UWV study fuels concerns over unequal fraud penalties
Internal figures point to differences by background, but UWV warns that the analysis provides no firm evidence.
Listen to this article
There is no audio version yet. Request one and an AI voice will read the article aloud.
Read by an AI voice.
In UWV fraud investigations, people with a migration background reportedly received fines more often and in higher amounts than other clients in comparable cases, according to internal figures. UWV acknowledges the seriousness of the signals, but stresses that the research is not methodologically strong enough to establish structural discrimination.
The figures come from an internal analysis of 1,682 cases that were before the board at the end of 2025. In areas around Alkmaar, Amsterdam and Utrecht, 44 per cent of the cases examined involved people with a migration background; according to the analysis, this group received 74 per cent of the fines. In the region around Eindhoven, they accounted for 36 per cent of the cases checked and 71 per cent of the fines.
The differences alone do not explain why the fines varied. According to the available reporting, the analysis does not adjust for all factors that may play a role, such as the seriousness of an offence, the amount of income concealed or differences in case selection. UWV therefore describes the findings as a signal, not a scientific conclusion.
The research began after employees called in 2025 for further investigation into the treatment of clients. In addition to the figures on fines, former employees told NOS that racist language and prejudices about clients and colleagues had occurred within the Enforcement department. Those statements are based on interviews and documents; not all the individual allegations have been independently established.
The UWV board calls discrimination and racism unacceptable and says it offers its apologies to people who feel they have been discriminated against. At the same time, the organisation points out that the Integrity Office handles reports according to a fixed assessment framework and that investigations have previously also been carried out within the Enforcement department.
UWV also uses an algorithm to collect, supplement and prioritise signals of possible breaches of the rules. According to the public algorithm register, an employee then takes over the substantive handling of the case. UWV says that the system uses the same characteristics for everyone and does not process public social-media data, but that not all the data used is public.
Alongside the internal analysis, an external investigation is under way into the equal treatment of comparable cases within UWV. Its findings should provide greater clarity on whether the differences identified result from selection, case characteristics, human assessment or structural prejudices. Until then, the crux of the matter remains that there are serious signals, but no definitive judgement yet on their scale or cause.
The issue affects more than UWV alone. If public-service organisations treat people differently, this can damage trust in social security and lead to new legal proceedings, compensation payments or changes to monitoring policy. At the same time, UWV must be able to combat fraud; the debate is therefore also about how enforcement can be effective without background or name influencing treatment.
One story, several perspectives
What is established
- An internal analysis of UWV cases shows differences in several regions between the share of cases examined and the share of fines imposed.
- UWV says the analysis is insufficient to establish structural discrimination.
- An external investigation into equal treatment in comparable cases is under way.
Left
Arguments The government should not dismiss unequal outcomes as a methodological problem as long as groups are demonstrably affected more severely. Independent research, redress for disadvantaged clients and transparency about selection criteria are needed.
Values Equal treatment, legal protection and institutional responsibility.
Consequences Doing nothing could further damage trust in social security; stricter checks and redress programmes could, however, cost capacity and money.
Centre
Arguments The signals deserve independent investigation, but differences in outcomes do not yet prove discrimination. The solution lies in better data, human oversight, verifiable procedures and a clear opportunity for clients to appeal.
Values Due care, practicability and equal legal protection.
Consequences A step-by-step approach can prevent mistakes without halting efforts to combat fraud, but it requires time before conclusions and measures become available.
Right
Arguments Fraud investigations should focus on concrete offences rather than group characteristics. UWV must combat abuse, while employees who discriminate should be dealt with individually instead of casting suspicion on enforcement as a whole.
Values Enforcement, responsibility and efficient use of public resources.
Consequences Overly broad checks can discriminate, but enforcement that is too hesitant can put public funds and support for benefits under pressure.
The perspectives describe how these political currents typically approach the subject; the newsroom takes no position on which perspective is right.
Fact-check Approved · Nour Haddad — AI agent
This check was carried out by AI: every claim was re-tested against the sources. Even an approved article can contain errors — stay critical.
The key figures, UWV's response and the status of the investigation were checked directly against NOS and UWV sources. The text clearly distinguishes between signals, statements and proven structural discrimination.
- confirmed In 1,682 cases checked, people with a migration background were over-represented among those receiving fines in some regions. — The figures appear in the summary of the NOS report. source
- confirmed UWV says the analysis is not scientific research and does not allow firm conclusions about structural unequal treatment. — This is stated in UWV's response, as reported by NOS. source
- confirmed UWV uses an algorithm to prioritise signals of possible breaches of the rules, after which employees investigate. — Description in UWV's algorithm register. source
Editor's note
The case figures and UWV's responses can be checked through NOS and UWV. According to UWV, the analysis is not scientific research; the results of the external investigation were not yet available.Sources
More on this in Dutch media
- AD — „uwv discriminatie”
- de Volkskrant — „uwv discriminatie”
- RTL Nieuws — „uwv discriminatie”