Global Value Chains

What does the survey comprise?

Purpose

The primary purpose of the ‘Global Value Chains' statistics is to capture the economic globalization of enterprises in the Netherlands. This is reflected in several key themes.

• Employment by business function
• Purchasing and supplying of goods and services abroad
• Sourcing of business functions
• Impact of recent global events on international value chains

The data are consistent with Eurostat’s ‘Global Value Chains’ statistics.

Target population

The target population consists of all market‑oriented enterprises in the Netherlands within the branches of manufacturing and energy, and commercial services according to the Standard Industrial Classification (SIC), with at least 50 employed persons on average in the final year of the reference period.

Inclusion is based on the General Business Register of Statistics Netherlands (CBS), and requires that enterprises:

• were classified as an ‘enterprise’ during the reference period
• were demonstrably active (in terms of turnover, production, employment, or investments)
• and were still active in the Netherlands at the end of the reference period.

Enterprises that fully relocated abroad and ceased operations in the Netherlands during the reference period are therefore excluded from the target population. In total, the population comprises approximately 11,500 enterprises per reference period.

Statistical unit

The statistical unit is the enterprise unit, as defined by Statistics Netherlands (CBS). An enterprise unit is understood to be the actual transactor in the production process, characterized by independence in decisions about the process and by providing products to others.

Date/year survey started

Starting from the 2021–2023 reference period, this statistic has become a mandatory business survey under European Union regulation. The current design builds on four earlier experimental studies (2001–2006, 2009–2011, 2014–2016, and 2018–2020). These studies are not fully comparable: initially, they focused solely on the sourcing of business functions abroad, but over time the scope has broadened to include additional themes related to international value chains.

The target population has also evolved. For example, the minimum enterprise size was reduced from 100 to 50 employees. In the 2014–2016 study, an attempt was made to collect data at the level of the enterprise group rather than the individual enterprise unit.

Frequency

Data are collected every three years. Each reference period covers three consecutive years, with the periods directly following one another (e.g. 2021–2023, 2024–2026, and so on). For certain topics, data are collected only for the final year within each reference period.

Publication strategy

The figures are published once, and only in their final form, on StatLine, no later than two years after the end of the final reference year. Only figures from mandatory variables required under the EU regulation are published on StatLine, except where there is uncertainty about whether the variable will remain mandatory in the future.

There are two exceptions to this rule:

1. Reference period (2021–2023): Figures from this period will retain a ‘provisional’ status until they can be sufficiently compared with the results of the subsequent reference period (2024-2026).
2. Figures based entirely on secondary data sources: Certain figures, such as Employment by business function, are produced exclusively from secondary data sources. These may retain a provisional status for a longer period, for example until official correspondence tables or other reference frameworks are considered final.

How is the survey conducted?

Survey type

The statistic is based on a sample of enterprises. A stratified sample was drawn from the target population using SBI classification, enterprise size, and multinational status as stratification criteria. Enterprises internally classified as ‘large enterprises’ are fully included in the survey.

Survey method

Data for this statistics are primarily gathered via an electronic questionnaire sent to enterprises. In the Netherlands, respondents know this as the ‘Your Enterprise in International Value Chains’ survey. In addition, secondary data sources are used. For instance, employment figures by business function are derived by combining results from the Labour Force Survey (LFS) and the Polis Administration (SPolisbus).

Respondents

Enterprises registered and active in the Netherlands.

Sample size

The sample comprises approximately 4,200 enterprises per reference period.

Checking and correction methods

Enterprise submissions are assessed for plausibility. The statistics are based largely on unaltered business data, as many variables are difficult to verify using internally available reference sources.

Data quality is ensured through automated checks in the electronic questionnaire, such as consistency validations. Manual corrections are applied sparingly. Where possible, existing data sources are reused as a preferred method. The results are internally reviewed for coherence and adjusted where necessary.

Weighting

Initial inclusion weights are calculated, inversely proportional to the probability of an enterprise being selected in the sample. These weights are scaled to match the total size of the target population and serve as starting weights.

To correct for non-response and improve the accuracy of estimates, reweighting is applied using auxiliary information. The weighting model aligns as closely as possible with the sample stratification and combines characteristics such as SBI classification, enterprise size and multinational status.

Final weights ensure that the response within each weighting cell adds up to the corresponding population count, enabling reliable population totals per publication cell.

Quality of the results

Accuracy

The outcomes are primarily derived from sample-based research, which entails a margin of statistical uncertainty. For key variables, 95% confidence intervals are therefore calculated.

Sequential comparability

The statistical processing method is designed to ensure comparability of results over time. The first mandatory survey builds on four earlier pilot studies (2001–2006, 2009–2011, 2014–2016, and 2018–2020), during which question formats, explanatory notes, and reference periods were refined. From the 2021–2023 reference period onward, the figures are expected to be comparable over time.

Quality strategy

To safeguard the quality of the outcomes, efforts are made to achieve the highest possible response rate. Respondents receive multiple reminder letters. Completed submissions are assigned a plausibility score, although interventions are currently limited due to a lack of reliable reference sources. The online survey includes automated consistency and validation checks. Results are compared over time. Where possible, information is drawn from existing sources, either as reference or preferably to fully supplement the data. In addition, research is conducted into the confidence margins of the key variables. Based on feedback from respondents and accumulated experience, the questionnaire may be further optimized, for example by clarifying instructions or improving question wording.