« Back to Results

Societal Impacts of Policing and Incarceration

Paper Session

Sunday, Jan. 5, 2020 10:15 AM - 12:15 PM (PDT)

Marriott Marquis, Grand Ballroom 2
Hosted By: American Economic Association & Committee on the Status of Women in the Economics Profession
  • Chair: Jennifer Doleac, Texas A&M University

Baby's Gone: The Effect of Increased Sentencing Severity on Fertility and Family Formation

Siobhan O'Keefe
Davidson College


The United States' incarceration rate has quintupled since 1970. By 2007, one percent of the adult population was incarcerated, and over 90% of those incarcerated were men. Since incarcerated men are physically separated from their communities, their absence may affect those who are left behind. Using the change in incarceration caused by a sentencing reform in North Carolina, together with an intensity of treatment research design, I show that incarceration policies have spillover effects on family formation patterns. In the wake of the policy change, unmarried and young black women reduce their fertility, and the composition of births shifts towards women of higher socioeconomic status. At the same time, I find that among those who gave birth, the quality of partner matches declines. White women are less likely to be married, but there is no effect on marriage rates of black women.

Housing Security and Domestic Violence: Evidence from Nuisance Ordinances

Sarah Kroeger
University of Notre Dame
Giulia La Mattina
University of South Florida


The majority of acts of violence against women, including homicides, are committed by intimate partners or ex-partners. Domestic violence imposes a burden on society by reducing victims’ physical and psychological wellbeing as well as their productivity in the labor market. Despite the heavy costs, the causes of domestic violence are still poorly understood. Previous research has uncovered some of the factors that affect domestic violence, including women’s potential wages relative to men, unilateral divorce laws and emotional cues. However, the relationship between housing security, homelessness and domestic violence has not been well disentangled. Domestic violence can be a cause of homelessness but also a consequence, if for example fear of homelessness reduces a victim’s ability to escape the abusive situation or to report incidents to the police. In this paper, we estimate the causal impact of local nuisance laws on reporting and actual incidence of domestic violence. Criminal activity nuisance ordinances (CANOs) penalize property owners when certain types of activity occur on or near their property, reported primarily through 911 calls to a residence. A large body of qualitative research links CANOs directly to tenant eviction. This eviction threat may create a strong disincentive for tenants to call for police assistance, even when their own physical safety is at risk. Our analysis examines the first order effect of CANOs on eviction risk, and the effect of CANOs on intimate partner violence reporting and incidence rates. Our data include police-reported municipal level crime incidents, municipal level evictions data, and CANO legislation for cities and towns in Ohio collected from municipal websites and municipal town hall offices.

In-group Bias and the Police: Evidence from Award Nominations

Nayoung Rim
United States Naval Academy
Bocar Ba
Duke University
Roman Rivera
Columbia University


This paper examines the impact of in-group bias on the internal dynamics of a police department. Prior studies have documented racial bias in policing, but little is known about bias against officers due to lack of available data. We construct a novel panel dataset of Chicago Police Department officers with detailed personnel information. Exploiting quasi-random variation in supervisor assignment, we find that white supervisors are less likely to nominate black officers than white or Hispanic officers. There is weaker evidence that male supervisors are less likely to nominate female officers than male officers. We explore several theories of discrimination that can explain our main findings. Requiring interaction between supervisors and officers reduces the minority nomination gap, but white supervisors still exhibit in-group favoritism. Our findings suggest departments should focus on policies that address in-group bias due to its effect on career advancement.

Do Arrests Deter or Promote Offending?

Emily Weisburst
University of California-Los Angeles
Felipe Goncalves
University of California-Los Angeles


Arrests are the primary entry point to the criminal justice system and they are commonplace; over 10
million arrests are made each year in the U.S. (UCR, 2018). This paper estimates the impact of an arrest
on the criminal outcomes of arrested individuals, an area where there has been limited quasi-experimental
research. Arrests may increase public safety by deterring future crime or incapacitating repeat offenders.
Alternatively, arrests resulting in convictions may decrease an individual’s ability to maintain employment
or receive public assistance (e.g. Pager, 2003; NICC, 2019). Even without a conviction, an arrest may have
negative consequences, as employers often have access to public arrest records and many arrests lead to
pretrial jail stays that disrupt employment.
We exploit unique data covering arrestees and individuals suspected of crimes but not arrested between
2007-2018 in Dallas, Texas. The data links dispatched 911 calls to responding officers, criminal suspects,
arrests, and jail stays. We plan to extend this data to include wage, employment and public benefits
In our first estimation approach, we use an event study design to measure the impact of an arrest. In our
second approach, we instrument for arrests, using the arrest propensity for the officer that responds to the
incident. This strategy is similar to the economics of crime literature using judge leniency as an instrument
for conviction, pretrial release and sentence (e.g. Dobbie et al., 2018; Aizer and Doyle, 2015; Mueller-Smith, 2015). We address selection into the suspect sample by estimating a selection correction term using the
responding officer’s propensity to record a suspect among incident responses.
In preliminary analysis, we find that arrests lead to a short-run reduction in criminal activity. This
effect is specific to jailed arrestees, suggesting the impact is due to incapacitation rather than a behavioral
Anne Piehl
Rutgers University
Erdal Tekin
American University
Benjamin Hansen
University of Oregon
Jennifer Doleac
Texas A&M University
JEL Classifications
  • K4 - Legal Procedure, the Legal System, and Illegal Behavior
  • J1 - Demographic Economics