Opinion
CASE NO. 3:16-cv-00018
04-20-2018
RODNEY HUBBARD, ET AL., Plaintiffs, v. ANDREW HOLMES, ET AL., Defendants.
MEMORANDUM OPINION
The plaintiffs in this case allege that Andrew Holmes, an Albemarle County police officer, violated the Fourteenth Amendment's Equal Protection Clause through racially selective enforcement of the law. Specifically, Rodney Hubbard and his mother Savannah Hubbard claim that Holmes stopped and searched their car because they are African-Americans. The defendants, Holmes and Albemarle County, have moved for summary judgment.
The plaintiffs have a Monell claim against Albemarle County. The claim's vitality rises and falls with the claim against Holmes.
The parties focus on a narrow question: Are certain statistics about Holmes' prior arrests and summonses sufficient to establish that Holmes' actions had a discriminatory effect? If not, plaintiffs' case falters because a selective enforcement claim requires the plaintiffs to establish both that the defendant's actions had a discriminatory effect and that those actions were motivated by a discriminatory purpose. Here, the statistics are the only evidence of discriminatory effect relied on by the plaintiffs. The defendants are entitled to summary judgment because this statistical evidence is insufficient to demonstrate discriminatory effect.
I. LEGAL STANDARD
Federal Rule of Civil Procedure 56(a) provides that a court shall grant summary judgment "if the movant shows that there is no genuine dispute as to any material fact." "As to materiality . . . [o]nly disputes over facts that might affect the outcome of the suit under the governing law will properly preclude the entry of summary judgment." Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986). In order to preclude summary judgment, the dispute about a material fact must be "genuine." Id. A dispute is "genuine" if "the evidence is such that a reasonable jury could return a verdict for the nonmoving party." Id. In considering a motion for summary judgment under Rule 56, a court must view the record as a whole and draw all reasonable inferences in the light most favorable to the nonmoving party. See, e.g., Celotex Corp. v. Catrett, 477 U.S. 317, 322-24 (1986).
II. FACTS
A. The stop
On September 11, 2015, Rodney Hubbard and his mother were driving north on Route 29 in a black sport utility vehicle. As they entered Albemarle County, Hubbard passed Holmes, who was standing outside his patrol car. Hubbard slowed down when passing Holmes. A mile or two later, Hubbard noticed that Holmes was now in his patrol car and following Hubbard. Holmes followed Hubbard for three to four miles before passing him. After another mile or two, Hubbard again saw that Holmes was again following him. At that point, Holmes turned on his lights and pulled Hubbard over. Holmes claimed Hubbard had driven 65 mph in a 60 mph zone.
Holmes then asked Hubbard for his driver's license and registration. Hubbard did not have his license, but gave him an identification card. Holmes asked Hubbard to step out of the vehicle. Holmes told Hubbard he stopped Hubbard for exceeding the speed limit, although Hubbard denies he was speeding. Holmes asked Hubbard whether his license was suspended in Virginia. Hubbard was unsure, but said his license in Maryland was expired. (During the stop, a different officer later informed Holmes that Hubbard's Virginia license was indeed suspended).
Holmes said he smelled marijuana and asked Hubbard why the vehicle smelled like marijuana. Hubbard denied the vehicle smelled like marijuana and said there was no marijuana in the car, but did admit that someone had smoked in the vehicle three or four days before the stop. However, Hubbard had placed an air freshener in his car to conceal the marijuana smell from his mother. In her deposition, his mother denied smelling marijuana. Holmes told Hubbard he was not under arrest, but searched him due to the marijuana odor Holmes claimed he smelled in Hubbard's vehicle. Holmes then placed Hubbard in handcuffs in the patrol car while Holmes searched Hubbard's vehicle. Other officers arrived to assist in the search of the vehicle. The officers did not find drugs. Holmes issued Hubbard a ticket for driving on a suspended license.
B. The statistics
In response to a discovery request in a related case, the defendants produced statistical data about Albemarle County and its police officers. The data show the racial breakdown of Albemarle County, a more specific racial breakdown of two sectors of Albemarle County where Holmes primarily worked, Holmes' total traffic stops (although these stops are not broken down by race), the racial breakdown of the summons and arrests issued by Holmes, and the racial breakdown of citations and arrests for other officers in Holmes' primary sectors for 2015.
The statistics state that the population of Albemarle County is 81% white and 10% black. Holmes normally does not cover the whole of Albemarle County, but is usually assigned to three police "sectors." The population of two of those sectors was approximately 68% white and 18% black. The parties do not have specific data on the racial breakdown of the third of the sectors.
While statistics are kept on total traffic stops, no accompanying demographic data are available for those stops unless a summons or citation is issued. In 2014 and 2015, 17% of summons issued by the Albemarle County Police Department were to black individuals. From 2009 to 2015, 49% of the summonses Holmes issued were to black individuals, and 51% were to white individuals. In 2015, 51% of the summonses Holmes issued were to black individuals and 48% were to white individuals. Other officers in two of Holmes' three sectors issued 22% of summonses to black individuals and 74% to white individuals in 2015. In 2015, 61% of Holmes' arrests were of black individuals and 39% were of white individuals.
III. ANALYSIS
The central question in this case is whether a jury could reasonably rely on these statistics to find that Holmes' actions had a discriminatory effect. The Court first takes a step back to situate that question within the broader law governing selective enforcement claims. The Court then responds to the plaintiffs' disagreements with that law. Finally, the Court asks whether the above statistics can satisfy the standard required by that law.
A. Applicable law
To prove a selective enforcement claim, plaintiffs must present evidence "demonstrat[ing] that the government's enforcement process 'had a discriminatory effect and that it was motivated by a discriminatory purpose.'" Cent. Radio Co. Inc. v. City of Norfolk, Va., 811 F.3d 625, 634-35 (4th Cir. 2016) (quoting Wayte v. United States, 470 U.S. 598, 608 (1985)); United States v. Mason, 774 F.3d 824, 829-30 (4th Cir. 2014). The parties here are only concerned with the first of these elements. There are two primary ways to demonstrate discriminatory effect: (1) by pointing to similarly situated individuals of a different race who were treated differently by law enforcement, or (2) through the use of statistics which address that question. See Farm Labor Organizing Committee v. Ohio State Highway Patrol, 308 F.3d 523, 534 (6th Cir. 2002) ("A claimant can demonstrate discriminatory effect by naming a similarly situated individual who was not investigated or through the use of statistical or other evidence which 'addresse[s] the crucial question of whether one class is being treated differently from another class that is otherwise similarly situated.'"); United States v. Olvis, 97 F.3d 739, 744-45 (4th Cir. 1996) (looking at both other unindicted co-conspirators and statistical studies). Here, the plaintiffs attempt to prove their claim through the use of statistics.
Not just any statistics will do. See United States v. Venable, 666 F.3d 893, 903 (4th Cir. 2012) ("[A]bsent an appropriate basis for comparison, statistical evidence of racial disparity alone cannot establish any element of a discrimination claim. . . . [T]here is not a presumption that unexplained statistical evidence of racial disparity proves racial animus." (citations and quotation marks omitted)); Chavez v. Illinois State Police, 251 F.3d 612, 638 (7th Cir. 2001) ("Of course, parties may not prove discrimination merely by providing the court with statistical analyses. The statistics proffered must address the crucial question of whether one class is being treated differently from another class that is otherwise similarly situated."). To support a finding of discriminatory effect, the statistics must show similarly situated individuals of another race were treated differently than the plaintiffs. See United States v. Armstrong, 517 U.S. 456, 465 (1996) ("To establish a discriminatory effect in a race case, the claimant must show that similarly situated individuals of a different race were not prosecuted."); United States v. Hare, 820 F.3d 93, 99 (4th Cir. 2016) ("This Court has adopted Armstrong's standard for proving selective prosecution as the standard for proving selective enforcement."). Individuals are similarly situated when there are "no distinguishable legitimate enforcement factors that might justify making different enforcement decisions with respect to them." Hare, 820 F.3d at 99. Put differently, statistics relevant to this inquiry will show "similarly situated white individuals who could have been targeted for [the enforcement actions] but were not." Id. at 100.
Accordingly, if the above statistics do not show similarly situated individuals of another race were treated differently than the plaintiffs, the plaintiffs do not have sufficient evidence to make out the first element.
B. Whether to treat selective prosecution and selective enforcement claims alike
Plaintiffs argue for a different standard; they think selective enforcement claimants should not be required to offer a comparison of similarly situated individuals who were treated differently. While the plaintiffs acknowledge the similarly situated standard governs selective prosecution claims, they argue "selective enforcement cases are different, largely because it would be impossible to prove that similarly situated white drivers were not stopped and/or searched as there are no records of who is not stopped and who is not searched." (Dkt. 43 at 3). Plaintiffs rely on four cases for this proposition, but none changes the binding law of this circuit.
The first of these cases is Farm Labor Organizing Committee v. Ohio State Highway Patrol, 308 F.3d 523 (6th Cir. 2002). There, the Sixth Circuit stated that "[a] claimant can demonstrate discriminatory effect by naming a similarly situated individual who was not investigated or through the use of statistical or other evidence which 'address[es] the crucial question of whether one class is being treated differently from another class that is otherwise similarly situated.'" Id. at 524 (emphasis added). Plaintiffs' reliance on this language is confusing; the language strongly reiterates that any use of statistics must provide information about whether similarly situated classes are being treated differently. This language and standard is very similar to the above language the Fourth Circuit has used in both the selective prosecution and selective enforcement claims.
Second, the plaintiffs likewise rely on Chavez v. Illinois State Police, 251 F.3d 612 (7th Cir. 2001). The language cited by the plaintiffs from that case also incorporates the similarly situated requirement: "[P]laintiffs do not have to provide the court with the name of an individual who was not stopped; instead they may attempt to use statistics to show that [law enforcement] treated them differently than other motorists who were similarly situated." Id. at 640 (emphasis added). As with Farm Labor Organizing Committee, this language cuts against the plaintiffs' desired standard.
The Chavez court did go on to note that there may be differences between some readings of the Supreme Court's selective prosecution cases and selective enforcement cases. But in doing so, it was addressing a "possible but unnecessary" reading of Armstrong. Id. at 639. This reading of Armstrong would require criminal defendants "to provide the precise name of a similarly situated defendant who was not prosecuted," i.e., statistical evidence would not suffice in the criminal context under this reading. Id. Importantly, this hypothetical reading of Armstrong has been rejected by the Fourth Circuit. See, e.g., Olvis, 97 F.3d at 745. In any event, the court went on to hold that even if that reading was correct, statistics could still be used in civil selective enforcement suits if they provided information about a similarly situated class of individuals. Id. at 639 ("[S]tatistics demonstrating that whites stopped for traffic violations were not detained and searched, even those who displayed indicators of drug trafficking, while similarly situated African-American or Hispanics drivers were detained and searched, would be sufficient to show discriminatory effect."). This standard is no different than the one presented above and used by this Court.
Third, the plaintiffs cite Pyke v. Cuomo, 258 F.3d 107 (2nd Cir. 2001). In that case, Native American plaintiffs alleged the state police did not enter onto a reservation to protect them because of their race. Id. at 108. The court held "[t]he Armstrong rule [i.e., the selective prosecution rule] has no application to their claim," differentiating between "discriminatory application" claims and selective prosecution claims. Id. at 109. The court held that "[s]o long as [the plaintiffs] allege and establish that the defendants discriminatorily refused to provide police protection because the plaintiffs are Native American, plaintiffs need not allege or establish the disparate treatment of otherwise similarly situated non-Native American individuals." Id.
The problem for the plaintiffs, however, is that the Fourth Circuit has recently and expressly disagreed with the Pyke's statement that "[t]he Armstrong rule has no application to [selective enforcement] claim[s.]" In United States v. Hare, 820 F.3d 93 (4th Cir. 2016), which is also the fourth and final case the plaintiffs rely on, the Fourth Circuit stated that it "has adopted Armstrong's standard for proving selective prosecution as the standard for proving selective enforcement." Id. at 99 (citing United States v. Bullock, 94 F.3d 896, 899 (4th Cir. 1996)). This statement and the court's application of the selective prosecution standard in the selective enforcement context in Bullock are fatal to the plaintiffs' reliance on Pyke. Hare's citation of Bullock belies the plaintiffs' argument that Hare "raises questions about the continuing validity of Bullock." (Dkt. 43 at 4). Instead, Hare relied on cases from the selective prosecution context in faulting the criminal defendant's statistical evidence because it "provide[d] no appropriate basis for comparison, as it contain[ed] no data on similarly situated white individuals who could have been targeted for stash house sting investigations but were not." 820 F.3d at 99. This is the same standard applied by this Court: Statistics can be used to prove selective enforcement if they provide information about similarly situated classes of people who were treated differently.
After announcing that the standard for proving a selective enforcement claim is the same as that for a selective prosecution claim, the Hare court did note in dicta the potential for different standards for getting discovery on the two types of claims. The court quoted language from the Seventh Circuit's Chavez opinion that "plaintiffs who allege that they were stopped due to racial profiling would not, barring some type of test operation, be able to provide the names of other similarly situated motorists who were not stopped." Id. at 101. But, as the original context in Chavez demonstrates, the import of this quote is only that a plaintiff need not point to a specific individual and can instead use statistics that provided information about similarly situated individuals. 251 F.3d at 640. Hare also suggested that law enforcement officers are not entitled to the same deference federal prosecutors receive. 820 F.3d at 101. But, as with the Chavez quote, the court never came to any conclusion about what these differences would mean for access to discovery (because, in Hare, the defendant had received all the discovery he was entitled to, regardless of the success of these arguments). Id.
This Court will take Hare, and the Fourth Circuit caselaw more generally, at its word and treat selective prosecution and selective enforcement claims alike. It was in Bullock, as recognized by Hare, that the Fourth Circuit adopted the Armstrong standard for selective enforcement claims, and this Court is required to follow that binding authority. As in the selective prosecution context then, the plaintiffs' statistical evidence must put forward evidence about similarly situated individuals, those with "no distinguishable legitimate enforcement factors that might justify making different enforcement decisions with respect to them." Hare, 820 F.3d at 99.
While not relying on Armstrong or working through statistical evidence, the Fourth Circuit cases has also incorporated a similarly situated analysis in the civil context. See Sylvia Dev. Corp. v. Calvert Cty., Md., 48 F.3d 810, 825 (4th Cir. 1995); Orgain v. City of Salisbury, Md., 305 F. App'x 90, 100 (4th Cir. 2008).
C. Application
While the parties spend the majority of their energy focused on the law that makes up that standard, they spend little on its application to the facts of this case. In truth, that application is relatively straightforward. There are two primary reasons why plaintiffs' statistical evidence is insufficient to support a finding of discriminatory effect. First, the plaintiffs do not have "a means of telling whether the data represent similarly situated individuals." Marshall v. Columbia Lea Reg'l Hosp., 345 F.3d 1157, 1168 (10th Cir. 2003) (internal citations to Chavez omitted); Hare, 820 F.3d at 99 ("Appellants' statistical evidence similarly provides no appropriate basis for comparison, as it contains no data on similarly situated white individuals who could have been targeted for stash house sting investigations but were not."). And second, the plaintiffs do not provide "a point of comparison to the actual incidence of crime among different racial or ethnic segments of the population." Marshall, 345 F.3d at 1168; Venable, 666 F.3d at 903 ("Without an appropriate basis for comparison, the percentage of African American crack cocaine defendants proved nothing, unless it could be presumed that crack cocaine violations were committed proportionately by all races, an assumption rejected by the Supreme Court in Armstrong.").
i. Whether the data represent similarly situated individuals
First, the plaintiffs have not put forward sufficient statistical evidence about individuals of a different race who were similarly situated to the plaintiffs but treated differently. Farm Labor Organizing Committee, 308 F.3d at 534. None of the statistics offered by the plaintiffs provide a racial breakdown of Holmes' traffic stops and searches. And while there is a demographic breakdown of Holmes' summonses and arrests, the plaintiffs have not produced any evidence that provides a reasonable basis to infer that this data concern only traffic summonses or arrests. Instead, the data appear to include all summonses and arrests instigated by Holmes and other officers located in specific police sectors.
The exhibits themselves simply refer to "summonses" and "arrests," without reference to traffic. There is no context in the record that would allow a jury to conclude that these statistics represent or include only traffic summonses and arrests instead of all general summonses and arrests made in these jurisdictions. In a related case, this Court also repeatedly treated these statistics as concerning all summonses and arrests. See Johnson v. Holmes, 3:16-cv-00016, dkt. 57 at *6 (W.D. Va. Oct. 19, 2017); Johnson, dkt. 84 at *7 (W.D. Va. Mar. 19, 2018); Johnson, dkt. 91 at *3 (W.D. Va. Mar. 22, 2018).
Even were the statistics about all the traffic summonses and arrests, this level of specificity would still likely be insufficient under Armstrong. After all, individuals can receive summons or be arrested for many different things, even at a traffic stop, and so it is impossible to know from these bald stats how similarly situated the other individuals were. See United States v. Bass, 536 U.S. 862, 864 (2002) ("[R]aw statistics regarding overall charges say nothing about charges brought against similarly situated defendants." (emphasis in the original)).
Statistics about the racial demographics of Holmes' general summonses and arrests do not provide information about classes of people similarly situated to the plaintiffs. This is because that data cover individuals with "distinguishable legitimate enforcement factors that might justify making different enforcement decisions." Hare, 820 F.3d at 99. Namely, that data are not broken down by the crime individuals were charged with or the reasons someone was arrested. Of course, officers would reasonably respond differently to a 911 call about domestic violence than someone speeding.
The similarly situated analysis set out above requires something more rigorous than comparing Holmes to officers who were responding to different types of crimes for different sets of reasons. See, e.g., Hare, 820 F.3d at 99-100 ("While a Department of Justice press release indicates that the [comparators offered by the criminal defendant] were involved in drug trafficking and armed home invasions, it is not known, for example, whether the crew members' criminal histories indicated that they would be receptive to a stash house robbery scenario, or whether ATF had the means of infiltrating this crew undercover."); Venable, 666 F.3d at 902 ("In sum, given all of the circumstances and differentiating factors between Venable's case on the one hand, and Turner's and Zechman's cases on the other hand, Venable has failed to make a credible showing that a similarly situated defendant of another race has evaded prosecution under Project Exile in order to obtain discovery on his selective prosecution claim."); Olvis, 97 F.3d at 744 ("Making decisions based on the myriad of potentially relevant factors and their permutations require[s] the very professional judgment that is conferred upon and expected from prosecutors in discharging their responsibilities.").
The plaintiffs have more generally presented insufficient information about what exactly is included in these statistics. Do the comparison officers include all officers that have spent any time in the same areas as Holmes? Even if 99% of their time is elsewhere? Did these officers have similar assignments? Or were some tasked with targeting certain crimes? Do the listed summonses and arrests only include interdictions instigated by the listed officer? Or do they also include responses to dispatcher calls? The Court does not mean to imply that each and every one of these answers would need to be answered, but the plaintiffs would need to be able to provide more information than they have about what exactly is included in the statistics.
Relatedly, reliance on the summary statistics without expert testimony or statistical analysis is problematic. In various types of racial discrimination cases, courts have frequently found that raw statistical evidence requires guidance from the proponent's expert. "[I]f a plaintiff offers a statistical comparison without expert testimony as to methodology or relevance to plaintiff's claim, a judge may be justified in excluding the evidence." Carter v. Ball, 33 F.3d 450, 457 (4th Cir. 1994); e.g., Anderson v. Westinghouse Savannah River Co., 406 F.3d 248, 261-63 (4th Cir. 2005) (affirming exclusion of expert testimony based on statistics that failed to compare similarly situated workers, categorizing them by "group" rather than single job title); Luh v. J.M. Huber Corp., 211 F. App'x 143, 149 (4th Cir. 2006) (affirming summary judgment where plaintiff had introduced raw statistics breaking down a reduction-in-force by race, because "[t]here was no expert testimony as to methodology justifying the sought conclusion of discrimination or relevance of the statistics to the plaintiff's claim."). Because so little is known about the content or input data of these statistics, the lack of expert testimony would help explain what sort of deviations are or are not explained by innocent reasons. C.f. Brown v. Nucor Corp., 785 F.3d 895, 908 (4th Cir. 2015) (citing Hazelwood Sch. Dist. v. United States, 433 U.S. 299, 308 n. 14 (1977) for the proposition "that anything greater than two or three standard deviations in racial discrimination cases is suspicious, at least for large sample sizes").
Without more, a jury could not find that Holmes' stops and searches had a discriminatory effect on the plaintiffs. See Chavez, 251 F.3d at 645 ("[W]ithout reliable data on whom [the] officers stop, detain, and search, and without reliable data indicating the population on the highways where motorists are stopped, detained, and searched, we cannot find that the statistics prove that the []officers' actions had a discriminatory effect on the plaintiffs.").
ii. Whether the data provide a point of comparison to the actual incidence of crime among different racial segments of the population
The second problem with the plaintiffs' statistics is that they do not provide information about the rates at which individuals of different races commit the crimes underlying the interdictions. In Armstrong, the Supreme Court said the "presumption that all people of all races commit all kinds of crime" proportionally was "at war" with statistics from U.S. Sentencing Commission. 517 U.S. at 469-70. Courts have applied this logic to traffic offenses. See United States v. Alabi, 597 F. App'x 991, 997 (10th Cir. 2015) (noting the criminal defendant "failed to demonstrate drivers of different races speed with the same frequency"). Here, the plaintiffs' statistical evidence does not compare the arrests and summonses arising from Holmes and other officers to demographic data about the incidence of the underlying crimes.
This second problem further demonstrates the depth of the first. Without more information about what the crimes underlying the various summonses and arrests included in the data are, a jury could not know whether the different officers were responding to the same sorts of incidents. Even if one assumed they were, the jury would then have to assume that individuals were committing the various crimes that led to these incidents at the same rates. But, again, this inference is foreclosed by Armstrong and the Fourth Circuit cases that follow it. See Hare, 820 F.3d at 99 (discussing Olvis and Venable).
These flaws cannot be saved by the plaintiffs' presentation of general population data of Albemarle County. See Chavez, 251 F.3d at 644 ("Even if it were entirely accurate, however, Census data can tell us very little about the numbers of Hispanics and African-Americans driving on Illinois interstate highways, which is crucial to determining the population of motorists encountered by the Valkyrie officers."). And "absent an appropriate basis for comparison, statistical evidence [of racial disparity] alone cannot establish any element of a discrimination claim." Olvis, 97 F.3d at 745. Accordingly, the plaintiffs' statistics cannot bear the weight the plaintiffs put on them.
IV. CONCLUSION
In 2014, the Fourth Circuit noted "the Armstrong burden is a demanding one and [the criminal defendant] has failed to identify any cases at the Supreme Court or in this circuit where an Armstrong violation for selective law enforcement has been found." United States v. Mason, 774 F.3d 824, 830 (4th Cir. 2014); id. at 829-30 ("A selective law enforcement claim asks a court to exercise judicial power over a special province of the Executive." (citations and quotation marks omitted)). While written in the criminal context, plaintiffs run into similar problems here in the civil context. The plaintiffs' claims fail as a matter of law because their statistical evidence cannot support a finding of discriminatory effect.
An appropriate order will issue. The Clerk of Court is requested to send a copy of this Opinion and the accompanying Order to the parties.
Entered this 20th day of April, 2018.
/s/ _________
NORMAN K. MOON
SENIOR UNITED STATES DISTRICT JUDGE