In Reply to Tsai and Venkataramani
We thank Tsai and Venkataramani for their comments, which raise important methodological issues and challenge the findings presented in our article. They reconstructed the data set and stated that the association found between psychiatric bed numbers and prison populations disappears once they adjusted for year fixed effects or linear time trends. Yet, they did not use the same data set as in our study. They interpolated the data, ending up with 104 observations, almost twice as many as the 53 observations in our analyis.