Discrimination Against Immigrants in the Criminal Justice System: Evidence from Pretrial Detentions

This paper tests for discrimination against immigrants in pretrial detention decisions using administrative data from Chile (2008–2017). Immigrant defendants are 8.5 percentage points less likely to be released than Chilean defendants with similar proxies for pretrial misconduct risk. We use two approaches to show that this gap is not driven by unobserved differences in misconduct potential, including a novel test using observed pretrial misconduct rates of released defendants. We show evidence that discrimination arises from an informational problem: judges cannot observe criminal records from immigrants' countries of origin, leading them to treat immigrants with no prior record in Chile similarly to those with criminal histories. Consistently, outcome tests reject accurate statistical discrimination. Discrimination is largest for drug offenses and increased over time in parallel with a growing immigration wave.