Discretion Within Automation: Prosecutorial Objections Under Colorado’s Clean Slate Act

• In Event: Automating Justice: State Policy Variations, Discretion, and the Impact of "Clean Slate" Record Relief

Abstract

Automatic record-sealing policies, often referred to as “Clean Slate” laws, have emerged as a major criminal justice reform aimed at reducing the long-term collateral consequences of criminal records. By removing the need for individuals to petition for relief, these policies are intended to expand access, improve administrative efficiency, and reduce disparities associated with discretionary decision-making. Despite this shift toward automation, many Clean Slate statutes retain provisions that allow prosecutors to object to the sealing of records that otherwise meet eligibility criteria. This feature introduces an important tension between automated processes and retained discretion, raising questions about whether these reforms fully achieve their intended equity goals.           
This study examines how prosecutorial discretion operates within an automated record-sealing framework by focusing on the implementation of Colorado’s Clean Slate Act. It asks whether prosecutorial objections are primarily shaped by legally relevant case characteristics and whether racial and ethnic disparities remain present in objection outcomes. The analysis uses administrative data on more than 135,000 individuals identified as eligible for automatic record sealing during the first two years of implementation, from 2024 to 2025. These data include demographic characteristics such as race and ethnicity, as well as case-level information including offense type and severity and contextual factors such as judicial district.            
To assess these relationships, the study employs generalized structural equation modeling (GSEM) to estimate both direct and indirect effects linking race and ethnicity to prosecutorial objections. The primary outcome is whether an objection was filed, measured as a binary indicator. Offense severity is modeled as a mediating factor, while additional controls account for age, sex, offense category, timing of eligibility, and variation across districts. This approach allows for a more detailed examination of how legally relevant factors and demographic characteristics intersect in shaping outcomes.             
The findings show that prosecutorial objections are relatively uncommon, occurring in about 6.6% of eligible cases, but they are strongly associated with offense severity. Individuals with felony convictions face more than twice the odds of receiving an objection compared to those with non-felony cases. At the same time, offense severity is not evenly distributed across groups. Hispanic and Black individuals are more likely than White individuals to have felony-classified cases within the eligible population. Mediation results indicate that this difference contributes to disparities in objection outcomes, with a statistically significant indirect effect for Hispanic individuals and a weaker pattern for Black individuals. When total effects are considered, both groups experience higher odds of objection relative to White individuals.              
The results suggest that automation reduces administrative barriers but does not fully eliminate disparities that are embedded in underlying case characteristics. Prosecutorial objections appear to operate as a mechanism that is responsive to legal distinctions while still reflecting discretionary processes. The findings point to the need for clearer standards, greater transparency, and continued evaluation to ensure that override provisions remain consistent with the broader equity goals of automated criminal justice reform.

Author

• Vivian Aranda-Hughes, Michigan State University
  Presenting Author