Were California’s Decarceration efforts smart? A quasi-experimental examination of racial, ethnic, and gender disparities

The study's evaluation of the California criminal justice reforms enacted between 2009 and 2015 provided evidence that the state's decarceration efforts reduced its overall rate of incarceration, but exacerbated racial, ethnic, and gender disparities in incarceration rates.

• California’s reforms substantially reduced the state's overall incarceration rate. Between 2009 and 2015, the incarceration rate in California decreased from 668 per 100,000 people to 520 per 100,000 people, while the incarceration rate of the synthetic control unit, a weighted combination of data from other U.S states, dropped from 666 per 100,000 people to 598 per 100,000 people. This means that, while the two rates were nearly identical in 2009, by 2015, California's incarceration rate was approximately 80 per 100,000 people lower than the comparison group.
• California's reforms resulted in an increase in Latinx-White incarceration disparities. While the Latinx-White incarceration rate ratios were identical in 2009, by 2015, the Latinx-White incarceration rate ratio in California was approximately 25% higher than its control counterpart's(1.69 vs 1.36).
• California's reforms resulted in an increase in Black-White incarceration disparities. Between 2009 and 2015, the Black-White incarceration rate ratio decreased from 7.15 to 6.97 Black people for every one white person, while the Black-White incarceration rate ratio in the comparison group decreased from 7.18 to 4.87 Black people for every one white person. This means that, while the two ratios were identical in 2009, by 2015, the Black-White incarceration rate ratio in California was approximately 43% higher than its control counterpart's.
• California's reforms resulted in an increase in incarceration disparities between men and women. Between 2009 and 2015, the incarceration rate ratio between men and women in California increased from 11.19 to 12.45 men for every one woman, while the incarceration rate ratio between men and women in the comparison unit decreased from 11.19 to 10.46 men for every one woman. This means that, while the two ratios were nearly identical in 2009, by 2015, the incarceration rate ratio between men and women in California was approximately 20% higher than its control counterpart's.

These findings suggest that future decarceration efforts can benefit from policies that, like California's reforms, target nonviolent offenses, but take additional care to ensure that reform efforts do not exacerbate incarceration disparities along race, ethnicity, and gender lines. This study suggests that reducing the rate of incarceration for nonviolent offenses is not enough to address disparities in the criminal justice system, but that reducing disparities should be a primary aim of reform efforts.

The study used a quasi-experimental approach to understanding the effects of California's criminal justice reform efforts on its incarcerated population. Instead of relying on any single state for comparison, the authors created a synthetic control unit, or a weighted combination of other states that best matched the pre-reform characteristics under study in California. The authors compared changes in California's and the control unit's incarceration rates from the period of 2000 to 2009, prior to the state's reform efforts, to the period of 2009 to 2015, the last year in which data was available. By using a control unit that closely matched the conditions in California prior to the reforms, the authors were able to isolate differences in outcomes between California and the control, and identify them as the direct results of California's reforms.

The authors used three key indicators to understand the incarceration trends in California and the control:

• Total incarceration rate, which was measured by the number of people incarcerated per 100,000 people. This measure did not include people in Federal prison, as they are not directly impacted by state-level policy.
• Racial and ethnic disparities, which were measured by both the Black-White total incarceration rate ratio and the Latinx-White total incarceration ratio. This measure did not include Asian-White and Native American-White disparities, as many states had few Asian or Native people in their incarcerated populations, which jeopardized the precision of the study. For racial and ethnic disparities, the study sample was limited to the 33 states that had data available on race and ethnicity incarceration rates for the period under study.
• Gender disparities, which were measured by the total incarceration rate ratio between men and women, for which data was available in all 50 states.

These measures relied on two sources of data. The Vera Institute of Justice’s Incarceration Trends data tool was used to compile state-level data on the total number of incarcerated people (in both prison and jail) and the racial, ethnic, and gender composition of incarcerated populations. U.S. Census Bureau sources were used to compile state-level data on state populations and their racial, ethnic, and gender composition.