Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.

Algorithm fine and transaction estimate Justice and Security

The algorithm predicts how many traffic fines the police issue in a year and how much of them the Ministry of Justice and Security collects. As a result, the ministry knows how many staff are needed and how much revenue is collected from fines.

Last change on 20th of August 2024, at 11:34 (CET) | Publication Standard 1.0
Publication category
Other algorithms
Impact assessment
Field not filled in.
Status
In use

General information

Theme

Public finance

Begin date

01-2021

Contact information

algoritmeregister.fez@minjenv.nl

Responsible use

Goal and impact

The algorithm helps make better estimates with fewer employees. Predictions previously differed greatly from the real numbers. And it took a lot of time to check where the differences were due to. The algorithm does not use personal data. The data shows the number of fines per offence (offence code) and the place of offence (offence location). JenV checks the accuracy and completeness of the algorithm. For example, by checking whether the figures used for the number of fines match the number of fines coming in. Annually, JenV determines which version of the algorithm it uses per enforcement tool (e.g. speed cameras, route checks or standing orders) and whether they need to adapt it to new policies.

Considerations

The algorithm helps make better estimates with fewer employees. Predictions used to differ greatly from the real numbers. And it took a lot of time to check where the differences were due to.

Human intervention

There is no automatic decision-making. Employees can look at the results and make decisions based on them. For example, adjust expected spending or allocate more staff. The algorithm provides Excel files stating that it is a forecast.

Risk management

The algorithm does not use personal data. The data shows the number of fines per offence (offence code) and the place of offence (offence location). JenV checks the accuracy and completeness of the algorithm. For example, by checking whether the figures used for the number of fines match the number of fines coming in before the model is trained. Annually, JenV determines which version of the algorithm it uses per enforcement tool (e.g. speed cameras, route checks or standing orders) and whether they need to adapt it to new policies.

Legal basis

The Directorate of Financial and Economic Affairs (DFEZ) is responsible for the financial management of the entire Ministry of Justice and Security (Article 4.1d and Article 4.1f of the State FEZ Decree).

Links to legal bases

artikel 4.1d en artikel 4.1f van het besluit FEZ van het Rijk: https://wetten.overheid.nl/BWBR0041910/2020-01-01

Operations

Data

Monthly reports provided by the CJIB were used, showing the number of fines issued per enforcement tool and agency over the past 5 years.

No personal data are used in this algorithm.

Technical design

A number of models are used to predict fines and transactions: The mean, last year and two time series algorithms (ETS and Arima). First, it is determined which model will be used. To determine which model to use, the different models are run over a full year for which the realisation figures are known. The model with the smallest deviation is finally chosen. To finally produce the estimate, the model finally selected is applied over the entire dataset (including the evaluation period used earlier). The result of this estimate is the so-called policy-neutral estimate. A number of further edits are then made to this, based on policy changes. In addition, a translation is made to revenues, which is done on the basis of an average fine amount. This amount is then adjusted based on established indexation.

External provider

Internally developed

Similar algorithm descriptions

  • The Directorate of Financial and Economic Affairs manages the finances of the Ministry of Justice and Security (JenV). To do so, it needs a good picture of expenditure and income. The 'probable course' algorithm predicts these.

    Last change on 20th of August 2024, at 11:26 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    DPIA
    Status
    In use
  • The algorithm is used by largest Dutch municipalities and determines, based on laws and regulations, whether there is a right to a benefit, type of benefit, amount and duration and gives this as advice to the counsellor.

    Last change on 5th of July 2024, at 9:00 (CET) | Publication Standard 1.0
    Publication category
    Impactful algorithms
    Impact assessment
    Field not filled in.
    Status
    In use
  • This algorithm helps Customs select goods for inspection based on risk. It uses declaration data from companies and considers whether or not there are risks of goods entering or leaving the European Union via the Netherlands infringing intellectual property rights.

    Last change on 10th of December 2024, at 8:57 (CET) | Publication Standard 1.0
    Publication category
    Impactful algorithms
    Impact assessment
    Field not filled in.
    Status
    In use
  • The algorithm is used to make an automated risk assessment for all Applications for Fixed Chargesgemoetkoming, prior to automated or manual granting and payment of the advance.

    Last change on 30th of May 2024, at 12:57 (CET) | Publication Standard 1.0
    Publication category
    High-Risk AI-system
    Impact assessment
    Field not filled in.
    Status
    In use
  • We use an algorithm to determine whether someone is entitled to AOW (General Old Age Pensions Act) and how much money that person will receive. This algorithm follows fixed rules.

    Last change on 28th of October 2024, at 10:25 (CET) | Publication Standard 1.0
    Publication category
    Other algorithms
    Impact assessment
    DPIA
    Status
    In use