The Economics of Discrimination in the Court System: Police, Technology, and Their Interaction

Abstract

This dissertation consists of three essays which utilize automated traffic enforcement data to investigate the existence of police discrimination in issuing speeding tickets and potential crime reduction as a secondary effect of using such programs. In the first chapter, I use tickets issued by automated traffic enforcement cameras as a measure of the population of speeders to compare with police-issued tickets. The novel dataset has an advantage over previous literature because data collection was not a result of suspected police bias. I find that a ticketed individual is more likely to be African-American and more likely to be female when ticketed by police as opposed to an automated camera. Though this implies some form of discrimination based on gender and race, it cannot be determined whether police are engaging in statistical or preference-based discrimination. Next, I extend the research question to determine whether the differential treatment of women and African-Americans by police should be characterized as preference-based or statistical discrimination. I use a detailed individual level dataset which follows individuals through the court process from receipt of a speeding ticket to trial. It seems that police are not engaging in statistical discrimination, because women and African-Americans are no more likely to immediately pay a speeding ticket. In fact, since African-Americans are actually more likely to attend a trial, police are targeting individuals who will utilize more court resources: contradictory to one motive of statistical discrimination. Individuals behave differently based on which judge they are assigned, but judges do not seem to be issuing fines discriminatorily. The final chapter aims to answer a different question regarding automated traffic enforcement: do automated traffic programs reduce crime? Many cities and companies which implement the automated systems cite crime reduction as a byproduct of adoption. They claim that these programs actually reduce crime rates by enabling police to focus on more serious offenders, increasing the marginal productivity of police. This is the first research to rigorously investigate these claims, and I find some supportive evidence, however, it seems that these companies may be exaggerating the extent of this effect.