Reference metadata describe statistical concepts and methodologies used for the collection and generation of data. They provide information on data quality and, since they are strongly content-oriented, assist users in interpreting the data. Reference metadata, unlike structural metadata, can be decoupled from the data.
INSEE - Direction générale 88 avenue Verdier CS 70058 92541 Montrouge Cedex
1.6. Contact email address
Confidential because of GDPR
1.7. Contact phone number
Confidential because of GDPR
1.8. Contact fax number
Confidential because of GDPR
2.1. Metadata last certified
12 June 2025
2.2. Metadata last posted
12 June 2025
2.3. Metadata last update
12 June 2025
3.1. Data description
Industrial production indices are used to measure the monthly changes of industrial and construction activities in France. As such, they are a primary information to monitor the business cycle in France and to identify turning points at an early stage, in parallel or along with other macroeconomic indicators such as employment statistics, prices indices, the Services Production Index or foreign trade statistics
Sources: They are calculated thanks to branch surveys conducted by INSEE (Institut National de la Statistique et des Études Économiques), the SSP (Service de la statistique et de la prospective), the SDES (Service de la Donnée et des Etudes Statistiques) and professional bodies. They also rely on prices indices for some activities.
Activities covered: Sections B to D of NACE Rev. 2.
3.4. Statistical concepts and definitions
The production indices measure the evolution of output (value added) in volume. The measure of production at finest levels rely on physical units, deflated turnover or worked hours (taking yearly changes in productivity into account).
3.5. Statistical unit
Reporting unit: Legal unit. Observation unit: Kind of activity unit (KAU).
3.6. Statistical population
Units employing 20 employees or more or having a turnover greater than 5 million euros in one industrial branch.
3.7. Reference area
France
3.8. Coverage - Time
Time series start in 1990.
3.9. Base period
Base (reference) year: 2021.
Index
Month
6.1. Institutional Mandate - legal acts and other agreements
Regulation (EU) 2019/2152 of the European Parliament and of the Council on European business statistics (EBS-Regulation) and its General Implementing Act (Regulation (EU) 2020/1197)
6.2. Institutional Mandate - data sharing
Agreements for collecting process with professional bodies and with the statistical services of the Ministry of Ecological Transition (data on energy) and the Ministry of Agriculture (data on food and beverages).
Data are transmitted to Eurostat
7.1. Confidentiality - policy
At the national level, Article 6 of Act No. 51-711 of 7 June 1951, as amended, on the obligation, coordination and secrecy of statistics determines what statistical secrecy is, its limits and the conditions for its application. These rules apply to surveys conducted by the official statistical service (SSP) whether or not they are mandatory. The derogations provided for in this article are governed by the Act. As it stands, the only exemptions that remain applicable are those relating to the status of public archives for surveys and censuses, which authorise the disclosure of individual information contained in the questionnaires and relating to personal and family life and, in general, to the information contained in the questionnaires. family life and, in general, to private facts and behaviour, after a period of 75 years for individuals and 25 years for legal entities. According to the Act, this communication may not be used for tax control or economic repression.
The obligations relating to statistical confidentiality also apply to administrative data that INSEE or ministerial statistical services may have access to under the terms of Article 7 bis of the aforementioned Act, as well as to private data communicated under the terms of Article 3 bis. under the terms of Article 3 bis of the Act. Generally speaking, with regard to access to public data, confidentiality obligations relating to the protection of privacy or business secrecy and the protection of personal data are guaranteed by Act (Article 1 of the Act for a Digital Republic).
At the European level, the confidentiality of statistical information is affirmed by Article 338 of the EU Treaty. The compilation of statistics shall respect (...) the confidentiality of statistical information. Statistical confidentiality is also the subject of Chapter V of Regulation 223/2009 as amended and Implementing Regulation No 557/2013 as regards access to confidential data for statistical purposes.
A Statistical Confidentiality Committee ensures that these statutory guarantees are maintained. Its powers are set out in Article 6 bis of Act No. 51-711 of 7 June 1951, as amended, on the obligation, coordination and secrecy of statistics and Chapter II of Decree No. 2009-318 of 20 March 2009 on the National Council for Statistical Information and the Statistical Confidentiality Committee. It is called upon to give its opinion on any question relating to statistical secrecy, and gives its opinion on requests for communication of individual data collected by means of a statistical survey or transmitted to the official statistical service, for the purpose of establishing a statistical report. for statistical purposes. Researchers can also ask the committee to give an opinion on access to various administrative data other than public statistics.
Chaired by a State Councillor, it includes representatives of the National Assembly and the Senate. The composition and operating procedures of the committee are set by decree in the Council of State. The beneficiaries of data communications resulting from ministerial decisions taken after the opinion of the Statistical Confidentiality Committee undertake not to communicate these data to anyone.
Any breach of the provisions of this paragraph shall be punishable by the penalties provided for in Article 226-13 of the Criminal Code.
7.2. Confidentiality - data treatment
The main rules for turnover indices are: no dissemination of data if they are based on the compilation of less than 3 enterprises or if a single entreprise represents more than 85% of the turnover of the field covered by the serie (dominance rule and p% rule).
Confidential treatment through TAU ARGUS (software designed to protect statistical tables) and confidential indices are not disseminated.
8.1. Release calendar
A monthly calendar is produced for the four following months. It can be found on the INSEE website (Publication calendars).
Internet users can consult and download all the available statistical data free of charge on the INSEE website, as well as the information needed to interpret them correctly.
The rules of access, particularly the embargo rules, are described on Insee website.
The transmission to Eurostat is made using the SDMX format, the day of the national release.
The results are available in the indices and time series category of the statistics and studies section on the Insee website. Data can be downloaded in xlsx or csv format.
They can also be retrieved via a web service, available on the Insee API portal and compliant with the SDMX standard.
A quality report has been drawn up in accordance with the quality assurance framework (see 11.1).
11.1. Quality assurance
Since 2005, the European Statistics Code of Practice has been the reference for assessing the quality of the output of national statistical institutes. Periodic reviews by European peers are organised to ensure that the principles of this reference framework are implemented and to ensure that each institute is committed to continuous improvement. Within this framework, INSEE has adopted a process-based approach. A range of tools, pooled within the Official Statistical Service (SSP), has been created to describe statistical production processes, analyse their strengths and weaknesses, assess the risks involved, examine their documentation (metadata) or assess a particular stage (analysis of users' needs, data validation, etc.). The diagnoses resulting from these quality approaches lead to the establishment of action plans that are regularly monitored in the context of process reviews. In addition, INSEE regularly conducts satisfaction surveys on the indicators and data it produces. The results of these surveys are available on the Insee website.
11.2. Quality management - assessment
Eurostat requirements are fulfilled and the variables used to calculate the indices are accurate and of good quality.
12.1. Relevance - User Needs
The Industrial production indices are meant to be used by many clients (i.e. users) including Eurostat, the official statistical service (INSEE and the SSM) and the general public (inseenauts, private or public institutions, media).
Different types of products are produced by the team on Industrial production indices:
monthly InfosRap publications
aggregated data files and series available on the on-line database
specific files responding to user needs (National Accounts, Business Cycle Analysis, Structural Surveys, SSM)
metadata: annual quality reports sent to Eurostat and made available on Insee website.
12.2. Relevance - User Satisfaction
This process is not subject to a specific satisfaction survey for external users but internal users are satisfied with the quality and punctuality of indices.
12.3. Completeness
Eurostat’s requirements in terms of completeness are fulfilled.
13.1. Accuracy - overall
There is a trade-off between accuracy and timeliness. Thus, the monthly production indices are provisional.
They comprise around 10% of estimated data.
Errors are mainly due to incorrect reports of the legal units and imputation errors.
The results are normally final 95 days after the end of the month.
13.2. Sampling error
Not available but limited
13.3. Non-sampling error
Coverage error
Units with fewer than 20 employees and those in which none of the industrial branches exceeds €5 million in annual turnover are not covered by the sample.
Measurement error
The measurement error is small: this would be due to inaccurate reporting not identified in the data clearance process.
Non response error
Response rate: Around 90% each month.
Processing error
The process is robust and uses cross-validations methods.
14.1. Timeliness
35 days after the end of the reference period (except for the publications related to the reference month of July and November, released 40 days after the end of the reference period).
14.2. Punctuality
Deadlines are respected and data have so far always been published on time.
15.1. Comparability - geographical
Methods are comparable to those of other European. They are compliant with the european regulation (EBS).
15.2. Comparability - over time
Consistency over time is ensured through the use of stable methods. In the case of methodological changes, or change of base period, past indices are backcast to ensure comparability over time.
15.3. Coherence - cross domain
Indices and microdata are confronted with the results of the annual production survey Enquêtes annuelles de production (structural business statistics).
A benchmarking with other STS data, structural data or national accounts can be carried out. In addition, national accountants use Industrial production indices for quarterly estimates or provisional annual estimates and provide feedback on consistency with past account data.
15.4. Coherence - internal
Internal consistency is ensured through the aggregation method.
Burden: 27 216 hours / Cost: 40 000 hours.
These figures correspond to the costs for Insee and the burden for enterprises to calculate IPI for Eurostat and for national purposes.
17.1. Data revision - policy
The same revision policy is applied to STS data released nationally and transmitted to Eurostat.
All revisions are taken into account.
Routine revisions are mainly due to available additional information, especially overdue respondents at the survey.
Then SA-WDA data are also revised each month as a consequence of the new calculation of SA-WDA coefficients. These revisions are taken into account each month.
Major revisions are mainly due to rebasing, new weights or update of productivity factors.
No revision calendar exists.
17.2. Data revision - practice
Routine revisions are explained in the national publication. There is a dedicated paragraph entitled revision of variations.
Warnings are published in the national publication Informations Rapides and on the insee website when the revisions are major (revisions like after rebasing or changes in methodology).
Generally, initial and final values are close, there are few revisions.
In the whole industry, the mean revision between january 2021 and december 2023 is -0.27 point for the year-on-year growth rate of the calendar adjusted serie (see Figure below).
In the whole industry, the mean absolute revision between january 2022 and december 2024 is 0.58 point for the year-on-year growth rate of the calendar adjusted serie (see Figure below).
These revisions include the impacts of rebasing or major methodological change.
Revisions in the whole industry (sections B to E) between january 2022 and december 2024 (36 months):
MR = +0,08
MAR = +0,58
RMAR = +0,54
Annual rebasing process
Since the previous 2015 rebasing in March 2018, the industrial production index is annually chained-linked (it was previously a fixed-weights index). In addition, the industrial production indices (IPI) are now reviewed at the rate of one seventh of the series each year. The aim of this annual update is to improve the relevance of the IPI with regards to the evolution of technologies and production processes, by adapting the way of measuring a product, by including emerging products or, on the contrary, removing old products. Each year, the annual review is implemented with the publication of March concerning the indices of January. The adjusted series are backward-reviewed. At an aggregated level, the revisions on indices are moderate. Nethertheless, some indices can be significantly revised resulting from an improving coverage or a modification of the measurement variable (for instance production sold in euros instead of quantities).
18.1. Source data
Type of source: Statistical survey, except for certain activities which are subject to specific data collection (e.g. energy data).
Sample for industry except manufacture of food products and beverages: about 4800 firms; the sample is selected in the almost 30000 firms of the Annual Production Survey (with thresholds of 20 employees and 5 million euros of turnover).
Sample for the manufacture of food products and beverages: more than 7000 firms. The survey is conducted by the SSP (Service de la Statistique et de la Prospective), the statistical service of the Ministry of Agriculture.
18.2. Frequency of data collection
Monthly
18.3. Data collection
Questionnaires used in the survey: Enquêtes de branche (branch surveys).
Planned changes in national questionnaires: Yearly.
Data collection media: Mainly (95%) electronic (Web questionnaire).
Data for energy and food and beverages industries are received from institutional partners.
Some data are also obtained from professional bodies.
18.4. Data validation
The data are validated before being sent to Eurostat. This validation is done in two stages: the checking (and if necessary correction) of individual data and then the checking of the aggregated indices. This second check may lead to further corrections of the individual data. We use selective editing.
18.5. Data compilation
Micro data is checked using selective editing methods.
Estimates for non-response on individual data: data from the previous period multiplied by a geometric mean between the average rates of change of the responding units and the changes between the same two months in the previous year, weighted by the response rate.
Type of index: Annual chain-linked Laspeyres indices with 2021 as reference year
Method of weighting and chaining: The weightings of the branches are based on raw value added at basic prices and are updated annually (chain-linked index). Within each branch the index is based on a representative sample of activities weighted by the sampling weights. The indices have a mean of 100 in 2021.
Planned changes in production methods: regular work is done to improve the construction of the sample and then the robustness and accuracy of estimates. In particular, work is carried out on the optimal choice, depending on activities, between cut-off and stratified sampling methods.
Actions to speed up or increase the rate of response: Phone or mail recall, (rarely) penalties and increased use of internet for data collection.
18.6. Adjustment
Adjustment
Indices are seasonally and working-day adjusted.
Seasonal adjustment
The raw indices are seasonnaly and working-day adjusted (SA-WDA) using the X13 ARIMA program available in JDemetra +. The WD adjustment (trading days, leap year) and the seasonal adjustment decomposition are calculated at the 4-digit level of the NACE Rev. 2. The upper levels are obtained by aggregating the series (indirect method), in the same way as the agregation of raw data.
The Reg ARIMA calendar adjustment is used by constructing working day regressors based on the French national calendar (which takes into account working days specific to France).
Outliers (additive outliers, temporary changes, level shifts, seasonal outliers) are fixed in the past and are detected automatically on the past 12 months onwards. The critical value for outlier detection, the filter length and the model/filter selection depend on the series and may have to be changed manually to improve the quality of the seasonal correction. This was the case to neutralize some particular points associated with the 2020-2021 health crisis (lockdowns for example), which would have induced an unjustified distortion of the seasonal coefficients over the past.
Either additive or multiplicative decomposition can be used. The seasonal adjustment models are reexamined every year (favouring stability) and the parameters are re-estimated every month.
Each month the SA-WDA data are revised from 2012. For the seasonal adjustment of indices in the recent past, the models are now estimated over a reduced sub-period (from 2012 onwards), in accordance with Eurostat guidelines, and in order to reinforce the robustness of the seasonal adjustment. The data before 2012 are fixed in evolution, in accordance with Eurostat's guidelines (avoid revisions over a too long period).
Not available
Industrial production indices are used to measure the monthly changes of industrial and construction activities in France. As such, they are a primary information to monitor the business cycle in France and to identify turning points at an early stage, in parallel or along with other macroeconomic indicators such as employment statistics, prices indices, the Services Production Index or foreign trade statistics
Sources: They are calculated thanks to branch surveys conducted by INSEE (Institut National de la Statistique et des Études Économiques), the SSP (Service de la statistique et de la prospective), the SDES (Service de la Donnée et des Etudes Statistiques) and professional bodies. They also rely on prices indices for some activities.
The production indices measure the evolution of output (value added) in volume. The measure of production at finest levels rely on physical units, deflated turnover or worked hours (taking yearly changes in productivity into account).
Reporting unit: Legal unit. Observation unit: Kind of activity unit (KAU).
Units employing 20 employees or more or having a turnover greater than 5 million euros in one industrial branch.
France
Month
There is a trade-off between accuracy and timeliness. Thus, the monthly production indices are provisional.
They comprise around 10% of estimated data.
Errors are mainly due to incorrect reports of the legal units and imputation errors.
The results are normally final 95 days after the end of the month.
Index
Micro data is checked using selective editing methods.
Estimates for non-response on individual data: data from the previous period multiplied by a geometric mean between the average rates of change of the responding units and the changes between the same two months in the previous year, weighted by the response rate.
Type of index: Annual chain-linked Laspeyres indices with 2021 as reference year
Method of weighting and chaining: The weightings of the branches are based on raw value added at basic prices and are updated annually (chain-linked index). Within each branch the index is based on a representative sample of activities weighted by the sampling weights. The indices have a mean of 100 in 2021.
Planned changes in production methods: regular work is done to improve the construction of the sample and then the robustness and accuracy of estimates. In particular, work is carried out on the optimal choice, depending on activities, between cut-off and stratified sampling methods.
Actions to speed up or increase the rate of response: Phone or mail recall, (rarely) penalties and increased use of internet for data collection.
Type of source: Statistical survey, except for certain activities which are subject to specific data collection (e.g. energy data).
Sample for industry except manufacture of food products and beverages: about 4800 firms; the sample is selected in the almost 30000 firms of the Annual Production Survey (with thresholds of 20 employees and 5 million euros of turnover).
Sample for the manufacture of food products and beverages: more than 7000 firms. The survey is conducted by the SSP (Service de la Statistique et de la Prospective), the statistical service of the Ministry of Agriculture.
Monthly
35 days after the end of the reference period (except for the publications related to the reference month of July and November, released 40 days after the end of the reference period).
Methods are comparable to those of other European. They are compliant with the european regulation (EBS).
Consistency over time is ensured through the use of stable methods. In the case of methodological changes, or change of base period, past indices are backcast to ensure comparability over time.
