Disclosure risk for firms in combined firm-datasets

Past studies have successfully shown that the level of anonymisation of scientific use files (SUF) is sufficiently high to protect against disclosure attacks that use data from traditional firm databases. However, with the increasing availability of online data about firms, new challenges for the provision of SUFs arise. In this paper, we therefore focus on a scenario, where an attack against the Mannheim Innovation Panel SUF is performed using data from the Mannheim Enterprise Panel and Mannheim Web Panel. We find that the disclosure risk of our attack is small and increases only slightly if data from the Mannheim Web Panel are considered. Data protection officers may use our findings when researchers want to publish SUFs.