Safely opening Pandora’s box: a guide for researchers working with leaked data

Abstract The use of leaked data is becoming increasingly common in empirical research, particularly in public finance. Although these data can be an enormously powerful tool for investigating behavior that is otherwise difficult to measure, their use also generates substantial ethical and legal risks. In this paper, we (i) present the growing body of social science research that relies on leaked data, (ii) discuss the ethical, legal and privacy hurdles faced by projects relying on such data and (iii) offer a practical roadmap for researchers looking to enter the space.