATISTICS (Hansard, 6 October 1986)

Employment Statistics

Employees in employment, the self-employed and HM Forces are estimated separately and together comprise the employed labour force. Since 1979 there have been changes in the methodology for estimating the first two elements of the employment labour force. In 1979 the estimates for employees in employment were based on the census of employment, updated by applying proportionate changes in the numbers of employees as estimated from sample surveys of employers. Estimates for the self-employed were obtained from the census of population, updated in 1975 by applying proportionate changes in self-employment from counts of national insurance cards; self-employment was assumed unchanged since 1975.

The first change in methodology since 1979 was the use of labour force survey (LFS) results to produce new estimates of self-employment for 1975 to 1979. These were published in the January 1982 Employment Gazette. The LFS data were used in place of the discontinued information from the national insurance card count.

When provisional results from the 1981 Census of Employment became available at the end of 1982 they showed that the application of changes estimated from sample survey data to the previous (1978) census figure was producing substantial underestimates of the number of employees in employment. The methodology was reviewed and a supplementary set of estimates, which included an undercounting allowance based on the average rate of shortfall which had developed between 1978 and 1981, was introduced. At the same time the conventional assumption that the level of self-employment had remained constant since the date of the latest LFS data was reviewed. As there were reasons for expecting some continuation of the upward movement in self-employment, a supplementary series which assumed a continuation of the rate of growth observed between the latest two surveys was introduced. The reasons for, and full details of, these changes in methodology were set out in the June 1983 Employment Gazette.

115WA

Data from the 1983 labour force survey confirmed that the supplementary figures provided more accurate estimates than the basic series. The estimates for employees in employment were revised, using LFS data for 1981 and 1983, to assess the current extent of underestimation from the sample survey of employers. The self-employment series was updated at the same time. As the department could now produce estimates in industrial and regional detail incorporating the adjustment for underestimation, estimates not incorporating the adjustment were no longer published and the term "supplementary" was no longer used. An article explaining the basis of the new estimates was published in the July 1984 Employment Gazette.

When the estimates of both employees in employment and the self-employed were updated to take account of results from the 1984 LFS and revised data from the 1983 LFS, the figures for self-employment showed exceptional growth between 1983 and 1984 and the department's statisticians considered it inappropriate to assume that this rate had continued. The estimates of both self-employment and employees in employment were further updated earlier this year to take account of the results of the 1985 labour force survey. For self-employment these revised estimates incorporate the assumption that the average rate of increase between 1981 and 1985 is continuing. An article describing the latest revisions was published in the May 1986 Employment Gazette.

Monthly Unemployment Count

The unemployment count is based on administrative records and is inevitably subject to occasional changes in coverage as a result of changes in the way benefits are paid or in signing-on arrangements. Only six changes since 1979 have had a discernible effect on the figures for adult unemployment and have been taken into account in a consistent series of seasonally adjusted estimates according to the coverage of the current count of benefit claimants. Following are the details of all the changes affecting the unemployment series.

In October 1979, fortnightly attendance of unemployment benefit offices was introduced and the estimated effect was to add about 20,000, both to the unemployment count used at the time, based on registration at jobcentres, and the claimant figures introduced later (see below).

In November 1981 the higher long-term rate of supplementary benefit was introduced for men over 60 who had been on supplementary benefit for over one year. Over the following 12-month period this removed an estimated 37,000 men, again from both the registrant and claimant series.

In October 1982 registration at jobcentres became voluntary, saving administrative costs and eliminating the need for unemployed people to attend both a jobcentre and an unemployment benefit office in order to get their benefits.

The previous count of registrants at jobcentres became incomplete and it was necessary to move to counting claimants at unemployment benefit offices. This reduced the count by 190,000 on average as a result of three factors: 116WA

(i) Computerisation of count and improved accuracy with more up-to-date record keeping of those becoming and ceasing to be unemployed (estimated effect–78,000);
(ii) exclusion of registrants not claiming benefits (–135,000);
(iii) inclusion of severely disabled (+ 23,000).

Details of the change were published in the September and December 1982 issues of Employment Gazette and figures on the new claimant basis back to 1971 were then published.

The 1983 Budget provisions enabled 162,000 men, mainly aged 60 and over, to receive national insurance credits or the higher long-term rate of supplementary benefit without attending an unemployment benefit office. The effect accumulated between April and August 1984.

In July 1985 a reconciliation between the Department of Health and Social Security's records and the Department of Economic Development's computer records of claimants showed discrepancies in the figures for Northern Ireland. The corrective action resulted in the adjusted figures being about 5,000 lower than would otherwise have been the case.

From March 1986, the compilation of the figures has been delayed by two weeks, to take place three weeks rather than one week after the specified count date. This excludes from the count an estimated average of 50,000 records of claimants who had already ceased to be unemployed before that date and who were previously over-recorded.

In addition, the change in school-leaving regulations in November 1980 affected the total claimant series later introduced and to a relatively minor extent the registrant series in use at the time but not the seasonally adjusted series of adult claimants. Young people leaving school are now assumed to be in full-time education until the beginning of the following school term and not entitled to benefit. From 1982 a separate count of non-claimant school-leavers registered at careers offices has been conducted in June, July and August when the numbers are significant.

There have also been a few other minor changes and some temporary distortions; for example, as a result of industrial action in the local offices. It was also thought that the introduction of taxation of unemployment benefits in July 1982 may have had some effect on the unemployment figures but none was evident. Similarly, the introduction of payment of unemployment benefit wholly in arrears from July 1985 has had no discernible effect.

Finally, from July 1986 new regional and national unemployment rates are being published which show the number of unemployed as a percentage of the working population (the sum of employees in employment, the unemployed, the self-employed and HM Forces). The considerable growth in self-employment in recent years has made it increasingly important for figures to be published showing unemployment as a proportion of the total workforce. However, unemployment rates on the former basis—showing unemployment as a percentage of the sum of employees in employment and the 117WA unemployed—will continue to be published and are the only rates available for counties and travel-to-work areas because self-employment data for these areas are not available. This latest change in no way affects the numbers of unemployed included in the count.