Two individual member states – France and Romania – collaborate to identify, arrest, detain and deport thousands of EU citizens within EU territory. While France operates a complex and massive deportation apparatus, the Romanian state has openly conceded to receive its deported citizens and to assist the French authorities in policing irregularised Romanian citizens, mainly of Roma ethnicity. Based on multi‐sited research conducted in Romania and France, this article scrutinises the challenges of ethnographic research within a shared deportation apparatus. It analyses the site of one‐to‐one collaboration between the two EU member states, which offers a unique perspective to make the state readable to researchers. In doing so, I document and expose the ways in which both states granted (or not) access to study state institutions and practices. The researcher's perceived identity and the states’ political interests contributed to the negotiations for access, never fully awarded and always partial. At the same time, access for research is conditioned by the practices, discretion and ambiguity of agents working for one state or another. I thus argue that within the present environment of securitisation and criminalisation of mobility, the police collaboration for deportation greatly informs the research and findings about the state and its institutions.