France (3)

Title of the survey

Enquête sur le coût de la main-d'oeuvre (Survey on Labour Costs). (Merged with the wage structure survey in 1992)

Organization responsible

L'Institut National de la Statistique et des Etudes Economiques (INSEE) is responsible for organizing and conducting the survey. The results are published under the responsibility of the INSEE and the Statistics Office of the European Communities (EUROSTAT).

Periodicity of the survey

Four-yearly (1984, 1988, 1992).

Objectives of the survey

To evaluate wages and labour cost as well as its components by economic activity, size of production unit and geographical location, thereby facilitating comparisons of labour cost in the countries of the European Union.

Main labour topics covered by the survey

Employment, earnings, hours of work and labour cost.

Reference period

The calendar year if the financial year begins in January and ends in December, otherwise a twelve-month period corresponding to the financial year.

Coverage of the survey

Geographical

The whole country (Metropolitan France).

Industrial

The 1992 survey covered all branches of economic activity in the private sector.

Establishments

Establishments of all sizes in the private sector.

Persons

All employees with a contract of employment.

Occupations

Occupations are not recorded in the survey on labour cost. However, data on occupations have been collected as part of a supplementary survey on wage structures.

Concepts and definitions

Employment

The data concern the employees of the establishment. These include full-time employees (working the normal stated hours of work, regardless of duration of employment in the year in question) and part-time employees (working a reduced timetable in relation to normal hours of work), as well as workers under contract to the establishment and workers without an exclusive contract of employment.

The following categories are identified separately:

Excluded from the survey are directors or managers who do not earn a salary (but remunerated on a profit-sharing or lump sum basis), workers paid in the form of fees, workers from temporary work agencies or other temporarily hired workers, and persons on early retirement, exempted from activity or under suspended contract. For employees under contract, data are collected on the average number per month of employees remunerated on a full-time basis and employees remunerated on a part-time basis, by category. Figures for other employees under contract are given in terms of number of contract-months for apprentices and young persons in training, and in terms of average monthly number of workers paid by commission. The number of persons without a contract of employment is expressed in terms of workers recorded who were engaged in the activity during the financial year.

Labour cost

This represents total expenditure on wages and salaries (gross remuneration paid by the establishment) and in employers' statutory and non-obligatory contributions made on behalf of employees during the reference period. Total gross remuneration comprises the basic wage, overtime payments and paid vacations, the various bonuses and allowances, including benefits in kind. It excludes the wages of medical workers if these are paid by the works council of the enterprise or establishment, vacations, fees, and flat-rate expense allowances of company directors. The following wage cost components, of which certain are included in the gross wage and others added to it, are identified separately:

Hours of work

The survey covers:

The hours actually worked for an establishment are calculated as follows: Total of hours paid for in respect of employees under contract (full-time or part-time), excluding apprentices, young persons in training and commercial travellers, representatives and brokers, adjusted by the average rate of absence in the establishment (calculated on the basis of individual data in the section wage structure) from which are subtracted: hours of training and hours of paid leave (calculated on the basis of the total of hours paid for, modified by a coefficient of paid leave; this is the ratio of annual leave entitlements - requested in the questionnaire - to the number of days normally paid for, i.e. 262).

International recommendations

The definition of labour costs is in line with that contained in the international recommendations. The concept of hours actually worked used in this survey for the two major groups of employees corresponds to the concept contained in the international recommendations. The other concept refers to hours paid for, including hours paid for but not worked.

Classifications

Components of labour cost / compensation of employees

The statistics on labour cost are classified by groups of components (see above, under Concepts and definitions). This classification corresponds to the International Standard Classification of Labour Cost (1966), with certain modifications such as, the cost of housing which is included among benefits in kind. Moreover, certain components of labour cost not elsewhere classified in the international classification are described as groups of components (cost of transport to and from work borne by the employer) or included under other major groups of components (cost of recruitment).

For the statistical requirements of the European Communities, the data on labour cost structure are classified as follows:

Industrial

The data relating to labour cost are classified according to principal activity engaged in (APE: Activité principale exercée) and the Standard Classification of Economic Activities of the European Communities (NACE) - 1990, Rev. 1. The NACE Rev. 1 is itself based on the International Standard Industrial Classification of All Economic Activities (ISIC), Rev. 3, 1990.

Occupational

Not relevant in the context of the survey on labour cost.

Others

The data are also classified according to size of establishment (establishments with fewer than 10 employees, 10 to 19, 20 to 49, 50 to 99, 100 to 199, 200 to 499, 500 to 999, 1000 or more).

In 1992, the distinction between manual and non-manual workers was dropped. A breakdown is now made (for data on employment and hours of work) between full-time and part-time employees. This distinction does not concern expenditure on wages and labour cost.

Sample size and design

Statistical unit

The statistical unit is the establishment, in principle. However, if an enterprise with several establishments does not have separate data for each of them, it may supply the required information by applying a distribution ratio to each establishment on the basis of direct wages or workers employed.

Survey universe / sample frame

The sample frame is the file on establishments, based on the Système informatique pour le Répertoire des Entreprises et des Etablissements (SIRENE) of the Institut National de la Statistique et des Etudes Economiques (INSEE). The SIRENE data file is updated daily to take into account the foundation, modification and closure of enterprises (over 10,000 daily updates), on the basis of official statements issued by the chambers of industry and commerce, the Clerks' Offices of the Commercial Tribunals, the tax offices, etc., and by INSEE surveys. In 1992, the SIRENE data base listed some 3,300,000 active enterprises and about 3,800,000 active establishments. The establishments listed in SIRENE employ both wage earners and non- wage earners. According to the employment survey of March 1993, the economically active population in employment (excluding conscripts called up for national service) at that date came to 22 million, including 2,6 million employers and self-employed persons and 0,5 million family helpers.

Sample design

In 1992, the survey covered a sample of 25,000 units employing over three million wage earners. The sample of respondents comprised 16,000 establishments employing 2,210,000. After stratification by region, economic activity and size of unit, the sampling units were selected with unequal probabilities, according to rates inversely proportional to the number of units in the universe.

Field work

Data collection

The survey is conducted by postal questionnaire. A preliminary notice is sent to the establishments selected before the survey begins, to inform them of the objectives and methods of the survey. The questionnaire is sent in January of the year following the reference year (January 1993 for the 1992 survey) and it must be returned to the INSEE by 1 March of the same year.

Survey questionnaire

Because the labour cost survey was merged with the wage structure survey in 1992, three separate questionnaires were used: In the first questionnaire, the establishments are asked about their workers, the main qualitative and quantitative aspects of the remuneration system, wage bills and wage costs, and certain characteristics about the organization of production, products and markets specific to the enterprise or establishment. Individual data on the employees or a sample of them (qualification, type of contract, composition of annual wage or salary, etc.) are also requested: for this purpose, a sample of employees is selected according to the procedure described in the explanatory notice. The purpose of the second questionnaire is to study the wage and salary structure, and it requests information on the employees of a sub-sample of establishments, such as last diploma obtained, professional experience, and other elements relating to the employee's professional career or family circumstances. The third questionnaire, addressed to enterprise management in cases where an establishment questioned is part of a multi-establishment enterprise, concerns in particular: employment in the enterprise, financial relationships, economic characteristics (market share, proportion of production located abroad), wage policy at enterprise level, restructuring, concentration or diversification projects over the last five years.

Detailed explanatory notes are enclosed with the questionnaires.

Substitution of sampling units

In the case of total non-response (because activities have ceased, change of address, refusal, etc.), establishments are not replaced.

Data processing and editing

Data are processed by computer after manual checking. In certain cases, the questionnaires are returned to the units surveyed for supplementary information.

Types of estimates

In order to measure hours of work and hourly labour cost, part-time employees are converted to full-time equivalents. The aggregated results are obtained by weighting, using the number of hours actually worked in the establishment.

Construction of indices

Not relevant.

Weighting of sample results

Ex ante weights are calculated as the inverse of the sampling fraction. To adjust for non-response, the weights of respondents are corrected according to the probability of their responding to the survey. This is estimated according to the characteristics of the establishment, available in the survey data file taken from the SIRENE directory. The weights for non-respondents are then reassigned to respondents with similar characteristics, in a varying manner according to the probability of the estimated response.

Adjustments

Non-response

These adjustments form part of the sample weighting procedure (see above).

Other bias

Not relevant.

Use of benchmark data

At the enterprise level, the data are benchmarked on the SIRENE directory. The data on wages and salaries are checked for consistency against the Déclarations Annuelles de Données Sociales and the Employment Structure Survey.

Use of other surveys

Not relevant.

Indicators of reliability of the estimates

Coverage of the sampling frame

The SIRENE data base is the official directory in which all enterprises in France and all their establishments are registered.

Sampling error / sampling variance

Not available

Non-response rate

32 per cent in 1988, 39 per cent in 1992.

Non-sampling errors

Not known.

Conformity with other sources

The data on employment collected in the survey on labour cost are compared with those derived from the survey on employment structure. The level of basic wages, bonuses etc. is compared with the results of the survey on the wage structure.

Estimates for non-survey years

These estimates are calculated by EUROSTAT and established as follows for each year between survey years (the following example applies to 1989): C89 = (s89/s88) * W88 * (1 + K89), where C89 = cost updated for the year 1989 s89 = average hourly (monthly) earnings in 1989 s88 = average hourly (monthly) earnings in 1988 W88 = average hourly (monthly) earnings in 1988 reported in the last cost survey K89 = updated average share of additional costs for 1989 (additional costs/direct earnings). s89/s88 gives the rate of increase in earnings according to national surveys. The source used is the six-monthly Survey on the activity and working conditions of the labour force (ACEMO).

(1 + K89) is the coefficient of the updated cost structure. If the weight of additional costs in proportion to direct earnings has varied because of changes in legislation or collective agreements, this is taken into account in the estimates.

Available series

by economic activity, size, region and employee category, and cross-classifications of these variables.

History of the survey

The survey on labour cost is part of the European Community system of statistics on wages and is conducted under the auspices of the Statistical Office of the European Communities (EUROSTAT). From 1966 to 1981, the survey took place every three years. Since 1984 it has been conducted every four years. The surveys relating to 1966, 1969, 1972, 1975, 1981 and 1984 covered industry, whereas those relating to 1970, 1974, 1978, 1981 and 1984 covered wholesale and retail trade, banking and insurances. The 1988 survey, conducted in 1989, collected data on labour cost only and covered industrial establishments and construction enterprises, wholesale and retail trades, banks, insurances and services with 10 or more employees. The 1992 survey on labour cost was merged with the survey on wage structure (which supplies individual information for a sample of employees) so as to enhance the use of data from each source through the information collected from the other, and also to rationalize the response burden on the establishments. The survey coverage was extended to include establishments with between 1 and 9 employees and all non-agricultural market activities. Since 1992, the distinction between manual and non-manual workers has been dropped from the results on the level and composition of labour cost.

Documentation

Institut national de la statistic et des études économiques (INSEE): Economie et statistique (11 issues per year; Paris); the survey results are published about 18 months after the reference period. idem: INSEE Résultats (irregular; ibid.). idem: INSEE Première (irregular; ibid.). In the last two mentioned publications, one issue every four years is devoted to the survey results.

Statistical Office of the European Communities (EUROSTAT) Labour Costs 1988 Vol. 1: Principal Results; Vol. 2: Results by size class and region; Series 3C (Luxemburg, 1992).

idem: Labour Cost - Updating 1989-1991 Series 3C (ibid., 1993).

At the time this methodological description was prepared, the results of the 1992 survey were not yet available.

Detailed results not appearing in the national publications can be obtained from the INSEE on request, and the data can be provided on tape or diskette.

Confidentiality / Reliability criteria

In accordance with Act No. 51-711 of 7 June 1951 (amended), concerning obligation, coordination and secrecy in statistical matters, the data collected are covered by statistical secrecy and may under no circumstances be used for purposes of tax inspection or economic repression.

The dissemination of the results is also subject to statistical secrecy whereby each cell of results must contain at least three units, none of these having a weight superior to a certain percentage of the total.

Other information

Data supplied to the ILO for publication

Statistics on hourly labour cost in manufacturing are published in Tables 22A and 22B of the Yearbook of Labour Statistics.