Racial disparities within the criminal justice system continue to be a pressing issue for the US and California. In the wake of the killing of George Floyd, discussions around police reforms have heightened and centered on how law enforcement engages with people of color.
In this report, we analyze data for almost 4 million stops by California’s 15 largest law enforcement agencies in 2019, examining the extent to which people of color experience searches, enforcement, intrusiveness, and use of force differently from white people. While it is important to caution the reader that analysis of these differences is not causal, our analysis—which focuses in particular on differences between Black and white Californians—reveals notable differences.
These disparities are driven primarily by traffic stops made by the 14 data-contributing police and sheriff departments (as compared with the California Highway Patrol). These findings can provide guidance for discussing which stops can safely be reduced to mitigate racial inequities, which may also reduce risks and injuries to both officers and civilians.
While the nation grappled with the greatest public health crisis in at least a century—a pandemic in which communities of color carried the heaviest burden—the killing of George Floyd, among others, sparked civil unrest around California and the country. This unrest further highlighted stark racial inequities in our criminal justice system and the need for reform.
Inequities are especially stark between Black and white individuals: while Black residents make up about 6 percent of California’s population, roughly 16 percent of all arrests are of Black residents. Disparities are even greater at later stages in the criminal justice process, where Black people account for about 25 percent of county jail populations, about 26 percent of the probation population, and 29 percent of the prison population.
A recent PPIC survey found that 62 percent of Californians believe that the criminal justice system is biased against African Americans. Among African Americans, 88 percent hold this view. And while 54 percent of adults in California say police treat all racial and ethnic minorities fairly “almost always” or “most of the time,” only 18 percent of African Americans share that view.
Recognizing the need for data and research on law enforcement stops, the California legislature passed the Racial and Identity Profiling Act (RIPA) in 2015. The legislation—which was rolled out in waves based on the size of the agency—will require all law enforcement agencies in California to collect officer-perceived demographic and other detailed data for all pedestrian and traffic stops by 2023. The most recent data available include nearly 4 million stops made in 2019 by the 15 largest law enforcement agencies in the state.
This report builds on our previous work on arrests in California that found that criminal justice reforms implemented over the last decade have reduced racial disparities in arrests, bookings, and incarceration (Lofstrom et al. 2020; Lofstrom, Martin, and Raphael 2020). However, wide gaps remain. Here we broaden the scope to law enforcement stops, which include the many interactions Californians have with law enforcement that do not lead to arrests.
Complementing the 2021 RIPA Board report, we analyze the most recent stop data to better understand how interactions with law enforcement vary across race and ethnicity. Given that the starkest disparities are between Black and white Californians, our research focuses on inequities between these groups in frequency of stops, reasons for stops, and outcomes to provide a more complete picture of what those experiences are like.
We examine the likelihood that the individual stopped is searched, whether the search yielded any contraband or evidence, and if the stop resulted in any enforcement measures. We also examine intrusiveness and use of force, measured by reported outcomes such as being asked to step out of the vehicle, being handcuffed, and the involvement of an officer’s weapon. We then separately analyze outcomes by statewide (California Highway Patrol) and local (police and sheriff’s departments) jurisdictions.
The California state legislature passed the Racial and Identity Profiling Act (RIPA) in 2015 (AB 953), which requires all law enforcement agencies in California to collect perceived demographic and other detailed data regarding all pedestrian and traffic stops by 2023 (see Technical Appendix A for more details). “Stop” is defined as any detention by a peace officer of a person, or any peace officer interaction with a person in which the officer conducts a search.
The data elements mandated by statute include person-level and stop-level information. For person-level data, which we refer to throughout as personal traits, officers are required to record their perception of the identity characteristics for each individual stopped, including
Officers are prohibited from asking the person stopped to self-identify these characteristics.
Stop-level elements include
The data do not allow for corroborating the accuracy of the reported information, including the race and identity of the individual stopped and the specific actions taken by the officer. Nor do the data include information on the race and ethnicity of the officer.
We examine the most recent available data from the 15 law enforcement agencies who submitted their first full year of stop data from 2019. This includes California Highway Patrol (CHP), eight police departments (Los Angeles, San Diego, San Francisco, Sacramento, Fresno, San Jose, Long Beach, and Oakland) and six county sheriff’s departments (Los Angeles, San Bernardino, Sacramento, San Diego, Riverside, and Orange County).
These agencies recorded 3,992,074 stops of motorists and pedestrians during the 2019 calendar year. Technical Appendix A provides details and a discussion of how the distribution of stops and outcomes vary across agencies.
The requirements for collecting California traffic and pedestrian stop data are arguably the largest and most expansive efforts in the United States, although other states have collection requirements as well. According to information made available by the NYU School of Law Policing Project, 20 states have laws that mandate collection of stop data on varying amounts of traffic stops. Of these 20 states, only California, Oregon, and Illinois mandate data collection for both traffic and pedestrian stops.
States such as California, Connecticut, Illinois, Oregon, and Texas make their data publicly available and include regular reports that analyze that data. The remaining states vary in their data availability and published reports. A number of cities and counties, in states with data collection laws and without, publish their own stop data.
For the interested reader, the Stanford Open Policing Project maintains a website that houses stop data from virtually every city, county, and state agency that reports data in the US, and produces its own research using these data.
A primary objective of this report is to examine disparities between the experiences and outcomes Black and white Californians have during a stop. To start, we examine racial disparities in the frequency of being stopped by law enforcement, and disparities in the reported reason for the stop.
When we compare shares each group represents in stops to shares by population, we find considerable disproportionality statewide. Black residents accounted for 16 percent of stops made by all participating law enforcement agencies during 2019 (Figure 1) but constituted only 7 percent of the state’s population. Residents identified by law enforcement as Middle Eastern or South Asian were also overrepresented in stops (5%) compared to their share of the state’s population (2%).
White residents were represented fairly proportionally in stops (33%), compared with their population share (34%), as were Latino residents (39% and 41%, respectively). Asian individuals were underrepresented in stops (6%) compared with their share of the population (12%), as were multiracial residents (1% and 3%, respectively).
Individuals identified as Pacific Islanders were overrepresented (0.5% of stops, compared with 0.3% of the population), and those identified as Native American were underrepresented (0.2% and 0.3%, respectively). The percentage-point differences are small, but as a proportion of the population share, these differences are considerable. Again, the racial/ethnic identification comes solely from the officer making the stop.
SOURCES: Author calculations using California Department of Justice, Racial and Identity Profiling Act (RIPA) Wave 2 data, 2019; RIPA Board Report 2021 population calculations using American Community Survey (2018).
The data also reveal differences between reasons for stopping people of different races. For example, while more than 90 percent of stops of individuals perceived to be Asian or of Middle East/South Asian origin are stopped for traffic violations, about 75 percent of Black Californians stopped are for traffic violations (Figure 2). Conversely, officers report reasonable suspicion in 21 percent of stops of Black people, while 11.7 percent of white people and 5.6 percent of Asian people are stopped for reasonable suspicion.
While fewer stops involve individuals known by the officer to be on parole or probation or to have an outstanding warrant, their status provides officers with rights to stop and search without consent or reasonable suspicion. The percent of Black residents stopped who are on parole or probation is twice that of white residents (1.2% vs. 0.6%), and it is notably higher than Latino (0.8%) and Asian (0.2%) residents stopped as well.
The share of stops for an outstanding warrant is also twice as high for Black compared to white residents, also at 1.2 percent versus 0.6 percent. Technical Appendix Table A2 details differences across race and ethnicity in officer-perceived gender, age, mental health status, and whether the officer was responding to a call for services.
SOURCE: Author calculations using California Department of Justice, Racial and Identity Profiling Act (RIPA) Wave 2 data, 2019.
A key way this report extends the 2021 RIPA Board report is by taking a closer look at racial disparities in the experiences and outcomes of individuals after they are stopped by law enforcement. More specifically, we analyze the following four stop outcomes:
Of the almost 4 million reported stops in 2019, slightly more than 452,000 led to a person or property being searched. In close to 97,000 of those searches—about 21 percent—the officer found some contraband or evidence. That is, officers found contraband or evidence in about 2.4 percent of all police stops (Table 1).
The most common contraband was drugs or drug paraphernalia, found in a little more than 60,000 searches. The second most common category is other, which includes alcohol and cell phones (presumably evidence). In 18,507 searches, the officer found a weapon or ammunition. In more than 11,000 instances, the officer found property, which includes money that was either illegally held or was evidence.
In the vast majority of stops, about 88 percent, the officer issued at least a warning (Table 1). The officer issued at least a citation in 64 percent of all stops. Officers made an arrest—either a cite and release, or a booking—in 11 percent of stops, and booked over 6 percent of stopped individuals into jail.
While most stops led to some level of enforcement, intrusive actions were less common: for example, individuals were at least asked to step out of the vehicle in about one in six stops. Officers report some physical contact in over 14 percent of stops, most of which involved detaining a person curbside or in the patrol car.
In about 8 percent of stops, the person was at least being handcuffed. The percentage of stops that involved an officer’s weapon (which captures an officer pointing a firearm as well as when the officer uses the firearm or other weapon) is relatively small, at 0.42 percent. However, while the percentage is small, there are 16,918 stops where the officer reports that their weapon was involved. In 1,930 instances (0.05% of stops), the officer used a weapon— meaning the officer discharged a firearm or electric device such as a Taser, used a chemical spray or a baton, or a canine bit the stopped individual. The stop data do not capture whether anyone was injured as a result of the use of a weapon, or any other action taken during the stop.
SOURCES: Authors calculations using California Department of Justice, Racial and Identity Profiling Act (RIPA) Wave 2 data, 2019.
Black Californians are notably overrepresented in police stops, and officers report reasons for stops that can vary across race and ethnicity, and across law enforcement agencies. We next consider the role of context of a stop, in an effort to better understand underlying factors to the patterns observed in outcomes.
California passed the RIPA legislation in 2015 based on concerns about bias in policing that leads to different groups having different experiences with law enforcement. Beyond mandating collection of stop data, the legislation expanded and clarified the definition of racial and identity profiling to consider and rely on protected group status, such as race and ethnicity, in “… deciding which persons to subject to a stop or in deciding upon the scope or substance of law enforcement activities following a stop…”
Research consistently finds evidence of racial bias, explicit and/or implicit, broadly in society (see for example Bertrand and Mullainathan 2004; Bayer et al. 2017; Rothstein 2017; Avenancio-Leon and Howard 2019; Chetty et al. 2020; Kline, Rose, and Walters 2021). Furthermore, research has also found racial discrimination within the criminal justice system in jury, judge, and prosecutor decisions (Anwar, Bayer, and Hjalmarsson 2012; Arnold, Dobbie, Yang 2018; Sloan 2019). It is perhaps unsurprising that these racial biases extend to policing (Fryer 2019; Luh 2019; Hoekstra and Sloan 2020; Ba et al. 2021; Feigenberg and Miller 2021; Goncalves and Mello 2021), providing support to concerns historically raised by communities of color—concerns renewed in the wake of the killing of George Floyd.
Many factors contribute to whether an officer stops someone and to the officer’s subsequent actions. And while the RIPA data quite strongly point toward differences in stop outcomes across race and ethnicity (RIPA Board Report 2021), these differences may echo circumstances that do not reflect an individual officer’s bias. The reason and context for the stop likely influence an officer’s decisions and actions—for example, an officer may simply warn a driver stopped for speeding. Hence, differences in stop experiences between Black and white people may reflect differences in the reasons for the stop (Figure 2).
Independent of race and ethnicity, if an officer observed a person committing a crime, if a person has a warrant, or has a weapon, that person likely will be detained and searched, and possibly booked into jail after a stop. Such situations may be more adversarial—including the potential for use of force—than a traffic stop. If an individual is acting erratic, possibly due to behavioral health issues, an officer may shift decisions and actions. The prevalence of such situations across race and ethnicity may contribute to differences in outcomes.
Additionally, younger/inexperienced drivers may be more likely to violate traffic laws, and hence are plausibly more likely to be stopped than older and more experienced drivers. With men and adolescents/younger adults engaging in relatively more criminal activity, officers may place more scrutiny on younger men when they are stopped than on older individuals or women, independent of race/ethnicity (e.g., Ulmer and Steffensmeier 2014).
On average, Black Californians stopped by law enforcement are perceived to be younger compared to white Californians who are stopped (Technical Appendix Table A2). People stopped are also more often males. Relatively higher shares of Black persons are stopped for reasonable suspicion, outstanding warrant for an arrest, or mandatory supervision of a parolee/probationer. The latter two categories make a search more likely, as a warrant for a search is not needed, nor is cause, in California. Officers also report visibly seeing contraband in a higher share of stops of Black people than of white people.
Among the 15 largest law enforcement agencies, California Highway Patrol made more than 60 percent of all stops in 2019 of white individuals, but only about 35 percent of stops of Black individuals. And while Los Angeles Police Department (LAPD) accounts for almost 31 percent of stops of Black Californians, the agency made only 10 percent of stops of white Californians; partly reflecting that a higher share of the Los Angeles population is Black (about 8%) compared to statewide (about 6%). Agency-level differences in policing strategies, missions and roles, as well as officer behavior and biases, are also possible contributing factors.
Differences in contexts, location, and agencies likely contribute to racial disparities in stop outcomes. Our goal here is to use regression models that adjust to account for differences across race/ethnic groups in such factors, and move us towards more “apples-to-apples” comparisons.
That is, we seek to compare stop outcomes across race/ethnicity for, say, individuals of the same age and gender, stopped for the same reasons by a given law enforcement agency. We also adjust for whether the officer reported seeing contraband, and whether the person had an outstanding warrant or is on parole or probation. (See Technical Appendix B for a detailed discussion of the analysis and regression model.)
This exercise should not be interpreted as a causal analysis of race and ethnicity on outcomes and experiences, where after we account for reported relevant factors and contexts, any remaining disparities between Black and white individuals represent police bias. It may overestimate actual officer bias because the data does not capture all relevant factors and contexts (e.g., history of violent crimes or substance abuse). Furthermore, crime rates are often higher in some areas of a jurisdiction than in others, and hence are likely to lead to different levels of police presence and activity. If people of color are overrepresented in low-income and high-crime areas of a jurisdiction, this difference can also contribute to racial disparities in police stops. Unfortunately, the data do not allow us to account for such a possibility.
The exercise might also underestimate the prevalence of police bias if the factors that we control for, such as the reported reason for the stop, themselves represent police bias. One example may be the higher likelihood of Black individuals being stopped for reasonable suspicion than white individuals. Additionally, while an individual’s status of being on parole or probation, or having an outstanding warrant, is a relevant factor to adjust for since officers have additional latitude in those cases, the status itself is influenced by myriad biases in the criminal justice system (Anwar, Bayer, and Hjalmarsson 2012; Arnold, Dobbie, and Yang 2018; Sloan 2019; Rose 2021).
It is also important to keep in mind that the data examined is based on information an officer reports after the stop is completed, and hence provides an opportunity to report information aimed at hiding biases (Luh, 2019). It is plausible that the fully adjusted racial gaps represent a conservative, lower-bound estimate of racial bias in policing, but further certainty on this point is beyond the scope of this report.
These estimated differences in outcomes across race and ethnicity serve as a starting point for understanding how experiences with law enforcement differ for people of color compared to white people. We aim to provide a more complete picture of what the data tell us these differences look like. As we will show, the approach provides information on how contexts affect an officer’s actions and decisions, which may contribute to different experiences during a stop. Moreover, it also directs us towards contexts that deserve closer examination as particular sources of disparity.
Over 10 percent of all stops involve a search of the person or property. For Black Californians the likelihood of being searched is more than twice that of white Californians—a search rate of 20.5 percent and 8.2 percent, respectively (Figure 3). The disparity between Black and white people stands out compared to all other race/ethnic groups; furthermore, individuals perceived to be Asian or of South Asian/Middle East origin are less likely to be searched than white individuals.
Focusing on the Black-white inequities, after we adjust for officer-perceived personal traits such as age, gender, and disabilities, the gap shrinks somewhat, to 10.9 percentage points (second columns in Figure 3). When we additionally adjust for differences in the reported reason for the stop, the gap in Black-white search rates drops to 7.2 percentage points (third columns in Figure 3).
This difference is driven by significantly higher search rates of individuals on parole/probation and with outstanding warrants compared to traffic violations, combined with more Black than white persons stopped for being on correctional supervision or having an outstanding warrant. Search rates vary notably across law enforcement agencies (see Technical Appendix Table A2), reflecting differences in factors such as mission of agencies (CHP vs. police department vs. sheriff’s departments), jurisdictional crime rates, and policing practices.
Furthermore, demographic characteristics differ across jurisdictions. For example, Los Angeles has a higher share of Black people than San Francisco, while San Francisco has a higher share of Asians than Los Angeles. If, for example, Black residents tend to live in cities where law enforcement may conduct searches more often across all racial and ethnic groups, then the Black-white disparity in search rates may partly reflect location.
When we adjust for average differences in search rate across law enforcement agencies and for fixed characteristics of a jurisdiction, the Black-white gap drops further to 4.1 percentage points. In other words, adjusting for perceived personal traits, context of stop, and location reduces the gap by about two-thirds. And while this suggests that these factors matter, we also find that Black people are still 1.5 times more likely to be searched during a stop than white people.
SOURCE: Author estimates using California Department of Justice, Racial and Identity Profiling Act (RIPA) Wave 2 data, 2019.
NOTES: The bars represent percentage point differences in the likelihood of being searched between white Californians and each of the racial/ethnic groups identified in the RIPA data. All estimates are statistically significant at the 95-percent confidence level. The stop data are limited to the state’s 15 largest law enforcement agencies (LEA): California Highway Patrol; the police departments of the cities of Los Angeles, San Diego, San Francisco, Fresno, Long Beach, Oakland, Sacramento, and San Jose; and the sheriff’s departments of Los Angeles, San Bernardino, Riverside, San Diego, Orange, and Sacramento Counties. Bars represent estimated gaps relative to white residents with consecutively added controls. Detailed regression results are presented in the Technical Appendix.
As a share of stops, contraband and evidence are found relatively rarely—in only about 2.4 percent of stops. So-called unconditional discovery or yield rates—the share of stops, as opposed to the share of searches, in which contraband or evidence was found—vary across perceived race and ethnic origin. For example, officers found some contraband and evidence in 4.4 percent of stops of Black persons, 2.5 percent with Latino persons, 1.8 percent with white persons, and 1 percent with Asian persons (Technical Appendix Table A7).
As a share of searches, contraband or evidence was found more often—in 21.4 percent of searches. The discovery rate, also known as conditional discovery or yield rate, varies little by race/ethnicity (Figure 4), from 19.3 percent for those perceived to be of South Asian/Middle Eastern origin to 23.9 percent for those perceived to be multi-race/ethnicity. Searches yield contraband or evidence in 0.6 percentage points fewer searches of Black people compared to searches of white people, or at rates of 21.6 percent and 22.2 percent, respectively.
The small Black-white gap increases as we control for personal traits and contexts reported by the officer. Officers are more likely to find contraband and evidence in searches of juveniles and young adults compared to older adults, and in searches of men compared to women—young men are represented more often in searches of Black individuals. Adjusting for perceived personal traits increases the gap to -0.8 percentage points, and adjusting for reason for the stop increases it to -1 percentage point.
The data also include information on the basis for the search, and searches most likely to yield contraband or evidence are searches when the officer reports either seeing or smelling (which includes canines) contraband or evidence. Officers more often report such basis for searches in stops of Black individuals than white (see Technical Appendix Table A2), which in turn means that when we adjust for the basis, the gap more than doubles to -2.3 percentage points. Put differently, officers discover contraband in their searches of Black people about 10 percent less than in searches of white people.
Adjusting for differences across law enforcement agencies does not appreciably change the Black-white gap in yield rates. Furthermore, when we break down contraband and evidence by separate categories (weapons, drugs, property, or other), the discovery rate gap is entirely driven by a lower likelihood of finding drugs in searches of Black than white Californians. The estimated gap, adjusted for all factors, is a statistically significant difference of -2.7 percentage points between the Black and white yield rate for drugs (Technical Appendix Table B5).
