United Kingdom (1)

Title of the survey

Short-Term (Monthly and Quarterly) Employment Statistics

Organization responsible

Employment Department, Statistical Services Division, Branch D2.

Periodicity of the survey

Monthly for production industries, and quarterly for agriculture, construction and service industries.

Objectives of the survey

To provide monthly and quarterly information on employment levels according to industrial activity and location of individual workplaces, at national and regional levels. This information is used by government ministries, Prime Minister, policy branches, statistical branches, Treasury, academics and others, for monitoring employment trends.

Main labour topics covered by the survey

Employment and hours of work.

Reference period

For employment, a specific day each month or quarter. For hours of work, a specific week each month or quarter.

Coverage of the survey

Geographical

The monthly survey of production industries covers Great Britain only. The quarterly surveys cover the whole of the United Kingdom and provide separate data for the whole country, Great Britain and standard regions.

Industrial

All divisions of economic activity, except the armed forces. However, some divisions are not covered by employer-based surveys, but data for them come from administrative centralized returns provided by government departments (e.g. construction is covered by the Department of Environment survey) and other large organizations (e.g. Post Office, the London and Scottish Clearing Banks Association).

Establishments

All types and sizes of establishments or data units.

Persons

Monthly estimates cover production workers (manual workers, operatives) in manufacturing. Quarterly estimates cover employees in employment. Excluded are working proprietors and working directors, unpaid family workers and young workers below the age of 16.

Occupations

All occupations are covered but none are separately identified. The only occupational groups that are separately identified are operatives and administrative, technical and clerical staff in manufacturing.

Concepts and definitions

Employment

The data refer to employees in employment. This is a count of civilian jobs of employees paid by employers who run a "Pay-As-You-Earn" (PAYE) income tax scheme. Employees include wage earners and salaried employees; permanent, temporary and casual employees; apprentices, trainees and workers on probation; persons temporarily absent from work because of paid or unpaid vacation, sickness or accident, temporary or indefinite lay off, industrial dispute, or any other reasons, whether authorized or not; and employees who work away from the workplace such as sales representatives and lorry drivers. Participants in Government employment and training schemes are included if they have an employment contract. As the estimates of employees in employment are derived from employers' reports of the number of people they employ, individuals holding two jobs with different employers will be counted twice. Excluded from employees in employment are private domestic servants, piece workers, home workers and commission agents; persons temporarily present on payroll during notice period preceding retirement, resignation or dismissal; former employees still on the payroll as pensioners; persons on temporary military service; those employed by outside contractors or agencies, and employees who normally work at another establishment such as temporary transfers and secondments. The following categories of workers are identified separately:

manual workers (operatives), who are production workers, maintenance, warehouse and transport workers and foremen (except works and other higher level foremen);
administrative, technical and clerical staff, who include managers, administrative, professional, scientific, technical, clerical and sales employees;
part-time operatives, who are persons normally working for not more than 30 hours a week except where otherwise stated;
operatives on lay off during the reference week;
operatives on short-time;
operatives working overtime.

Earnings

Not relevant.

Wage/salary rates

Not relevant.

Hours of work

Data are collected on hours lost on short-time and on overtime hours worked by operatives in manufacturing and in Great Britain only. Hours lost on short-time refer to those arrangements made by an employer for working less than regular hours (i.e. less than the normal weekly hours specified in national collective agreements or statutory wages orders for manual workers). Short-time working excludes time lost through sickness, holidays, absenteeism and the direct effect of industrial disputes. Overtime hours refer to hours outside normal hours, for which a premium rate is paid. Data on hours lost on short-time working and on overtime hours are collected separately for full-time and part-time operatives. Data are not collected on normal hours of work.

Classifications

Industrial

Employees are allocated to the activity code of the establishment or data unit in which employed, and classified according to the Standard Industrial Classification (SIC), 1980. This classification is not linked to the International Standard Industrial Classification of all economic activities (ISIC), Revision 2, 1968. However, the recent revision (SIC-1992) is convertible to ISIC, Revision 3, 1990.

Occupational

Not relevant.

Others

The survey data are also classified according to:

standard region (Borough of Greater London, county in England and Wales, administrative region in Scotland);
sex (only employment data);
full-time and part-time workers.

Sample size and design

Statistical unit

The sampling and reporting unit is the data unit, i.e. a specific PAYE paypoint at a single site which belongs to a single industry. So a company with a corporate headquarters and four branch offices is regarded as five separate sampling units, each of which could be chosen independently for inclusion in the panel survey. If one of the branch offices has two paypoints (one for salaried staff and one for weekly paid staff) there would be six units and again each unit would have an independent chance of selection.

Survey universe / sample frame

The sampling frame consists of the Pay-As-You-Earn (PAYE) income tax records of employers, maintained by the Inland Revenue. Income tax records are updated constantly, but with a certain time-lag, when the tax authorities are notified by employers.

Sample design

The sample is designed around a core, called the main panel, augmented with a quarterly top-up of newly created business units, after recording units which have closed. The main panel is designed as a sub-set of all companies which responded to the last Census of Employment (the present main panel is based on the 1991 census), stratified by region, industry and number of employees. Each panel member is then approached for up-to-date information on the number of employees. A random sample of data units is drawn within each stratum. In order to take account of the creation of new companies, the main panel is supplemented by quarterly Inland Revenue information on the quarterly registration of PAYE schemes by employers. This information identifies new PAYE paypoints which are examined and sometimes approached to see if they fall within the boundary of the new units. Those which do are sampled to provide a suitable top-up of new units to add to the main panel. The sampling method is similar to that used to identify the original main panel, i.e. a stratified sample by region, industry and size. Each quarter, approximately 30,000 units are in the sample (i.e. about three per cent of all units and 33 per cent of employment), with a monthly sub-sample of 6,000 units in the manufacturing and production industries. The sample consists of over 6,000 strata and the sampling fraction varies with each stratum. Large companies and organizations (generally those with 1,000 or more employees, sometimes those with 250 or more employees) are included with certainty. The sample drawn from businesses accounts for about two thirds of all employees. The remainder are covered by a separate system of centralized returns whereby complete details on employment in certain industries are collated each quarter and sent in by one central point within the industry. The sample is updated every quarter, using the information provided by the Inland Revenue tax records. Sample rotation is not used.

Field work

Data collection

This takes place each month or quarter. Questionnaires are mailed to the sample units up to two weeks prior to the reference date. Reminder forms are issued for non-responding units and phone contacts are made for queries.

Survey questionnaire

This incorporates the statutory notice under the Statistics of Trade Act, 1947, and instructions on definitions, inclusions and exclusions with regard to each question. Different questionnaires are used for different industries; in general, they include questions on:

number of employees, by sex, full-time and part-time, and reasons for important changes in numbers since last return;
number of operatives, hours lost on short-time working and number of overtime hours (only in manufacturing);
number of administrative, technical or clerical staff, by sex;
reason(s) why the business has stopped;
contact name for further queries.

Substitution of sampling units

There is no substitution of sampling units in the case of total non-response.

Data processing and editing

Data are processed by computer, and the forms are entered by batch onto the data base. Data checking and editing involves computer-generated queries with respect to credibility or comparison with data for the previous month or quarter. Part-time workers are counted as a separate category contributing to the total in the same way as full-time workers. No allowance is made for differing amounts of hours worked. Missing data are dealt with by imputation or allowed for in the grossing-up procedure.

Types of estimates

total of employees, for the month or quarter;
total of operatives (in manufacturing);
totals and averages of overtime hours and hours lost by short-time work.

Construction of indices

Overtime, short-time and employment data are combined with data on normal hours of work derived from the New Earnings Survey, to produce an index of average and total hours worked by operatives in manufacturing.

Weighting of sample results

The employment estimates are weighted on the basis of the latest Census of Employment, according to the following formula:

where,

E_ir(t)=: employment estimate for industry i, region r, at time t,
C_ir(t)=: Census estimate,
s=: size range,
W=: weight,
e_irs(t)=: sample employment at time t,
e_irs(0)=: sample employment at Census time.

Adjustments

Non-response

There is no adjustment for non-responses. Employers are required by law (under the 1947 Statistics of Trade Act) to furnish replies to the questionnaires.

Other bias

No adjustments are made for any other bias.

Use of benchmark data

The estimation formula (see above) includes benchmark information on the basis of the latest Census of Employment.

Seasonal variations

Data are seasonally adjusted, using the X-11 programme. Additive factors are produced for employment, and multiplicative factors are produced for hours. Forward factors are used for one quarter, then the programme is re-run and the previous quarter is revised. There is also an annual review of seasonal factors to produce consistent series.

Indicators of reliability of the estimates

Coverage of the sampling frame

Coverage is virtually complete. PAYE income tax records are updated constantly, but with a certain time-lag, when the tax authorities are notified by employers.

Sampling error / sampling variance

Not available.

Non-response rate

In manufacturing, the monthly non-response rate averages 13 per cent and the quarterly non-response rate is approximately 15 per cent. In the services, it is around 13 per cent.

Non-sampling errors

The main known sources of bias relate to possible duplications or omissions from the benchmark (Census of Employment), incorrect data received from employers, errors in keying the data or in clearing the queries.

Conformity with other sources

In the past (up to 1991), the survey results were compared with the results of the annual Labour Force survey and adjustments were made to the survey results to bring the annual rates of change in line with the LFS. This practice has been discontinued and the new panel of employers designed in 1992 is expected to track the real economy accurately and provide estimates that are both timely and reliable. The series of employees in employment are then revised retroactively on the basis of the next Census of Employment (as this was the case in 1993, on the basis of the results of the 1991 census).

Available series

Published tables include estimates of employees in employment

by industry, sex and full- or part-time;
by industry and region;
by industry, sector and full- or part-time (seasonally adjusted and non-adjusted);

and estimates of overtime and short-time operatives and hours in manufacturing, including monthly indices.

History of the survey

The Short-Term Statistics of Employees in employment started in 1971. Various changes were made to the layout of the questionnaires at various times, but the same information has been collected. Periodically, estimates of employees in employment are revised on the basis of the latest Census of Employment. The 1991 census introduced a change in the collection of data for employees working for local authorities, which had implications on the estimates for industrial series. In March 1992 a new panel of employers was selected. At the same time, estimation techniques and allowance for non-response were also modified. Changes planned for the next few years include:

improvements to the Register of Employers and adoption of a business register (EDBR) common to all Employment Department employment surveys;
change in industrial classification in 1995;
faster processing of survey documents by optical character recognition technology;
new short-term estimates rebased on the 1993 Census of Employment;
earlier close-down of short-term surveys.

Documentation

Employment Department: Employment Gazette (monthly, London). Monthly employment and hours data in manufacturing are published some eight to ten weeks after the survey reference period. Quarterly data in services and data for the whole economy, including that part of the workforce in employment not covered by the surveys, 14 to 16 weeks after the reference period. Additional data which do not appear in publications can be made available upon request, in the form of ASCII files, LOTUS-123 or SUPERCALC spreadsheets, CSV, etc., on diskette.

Confidentiality / Reliability criteria

The Statistics of Trade Act, 1947, under which the survey is conducted, prohibits the publication of information about indentifiable individuals or units. The information supplied by employers is never disclosed to any unauthorised person without their consent.

Other information

Data supplied to the ILO for publication

Estimates of employees in employment form part of the employment estimates published in Tables 3A to 5B of the Yearbook of Labour Statistics. Quarterly employment series are published in Tables 1 to 3 of the Bulletin of Labour Statistics. Statistics.