We calculated change in both absolute and relative differences across race/ethnicity, because relative differences can increase, even as absolute differences narrow. Relative differences depend on the rate in the reference group ; as rates in the reference group decline, a shrinking absolute difference may correspond to a widening relative difference. To calculate the difference in expected vs. observed counts, we subtracted model-predicted counts from actual arrests during the corresponding period. Confidence intervals were obtained by bootstrapping using 1,000 replications. All analyses were conducted using Stata Version 15.Felony drug arrest rates dropped immediately and precipitously for all racial/ethnic groups during the first month, and continued to decrease over time . In the year prior to Prop 47, the proportion of drug arrests that were felonies was highest for Blacks , suggesting that Blacks would have the most to gain from reclassification of drug felonies to misdemeanors. However, Prop 47 did not reclassify all felony drug offenses, only drug possession. In the year before Prop 47 was passed, the proportion of felony drug arrests for offenses that would be reclassified was 73% among Blacks, vs. 86% among Whites and 83% among Latinos. Thus, relative to Whites and Latinos, we would expect a smaller percentage decline in felony drug arrests of Blacks resulting from Prop 47. Consistent with these predictions, the models adjusted for secular and seasonal trends showed a significant, immediate decline in monthly felony drug arrest rates of 81 per 100,000 among Blacks, compared to declines of 44 among both Whites and Latinos . As a result, the absolute disparity in Black-White felony arrest rates declined by 37 per 100,000 in the first month, from a difference of 81 per 100,000 in the month pre-passage ,vertical grow systems for sale and continued to decrease over time. Felony drug arrest rates among Whites and Latinos were nearly equivalent in the month pre-passage and no difference was evident by one year post-passage.
Though Blacks had the largest decline in absolute felony drug arrest rates, they had the least proportional decline, likely because they had lower proportions of felony drug offenses that were reclassified to misdemeanors by Prop 47. In the first month, felony drug arrests among Blacks declined by 60% , compared to 69% among Whites and Latinos . With a larger percentage decline among Whites, the relative Black-White disparity increased by 27% . Though proportional declines in felonies continued to grow over time for all racial/ethnic groups, differences in declines across groups grew as well. By one year post-passage, felony rates among Blacks were 3.09 times higher than those of Whites , compared to 2.16 times higher in the month pre-policy . The difference in the number of felony arrests made, compared to the number expected based on secular and seasonal trends, was substantial. During the first year of adoption, Prop 47 led to an estimated 51,985 fewer felony arrests among Whites, 15,028 fewer among Blacks, and 50,113 fewer among Latinos . These represent reductions of 75.9%, 66.1%, and 73.7%, respectively. Our analysis of 60 months of county, race, and offense-specific arrest rates in California point to several notable effects of Prop 47. There was little indication of elevated drug arrest rates in Latinos compared to Whites before or after Prop 47, while the large absolute Black-White difference in felony drug arrest rates was reduced. With a higher proportion of felony drug offenses affected by Prop 47 , Whites had the greatest proportional decline in drug felonies, contributing to an increase in the relative Black-White disparity. Prop 47 appears to have led to reductions in arrests for drug offenses overall , which saw a decrease in the absolute Black-White difference, while relative disparities remained the same. There has been little study of the impacts of reducing offense severity on racial disparities in criminal justice involvement. In one exception, researchers found that reforming marijuana laws reduced arrests across all racial/ethnic groups, with no change in relative disparities between Blacks and other racial/ethnic groups . This aligns with our finding on Prop 47’s effect on total drug arrests, though we find increases in relative disparities for felonies.
Why did reducing the classification of some drug offenses reduce drug arrests overall? Reductions in drug arrests were unlikely a reflection of underlying crime trends – violent and property crime rates increased during this period . Prop 47 was a ballot initiative and law enforcement may be responding to perceptions of public opinion about public safety priorities. In areas with high rates of violent crime, police agencies may welcome a freeing up of resources to focus on these offenses. Officers may use their discretion to opt out of drug arrests and focus on offenses their department prioritizes. The initial drop in total drug arrests followed by the rise in month three suggests officers may also be responding to feedback from the courts regarding how to interpret and act on the legislative change. Reductions in arrests for all drug offenses may also reflect fluid lines between drug possession and sale. One lieutenant explained to the first author that in his city, sellers typically plead out to possession up to the third arrest, while arrests for possession were used to get information on sellers . Reducing possession to a misdemeanor may have diminished tools police and prosecutors used to enforce drug laws, contributing to a de-emphasis on arrests. These impacts warrant further investigation. While the absolute Black-White disparity in felony drug arrests decreased, the relative disparity increased, in part because of differences in pre-existing felony offense composition by race/ethnicity. Blacks had larger proportions of felony drug offenses unaffected by Prop 47 did not alter, such as sale. Whether this reflects racial differences in offending, or racial biases and practices in drug law enforcement, cannot be determined from our data, but other studies point to the latter . Prop 47 targeted drug possession, with the aim of decriminalizing substance use disorder. However, distinctions between sale and possession can be murky and influenced by prosecutorial discretion concerning which charges to file. Further study of racial inequalities in drug charges could help to address this unintended effect in California and states considering similar policies. Given substantial evidence of the role of social and economic factors in health outcomes , reducing incarceration and felony convictions through policy reform may be a critical component to addressing racial disparities in health. Our findings suggest that reclassifying drug offenses to misdemeanors is an effective approach to decreasing felony arrests across racial/ethnic groups, and absolute differences between Blacks and Whites.
However, a full assessment of how reducing criminal penalties affected racial/ethnic disparities in criminal justice involvement must go beyond the stage of arrest, particularly since groups may differ in the prevalence of prior convictions, which affect the likelihood of prosecution. Still, there is clear evidence that on a population level, there were declines in incarceration resulting from Prop 47,vertical grow system providing an opportunity to evaluate how reducing exposure affects health and associated racial/ethnic disparities, including the health of families and communities most affected by high rates of incarceration. Regarding more direct health impacts, a core component of Prop 47 was to reinvest savings from reduced incarceration to buttress substance use disorder and mental health treatment, with grants totaling $103 million awarded to 23 city and county agencies in mid-2017 . Prop 47 generated debate about whether arrestees would lose the incentive to enroll in treatment without a felony threat, which remains to be evaluated . Alternatively, populations accessing treatment or the proportions entering through voluntary vs. court-referred admissions may change . Racial disparities in substance treatment access could be impacted by Prop 47 as well. Blacks and Latinos arrested for drug offenses are more likely than Whites to receive incarceration, rather than drug treatment diversion . Sentence standardization initiated by Prop 36 in 2001 reduced disparities, but had a greater impact on Latinos than Blacks, perhaps because Blacks had more prior drug and violent offenses that precluded eligibility for diversion. Treatment resources generated by Prop 47 may have more promise for reducing disparities, given broader participant eligibility criteria stipulated in grant requirements . Critical questions remain regarding how shifting funds from a criminal justice to a public health approach to substance use disorders will influence treatment enrollment and outcomes for health, well-being, productivity, and public safety.
New programs funded by Prop 47 offer opportunities to evaluate these questions, and identify the most effective models for improving public health. In arrests with multiple offenses, only the most severe is included in the dataset. Since Prop 47 reduced the severity of drug possession offenses, one concern is whether some reduction in arrests could be attributed to co-occurring offenses that became comparatively more severe than drug possession postProp 47. This could occur only in measures incorporating offenses classified as felonies pre-Prop 47 and misdemeanors post-Prop 47. These would include drug arrests reclassified by Prop 47 and total drug arrests, but would exclude felony drug arrests, misdemeanor drug arrests, and felony drug arrests unaffected by Prop 47. We used data on juvenile arrests, which contain up to five co-occurring offenses, to estimate possible bias. We found that potential masking was minimal and would not alter findings. Approximately five percent of “drug arrests reclassified by Prop 47” may have been masked post-policy – far less than the 46-51% declines in this category 12 months post-policy. Masking in the “total drug arrests” measure was estimated at three percent, compared to declines of 17-22% at 12-months post policy. Data are also event-, rather than person-level. Some arrests may represent the same person, though we minimized this possibility by using monthly rates. We assessed the extent of possible bias by linking individuals on name, date of birth, and jurisdiction for July 2013; just 1.2% were arrested more than once and 0.1% more than twice. We also lacked data on prior convictions; Prop 47 offenses retain felony classification for individuals with serious and/or violent convictions such as homicide and sexually violent offenses, or convictions requiring registration as a sex offender. The history of more frequent arrests and severe charges among Blacks arrested for drug offenses , may have minimized the effects of Prop 47 on reducing Black-White disparities in felony convictions. Given that other states are pursuing similar policy changes to reduce racial disparities , this effect should be further explored, and limiting prior record exclusion criterion considered. Race/ethnicity in arrests data may be based on officers’ observations, rather than self-report, in population denominators. This could lead to misclassification of the numerator in arrest rates, though sensitivity analyses indicated findings were robust. We also only analyzed three racial/ethnic groups; though these groups make up 95% of arrests in California, further research could assess disparities and impacts among other populations.State-level criminal justice reforms often leave a great deal of room for interpretation and discretion. There can be a tension between the goals of state legislators enacting criminal justice laws and county level officials who administer them, leading to highly county-specific implementation . As an example, mandatory minimum sentencing laws were seen as tough on crime and therefore historically supported by state legislators as a symbolic statement. They were opposed by courts, however, because they increased trial rates and case processing times, and penalties were considered disproportionately severe . In his review of two centuries of mandatory minimum sentencing laws, Tonry found a long history of courts using devices to circumvent them: prosecutors refused to file charges, plea bargaining was used to reduce charges, and judges refused to convict or ignored the statute and imposed a different sentence . When severe mandatory minimums for drug sale were in place in Michigan, for example, nearly every charge was reduced to possession , while harsh minimums for felony possession during the Rockefeller drug law era in New York were circumvented by reducing charges to misdemeanors or referring defendants to drug courts . Scholars have proposed that differences in local contexts determine how discretionary options are used within locales, producing geographic variation in case dispositions. Specifically, cases are prosecuted in accordance with personal and local principles of proportionality , local political leanings, resources for prosecution, and community priorities .