Electoral cycles in macroprudential regulation : a replication of Müller (2023)

Müller (2023) presents evidence for electoral cycles in macroprudential policy in a sample of 58 countries from 2000 through 2014. Consistent with theoretical arguments, the pattern of looser regulation is larger when election outcomes are uncertain and institutions are weak. In this replication, we first conduct a fully successful computational reproduction, using the provided replication package. We then subject the paper's main results to a series of robustness tests, involving measuring the dependent variable differently, bootstrapping standard errors, and applying different specifications of the main estimations. We also use new data, extending the covered time period, and re-examine the results. We find that the main results are robust to our robustness tests, but vanish using newer data. In an additional analysis, we provide suggestive evidence that the original results are based on rather limited variation in the dependent variable.