Lessons to Be Learned from the Dutch Childcare Allowance Scandal: A Comparative Review of Algorithmic Governance by Tax Administrations in the Netherlands, France and Germany

Abstract

In January 2021, the cabinet of Mark Rutte, Prime Minister of the Netherlands resigned after the revelation that the Dutch tax administration had infringed fundamental rights of an estimated 35,000 welfare recipients. The Dutch childcare allowance scandal or toeslagenaffaire unveiled the ways in which the Dutch tax administration made use of artificial intelligence algorithms to discriminate against and violate the rights of welfare recipients. The Netherlands is not the only EU Member State using artificial intelligence for the selection of taxpayers for audits. Accordingly, the toeslagenaffaire can offer valuable lessons for the protection of taxpayers’ rights in the context of algorithmic governance in other Member States. Following the introductory section 1 of this article, section 2 outlines the important elements of the toeslagenaffaire and identifies qualitative requirements for the protection of taxpayers’ rights. In section 3, a joint comparative study of French and German algorithmic governance is conducted to determine whether lessons were learned from the toeslagenaffaire.

Keywords: Artificial Intelligence, Discrimination, Tax Policy, Machine-learning, Machine Bias

JEL Classification: K34; K42