Institut National de la Statistique et des Etudes Economiques (INSEE).

Labor market overview and outlook division (in French : Synthèse et conjoncture du marché du travail)

88 avenue Verdier, 92120 Montrouge, FRANCE

INSEE employment estimates track the number of persons of employees and self-employed persons as registered in administrative records, following as closely as possible the concepts defined by the International Labor Organization (ILO), after processing multiactivity:

- employment is measured on the last week of the year and any employment declared in the administrative data (social declarations, payroll files) is recorded;

- Processing multiactivity means that persons who have more than one job during the last week of the year are counted only once for their main job, which corresponds to the one that provides the highest remuneration over the year.

NACE Rev. 2

This metadata file is applicable to all sectors.

INDUSTRY
Economic activities listed in sections B to E of NACE (B-Mining and quarrying, C-Manufacturing, D-Electricity, gas, steam and air conditioning supply, E-Water supply; sewerage, waste management and remediation activities).

CONSTRUCTION
Economic activities listed in section F of NACE (Construction).

WHOLESALE AND RETAIL TRADE
Economic activities listed in section G of NACE (Wholesale and retail trade; repair of motor vehicles and motorcycles).

SERVICES
Economic activities listed in sections H, I, J, M and N of NACE (H-Transportation and Storage, I-Accommodation and food service activities, J-Information and communication, M-Professional, scientific and technical activities, N-Administrative and support service activities).

INSEE employment estimates track the he number of persons of employees and self-employed persons as registered in administrative records, following as closely as possible the concepts defined by the International Labor Organization (ILO), after processing multiactivity:

- employment is measured on the last week of the year and any employment declared in the administrative data (social declarations,payroll files) is recorded;

- Processing multiactivity means that persons who have more than one job during the last week of the year are counted only once for their main job, which corresponds to the one that provides the highest remuneration over the year.

Trainees, unpaid family help, and those persons involved in national service in the military are excluded (although the latter are included in the labor force). In the Eurostat database, the series provided cover only wage employment, contrary to Insee annual estimates which cover both wage earners and own-account workers. For the purposes of employment statistics, managers and company directors are considered to be wage earners.

Number of payroll employees and self-employed persons

Industrial coverage: Quaterly employment is calculated for all sectors of activity (agriculture, industry, construction, market services and non-market services).

Occupational coverage: There is no selection by occupation. Temporary workers are included in the business services sector (Nace section N).

Geographical coverage: Employment is calculated for the whole of France (including the overseas departments, except Mayotte).

Population coverage: All residents on French territory are included. 

Employment is calculated for the whole of France (including the overseas departments, except Mayotte).

• French employees working in French companies abroad are not covered
• Foreign employees working in foreign companies on French territory are covered.

The data on payroll employment are published using the French activities nomenclature (NAF) with 38 categories or 17 categories, depending on the time period :

• On non-farm market sectors, data are available from 1989 in NAF38 and from 1970 in NAF17
• On non-farm non-market sectors, data are available from 2010 (NAF38)
• On agriculture, data are available from 2010 (NAF38)
• The distinction between public and private employment is available since 2010

2015 average

Data are published at national level in absolute values corrected of seasonal variationand at european level in indices in percent

Quarter

All statistics collected and published by Insee are governed by the statistical Law (« Loi (modifiée) du 7 juin 1951 sur l'obligation, la coordination et le secret en matière de statistique »), published in the « Journal officiel de la République française » and the data processing and privacy protection Law (« Loi du 6 janvier 1978 relative à l'informatique, aux fichiers et aux libertés »), published in « Journal officiel de la République française ».

All statistics transmitted to Eurostat are under the EBS regulation (EU 2019/2152) and EBS GIA (2020/1197).

The quaterly employment time series is being produced by Insee with the collaboration of :

• Acoss (organization in charge of collecting employer’s social contributions) for the private sector, excluding temporary employment
• Dares (statistical department of the Ministry of Labor) for temporary employment data.

All statistics collected and published by INSEE are governed by the statistical (amended) Law of the 7th June 1951 ”concerning the obligation, the coordination and the secrecy as regards statistics”. The protection of private information during data processing is covered by the Law of the 6th January 1978 “concerning data processing, files and freedoms”.

Employment quaterly data are published using the French activities nomenclature (NAF) with 38 categories for wage earners, and 5 of these 38 categories are also given at a more precise level (in NAF 88). These categories cover a large enough number of individuals, even when split by geographical areas. No particular treatment is therefore required to ensure confidentiality.

A calendar of precise release dates is published 4 months in advance by INSEE on the 25th of each month.

The general public can gain direct access to this calendar on INSEE's website: (https://www.insee.fr/fr/information/1405540?conjoncture=27&debut=0). This calendar is also published once a month in INSEE's "Tableau de bord hebdomadaire" (in French).

Data are published:

• by INSEE in the quarterly publication: "Informations rapides - emploi salarié à la fin du Xième trimestre". The quarterly series are first published in a flash estimate (45 days after the end of the quarter), partly based on econometric estimations. The first estimation relying fully on administrative data is released 70 days after the end of the quarter and is broken down by sectors. Both publications are available in French and in English. Data are then transmitted to Eurostat in indice of percent at each period using the transmission handler of Eurostat to send series under an xml file.
• by the Ministry of Labour in its quarterly publication "Premières informations et premières synthèses - Évolution de l'emploi salarié au xième trimestre 200x" (in French).

Both "Informations rapides" and "Premières informations et premières synthèses - Évolution de l'emploi salarié au xième trimestre 200x" are sent to all press agencies as a press release at 7:15 a.m., under a 15 minutes embargo, or at 8:30 a.m. for provisionnal (flash) results.

The frequency of data transmission to Eurostat and the frequency of national dissemination are quarterly.

None

The quarterly series are first published in a flash estimate (45 days after the end of the quarter). The flash estimate only covers the private field. It is broken down by major activity sector: agriculture, industry, construction, market services, non-market services. In market services, temporary work is isolated. When a quarter flash employment is published, data for previous quarters are not revised excepting for temporary employment. Since the first quarter of 2020, the flash estimate for the private sector is mainly based on the “Declarations sociales nominatives” (DSN) received and processed on an early basis by the URSSAFs (social security contribution collection offices), DARES (statistical service of the Ministry of Labor) and INSEE constitute the main part of the information used.

The first estimation relying fully on administrative data is released 70 days after the end of the quarter and is broken down by sectors (NAF38). Data can be revised over a period of three years due to final annual estimates (see below). All revised data are published in Insee's "Informations rapides". Both publications are available in French and in English.

Annual statistics on employment are published twice a year : in the fall of year N+1 for annual estimate of employment at the end of year N (provisional data), and every march of year N+3 annual estimate of employment at the end of year N (final data).

All time series are available throuh the "Base de données macroéconomique" (BDM) of Insee.fr

Data available on Insee.fr are updated simultaneously with the publications (quarterly frequency, T+45 and T+70).

Not available

Data are sent to Eurostat

See the documentation in annex below, avalaible on Insee.fr

Not available

Quality checks are performed during the process of data validation : detection of gaps in data collection, incorrect inputs by employers, inconsistency between employees count and total payroll etc.

Not available

Need for information on labour market developments, in particular on employment, for the general public, the media, regional and national public actors, etc.

Not available

Quarterly statistics on employment cover :

• all payroll employment ;
• all economic sectors ;
• the whole of France, excluding Mayotte.

Non-wage earning employment is covered only on an annual basis, by lack of relevant quarterly administrative records.

Employment estimates are based on administrative records. These records are deemed exhaustive. Inaccuracy can therefore come from :

• errors in records (in employers declaration or in the collection chain) ;
• inadequate statistical treatments.

Administrative data is supposed to be exhaustive. Gaps in collectected data are corrected if any is detected.

Error in administrative records or in statistical treatments.

Provisional data (flash estimation) are disseminated 45 days after the end of the reference period.

Final data are available 70 days after the end of the reference period.

National publication deadlines have always been met. Data is also delivered at the Eurostat’s deadline (approximately T + 3 months).

Employment data is disseminated on Insee.fr at a regional level (13 "régions" and 97 "départements"). Geographical comparability is ensured by the use of a single information system for the whole of France.

The time series do not show any significant breaks. Major methodological corrections due to improvements in data collection or statistical treatment are always back-casted to ensure time-constency.

Long times series on employment in France excluding Mayotte, using the results of the Labour Force Survey, are published in INSEE's "INSEE - Résultats: Marché du travail, séries longues" (in French). LFS provide measure on employment as defined by the International labour office (ILO). An employed person as defined by the ILO is a person aged 15 or over who worked at least one hour for pay or profit during the reference week or was temporarily absent from such work (because of annual leave, sick leave, maternity leave, etc.).
 
Insee's "Employment estimates" are based on admistrative data ("registered employment"). The central concept of "employment" is similar to that of the ILO but, apart from the usual differences between surveys and admistrative data, several differences remain. In particular  :

• Measure of Insee's "Employment estimates" is dated  at the end of the year, whereas LFS employment series are calculated as an annual mean

• LFS employment is "resident" employment : contrary to LFS, Insee's employment include persons who live outside France (Belgium, Spain, Italy, etc.) and work in France, and exclude the ones who live in France and work abroad

• Other adjustments are made regarding persons who have several jobs

These differences are explained in the note in Annex, available on Insee.fr

The aggregates are always consistent with their main sub-aggregates.

Not available

When a quarterly flash employment is published (T+45), data for previous quarters are not revised except for temporary employment.

When the 70-day employment estimate is published, data for previous quarters are subject to revision.

• First, the raw data can be revised, either by incorporating new information (delayed declarations, etc.), or additional processing, or by taking into account new annual estimates (in autumn n+1 for the provisional annual estimate of the year n, in March n+3 for the final estimate).
• Additionnaly, every quarter, the updating of the seasonal variation coefficients leads to a slight revision of the series published. Specifically, seasonal adjustment models are recalculated once a year, when the results for the first quarter are released. In the publications for the other quarters, the seasonal coefficients are revised but the models remain unchanged.

Same as "Policy" above.

Since the first quarter of 2020, the flash estimate for the private sector (excluding agriculture and private individuals employees) is mainly
based on an arbitrage made from several sources:

-          The “Declarations sociales nominatives” (DSN) received and processed on an early basis by the URSSAFs (social security contribution
collection offices), DARES (statistical service of the Ministry of Labor)
and INSEE constitute the main part of the information used.

-          Calibration models estimated by INSEE from INSEE business tendency surveys are used for information purposes.
For agricultural employment, estimates are produced by INSEE based on the first information extracted from the DSN. A forecast is made by INSEE
on private individuals employees. Finally, temporary employment
is compiled by Dares.

The second estimate of payroll employment published 70 days after the end of the quarter is drawn up from different sources:

-          declarations of social contributions made to the URSSAFs (social security contribution collection offices), either in the form of “bordereaux récapitulatifs

de cotisations” (BRC) or nominative declarations, namely “déclarations sociales nominatives” (DSN);

-          payroll files for central governement civil servants;

-          declarations of social contributions collected by the agricultural social-security mutual fund (Mutualité Sociale Agricole, or MSA);

-          declarations of private individual employers from three administrative requirements: simplified nominative declaration (DNS), universal employment service

check (chèque emploi service universel, CESU) and childcare allowance ( prestation d'accueil du jeune enfant, PAJE);

-          temporary employment agencies “déclarations sociales nominatives” (DSN).

 In the private sector excluding temporary employment, DSNs are processed by the central agency of social security organisations, ACOSS, within the SEQUOIA system (see "Sources et méthodologie", p11 of this publication,

in French only). On the civil service scope, INSEE processes the BRCs and DSNs within the Épure system, as well as the central government civil service payroll files (which were previously integrated

into the Siasp system). INSEE also processes MSA data via the Épure system. The declarations about private individuals' employees are processed throuh the system for information on private individuals' employees

set up by INSEE. Finally, the DSNs of the temporary employment agencies are processed by DARES (see DARES methodological note, in French only).

Quarterly

Combination of paper and (mostly) electronic methods

Intra dataset checks : gaps in data collection, time inconsistencies, inconsistencies between employees count and payroll.

Treatment of missing values: As the main source is administrative data, data is corrected for total non response (lack of declaration for a given quarter).

Seasonal adjustments: The quarterly data are seasonally adjusted using procedure X13-ARIMA, using the JDemetra+ software. The seasonal adjustment coefficients are recalculated every quarter, and seasonal adjustment models once a year, when Q1 results are released.

Not available