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.
The GVC dataset contains ten variables defined in Implementing Regulation (EU) 2022/918. These cover:
(1) employment by business function
(2-5) enterprises purchasing or supplying goods and services abroad
(6) the number of enterprises carrying out international sourcing
(7-8) jobs created and lost due to international sourcing
(9) motivations and barriers for sourcing, and
(10) events affecting GVC arrangements.
For variables (2)-(5), the dataset includes only enterprises that report at least EUR 100 000 in goods or services purchased from or supplied abroad in the final year of the reference period.
All data are provided as absolute values and presented by NACE activity, business function, geographical partner area, size class, motivation and barrier types, and other breakdowns required by the regulation.
Annexes: Implementing Regulation (EU) 2022/918.
3.2. Classification system
For classification (enterprises) statistical classification of economic activities in the European Community (NACE Rev. 2) is used.
Business functions are created following the Manual on the Classification of Business Function and follow the harmonised list defined in the GVC implementing act, distinguishing core and support functions as well as specific functional categories such as production of goods, ICT services, marketing, engineering, and R&D.
Geographical classifications follow the breakdowns prescribed in Regulation (EU) 2022/918, including EU Member States, Extra-EU, and specific partner regions (e.g. China, India, USA and Canada, Central and South America, Africa).
Classifications for job skills follow the high-skilled / not-high-skilled distinction set out in the GVC Compilers’ Manual.
Annexes: Implementing Regulation (EU) 2022/918
3.3. Coverage - sector
Market producers with 50 or more employees and self-employed persons in the final year of the reference period (2023) of NACE Rev. 2 Sections B to N (industry and construction, business services). This sector coverage is defined in Implementing Regulation (EU) 2022/918 and applies uniformly to all variables.
Annexes: Implementing Regulation (EU) 2022/918
3.4. Statistical concepts and definitions
All concepts and definitions follow Implementing Regulation (EU) 2022/918 and the GVC Compilers’ Manual. According to the EU enterprise definition (Council Regulation (EEC) No 696/93), an enterprise is the smallest combination of legal units that is an organisational unit producing goods or services, which benefits from a certain degree of autonomy in decision-making, especially for the allocation of its resources. An enterprise carries out one or more activities at one or more locations. An enterprise may correspond to a single legal unit ("simple enterprise") or consist of several legal units ("complex enterprise"). Global value chains indicate any cross-boarder activity of an enterprise for production or services from conception to delivery to the final customer. The creteria for this concept is wheather a company has supplied or purchased goods or services (from) abroad in the reference period. The lower bound for a particular type of goods or services when these criteria are met equals EUR 100 000. International sourcing is the partial or full movement of business activities that were previously based within the company into foreign countries. International sourcing is possible both within the same enterprise group and to an external company. Business functions are groupings of various parts of business activity with relevance for the enterprises. The core business function is the business function through which an enterprise generates its revenue. All the other business function of the enterprise are then considered as supporting business functions. Persons employed means the sum of all employed persons in an enterprise on 30.09. of the reference year, including employed owners, assisting family members, employees assigned to other companies and home workers, but excluding temporary agency workers assigned to the corresponding company. Job losses cover every single job lost in the home country caused by international sourcing. Job increase covers every single job created in the home country in connection with international sourcing. High-skilled job status is used in terms of the formal qualification, i. e. tertiary education. Examples for high-skilled employees are qualified technicians or persons with a university degree. Geografical areas of international sourcing cover Europe, Asian and Oceanian countries, America and Africa. Within Europe there is a distinction between EU Member States (as of 2023), the United Kingdom and other European countries.
3.5. Statistical unit
The statistical unit for the GVC survey is the enterprise as defined in line with Regulation (EU) 2019/2152 on European business statistics and the Business Register Regulation and Regulation (EEC) No. 696/93 on the statistical units for the observation and analysis of the production system in the Community.
Annexes: Regulation (EU) 2019/2152 on European business statistics Regulation (EEC) No. 696/93 on the statistical units
3.6. Statistical population
Market producers with 50 or more employees of NACE Sections B to N (industry and construction, business services).
3.7. Reference area
Germany
3.8. Coverage - Time
2021-2023 or 2023 respectively. There is no real time series for former periods. Comparisons with data from 2018-2020 period is possible, but only with limitations.
3.9. Base period
Not applicable
All variables in the GVC dataset are compiled and transmitted in absolute values.
Number of enterprises are expressed in units.
Employment variables (number of employees) are expressed in units.
Reference period 2021 - 2023 or reference year 2023 depending on the characteristic.
Reference year 2023 for variables (1)-(5) where the number of employees of that year is defined as the number at a particular date (30 September 2023).
The whole reference period 2021-2023 for variables (6)-(10).
6.1. Institutional Mandate - legal acts and other agreements
National law (Germany): Act on Statistics for Global Value Chains (GWStatG) from 17 January 2024 (published in BGBl. 2024 I No. 13) in connection with the Federal Statics Act (BStatG).
Community law (EU): EBS Regulation (EU) 2019/2152 and EBS Implementing Regulation (EU) 2022/918.
Annexes:
EBS Implementing Regulation (EU) 2022/918
EBS Regulation (EU) 2019/2152
Statistics for Global Value Chains (GWStatG)
Federal Statics Act (BStatG)
6.2. Institutional Mandate - data sharing
Data is transmitted to Eurostat under ESS confidentiality rules where only non-confidential data is available on its dissemination website. According to the Federal Statistics Act (Bundesstatistikgesetz [BStatG]) only aggregated data are disseminated. Art. 16 VI BStatG states that only universities and other institutions who do independent scientific research are allowed to use individual statistical data. The data is anonymised to make it impossible to identify units. Furthermore, due to Art. 9 of GWStatG it is allowed to send tables with statistical results to Federal or Federal State authorities in Germany. For a Statistical State Office as the reciever this paragraph also allows to send individual data of its Federal State. Legal obligation for confidentiality applies to everyone within this (limited/exceptional) permission - legislated in BStatG or GWStatG - to use individual statistical data.
Annexes:
Statistics for Global Value Chains (GWStatG)
Federal Statics Act (BStatG)
7.1. Confidentiality - policy
Article 16 of the Federal Statistics Act (Bundesstatistikgesetz [BStatG]) guarantees confidentiality. It is strictly followed. Respondents must be informed about their rights and obligations when providing statistical information. Furthermore, individual data are used for statistical purposes only and, provided non-disclosure and data protection measures are taken, for scientific purposes.
Annexes: Article 16 of the Federal Statics Act (BStatG)
7.2. Confidentiality - data treatment
Methods of primary and secondary data confidentiality are applied by automatic procedures. Furthermore, additional checks are implemented.
In primary confidentiality, a cell is suppressed amongst others for the following cases:
The cell contains data for three or less than three units (minimum-case-rule).
The largest contribution value can be estimated knowing the second largest contribution value, under the condition that the deviation of the estimated contribution value is less than p percent of the largest contribution value (p-percent-rule).
There is no cell where the p-percent-rule had to be applied.
Secondary confidentiality has also been applied to guarrantee that suppressed cells cannot be calculated from other cells via differences and sums.
Only aggregated data are transmitted to Eurostat, and microdata are not disseminated. Eurostat applies its own confidentiality rules to the aggregated data received, ensuring secure handling and safe dissemination.
8.1. Release calendar
Not applicable
8.2. Release calendar access
Not applicable
8.3. Release policy - user access
GVC data was published on 19 November 2025 and simultaneously released to all users. There is no privileged pre-release access to any outside user according to Principle 6.7 of the European Statistics Code of Practice. Since all users have equal access to statistical releases, impartiality is respected. GVC data are transmitted to Eurostat before they are disseminated at national level. Dissemination follows the NSI's standard user access and embargo rules, which are aligned with national statistical legislation and ESS guidelines. Metadata and documentation are made available at the time of release. GVC data are available on database “GENESIS-Online” (code 52931) provided by the Federal Statistical Office (Destatis) for all users. After registration, users are informed via email when new data are available. Additional information can be provided on request.
Detailed data for 2021-2023 GVC-statistics are available in a statistical report for download the destatis hompepage.
Annexes: Database of the Federal Statistical Office of Germany destatis hompepage European Statistics Code of Practice
GVC statistics are collected, compiled and transmitted every three years in accordance with Implementing Regulation (EU) 2022/918. National dissemination follows the same triannual cycle and usually takes place after the transmission of data to Eurostat.
Annexes: Implementing Regulation (EU) 2022/918
10.1. Dissemination format - News release
A news release was issued on 19 November 2025.
10.2. Dissemination format - Publications
A statistical report for download, web articles, tables and graphics are presented on the homepage for GVC statistics. An English version of this homepage is planned. A paper on methodical explanations will be published in 2026 in the statistics journal "Zeitschrift Wirtschaft und Statistik" (WiSta).
Destatis has committed itself to assure and further enhance the quality already attained. Consequently, Destatis implemented comprehensive quality guidelines. Therefore before releasing data a quality report has been published on the homepage for GVC-statistics. To assure quality in longer terms, Destatis applies a variety of systematic quality assurance measures, e. g. the European Statistics Code of Practice.
11.1. Quality assurance
In the process of compiling statistics, a wide range of measures are carried out that contribute to ensuring the quality of our data. The quality assurance measures, which are applied at different stages during the compilation of statistics, are adjusted where necessary and supplemented by standardised methods of quality assessment and assurance (as set out, for example, in the quality manual of the Federal Statistical Office and Statistical Offices of the Federal States). These standardised methods also include the publication of a national quality report. In order to ensure the most uniform methodological approach possible at the EU level, EUROSTAT provides detailed guidelines on data collection, processing and delivery. In addition, the GVC-statistics (or its predecessors, as the methods have changed over time) have undergone comprehensive quality development since 2006 as a result of intensive piloting, including various method tests. This also includes an overall evaluation. The online questionnaire includes extensive plausibility checks to ensure the internal coherence of the reported data and to give immediate feedback to the enterprises when their data are inconsistent, asking them to correct their data. If there is still implausible data after submitting the questionnaire, this is corrected mainly by automatic plausibility checks (applying SAS-code) and, in few selected cases, by means of queries to the company. Missing data does not occur due to the questionnaire programming. In one special case where it nevertheless has been permitted (enterprises with no sourcing but considered carrying out international sourcing in question 4), it could be replaced by the plausible value “not applicable”. To ensure coherence to employment figures from SBS, the employment figures from GVC are calibrated to SBS when grossing up the sample.
11.2. Quality management - assessment
Overall, the results of the GVC-statistics can be classified as good in terms of accuracy, reliability and timeliness. However, due to the rarity of relocations very detailed evaluations of some characteristics cannot be reliably estimated. Nonresponse is compensated for by means of an extrapolation procedure based on aggregates of economic activity sectors and size classes. The GVC-statistics provides users with relevant results about integration of German enterprises into international sourcing and meet a new demand for data. Extensive plausibility checks during data collection and processing keep non-sampling errors to a minimum. Missing data can almost be excluded due to questionnaire programming (see 11.1).
12.1. Relevance - User Needs
Users of GVC are for instance, European Commission services, the German Ministry for Economic Affairs and Energy, international organisations, chambers of commerce, business associations, trade unions, journalists and researchers. User needs can go beyond standard publications on GVC data. In this case, special analysis can be offered to users depending on their well-defined requests.
12.2. Relevance - User Satisfaction
The characteristics and their specifications (dataset with ten variables) are determined by EUROSTAT in close consultation with all statistical offices of the EU Member States participating in the survey. User consultations take place within the framework of various user committees. Especially some feedback from trade associations could be received in the committee for manufacturing business.
12.3. Completeness
Nothing is missing. Data cover all required combinations of NACE sections B-N (market producers: manufacturing and services), all size classes for complex enterprises with 50 or more employees, all business functions and all the geographical areas as specified in the Implementing Regulation (EU) 2022/918. The 1 percent rule has not been applied for Germany.
Annexes: Implementing Regulation (EU) 2022/918
13.1. Accuracy - overall
The survey is designed in such a way that potential errors can be minimised and controlled. The mathematical and statistical methods used in sampling and data processing enable representative results on international sourcing of Germany's enterprises and their integration into global value chains respectively. Most of the results can be published, because the values are statistically reliable. Only for detailed results the potential of the statistics cannot be fully exploited due to the rarity of some combinations of characteristics, as the data quality will be insufficient in this case (see 11.2). The sample can be considered representative, as the number of non-responses is very low and there are no indications of any related bias.
13.2. Sampling error
The survey relies on a stratified sample using the variables industry, employment size class and a dummy whether an enterprise likely was part of a global value chain as stratification variables. The dummy variable was determined ex ante based on ITGS, business register, and other statistics. This ensures to target the survey mainly to those enterprises which are engaged in GVCs and increases the precision of the estimates for GVC variables. In principle, random sampling errors are estimated by determining relative standard errors. Results that are not sufficiently reliable due to an insufficient standard error are not published. In the GVC-statistics for 2021-2023, this applies to results with a relative standard error of more than 20 percent. Usually, only selected values in low-level subdivisions are affected of that. For this reason, the results are largely presented at a more aggregated, reliable level. The number of enterprises by types of goods and services relevant for global value chains has an average sampling error of 3.4 percent. The characteristic “number of enterprises by carrying out international sourcing” has an average relative sampling error of 9.4 percent. In general, the relative standard errors for characteristics “persons employed” and “number of jobs lost” are low. For characteristic “number of jobs created”, however, higher standard errors occur due to the low number of cases for some combinations of characteristics. The relative standard error for the characteristic “number of employees and self-employed persons by business function” (persons employed) has a range of 0.6 to 2.8 percent.
13.3. Non-sampling error
Item-Non-Response Since the relevant questions on the electronic form are programmed as mandatory questions, missing values in the survey units are virtually non-existent. Implausibilities in the answers provided are checked automatically and, for the most part, also corrected automatically (applying SAS-code). In this process, the distribution of persons employed over business functions was adjusted in 19 percent of all cases. The total of this variable (employees and self-employed persons) was adjusted in 1 percent of all cases. In few cases automatic correction was not possible, so the survey units (i. e. the companies) were asked for clarification.
Unit-Non-Response The number of non-responses is low and does not result in a systematic bias or something of that kind. Among 9813 enterprises drawn into the sample, we received plausible responses from 8607 enterprises, corresponding to an effective response rate of 87.7 percent. Non-responses are taken into account in the extrapolation.
Representation approach for qualitative characteristics In enterprises consisting of more than one legal unit, the main legal unit representative of the whole enterprise in terms of EU definitions has been selected for the survey. In particular, the main business function of this unit corresponds to the economic activity of the enterprise. For extrapolation of quantitative characteristics to the level of the enterprise the so-called representation approach was applied: All qualitative characteristics, i. e. the answers to those questions, of the surveyed legal unit have been assumed as representative for the entire enterprise. Overall, this pragmatic approach allows to derive plausible results on the enterprise level, as in most cases the main legal unit can be considered representative of the total enterprise. However, there can be inaccuracies in some cases if e.g. the main legal unit does not source internationally, while another legal unit within the enterprise does.
14.1. Timeliness
Data for GVC is collected and calculated triannually. Data collection for legal units starts t+10 months after the end of the reference period. Data transmission to Eurostat at t+21 months after the reference period covers the final results for enterprises. Publication at national level is scheduled at t+23.
14.2. Punctuality
Generally, all transmission deadlines are met. Data dissemination at national level took place during the second month following data transmission to Eurostat.
15.1. Comparability - geographical
The same statistical concepts are applied in the entire national territory. There is no publication of GVC results below the national level.
Harmonised concepts for the EU enable geographical comparability for most characteristic. Comparability is limited when there are some optional characteristics that are not published by all EU member states.
15.2. Comparability - over time
Comparability over time is limited, as 2021-2023 is the very first reporting period where answering the GVC questionnaire was obligatory for all legal units contained in the sample. Responses for preceding surveys were voluntary, so there is less usable data available. There had also been a change in the presentation unit as an enterprise was represented by the legal unit before the EU concept had been introduced. Furthermore, the questionnaire can change over time also limiting the comparability of some variables.
15.3. Coherence - cross domain
Calibration to employment figures from SBS ensures coherence of the employment figures between GVC and SBS statistics. Both "International Trade in Goods Statistics (ITGS)" and "Statistics on International Trade in Services (SITS) and Transfers" are not comparable to GVC, although related. The latter statistics is published by the German Central Bank and contains transactions with a value not less than EUR 12 500 per year for each enterprise, while for GVC statistics the lower limit of EUR 100 000 per business function per year must be reached. On the contrary, in ITGS the yearly thresholds for the internal trade in the EU (i. e. when an enterprise must report) are equal to EUR 500 000 for supplied goods and EUR 800 000 for received goods. Another important difference are the units used in the respective statistics. GVC statistics records the number of enterprises only, while there is no information on the total monetary value of exports or imports in contrast to the related statistics. Additionally, in ITGS the quantity of goods is recorded. Both ITGS and SITS use a different classification for goods and services than GVC.
15.4. Coherence - internal
GVC can be described as internally coherent.
Costs related to GVC data collection could be reduced by several organisational arrangements. Survey data is transmitted by a questionnaire that is part of a digital reporting system (IDEV) used by statistic authorities in Germany. The questionnaire is programmed to support users while filling in the answers step by step. A filter question on international sourcing makes it possible to skip irrelevant questions for those participants who negated international sourcing. Small enterprises are not part of the statistical population. Of all the others, only seven percent have been surveyed. Furthermore, the used sampling model includes those enterprises with a higher probability who are expected to be involved in global value chains. Therefore, units expected to be more relevant for GVC statistics are preferred when sampling, while the costs for those units whose data seems to be less relevant for the survey can be reduced.
17.1. Data revision - policy
For GVC no preliminary results are published. Published data is considered as final. There are no revisions to consider.
17.2. Data revision - practice
Not applicable.
18.1. Source data
The GVC data is obtained by conducting a survey. For this purpose, a stratified sample is drawn. Stratification criteria are NACE sections, number of persons employed (50-249, 250-749, 750 or more) and estimated integration in global value chains (yes/no). In correspondence with the Implementation Act, only companies with 50 or more employed persons in NACE sections B-N are drawn. 9813 units were drawn.
Furthermore, data from SBS, German enterprise Register and earnings statistics were used for extrapolating the sample as well as validating and imputating data.
18.2. Frequency of data collection
Triannual.
18.3. Data collection
Data is collected from those enterprises which are in the sample (9813 units covering ca. 7 percent of the statistical population) and that have completed and submitted the electronic questionnaire.
18.4. Data validation
Please refer to Chapter 11.1.
18.5. Data compilation
Imputation and adjusted grossing-up (see also 13.3).
18.6. Adjustment
GVC refers to the calendar year or a three years period, so no further adjustment is required. Adjustments in the extrapolation process were described in section 13.3.
No comment.
The GVC dataset contains ten variables defined in Implementing Regulation (EU) 2022/918. These cover:
(1) employment by business function
(2-5) enterprises purchasing or supplying goods and services abroad
(6) the number of enterprises carrying out international sourcing
(7-8) jobs created and lost due to international sourcing
(9) motivations and barriers for sourcing, and
(10) events affecting GVC arrangements.
For variables (2)-(5), the dataset includes only enterprises that report at least EUR 100 000 in goods or services purchased from or supplied abroad in the final year of the reference period.
All data are provided as absolute values and presented by NACE activity, business function, geographical partner area, size class, motivation and barrier types, and other breakdowns required by the regulation.
Annexes: Implementing Regulation (EU) 2022/918.
28 November 2025
All concepts and definitions follow Implementing Regulation (EU) 2022/918 and the GVC Compilers’ Manual. According to the EU enterprise definition (Council Regulation (EEC) No 696/93), an enterprise is the smallest combination of legal units that is an organisational unit producing goods or services, which benefits from a certain degree of autonomy in decision-making, especially for the allocation of its resources. An enterprise carries out one or more activities at one or more locations. An enterprise may correspond to a single legal unit ("simple enterprise") or consist of several legal units ("complex enterprise"). Global value chains indicate any cross-boarder activity of an enterprise for production or services from conception to delivery to the final customer. The creteria for this concept is wheather a company has supplied or purchased goods or services (from) abroad in the reference period. The lower bound for a particular type of goods or services when these criteria are met equals EUR 100 000. International sourcing is the partial or full movement of business activities that were previously based within the company into foreign countries. International sourcing is possible both within the same enterprise group and to an external company. Business functions are groupings of various parts of business activity with relevance for the enterprises. The core business function is the business function through which an enterprise generates its revenue. All the other business function of the enterprise are then considered as supporting business functions. Persons employed means the sum of all employed persons in an enterprise on 30.09. of the reference year, including employed owners, assisting family members, employees assigned to other companies and home workers, but excluding temporary agency workers assigned to the corresponding company. Job losses cover every single job lost in the home country caused by international sourcing. Job increase covers every single job created in the home country in connection with international sourcing. High-skilled job status is used in terms of the formal qualification, i. e. tertiary education. Examples for high-skilled employees are qualified technicians or persons with a university degree. Geografical areas of international sourcing cover Europe, Asian and Oceanian countries, America and Africa. Within Europe there is a distinction between EU Member States (as of 2023), the United Kingdom and other European countries.
The statistical unit for the GVC survey is the enterprise as defined in line with Regulation (EU) 2019/2152 on European business statistics and the Business Register Regulation and Regulation (EEC) No. 696/93 on the statistical units for the observation and analysis of the production system in the Community.
Annexes: Regulation (EU) 2019/2152 on European business statistics Regulation (EEC) No. 696/93 on the statistical units
Market producers with 50 or more employees of NACE Sections B to N (industry and construction, business services).
Germany
Reference period 2021 - 2023 or reference year 2023 depending on the characteristic.
Reference year 2023 for variables (1)-(5) where the number of employees of that year is defined as the number at a particular date (30 September 2023).
The whole reference period 2021-2023 for variables (6)-(10).
The survey is designed in such a way that potential errors can be minimised and controlled. The mathematical and statistical methods used in sampling and data processing enable representative results on international sourcing of Germany's enterprises and their integration into global value chains respectively. Most of the results can be published, because the values are statistically reliable. Only for detailed results the potential of the statistics cannot be fully exploited due to the rarity of some combinations of characteristics, as the data quality will be insufficient in this case (see 11.2). The sample can be considered representative, as the number of non-responses is very low and there are no indications of any related bias.
All variables in the GVC dataset are compiled and transmitted in absolute values.
Number of enterprises are expressed in units.
Employment variables (number of employees) are expressed in units.
Imputation and adjusted grossing-up (see also 13.3).
The GVC data is obtained by conducting a survey. For this purpose, a stratified sample is drawn. Stratification criteria are NACE sections, number of persons employed (50-249, 250-749, 750 or more) and estimated integration in global value chains (yes/no). In correspondence with the Implementation Act, only companies with 50 or more employed persons in NACE sections B-N are drawn. 9813 units were drawn.
Furthermore, data from SBS, German enterprise Register and earnings statistics were used for extrapolating the sample as well as validating and imputating data.
GVC statistics are collected, compiled and transmitted every three years in accordance with Implementing Regulation (EU) 2022/918. National dissemination follows the same triannual cycle and usually takes place after the transmission of data to Eurostat.
Annexes: Implementing Regulation (EU) 2022/918
Data for GVC is collected and calculated triannually. Data collection for legal units starts t+10 months after the end of the reference period. Data transmission to Eurostat at t+21 months after the reference period covers the final results for enterprises. Publication at national level is scheduled at t+23.
The same statistical concepts are applied in the entire national territory. There is no publication of GVC results below the national level.
Harmonised concepts for the EU enable geographical comparability for most characteristic. Comparability is limited when there are some optional characteristics that are not published by all EU member states.
Comparability over time is limited, as 2021-2023 is the very first reporting period where answering the GVC questionnaire was obligatory for all legal units contained in the sample. Responses for preceding surveys were voluntary, so there is less usable data available. There had also been a change in the presentation unit as an enterprise was represented by the legal unit before the EU concept had been introduced. Furthermore, the questionnaire can change over time also limiting the comparability of some variables.
