Arresting Developments
The Racial Politics of Mass Incarceration
John Clegg & Adaner Usmani
Perspectives on Politics, forthcoming
Abstract:
Many argue that America’s punitive turn was the result of racial backlash to the Civil Rights Movement. Yet some have noted support among black people for the policies attributed to this backlash, citing the influence of rising crime on black voters and politicians. In this article we gather new evidence and examine what it implies. Public opinion data show that not just the white but also the black public became more punitive after the 1960s. Voting data from the House show that most black politicians voted punitively at the height of concern about crime. In addition, an analysis of federally mandated redistricting suggests that in the early 1990s, black political representation had a punitive impact at the state level. Together, our evidence suggests that crime had a profound effect on black politics. It also casts some doubt on the conventional view of the origins of mass incarceration.
When insignificance is significant: Rethinking race, immigration, and the myth of victim reluctance to report to police and use victim services
Hyunjung Shim & Sarayu Cheemalapati
Journal of Criminal Justice, November–December 2025
Abstract:
Public discourse often portrays victims of racial, ethnic, or immigrant minorities as more reluctant to engage in police or seek further help. Yet, this perception remains largely unexamined through empirical research. This study tests the correlates of police notification and victim service utilization among victims of violence, with particular focus on how race/ethnicity and immigration status — key components of sociostructural positioning — interact. Drawing on the Multilevel, Contextualized Help-Seeking Model, we analyzed data from the National Crime Victimization Survey (NCVS) from 2017 to 2023. We estimate multilevel logistic models of police reporting and service use, while accounting for selection bias in the victim sample. Findings reveal that Black victims generally had higher odds of notifying police than White victims. However, this trend reverses for Black foreign-born citizens, who were significantly less likely to report. Asian non-citizens, conversely, were nearly ten times more likely to use victim services than White citizens. Weapon involvement was linked to higher odds of police reporting, while incident severity was associated with greater service use. Series-victimization was associated with lower odds of police reporting, but higher odds of service use. These results challenge the prevailing assumption that racial and ethnic minorities are uniformly less likely to seek help and suggest that targeted public policy solutions can effectively promote help-seeking, especially among immigrant populations.
Migration and the persistence of violence
Martin Vinæs Larsen, Gabriel Lenz & Anna Mikkelborg
Proceedings of the National Academy of Sciences, 2 December 2025
Abstract:
Using data on millions of internal US migrants, we document that historical homicide rates follow migrants around the United States. Individuals born in historically safe states remain safer wherever they go, while individuals born in historically dangerous states face a greater risk, including from police violence. This pattern holds across demographic characteristics such as age, gender, and marital status, across migrant groups with different average levels of education, income, and even when comparing migrants from different states who reside in the same county. To help understand why, we conducted a large national survey that oversampled internal White US migrants. The results suggest this persistence may reflect a sociocultural adaptation to dangerous settings. Residents and migrants from historically unsafe states — mainly former frontier states and the deep South — see the world as more dangerous, react more forcefully in aggressive scenarios, value toughness, distrust law enforcement, and say they rely on self and family in violent situations. These adaptations may have kept them safe in historically dangerous states, but may increase their vulnerability to harm in safer states.
Not a Sip: Effects of Zero Tolerance Laws on Road Traffic Fatalities
Andres Ramasco
Health Economics, forthcoming
Abstract:
A substantial proportion of alcohol related fatalities and their consequences are preventable, prompting policymakers to implement measures aimed at reducing these deaths. I exploit time and geographic variation in the adoption of zero-tolerance laws in a difference-in-differences design to study the impact of these regulations on traffic-related incidents. Using county-level data, I find no sizable reductions in fatalities and an increase in injury counts after the adoption of such laws. I do not find significant changes in several measures of alcohol consumption, consistent with the lack of reduction in driving fatalities.
Police department design, political pressure, and racial inequality in arrests
Andrew McCall
American Journal of Political Science, forthcoming
Abstract:
This paper theorizes a source of bias in discretionary arrests: strategic limits on police officer learning. Officers have a variety of tactics at their disposal besides arrest that they use for less serious offenses when they judge the underlying behavior to be less severe. In departments led by a chief with special expertise in crime control, the chief's directives to change the severity threshold at which officers make arrests are a source of information about the most effective practices. However, if officers are uncertain whether their chief is swayed by political pressures, those directives will be less persuasive, especially when they align with what influential advocates want. This mechanism represents a constraint on the effectiveness of departments where chiefs have limited means to force subordinate compliance. With increasing Black political influence in cities since World War II, this inefficiency would have generated a form of anti-Black structural racism in policing.
No sales after midnight: Evaluating the impact of a business curfew on drug-related crime in San Francisco’s tenderloin
Mirko Nazzari, Marco Calaresu & Moris Triventi
Security Journal, November 2025
Abstract:
Business curfews are emerging as regulatory policy instruments to reduce crime in high-risk areas, yet rigorous evaluations remain limited. This study examines San Francisco’s Tenderloin Retail Hours Restriction Pilot, which required select businesses to close from 12:00 a.m. to 5:00 a.m. starting July 2024. Using a customized Bayesian Structural Time Series model, we estimate a 56% reduction (95% credible interval: −72% to −27%) in drug-related incidents during curfew hours over nine months, with no evidence of spatial displacement to nearby areas or temporal displacement within the Tenderloin Public Safety Area. Results hold under Causal-ARIMA sensitivity tests. Findings suggest curfews may reduce opportunities for street-level drug activity, but potential economic costs and questions about long-term sustainability underscore the need for careful policy design.
Vulnerability of the Public Safety System: Evidence from Micro-Shocks
Shooshan Danagoulian et al.
NBER Working Paper, October 2025
Abstract:
The 911 system, a vital component of public safety, responds to over 210 million emergency calls each year. We examine the effect of small exogenous public health stressors — pollen emissions — on the operation of the 911 call system in Seattle, Washington. Leveraging daily variation in pollen levels, we evaluate pollen’s impact on the volume of 911 calls. We find that a 100% increase in pollen levels — the modal day-over-day variation we observe in our data — results in six additional 911 calls per day. We further show that elevated pollen levels influence the quality of emergency and non-emergency response. The 911 system responds to increased strain by prioritizing more urgent calls — and as a result, increasing response time to non-urgent calls. Though all calls are connected to a dispatcher, there are increased missed connections — canceled emergency response, missing records of response, and false reports. Despite the prioritization of urgent calls, these additional daily calls lead to reallocation of resources in ways that potentially impede service quality. Thus, our findings suggest that common and small shocks may lead to under-provision of valuable services to communities across the U.S.
Can I See Some Identification? The Impact of Receiving Identification on Recidivism Upon Release From Prison
Elizabeth Steffensmeier
Social Science Quarterly, November 2025
Methods: Using individual-level data from the National Corrections Reporting Program (2000–2020), the effects of state-level identification policies on recidivism were estimated using a difference-in-differences design. The analysis exploits variation in the timing and type of identification-related legislation across states.
Results: States that have mandates to provide an identification card to inmates lacking one upon release or a policy in place stating that an inmate may receive assistance with obtaining identification experience statistically significant reductions in recidivism: a 7.7% decrease within 3 years, 8.6% within 5 years, and 7.7% within 8 years. Policies that mandate the provision of identification upon release show similar results for recidivism within 5 and 8 years.
Predictors of Successful Resolutions and Avoidance of Harms in Crisis Negotiations: An Analysis of the FBI’s HOBAS Database from 1982 to 2023
Steven Pace et al.
Journal of Police and Criminal Psychology, December 2025, Pages 803-815
Abstract:
Crisis negotiation incidents can be emotionally charged events and often prove to be highly dynamic. Using the Hostage and Barricade Database System (HOBAS), we used correlations and logistic regressions to examine the situational characteristics that predicted successful negotiation resolution and successful avoidance of violence during crisis negotiations. Characteristics related to incidents were typically more predictive than those related to the subjects. Situational characteristics that reduce the likelihood of successful negotiations include weapons or property damage to the scene. Communication initiated by a trained responder or conducted with a technological voice uniquely predicted a successful negotiated resolution. Communication with a bullhorn predicted less success in negotiations and more successful tactical resolution. When trained negotiators started communicating with a subject, they were more likely to result in successful negotiations; in contrast, when an incident commander routinely trained with SWAT, violence was more likely to occur after the onset of an incident. These findings suggest that training is a cornerstone of crisis negotiation and, although tactical responses may be appropriate in certain contexts, negotiators must know the situations in which it should be applied.
Street-Level Bureaucrats and Political Control: Do Los Angeles Police Change Their Behavior in Response to Policy Changes?
Nicholas Weller et al.
University of California Working Paper, June 2025
Abstract:
Street-level bureaucrats wield broad discretion, and elected principals periodically try to limit it. We test whether such attempts constrain agent behavior. We study a 2022 Los Angeles Police Department policy change that narrowed officers' discretion in conducting "pretextual" stops. Using both regression discontinuity in time and synthetic differences-indifferences methods, we estimate that in the short-term the policy reduced pretextual stops, but the effect disappears within a little more than a year. We then turn to interviews with LAPD officers and supervisors to better understand these empirical patterns. Officers describe misunderstanding the scope of the policy, an absence of sanctions for non-compliance, and learning and adaptation that led pretextual stops to return to pre-policy levels. These findings illustrate the difficulty in exerting political control over street-level bureaucracies and therefore the limits of reform attempts.