In the United States, policy-making frequently occurs across more than one level of government. For instance, federal officials make policies concerning Medicaid that are then managed by state agencies. States retain broad discretion in how Medicaid policies are applied, including who is eligible to enroll, which can lead to wide variation in how successful the policy is in achieving its stated goals. Similarly, in education, states make policies around curricular changes that are then implemented and enforced by local administrators. School districts might take sizably different approaches to teacher training, instructional practice, and evaluation, which can shape success towards achieving intended outcomes.

Some of the variation that occurs during the implementation process is likely to be beneficial. Flexibility in how policies are implemented affords lower-level jurisdictions the ability to ensure that policies can be contextually appropriate, innovative, and effective for their specific populations. In some cases, however, variation can result in unequal provision of services, lack of enforcement, and unevenly distributed social and economic outcomes. Bureaucratic discretion, along with both political and structural factors, frequently means that these unequal outcomes map onto existing demographic inequalities, with disproportionate harm (or lack of proportional benefits) to minority racial groups.

In partnership with the California Department of Justice, we examine the local impacts of a major state policy reform designed to reduce incarceration: AB 109. We find that although the policy resulted in overall reductions in imprisonment, its implementation at the local level varied widely. To illustrate, we estimate that if all arrests had been made in the most lenient county, the result would have been 80% fewer prison sentences. In contrast, the state would have had 150% more prison sentences if all arrests had been made in the most severe county. 

As our results make clear, the impact of reforms on the likelihood of a person’s exposure to prison—and its potentially lifelong harms—can depend largely on where they happen to live or have their case adjudicated, and can vary substantially across racial groups. More broadly, these results make evident the importance of research not only on how policies are passed, but on how they are implemented. 

This project was made possible with funding from the National Institute of Justice.