Algorithmic Governance: The Implications of the Case Matcher Algorithm for the Refugee Status Determination Process in The Netherlands

This paper focuses on the Case Matcher algorithm used by asylum decision-officers at the Dutch Immigration and Naturalisation Services (IND). The algorithmic support tool enables users to consult older cases with similar characteristics, allowing for comparison-based decision-making. Case Matcher aims to make the refugee decision process (RSD) quicker and reduce inconsistencies in asylum assessment. The increase in use of algorithms in public administration have resulted in the concept of Algorithmic Governance, a topic concerned with the decisions made by algorithms about refugees and asylum-applicants as a way for institutions to control movement using technology. With Case Matcher, Algorithmic Governance becomes increasingly relevant within Dutch government bodies. This paper therefore questions: How does Algorithmic Governance influence the decision-making process of IND asylum decision-officers in the Netherlands? From interviews conducted with 14 asylum decision-officers, the implications of Case Matcher for the work of employees were analysed through the lens of ethical risks, responsibility & accountability, and efficiency & effectiveness. While Case Matcher was expected to significantly alter credibility thresholds and make comparison the default mode of judgment, the findings reveal more complex implications. Participants highlight a strong sense of responsibility regarding case assessment and emphasise the need to remain critical and self-aware when consulting Case Matcher to compare cases. Frequent use reduces the perceived efficiency of refugee status determination, and workload is seen as minimally affected by Case Matcher. Ethical risks such as over-reliance and misinterpretation are at risk when employees fail to receive coherent training sessions. However, these ethical risks appear to be mostly an effect of the organisational environment of the IND. That is; ethical risks are dependent on how employees choose to interact with the tool and how the organisational environment of the IND frames its use. This research shows that Algorithmic Governance does not have to erode human judgement and critical thinking in migration governance but can be a way of governing where human agency, institutional culture and tool design intersect. These findings help to move the conversation of Algorithmic Governance in the context of RSD beyond technological determinism and credibility discretion towards one that is more closely related to organisational behaviour and the collaboration of humans and algorithms. Future research should decentre algorithms and instead research how the use of algorithms such as Case Matcher depends on how the algorithm is interpreted and institutionalised within migration government bodies like the IND.