Narrow and wide replication of Chalfin and McCrary (REStat, 2018)

We undertake a narrow and wide replication of "Are US Cities Underpoliced? Theory and Evidence" by Chalfin and McCrary. Using data from medium to large cities in the United States from 1960 to 2010, the authors estimate the effect of police on crime. To correct for the presence of measurement error, they propose to combine the information from two proxies of the police variable within the generalized method of moments framework. Throughout our replication exercise, we find that, in general, the original results are robust to computation in R compared to STATA and to the inclusion of more recent data (until 2019).