Who gets caught?

Some people are more likely to be convicted of a crime than others. Some argue that differential arrest and conviction rates largely reflect differences in the rates at which groups commit crimes. Others argue that law enforcement agents discriminate against certain groups. In this paper, we set out a simple model of 'statistical discrimination' where law enforcement agents use information on personal characteristics to reduce search costs. We then use a novel procedure and data set to test whether or not police overuse information on individuals. We find that a person's characteristics do indeed influence the probability of appearing in court, while a number of forms of delinquent activity have no or even negative impacts in court appearances.