Algorithm sampling script

Dutch Tax and Customs Administration

This algorithm takes a statistical sample from a given population. The algorithm supports drawing a statistical sample and recording its steps.

Last change on 23rd of September 2025, at 13:30 (CET) | Publication Standard 1.0

• Publication category: Impactful algorithms
• Impact assessment: Field not filled in.
• Status: In use

General information

Theme

Public finance

Begin date

2018

Contact information

[e-mail]

Link to publication website

https://over-ons.belastingdienst.nl/onderwerpen/omgaan-met-gegevens/algoritmeregister/

Link to source registration

https://over-ons.belastingdienst.nl/onderwerpen/omgaan-met-gegevens/algoritmeregister/steekproefschrift/

Responsible use

Goal and impact

The sample script ensures uniform usage within the Tax Administration. Changes in national policy can be made centrally so that the most up-to-date version is used. Alternatively, one could define the statistical measures used to draw the sample oneself, and take the sample oneself in Arbutus. This is more error-prone than a centralised approach.

Also, the script leads to standard reporting for audit file, time saving for IT auditor and standard output to the companies.

The script ensures random draw of cash units to be audited, making the sample demonstrably representative of the population from which it was drawn. The use of the sampling script excludes any human bias when performing correctness checks.

Considerations

Drawing money samples is important for objective, efficient and transparent execution of book audits of taxpayers. The algorithm can support a Tax Administration employee in doing so. As a result, the assessment is more careful and efficient.

Human intervention

Human intervention in the Tax Administration context implies that a competent and knowledgeable employee plays a substantial role in decision-making.

Human intervention is always involved in the operation of the algorithm. The algorithm detects and selects the entries to be checked in the financial records. It is the Inland Revenue employee who makes the decision. The designated entry is checked for tax accuracy by the Inland Revenue employee, with the sample script playing no role at all.

Risk management

The General Administrative Law Act (Awb) requires the government's actions to be transparent and lawful. The Tax and Customs Administration observes the general principles of good governance when applying and developing algorithms.

The selection rules are reviewed periodically and adjusted if necessary to remain in compliance with laws and regulations.

Legal basis

1. General State Tax Act:
2. Payroll Tax Act 1964:
3. Income Tax Act 2001:
4. Corporation Tax Act 1969:
5. Turnover Tax Act 1968:

Links to legal bases

• General State Tax Act:: https://wetten.overheid.nl/BWBR0002320/
• Payroll Tax Act 1964:: https://wetten.overheid.nl/BWBR0002471/
• Income Tax Act 2001:: https://wetten.overheid.nl/BWBR0011353/
• Corporation Tax Act 1969:: https://wetten.overheid.nl/BWBR0002672/
• Turnover Tax Act 1968:: https://wetten.overheid.nl/BWBR0002629/

Elaboration on impact assessments

• Privacy and AVG

The use of data must be assessed against the General Data Protection Regulation (AVG). Reviewing personal data reveals any privacy risks and allows appropriate measures to be taken.

The AVG prescribes that no more data should be used than necessary. This is called data minimisation. The Tax Administration regularly examines whether the data used are still necessary and may therefore be used.

• Use of special personal data

No special personal data are used in the algorithm.

• Equality and non-discrimination

The algorithm is assessed in line with applicable non-discrimination principles for direct and indirect discrimination. Processing as little personal data as possible reduces the risk of direct discrimination. Employees involved in developing and managing the algorithms receive training on data protection and bias.

Operations

Data

Mutations from financial records

Links to data sources

Mutations from financial records: Bedrijf zelf

Technical design

The sampling script produces a statistical sample. The control officer assesses the mutations designated by the sample (at random).

The algorithm is not self-learning. This means that the algorithm does not develop itself during its use.

External provider

The algorithm was developed by staff at the Inland Revenue and is also maintained internally.