This paper replicates the results in Schularick and Taylor (American Economic Review 2012; 102(2): 1029-1061; ST hereafter). Specifically, I replicate ST's results in the narrow sense by reproducing their calculations in the open source econometrics package gretl. (Gretl is an acronym for Gnu Regression, Econometrics and Time-series Laboratory. It is available for Windows, Mac and Linux at www.gretl.sourceforge.net.) I also demonstrate the robustness of ST's findings to different estimation methods. I obtain qualitatively similar results to ST via Bayesian estimation of both static and dynamic panel probit models. Finally, I show that the marginal effects of credit growth on the probability of a financial crisis vary considerably across the countries in the dataset.
