Directive on Administrative Coorperation-6 (DAC6)

Dutch Tax and Customs Administration

This page contains information about the 'Directive on Administrative Coorperation-6' algorithm. This algorithm helps Tax and Customs Administration staff monitor DAC6 notifications, which are received by the Tax and Customs Administration under this European directive.

Last change on 13th of May 2024, at 11:58 (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

Field not filled in.

Contact information

[e-mail]

Link to publication website

https://www.belastingdienst.nl/wps/wcm/connect/nl/intermediairs/content/mandatory-disclosure-rules-dac6

Link to source registration

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

Responsible use

Goal and impact

In 2018, the European Commission adopted the DAC6 ("Directive on Administrative Cooperation") directive. This directive aims to encourage cooperation between European Union member states on unwanted tax avoidance. In the Netherlands, this directive has been incorporated into the International Tax Assistance Act (WIB). This introduced an obligation to report 'cross-border constructions' by intermediaries in particular, for example tax advisers, lawyers or accountants.

For more information on the DAC6 directive, see.

https://www.belastingdienst.nl/wps/wcm/connect/nl/intermediairs/content/mandatory-disclosure-rules-dac6

Around 1,000 notifications per year are expected to involve one or more Dutch taxpayers. A multitude of these have also been received about constructions implemented in the period 25 June 2018 to 31 December 2021. The algorithm puts all these notifications in order of importance. Notifications with the highest risk of tax evasion are put at the top. Conducting supervision is more efficient and effective with the algorithm.

Considerations

Around 1,000 notifications per year are expected to involve one or more Dutch taxpayers. A multitude of these have also been received about constructions implemented in the period 25 June 2018 to 31 December 2021. The algorithm puts all these notifications in order of importance. Notifications with the highest risk of tax evasion are put at the top. Conducting supervision is more efficient and effective with the algorithm.

Human intervention

Field not filled in.

Risk management

The algorithm was developed by the Tax Administration and is also maintained internally. Qualitative evaluation of the algorithm takes place periodically. This makes it possible to assess whether the algorithm needs to be improved.

The use of the data is tested against the General Data Protection Regulation (AVG). The Tax Administration prevents direct discrimination with the algorithm by not using protected personal data, such as ethnic origin. However, the data 'country of residence' and 'tax resident' are processed. These are not used in the algorithm as an alternative to ethnic origin.

For the development of algorithms, the Tax and Customs Administration has drawn up conditions, a quality framework. This contains rules and agreements that were followed when developing the algorithm. The conditions of the Audit Service Rijk (ADR) are leading in this respect. When changes are made to the algorithm, the Tax and Customs Administration checks whether the algorithm still meets the quality requirements.

The use of the relevant data is tested against the relevant legislation.

The AVG and WIB stipulate that we cannot use more data than necessary. The Inland Revenue regularly reviews whether the data used is still necessary and therefore may be used. If this is not the case, the algorithm is adjusted and this data is also no longer used.

The algorithm puts reports about individuals and companies in a specific order. This order is only used in the described process. Outside of that, they are not used, to avoid profiling that is not allowed.

Legal basis

The collection and use of the data described above is governed by the:

Links to legal bases

• Algemene wet inzake rijksbelastingen: https://wetten.overheid.nl/BWBR0002320/
• Wet op de internationale bijstandsverlening bij de heffing van belastingen (WIB): https://wetten.overheid.nl/BWBR0003954/
• Wet Inkomstenbelasting: https://wetten.overheid.nl/BWBR0011353/
• Wet op de Vernnootschapsbelasting: https://wetten.overheid.nl/BWBR0002672/
• Richtlijn 2011/16/EU van de Raad van 15 februari 2011: https://eur-lex.europa.eu/legal-content/NL/ALL/?uri=celex%3A32011L0016
• Richtlijn (EU) 2018/822 van de Raad van 25 mei 2018: https://eur-lex.europa.eu/legal-content/NL/TXT/?uri=CELEX%3A32018L0822

Operations

Data

• Personal data (addresses and relations between natural persons, tax liability)

Basic Registration of Persons (BRP)

• Company data (addresses and relations, tax liability, tax number)

Chamber of Commerce

• Financial products

Banks

• Income tax return details (IH)

Inland Revenue

• Supervisory data (tax audits)

Tax Office

• Corporate income tax (VPB) return details

Tax Office

• DAC-6 reports and statements

Tax Office

Technical design

The algorithm consists of decision rules created in collaboration with content experts. There are general decision rules and decision rules that apply to a substantial feature. A substantial feature is a feature that, according to the directive, indicates potential tax avoidance. This may involve looking at data from multiple tax resources at the same time, e.g. corporate tax and income tax. Based on the decision rules, the algorithm gives each report a risk score. For each creature characteristic, the notifications with the highest risk score are signalled to a Tax Administration employee who assesses the notification for the need for further investigation. As a result, follow-up actions can be taken, such as asking questions of the intermediary or dealing with the notification when assessing a return of the taxpayer included in the notification.

The algorithm is not self-learning. This means that it does not evolve during its use.

External provider

Internally developed