Episode Summary
Executive Summary: Jennifer Doliak argues that U.S. criminal justice problems are driven by excessive police discretion, weak accountability, racial bias, and overreliance on punishment. She emphasizes evidence-based reforms that improve trust and prevention—better training, diversity, non-police responders, lead abatement, CBT, lighting, and DNA databases—while warning against overstating weak evidence for popular reforms like body cams or Eight Can’t Wait.
Main Topics: Core failures in U.S. policing: The conversation centers on unnecessary escalation, racial disparities, weak accountability, and broad police discretion that can enable misconduct and distrust. Evidence on racial bias and discretion: Doliak highlights studies showing racial bias in stop-and-search, traffic enforcement, and judicial decisions, especially where officers or judges have discretion. Reform ideas that may work: Promising reforms include procedural justice training, increased diversity in policing, stronger accountability, clearer rules limiting discretion, and better recruitment. Crime prevention outside policing: The discussion emphasizes interventions that reduce crime before it happens, including CBT, lead abatement, lighting, and financial/educational support. Limits of popular reforms and evidence standards: She critiques oversold claims around body cameras and Eight Can’t Wait, arguing for transparency, humility, and post-implementation evaluation rather than false certainty. Police unions and institutional barriers: Police unions are presented as a major obstacle to discipline, firing bad actors, and implementing reforms, though political dynamics may be shifting. Diversity and representation in economics: The interview briefly shifts to the economics profession, where Doliak argues that more women and people of color would improve research questions, culture, and policy relevance.
Key Arguments: Police problems are not just isolated 'bad apples'; they are often structural and culturally pervasive, especially where broad discretion is exercised without accountability. Racial bias is easiest to detect in low-discretion settings; the more discretion officers or judges have, the more likely disparities emerge. Procedural justice training appears surprisingly effective, reducing complaints and use of force in Chicago in a randomized rollout. Increasing police diversity historically reduced victimization and improved reporting, especially among Black residents and women dealing with domestic violence. Body-worn cameras are not a reliable fix; across many RCTs, effects are often null and sometimes negative. Eight Can’t Wait and similar campaigns may have sensible policy proposals, but they overclaimed causal evidence. Hiring more police can reduce crime, mainly through deterrence, but the social costs of aggressive policing and mistrust may offset some benefits. A large share of crime prevention is likely cheaper and more humane outside the criminal justice system, especially via CBT, lead reduction, lighting, and early-life supports. DNA databases can deter reoffending substantially by increasing perceived detection risk. Police unions and rigid bargaining structures make it hard to discipline or remove officers, limiting reform impact. Local variation across thousands of police departments is a research opportunity and may speed learning about what works. Economics and social science should be transparent about uncertainty; good policy evidence requires clear counterfactuals, not just complex models.
Data Points: People killed by police annually in the U.S.: around 1,000+ per year - Used as part of discussion of the visible tip of the iceberg of police violence. People incarcerated in the U.S.: over 2 million - Cited in the discussion of mass incarceration and the limits of drug decriminalization as a solution. Black driver / white officer disparity in write-ups: much more likely to be written up for low-level offenses - From Jeremy West’s highway crash study using random dispatch of officers. Procedural justice training rollout: 8,000+ Chicago officers trained 25 at a time monthly - Staggered rollout enabled a natural experiment in training effects. Procedural justice training impact: reduced complaints and use of force - Chicago study found significant beneficial effects after a one-day training. Body-worn camera evidence base: over a dozen randomized controlled trials - Summarized as mostly null overall, with some positive and some negative findings. Police effectiveness on crime: 1% increase in police can reduce crime by roughly 0.1% to 1% - Rule-of-thumb from the literature, especially Chelin and McCrary. DNA database effect in Denmark: 43% reduction in reoffending - Result from expansion of the DNA database to broader offender groups. DNA database effect in the U.S.: 17% reduction in new conviction within five years for serious violent offenders; 6% for serious property offenders - Estimated from database expansion thresholds. Black seller discrimination online: 18% fewer offers; offers 11% lower - A prior paper on discrimination in second-hand goods markets. Ban the Box result: increased discrimination against African-American men with high school degree or less - Policy intended to help ex-offenders get jobs but had unintended consequences. Street lighting impact: at least 36% reduction in nighttime outdoor crimes - Referenced from an episode discussing bright lights in crime hotspots. Diversity and policing outcomes: Black victimization fell; domestic violence reporting rose; intimate partner homicides fell - From 1970s court-mandated hiring of Black and female officers. Police departments in the U.S.: about 18,000 departments - Used to explain decentralization and variation in quality and reform experimentation. Effect of more police: hiring more officers reduces crime - Repeatedly stated as established evidence, mainly via deterrence.
Pivotal Quotes: "“What we know at this point is that hiring more police officers reduces crime.”" — Jennifer Doliak: Used to acknowledge police effectiveness even while discussing harms and reform needs. "“The problem is that they sold it as an evidence-based, data-driven… data proves that these eight changes will reduce police killings by 70% or something.”" — Jennifer Doliak: Her critique of Eight Can’t Wait and overclaiming causal evidence. "“We should just be very transparent about what we are showing and what we’re not.”" — Jennifer Doliak: Her preferred norm for advocacy groups and researchers alike.
Implications: Listeners should expect reform to be local, iterative, and evidence-seeking rather than slogan-driven. The biggest gains likely come from prevention, accountability, and trust-building—not just punishment or technology.