INTRODUCTION
Eastpointe, Michigan borders on the City of Detroit; Eight Mile Road, eastbound, lies in Detroit, while the westbound lanes are in the City of Eastpointe. The Detroit neighborhoods closest to Eastpointe are racially integrated, but when one crosses Eight Mile Road and enters Eastpointe, Black residents are quite uncommon. The pattern of residence, by race of driver, is predictably polarized; nearly all Blacks ticketed by Eastpointe police live in Detroit while among Whites who are ticketed, most live in Eastpointe or Warren, but quite a few are Detroit residents. Motorists in Eastpointe are often passing through on their way to jobs or shopping complexes in neighboring suburbs.
Eastpointe is not home to such shopping amenities as large regional malls. Nor are huge manufacturing complexes to be found in Eastpointe, although they are commonplace in the neighboring industrial suburb of Warren, directly to the west. It is a residential suburb of modest homes and strip malls. This particular part of suburbia inspired Reynolds Farley and his co-authors (Reference Farley, Danziger and Holzer2000), in their book Detroit Divided, to put forth their “hostile suburb” hypothesis to explain Black worker reluctance to seek employment in working-class suburban Detroit. This hostility is expressed, in part, by the police practice of profiling Black drivers for arbitrary stops “if they drive into White neighborhoods” (Farley et al., Reference Farley, Danziger and Holzer2000, p. 8).
The U.S. Department of Justice guide defines racial profiling as “any police-initiated action that relies on the race, ethnicity, or national origin rather than the behavior of an individual” (Ramirez et al., Reference Ramirez, McDevitt and Farrell2000, p. 3). Asked by a Gallop Poll if they disapproved of racial profiling, 81.9 percent of Americans responded that they did. Asked if they felt racial profiling was widespread, fifty-six percent of Whites and most Blacks indicated that it was (Ramirez et al., Reference Ramirez, McDevitt and Farrell2000). In a review of the anecdotal evidence on racial profiling, Ramirez et al. observed, “a common complaint is that police stop people of color traveling through predominantly White areas because the police believe that people of color do not belong in certain neighborhoods” (Reference Ramirez, McDevitt and Farrell2000, p. 5).
Lefebvre (Reference Lefebvre1991) describes urban space as a social product, the result of a mix of legal, political, economic, and social practices, and a common outcome of this mix is conflict over who has what rights to occupy which spaces. This study utilizes Eastpointe police records of traffic stops, ticketing, driver and vehicle searches, and arrests to investigate the reality of racialized spaces in suburban Detroit. Black Detroiters, observes Meiklejohn (Reference Meiklejohn2002), often “perceive the larger world outside the city's boundaries to be a hostile and potentially harmful place” (pp. 343–4). A Black executive, asked to identify suburban Detroit neighborhoods where African Americans are harassed by the police, responded “any of them that you don't belong in” (Turner Reference Turner1997, p. 133). The “conflicts about what (and who) belongs where and about what (and who) is out of place” (Cresswell Reference Cresswell1996, p. 99) are well defined in blue-collar suburban Detroit: Black Americans belong in the central city; those venturing into suburbia are often viewed as out of place.
SELECTIVE OVERVIEW OF DRIVING-WHILE-BLACK STUDIES
The variant of racial profiling, popularly known as “driving while Black”, has inspired numerous studies over the past decade of police treatment of African American motorists. These studies commonly find that police disproportionately stop, ticket, search, and arrest minority motorists, in comparison to Whites. After documenting this empirical fact, the various studies tend to polarize into pro- and anti-police subgroups; the former defend disproportionate ticketing, searching, and arresting of Black drivers as efficient police practices, while the latter criticize them as abusive and/or ineffective police actions. The City Journal, a quarterly journal of urban affairs published by the Manhattan Institute, has enthusiastically and frequently portrayed the case in favor of racial profiling. Because minorities are simply assumed to be more inclined than Whites to break laws, they attract disproportionate police attention (MacDonald Reference MacDonald2001). Seen in this light, racial profiling is an effective law-enforcement strategy and its critics are an anti-police crusade that travels under the banner of ending racial profiling (MacDonald Reference MacDonald2002).
Scholarly studies criticizing racial profiling as abusive are often based on mainstream scholarly empirical research methodologies rather than a priori assumptions; they commonly utilize multivariate statistical analyses of citizen responses to national surveys probing police-citizen interactions in various circumstances, including traffic stops. Engel and Calnon's (Reference Engel and Calnon2004) analysis of drivers involved in recent traffic stops typifies such studies; their findings dispute the accuracy of the “good policing” argument supporting racial profiling. Their analysis of 7054 drivers stopped by the police found that Blacks, controlling statistically for situational factors and driver traits (other than race), were forty-seven percent more likely than Whites to be ticketed. Similarly, the odds of arrest were seventy-nine percent higher for Black drivers than for Whites, and, controlling for arrest, the odds of being searched were fifty percent higher for Blacks than for White drivers. On a per search basis, however, contraband finds were over twice as frequent for Whites than among Blacks. High search and arrest rates for Black motorists coexisted with low contraband find rates, yet the “efficient policing” rationale for searching Blacks at higher rates than Whites is based on the premise of their greater frequency of contraband possession.
Utilizing survey data and methodological techniques similar to those employed by Engel and Calnon (Reference Engel and Calnon2004), Lundman and Kaufman (Reference Lundman and Kaufman2003) found that African American drivers stopped by the police were significantly more likely than Whites to report that the police had no legitimate reason for making the traffic stop. Further, Blacks drivers were significantly less likely than Whites to report that the police acted properly during the traffic stop encounter. Methodologically-sophisticated analyses of interactions between police and African American motorists have rarely addressed the hypothesis that police attention to Black drivers varies in intensity in Whiter as opposed to Blacker urban areas. An exception is Meehan and Ponder's (Reference Meehan and Ponder2002) analysis of variations of police behavior by ecological or neighborhood context. Utilizing data collected from a suburban police department serving a city characterized as “blue collar and predominantly White (98%)” and sharing a “border with a city of predominantly African American residents (more than 75%)” (p. 405), Meehan and Ponder's study focused upon police behavior in a city much like Eastpointe.
Comparing police stops of African American motorists in the areas bordering the predominantly Black city, versus areas described as non-border neighborhoods, Meehan and Ponder found that Black drivers were three times more likely to be stopped by the police in non-border (as opposed to border) areas. The police practice of making computer queries (based on vehicle license plate data) of Black drivers, furthermore, varied widely in border and non-border areas. Query rates targeting Black drivers were highest in non-border, more affluent White neighborhoods, leading the authors to note that the police appeared to be “hunting for, or clearly noticing, African American drivers in these sectors” (p. 422). Meehan and Ponder (Reference Meehan and Ponder2002) conclude that race and place shape police interactions with motorists; specifically, being Black and out of place is noticed.
TRAFFIC FLOWS, STOPS, AND TICKETING PATTERNS: RACIAL COMPOSITION OF DRIVERS IN EASTPOINTE MICHIGAN
An interstate highway (I-94) runs within one mile of Eastpointe and exits from this freeway connect to three major Eastpointe streets—Eight Mile Road, Nine Mile Road, and Ten Mile Road. According to this study, Black drivers were most prevalent on Eight Mile Road. Yet, this is a boundary street dividing Eastpointe and Detroit, not a street in a White neighborhood. This study proceeds by comparing treatment of Black motorists on Eight Mile Road to the interior streets of Nine and Ten Mile Roads to investigate whether enforcement of racialized space by the police can be detected by comparing how drivers are treated on major thoroughfares that are more, as opposed to less, White.
Whether or not Blacks are being disproportionately harassed depends not only upon ticket and arrest patterns themselves, but upon the frequency of Black motorists in Eastpointe. To establish the racial composition of drivers on major streets, this study used observers who conducted eighty two-hour sessions over a three-month period, noting race of the driver for nearly 41,000 cars. This methodological approach for establishing the racial composition of the traffic flow on specific streets was first developed and utilized by Lamberth (Reference Lamberth1994) to investigate whether Black motorists traveling on the New Jersey Turnpike were disproportionately stopped and arrested by the state police.
Nearly ninety-seven percent of the drivers of vehicles in Eastpointe were classified as either Black or White; remaining drivers were either neither, or their racial identity was unclear to the observer (these cases were excluded from the analysis). Observed relative incidence of Black drivers on the following three Eastpointe areas (Table 1) was: (1) Eight Mile Road—39.3 percent Black drivers, (2) Nine and Ten Mile Roads—5.3 percent Black drivers.
Table 1. Percent of All Black and White Drivers by Race on Selected Streets, 1999 and 2000
Table 2 summarizes race of driver, by Eastpointe street location, for 13,443 police stops in which traffic tickets were issued during the 1996–1999 time period.Footnote 1 Although drivers ticketed in 21.8 percent of all stops were Black, their proportionate share varied greatly from street to street. On Eight Mile Road, for example, Black drivers accounted for thirty-five percent of the traffic stops, while on Nine and Ten Mile Roads, the Black motorist share was 7.8 percent (Table 2).
Table 2. All Stops of Black and White Drivers Resulting in Tickets Being Issued*
* Cutoff: Black → 20% or less chance of being White
Whether or not the Black share of police stops (tickets and arrests) is high or low is defined relative to the proportion of all drivers who are Black on the corresponding streets: the differential—the share of stops minus the share of Black drivers—is summarized below.
For Eight Mile Road, Black motorists accounted for 39.3 percent of the drivers observed and thirty-five percent of the stops. The probability of being stopped and ticketed on this road was actually lower for Black than for White motorists. On Nine and Ten Mile Roads, Blacks were 5.3 percent of the drivers observed and 7.8 percent of the drivers stopped and ticketed. Determination of the over-representation or under-representation of Black drivers stopped is measured, in turn, by comparing these differentials to the percent of Black motorists, as follows:
A “stop” is defined as taking place if a motorist receives one or more traffic tickets after being stopped by the Eastpointe police. Number of tickets, as such, does not matter; one stop has occurred if one ticket is issued and one stop has occurred if two or three tickets are issued (by definition).
Types of ticketing patterns vary racially in Eastpointe. Black drivers stand out particularly in terms of receiving multiple tickets (two or three traffic citations) when stopped by the police. Table 3 looks solely at multiple ticket recipients, summarizing race of driver, by street groups. Black drivers accounted for 32.3 percent of all multiple tickets issued in Eastpointe in the relevant 1996–1999 time period. On Eight Mile Road, Blacks received 49.1 percent of the multiple tickets, and on Nine and Ten Mile Roads, the Black motorist share was 10.6 percent (Table 3).
Table 3. Multiple Tickets: All Multiple Tickets Going to Black and White Drivers, by Street Where Ticket Was Issued*
* Cutoff: Black → 20% or less chance of being White
Ticket differentials, corresponding to those presented on the preceding page for all tickets, are summarized below for multiple tickets only:
The heavy overrepresentation of Black drivers receiving multiple (two or three) tickets when stopped was measured by comparing the Black differential to the percent of Black motorists, as follows:
These substantial overrepresentations are statistically significant, indicating that Black drivers received multiple tickets in the course of traffic stops significantly more frequently than White drivers.
In contrast to their multiple ticket over-representation, Black drivers are generally underrepresented among recipients in Eastpointe of the dominant ticket issued—the speeding ticket. More speeding tickets are issued than all other moving violations combined, and over seventy-six percent of the speeding tickets involved radar stops. Black motorists received 19.9 percent of all speeding tickets issued during radar stops, less than their 21.8 percent share of all Eastpointe traffic stops. Their overrepresentation among ticket recipients was most pronounced in situations involving discretion among the ticketing police officers.
The wide variance in racial ticketing patterns from officer to officer brings home the discretionary nature of ticketing. Comparing Officer L and Officer K, the two officers on the Eastpointe police force who issued the most radar speeding tickets, Officer L wrote 1883 tickets and seventy-six (about four percent) were multiple tickets. Officer K, in contrast, wrote 1943 tickets, 456 of them (23.5 percent) were multiple tickets.
OTHER FACTORS MAY INFLUENCE TICKETING PATTERNS
Thus far, two patterns stand out. First, Black motorists on Nine and Ten Mile Roads were ticketed out of proportion to their numbers, and this was particularly true for multiple tickets. Second, Black drivers across the board were much more likely than White motorists to receive multiple tickets when stopped by the Eastpointe police. These findings may or may not, by themselves, indicate discriminatory treatment of Black motorists by the police. A possible explanation for the disproportionate ticketing of Black motorists in Eastpointe is that Black drivers, perhaps being lower income, on average, than Whites were disproportionately ticketed because they were low income (and not because they are Black). The proxy measure of low income examined here was age of vehicle. This proxy, however, suggests that Black drivers in Eastpointe may be broadly similar to ticketed White drivers in terms of economic status.
Driving older vehicles, in turn, may itself be a cause of being ticketed, quite irrespective of driver race. Car traits such as faulty mufflers are more likely to occur on older than newer cars, and these deficiencies may lead to tickets. Indirectly, older vehicles are a reflection of driver economic status in many instances, and it is reasonable to expect that low-income drivers disproportionately drive older cars. Yet the data collected from traffic tickets included vehicle age, revealing minimal differences in the age distribution of cars being driven by Black and White motorists ticketed in Eastpointe during the time period under consideration. Among cars in the less-that-six-years-old category (“new” cars), the proportions of stopped and ticketed Black and White drivers was virtually identical. Defining “old” cars as those in operation for eleven or more years in the year they were stopped/ticketed in Eastpointe, 31.2 percent of Whites stopped had old cars, versus 33.7 percent of Blacks.
Beyond vehicle age, low-income drivers may be less likely to have current registration for their vehicles, current insurance, and the like, since their low incomes may constrain their ability to purchase these items. Over ten percent of all tickets issued to White drivers and nearly fifteen percent of tickets issued to Black drivers were for the offense of having no proof of auto insurance, making this the second most common type of ticket issued by the Eastpointe police. The “no-proof-of-insurance” tickets were further stratified according to age-of-vehicle categories, revealing that drivers of old cars were more likely to be ticketed for no proof of insurance than people driving new cars. Among Black drivers of new cars, for example, 10.7 percent of all tickets received were for no proof of insurance; among Blacks driving old cars, the corresponding percentage was 18.3 percent. Yet the data revealed a pattern of racial disparity: Black drivers of new cars were slightly more likely to get no-proof-of-insurance tickets than Whites with cars six to ten years old. Whites with old cars, similarly, were less likely to get no-proof-of-insurance tickets than Blacks with cars six to ten years old (12.7 percent for Whites versus 14.5 percent for Blacks). Quite irrespective of age of car, Blacks were more likely to be ticketed for no proof of insurance than White drivers.
Black and White motorists in Eastpointe tended to be ticketed on different streets; Blacks were more likely than Whites to receive multiple tickets when stopped by the police, particularly when the stops were on internal Eastpointe streets. Other traits that differ systematically by race of driver might influence ticketing patterns. Relative to the number of cars on the road, for example, those driving at night were more likely to be ticketed than those driving during the day. Could this partially explain the higher incidence of ticketing of Black drivers in Eastpointe? In fact, it clearly does not.
During the late night and early morning hours, the proportion of tickets going to Black motorists declined somewhat. In the off-hour times (9:00 pm to 6:00 am), the share of tickets going to Black motorists was relatively low. Relative to the number of cars on the road, drivers were heavily over ticketed in the late night and early morning hours, and the evidence consistently shows that this pattern was not pushing up Black motorist ticketing incidences. Driver race, after all, is harder for the police to determine prior to stopping a vehicle when it is dark outside. Like the age of vehicle factor, time of day tells us little about Black/White ticketing patterns.
DETERMINING RACE OF DRIVER
When Eastpointe police issue traffic tickets, they may or may not record race of driver. Using decennial census of population block data from the U.S. Bureau of the Census, the address of each driver not racially identified by the police was traced to the specific block where the driver resided, and the racial composition of that block was recorded from the census data. Assume, for the sake of illustration, that the block where a driver resided had 193 residents, 127 of whom were Black, sixty-three were White, and three were “Other”. The percentage of Black residents on this block (127 ÷ 193) was 65.8; the percentage of White residents on this block (63 ÷ 193) was 32.6. Finally, the percentage of Black residents divided by the percentage of White residents (65.8 ÷ 32.6) equals 2.018, meaning that a resident of this block was 2.018 times more likely to be Black than White. If the percentage of Black residents exceeded the percentage of White residents, then this Black-to-White ratio exceeds one. This can be thought of as a fifty percent cutoff: if the ratio exceeds one, then residents had a greater likelihood of being Black than White.
In Detroit, Eastpointe, Warren, and the Detroit area generally, racially integrated blocks are not common, simplifying the task of accurately assigning driver race. Because of this reality, a higher cutoff point than fifty percent was used to assign driver race in this study: an eighty percent cutoff was adopted. For drivers with less than an eighty percent chance of being Black and less than an eighty percent chance of being White, race was unclear and such drivers were dropped from the analysis. This reduced the number of traffic stops analyzed in this study by nearly ten percent. Consequences for applying other driver racial assignment rules are investigated later in this study. Applying different cutoffs has a major benefit: the robustness of the findings of over-ticketing of Blacks is demonstrated. The findings derived from applying higher or lower cutoffs are extremely robust.
LOGISTIC REGRESSION ANALYSIS OF TRAFFIC TICKET PATTERNS
Regression analysis is used to sort through possible explanations for the disproportionate multiple ticketing of Black drivers in Eastpointe. Of the various factors considered in this study—driver race, age of vehicle, street where one was ticketed, and so forth—which are the important determinants of multiple ticketing patterns? The purpose of logistic regression analysis is to consider all of these factors simultaneously so that one can tease out the separate impact of driver race on ticketing patterns, while controlling statistically for ticketing location (street), time of day, age of vehicle, and other factors. Precise definitions of all explanatory definitions used in the Table 4 regression analysis exercise are spelled out in the attached appendix: Traffic Tickets. The hypothesis of interest is this: quite irrespective of the influence of ticketing location, time of day, age of vehicle, and other factors, being Black is hypothesized to increase one's chance of receiving multiple traffic tickets. The logistic regression model presented in Table 4 tests this hypothesis. Results confirm that being Black significantly increases one's likelihood of getting multiple tickets.
Table 4. Logistic Regression Analysis Delineating Drivers Receiving Multiple Tickets When Stopped from Drivers Getting One Ticket**
* Statistically significant, five percent significance level
** Cutoff: Black → 20% or less chance that the person is White
Source: Traffic tickets issued by Eastpointe police
In the Table 4 logistic regression model, the dependent variable is equal to zero for drivers receiving one ticket; if the driver received a second or third ticket, the dependent variable equals one. Thus, a negative regression coefficient attached to the Black-driver variable would indicate that Black drivers, other things equal, are less likely to get single tickets than White drivers while a plus sign indicates an increased likelihood of receiving multiple tickets.
Among the statistically significant determinants of multiple ticketing patterns in Table 4, two factors are associated with an increased chance of getting multiple tickets. They are (1) being a Black driver, and (2) driving an old car (compared to a middle-aged car). Similarly, four factors (all statistically significant) are associated with a reduced chance of getting multiple tickets, including (1) driving a new car (relative to a middle-aged car), (2) being ticketed for speeding in a radar stop, (3) being stopped late (between 9:00 pm and 6:00 am), and (4) being stopped on Eight Mile Road. Being a Black driver, as expected, increases one's chance of getting multiple tickets.
To explore in further detail the multiple ticketing patterns in Eastpointe, logistic regression analyses (along the lines of the analysis in Table 4) were conducted separately for White drivers only and for Black drivers only. This analysis is designed to highlight possible differences in multiple ticketing patterns across the two racially-defined groups.
Applying regression analysis to White drivers only, one factor was statistically significant and associated with an increased chance of getting multiple tickets: driving an old car. Four factors (all statistically significant) are linked to a reduced chance of getting multiple tickets: (1) driving a new car, (2) being ticketed for speeding in a radar stop, (3) being stopped late, and (4) being stopped on Eight Mile Road. When the logistic regression analysis was repeated for Black drivers only, noteworthy racial differences emerged. The time and the location of police stops involving Black motorists had no impact on the probability of receiving multiple tickets, nor did age of the vehicle. Only one statistically significant factor was linked to a reduced chance of getting multiple tickets—being ticketed for speeding in a radar stop.
Additional regression analyses were conducted to test the robustness of the finding that being Black leads to increased multiple ticketing. Under varying approaches, assumptions, and variable specifications, the same finding repeatedly emerged: Black drivers were more likely than Whites to get multiple tickets, other things equal. In cases where race of driver was not recorded by the Eastpointe police, race of driver was inferred, as described above. Utilizing geocoding, I created variables identifying Black drivers as “less” Black and “more” Black. Further, impacts of applying different cutoff values to infer driver race—fifty percent, 66.7 percent, eighty percent—were explored. None of these variations had even minor impacts on the regression-analysis finding that driving while Black substantially increased one's likelihood of receiving multiple traffic tickets.
ARRESTS
Routine Eastpointe traffic stops typically result in issuance of traffic tickets, after which the driver goes on his or her way. Yet, for roughly nine percent of White drivers and thirty percent of Blacks, these stops lead to arrests. A review of the police arrest narratives originating in traffic stops between 1996 and 1998 indicates that African Americans accounted for the majority of those arrested by the Eastpointe police. Arrest narrative reports are required to include race-specific information and nearly all did. The applicable 1965 police reports described 1879 arrests: 989 of those arrested were Black and 890 were White. Black motorists overall accounted for 21.8 percent of the traffic stops and 52.6 percent of ensuing arrests.
This very large racial differential in arrest rates is clarified by analyzing street-specific arrest patterns. On Eight Mile Road, Black drivers accounted for thirty-five percent of the traffic stops (Table 2), and 67.1 percent of the arrests. On Nine and Ten Mile Roads, in contrast, where 7.8 percent of the motorists stopped and ticketed by the police were Black (Table 2), 33.6 percent of those arrested were Black. While Black motorists on Eight Mile Road were arrested ninety-two percent more often than Whites in the course of routine traffic stops, Black drivers stopped on Nine and Ten Mile Roads were 331 percent more likely than Whites to be arrested.
How much discretion did the Eastpointe police exercise in arrest situations? Once motorists are stopped, the police are required to initiate a computer search to explore such issues as driver's license and registration validity, as well as the existence of outstanding warrants. Outcomes of this computer search exercise accounted for over fifty percent of the charges filed against arrested drivers. Nonetheless, the discretionary range was extraordinarily broad. Consider the case of searches of motorists and their vehicles: Eastpointe police were not hindered by probable cause constraints in deciding whom to search. Absent probable cause, the police were permitted to search drivers, passengers, and/or vehicles whenever they felt that their personal safety might be at risk. The Eastpointe arrest narratives described 1027 searches (611 person searches and 416 vehicle searches) and 45.2 percent of those searches targeted Black motorists. Note that whenever a search was conducted, the police were required to complete an arrest narrative, irrespective of the outcome. For this reason, arrest narratives were written when search outcomes were negative and no arrest was made; no arrest took place in eight percent of the narratives describing White motorists and six percent of those involving Blacks.
Based upon the initial traffic violation causing the Eastpointe police to stop motorists, it was difficult to predict who would be the target of police searches. Only two types of common traffic violations—swerving and improper lane use—were positively correlated to subsequent person and vehicle searches. Improper lane use refers to situations in which a vehicle is straddling two lanes. Swerving, in particular, was disproportionately cited by the police for initiating stops resulting in drunk driver arrests. Analysis of arrest patterns indicated that drunk drivers were much more likely to be searched than sober ones. Having an outstanding warrant did not correlate closely with being searched.
The fact that Blacks made up less than twenty-two percent of motorists stopped by the Eastpointe police and 45.2 percent of those searched, finally, suggests the police believed that Blacks were more likely than Whites to possess illegal contraband. While search and arrest differentials indicated disproportionate focus upon Blacks, an alternate explanation was that Eastpointe police were colorblind and simply arrested lawbreakers. Disproportionate Black criminality rather than racial bias might explain police actions. Their analysis of state police searches of drivers along Interstate 95 in Maryland led Knowles et al. (Reference Knowles, Persico and Todd2001) to conclude that disproportionate searches of Black motorists represented efficient use of scarce police resources, not racial discrimination. The fact that Black drivers were only eighteen percent of the drivers on I-95, yet were sixty-three percent of those searched was interpreted as nondiscriminatory because searches of Blacks and Whites yielded roughly equal contraband find rates. “If police are prejudiced, the equilibrium returns to searching members of the group that is discriminated against will be below average”, assert Knowles and his co-authors (Reference Knowles, Persico and Todd2001, p. 207).
This “ends justify the means” rationale is not without its critics (Myers Reference Myers2002). Ramirez et al. (Reference Ramirez, McDevitt and Farrell2000) contend that equity should trump efficiency in guiding police actions: “Regardless of whether the perception that Blacks and Latinos are more likely to be found in possession of contraband could be empirically verified, United States laws do not, and should not, permit race to be used as a basis for stopping and searching individuals” (p. 11). Knowles et al. (Reference Knowles, Persico and Todd2001), to their credit, have provided a concrete standard for evaluating whether police behavior is, on economic efficiency grounds, prejudiced regarding search of suspects. Racial prejudice is present, under their standards, if the police are sacrificing arrests by disproportionately searching members of a racial group characterized by a below-average contraband find rate. This standard, of course, assumes that police behavior is guided by their desire to arrest lawbreakers when, in fact, the police themselves may not embrace this as their sole objective. Possible alternative motivations include harassing Black motorists and/or enforcing racialized urban space.
Eastpoint's arrest narratives were sufficiently detailed to permit investigation of what police actions implied about their guiding motivations. Straightforward racial comparisons of contraband find rates are the first order of business. The 1027 searches of persons and vehicles described in the arrest narratives of the Eastpointe police produced 554 contraband finds, with possession of illegal drugs accounting for more charges filed against drivers and their passengers than all other contraband sources (open intoxicants, drug paraphernalia, weapons) combined. Numbers of searches and search outcomes are summarized, for Whites and Blacks, in Table 5. Of the 611 person searches, 327 involved Whites, producing 181 finds of illegal contraband; 284 searches of Blacks generated ninety-four contraband finds (Table 5). The find rates of illegal contraband, expressed in percentage terms, were 55.4 percent for Whites and 33.1 percent for Blacks. Stated differently, 611 person searches revealed that Whites were 1.7 times more likely than Blacks to be found in possession of illegal contraband.
Table 5. Police Searches and Findings of Contraband, by Race
Source: Eastpointe police arrest narratives.
Vehicle searches produced higher contraband find rates, 74.2 percent for Whites and 57.8 percent for Blacks, indicating that Whites were 1.3 times more likely than Blacks to be found possessing illegal contraband. Some contraband finds did not ultimately lead to criminal charges being filed, and Blacks found in possession of contraband were slightly more likely than Whites not to be charged. In the case of illegal drug possession, 284 charges were ultimately filed against arrested drivers and their passengers—210 Whites were thusly charged as were seventy-four Black motorists. In this example, Blacks—targeted by over forty-five percent of the applicable police searches—ultimately accounted for 26.1 percent of charges for possession of illegal drugs found in the course of person and vehicle searches.
The search outcome data summarized in Table 5 destroy the “efficient policing” rationale for disproportionate police searches of Black motorists in Eastpointe. Ayres (Reference Ayres2001) has observed that “a defense that police searching decisions were driven by the underlying criminality of those searched—and that minorities make up a larger proportion of those deserving to be searched—would be contradicted by systematically lower success rates when such searches were in fact completed” (p. 134). Knowles and his co-authors (Reference Knowles, Persico and Todd2001), applying their standards concerning prejudice and efficient allocation of police resources to the Eastpointe police, would conclude that they utilized their scarce resources inefficiently by over-searching the low-find group, Black motorists, thus sacrificing arrests for the sake of practicing racial discrimination. Ayres and Borowsky (Reference Ayres and Borowsky2008) observed this same discriminatory pattern in their analysis of searches conducted by the Los Angeles police department. Black motorists driving in Eastpointe were probably aware of this long ago, which may explain why they exercise caution when driving through suburbs like Eastpointe.
Examining all charges filed against arrested drivers, charges for possession of illegal contraband were proportionally over twice as frequent for arrested Whites, in comparison to Black arrestees. If actual contraband find rates guided future police behavior, the Eastpointe police, over time, would reallocate their search efforts toward Whites and away from Blacks. Yet, search patterns may not be motivated by police desires to maximize arrest rates. The data on police search patterns and outcomes certainly clarify one issue. The notion that Eastpointe police were colorblind and simply strived to identify and arrest lawbreakers is untenable.
Recall that Black drivers stopped and ticketed by the Eastpointe police on Eight Mile Road accounted for thirty-five percent of all traffic stops and 67.1 percent of all motorist arrests. Relative to White drivers, overrepresentation among those arrested was huge. Yet, in an important sense, the Black arrest incidence observed on Eight Mile Road was low. Specifically, 46.5 percent of all police stops of Black drivers in Eastpointe occurred on Eight Mile Road, yet only 32.6 percent of Black motorist arrests occurred on that same street. Interior streets of Eastpointe, in contrast, accounted for 53.5 percent of Black drivers stopped and 67.4 percent of arrests. Police searches followed this same pattern.
Predictably, Black motorist searches on Nine and Ten Mile Roads were proportionally much more numerous, relative to Eight Mile Road patterns. Blacks stopped on Nine and Ten Mile Roads were nearly six times more likely than Whites to be subjected to person searches. Recall that drivers charged with intoxication offenses were more likely than others to undergo searches by the Eastpointe police. A possible inference is that Black drivers stopped on Nine and Ten Mile Roads were intoxicated more often than Whites. Yet, the facts from police narratives indicated otherwise: Black drivers accounted for under three percent of the intoxication arrests on Nine and Ten Mile Roads, yet they were nonetheless searched disproportionately. Contraband find rates—for Blacks searched after being stopped by the police on Nine and Ten Mile Roads—were actually lower (under twenty-five percent for person searches) than the overall find rates reported in Table 5. The fact that police searches targeted Black motorists most heavily precisely in areas of Eastpointe where contraband find rates were lowest suggests that the Eastpointe police pursued objectives other than law enforcement.
If we assume that the Eastpointe police are merely enforcing laws (as opposed to enforcing racialized space), we are left with a mystery in attempting to understand why Black motorists, particularly those stopped on Nine and Ten Mile Roads, were arrested much more frequently than Whites. Eastpointe police narratives did not specifically identify causes of arrests. Rather, they listed outcomes of police search activities as well as computer searches of driver registration, driver's license validity, outstanding warrants, and the like, and they listed charges filed at the time of arrest. Individual arrest narratives often listed offenses that automatically triggered arrests, yet they frequently listed only the types of charges where arrest was discretionary. Thus, drivers with invalid driver's licenses might have been arrested or they might merely be ticketed, after which they were free to proceed. No concrete rules delineated whether arrests would or would not occur in such instances. In this discretionary arrest realm, we finally locate causes of the disproportionate arrests of Black motorists by the Eastpointe police.
Table 6 summarizes the most common offenses described in police arrest narratives, comparing charges filed against Black and White drivers. These data exclude passenger arrests. Among arrested drivers, Whites and Blacks were charged with driver's license violations more often than any other offense (Table 6). The second most common charge for arrested Whites was illegal contraband possession, while among Black drivers, contraband offenses were the fifth most common charge; second in frequency was vehicle registration offenses. Overall, White overrepresentation was most pronounced for serious charges, including driver intoxification, contraband possession, and outstanding warrants, while charges against Blacks disproportionately involved driver's license issues, vehicle registration problems, and no proof of insurance. Whites disproportionately were arrested for crimes generating arrests automatically; Blacks were disproportionately arrested where arrest decisions were discretionary.
Table 6. Charges Most Often Filed against Arrested Drivers
Source: Eastpointe police arrest narratives
At the time of arrest, a final discretionary decision was whether to handcuff the arrestee. Despite the clear pattern of lesser charges, Blacks were handcuffed relatively more often than Whites. Indeed, Whites arrested on weapons violations were less likely to be handcuffed than Blacks arrested for having suspended driver's licenses.
CONCLUSIONS
In metropolitan Detroit, scholars have often observed that geographic space is racialized in the sense that Black Americans are not welcome in many suburban communities. This extends beyond housing segregation: Black drivers are not wanted on many suburban streets. While such racial animosity has been noted, the mechanisms utilized to enforce spatial restrictions—to “protect” specific areas from Black presence—have rarely been quantified and analyzed. Utilizing ticketing and arrest data collected from observation and Eastpointe Michigan police records, this study investigated whether police behavior could accurately be characterized as enforcing racialized space. Apart from their broader tendency to harass Black motorists across the broad, the police in this working-class suburb differentiated their treatment of Black motorists, depending upon whether they were stopped in more-White as opposed to less-White areas of Eastpointe. Police behavior was consistent with enforcing racialized space, signaling where Black presence was tolerated and where it was discouraged.
These conclusions are broadly consistent with findings of previously discussed studies of police/Black motorist interactions utilizing different data sources and methodologies. In the national survey data analyzed by Engel and Calnon (Reference Engel and Calnon2004), Black drivers were found to be more likely than Whites to be stopped, cited, searched, and arrested by the police in the course of traffic stops. Based upon their analysis of police records, Meehan and Ponder (Reference Meehan and Ponder2002) reported that Black drivers attracted greater police attention in suburban areas described as more as opposed to less White.
Findings of Knowles et al. (Reference Knowles, Persico and Todd2001) indicate that disproportionate police searches of Black motorists traveling on I-95 in Maryland were nondiscriminatory in the limited sense that contraband find rates were roughly equal for White and minority motorists searched by the police. While this may appear to be inconsistent with the findings of this study, in fact the context in which Knowles et al. (Reference Knowles, Persico and Todd2001) conducted their analysis was markedly different. The notion of enforcing racially defined space—a primary focus of this study—could not be explored by Knowles and his co-authors in the context of a statewide stretch of interstate highway.
Indeed, a specific contribution of this study is its precise specification of geographic context and its relevance to understanding the nature of police responses to Black motorists. Clearly, police treatment of African American drivers varies markedly in differing contexts. In contrast, studies using nationwide, statewide, even citywide data to probe the nature of interactions between Black drivers and the police cannot address the issues of police motivation addressed here. Driving while Black in a White neighborhood in suburban Detroit differs substantively, regarding police reaction, from being Black in an area where such presence is not contested. Beyond mere enforcement of the law, the police in Eastpointe include among their duties the enforcement of racialized space.
Three findings stand out. First, Black motorists driving in Eastpointe, when stopped and ticketed for traffic violations, were more likely than Whites to receive multiple traffic tickets. Logistic regression model results demonstrated that Black drivers were significantly more likely than Whites to receive multiple tickets when statistical controls for variety of factors—age of vehicle, time of day, and type of traffic stop—were introduced into the analysis of ticketing patterns.
Second, Black motorists were stopped, ticketed, searched, and arrested on Nine and Ten Mile Roads far out of proportion to their presence on those streets. Although Blacks made up 5.3 percent of the drivers on Nine and Ten Mile Roads, they accounted for 7.8 percent of the stops, received 10.6 percent of the multiple tickets, and accounted for 33.6 percent of motorists arrested by Eastpointe police officers on these streets. Under a variety of alternative examinations of ticketing, search, and arrest patterns on Nine and Ten Mile Roads, disproportionate ticketing, searches, and arrests were consistently the fate of Black drivers.
Third, disproportionate arrests of Blacks coexisted with White motorist dominance of the most common serious offenses causing driver arrests by the Eastpointe police. Evidence of disproportionate Black criminality in both the police search and arrest data was utterly lacking. Analysis of police searches of persons and vehicles originating in traffic stops documented that Black motorists, on a per search basis, were much less likely than Whites to possess illegal contraband. The offense of driving while Black on interior streets of Eastpointe clearly increased one's chances of attracting unwanted police attention.
ACKNOWLEDGMENTS
The comments and suggestions of George Galster materially improved this study, as did the research assistance of Robin Smith, the staff of the Center for Urban Studies at Wayne State University, and Yolanda Jones.
APPENDIX A: CAR COUNTING
Car-counting sessions conducted along Eight, Nine, and Ten Mile Roads began on October 26, 1999 and ended on January 9, 2000. Altogether, eighty two-hour sessions were conducted on Eastpointe streets, and the race of the driver was noted for 41,830 cars. Drivers who could not be classified as Black or White (the unclassified cases) were 2.8 percent of the 41,830 drivers. In these sessions, one person recorded driver race for vehicles passing in the outside lane and one person recorded driver race for vehicles in the inside lane.
The methodology of recording race of driver for all cars driving on a particular road during specific two-hour time periods, in order to establish the racial composition of the traffic flow, was originated by John Lamberth (Reference Lamberth1994) and applied by Lamberth in an analysis of police stops and arrests of Black motorists traveling on the New Jersey Turnpike. This methodology is used to generate a frame of reference for testing whether the police disproportionately ticket, search, and arrest Black motorists on specific streets, relative to the percentage of Blacks in the overall motorist population on those streets.
Fifty-six car-counting sessions were conducted on Nine Mile and Ten Mile Roads and 23,225 drivers were classified as Black or White. On Eight Mile Road, finally, 17,437 drivers were classified as Black or White in twenty-four sessions. In the case of commercial vehicles and government vehicles, such as UPS trucks and police cars, race of driver was not recorded.
Because driver race patterns varied by time of day and day of week, the car-counting sessions were scheduled across different times of the day and different days. Because driver race patterns may vary, as well, according to the direction of traffic or the location along a street, car-counting sessions were conducted at different locations along each street under consideration. Car-counting sessions were conducted on twenty-two different days, including four days in October, seven days in November, nine days in December, and two days in January. Car-counting sessions were scattered to cover different times of day. On Nine Mile Road, for example, car-counting sessions began at 2:00 am, 6:00 am, 7:15 am, 8:00 am, 10:00 am, 12:30 pm, 2:00 pm, 3:30 pm, 5:00 pm, 5:40 pm, 7:00 pm, 8:40 pm, 10:00 pm, and 11:00 pm. The purpose of spreading the car-counting sessions over different times of the day, days of the week, months of the year, of course, was to measure the racial composition of drivers on the Eastpointe streets under a wide variety of conditions.
APPENDIX B: TRAFFIC TICKETS
Information coded from Eastpointe traffic tickets and arrest narratives was used to construct explanatory variables that were used in the logistic regression models summarized in Table 4 of this report. Exact definitions of those explanatory variables are spelled out below:
Age of Vehicle
a. New car: for vehicles that were less than six years old when the traffic ticket was issued in Eastpointe, new car = one, zero otherwise.
b. Middle-age car: for vehicles that were no more than ten years old, but at least six years old when the traffic ticket was issued, middle-age car = one, zero otherwise.
c. Old car: for vehicles that were eleven or more years old when the traffic ticket was issued, old car = one, zero otherwise.
Radar Stop
For drivers getting speeding tickets as a result of being pulled over in a radar stop in Eastpointe, this variable equals one, zero otherwise.
Late Stop
For drivers ticketed between the hours of 9:00 pm and 6:00 am, late stop = one, zero otherwise.
Weekend Stop
For drivers ticketed on Saturday or Sunday in Eastpointe, weekend stop = one, else zero.
Street of Stop
Eight Mile Road: for drivers ticketed on this street, this variable = one, else zero.
Nine or Ten Mile Road: for drivers ticketed on either Nine Mile Road or Ten Mile Road, this variable = one, else zero.
Black Driver
Arrest narrative data in all cases identified drivers as Black, White, or other. Traffic tickets sometimes identified race of driver and sometimes not. Drivers not identified by race on traffic ticket or arrest narrative data were identified as Black through the technique of address geocoding at the residential block level; this variable is defined for drivers who have at least an eighty percent chance of being Black. Drivers identified as Black by the above techniques lead to the definition of this variable as Black = 1, else zero.
Note, finally, that my analysis of traffic tickets suffers from a consistent bias, causing, most likely, a systematic understating of Black/White ticketing differentials. Drivers are frequently stopped by the police and given warnings but no tickets. Because I had no data on stops resulting in warnings but not tickets, I implicitly assumed that this phenomenon equally impacted Black and White motorists stopped by the Eastpointe police. In fact, evidence from national survey data indicate that White drivers are more likely than Blacks to receive warnings but no tickets in the course of traffic stops (Engel and Calnon, Reference Engel and Calnon2004).
