Commonwealth v. Lora

Cordy, J.

In this case, in which we granted the defendant’s application for direct appellate review, we conclude that evidence of racial profiling1 is relevant in determining whether a traffic stop is the product of selective enforcement violative of the equal protection guarantee of the Massachusetts Declaration of Rights; and that evidence seized in the course of a stop viola-tive of equal protection should, ordinarily, be excluded at trial.2 We also conclude that statistical evidence demonstrating disparate treatment of persons based on their race may be offered to meet the defendant’s burden to present sufficient evidence of impermissible discrimination so as to shift the burden to the Commonwealth to provide a race-neutral explanation for such a stop. Finally, we conclude that the evidence proffered by the defendant fell short of what is necessary to overcome the presumption that a law enforcement officer making a traffic stop, based on probable cause, has acted in good faith and without intent to discriminate. Consequently, the evidence seized *427in this case should not have been suppressed.3

1. Background, a. Traffic stop.4 On the evening of December 20, 2001, State Trooper Brendhan Shugrue was patrolling Interstate Route 290 in Auburn (Route 290).5 At approximately 9:10 p.m., he approached a motor vehicle traveling ahead of him in the left lane. At the time, there was no traffic in the center lane or the right lane. The vehicle traveled within the speed limit, did not swerve, and made no erratic movements. It did not pass any vehicles because there was no traffic at the time.

Shugrue followed the vehicle for three-quarters mile to an area within several hundred yards of the border between Auburn and Worcester. He observed that the two occupants of the vehicle were dark skinned (the defendant, Andres Lora, is Hispanic). He activated the cruiser’s blue lights and stopped the vehicle for traveling in the left lane while the center and right lanes were unoccupied.6 Shugrue approached the vehicle from the passenger’s side and asked the driver for his license and registration. The driver explained that his license had been suspended and that the vehicle belonged to the passenger, Lora. Lora then produced his license and the vehicle’s registration.

Shugrue asked the driver to step out of the vehicle, and instructed him to sit in the back of his cruiser. Lora remained in the vehicle. Shugrue then checked the status of the driver’s license, as well as the status of Lora’s license and registration. He confirmed that the driver’s license was, in fact, suspended, but that Lora’s license and registration were both valid. As Shugrue was retrieving this information, he observed Lora get*428ting out of his vehicle while talking on his cellular telephone. Shugrue got out of the cruiser intending to instruct Lora to get back into his vehicle; as Shugrue approached, Lora got back into the vehicle and shut the door. Shugrue then directed his flashlight inside the vehicle, where he observed a small glossine bag on the driver’s side floor containing white powder.

Shugrue immediately asked Lora to step out of the vehicle and frisked him. He then retrieved the glossine bag, which appeared to contain cocaine, and radioed the State police barracks to request assistance. Trooper William Pinkes arrived at the scene ten to fifteen minutes later. The troopers then proceeded to search the vehicle.7 Shugrue discovered substantially more cocaine in the trunk.8 Pinkes then placed Lora and the driver under arrest.

A grand jury subsequently returned an indictment charging Lora with trafficking in cocaine in violation of G. L. c. 94C, § 32E (b).

b. Motion to suppress. On March 27, 2003, Lora filed a motion to suppress the cocaine as the fruit of an unconstitutional search.9 He contended that Shugrue initiated the traffic stop because the occupants of the vehicle were dark skinned, and not solely because the operator of the vehicle committed a traffic violation by driving in the left lane. In other words, he claimed that Shugrue impermissibly engaged in the practice of racial profiling. Thus, Lora asserted, the stop violated his right to equal protection under the law, as guaranteed by the Fourteenth Amendment to the United States Constitution and arts. 1 and 10 of the Massachusetts Declaration of Rights. He further asserted that a traffic stop predicated solely on the race of the passengers of a vehicle was inherently unreasonable and constitutionally *429infirm under the Fourth Amendment to the United States Constitution and art. 14 of the Massachusetts Declaration of Rights.

Lora sought to prove that the stop of the vehicle in which he was traveling was the product of racial profiling by establishing that Shugrue had a history of disproportionately stopping and citing nonwhite motorists for motor vehicle violations. To that end, defense counsel filed an affidavit stating that he had reviewed 25610 citations issued by Trooper Shugrue between August 22, 2001, and February 18, 2002.11,12 The affidavit pointed out that during this time period, Shugrue cited the operators of fifty-one vehicles driving on the stretch of Route 290 that passes through Auburn. Sixteen of the operators cited (or 31.37 per cent) were identified as Hispanic, and six (or 11.76 per cent) as African-American.13,14

Defense counsel then compared the percentage of citations issued to each racial group with the racial composition of the town of Auburn, as tabulated by the 2000 United States census (census). White residents accounted for 97.5 per cent of the population of Auburn; Hispanic residents, one per cent; and African-American residents, .6 per cent. Implicitly assuming that the demographics of the town of Auburn mirror the demographics of those driving on Route 290 through Auburn, defense counsel argued that Shugrue cited minority drivers at a rate wildly disproportionate to their representation in the local Auburn population. This type of comparison is known as census benchmarking. Using this analytical framework, defense counsel contended that a Hispanic driver would be 31.37 times more *430likely to be cited than an average motorist15,16; and an African-American driver would be 19.60 times more likely to be cited than an average motorist. By contrast, white drivers were only .5631 times as likely to be pulled over as an average motorist (or, in other words, about one-half as likely).

Defense counsel’s affidavit also included information regarding traffic citations issued by Shugrue along the stretch of Route 290 passing through the city of Worcester. In that same period, i.e., between August 22, 2001, and February 18, 2002, Shugrue cited eighty-nine motorists. Seventy-four, or 83.15 per cent, of the motorists cited were white; seven, or 7.87 per cent, were Hispanic; and six, or 6.74 per cent, were African-American. According to the census, white residents account for 77.1 per cent of the population of Worcester; Hispanic residents, 15.1 per cent; and African-American residents, 8.0 per cent. Defense counsel again used census benchmarking to compare the racial composition of the citations to the racial composition of the inhabitants of Worcester. Using the same implicit assumption that the demographics of Worcester accurately reflected the demographics of motorists driving on Route 290 in Worcester, defense counsel calculated that white motorists were 1.08 times more likely to be cited than an average motorist (that is, slightly more likely). Hispanic motorists were .52 times as likely to be cited as an average motorist (that is, about one-half as likely), while African-American drivers were .84 times as likely to be cited (that is, slightly less likely).17

*431At the hearing on the motion to suppress, Lora introduced Shugrue’s citation history, which was admitted de bene by the motion judge.18 Troopers Shugrue and Pinkes testified on behalf of the Commonwealth, contending that the traffic stop was motivated solely by the operator’s failure to keep to the right, and that the subsequent search was justified by the cocaine-filled glossine bag within the passenger compartment of the vehicle.19 After taking the matter under advisement, the judge allowed Lora’s motion to suppress. In his decision, the judge found that although “[a] strict reading of Massachusetts law supports that Shugrue was authorized to stop the [vehicle] for traveling in the left-hand lane,” the record left “no reasonable conclusion but that Shugrue stopped the motor vehicle in which the Defendant was a passenger because of the race of the operator and the race of the Defendant.” Therefore, the trooper “violated the Defendant’s rights pursuant to the Fourteenth Amendment . . . and Articles 1 and 10 . . . when he stopped and searched his vehicle.”

In reaching his conclusions, the judge noted that “Mas*432sachusetts appellate courts have yet to resolve the conflict that arises when a well-substantiated allegation of racial profiling is made within the context of a legitimate traffic stop” (emphasis in original). He ruled that the authorization approach to motor vehicle stops, under which a stop is deemed to be constitutionally valid as long as it is one that police are “legally permitted and objectively authorized” to make, Commonwealth v. Santana, 420 Mass. 205, 209 (1995), is inapplicable when a defendant is able to present objective evidence that the police officer who effectuated the stop has engaged in racial profiling. In this respect, the judge found that the evidence introduced by Lora, including Shugrue’s past citations and the statistical evidence derived therefrom, was “both credible and relevant to the case at bar,” creating an “inference of purposeful discrimination, which [was] not . . . rebutted by the Commonwealth” either with “a race-neutral explanation for the statistical disparities or [by] explaining] a compelling government interest in treating members of one race differently from another.” The judge found that Lora had sustained his burden of proof of discrimination “both by a preponderance of the evidence as well as clear and convincing proof.”

c. Motion for rehearing and the joint motion to vacate. Following the judge’s decision, the Commonwealth filed a motion for reconsideration.20 In its motion, the Commonwealth contended that the judge based his decision on documents (the citations of Shugrue and accompanying statistical analysis) that had not in fact been admitted in evidence; the judge wrongly admitted the documents in any event; and the judge’s findings as to racial profiling were based on mathematical and analytical errors. The motion for reconsideration was allowed, and the judge ordered the scheduling of another evidentiary hearing.

Before the evidentiary hearing took place, the Commonwealth shared with defense counsel the evidence that it anticipated presenting at the hearing, most prominently an affidavit by Donna M. Bishop of Northeastern University’s College of Criminal Justice, who holds a doctorate in criminal justice. In *433that affidavit, Bishop stated that she would testify that the “defendant’s data collection methodology [did] not satisfy any accepted scientific standard,” and that the “statistics [the defendant introduced] failed to measure the demographics of the motorists traveling on the interstate highway.” She further averred that “[t]he driving population, rather than population demographics for the cities through which roads pass, is a more appropriate benchmark in evaluating police stop data.” After defense counsel had the opportunity to review Bishop’s affidavit and discuss its contents with Lora, he joined the Commonwealth in filing a motion to vacate the suppression order, expressing the parties’ “agreement that the suppression order should not have entered,” that “the statistics relied upon in the suppression order do not show that Trooper Shugrue’s decision to stop the defendant’s car was motivated by the race of the occupants,” and that “any further litigation [on the issue] is unnecessary.”

The judge held a hearing on the joint motion to vacate the suppression order, during which he suggested that both parties were “looking to duck the issue”; that the Commonwealth was looking to avoid “an adverse judgment against it”; and that the defendant was using the Commonwealth’s reluctance to litigate the issue to “bargain” for a “good deal.” Consequently, the judge declined to rule on the joint motion and directed both parties to present expert testimony at the rehearing.

At the rehearing, the defense called Scott Evans, a research scientist in the department of biostatistics at Harvard School of Public Health. In preparing for his testimony, Evans reviewed only the data regarding the citations issued by Shugrue in Auburn, and compared that information to the population demographics of that town. On the basis of that data, he concluded that the chance that race was not a factor in the disparate stop rates in Auburn was “very, very minute.”21 On cross-examination by the Commonwealth, Evans admitted that the assumption underlying his conclusion was that the demographics of the town of *434Auburn reflected the demographics of motorists on Interstate Route 290, and admitted that he had not considered the stops that occurred in Worcester.

The Commonwealth called Bishop, who testified that the census benchmarking technique used by Lora is “outmoded and no longer accepted within the scientific community because ... it is highly likely to yield misleading and erroneous conclusions.” She testified that census benchmarking assumes that the “demographic profile of the community [is] nearly identical to the demographic profile of drivers on the road where the officer was patrolling,” an assumption that is “not accurate.”22 Bishop further testified that Shugrue’s citations themselves demonstrate the flaw in census benchmarking: “I examined each of the citations that were given by Trooper Shugrue of [sic] motorists in Auburn . . . ninety per cent of those motorists were not from Auburn, which is conclusive evidence that a residential or a census benchmark coming from Auburn is totally inappropriate [as a point of comparison].” She went on to explain that an appropriate benchmark must include “a profile of the people who should be legitimately at more risk of being stopped and then compare that to the profile of who is actually stopped,” and ideally “roadside observers [would be used to] develop a demographic profile of people who violate traffic laws in a particular location where police are patrolling and stops are being made. In that instance, the benchmark would represent fairly well the group of people who should be at risk for being stopped if no bias exists.”23

*435The Commonwealth also called Joseph Petruccelli, a professor of mathematical sciences at Worcester Polytechnic Institute with a Ph.D. in statistics. Petruccelli testified that while the methodology used by Lora’s expert is an established tool of statistical analysis, its results are valid only when the appropriate data are fed into it. The census data used here made the statistical study scientifically unacceptable.

On December 2, 2005, the judge issued his decision affirming the earlier suppression order and reiterating his conclusion that Lora “met his burden by offering evidence of an inference of selective law enforcement and that the Commonwealth has not sufficiently rebutted [that] evidence with a neutral explanation.” The Commonwealth filed a notice of appeal, as well as a supplemental application for an interlocutory appeal. The single justice allowed the Commonwealth’s interlocutory appeal.

2. Discussion. On appeal, the Commonwealth first argues that even assuming Lora could prove that Shugrue stopped the vehicle in which Lora was riding because of the race of its occupants, his subjective intent is irrelevant to the legality of the stop and the subsequent search. In support of its position, the Commonwealth points to our holding that “it is irrelevant whether a reasonable police officer would have made the stop but for [an] unlawful motive; the stop is valid ‘so long as the police are doing no more than they are legally permitted and objectively authorized to do.’ ” Commonwealth v. Santana, 420 Mass. 205, 209 (1995), quoting United States v. Trigg, 878 F.2d 1037, 1041 (7th Cir. 1989), cert. denied sub nom. Cummins v. United States, 502 U.S. 962 (1991). The Commonwealth also argues that the United States Supreme Court’s decision in Whren v. United States, 517 U.S. 806, 813 (1996), “foreclose^] any argument that the constitutional reasonableness of traffic stops depends on the actual motivations of the individual officers involved.”

The Commonwealth reads too much into these cases. Both involved traffic stops that were objectively valid but motivated by an interest in investigating more serious crimes for which the officers had insufficient suspicion to justify an investigative stop. The defendants contended that if the traffic stops would not have been made by a reasonable police officer motivated to *436make the stop solely by a desire to enforce the traffic laws, they were “unreasonable” and therefore in violation of the Fourth Amendment right to be secure from unreasonable searches and seizures. Both courts rejected this argument, concluding essentially that the decision to stop an automobile is reasonable for Fourth Amendment purposes “where the police have probable cause to believe that a traffic violation has occurred.” Whren v. United States, supra at 810. See Commonwealth v. Santana, supra.

Neither case involved a challenge to the traffic stops based on equal protection grounds. Indeed, in Whren v. United States, supra at 813, the Court specifically noted, “We of course agree with petitioners that the Constitution prohibits selective enforcement of the law based on considerations such as race. But the constitutional basis for objecting to intentionally discriminatory application of laws is the Equal Protection Clause, not the Fourth Amendment. Subjective intentions play no role in ordinary, probable cause Fourth Amendment analysis.” Our holding in Commonwealth v. Santana, supra, is not to the contrary.

a. Selective enforcement. Although Lora contends that the issue of racial profiling is novel, it is at base a claim that Shugrue selectively enforced the laws in contravention of the Fourteenth Amendment and arts. 1 and 10, a claim that we have considered in other cases.

It is well established that “[tjhe equal protection principles of the Fourteenth Amendment. . . and arts. 1 and 10 . . . prohibit discriminatory application of impartial laws.” Commonwealth v. Franklin Fruit Co., 388 Mass. 228, 229-230 (1983). See Cote-Whitacre v. Department of Pub. Health, 446 Mass. 350, 376 (2006). See also Yick Wo v. Hopkins, 118 U.S. 356, 373-374 (1886).24 “It is equally well established, however, that prosecutors and other law enforcement officers enjoy considerable *437discretion in exercising some selectivity for purposes consistent with the public interest.” Commonwealth v. Franklin, 376 Mass. 885, 894 (1978). See Commonwealth v. King, 374 Mass. 5, 22 (1977) (law enforcement officials necessarily have wide discretion in determining whether to prosecute under prostitution statutes). The “ ‘conscious exercise of some selectivity’ in criminal law enforcement” is permitted “as long as the selectivity is not based on ‘an unjustifiable standard such as race, religion or other arbitrary classification.’ ” Id. at 20, quoting Oyler v. Boles, 368 U.S. 448, 456 (1962). “Following these principles, courts have reversed criminal convictions and dismissed criminal complaints as violative of equal protection principles when they found selectivity in prosecution based on impermissible standards.” Commonwealth v. King, supra at 20. See Commonwealth v. Franklin, supra at 906-907 (cases remanded for hearing on defendants’ motions to dismiss based on allegation that only blacks and not whites were prosecuted for violence in housing project).

An arrest or prosecution based on probable cause is ordinarily cloaked with a presumption of regularity. “Because we presume that criminal prosecutions are undertaken in good faith, without intent to discriminate, the defendant bears the initial burden of demonstrating selective enforcement.” Commonwealth v. Franklin, supra at 894. See Commonwealth v. King, supra at 22. In order to meet this burden, the defendant must first “present evidence which raises at least a reasonable inference of impermissible discrimination,” including evidence that “a broader class of persons than those prosecuted has violated the law, . . . that failure to prosecute was either consistent or deliberate, . . . and that the decision not to prosecute was based on an impermissible classification such as race, religion, or sex” (citations omitted). Commonwealth v. Franklin, supra at 894. See Commonwealth v. Franklin Fruit Co., 388 Mass. 228, 230-231 (1983) (adopting *438same “tripartite burden” in selective enforcement case involving national origin). Once a defendant has raised a reasonable inference of selective prosecution by presenting credible evidence that persons similarly situated to himself have been deliberately or consistently not prosecuted because of their race, “the Commonwealth must rebut that inference or suffer dismissal of the underlying complaint.” Commonwealth v. Franklin, supra at 895.

b. Suppression as remedy. The Commonwealth objects to the use of the Franklin “tripartite burden” in this case, contending that claims of selective enforcement must be raised in a motion to dismiss. We disagree. Lora does not contend that he was charged with trafficking cocaine because of his race, and consequently has not moved to dismiss that charge on the ground of selective enforcement. Rather, his contention is that the traffic stop that led to the discovery of the cocaine was unconstitutional because it was racially motivated. Thus, he argues, the evidence should be suppressed as the product of an unconstitutional stop. In these circumstances, the contention is properly brought in Lora’s motion to suppress.

The suppression of evidence under the exclusionary rule is a “judicially created remedy,” whose “prime purpose is to deter future unlawful police conduct.” United States v. Calandra, 414 U.S. 338, 347, 348 (1974). See Stone v. Powell, 428 U.S. 465, 486 (1976) (“primary justification for the exclusionary role” is deterrence of unconstitutional police conduct). “The rule is calculated to prevent, not to repair. Its purpose is to deter — to compel respect for the constitutional guaranty in the only effectively available way — by removing the incentive to disregard it.” United States v. Calandra, supra at 347, quoting Elkins v. United States, 364 U.S. 206, 217 (1960).25 The exclusionary role is typically used in cases involving violations of the Fourth, *439Fifth, and Sixth Amendments to the United States Constitution, and their counterparts in the Massachusetts Declaration of Rights.26 See generally United States v. Smith, 196 F.3d 1034, 1040 (9th Cir. 1999), cert. denied, 529 U.S. 1028 (2000) (“exclusionary rule is an exceptional remedy typically reserved for violations of constitutional rights”).

We conclude that the application of the exclusionary rule to evidence obtained in violation of the constitutional right to the equal protection of the laws is entirely consistent with the policy underlying the exclusionary rule, is properly gouged to deter intentional unconstitutional behavior, and furthers the protections guaranteed by the Massachusetts Declaration of Rights. See State v. Segars, 172 N.J. 481, 493 (2002) (rationales that support suppression of evidence, “namely, deterrence of impermissible investigatory behavior and maintenance of the integrity of the judicial system, apply equally, if not more so, to cases of racial targeting”). See also United States v. Navarro-Camacho, 186 F.3d 701, 711 (6th Cir. 1999) (Moore, J., concurring) (remedy of suppression may be available to defendant who demonstrates that investigators engaged in selective enforcement).27 See generally Holland, Safeguarding Equal *440Protection Rights: The Search for an Exclusionary Rule Under the Equal Protection Clause, 37 Am. Crim. L. Rev. 1107, 1123 (2000). Consequently, if a defendant can establish that a traffic stop is the product of selective enforcement predicated on race, evidence seized in the course of the stop should be suppressed unless the connection between the unconstitutional stop by the police and the discovery of the challenged evidence has “become so attenuated as to dissipate the taint.” Wong Sun v. United States, 371 U.S. 471, 487-488 (1963), quoting Nardone v. United States, 308 U.S. 338, 341 (1939). See Commonwealth v. Benoit, 382 Mass. 210, 216 (1981).

c. Statistical evidence. “Statistics may be used to make out a case of targeting minorities for prosecution of traffic offenses . . . .” State v. Soto, 324 N.J. Super. 66, 83 (1996) (unrebutted statistical evidence of disproportionate traffic stops of African-American motorists established de facto policy of targeting them for investigation and arrest). See Chavez v. Illinois State Police, 251 F.3d 612, 637-640 (7th Cir. 2001) (differential treatment among races in traffic stops, and discriminatory effect, may be shown through statistics). “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.” Id. at 638. “Further, ‘statistics are not irrefutable; they come in infinite variety and, like any other kind of evidence, they may be rebutted. In short, their usefulness depends on all of the surrounding facts and circumstances.’ ” Id., quoting International Bhd. of Teamsters v. United States, 431 U.S. 324, 340 (1977). If the evidence is introduced through expert testimony, the “expert analysis must be both relevant and reliable.” Chavez v. Illinois State Police, supra at 641.

In State v. Soto, supra at 69-73, a New Jersey case examining racial profiling in traffic stops, the judge was presented with rigorously prepared surveys regarding both the racial makeup of motorists traveling along a stretch of the New Jersey Turnpike encompassing three exits (traffic survey) and the racial makeup of motorists observed violating speeding and other moving violation laws along that same stretch of highway (violation *441survey). The data from these surveys demonstrated that the racial makeup of the motorists traveling the road was 13.5 per cent black, and that the observed violators were approximately fifteen per cent black. Id. at 70-71. This data were used as the benchmark for comparison with the number of “race identified” traffic stops made by the State police along the same stretch of highway. Id. at 70-71.28 The stop data showed that 46.2 per cent of the stops were of black motorists. Id. This disparity was calculated to be more than sixteen standard deviations over the expected norm. Id. at 70. Further expert testimony presented by the defendants established that the surveys were well designed and performed, and statistically reliable for the analysis, and that a similarly situated black motorist was 4.85 times as likely as a white motorist to be stopped on the roadway. Id. at 71-72.

Ultimately, the judge concluded that the statistical data were sufficient to support the finding of a “de facto policy” on the part of certain State troopers of targeting black motorists, which was not adequately rebutted by the State. Id. at 84. In reaching this conclusion, the judge found that the stark patterns revealed in the data established a prima facie case of selective enforcement, that is, that similarly situated motorists were intentionally treated differently because of their race. Accordingly, the judge granted the defendants’ motions to suppress the evidence garnered as a result of their traffic stops. Id. at 84-85.29

In contrast, in Chavez v. Illinois State Police, supra, the United States Court of Appeals for the Seventh Circuit concluded *442that the statistical data presented by the defendants against the Illinois State police failed to establish the elements of a “prima facie” case of selective enforcement of traffic laws based on race. Id. at 641. The data in that case, which were drawn largely from field reports of traffic stops, were problematic for a number of reasons, but most importantly, the court found that the benchmark data used for comparison with the traffic stop data were census data for the entire State, which “can tell us very little about the numbers of Hispanics and African-Americans driving on [the] Illinois interstate highways [at issue], which is crucial to determining the population of motorists encountered by the . . . officers.” Id. at 644.

We are of the view that statistical evidence may be used to meet a defendant’s initial burden of producing sufficient evidence to raise a reasonable inference of impermissible discrimination. Cf. Sullivan v. Liberty Mut. Ins. Co., 444 Mass. 34, 55-56 (2005) (statistical evidence may be used to raise inference of intentional discrimination in employment discrimination claim); Smith College v. Massachusetts Comm’n Against Discrimination, 376 Mass. 221, 228 & n.9 (1978) (same). Cf. Lipchitz v. Raytheon Co., 434 Mass. 493, 508-509 (2001) (statistical evidence may support inference that employment decision was tainted by unlawful bias). At a minimum, that evidence must establish that the racial composition of motorists stopped for motor vehicle violations varied significantly from the racial composition of the population of motorists making use of the relevant roadways, and who therefore could have encountered the officer or officers whose actions have been called into question.30

*443d. Lora’s evidence.31 Having concluded that properly gathered, analyzed, and relevant statistical data may be used to meet a defendant’s burden, we turn to Lora’s evidence. For the reasons set out below, Lora has failed to present credible evidence establishing a reasonable inference of impermissible discrimination sufficient to rebut the presumption that the stop of his motor vehicle was “undertaken in good faith, without intent to discriminate.” Commonwealth v. Franklin, 316 Mass. 885, 894 (1978).

An assessment of the evidence admitted at the rehearing on Lora’s motion to suppress leads us to the inescapable conclusion that the use of census benchmarking to compare the demographics of a small community with citation ratios on a major interstate highway, which happens to pass through it, is unreliable and not accepted in the scientific community. Such benchmarking data do not provide an adequate basis for assessing the racial composition of the drivers encountered by Shugrue on Route 290 and is inadequate to establish that similarly situated drivers of different races were treated differently. Indeed, Lora’s own evidence disproves his premise of comparability: of the fifty-two motorists that Shugrue ticketed on Route 290 in *444Auburn, ninety per cent were not residents of Auburn. Lora therefore failed to present sufficient credible evidence of discriminatory effect. The judge’s determination to the contrary was clearly erroneous, and the motion to suppress should not have been granted.

Even were we to consider census benchmarking data as providing some evidence of selective enforcement, the use of the population demographics of the inhabitants of the town of Auburn was too limited. The population of Auburn is 15,901; the population of Worcester is 172,648. Lora introduced evidence of Shugrue’s citation history in both Worcester and Auburn, but compiled the data in isolation, comparing only the stops effectuated on Route 290 in Auburn with the demographics of Auburn, and the stops effectuated on Route 290 in Worcester with the demographics of Worcester. Logically, as Bishop testified, the combined census demographics of Worcester and Auburn are more likely to reflect the demographics of motorists on Route 290 near the border of Worcester and Auburn than the demographics of Auburn only. When the citations issued by Shugrue in Worcester and Auburn are compared to their combined demographics, a more balanced picture emerges. Shugrue cited 140 motorists in Worcester and Auburn during the relevant eight-month period, 102 of whom were white (72.9 per cent), twenty-three Hispanic (16.4 per cent), and twelve African-American (8.6 per cent). The combined demographics of the area reflect a population that is 77.2 per cent white, 13.7 per cent Hispanic, and 6.2 per cent African-American. Thus, when viewed through an appropriately broad lens, even the questionable census benchmarking methodology used by Lora fails to establish any significant disparity of treatment based on race.

3. Conclusion. Justices of this court have expressed considerable concern about the practice of racial profiling in prior decisions. See Commonwealth v. Feyenord, 445 Mass. 72, 88 (2005) (Greaney, J., concurring), cert. denied, 546 U.S. 1187 (2006) (“A motorist must never be stopped based on his or her race or ethnicity without legally sufficient cause. Getting a traffic ticket is never a happy experience. Getting a traffic ticket if you are a black or Hispanic person who has committed a minor traffic violation and then been questioned in public view by an *445armed police officer determined to find a basis ... to bring in a police dog, is humiliating, painful, and unlawful” [emphasis in original]); Commonwealth v. Gonsalves, 429 Mass. 658, 670 (1999) (Ireland, J., concurring) (“The widespread public concerns about police profiling, commonly referred to as ‘DWB — driving while black,’ has been the subject of much discussion and debate both across the country and within the Commonwealth”). See also United States v. Montero-Camargo, 208 F.3d 1122, 1135 (9th Cir.), cert. denied sub nom. Sanchez-Guillen v. United States, 531 U.S. 889 (2000) (“Stops based on race or ethnic appearance send the underlying message to all our citizens that those who are not white are judged by the color of their skin alone”). These concerns would not be alleviated by a standard that nominally allows a defendant to make claim of selective enforcement of traffic laws, but forecloses such a claim in practice.

On the other hand, the standard must be sufficiently rigorous that its imposition does not unnecessarily intrude on the exercise of powers constitutionally delegated to other branches of government. Balance is therefore important. While racial profiling evidence is relevant to assessing the constitutionality of a traffic stop, and evidence seized during a traffic stop that violates the equal protection guarantees of the Massachusetts Declaration of Rights may be suppressed, the initial burden rests on the defendant to produce evidence that similarly situated persons were treated differently because of their race. The practical weight of this burden is admittedly daunting in some cases, but not impossible. It was done, and done well, in New Jersey.32 *446Data now being collected in Massachusetts,33 and the work of academic and other institutions to develop more sophisticated analytic tools with which to identify and measure the use of race in the context of traffic stops34 may be of assistance in meeting this burden.

Of necessity, the important responsibility of eliminating racial considerations in the day-to-day enforcement of our laws lies principally with the executive branch of government, and no evidence was presented in this case to suggest that this is a responsibility that is being ignored.35 While the judicial branch shares the responsibility of ensuring that the protections of the *447Constitution are afforded to all residents, it can only exercise that responsibility when proper and sufficient evidence has been presented to it. The judge’s order suppressing the evidence in this case is reversed.

So ordered.

Racial profiling, as defined by St. 2000, c. 228, § 1, is “the practice of detaining [or stopping] a suspect based on a broad set of criteria which casts suspicion on an entire class of people without any individualized suspicion of the particular person being stopped.” Racial profiling “is generally understood to mean the improper use of race as a basis for taking law enforcement action.” Chavez v. Illinois State Police, 251 F.3d 612, 620 (7th Cir. 2001). In the context of traffic enforcement, officers who engage in racial profiling “utilize impermissible racial classifications in determining [which motorists] to stop, detain, and search.” Id. at 635. See Alschuler, Racial Profiling and the Constitution, U. Chi. Legal F. 163, 168 n.24 (2002) (surveying definitions of “racial profiling” in State law and academic literature).

We reach this conclusion under arts. 1 and 10 of the Massachusetts Declaration of Rights and do not consider whether the same result would be reached under the Fourteenth Amendment to the United States Constitution.

We acknowledge the amicus brief of the Committee for Public Counsel Services, the American Civil Liberties Union of Massachusetts, and the Massachusetts Association of Criminal Defense Attorneys.

The facts with regard to the traffic stop, which are not disputed by the parties, are drawn from the judge’s findings made after an evidentiary hearing on the defendant’s motion to suppress, supplemented by undisputed testimony.

Trooper Shugrue was assigned to patrol Interstate Route 290, between Interstate Route 90 and Interstate Route 495.

General Laws c. 89, § 4B, states in relevant part: “Upon all ways the driver of a vehicle shall drive in the lane nearest the right side of the way when such lane is available for travel, except when overtaking another vehicle or when preparing for a left turn. . . .” Andres Lora concedes that the vehicle in which he was traveling was in violation of this statutory provision.

During the search, Trooper William Pinkes acted as a “cover officer,” or a trooper that provides backup support by monitoring a suspect while the investigating trooper effectuates the search.

In the trunk, Shugrue found a package wrapped in cellophane and tape containing approximately one kilogram of cocaine.

Lora also moved to suppress statements made to Pinkes at the police barracks after his arrest on the ground that he did not knowingly or voluntarily waive his Miranda rights. The statements were suppressed, and the Commonwealth does not appeal from that ruling.

In the chart accompanying the affidavit, the number was 253.

The records reviewed were Shugrue’s citations, which demonstrate only which motorists were cited, and do not account for motorists who may have been stopped but not cited.

Defense counsel also reviewed 115 citations issued by Trooper Pinkes between June 16, 2001, and August 3, 2001, determining that fifty-two were issued in Worcester or Auburn, and analyzing those citations to determine the race of the cited motorist. The motion judge apparently found these resulting statistics to be persuasive, as he cited them in his decision. However, there is no question that Shugrue initiated the traffic stop on his own, and had decided to search the vehicle before Pinkes arrived.

The records denote only the race of the operator of the vehicle, not the race of any other occupants.

One citation was given to an operator identified as Asian.

The average motorist refers to a random motorist whose race cannot be ascertained by a patrolling officer. He or she is, therefore, race neutral.

These probabilities are determined by comparing the percentage of Shugrue’s citations issued to drivers of a particular minority group with the percentage of the town’s residents of that same race. For example, Hispanics accounted for 31.37 per cent of Shugrue’s citations on Route 290 in Auburn, but only one per cent of the town’s population. Thus, Hispanics are pulled over at a rate 31.37 times greater than their one per cent representation in Auburn’s population would suggest

In his affidavit, defense counsel cast these numbers as percentage increases, labeled “differentials” rather than multiples. The discrepancy between the proportion of Hispanics living in Auburn and the proportion of the Hispanics stopped by Shugrue was not, therefore, described as 31.37 times greater than expected, but rather was described as resulting in a 3,137 percent “differential.” The motion judge included these “differentials” in his decision.

In granting Lora’s motion to suppress, the judge misunderstood the *431significance of portions of the Worcester data. Most prominently, the judge concluded that “[t]he percentage of Hispanics that Shugrue stopped in Worcester [was] 53% higher than the general population of the city [suggests should have been stopped].” In actuality, Hispanic motorists were fifty-three per cent as likely to be stopped as an average motorist; that is substantially less likely.

The judge also admitted testimony from Aaron Jones, a motorist who was stopped by Shugrue on February 2, 2002, at a similar area on Route 290, and cited, inter alia, for operating a motor vehicle with a suspended license, operating an unregistered motor vehicle, and having an inoperative registration plate light. Defense counsel suggested, but did not establish, that Jones was stopped (and subsequently searched) not because of his failure to follow the relevant traffic laws, but because he was a nonwhite motorist. In response, Shugrue testified that the stop and subsequent citation of Jones was consistent with police practice. The motion judge did not reference Jones’s testimony in his decision granting Lora’s motion to suppress.

Defense counsel also contended that Shugrue conducted noninventory searches of vehicles more frequently when the operator cited was Hispanic or African-American than when the operator was white. To that end, he tabulated the number of Shugrue’s citations that later led to motor vehicle searches, and argued that the vehicles of nonwhite drivers were searched disproportionately. Lora’s motion to suppress, however, was centered on his claim that the traffic stop leading to the search was improperly influenced by racial considerations, and the motion judge focused his analysis on that issue. The matter of the disparity in conducting searches has not been argued on appeal and is not before us.

The Commonwealth also filed an application for an interlocutory appeal and a notice of appeal. A single justice of this court deferred action on the Commonwealth’s application while the motion for reconsideration was pending.

Lora also introduced testimony from Dominque Haughton, a professor of mathematical science at Bentley College. Her testimony was limited to whether Shugrue searched nonwhite motorists at a higher rate than white motorists, a subject not raised in this appeal. See note 18, supra.

Bishop elaborated on the reasons for this inaccuracy, testifying that “[t]he stops in this particular case took place near a turnpike exit. Drivers on the turnpike include many nonresidents. There are commuters from other communities who are going to and from work, there are tourists and people on vacation who are passing through .... There is reason to believe that the resident population is not an accurate reflection of the people who are on the roadway.”

Bishop also testified that in the Northeastern University study on racial profiling, “driving observation data” were used as the benchmark. The driving observation data were obtained by “posting] observers on the roadway at various times of day, at different roadways . . . and have them observe, as best they could, the race of the motorists at different times of day on each of those roadway segments.” Although this data did not include observations of traffic law violators, Bishop testified the difference was only a “matter of degree.”

The seminal selective enforcement case is Yick Wo v. Hopkins, 118 U.S. 356 (1886). The petitioner, Yick Wo, filed a writ of habeas corpus, alleging that he had been improperly deprived of his personal liberty by the city of San Francisco because, he argued, the city selectively enforced a law prohibiting the operation of laundromats in wooden buildings against Chinese immigrants. All eighty non-Chinese individuals who applied for an exemption from law were granted that exemption, while all 200 Chinese immigrants who applied for the same exemption were denied it and were thus subjected to a fíne or *437imprisonment. Id. at 374. In reversing the petitioners’ convictions, the Court stated that “[tjhough the law itself be fair on its face and impartial in appearance, yet, if it is applied and administered by public authority with an evil eye and an unequal hand, so as practically to make unjust and illegal discrimina-tions between persons in similar circumstances, material to their rights, the denial of equal justice is still within the prohibition of the Constitution.” Id. at 373-374.

There are, of course, other considerations justifying exclusion, among them the “imperative of judicial integrity.” Mapp v. Ohio, 367 U.S. 643, 659 (1961), quoting Elkins v. United States, 364 U.S. 206, 222 (1960). But see Stone v. Powell, 428 U.S. 465, 485 (1976) (“While courts, of course, must ever be concerned with preserving the integrity of the judicial process, this concern has limited force as a justification for the exclusion of highly probative evidence”). The exclusion of evidence has been applauded as just “because it puts both the State and the accused in the positions they would have been in *439had the Constitution not been violated — neither better nor worse.” Norton, The Exclusionary Rule Reconsidered: Restoring the Status Quo Ante, 33 Wake Forest L. Rev. 261, 262 (1998).

The exclusionary rule has been applied to evidence seized in violation of the Fourth Amendment to the United States Constitution and art. 14 of the Massachusetts Declaration of Rights. See, e.g., Mapp v. Ohio, supra; Commonwealth v. Guaba, 417 Mass. 746, 755-756 (1994). It has similarly been applied to confessions exacted by police in violation of the right against compelled self-incrimination, as protected by the Fifth Amendment to the United States Constitution and art. 12 of the Massachusetts Declaration of Rights. See, e.g., Dickerson v. United States, 530 U.S. 428, 432 (2000); Commonwealth v. Smith, 412 Mass. 823, 836-837 (1992). Evidence obtained from government interrogation of a defendant after he has been charged and has asserted his right to counsel has also been suppressed to deter violations of the Sixth Amendment to the United States Constitution, and the right to counsel as protected by art. 12. See, e.g., Maine v. Moulton, 474 U.S. 159, 179-180 (1985); Commonwealth v. Murphy, 448 Mass. 452, 465-468 (2007).

But see United States v. Nichols, 512 F.3d 789, 794 (6th Cir. 2008) (“While we, of course, agree with the general proposition that selective enforcement of the law based on a suspect’s race may violate the Fourteenth Amendment, we do not agree that the proper remedy ... is necessarily suppression of evidence . . .”).

At the time, not all the reports of stops by the New Jersey State police contained information regarding the race of the motorist. State v. Soto, 324 N.J. Super. 66, 72 n.5 (1996).

The Soto decision had far-ranging effects within New Jersey. It led to a review of the law enforcement practices of the New Jersey State police (see, e.g., Interim Report of the State Police Review Team Regarding Racial Profiling Allegations, April, 1999), which led the New Jersey Attorney General to conclude “that defendants perceived to be African-American, Black or Hispanic are entitled to discovery [regarding racial profiling] for motor vehicle stops that originated as a result of observations made by [New Jersey] State Troopers.” State v. Lee, 190 N.J. 270, 274-275, 280 (2007). In 2000, the Supreme Court of New Jersey issued an administrative order, at the request of the New Jersey Attorney General, assigning one judge to hear all motions for discovery relating to racial profiling by the New Jersey State police to ensure “centralized *442judicial management” of the rapidly emerging issue. Id. at 279.

The decision also became the center of attention in the ongoing legal debate about how to confront the problem of racial profiling. See, e.g., Garrett, Remedying Racial Profiling, 33 Colum. Hum. Rts. L. Rev. 41, 42 & n.1 (2001) (describing Soto case as “[t]he most dramatic example of a lawsuit drawing attention to the racial profiling problem”). Other litigants (criminal defendants and civil plaintiffs) sought to mirror the methodology used by the Soto defendants in their efforts to demonstrate selective enforcement. See Harris, The Stories, the Statistics, and the Law: Why “Driving While Black” Matters, 84 Minn. L. Rev. 265, 277-288 (1999).

While we do not presume that all races violate all laws equally, see United States v. Armstrong, 517 U.S. 456, 479-480 (1996), we are unaware of any reliable study establishing that motor vehicle violations are more frequently committed by any particular race of driver. See State v. Soto, supra at 74.

The Commonwealth contends that Shugrue’s citations were never actually admitted in evidence at the original suppression hearing, despite the motion judge’s conclusion that they were admitted de bene. Regardless of whether the evidence was admitted at the original motion to suppress hearing, all four experts testified to the content of the evidence at the subsequent rehearing.

At that rehearing, the Commonwealth again objected to the admission of the citation data on two grounds — namely, relevance and that St. 2000, c. 228, § 9, prohibits the introduction of citations issued by Shugrue. The Commonwealth reasserts both grounds on appeal. The objections are without merit. The evidence was relevant to the issue of selective enforcement. Additionally, St. 2000, c. 228, § 9, is irrelevant to the issue of admissibility. Section 9 states that “data acquired [by the registry of motor vehicles] shall be used only for statistical purposes and may not contain information that may reveal the identity of any . . . law enforcement officer.” See Boston Police Patrolmen’s Ass’n v. Police Dep’t of Boston, 446 Mass. 46, 51 (2006) (limiting language in § 9 applies only to registry of motor vehicle’s collection of data and subsequent statistical analysis). That provision cannot be read to preclude the introduction of Shugrue’s citations, particularly where the defendant’s evidence did not include any of the registry of motor vehicle’s data or subsequent statistical analysis. Instead, Shugrue made copies of his own past citations pursuant to a pretrial discovery order. That order was not challenged at the time, and the issue is not before us.

In addition to the survey approach used in New Jersey, Bishop testified to the use of other forms of benchmarks, such as “internal benchmarking,” which compares the traffic stop rates among officers within the same geographic area, time, and assignment. If one officer has a demonstrable history of pulling over minority drivers at a higher rate, that is relevant to a showing of selective enforcement because the officer involved is compared to other officers patrolling the same portion of highway, with logically consistent demographics and traffic violation rates. See Smith, Depoliticizing Racial Profiling: Suggestions for the Limited Use and Management of Race in Police Decision-Making, 15 Geo. Mason U. Civ. Rts. LJ. 219, 247-256 (2005). There is, of course, one crucial limitation in the ability of internal benchmarking to detect racial profiling: “If racial profiling discrimination is a department-wide problem, then this method will only pick up the exceptionally horrific offender because one will be comparing a bad apple to other bad apples.” Note, *446The Statistical Evidence of Racial Profiling in Traffic Stops and Searches: Rethinking the Use of Statistics to Prove Discriminatory Intent, 49 B.C. L. Rev. 263, 278 (2008).

See, e.g., “An Act providing for the collection of data relative to traffic stops,” St. 2000, c. 228 (Act), which empowers the registry of motor vehicles and the Secretary of the Executive Office of Public Safety (Secretary) to collect data to determine whether State and local police engage in the practice of racial profiling. Section 8 of the Act requires the registry to collect data from any issued Massachusetts uniform citation regarding the race and gender of the motorist, the traffic infraction, whether a search was initiated, and whether the stop resulted in a warning, citation, or arrest. The registry is then required to report the statistical information it collects every month to the Secretary.

Pursuant to § 10 of the Act, the Secretary shall transmit the collected data to a “university in the commonwealth with experience in the analysis of such data, for annual preparation of an analysis and report of its findings.” If the analysis suggests that a State police barracks or a municipal police department “appears to have engaged in racial or gender profiling,” that department or barracks is thereafter required to “collect information on all traffic stops, including those not resulting in a warning, citation, or arrest.” Id. Individual police departments may mandate that police officers disclose additional information, including their identification, in order to allow those departments to “view data at the officer level” and “manage officers” accordingly. See Boston Police Patrolmen’s Ass’n, Inc. v. Police Dep’t of Boston, 446 Mass. 46, 49, 52-53 (2006) (“In order to fulfil the Act’s objective of eliminating profiling by police officers, the only statutory interpretation that renders the legislation in harmony with common sense and sound reason is one that allows the collection of officer identification information [(if) statistical analysis has already indicated that at least some unidentified officers in a particular police department may be engaging in improper profiling”]).

See, e.g., Farrell, Massachusetts Racial and Gender Profiling Study: Final Report, Inst, on Race and Justice of Northeastern University (2004). See generally Weatherspoon, Ending Racial Profiling of African-Americans in the Selective Enforcement of Laws: In Search of Viable Remedies, 65 U. Pitt. L. Rev. 721, 730 (2004).

Contrast State v. Lee, 190 N.J. 258 (2007) (failure of police to change its *447policies in aftermath of findings in State v. Soto, 324 N.J. Super. 66 [1996], and New Jersey Attorney General’s report regarding racial profiling by State police justified remanding for postconviction discovery).