Establish regulatory capture phenomena.

Summary The article sheds light on the phenomena of regulatory capture by examining an empirical configuration where quantitative results are available for a large sample, and where we seek to identify causal combinations by examining a small number of cases qualitatively. A Bayesian interpretation of process tracing is then proposed and applied to the study of the genesis of drug regulatory agencies in France and England. The approach allows the identification of institutional dependencies that are difficult to capture by the initial statistical analysis. They explain, more favorably than the capture thesis, the orientation of the regulatory schemes studied.