Why Longer Prison Sentences Don’t Work

What It Covers

Economist Jennifer Doleac, author of *The Science of Second Chances*, explains why longer prison sentences fail to deter crime, why swift and certain punishment outperforms severe punishment, and how policies like Ban the Box produce racially discriminatory unintended consequences — drawing on natural experiments and randomized trials across U.S. states and Denmark.

Key Questions Answered

• •Certainty vs. severity of punishment: Criminals with short time horizons — thinking days or weeks ahead, not years — are unresponsive to longer sentences but highly responsive to increased probability of arrest. Shifting criminal justice resources from extending sentences toward improving clearance rates and police presence produces measurably larger reductions in crime, particularly violent and property offenses, at lower cost to taxpayers.
• •DNA databases as deterrence tools: Entering first-time offenders into law enforcement DNA databases reduces recidivism by approximately 40%, based on Danish natural experiment data comparing otherwise identical offenders just before and after database expansion. The mechanism is purely deterrence — raising the perceived probability of being caught if reoffending — without changing sentence length or punishment severity at all.
• •Ban the Box backfires for Black applicants: Removing criminal history checkboxes from job applications causes employers to statistically discriminate, reducing callback rates for young Black men with limited education by roughly 5%. Field experiments in New York City and New Jersey confirm the racial gap widens post-policy. People with actual criminal records show no employment benefit, making the policy net harmful.
• •First-offense diversion cuts recidivism by 50%: Natural experiments comparing defendants whose first nonviolent felony charges were dropped versus prosecuted show a roughly 50% reduction in reoffending among those who avoided conviction. Arrest, booking, and case proceedings alone function as sufficient deterrence for a meaningful share of first-time offenders, making criminal records unnecessary and counterproductive for this group.
• •Street lighting reduces crime measurably: A randomized trial in New York City public housing found ambient lighting improvements reduced crime by approximately 10%. Daylight saving time data corroborates this — shifting daylight to evening hours reduces crime during that hour. Overlaying high-crime area maps with poor-lighting maps gives city officials a low-cost, data-driven tool for targeted infrastructure investment.
• •Electronic monitoring outperforms incarceration on cost and recidivism: Using electronic monitoring — house arrest with GPS tracking — as a partial or full alternative to prison sentences reduces reoffending while costing less than incarceration. The First Step Act created federal authority to expand early release to home confinement, and the current Bureau of Prisons leadership is working toward full implementation of this underutilized provision.