Algorithms and child benefit

By using algorithms, we can arrive at automated decisions on whether or not you are entitled to child benefit. We want to make it easier for you by paying your child benefit quickly and correctly first time.

When a child is registered with the municipal population register, the ‘gemeente’ sends us the details of the child and their parents. These details are combined by an algorithm. This shows us whether we need to send you a claim form for child benefit. If you do not qualify for child benefit, we will not send you a claim form.

After child benefit has been awarded, algorithms are used to ensure the child benefit payments are sent automatically each quarter. Simple changes in the family situation such as moving to a new address within the Netherlands or the birth of a second child are processed automatically.

The use of algorithms means that 90% of our pension and benefit decisions are fully automated. We do not need as much information from you and the costs per pension or benefit are lower.

If we do not calculate your child benefit correctly, it could lead to financial problems. You might find that you do not have enough money to live on, or that you have received too much and have to pay it back. These are difficult situations that can result in debt. By using automated data processing and algorithms, we can avoid human error and prevent debt.

Algorithms help us to predict the total cost of our pensions and benefits for the national budget. A good algorithm reduces the number of client cases processed manually. This saves money and gives us more time for individual cases where necessary.

The rules and legislation governing each of our pension and benefit schemes have been converted into a set of questions. The algorithms go through a question set step by step to arrive at a correct decision. We programme the questions in our system ourselves.

Our algorithms are prescriptive. This means decisions are taken automatically. We check the algorithms and make any necessary improvements. They cannot do this themselves.

The data we process is very sensitive. That is why our most experienced staff help develop the algorithms. We also have client panels to give us feedback during the development stages.

In addition, we have a very detailed set of internal guidelines on how to process data. We regularly check and cross-check to make sure we are using the data securely and keeping to the rules.

Our main sources of data are: 

  • the Dutch municipal population registers
  • the insurance records we keep in connection with the Dutch national insurance schemes.
  • international records systems for clients living outside the Netherlands