Signal Model Sales Tax Large Enterprises (SOB GO)

Dutch Tax and Customs Administration

The algorithm 'Signal Model OB GO', hereafter abbreviated as SOB GO, helps Tax Administration staff to assess the risk of turnover tax returns that fall within the target group of Large Enterprises. About 7% of the returns involve returns by natural persons.

Last change on 26th of November 2024, at 15:24 (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://over-ons.belastingdienst.nl/onderwerpen/omgaan-met-gegevens/algoritmeregister/

Link to source registration

https://over-ons.belastingdienst.nl/onderwerpen/omgaan-met-gegevens/algoritmeregister/signaalmodel-omzetbelasting-grote-ondernemingen-sob-go/

Responsible use

Goal and impact

Almost every entrepreneur submits sales tax (OB) returns. Some of the entrepreneurs who are required to file OB returns in the Netherlands fall under the supervision of the Large Enterprises Directorate.

Since 1 October 2024, the employees of the Large Companies Directorate have been supported by SOB GO in assessing these OB returns.

The algorithm detects possible discrepancies in the returns based on relevant data available to the Tax Administration, and on business rules formulated by experienced staff.

The purpose of the algorithm is to help the Tax Administration employee with advice on which returns are potentially risky and therefore may need to be handled manually. In addition, the algorithm provides information for (in particular) client coordinators, who are the contact persons for large businesses at the Tax Administration. This information is important for the client picture that the Large Enterprises Directorate carefully builds up. The actual decision to assess a return, or to use information in client handling, lies with the specialised staff and client coordinators.

Considerations

The algorithm supports Tax Administration staff with advice on which returns may require manual processing. This makes better use of employees' available capacity. Employees need to spend less time on declarations that are (probably) completely correct and do not need to be checked. To keep track of the returns that are (probably) fully correct as well as to check the business rules used, samples are also taken from the returns that are not selected for assessment.

The sales tax return process has become more effective with the use of the algorithm. As a result, businesses get faster feedback on incorrect returns. In addition, the algorithm provides signals that help in individual client handling at Large Enterprises.

Human intervention

Human intervention means that human oversight of decision-making is substantial and carried out by someone who is competent and knowledgeable.

The operation of the algorithm involves human intervention, but decisions are also made by the algorithm. The algorithm detects, selects and makes decisions. In situations where the algorithm cannot make the decision (more complex situations or deviation from the declaration), there is human intervention by an employee.

Risk management

The use of the data is reviewed against the General Data Protection Regulation (AVG). This review 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 and Customs Administration regularly examines whether the data used are still necessary and may therefore be used.

The algorithm was developed in-house at the Tax Administration and is also maintained internally. Every year, the business rules are evaluated, also taking into account the results of the sample declarations that, among other things, serve to assess the correctness and completeness of the business rules used. If necessary, the rules are adjusted.

Evaluation is carried out by the project group/product owner in coordination with the regional Subject Group OB GO Coordinators.

The selection rules and algorithm are tested against non-discrimination legislation. 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 prejudice.

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

Conditions, a quality framework, have been drawn up by the Tax and Customs Administration for the development of algorithms. This contains rules and agreements that are followed during algorithm development. The conditions of the Audit Service Rijk (ADR) are leading in this respect. At regular intervals, the Tax and Customs Administration checks whether the algorithm still meets the quality requirements.

The algorithm uses data from various tax laws. As required by the AVG, no more data is used than necessary.

Legal basis

• General State Tax Act
• General Administrative Law Act
• General Data Protection Regulation
• General Data Protection Regulation (Implementation) Act
• Turnover Tax Act 1968
• General provisions Citizen Service Number Act
• Archives Act 1995

Links to legal bases

• Algemene wet inzake rijksbelastingen: : https://wetten.overheid.nl/BWBR0002320/
• Algemene wet bestuursrecht: : https://wetten.overheid.nl/BWBR0005537/
• Algemene verordening gegevensbescherming: : https://eur-lex.europa.eu/legal-content/NL/TXT/HTML/?uri=CELEX:32016R0679
• Uitvoeringswet algemene verordening gegevensbescherming: : https://wetten.overheid.nl/BWBR0040940/
• Wet op de Omzetbelasting 1968: : https://wetten.overheid.nl/BWBR0002629/
• Wet algemene bepalingen Burgerservicenummer:: https://wetten.overheid.nl/BWBR0022428/
• Archiefwet 1995: : https://wetten.overheid.nl/BWBR0007376/

Operations

Data

The following data are entered into the algorithm:

• Personal data;
• Turnover tax (OB) return data and assessment data;
• Corporate income tax (VpB) return data and assessment data;
• Recovery data.

The algorithm does not use any special personal data.

Links to data sources

• Persoonsgegevens : Basisregistratie Personen (BRP)
• Aangiftegegevens en aanslaggegevens Omzetbelasting (OB): Belastingdienst
• Aangiftegegevens en aanslaggegevens Vennootschapsbelasting (VpB): Belastingdienst
• Gegevens Invordering: Belastingdienst

Technical design

The algorithm is a model based on business rules. These business rules are taken from legislation, content expertise and statistically compiled by content experts. Through these predefined rules, the algorithm detects possible anomalies in the declaration. The algorithm divides the signals into different priority categories. These priority categories help employees determine when and by whom which signal is handled.

A business rule can look like this: 'An X amount of negative turnover has been declared in section X in the past X returns'.

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

External provider

The algorithm was developed in-house at the Inland Revenue and is also maintained internally.