Preventing crime without police & prisons | Jennifer Doleac (2020)

U.S. Incarceration and Demographics

• The U.S. has the highest incarceration rate globally, approximately four times the typical global average, peaking at 1 in 100 adults in 2007.
• Currently, 1 in 110 U.S. adults (2.3 million people) is incarcerated: 1.3 billion in state prisons, 630,000 in local jails, and 225,000 in federal prisons.
• Racial disparities are stark; Black Americans comprise 40% of the incarcerated population despite representing only 13% of the total U.S. population.
• Jails hold 630,000 people on any given day, but see 10.6 million admissions annually; roughly 75% of jail inmates have not been convicted of a crime.
• Unconvicted jailers are detained primarily due to inability to afford bail and judicial assessments of flight risk or public danger.
• Approximately 13 million misdemeanor charges are filed annually, with low-level offenses (penalties under one year) accounting for nearly 25% of the daily jail population.
• Among the convicted incarcerated population, 47% are imprisoned for violent crimes, 19% for drug crimes, 17% for property crimes, and 17% for public order offenses.
• Private prisons house less than 9% of all incarcerated individuals, suggesting private sector lobbying is not the primary driver of mass incarceration.
• Robbery was the most common violent crime leading to incarceration until 2016, when it was narrowly surpassed by murder.
• The clearance rate for serious violent crimes is 46%, while the clearance rate for property crimes is only 18%.
• Average time served for drug possession is 15 months (median 10 months), compared to 4.7 years (median 3.2 years) for robbery.
• Sentence lengths increased until roughly 2000 but have remained relatively flat since then.
• The probability of being arrested for a serious violent crime reported in 2017 was approximately 22%, indicating a significant gap between crimes committed and arrests made.

Problems with Policing and Racial Bias

• Primary issues include unnecessary escalation of police encounters (arrests/force that could have been avoided) and systemic racial bias.
• Jeremy West's research using Texas highway crashes found that white officers are significantly more likely to issue citations for minor violations (e.g., expired registration) to Black drivers than White drivers, showing no bias in felony arrests.
• Mark Hoekstra and Carly Will Sloan found that dispatching white officers to 911 calls in Black neighborhoods results in a higher likelihood of force compared to dispatching Black officers.
• Randomization of defendants to different judges reveals that judges vary dramatically in their propensity to detain pretrial, leading to racial disparities where Black defendants are more likely to be locked up than White defendants in similar circumstances.
• Misconduct by a small minority of officers is predicted by previous complaints, but current systems often lack the mechanisms to fire officers based on this predictive data.
• Qualified immunity and civil asset forfeiture allow police to seize property with low evidentiary standards and often provide financial incentives for departments to retain seized assets.
• Police unions are cited as a major impediment to reform, with studies indicating unionization correlates with increased police killings without a corresponding reduction in crime.
• "Ban the Box" policies intended to help formerly incarcerated people find jobs were found by Jennifer Doliak to have increased discrimination against Black men with high school diplomas or less.

Evidence-Based and Experimental Reforms

• A one-day "procedural justice" training program for 8,000+ Chicago police officers, implemented via staggered rollout, significantly reduced both citizen complaints and use of force.
• Court-mandated affirmative action hiring in the 1970s led to reduced victimization of Black residents and female residents, as trust in police increased and reporting rates for crimes rose.
• Body-worn cameras have shown null or mixed effects in randomized controlled trials, with some studies finding no reduction in force and others finding increases in use of force.
• The "Eight Can't Wait" reform agenda's claims of a 70% reduction in police killings are criticized by researchers for relying on raw correlations rather than causal evidence.
• Hiring more police officers reduces crime, with elasticity estimates suggesting a 1% increase in officers leads to a 0.1% to 1% reduction in crime, primarily through deterrence rather than incapacitation.
• Expanding DNA databases reduced reoffending by 43% for serious offenders in Denmark and 17% for violent offenders in the U.S., likely due to increased perceived likelihood of capture.
• Increased ambient lighting (via daylight saving time or streetlights) resulted in a net reduction in nighttime outdoor crimes, with no evidence of crime substitution to other times.
• Cognitive Behavioral Therapy (CBT) for at-risk youth and juvenile offenders demonstrated large, persistent reductions in recidivism and reincarceration.
• Lead abatement is identified as a high-impact prevention strategy; historical data links the phase-out of leaded gasoline to the massive drop in crime rates starting in the mid-1990s.
• Early childhood interventions (e.g., Head Start) and adult government transfers are associated with lower crime rates, whereas excluding individuals with criminal records from transfers increases recidivism.

Structural and Systemic Observations

• The U.S. incarcerates more people than any other nation but spends less on police per capita and employs fewer officers than many peer countries.
• The U.S. violent crime rate is significantly higher than in Europe and Asia, a difference attributed partly to cultural factors and gun prevalence rather than policing alone.
• The 18,000 decentralized U.S. police departments create variation in quality and allow for natural experimentation, though it also hinders consistent oversight.
• Limiting police discretion (e.g., strict sentencing guidelines) can reduce racial disparities, whereas broad discretion often allows bias to influence outcomes.
• Implicit bias training lacks robust evidence of effectiveness and is often criticized as a "PR solution" rather than a substantive fix.
• Economics is underrepresented in the diversity of its researchers, leading to potential blind spots in research questions and hypotheses regarding crime and justice.
• Women in economics face higher citation thresholds and are often credited less than male co-authors, despite producing equivalent or superior marginal work.
• Economists prioritize transparency, robustness checks, and simple graphical visualizations to validate causal claims, warning against overly complex models without clear intuition.
• Jennifer Doliak expresses optimism that the current momentum for criminal justice reform, driven by diverse stakeholders and a lack of strict partisan divides, will lead to a more effective and fair system within a decade.