Indictments found and returned in the Superior Court Department on February 9, 2018.

      A pretrial motion to suppress evidence was heard by Joseph F. Leighton, Jr., J.

      An application for leave to file an interlocutory appeal was allowed by Lenk, J., in the Supreme Judicial Court for the county of Suffolk, and the appeal was reported by her.

      John P. Warren for the defendant.

      Cailin M. Campbell, Assistant District Attorney, for the Commonwealth.

      Rebecca Kiley, Committee for Public Counsel Services, Matthew R. Segal, Jessica Lewis, & Jessie J. Rossman, for Committee for Public Counsel Services & another, amici curiae, submitted a brief.

      Oren N. Nimni, Katharine Naples-Mitchell, Chauncey B. Wood, & Radha Natarajan, for Massachusetts Association of Criminal Defense Lawyers & others, amici curiae, submitted a brief.

      GAZIANO, J.  At about eleven o'clock on a November morning, two members of the Boston police department's youth violence strike force, who had been driving an unmarked vehicle, noticed a maroon Mercedes pass in front of them on a residential street.  The driver was a Black man.  The officers decided to query the vehicle's license plate in their onboard computer.  The results returned indicated that the vehicle was registered to a Black woman and that it lacked an inspection sticker.  The officers stopped the vehicle.  When they learned that the driver, the defendant, had outstanding warrants and his driver's license was suspended, they searched the vehicle and found a gun in a bag on the rear passenger seat.

      The defendant subsequently was charged with several firearms offenses.[2]  He moved to suppress the evidence seized from the vehicle, on the ground that the motor vehicle stop was the product of selective enforcement based on race, and the inventory search of the vehicle was impermissible.  A Superior Court judge determined that the defendant had not met his initial burden to raise a reasonable inference that the stop had been made been motivated by race, and that the decision to impound the vehicle was reasonable in the circumstances; he therefore denied the motion.  The defendant sought leave in the county court to pursue an interlocutory appeal; the single justice allowed the application and ordered the appeal to be conducted in this court.

      We conclude that the Superior Court judge abused his discretion in denying the motion to suppress, because the defendant produced sufficient evidence to raise a reasonable inference that the stop was racially motivated.  Nonetheless, we are persuaded that, in our efforts in Commonwealth v. Lora, 451 Mass. 425, 436-438 (2008), to ease the burden on defendants, we set the bar too high for defendants attempting to establish a reasonable inference of a discriminatory stop.  In practice, providing statistical evidence sufficient to raise a reasonable inference that a motor vehicle stop was racially motivated, given the limitations of available police data, has proved infeasible for defendants.  The judge's ruling well illustrates the concerns repeatedly raised about the difficulty of meeting the requirements set forth in Lora, supra at 447-449.  See Commonwealth v. Buckley, 478 Mass. 861, 879-880 (2018) (Budd, J., concurring), and cases cited.

      Thus, in order to ensure that drivers who are subjected to racially motivated traffic stops have a viable means by which to vindicate their rights to the equal protection of the laws, as provided by the Massachusetts Declaration of Rights, we today establish a revised test.  A defendant seeking to suppress evidence based on a claim that a traffic stop violated principles of equal protection bears the burden of establishing, by motion, a reasonable inference that the officer's decision to initiate the stop was motivated by race or another protected class.  To raise this inference, the defendant must point to specific facts from the totality of the circumstances surrounding the stop; the inference need not be based in statistical analysis.  If this inference is established, the defendant is entitled to a hearing at which the Commonwealth would have the burden of rebutting the inference.  Absent a successful rebuttal, any evidence derived from the stop would be suppressed.[3]

      1.  Background.  We present the facts as found by the motion judge, supplemented by uncontroverted facts from the record that the judge "explicitly or implicitly credited," reserving certain details for discussion.  See Commonwealth v. Jones-Pannell, 472 Mass. 429, 431 (2015) ("Although an appellate court may supplement a motion judge's subsidiary findings with evidence from the record that is uncontroverted and undisputed and where the judge explicitly or implicitly credited the witness's testimony, . . . it may do so only so long as the supplemented facts do not detract from the judge's ultimate findings" [quotations and citations omitted]).

      At approximately 11 A.M. on November 28, 2017, the defendant, a young Black man, was driving a Mercedes sport utility vehicle (SUV) on a well-traveled and largely residential road in the Clam Point section of Boston.  Two plainclothes officers from the Boston police department's youth violence strike force (gang unit)[4] were sitting in an unmarked vehicle on a side street, where they were waiting to make a right turn onto the road on which the defendant was driving.

      The defendant drove past the side street, and the officers turned onto the main road directly behind his vehicle.  They did not observe a traffic violation, but one of the officers decided to enter the vehicle's license plate number into the Criminal Justice Information Services (CJIS) database.  The query revealed that the defendant's vehicle was registered to a Black woman, later identified by the defendant as his girlfriend.  The query also showed that the vehicle did not have a current inspection sticker.

      The officers decided to stop the vehicle on the basis of the missing inspection sticker.  When the officers activated their lights and siren, the defendant pulled into a lawful parking spot on the side of the street.  The officers requested his driver's license; the defendant explained that he did not have a driver's license, only a permit, which he provided the officers.  Although they had never encountered each other, one of the officers recognized the defendant's name and photograph from the gang unit's database.  After conducting a query of the defendant's information in the CJIS database, the officers learned that his driver's license was suspended, and that he had two default warrants for operating without a license and failure to identify himself.  The officers ordered him out of the vehicle and handcuffed him.

      The officers, who testified that they were aware of thefts, vandalism, and shootings in the vicinity, decided to tow and impound what they deemed to be a "high-end" vehicle, which did not belong to the defendant.[5]  Before towing the vehicle, one officer began to search it.  During the search, he observed what he believed was the handle of a gun inside an open bag on the rear passenger seat.  Once he confirmed that the object indeed was a firearm, he read the defendant the Miranda warnings and then inquired whether the defendant had a license to carry a firearm.  When the defendant responded that he did not, the officers called dispatch to transport the defendant to the police station.

      2.  Prior proceedings.  The defendant filed a motion to suppress the evidence seized from the vehicle on the ground that the stop was impermissible because it was the result of selective enforcement of the traffic laws based on race, and the inventory search was an unlawful search for investigatory purposes, as impoundment of the vehicle was not necessary.  The defendant obtained an expert in statistics (a mathematics professor who has published numerous articles and reports, and had testified previously in the Superior Court and the District Court).  The expert testified at an evidentiary hearing on the motion, and introduced a report on her findings.  The judge found that the datasets used by the expert were "insufficiently reliable to yield results that could raise a reasonable inference of impermissible discrimination," and denied the motion.  The defendant filed a timely notice of appeal and an application in the county court for leave to pursue an interlocutory appeal.  See Mass. R. Crim. P. 15 (a) (2), as amended, 476 Mass. 1501 (2017).  The single justice allowed the appeal to proceed in this court.

      3.  Discussion.  In Lora, 451 Mass. at 437-438, we adapted general principles of our jurisprudence on selective prosecution in an attempt to address the persistent and pernicious problem of racial profiling in traffic enforcement.  Today, we conclude that that decision placed too great an evidentiary burden on defendants, and that we must lower this burden in order to create a viable path for individuals to present and demonstrate their claims of racial profiling in traffic stops.  While defendants still may raise a reasonable inference of racial profiling by demonstrating consistent patterns of racially disparate traffic enforcement by the officer involved, they also may raise a reasonable inference that a stop was racially motivated based on the totality of the circumstances surrounding the particular traffic stop at issue.

      Furthermore, in our view the problem of discriminatory traffic stops continues to be best addressed under our equal protection jurisprudence and not, as Justice Budd's concurrence suggests, the search and seizure doctrine of art. 14 of the Massachusetts Declaration of Rights.  As to the stop of this defendant, even under the overly heavy evidentiary burden that resulted from our decision in Lora, we conclude that he presented more than adequate data to support his claim, and thus that the judge erred in denying his motion to suppress.

      a.  Equal protection framework.  Our guarantee of equal protection under the law derives both from the Fourteenth Amendment to the United States Constitution and arts. 1 and 10 of the Massachusetts Declaration of Rights.  When the Fourteenth Amendment was ratified after the Civil War, its "primary objective . . . was the freedom of [African-Americans], the security and firm establishment of that freedom, and the protection of the newly-made freeman and citizen from the oppressions of those who formerly had exercised unlimited dominion over him" (quotation and citation omitted).  Flowers v. Mississippi, 139 S. Ct. 2228, 2238 (2019).  Art. 1 has a similar purpose.  Ratified as part of the Massachusetts Declaration of Rights in 1780, the article was the basis of the judicial abolition of slavery in 1781, see Jackson v. Phillips, 14 Allen 539, 563 (1867), and subsequent decisions applying the guarantee of equal protection to African-Americans.[6]  See, e.g., Commonwealth v. Aves, 18 Pick. 193, 210 (1836).

      Under these constitutional guarantees, the racism in which our nation had been steeped was to yield to the promise of equality.  See Williams v. Illinois, 399 U.S. 235, 245 (1970) ("the constitutional imperatives of the Equal Protection Clause must have priority over the comfortable convenience of the status quo").  And, indeed, many explicitly discriminatory laws did fall.  See, e.g., McLaughlin v. Florida, 379 U.S. 184, 184 (1964).

      All too frequently, however, the prohibition against facially discriminatory laws has been inadequate to address the role played by racism and other invidious classifications in the way facially neutral laws actually are enforced.  See Buckley, 478 Mass. at 871 (sharing "considerable, legitimate concerns regarding racial profiling and the impact of such practices on communities of color"); Lora, 451 Mass. at 449 (Ireland, J., concurring); Commonwealth v. Feyenord, 445 Mass. 72, 88 (2005), cert. denied, 546 U.S. 1187 (2006) (Greaney, J., concurring); Commonwealth v. Gonsalves, 429 Mass. 658, 670 (1999) (Ireland, J., concurring).

      Thus, it long has been held that "[t]he equal protection principles of the Fourteenth Amendment . . . and arts. 1 and 10 . . . prohibit discriminatory application of impartial laws."  See Lora, 451 Mass. at 436, quoting Commonwealth v. Franklin Fruit Co., 388 Mass. 228, 229–230 (1983).  See also New York Times Co. v. Commissioner of Revenue, 427 Mass. 399, 406 (1998), citing Yick Wo v. Hopkins, 118 U.S. 356, 373 (1886).  While some selectivity or discretion must be tolerated in criminal law enforcement, that selectivity is permissible only so long as it "is not based on an 'unjustifiable standard such as race, religion or other arbitrary classification.'"  Lora, supra at 437, quoting Commonwealth v. King, 374 Mass. 5, 20 (1977).  See Oyler v. Boles, 368 U.S. 448, 456 (1962).

      Consistent with Federal equal protection law, we have held that a prosecution brought based on an impermissible classification must be dismissed.  See King, 374 Mass. at 22.  "Because we presume that criminal prosecutions are undertaken in good faith, without intent to discriminate, the defendant bears the initial burden of . . . present[ing] evidence which raises at least a reasonable inference of impermissible discrimination."  Commonwealth v. Franklin, 376 Mass. 885, 894 (1978), and cases cited.  To support this inference, a defendant must show 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).  See id.  "If a defendant meets this prima facie showing, the case must be dismissed unless the Commonwealth is able to rebut the inference of selective prosecution."  Commonwealth v. Wilbur W., 479 Mass. 397, 409 (2018), citing Commonwealth v. Bernardo B., 453 Mass. 158, 168 (2008).

      This court has identified the discriminatory enforcement of traffic laws as particularly toxic.  "Years of data bear out what many have long known from experience:  police stop drivers of color disproportionately more often than Caucasian drivers for insignificant violations (or provide no reason at all)."  Buckley, 478 Mass. at 876-877 (Budd, J., concurring).  See Pierson, Simou, Overgoor, Corbett-Davis, Jenson, Shoemaker, Ramachandran, Barghouty, Phillips, Shroff, & Goel, A Large-Scale Analysis of Racial Disparities in Police Stops Across the United States, 4 Nature Human Behavior 736, 736, 737 (2020) (analysis of approximately 95 million stops nationwide found that "[r]elative to their share of the residential population, . . . [B]lack drivers were, on average, stopped more often than white drivers," and that Black drivers comprised a smaller share of drivers stopped at night, when it is harder for officers to detect race, "suggest[ing] [B]lack drivers were stopped during daylight hours in part because of their race").

      The discriminatory enforcement of traffic laws is not a minor annoyance to those who are racially profiled.  To the contrary, these discriminatory practices cause great harm.  See Buckley, 478 Mass. at 877 (Budd, J., concurring), citing Feyenord, 445 Mass. at 88 (Greaney, J., concurring) ("to a Caucasian driver a traffic stop may be annoying or embarrassing, but for a driver of color, such a stop can be humiliating and painful. . . . Further, recent tragic events have shown that the fear people of color have of being stopped by police is justified:  African–Americans have been killed during routine traffic stops").  See also Utah v. Strieff, 136 S. Ct. 2056, 2069 (2016) (Sotomayor, J., dissenting) ("unlawful 'stops' have severe consequences much greater than the inconvenience suggested by the name. . . . When we condone officers' use of these devices without adequate cause, we . . . risk treating members of our communities as second-class citizens").

      While the constitutional principle at stake in this case is exceedingly clear -- police may not target drivers for traffic stops, citations, and further investigation because of their race -- the evidentiary difficulties in identifying racially motivated traffic stops are profound.  The traffic stop often constitutes the first and only interaction between a police officer and the occupants of a stopped vehicle; the interaction thus generally provides a minimal amount of direct evidence of the officer's motivations for the particular stop.  Additionally, the plethora of potential traffic violations is such that most drivers are unable to avoid committing minor traffic violations on a routine basis, thereby affording officers wide discretion in the enforcement of traffic laws.  See State v. Ladson, 138 Wash. 2d 343, 358 n.10 (1999), citing Shakow, Let He Who Never Has Turned Without Signaling Cast the First Stone:  An Analysis of Whren v. United States, 24 Am. J. Crim. L. 627, 633 (1997).

      b.  The Lora selective prosecution analysis.  Due to these challenges, in combination with the urgent need to deter discriminatory policing, in Lora, 451 Mass. at 437, we attempted to adapt the principles of selective prosecution to the unique evidentiary challenges posed by claims of discriminatory traffic stops.  We did so by relying upon the preexisting burden-shifting analysis of our selective prosecution framework, but expanding the ways in which a defendant could meet his or her initial burden.

      In the first stage of that analysis, a defendant must "'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.'"  Lora, 451 Mass. at 437, quoting Franklin, 376 Mass. at 894.  Because of the difficulty of showing that a particular officer's intent in making a specific motor vehicle stop was racially motivated, however, we held that the defendant's burden could be met through the presentation of evidence of that officer's motor vehicle stops in other cases.  See Lora, supra at 442.

      By allowing reasonable inferences based on broader patterns involving other defendants, our holding in Lora avoided the limitations that have been placed on Federal equal protection claims based on variances in statistics since the United States Supreme Court's decision in McCleskey v. Kemp, 481 U.S. 279, 292-293, 297-298, 311-313 (1987).  In that case, the Court determined that the statistical data presented by the defendant demonstrated a correlation between race and the imposition of the death penalty, but was insufficient to show causation in the defendant's specific case.  With respect to the protections against selective enforcement of traffic laws under our State constitution, we stated that the evidence presented by a defendant, "[a]t a minimum, . . . 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."  Lora, 451 Mass. at 442.

      As with any other selective enforcement claim, if a defendant raises a reasonable inference that a stop was motivated by race, the burden shifts to the Commonwealth to rebut the inference.  See Lora, 451 Mass. at 438.  Unlike other types of selective prosecution cases, however, if the Commonwealth fails to rebut that inference in the context of a traffic stop, the remedy is not dismissal.  Because the discriminatory enforcement of traffic laws is more closely tied to the evidence obtained as a result of the stop, rather than the decision to bring criminal charges based on that evidence, we concluded that suppression was the correct remedy for a traffic stop that violated the guarantees of equal protection in arts. 1 and 10.  See id. at 438-439.

      c.  Revising the evidentiary requirements of Lora.  Our decision in Lora was intended to make it easier for defendants to establish racial discrimination by allowing them to raise a reasonable inference of racial profiling based on an officer's conduct in other traffic stops.  From this pattern of unequal treatment, and in the absence of explicit "smoking gun" evidence concerning that particular stop, a judge could infer that the challenged stop of an individual defendant was motivated by race.  Importantly, this mechanism also allows defendants a means by which to detect and challenge implicit bias, by demonstrating that an officer's pattern of behavior toward members of the protected class of which the defendant is a member showed discrimination, regardless of the officer's lack of awareness of any bias.

      When Lora was decided, it was believed that data regarding the traffic stops made by individual police officers throughout the Commonwealth, and the demographics of the individuals stopped, would be readily available to defendants.  See Lora, 451 Mass. at 446 & n.33, n.34.  See also Boston Police Patrolmen's Ass'n, Inc. v. Police Dep't of Boston, 446 Mass. 46, 48-49 (2006), discussing St. 2000, c. 228, § 10.  Unfortunately, that assumption has not been borne out in practice.  A statute enacted in 2000, effective for a limited period of time, mandated that police departments that appeared to have engaged in racial profiling in motor vehicle stops collect detailed data on stops made by each officer, and the races of the driver stopped.  See St. 2000, c. 228, § 10.   That statute, however, expired by its terms, and a replacement was not enacted.  See id. (requiring one year of data collection, followed by additional year of broader data collection for departments found to have engaged in racial or gender profiling).  Instead, our effort to ease the burden on defendants has been unsuccessful due to inadequate or inaccessible data.  See Buckley, 478 Mass. at 880 (Budd, J., concurring); Lora, 451 Mass. at 449 (Ireland, J., concurring).  Consequently, since 2008 when Lora was decided, we are aware of only one case in which a defendant successfully moved to suppress evidence under it.  See Commonwealth vs. Vargas, Middlesex Superior Court No. 1481CR1135, slip op. at 1, 16 (Aug. 16, 2019).

      Although we conclude that the defendant here also met his burden under the existing Lora standard, it is clear that Lora has placed too great an evidentiary burden on defendants.  The right of drivers to be free from racial profiling will remain illusory unless and until it is supported by a workable remedy.  Therefore, the time has come to address Lora's practical shortcomings.

      i.  Defendants' revised burden.  Although Lora, 451 Mass. at 440-442, focused on how statistical evidence could be used to meet a defendant's initial burden of raising a reasonable inference of discrimination, we did not say in that case that statistical evidence would replace the previous means of establishing a violation of equal protection, and would become the only way in which an inference that a stop was motivated by race could be raised.  See id. at 442 ("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" [emphasis added]).

      Indeed, in the broader jurisprudence on selective enforcement, both nationally and in Massachusetts, the evidence necessary to raise a reasonable inference of discrimination need not be statistical.  See, e.g., Wilbur W., 479 Mass. at 409, quoting Bernardo B., 453 Mass. at 168 ("defendant raising a selective prosecution claim may do so 'by introducing statistical evidence or other data demonstrating that similarly situated suspects or defendants are treated differently by the prosecutor on the basis of impermissible categorizations'" [emphasis supplied]); Franklin, 376 Mass. at 894 (testimonial evidence was used to support claim of selective prosecution); Marshall v. Columbia Lea Reg'l Hosp., 345 F.3d 1157, 1168 (10th Cir. 2003) ("[the defendant] seeks to prove the racially selective nature of his stop and arrest not by means of statistical inference but by direct evidence of [the officer's] behavior during the events in question").  Nor must the asserted discrimination be "systematic [or] long-continued."  See Snowden v. Hughes, 321 U.S. 1, 9–10 (1944).

      In light of the persistent difficulties attendant to using statistical data to meet a claim under Lora, however, we now must develop more fully the other ways in which defendants may show that a stop was based on an impermissible classification.  Cf. Batson v. Kentucky, 476 U.S. 79, 92 (1986) (previous decisions that required comprehensive statistics showing prior discriminatory action amounted to "crippling burden of proof" on defendants attempting to vindicate rights to equal protection).  Therefore, while the use of statistical data continues to be one means by which a defendant may raise a reasonable inference that the challenged traffic stop was racially motivated, we today expand and clarify the other ways in which such an inference may be raised, by evidence of the totality of the circumstances surrounding the stop itself.

      Moreover, not only must the categories of permissible evidence be altered; the way in which defendants may establish a reasonable inference of discrimination also requires modification.  Under general selective prosecution analysis, a defendant's initial showing must include evidence "that a broader class of persons than those prosecuted 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."  See Bernardo B., 453 Mass. at 168, quoting Lora, 451 Mass. at 437.  These first two requirements generally would be difficult or impossible to prove with circumstantial evidence.  If a defendant sought to present evidence only concerning the stop itself, the defendant would not be able to show that a broader class of persons violated the law.  Furthermore, if a defendant did not point to consistent patterns of conduct or police admissions, it usually would not be possible to show that the failure to enforce the traffic laws was deliberate or consistent.

      In the context of racially biased motor vehicle stops, purportedly to enforce traffic laws, however, these first two requirements are unnecessary.  As stated, because of the ubiquity of traffic violations, only a tiny percentage of these violations ultimately result in motor vehicle stops, warnings, or citations.  Thus, it virtually always will be the case "that a broader class of persons" violated the law than those against whom the law was enforced.  See Bernardo B., 453 Mass. at 168.  Similarly, in stopping one vehicle but not another, an officer necessarily has made a deliberate choice.  In the context of a motor vehicle stop, we therefore conclude that the first two requirements are not needed.  The totality of the circumstances test, described infra, requires only the evidence necessary to support a reasonable inference that the stop was based on race or membership in another constitutionally protected group.

      Additionally, in the context of traffic stops, we must depart from our prior interpretations of the meaning of a "reasonable inference," to the extent that the phrase was used to represent an onerous standard in other areas of selective enforcement law.  In King, 374 Mass. at 17, for instance, a female defendant argued that her arrest and prosecution for prostitution was based on her gender, in violation of equal protection principles.  Notwithstanding the arresting officer's testimony "that he had never arrested a male prostitute and that it was the policy of the Boston police department vice squad to arrest only female prostitutes," we held that the defendant had fallen "far short of establishing any evidence of a denial of equal protection since other criminal statutes may be employed to punish male conduct equivalent to female prostitution."  Id. at 18.

      Similarly, in Franklin Fruit Co., 388 Mass. at 229-230, the defendant supermarket claimed that the prohibition against the operation of a supermarket on Sundays was applied against it, and not others, because its owners were Greek.  Relying upon the doctrine of selective enforcement, the judge dismissed the charges against the supermarket after he found that the police chief had refused to enforce the prohibition against the defendant's competitors, and had called the owner of the supermarket a "money hungry Greek."  Id. at 230.  Yet, this court reversed that decision because it concluded that there was "nothing in the record to indicate that citations were issued to [the supermarket] simply because [its owner] was of Greek heritage."  Id. at 234.

      As has been demonstrated, these holdings would set a nearly impossible bar for victims of discriminatory traffic stops to clear in order to establish their claims, whether through statistics or other circumstantial evidence.  See Lora, 451 Mass. at 445 (biased policing "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").  Rather than requiring a "reasonable inference," these cases actually demand something much more.  Accordingly, we conclude that our past interpretations of a reasonable inference do not control in the context of traffic stops.  While a defendant must show more than the fact that he or she was a member of a constitutionally protected class and was stopped for a traffic infraction, the burden must not be so heavy that it makes any remedy illusory.  The requirement that a defendant establish a reasonable inference that a traffic stop was motivated by racial bias means simply that the defendant must produce evidence upon which a reasonable person could rely to infer that the officer discriminated on the basis of the defendant's race or membership in another protected class.  Conclusive evidence is not needed.

      ii.  The revised test.  The burden shifting framework under Lora remains the same, even as we elaborate on the ways in which a defendant can present nonstatistical evidence of a race-based pretextual stop.  A defendant first should raise a reasonable inference of racial profiling through a motion to suppress.  The motion should describe all of the circumstances of the traffic stop that support a reasonable inference that the decision to make the stop was motivated (whether explicitly or implicitly) by race.  The defendant need not submit admissible evidence; rather, the motion simply must point to specific facts about the stop that support such an inference.  These facts, including statements by the defendant and others, may be based on the defendant's personal knowledge, the defendant's own investigation, evidence obtained during discovery, and other relevant sources.  If the defendant's motion establishes such an inference, the defendant is entitled to a hearing, at which the Commonwealth would bear the burden of rebutting the inference.  Of course, a traffic stop motivated by race is unconstitutional, even if the officer also was motivated by the legitimate purpose of enforcing the traffic laws.

      When examining the totality of the circumstances, judges should consider factors such as:  (1) patterns in enforcement actions by the particular police officer;[7] (2) the regular duties of the officer involved in the stop;[8] (3) the sequence of events prior to the stop;[9] (4) the manner of the stop;[10] (5) the safety interests in enforcing the motor vehicle violation;[11] and (6) the specific police department's policies and procedures regarding traffic stops.[12]  These factors are not exhaustive; any relevant facts may be raised for the judge's consideration.

      A defendant has a right to reasonable discovery of evidence concerning the totality of the circumstances of the traffic stop; such discovery may include the particular officer's recent traffic stops and motor vehicle-based field interrogations and observations (FIOs).  To the extent that the relevant information exceeds the automatic discovery requirements of Mass. R. Crim. P. 14 (a) (1) (A), as amended, 444 Mass. 1501 (2005), a defendant may seek such discovery by means of a motion filed pursuant to Mass. R. Crim. P. 14 (a) (2), as appearing in 442 Mass. 1518 (2004).  See Commonwealth v. Durham, 446 Mass. 212, 234, cert. denied, 549 U.S. 855 (2006) (Cordy, J., dissenting) ("Rule 14 (a) (2) gives a judge discretion to authorize a defendant to discover from the Commonwealth 'relevant evidence'").  "At the discovery stage, the question is whether the defendant has made a threshold showing of relevance."  Bernardo B., 453 Mass. at 169, discussing Mass. R. Crim. P. 14 (a) (2).  Where relevant and material, discovery also would include information regarding the policies and procedures pertaining to the officer's unit, as well as the officer's typical duties and responsibilities.  Of course, this right to discovery applies equally to all claims of racially motivated stops, regardless of whether a defendant is pointing to the circumstances of the stop to raise a claim of discriminatory enforcement or is presenting the type of broader statistics contemplated by Lora.

      Once a reasonable inference of racial profiling has been established, the Commonwealth would bear the burden of rebutting that inference.  See Lora, 451 Mass. at 438.  To meet its burden, the Commonwealth would have to do more than merely point to the validity of the traffic violation that was the asserted reason for the stop.  Rather, it would have to grapple with all of the reasonable inferences and all of the evidence that a defendant presented, and would have to prove that the stop was not racially motivated.  If the Commonwealth does not rebut the reasonable inference that the stop was motivated at least in part by race, the defendant would have established that the stop violated the equal protection principles of arts. 1 and 10, and therefore was illegal, and any evidence derived from the stop would have to be suppressed.  See id.

      d.  Equal protection is the appropriate constitutional provision.  Justice Budd would use art. 14 to address the affliction of racial profiling in traffic enforcement, thereby grafting the equal protection inquiry onto our jurisprudence on searches and seizures.  This court and the United States Supreme Court, however, consistently have held that "the constitutional basis for objecting to intentionally discriminatory application of laws is the Equal Protection Clause, not the Fourth Amendment" to the United States Constitution or art. 14 of the Massachusetts Declaration of Rights.  See Lora, 451 Mass. at 436, quoting Whren v. United States, 517 U.S. 806, 813 (1996).

      While the justices differ as to which constitutional principles to use as the basis of a system intended to eliminate racial profiling in traffic stops, both approaches share this common goal.  It is critical to bear in mind that a disagreement about the best legal analysis to use to redress the fundamental problem of racial bias in traffic stops is not a disagreement about the importance of systemic change to attempt to reach this goal.

      Justice Budd contends that the use of a "would have" test, based solely on a reasonable officer standard, without inquiry into the true motivations of the officer who conducted the stop, would address the evidentiary challenges inherent in establishing that a traffic stop was based on racial profiling.  In her view, an analysis under equal protection is susceptible to manipulation, whereas an analysis under art. 14, which asks whether a reasonable officer would have made the stop, somehow magically would eliminate all concerns of police manipulation or lack of candor.  Similarly, in this view, analyzing a traffic stop under art. 14 would eliminate judicial reluctance to inquire into an officer's motive in making the stop whereas, for a driver to prevail on an equal protection claim, necessarily requires a finding that an officer indeed was racially biased.  But these complex and nuanced subjective inquiries are not so easily avoided.  Given the significant difficulties in discerning the characteristics of a "reasonable officer," in conjunction with the justified trepidation of trial judges in second guessing discretionary law enforcement decisions where no discriminatory motivation was involved, any version of the "would have" inquiry, regardless of whether it explicitly includes only an objective component, inevitably would slide into the subjective motivation of the officer.  A determination that a reasonable (nonracist) officer would not have made the stop may be worded more palatably, but the underlying conclusion is the same as a determination that a stop was in violation of equal protection:  a judge allowing a motion to suppress evidence seized after both stops has inquired, however obliquely, into the officer's intent, and has determined that the stop was motivated by racial bias.

      That the examination of subjective motives is essentially unavoidable under the "would have" test, and therefore also subject to the "fraught" judicial inquiry into subjective intent, and potential manipulation by law enforcement, is illustrated by the ways in which the test has been used in practice.  That it has not, in fact, accomplished its goal of removing officer intent and implicit bias from a reviewing judge's decision-making process suggests that in Massachusetts, too, the "would have" test will not be the magic wand that Justice Budd anticipates.  A review of its history in instructive.

      In 1995, the United States Supreme Court rejected the "would have" test and declared that a motor vehicle stop based on probable cause that a traffic violation had occurred complies with the Fourth Amendment, regardless of any ulterior motives for the stop.  See Whren, 517 U.S. at 813, 816-819.  In subsequent years, forty-seven States have followed the Court in rejecting the "would have" test.[13]

      New Mexico and Washington, the only two States to have adopted the "would have" test post-Whren, explicitly have included the subjective motivations of the officer conducting the stop as part of their version of the test.[14]  See State v. Gonzales, 257 P.3d 894, 898 (N.M. 2011) (question is whether, based on totality of circumstances, "the officer who made the stop would have done so even without the unrelated motive"); Ladson, 138 Wash. 2d at 358–359 ("When determining whether a given stop is pretextual, the court should consider the totality of the circumstances, including both the subjective intent of the officer as well as the objective reasonableness of the officer's behavior."

      Under these tests, courts have focused on whether the circumstantial evidence was sufficient to establish that the officer had the requisite subjective motive.[15]  See, e.g., Schuster v. State Dep't of Taxation & Revenue, Motor Vehicle Div., 283 P.3d 288, 298 (N.M. 2012) (approaching motorcycle operator whose motorcycle had fallen to ground was not pretext because motive was community caretaking); State vs. Deleon, N.M. Ct. App., No. 30,813, slip op. at 6-8 (Feb. 14, 2013) (wide turn violation was pretext for investigation of driving under influence of alcohol, where pattern was established by six witnesses who testified that they had been pulled over for minor violations in previous months after leaving same bar, asked about drinking, and released without traffic citation); State v. Jones, 163 Wash. App. 354, 363 (2011) (stop for failure to use seat belt was not pretext, where officer cited defendant for infraction); State v. Montes-Malindas, 144 Wash. App. 254, 262 (2008) (stop of driver who turned on headlights one hundred yards after starting to drive, conducted by officer who was not on traffic patrol, was pretext).[16]  These inquiries are quite similar to the inquiry described, supra, under a selective prosecution analysis.

      A move to art. 14 would confuse this inquiry, because our jurisprudence on search and seizure provides no guidance regarding which officer motivations render unreasonable an otherwise permissible traffic stop.  Our equal protection jurisprudence, by contrast, provides clear guidelines.  A governmental action based on membership in a suspect class, including "sex, race, color, creed, or national origin," is unconstitutional unless the action survives strict scrutiny.  See Finch v. Commonwealth Health Ins. Connector Auth., 459 Mass. 655, 662 (2011), S.C., 461 Mass. 232 (2012), quoting King, 374 Mass. at 21.  Thus, unlike art. 14, our equal protection doctrine provides a clear definition of an unlawful traffic stop:  any stop based on a suspect classification.

      In practice, the additional methods of establishing racial profiling that the court adopts today, and the "would have" test that Justice Budd propounds, often could look similar at a hearing on a motion to suppress.  In some situations, however, the "would have" test might not be effective where an analysis under equal protection would.  By focusing on what a reasonable officer would have done in those particular circumstances, for example, and disregarding any overt discussion of intent, the "would have" test would conclude that a stop for driving twenty miles per hour over the speed limit in a residential neighborhood was reasonable to protect public safety, and would not take into account that the traffic enforcement officer making the stop largely was stopping Black drivers.  Similarly, if an officer targeted intensive traffic enforcement efforts only at neighborhoods where most residents are people of color, motor vehicle stops such as for running red lights or stop signs would be reasonable under the "would have" test to protect pedestrians and other vehicles.  Yet, under an equal protection analysis, the municipality-wide statistics that will be available for all municipalities in the Commonwealth on an ongoing basis, as a result of a 2019 statute, see G. L. c. 90, § 63, inserted by St. 2019, c. 122, § 10, and part 3.f., infra, could be used to support a reasonable inference that the stops unfairly targeted communities of color, and were made due to racial bias.

      In sum, the deficiency in our response to race-based traffic enforcement has not been the basic principles of equal protection, but, rather, the burdens we have placed on defendants to establish their claims.  The way to address that issue, as we do today, is to modify the ways in which claims of racial profiling can be demonstrated, and not to change the constitutional protection under which claims are analyzed.

      e.  The defendant's evidence in this case.  Even as we create new methods that defendants may use to establish a reasonable inference of discrimination in traffic stops, we recognize that statistical evidence, if available, has unique advantages for reaching the thorny question of intent, particularly when implicit bias is at issue.  This is so regardless of whether the officers' actions are analyzed under an equal protection framework or under art. 14.  In terms of systemic change, statistics provide potentially the strongest tool to demonstrate that bias, particularly where it is implicit; indeed, the many journal articles relied upon by the concurrence derive their power from presenting staggering inequity in numbers.

      There generally are two components to the statistical data that defendants have used to establish a reasonable inference that the decision to conduct the traffic stop was motivated by race:  (1) information about how the statute was enforced against other drivers of the defendant's race by the officers or department in question, often involving numbers of stops, citations, and FIOs for drivers of specific races (enforcement data); and (2) statistical data that estimate the demographic distribution of drivers on the roads in the area of the stop (benchmark data).  The two are then compared, under the assumption that, absent impermissible discrimination, the enforcement rates should reflect the demographic composition of all drivers.  See Lora, 451 Mass. at 442.  More specifically, in this case, the percentage of citations and FIOs involving Black drivers should have been similar to the estimated percentage of Black drivers on the road in the area of the stop.

      In support of his claim, the defendant obtained FIO reports issued by the officers who stopped him, for a period spanning from January 1, 2011, to November 28, 2017, the day of the stop; he also obtained citation data from December 14, 2011, to November 11, 2017.  The defendant then engaged a statistical expert to compare the citations and FIO reports involving motor vehicle stops to the population of drivers on the road on which the defendant was stopped, that is, the "benchmark" population.  The expert estimated the demographics of the benchmark population based on data from the 2010 United States Census concerning "census blocks" -- geographical subunits containing from 600 to 3,000 people -- for areas in which these officers either had issued a citation or had reported a motor vehicle-related FIO.  Because the population of motorists on these largely residential roads may include motorists who reside in areas outside the pertinent census blocks, the expert also estimated the demographics of benchmark populations in additional census blocks within 300 feet, 600 feet, and 1,000 feet of an FIO or citation.  Ultimately, the expert concluded that the officers in this case were significantly more likely to conduct an FIO of a driver based on a motor vehicle infraction if the driver was Black than if the driver was not Black.[17]  She also concluded that there was strong statistical evidence that the officers issued citations to Black drivers in Boston at rates consistent with racial profiling.[18]

      The judge, however, found that this evidence did not satisfy the defendant's initial burden to raise a reasonable inference of discrimination under Lora.  Although the judge recognized that it was unclear whether the data that he concluded were necessary even existed, he reasoned that the FIO and citation data presented were unreliable because not every traffic stop results in the production of an FIO report or the issuance of a citation.[19]  As the available data did not reflect the "larger, unknown total number of motor vehicle stops," the judge found that it "f[e]ll well short" of the data analyzed in State v. Soto, 324 N.J. Super. 66 (1996), a New Jersey case we cited approvingly in Lora, 451 Mass. at 440-441, where the defendants had access to a police database of all stops made on the relevant roadway.  See Soto, supra at 69.  The judge also concluded that, as compared to the observational data presented in Soto, the census data here were an unreliable benchmark for the demographics of motorists on the roads patrolled by the officers without some "independent verification" that the data were reflective of the population of relevant motorists.  See Lora, 451 Mass. at 444.

      The judge erred in discounting both aspects of the defendant's data.  With regard to benchmark data, Lora, 451 Mass. at 443-444, does not stand for the categorical rule that census data is never an appropriate proxy for the actual population of motorists on the relevant roadway.  In Lora, supra, we concluded that it was inappropriate to use census data from the town of Auburn as a benchmark for the demographics of drivers passing through Auburn on a major interstate highway; we noted that, of the fifty-two motorists ticketed by the officer in question on that stretch of highway, ninety percent were not residents of Auburn.  Here, where the relevant roadways are urban residential roads, as opposed to an interstate highway, we have much greater confidence in the accuracy of residential demographics from United States Census data as representative of those making use of the residential roads.  Moreover, in this case, the defendant's expert statistician took pains to account in sophisticated ways for the possible presence of nonresident drivers.  This benchmarking data was more than sufficiently reliable to support a claim of selective enforcement and racial discrimination in making the traffic stop, under the defendant's then-existing burden.  The judge abused his discretion in rejecting it as insufficient.

      The same is the case with respect to the defendant's use of data on citations and motor vehicle stop-generated FIOs.  If, as here, the data on the officers' citations and motor vehicle FIOs show that these interactions are racially skewed, it is a reasonable inference that the rate at which those officers' stops of drivers of a particular race is similarly disproportionate to stops of other drivers.  In declining to take this inferential step, the judge erred.  His insistence that this inference had to be all but unescapable, and not merely reasonable, was an abuse of discretion.

      In support of his motion to suppress, the defendant submitted the expert's statistical analysis, which established a reasonable inference of impermissible discrimination.  Having made this implicit determination, the judge properly decided to conduct an evidentiary hearing, where the burden was on the Commonwealth to rebut the reasonable inference established by the defendant.  See Lora, 451 Mass. at 438.  The Commonwealth offered the testimony of the arresting officers, who testified that they did not conduct the stop due to the defendant's race.  Because implicit bias may lead an officer to make race-based traffic stops without conscious awareness of having done so, such a simple denial is insufficient to rebut the reasonable inference.  See Commonwealth v. McCowen, 458 Mass. 461, 499 (2010) (Ireland, J., concurring) ("people possess [implicit racial biases] over which they have little or no conscious, intentional control" [citation omitted]); Givelber, The Application of Equal Protection Principles to Selective Enforcement of the Criminal Law, 1973 U. Ill. L.F. 88, 114 (1973) (where "there is no legitimate justification for a given instance of selective enforcement, then the unjust treatment by the prosecutor should violate the equal protection clause regardless of whether the prosecutor knew he was abusing his office").

      The Commonwealth did not call an expert or present any statistical evidence.  As discussed, supra, the prosecutor's primary argument was that analyses based on FIOs and United States Census tracts were unreliable.  Additionally, the Commonwealth argued that Black drivers were overrepresented in the statistical data because Black individuals commit more crimes.  "[W]e are unaware of any reliable study establishing that motor vehicle violations are more frequently committed by any particular race of driver."  Lora, 451 Mass. at 442 n.30, citing Soto, 324 N.J. Super. at 74.  Thus, the Commonwealth clearly failed to rebut the reasonable inference of impermissible discrimination raised by the defendant, and the denial of the motion to suppress must be reversed.[20]

      f.  The potential for widely available statistics.  In 2019, the Legislature enacted G. L. c. 90, § 63, see St. 2019, c. 122, § 10, which requires, on an ongoing basis, an annual report of consolidated traffic stop data by town, and consolidation of that data to the registry of motor vehicles.  A bill currently under consideration by the Legislature would repeal G. L. c. 90, § 63, and instead would require law enforcement officers to collect more comprehensive data on each traffic stop they make, such that data by officer would be available to the municipality as well as the registry of motor vehicles.  See 2020 Senate Doc. No. 2820, § 52.  This bill would require all officers, not only those in police departments found to have engaged in racial profiling, to record information on any traffic stop (not just those resulting in an issued citation), including the reason for the stop and the age, race, ethnicity, and gender of the individual stopped, among other information.  See 2020 Senate Doc. No. 2820, § 52 (d) (1).  Further, the bill would require each municipal law enforcement department semiannually to publish a statistical analysis of the department's stop and search data.  See 2020 Senate Doc. No. 2820, § 52 (d) (5).  If enacted, the bill likely would enable defendants to access publicly available, department-wide data on the demographics of all traffic stops, by officer, in the relevant municipality, and would provide a plethora of relevant data available to support (or weaken) equal protection claims.  The House of Representatives, however, recently passed a revised version of the bill that completely omits the provisions requiring collection of data on traffic stops.  See 2020 House Doc. No. 4860.

      We urge the Legislature to require the collection and analysis of officer-specific data, such as set forth in 2020 Senate Doc. No. 2820.  This type of data collection would help protect drivers from racially discriminatory traffic stops, and also would protect police officers who do not engage in such discriminatory stops.

      4.  Conclusion.  The matter is remanded to the county court, where an order shall issue reversing the Superior Court judge's order denying the defendant's motion to suppress, and remanding the matter to the Superior Court for such other proceedings as are necessary, consistent with this decision.

So ordered.

       GANTS, C.J. (concurring).  I agree with the court that a motor vehicle stop motivated to any extent by the race of the driver or passenger (or by the driver's or passenger's membership in any suspect class) violates our guarantee of equal protection under arts. 1 and 10 of the Massachusetts Declaration of Rights, and that the fruits of any such unconstitutional stop must be suppressed.  See ante at    .  I also agree with the court that a reasonable inference of racial profiling does not require statistical evidence but may instead be based on the totality of evidence surrounding the stop.  See ante at    .

      I agree with Justice Budd's concurrence that a motor vehicle stop that is found unconstitutional as a violation of equal protection would also be an unreasonable stop in violation of art. 14 of our Declaration of Rights.  See post at    .  Specifically, I agree that a stop that is motivated by the race of the driver is not constitutionally reasonable.  I note that, despite our authorization rule, we have long considered the motivation of the law enforcement officer who conducts a search where it is claimed to be an inventory or administrative search; if the officer's motivation is investigative and the search is not lawful as an investigative search, the search is unconstitutional and the fruits of the search are suppressed.  See Commonwealth v. Rostad, 410 Mass. 618, 620 (1991) (inventory search "may not be allowed to become a cover or pretext for an investigative search"); Commonwealth v. Eagleton, 402 Mass. 199, 207 n.13 (1988) ("An administrative search may not be used as a subterfuge to avoid the burden of establishing probable cause to support a criminal investigative search").  These cases demonstrate that courts can, and do, successfully explore the true motivation of an officer in examining the lawfulness of a search.

      I do not join Justice Budd's concurrence because I would not declare in this case that any pretextual motor vehicle stop, even where the pretext is not based on the membership of the driver or passenger in a suspect class, is also a violation of art. 14.  See post at    .  Here, the only claim of pretext is based on the race of the driver; we need not decide in this case whether to extend the reach of our opinion to apply to all pretexts.

      In closing, I note that, despite our jurisprudential differences reflected in the various opinions in this case, the court is unanimous in concluding that a motor vehicle stop that arises from racial profiling is unconstitutional, that the burden placed on defendants in criminal cases to establish an inference of racial profiling need not be based on statistical evidence, that the burden may be met by inferences arising from the totality of the evidence of the stop, and that the burden should not be so high that a remedy is, in practice, mostly illusory.  In short, it is the unanimous view of this court that the prohibition against racial profiling must be given teeth and that judges should suppress evidence where a motor vehicle stop is motivated, even in part, by the race of the driver or passenger.

      BUDD, J. (concurring, with whom Lenk, J., joins).  Racial profiling in traffic stops has proven to be an intractable problem, the devastating consequences of which members of this court have recognized for many years.  See Commonwealth v. Buckley, 478 Mass. 861, 871 (2018); id. at 876 (Budd, J., concurring); Commonwealth v. Lora, 451 Mass. 425, 449 (2008) (Ireland, J., concurring); Commonwealth v. Feyenord, 445 Mass. 72, 88 (2005) (Greaney, J., concurring), cert. denied, 546 U.S. 1187 (2006); Commonwealth v. Gonsalves, 429 Mass. 658, 670 (1999) (Ireland, J., concurring).  Today, the court expressly acknowledges that discriminatory motor vehicle stops are profoundly harmful to persons and communities of color, and adjusts our existing equal protection framework for addressing such stops.

      I agree that the statistical evidence provided by this defendant was more than sufficient to show that the traffic stop was racially motivated under the requirements set forth in Lora, supra at 436-438.  I also agree that Lora inadvertently set the bar too high for defendants to meet their initial burden of raising an inference of racial discrimination.  However, the problem of racially discriminatory motor vehicle stops is a systemic product of the criminal justice system as it currently functions -- it thus requires a systemic solution.

      In looking at the issue more broadly, it is clear that the root of the problem is pretextual stops, which allow police to utilize traffic stops as a means to act on hunches that are unsupported by reasonable suspicion and often based on the race of the driver.  I conclude that pretextual stops are unconstitutional under art. 14 of the Massachusetts Declaration of Rights[1] because they allow for the investigatory stop of an individual without reasonable suspicion of the crime sought to be investigated.  For that reason, I would prohibit all pretextual stops.[2]

      The court seeks to solve the problem of race-based traffic stops by improving upon the traditional equal protection analysis.  However, the ability to challenge alleged race-based stops on both equal protection and art. 14 grounds would enhance the ability of people of color to pursue an effective remedy against discrimination.  That is, I fear that the efficacy of the equal protection test remains blunted by the ability of police to use pretextual stops to disguise stops based on racial bias.  The long, difficult history of racial discrimination in law enforcement demonstrates that, without more, making it easier for defendants to raise an inference that race was the basis for their stops in discrete cases will not be enough to dismantle the practice of racial profiling.  Acknowledging the unconstitutionality of pretextual stops has the added systemic benefit of removing, in the first instance, the means by which racial profiling is accomplished.

      1.  Pretext and the systemic problem of racially motivated motor vehicle stops.  Under the authorization test, a traffic stop is valid "so long as the police are doing no more than they are legally permitted and objectively authorized to do."  Buckley, 478 Mass. at 865, quoting Commonwealth v. Santana, 420 Mass. 205, 209 (1995).[3]  In practice, this means that an officer is permitted to stop a vehicle so long as he or she has observed a motor vehicle violation (or otherwise has either probable cause or reasonable suspicion to believe that one was committed).  See Commonwealth v. Rodriguez, 472 Mass. 767, 774 (2015).  The authorization test therefore permits police to perform pretextual motor vehicle stops, i.e., stops ostensibly made on the basis of a motor vehicle violation, but actually made for the purpose of investigating suspicions of unrelated criminal activity.  See State v. Ladson, 138 Wash. 2d 343, 349 (1999) ("[T]he essence of [a] pretextual traffic stop is that the police are pulling over a citizen, not to enforce the traffic code, but to conduct a criminal investigation unrelated to the driving.  Therefore the reasonable articulable suspicion that a traffic infraction has occurred which justifies an exception to the warrant requirement for an ordinary traffic stop does not justify a stop for criminal investigation").

      "[D]riving is an indispensable part of modern life."  Commonwealth v. McCarthy, 484 Mass. 493, 507 (2020).  In addition, our traffic code is comprehensive and detailed.[4]  As a result, "[v]ery few drivers can traverse any appreciable distance without violating some traffic regulation" (quotation and citation omitted).  LaFave, The "Routine Traffic Stop" from Start to Finish:  Too Much "Routine," Not Enough Fourth Amendment, 102 Mich. L. Rev. 1843, 1853 (2004).  See Caldwell v. State, 780 A.2d 1037, 1048 n.25 (Del. 2001), quoting Whitehead v. State, 116 Md. App. 497, 507 n.4 (1997) ("studies conducted on a stretch of Interstate 95 between Baltimore and Delaware revealed that 93% of all drivers committed some type of traffic violation").  See also Caldwell, supra, citing Harris, Whren v. United States:  Pretextual Traffic Stops and "Driving While Black," The Champion (March 1997), at 41; State v. Soto, 324 N.J. Super. 66, 70 (1996) (observational study over four days on three-exit stretch of New Jersey turnpike showed 98.1 percent of 2,096 observed vehicles exceeded speed limit).

      Pretextual stops are an attractive investigatory technique because a traffic stop can lead to a multitude of additional investigatory actions.  See, e.g., Commonwealth v. Goncalves-Mendez, 484 Mass. 80, 81 (2020) (traffic stop led to discovery of bench warrant, arrest, impoundment, and inventory search); Commonwealth v. Amado, 474 Mass. 147, 150 (2016) (exit order and patfrisk during traffic stop).  See also Carbado, From Stopping Black People to Killing Black People:  The Fourth Amendment Pathways to Police Violence, 105 Cal. L. Rev. 125, 151 (2017) (Carbado) (describing "the ways in which traffic stops function as gateways to more intrusive searches and seizures").  Thus, law enforcement officers have powerful incentives to use traffic violations as a pretext to conduct investigatory stops.

      It is no secret that this combination of factors has allowed racially motivated motor vehicle stops to flourish.  See, e.g., Carbado, supra at 129, 152 (describing "de facto legalization" of racial profiling via cases "in which Fourth Amendment law turns a blind eye to racial profiling or makes it easy for the police to get away with the practice," including Whren v. United States, 517 U.S. 806 [1996]); Capers, Rethinking the Fourth Amendment:  Race, Citizenship, and the Equality Principle, 46 Harv. C.R.-C.L. L. Rev. 1, 34 (2011) (Capers) ("Given that drivers routinely violate traffic laws . . . [Whren] virtually gives officers carte blanche to engage in race-based pretextual stops.  And if the driving while [B]lack statistics . . . show anything, this is what officers do"); Johnson, How Racial Profiling in America Became the Law of the Land:  United States v. Brignoni-Ponce and Whren v. United States and the Need for Truly Rebellious Lawyering, 98 Geo. L.J. 1005, 1069 (2010) (summarizing critiques of Whren); Glasser, American Drug Laws:  The New Jim Crow, 63 Alb. L. Rev. 703, 708 (2000) ("We are talking about a national policy which is training police all over this country to use traffic violations, which everyone commits the minute you get into your car, as an excuse to stop and search people with dark skin").

      If "systemic racism" is defined as a "system[ or] institution[] that produce[s] racially disparate outcomes, regardless of the intentions of the people who work within [it]," then our criminal justice system is rife with it.  See There's Overwhelming Evidence That the Criminal Justice System is Racist.  Here's the Proof, Washington Post, June 10, 2020, https://www.washingtonpost.com/graphics/2020/opinions

/systemic-racism-police-evidence-criminal-justice-system/ [https://perma.cc/8YLM-KWSY].  As applicable here, allowing ostensibly routine traffic stops produces racially disparate outcomes.  In order adequately to address this systemic problem, we must take a clear-eyed look at pretextual stops.

      2.  The unconstitutionality of pretextual stops.  The requirement that government searches or seizures require a warrant supported by probable cause is the touchstone of art. 14's protection against arbitrary searches and seizures.  See Commonwealth v. Bostock, 450 Mass. 616, 623–624 (2008), quoting Commonwealth v. Cast, 407 Mass. 891, 901 (1990) ("It is a cardinal principle that searches conducted outside the judicial process, without prior approval by judge or magistrate, are per se unreasonable [under both the Fourth Amendment to the United States Constitution and art. 14] -- subject only to a few specifically established and well-delineated exceptions" [quotations omitted]).  See also Commonwealth v. Antobenedetto, 366 Mass. 51, 57 (1974) ("under the Fourth Amendment searches conducted without valid warrants are presumed in the first instance to be unreasonable.  It is then up to the government to show that a particular search falls within a narrow class of permissible exceptions").

      The various exceptions to the warrant requirement, and the exceptions to those exceptions, make up our search and seizure jurisprudence.  We carefully scrutinize warrantless actions taken by law enforcement, recognizing that, where the power of police to conduct warrantless searches or seizures is "too permeating," it necessarily is "too susceptible to being exercised arbitrarily by law enforcement -- precisely the type of governmental conduct against which the framers sought to guard" (quotations and citation omitted).  Commonwealth v. Almonor, 482 Mass. 35, 47 (2019).

      Although a motor vehicle stop constitutes a seizure for the purposes of art. 14, we have held that an investigatory stop of a vehicle, like a Terry-type stop of a pedestrian, does not require a warrant because it is by its nature a threshold, on-the-spot inquiry that is less intrusive than an arrest.  See Commonwealth v. Willis, 415 Mass. 814, 818 (1993); Commonwealth v. Bacon, 381 Mass. 642, 643 (1980).  See generally Terry v. Ohio, 392 U.S. 1, 20 (1968).  Motorists nevertheless are ostensibly protected from unreasonable searches and seizures because, ordinarily, to pass constitutional muster under art. 14 a warrantless investigatory stop (seizure) of a motor vehicle and its occupants requires "reasonable suspicion, based on specific, articulable facts and inferences therefrom, that an occupant . . . had committed, was committing, or was about to commit a crime" (citation omitted).  Commonwealth v. Manha, 479 Mass. 44, 46 (2018).  However, the authorization test strips away that protection because it substitutes reasonable suspicion of a traffic violation for reasonable suspicion of the separate criminal conduct that the officer seeks to investigate.[5]  That is, as discussed supra, the rule permits officers to make an investigatory stop without the reasonable suspicion normally required as long as they have observed a traffic violation that can be used as pretext for the stop.

      In practice, if an officer wants to investigate a hunch about suspected criminal activity in connection with a particular motor vehicle, he or she has two options.  If the officer has reasonable suspicion of the suspected criminal activity, he or she may conduct an investigatory stop.  If not, the officer can simply wait until the driver commits a traffic violation, stop the vehicle based on that violation, and then attempt to get more information during the stop to corroborate that hunch.  See Commonwealth v. Cordero, 477 Mass. 237, 241 (2017) ("A routine traffic stop may not last longer than reasonably necessary to effectuate the purpose of the stop" [quotations and citation omitted]).

      The problem with the authorization test is that it automatically categorizes any stop preceded by a traffic violation as "authorized," and therefore presumptively reasonable, regardless of the actual motivation for the stop.  See Buckley, 478 Mass. at 869.  Because the test predetermines the outcome of any claim of unreasonableness for the purposes of art. 14, no actual reasonableness analysis is required (or allowed).  Incredibly, this is true even if the motive is unlawful.  See Santana, 420 Mass. at 209 ("Under [the authorization] test, it is irrelevant whether a reasonable police officer would have made the stop but for the unlawful motive" [citations omitted]).  In this case, for example, the officers discovered the inspection sticker violation prior to making the stop.[6]  Pursuant to Santana, the stop would likely be considered constitutionally reasonable under art. 14, even though the court concludes, and I agree, that the defendant successfully showed that it was racially motivated.  Such an outcome clearly is indefensible -- it hardly can be argued that a motor vehicle stop predicated on one's race is reasonable in any circumstance.

      The authorization test's reflexive and inflexible rule is at odds with the origins and purpose of art. 14.  Our cases acknowledge that the framers "wrote constitutional search protections in 'response to the reviled general warrants and writs of assistance of the colonial era, which allowed British officers to rummage through homes in an unrestrained search for evidence of criminal activity'" (quotations omitted).  McCarthy, 484 Mass. at 498-499, quoting Carpenter v. United States, 138 S. Ct. 2206, 2213 (2018).  See Commonwealth v. Blood, 400 Mass. 61, 71 (1987) (adoption of art. 14 motivated by "[o]pposition to the search policies . . . which allowed officers of the crown to search, at their will, wherever they suspected [evidence of criminal activity] to be" [citation omitted]).  The general warrant epitomized the type of government intrusion that is intolerable under our Constitution:  invasions of privacy which technically are cloaked in legal authorization but are exercised in an arbitrary or unfair manner.  Given the broad discretion afforded to police officers by way of the authorization test and the ample opportunities they have to exercise that discretion, pretextual investigatory stops are comparable to the general warrants that were the impetus for art. 14.  And like general warrants, the fact that pretextual investigative stops are legally "authorized" does not make them tolerable.  See Blood, supra.

      Notably, the nearly unlimited discretion granted to law enforcement officers to make pretextual traffic stops is an anomaly under our art. 14 jurisprudence.  For example, we do not permit investigatory stops of pedestrians unless police have reasonable suspicion "that a person has committed, is committing, or is about to commit a crime," based on "specific and articulable facts" (quotations and citations omitted).  See Bacon, 381 Mass. at 643.  "A mere 'hunch'" and "[s]imple good faith on the part of the officer" are not enough to justify an investigatory stop of a pedestrian (citation omitted).  Id.  Yet under the authorization test, the moment a driver commits (or the police discover) a motor vehicle violation, the occupants of a vehicle are exposed to the very same investigatory stops we rightly prohibit when they are on foot -- stops based on unsupported hunches, discrimination, harassment, or any other purpose lacking reasonable articulable suspicion of criminal activity.  See United States v. Botero-Ospina, 71 F.3d 783, 790 (10th Cir. 1995), cert. denied, 518 U.S. 1007 (1996) (Seymour, C.J., dissenting) ("Given the multitude of applicable traffic and equipment regulations in any jurisdiction . . . upholding a stop on the basis of a regulation seldom enforced opens the door to the arbitrary exercise of police discretion condemned in Terry and its progeny").

      We also previously have recognized, and limited, the discretion law enforcement officers may exercise in the context of exit orders during motor vehicle stops.  In Gonsalves, 429 Mass. at 660, 662, we concluded that although the Supreme Court held that the Fourth Amendment allows a police officer to issue exit orders during any lawfully executed traffic stop, see Pennsylvania v. Mimms, 434 U.S. 106, 111 (1977) (drivers), Maryland v. Wilson, 519 U.S. 408, 415 (1997) (passengers), art. 14 prohibits the practice unless the officer has a reasonable belief that the officer's safety, or the safety of others, is threatened.  We rejected the Federal rule that "permit[s] automobile exit orders during any traffic stop, but which do not require that such orders be given," because we recognized that this unfettered discretion is "a clear invitation to discriminatory enforcement of the rule" (emphasis added).  See Gonsalves, supra at 664.  We further noted that "[c]itizens do not expect that police officers handling a routine traffic violation will engage, in the absence of justification, in stalling tactics, obfuscation, strained conversation, or unjustified exit orders, to prolong the seizure in the hope that, sooner or later, the stop might yield up some evidence of an arrestable crime."  Id. at 663.

      Thus, we take care to protect pedestrians from investigatory stops without reasonable suspicion, and drivers from being unduly detained by an exit order during a traffic stop barring safety concerns.  However, the authorization test has created a markedly different standard when it comes to initiating a motor vehicle stop, as it provides officers with wide discretion to substitute pretext for reasonable suspicion as an investigatory tactic in the hopes that a hunch regarding criminal activity might bear fruit.

      The reasonable suspicion requirement is the linchpin of a valid investigatory stop under art. 14.  Using pretext to circumvent it breaks with our fundamental rules of search and seizure.  As observed by the Supreme Court of Washington:

"[T]he problem with a pretextual traffic stop is that it is a search or seizure which cannot be constitutionally justified for its true reason (i.e., speculative criminal investigation), but only for some other reason (i.e., to enforce traffic code) which is at once lawfully sufficient but not the real reason.  Pretext is therefore a triumph of form over substance; a triumph of expediency at the expense of reason.  But it is against the standard of reasonableness which our constitution measures exceptions to the general rule, which forbids search or seizure absent a warrant.  Pretext is result without reason."

Ladson, 138 Wash. 2d at 351.  If art. 14 is meant to protect individuals against the arbitrary exercise of power by agents of the Commonwealth, pretextual investigatory stops are in direct conflict with this objective.  As the authorization test creates a gaping hole in the foundational principle that a stop must be backed by reasonable suspicion, I would abandon it.[7]  Instead, I would hold that a traffic violation cannot replace the reasonable suspicion required to make an investigatory stop under art. 14.

      [3].  Detecting pretext:  the "would have" test.  Identifying a pretextual traffic stop may not be a straightforward task, but other courts and commentators have outlined what I conclude is a workable test.  Prior to Whren, many State courts, including courts in Arkansas, Illinois, Iowa, New York, Ohio, and Washington, employed various forms of the "reasonable officer," or "would have," test to determine the validity of a stop.  See State v. Brown, 930 N.W.2d 840, 902 (Iowa 2019) (Appel, J., dissenting) (collecting cases).  Although there are varying versions of that test, their common feature is that an alleged pretextual stop is valid only if a reasonable police officer "would have" made the stop in the absence of an ulterior motive; that is, a reasonable officer would have made the stop solely to enforce the motor vehicle infraction.

      A version of the "would have" test has been adopted in Washington and New Mexico.  See State v. Arreola, 176 Wash. 2d 284, 298 (2012).  However, it has been criticized as requiring a judge to discern an officer's subjective motives.  See Whren, 517 U.S. at 815; Buckley, 478 Mass. at 867-868.  See also Brown, 930 N.W.2d at 869 n.16 (Cady, C.J., dissenting) ("despite adopting the ['would have'] test, Washington courts may be reluctant to find that a police officer is lying about their motivations or 'have difficulty discerning pretextual behavior without an admission'" [citation omitted]); Lawton, The Road to Whren and Beyond:  Does the "Would Have" Test Work?, 57 DePaul L. Rev. 917, 935-936, 956-957 (2008).

      Under my proposed formulation of the test, the defendant need not prove, and the motion judge is not required to determine, the officer's true motive.  Instead, the question would be whether a reasonable officer would have made the stop solely for the purpose of traffic enforcement.[8]  See State v. Sullivan, 340 Ark. 315, 318 (2000), cert. granted, judgment rev'd, 532 U.S. 769 (2001) ("Claims of pretextual arrest raise a unique problem in law -- deciding whether an ulterior motive prompted an arrest which otherwise would not have occurred.  Confusion can be avoided by applying a 'but for' approach, that is, would the arrest not have occurred but for the other, typically, the more serious crime. . . .  The question then becomes whether [the defendant] would have been arrested simply for" purported traffic violation).  See also State v. Ochoa, 146 N.M. 32, 44 (2008); Abramovsky & Edelstein, Pretext Stops and Racial Profiling After Whren v. United States:  The New York and New Jersey Responses Compared, 63 Alb. L. Rev. 725, 734-735 (2000) (describing objective circumstantial factors New York courts considered when identifying pretext).

      Like the revised equal protection test announced by the court today, this version of the "would have" test allows a defendant to raise an inference that a stop was a pretext based on the circumstances of the stop.  It would allow for consideration of the same factors and likewise would operate under a burden-shifting framework.  See Ochoa, 146 N.M. at 44 (listing factors).  See also State v. Heath, 929 A.2d 390, 403 (Del. Super. Ct. 2006).  The mechanics of this test to detect pretext, then, are very similar to the one the court presents to detect racial profiling.[9]

      Importantly, because all underlying motives lacking reasonable suspicion would be considered unreasonable under art. 14, and therefore impermissible,[10] this version of the test would avoid the fraught inquiry into whether an officer was in fact motivated by racial bias.[11]  The advantage of an objective "would have" test is that whatever purportedly race-neutral justification the Commonwealth might proffer as the true motive for the stop would be irrelevant to the question whether a reasonable officer would have made the stop as a traffic stop, absent any motive other than traffic enforcement.  Thus, the Commonwealth would not be able to justify a pretextual stop under the "would have" test I propose simply by substituting a facially neutral motive for the alleged race-based motive.[12]

      We have remarked that, generally, "police conduct is to be judged 'under a standard of objective reasonableness without regard to the underlying intent or motivation of the officers involved.'"  Buckley, 478 Mass. at 867, quoting Santana, 420 Mass. at 208.  The "would have" test, the hallmark of which is the reasonableness of an officer's actions pursuant to art. 14, squares neatly with art. 14's focus on objective reasonableness.

      4.  The court's approach.  To be sure, the revised equal protection test announced today will provide a measure of relief to defendants who raise claims of having been racially profiled.  Indeed, as the court indicates, there are likely instances in which the revised equal protection analysis will be able to detect of racial profiling where an art. 14 inquiry would not.  However, this revised test strengthens what is, in the end, a partial work-around to a complex problem that we have yet to solve.

      a.  Reliance on the equal protection approach alone leaves in place the unconstitutional practice of pretextual stops.  As discussed, pretextual stops are unreasonable pursuant to art. 14's prohibition against unreasonable seizures.  Because the equal protection approach taken by the court prohibits only those pretextual stops motivated by race, the court's solution preserves the permissibility of pretextual stops based on any other ulterior motive.  And because the court concludes that it is unnecessary to analyze such stops under art. 14, investigatory stops made without reasonable suspicion of the criminal activity sought to be investigated will continue unabated.

      It appears that the court previously has not examined the constitutionality of pretextual stops from an art. 14 perspective.  When we essentially adopted the authorization test in Santana, 420 Mass at 208-209, we did not consider specifically the art. 14 consequences of automatically validating pretextual stops.  In that case, although the defendant argued that a broken tail light was used as a pretext to stop and search the vehicle, the motion judge did not so find.  Instead, the judge found that "the stop of the vehicle for defective equipment was a matter of routine standard police procedure."  Santana, supra at 209, citing Commonwealth v. Matchett, 386 Mass. 492, 510–511 (1982).  The Santana court accepted that finding and simply concluded that "the record does not support the contention that the troopers stopped the automobile in order to search it or to interrogate the defendants regarding illegal drug activities."  Id. at 209.  Thus, pretext was discussed only briefly and was not truly at issue.

      Further, although the observation made in Santana, supra, that "Massachusetts cases follow the authorization approach" was supported with citations to prior Massachusetts cases, see id. at 208, pretext was not at issue in any of those cases, either.  See Commonwealth v. Petrillo, 399 Mass. 487, 489 (1987) (arrest for trespass); Commonwealth v. Ceria, 13 Mass. App. Ct. 230, 235 (1982) (Terry-type analysis of stop of man riding moped in park by officers on foot); Commonwealth v. Tisserand, 5 Mass. App. Ct. 383, 386 (1977) (automobile double-parked on public street appropriately approached by police).

      Since Santana was decided, we specifically have affirmed the principle that a stop legally authorized by the observation of a motor vehicle violation, including where the observed violation is pretext for an unrelated motivation, is per se reasonable for the purposes of art. 14, but we have not explained why.  See Buckley, 478 Mass. at 869 ("a traffic stop cannot be 'arbitrary,' because it is predicated on a driver violating a traffic law").  See also, e.g., Larose, 483 Mass. 323 327 (2019); Buckley, supra at 865-866; Feyenord, 445 Mass. at 75-76.

      In Lora, although the defendant challenged his stop on both equal protection and art. 14 grounds, we dispensed with the art. 14 argument by quoting the authorization test, and then went on to analyze only the equal protection claim.  See Lora, 451 Mass. at 435-436.  In Buckley, 478 Mass. at 781, because the defendant had "expressly disavowed any . . . argument that race was a factor in the stop at issue," we declined to reexamine "our general art. 14 standard governing the reasonableness of traffic stops."  Instead, we reviewed Lora's discussion of Santana and Whren, and summarily stated that, "to the extent we do consider the purpose of a stop when assessing its validity, we do so pursuant to the equal protection principles of arts. 1 and 10 [of the Massachusetts Declaration of Rights] -- not art. 14's guarantee against unreasonable seizures."  Id. at 870.  Thus, despite the fact that Santana often is cited for the proposition that pretextual stops are valid, see, e.g., Larose, 483 Mass. at 327; Buckley, 478 Mass. at 865; Commonwealth v. Blevines, 438 Mass. 604, 608 (2003); Commonwealth v. Gentile, 437 Mass. 569, 576 (2002); Commonwealth v. Barros, 435 Mass. 171, 180 n.4 (2001), there is no case of which I am aware that specifically has considered whether using pretext to make an investigatory stop without reasonable suspicion of the crime sought to be investigated is a violation of art. 14, and if not, why not.[13]

      Today, the court reiterates that "the constitutional basis for objecting to intentionally discriminatory application of laws is the Equal Protection Clause, not the Fourth Amendment" or art. 14, but fails to explain why.  See Lora, 451 Mass. at 436, quoting Whren, 517 U.S. at 813.  However, a pretextual stop also properly is analyzed under art. 14, just like any other warrantless stop.  Without reasonable suspicion of criminal activity (other than the traffic violation), I conclude that such stops are not justifiable under art. 14.  For that reason, I conclude that the court's decision today leaves in place an investigatory practice that is unconstitutional.

      b.  The revised equal protection test is necessary but is not alone sufficient.  The revisions to the test the court announces today undoubtedly will make it easier for a defendant to raise an inference of discriminatory intent.  However, because the authorization test is left intact, it will be possible for the Commonwealth to rebut an inference of racial profiling by pointing to a hunch of criminal activity, based on factors that ostensibly are race-neutral but are in fact proxies for race.[14]

      For example, an officer could testify that the stop was based on the officer's recognition of the vehicle or an occupant from a prior interaction or observation, conversations with other officers, or information in a gang database.  These are the same factors currently used by police to racially profile people of color.  See generally Commonwealth v. Warren, 475 Mass. 530, 539-540 (2016) (according to study of Boston police field interrogation and observation reports, "[B]lack men in the city of Boston were more likely to be targeted for police-civilian encounters such as stops, frisks, searches, observations, and interrogations.  Black men were also disproportionately targeted for repeat police encounters"); Inside the Boston Police Gang Database, WGBH, July 30, 2019, https://www.wgbh.org/news/local-news/2019/07/30/inside-the-boston-police-gang-database [https://perma.cc/N475-MS5K] (noting stark racial disparities and errors in individuals included in police gang database).

      In addition to well-disguised proxies for conscious racial bias, unconscious bias is also at play and by definition may not be easily identified.  See generally D. Weisburd & M.K. Majmundar, eds., Proactive Policing:  Effects on Crime and Communities, ch. 7, Racial Bias and Disparities in Proactive Policing, 251, 277-280 (2018) (collecting studies on unconscious racial bias); id. at 277 ("Although overt expressions of biased behavior have declined in society and among police, racial animus has not disappeared.  Rather, it has evolved"); Norton, Vandello, Sommers, & Darley, Mixed Motives and Racial Bias, 12 Psychol. Pub. Pol'y & L. 36, 39 (2006), and studies cited ("people are quite good at masking their biased behavior by couching it in more acceptable terms, both to avoid the appearance of impropriety and as part of a more general effort to view themselves and their choices positively").

      Furthermore, we know that judges traditionally are reluctant to reject race-neutral explanations based on what happens when they are required to rule on them in the context of jury selection.[15]  In fact, the Batson framework has been criticized for this very reason, i.e., the unwillingness of judges to make a finding that the nondiscriminatory reason proffered to explain a peremptory strike is not the actual reason for the strike.  See Batson v. Kentucky, 476 U.S. 79, 93-96 (1986); Bellin & Semitsu, Widening Batson's Net to Ensnare More Than the Unapologetically Bigoted or Painfully Unimaginative Attorney, 96 Cornell L. Rev. 1075, 1092, 1098-1099, 1127 (2011) (analysis of Federal court opinions and orders evaluating race-based Batson challenges from 2000 through 2009 revealed "[a] broad subset of reasons [accepted by courts to justify a challenged strike], while not race-based per se, seem to correlate with race, suggesting that an attorney seeking to eliminate all the members of a certain race from the jury could achieve much of that goal by focusing on purportedly 'race-neutral' factors that happen to correlate with race"); Cavise, The Batson Doctrine:  The Supreme Court's Utter Failure to Meet the Challenge of Discrimination in Jury Selection, 1999 Wis. L. Rev. 501, 545 (1999) ("The savvy litigator can succeed with the most blatant discriminatory purpose by a simple manipulation of the neutral explanation coupled with a dose of disingenuousness").

      Statistical data are not as susceptible to being explained away with race-neutral justifications and therefore will, as the court indicates, continue to have "unique advantages" over other types of evidence.  However, the available data on the racial demographics of motorists involved in traffic stops continues to be severely limited, which is the very reason that Lora did not work for defendants alleging racial profiling.[16]  Although this defendant was able to collect sufficient data (consisting of the six most recent years of officer-specific field investigation and observation reports and citations), it is unlikely that other defendants will be able to follow suit.[17]

      The most recent Statewide dataset on racial profiling in traffic stops of which we have been made aware is a May 2004 report prepared by experts at Northeastern University pursuant to St. 2000, c. 228.  Analyzing motor vehicle citations issued by every police department in the Commonwealth from April 1, 2001, through June 30, 2003, the experts concluded that 249 of the Commonwealth's 366 police departments appeared to have engaged in racial or gender profiling.[18]  See Farrell, McDevitt, Bailey, Andresen, & Pierce, Massachusetts Racial and Gender Profiling Study:  Final Report, Institute on Race and Justice of Northeastern Univ., at 1, 30 (May 4, 2004).

      For a defendant challenging a traffic stop by an officer in one of those 249 departments, this 2004 study may be sufficient to raise an inference of racial profiling if unrebutted by more recent data.  But for defendants stopped by officers of any of the 117 departments that were not found to have engaged in racial profiling in 2004, it is unclear whether any data currently exists that could raise a statistical inference of racial profiling.  More generally, data showing an over-all pattern of racially evenhanded stops by a department (or by an individual officer) may belie particular instances of racial profiling that are not apparent in the aggregation of numerous stops.  The vast majority of stops by an officer or department may be race-neutral; however, data demonstrating an over-all pattern of fairness should not preclude a remedy for an outlier racially motivated stop.

      Despite the foregoing, the court's revised test unquestionably increases a defendant's chances of demonstrating racial discrimination.  Indeed, the scenarios the court foresees in which traffic stops that pass muster under the "would have" test are nonetheless products of racially discriminatory decisions about where to deploy police officers, illustrate the continuing importance of the equal protection remedy.  Nevertheless, the ability to proceed on both equal protection and art. 14 grounds would give defendants an even greater opportunity to establish illegal discrimination.

      c.  Equal protection is not the only constitutional guarantee available to address racial discrimination.  Today the court continues to analyze claims of racial profiling solely through an equal protection lens, without explaining why such claims do not also implicate art. 14.  The court's failure to engage this question at all is particularly bewildering because the answer seems straightforward:  surely a stop based on race is an unreasonable seizure under art. 14.

      Historically, equal protection of the laws, enshrined in both the Declaration of Rights and the United States Constitution, never has been the only constitutional avenue for redressing racial injustices.  In fact, many Supreme Court cases that granted to all criminal defendants what are now basic procedural rights originated from the mistreatment of people of color.  See, e.g., Miranda v. Arizona, 384 U.S. 436, 444-445 (1966) (right to Miranda warnings); Gideon v. Wainwright, 372 U.S. 335, 344-345 (1963) (right to counsel); Powell v. Alabama, 287 U.S. 45, 68-69 (1932) (same); Brown v. Mississippi, 297 U.S. 278, 287 (1936) (right to be interrogated free from coercion); Capers, supra at 7 n.45 (noting that "an earlier draft of [Miranda] was explicit about the racial dynamics of police interrogations"); id. at 8 (noting that in Gideon, "relying heavily on the racially-tinged Scottsboro Boys case . . . the Court clearly recognized the impact its decision would have on minority defendants especially, given the correlation, at the time [and today], between race and indigence").  See also Klarman, The Racial Origins of Modern Criminal Procedure, 99 Mich. L. Rev. 48, 48(2000) ("the linkage between the birth of modern criminal procedure and southern [B]lack defendants is no fortuity"); Kahan & Meares, Foreword:  The Coming Crisis of Criminal Procedure, 86 Geo. L.J. 1153, 1153 (1998) ("The need that gave birth to the existing criminal procedure regime was institutionalized racism"); Pye, The Warren Court and Criminal Procedure, 67 Mich. L. Rev. 249, 256 (1968) ("The Court's concern with criminal procedure can be understood only in the context of the struggle for civil rights").  As succinctly put by Kahan & Meares, supra:

"Law enforcement was a key instrument of racial repression, in both the North and the South, before the 1960's civil rights revolution.  Modern criminal procedure reflects the Supreme Court's admirable contribution to eradicating this incidence of American apartheid.  Supplanting the deferential standards of review that had until then characterized its criminal procedure jurisprudence, the Court, beginning in the 1960's and continuing well into the 1970's, erected a dense network of rules to delimit the permissible bounds of discretionary law-enforcement authority.  Although rarely couched as such, the unmistakable premise of these doctrines was the assumption that communities could not be trusted to police their own police because of the distorting influence of racism.

      There is no reason why a defendant claiming that a stop was motivated by race should be barred from seeking a remedy for an unreasonable seizure violative of art. 14 in addition to an equal protection claim.  Cf. Goodridge v. Department of Pub. Health, 440 Mass. 309, 320 (2003) (analyzing restrictions on marriage under equal protection and due process doctrines, "two constitutional concepts [which] frequently overlap").  See Brown, 930 N.W.2d at 920 (Appel, J., dissenting) ("Certainly, the theoretical availability of an equal protection claim should not preempt the possibility of a claim under search and seizure principles").[19]

      5.  Conclusion.  In 1999, then-Associate Justice Ireland broached the issue of racially discriminatory motor vehicle stops, noting that, by that time, the "widespread public concerns about police profiling, commonly referred to as 'DWB -- driving while [B]lack,' ha[d] been the subject of much discussion and debate both across the country and within the Commonwealth."  See Gonsalves, 429 Mass. at 670 (Ireland, J., concurring).  It would be nearly ten years before the court addressed the problem directly.  See Lora, 451 Mass. 426.  Unfortunately, for the reasons previously discussed, Lora was not effective in solving the problem of racially motivated motor vehicle stops.  Twelve years after Lora was decided, and more than two decades since the issue first was raised by Justice Ireland, we have a chance to revisit the question of how to craft a remedy that will provide relief to people in the Commonwealth who are targeted on the basis of the color of their skin.

      This time around we should confront squarely the fact that the phenomenon of racial profiling is a product of more than one-off cases of individual bias or animus -- it is a systemic problem that has flourished under the rules that this court has set.  A systemic solution requires more than an improved test for identifying individual instances of bias when they come before courts.  It requires a reevaluation of the rules that enable and incentivize officers to make pretextual race-based stops in the first place.[20]  Combining the court's improved test for identifying particular cases of race-based stops with a broader prohibition on pretextual stops would deter racial profiling and eliminate the tool most often used to accomplish it.  More generally, as long as we continue to allow pretextual stops, the search and seizure protections of art. 14 will continue to ring hollow.

      In the twenty-five years since deciding Santana, the court has not examined the art. 14 implications of the pretextual stops that are legitimized by the authorization test.  Given the opportunity to broaden the options available to combat racial profiling, it is disappointing that the court is willing to stand behind a rule that allows for pretextual stops without considering whether, and how, such stops are reasonable from an art. 14 standpoint.  See Amado, 474 Mass. at 151 n.4 (pretextual stops, "though lawful under our current jurisprudence, implicate important policy concerns about racial profiling in encounters between the police and persons of color"); Lora, 451 Mass. at 447 (Ireland, J., concurring), quoting Feyenord, 445 Mass. at 87 (Greaney, J., concurring) ("I repeat the observation of Justice Greaney that poorer citizens, who likely would include minorities, are more likely to be 'driving vehicles with defective equipment,' thus providing police with a legitimate reason to exercise discretion to stop them").

      As evidenced by the letter that the full court issued recently, we are united in the goal to "ensure that the justice provided to African-Americans is the same that is provided to white Americans."  Letter from the Seven Justices to Members of the Judiciary and Bar (June 3, 2020).  It is worth reiterating, however, that systemic or institutional racism produces racially disparate outcomes regardless of the intent of the people who work within the institution.  See id.  Our current state of affairs is not what any one of the Justices who comprise this court chose or would choose.  It nevertheless is a painful fact.

      I conclude that evaluating this form of racial profiling under art. 14 not only would resolve an inconsistency in our search and seizure jurisprudence but also would go a long way, at long last, toward addressing the systemic problem of racial profiling with a systemic solution.

      CYPHER, J. (concurring).  I agree with Justice Gaziano that lessening the burden on the defendant in the Lora test is the best approach to address the disparate effects of automobile stops on minority communities.  See Commonwealth v. Lora, 451 Mass. 425, 440-442 (2008).  I write separately to emphasize two points.

      First, the officers' stop of the defendant did not result from an observed traffic violation, but rather from conducting a query regarding the defendant's license plates.1  There was no observed traffic violation that motivated the officers to query the plates; however, there is no expectation of privacy in a license plate number, which is required by law to be displayed conspicuously on the exterior of a vehicle.  G. L. c. 90, § 6.  See Commonwealth v. McCarthy, 484 Mass. 493, 502 & n.8 (2020), discussing Commonwealth v. Starr, 55 Mass. App. Ct. 590, 593-594 (2002).2

      Although the random stop of a motor vehicle has been held to be a violation of the Fourth Amendment to the United States Constitution, a random number plate check is not.  Starr, supra at 594.  See United States v. Diaz-Castaneda, 494 F.3d 1146, 1151-1152 (9th Cir.), cert. denied, 552 U.S. 1031 (2007).

      That does not mean, however, that a race-based decision to query plates is permissible.  See Starr, supra at 594 n.8.  Much like a decision to selectively perform a traffic stop based on race, the querying of plates based on race is a potential violation of the principles of equal protection.  Even though there are no limitations under art. 14 of the Massachusetts Declaration of Rights or the Fourth Amendment on querying license plates, the start of the analysis in some cases should be the decision to query the plates.  In the event that a judge determines, based on the totality of the circumstances, that an officer's decision to query plates was motivated by racial or ethnic bias, the remedy remains suppression of evidence obtained during the subsequent stop.  See Lora, 451 Mass. 425 439-440.

      Second, Massachusetts places the enforcement of motor vehicle laws with the police.  Under G. L. c. 90C, § 3 (A) (1), "if a police officer observes or has brought to the officer's attention the occurrence of a civil motor vehicle infraction," the officer may issue a warning or citation.  Indeed, in some circumstances, a police officer or his or her employer may face liability for failing to address a public safety hazard created by a driver.  See Irwin v. Ware, 392 Mass. 745, 764 (1984) (jury could have found that town's police officers had duty to plaintiffs and that police officers' negligence in failing to remove intoxicated driver proximately caused plaintiffs' injuries).  As Justice Budd's opinion has demonstrated, the problem we are confronting today is systemic in nature and therefore requires a systemic solution.  Such a solution, however, requires the full engagement of all of the relevant interests, including the public and the police, and perhaps the Legislature.  See, e.g., Berkeley council approves "omnibus motion" on police reform, Berkeleyside, July 15, 2020, https://www.berkeleyside.com/2020/07/15/berkeley-council-approves-omnibus-motion-to-reform-policing [https://perma.cc/S7UP-VGZL].

footnotes

      [1]Chief Justice Gants participated in the deliberation on this case and authored his concurrence prior to his death.

      [2]The defendant was charged with unlawful possession of a firearm, G. L. c. 269, § 10 (a), as well as an enhancement for two previous convictions of violent crimes or serious drug offenses, G. L. c. 269, § 10G; carrying a loaded firearm, G. L. c. 269, § 10 (n); possession of ammunition, G. L. c. 269, § 10 (h); possession of a large capacity feeding device, G. L. c. 269, § 10 (m); and receiving a firearm with a defaced identification number, G. L. c. 269, § 11C.

      [3]We acknowledge the amicus briefs submitted by the Committee for Public Counsel Services and the American Civil Liberties Union of Massachusetts, Inc.; and by the Massachusetts Association of Criminal Defense Lawyers, The New England Innocence Project, Lawyers for Civil Rights, and the Charles Hamilton Houston Institute for Race and Justice at Harvard Law School.

      [4]The primary purpose of the Boston police department's youth violence strike force is proactively to reduce gang and gun violence and drug activity in Boston.

      [5]At the hearing on the defendant's motion to suppress, one of the officers testified that, in cases where he was arresting the driver and the driver was not the owner of the vehicle, he routinely had the vehicle towed unless the owner arrived at the scene.

      [6]These protections, of course, also apply to suspect classifications other than those based on race.  "The Equal Protection Clause was intended to work nothing less than the abolition of all caste-based and invidious class-based legislation."  Plyler v. Doe, 457 U.S. 202, 213 (1982).  Under the Federal Constitution, the guarantee of equal protection forbids the government from making suspect classifications, include those based on race, religion, nationality, alienage, or membership in another discrete and insular minority, unless the governmental action survives strict scrutiny.  See, e.g., New Orleans v. Dukes, 427 U.S. 297, 303-304 (1976); Graham v. Richardson, 403 U.S. 365, 372 (1971), citing United States v. Carolene Prods. Co., 304 U.S. 144, 152-153 n.4 (1938); Graham, supra at 376.  See also Bolling v. Sharpe, 347 U.S. 497, 500 (1954) (applying equal protection guarantee against Federal government under Fifth Amendment to United States Constitution).  Our State Constitution goes further and applies strict scrutiny review to sex and gender classifications as well.  See Finch v. Commonwealth Health Ins. Connector Auth., 459 Mass. 655, 665-666 (2011), S.C., 461 Mass. 232 (2012).

      [7]To make such a demonstration, a defendant might point to an officer's patterns of enforcement before and after the stop at issue.  It could be probative, for example, if a significant percentage of stops made by the officer in the preceding weeks or months involved drivers of the same race being stopped for minor traffic infractions, while those of other races were not.  Or, if the officer repeatedly noted the same minor infraction, such as a failure to signal a lane change, while stopping drivers who shared the same protected class as the defendant.  Such evidence need not be demonstrated to be statistically valid in order to support a reasonable inference.

      [8]Traffic stops initiated by officers whose primary assignment does not involve the enforcement of traffic laws might warrant particular scrutiny.  For example, an officer working routine patrol might write many tickets as part of ordinary duties, as compared to an officer working a specialized assignment such as drug enforcement task force, hostage rescue, or a domestic violence unit.

      [9]It could be relevant that officers observed or followed a vehicle for an extended period of time prior to making the stop, see State vs. Deleon, N.M. Ct. App., No. 30,813, slip op. at 8 (Feb. 14, 2013) (officer followed defendant for multiple miles before stop); State v. Arreola, 176 Wash. 2d 284, 301 (2012) (Chambers, J., dissenting) ("officer noticed, after following the car he wished to stop for a half mile or so, that its exhaust system was not in compliance with traffic regulations"); a judge also might consider whether the circumstances would have allowed the officer to note the defendant's race, see State v. Snapp, 174 Wash. 2d 177, 199-200 (2012) (on dark night, officer could not see race of defendant).

      [10]A judge might examine whether the officer's conduct during the stop was consistent with, and limited to, that necessary to enforce the motor vehicle violation.  See, e.g., People v. Roundtree, 234 A.D. 2d 612, 613 (N.Y. 1996) (court considered fact that officer "did not question [the driver] about the [traffic] infraction or issue a traffic summons").  Cf. Commonwealth v. Cordero, 477 Mass. 237, 242 (2017) ("police inquiry in routine traffic stop must end upon production of valid license and registration" [citation omitted]).

      [11]For example, where the traffic infraction clearly implicated significant public safety concerns, such as operating under the influence of alcohol, those concerns would weigh against drawing an inference of discriminatory intent.

      [12]If an officer's actions in making the stop deviated from the policies and procedures of his or her department, such as a stop by an undercover officer where a department had a policy against making routine traffic stops in unmarked vehicles, the deviation might support an inference that the stop involved racial profiling.

      [13] See Commonwealth v. Buckley, 478 Mass. 861, 870 (2018); State v. Williams, 249 So.3d 527, 532–533 (Ala. Crim. App. 2017); Hamilton v. State, 59 P.3d 760, 765–766 (Alaska Ct. App. 2002), cert. denied, 540 U.S. 915 (2003); Jones v. Sterling, 210 Ariz. 308, 311 (2005); State v. Mancia-Sandoval, 361 S.W.3d 835, 839 (Ark. 2010); People v. Woods, 21 Cal. 4th 668, 680 (1999), cert. denied, 529 U.S. 1023 (2000); People v. Rodriguez, 945 P.2d 1351, 1359–1360 (Colo. 1997); State v. Jones, 113 Conn. App. 250, 265 (2009) (noting Federal rule and declining to address under State Constitution); Dobrin v. Florida Dep't of Highway Safety & Motor Vehicles, 874 So. 2d 1171, 1173 (Fla.), cert. denied, 543 U.S. 957 (2004); State v. Holler, 224 Ga. App. 66, 70 (1996); State v. Bolosan, 78 Haw. 86, 94 (1995); State v. Myers, 118 Idaho 608, 610 (Ct. App. 1990); People v. Gray, 305 Ill. App. 3d 835, 839 (1999); Mitchell v. State, 745 N.E.2d 775, 787 (Ind. 2001); Brown, 930 N.W.2d at 848-855, and cases cited; State v. Hardyway, 264 Kan. 451, 456 (1998); Moberly v. Commonwealth, 551 S.W.3d 26, 29 (Ky. 2018), reh'g denied (Aug. 16, 2018); State v. Waters, 780 So. 2d 1053, 1056 (La. 2001); State v. Bolduc, 722 A.2d 44, 45 (Me. 1998); Thornton v. State, 465 Md. 122, 135 (2019); People v. Labelle, 478 Mich. 891, 891 (2007), overruled on other grounds by People v. Mead, 503 Mich. 205 (2019); State v. George, 557 N.W.2d 575, 578-579 (Minn. 1997); Martin v. State, 240 So. 3d 1047, 1051–1052 (Miss. 2017), cert. denied, 138 S. Ct. 2592 (2018); State v. Smith, 595 S.W.3d 143, 145-146 (Mo. 2020); Brunette v. State, 383 Mont. 458, 465 (2016); State v. Draganescu, 276 Neb. 448, 460–461 (2008); Doyle v. State, 116 Nev. 148, 155 (2000); State v. McBreairty, 142 N.H. 12, 15 (1997); State v. Bacome, 228 N.J. 94, 103 (2017); People v. Robinson, 97 N.Y.2d 341, 349 (2001); State v. McClendon, 350 N.C. 630, 635–636 (1999); State v. Bartelson, 704 N.W.2d 824, 827-828 (N.D. 2005); Dayton v. Erickson, 76 Ohio St. 3d 3, 11 (1996); Lozoya v. State, 932 P.2d 22, 32 (Okla. Crim. App. 1996); State v. Olaiz, 100 Or. App. 380, 386-387 (1990); Commonwealth v. Chase, 599 Pa. 80, 102 (2008); State v. Bjerke, 697 A.2d 1069, 1073 (R.I. 1997); Milledge v. State, 422 S.C. 366, 375 (2018); State v. Sleep, 590 N.W.2d 235, 237-238 (S.D. 1999); State v. Donaldson, 380 S.W.3d 86, 92 (Tenn. 2012); Holder v. State, 595 S.W.3d 691, 698 (Tex. Crim. App. 2020); State v. Lopez, 873 P.2d 1127, 1134-1135 (Utah 1994); State v. Trudeau, 165 Vt. 355, 359 n.3 (1996); Thomas v. Commonwealth, 57 Va. App. 267, 273-274 (2010); Miller v. Chenoweth, 229 W. Va. 114, 120 (2012); State v. Houghton, 364 Wis. 2d 234, 250 (2015); Fertig v. State, 146 P.3d 492, 501 (Wyo. 2006).

      [14]Delaware has left the question unresolved, with one trial court decision applying the test and others rejecting it.  See Turner v. State, 25 A.3d 774, 777 (Del. 2011), citing State v. Heath, 929 A.2d 390 (Del. Super. Ct. 2006).

      [15]Moreover, Washington has deemed constitutional certain "mixed-motive" stops, in which an officer was motivated both by the need to address the traffic violation and by a desire to investigate other suspected criminal activity.  In these cases, Washington courts delve even further into the subjective motivations of the officer, and will consider a stop permissible if the officer "ma[de] an independent and conscious determination that a traffic stop to address a suspected traffic infraction [was] reasonably necessary in furtherance of traffic safety and the general welfare."  See Arreola, 176 Wash. 2d at 298–299.  Of course, under equal protection principles, a traffic stop motivated by race is unconstitutional even if the officer also is motivated by the legitimate purpose of enforcing traffic laws.

      [16]See also State v. Scharff, 284 P.3d 447, 451 (N.M. Ct. App. 2012); State vs. Shindledecker, N.M. Ct. App., No. 34,442, slip op. at 5 (Aug. 11, 2016); State vs. Gonzales, N.M. Ct. App., No. 30,188, slip op. (May 1, 2012); State vs. Dominguez, N.M. Ct. App., No. 29,741, slip op. at 5 (Oct. 20, 2010); State v. Wright, 155 Wash. App. 537, 559 (2010), reversed on other grounds sub nom. State v. Snapp, 174 Wash. 2d 177 (2012); State v. Nichols, 161 Wash. 2d 1, 11-12 (2007); State v. Myers, 117 Wash. App. 93, 96-98 (2003); State v. DeSantiago, 97 Wash. App. 446, 452 (1999).

      [17]Of the total number of the officers' vehicle-related field investigation and observation (FIO) reports, 80.62 percent involved Black drivers, as compared to a benchmark population of the census blocks containing those stops in which 44.67 percent of residents were Black.  The benchmark estimate that 44.67 percent of motorists on these roads were Black likely was an overestimate of the percentage of Black motorists on them, given that the proportion of Black residents was lower in all of the surrounding areas and municipalities.  Nonetheless, the expert concluded that even if 77 percent of the actual motorists were Black, there was still a statistically significant probability that Black drivers were more likely to be the subject of an FIO than other drivers.

      [18]Of all the traffic citations the officers issued in Boston, 56.59 percent were issued to Black drivers.  The expert compared this to a benchmark population for all of Boston, in which, in 2010, 24.38 percent of residents were Black.  In the absence of racial profiling, the odds of this disparity in citations occurring randomly is less than one in 100,000.  According to the expert's analysis, this data would support a statistically significant inference of discrimination in citations even if fifty percent of the drivers in Boston were Black.

      [19]In 2005, certain police departments, including the Boston police department, were required to collect data on "all traffic stops, including those not resulting in a warning, citation or arrest," for a period of one year after an initial review of citation data suggested they had engaged in racial or gender profiling.  See St. 2000, c. 228, § 10; Boston Police Patrolmen's Ass'n, Inc. v. Police Dep't of Boston, 446 Mass. 46, 48 (2006) ("249 of 366 Massachusetts law enforcement agencies appeared to have engaged in racial or gender profiling").  This case, however, arose more than a decade after the end of the statutory mandate to collect this "all stops" data.  It appears that no such data were available for the relevant time period and the particular officers who stopped the defendant.

      [20] Because of this conclusion, we need not reach the defendant's contention that the subsequent inventory search was unconstitutional.

footnotes for concurring

      [1]Article 14 of the Massachusetts Declaration of Rights states in pertinent part:  "Every subject has a right to be secure from all unreasonable searches, and seizures, of his person, his houses, his papers, and all his possessions."

      [2]The defendant has not challenged the stop under art. 14; however, as we declined to evaluate the art. 14 standard for determining reasonableness of a motor vehicle stop (i.e., the authorization test) less than two years ago in Commonwealth v. Buckley, 478 Mass. 861, 866 (2018), it is understandable that the defendant here did not raise what reasonably might have appeared to be a futile argument.  See generally Commonwealth v. Vasquez, 456 Mass. 350, 358-359 (2010).  In any case, the constitutionality of pretextual stops was well and fully briefed by both parties and by amici in Buckley.  The fundamental points I raise today were laid out before us then, together with citations to dozens of studies, law review articles, and statistical reports relating to racial profiling in traffic stops and pretextual stops.  See, e.g., Harris, The Stories, the Statistics, and the Law:  Why "Driving While Black" Matters, 84 Minn. L. Rev. 265, 273-274 (1999); Sklansky, Traffic Stops, Minority Motorists, and the Fourth Amendment, 1997 Sup. Ct. Rev. 271, 312-316 (1997); Davis, Race, Cops, and Traffic Stops, 51 U. Miami L. Rev. 425, 432 (1997).

      As is obvious from this concurrence, since Buckley was decided I have changed my view that it is "unworkable" to shift our jurisprudence away from the authorization test announced in Commonwealth v. Santana, 420 Mass. 205, 208-209 (1995).  Buckley, 478 Mass. at 876 (Budd, J., concurring).  See Commonwealth v. Larose, 483 Mass. 323, 336 (2019) (Lenk, J., concurring).

      [3]As discussed in more detail infra, this is the same approach that the United States Supreme Court adopted with regard to the Fourth Amendment to the United States Constitution.  See Whren v. United States, 517 U.S. 806, 813 (1996).

      [4] Our traffic code regulates nearly every aspect of operating a motor vehicle, from obvious moving violations such as speeding and failing to stop at red lights, see G. L. c. 90, § 17; G. L. c. 89, § 9, to less conspicuous moving violations such as failing to come to a full stop at a stop sign, see G. L. c. 89, § 9, or crossing marked lanes, see G. L. c. 89, § 4A, and subtle equipment deficiencies including window tint, lights, and passenger restraints, see G. L. c. 90, §§ 7, 7AA, 9D, 13A.

      [5] It is worth noting that many types of traffic violations are civil in nature, not criminal.  See, e.g., Commonwealth v. Rodriguez, 472 Mass. 767, 770-771 (2015).

      [6] As Justice Cypher notes, the stop here was not based on a moving or equipment violation.  Instead, the officers discovered the inspection sticker violation only after deciding to check the vehicle's license plate number in a database.  However, just like a traffic stop based on an observed motor vehicle violation, a traffic stop resulting from a plate check is a seizure pursuant to art. 14 and the Fourth Amendment.  See generally Buckley, 478 Mass. at 865, quoting Rodriguez, 472 Mass. at 773 ("Because '[a] police stop of a moving automobile constitutes a seizure,' . . . that stop must be reasonable in order to be valid under the Fourth Amendment and art. 14").

      [7] Although the court notes that after Whren was decided, a majority of States adopted a version of the authorization test, ante at    , it perhaps goes without saying that, regardless of what other jurisdictions have to say about their State constitutions, "ultimately we must accept responsibility for interpreting our own Constitution as text, precedent, and principle seem to us to require."  Planned Parenthood League of Mass., Inc. v. Attorney Gen., 424 Mass. 586, 590 (1997).

      [8]Although the court foresees difficulty in determining the characteristics of a reasonable officer, this standard is applied successfully in other circumstances.  See Commonwealth v. Barreto, 483 Mass. 716, 722 (2019), quoting Commonwealth v. Gonsalves, 429 Mass. 658, 661 (1999) (exit order based on safety concerns is justified where "a reasonably prudent [person] in the [officer's] position would be warranted in the belief that the safety of the police or that of other persons was in danger"); LaChance v. Commissioner of Correction, 463 Mass. 767, 777-778 (2012), quoting Longval v. Commissioner of Correction, 448 Mass. 412, 419 (2007) (qualified immunity standard asks whether "it would be clear to a reasonable officer that his conduct was unlawful in the situation he confronted").

      [9]These similarities are not surprising, because many of the factors the court identifies as relevant to the totality of the circumstances analysis, ante at notes 9 and 10, are drawn from cases applying other jurisdictions' versions of the "would have" test.  See State vs. Deleon, N.M. Ct. App., No. 30,813 at 4-5 (Feb. 14, 2013); People v. Roundtree, 234 A.D. 2d 612, 613 (N.Y. 1996); State v. Arreola, 176 Wash. 2d 284, 301 (2012) (Chambers, J., dissenting); State v. Snapp, 174 Wash. 2d 177, 200-201 (2012).

      [10]The court cautions that analyzing pretextual stops under art. 14 would confuse the inquiry into whether a stop is unlawful because "our jurisprudence on search and seizure provides no guidance regarding which officer motivations render unreasonable an otherwise permissible traffic stop."  Ante at    .  However, our art. 14 and Fourth Amendment jurisprudence is quite clear that an "unlawful" search or seizure is an unreasonable one -- that is, any motive lacking reasonable suspicion of criminal activity would render a pretextual stop unreasonable.  See generally Commonwealth v. Alvarado, 423 Mass. 266, 268 (1996); Commonwealth v. Bacon, 381 Mass. 642, 643 (1980).

      [11] The court posits that, like other versions of the "would have" test, this one inevitably would lead to an examination of subjective motives.  Ante at    .  I do not agree with this assessment; I conclude that much can be learned from the experiences in Washington and New Mexico where each employed an explicitly subjective test.  At any rate, as the court acknowledges, those subjective inquiries are "quite similar" to the court's equal protection analysis, which requires a subjective determination of whether racial bias was the true motive for the stop.  Ante at    .  Even assuming for the sake of argument that the "would have" test necessarily entails a subjective inquiry, the added benefits of undoing the authorization test, described infra, would justify the potential difficulty of formulating a workable test to identify pretext.

      [12] The court points out that, like the revised equal protection test, this formulation of the "would have" test would not eliminate issues of less than candid police officers, or of judges reluctant to reject an officer's proffered motive for a stop.  I am under no illusions that an art. 14 approach is "magic."  See ante at    .  Righting the wrongs of the racism that have been plaguing this country since its inception has been slow-going -- some might say glacial.  Although I conclude that the approach I suggest could be part of a larger solution, it will take much more than any single judicial pronouncement (or concurrence) to increase the pace of progress.  Despite any perceived flaws in my suggested approach, I conclude that defendants who pursue claims of racial discrimination in connection with a traffic stop should be entitled to all available remedies, especially where, as discussed, such a stop plainly is unconstitutional under art. 14.

      [13]Similarly, the Supreme Court's decision in Whren did not fully elaborate on the reasoning behind its oft-cited holding that "the constitutional basis for objecting to intentionally discriminatory application of laws is the Equal Protection Clause, not the Fourth Amendment" and that "[s]ubjective intentions play no role in ordinary, probable-cause Fourth Amendment analysis."  See Whren, 517 U.S. at 813.  The Whren Court rested its conclusion almost entirely on precedents that had not addressed the reasonableness of traffic stops made for a reason other than routine traffic enforcement.  See LaFave, The "Routine Traffic Stop" from Start to Finish:  Too Much "Routine," Not Enough Fourth Amendment, 102 Mich. L. Rev. 1843, 1856 (2004) ("By this reckless use of its own precedents, the Court in Whren makes it appear that the issue raised by the petitioners was already settled, while in fact it was very much an open question.  The fact that the Court created this false appearance perhaps explains why the Court in Whren had so little to say about the merits of the petitioners' claim").     

      [14]As discussed infra, even if race-neutral, any factors that together fail to establish reasonable articulable suspicion of criminal activity unrelated to a traffic violation are insufficient to conduct an investigatory stop under art. 14.

      [15] Under the Batson framework, a party may challenge a peremptory strike by raising an inference of discrimination which the opposing party may rebut by proffering a nondiscriminatory reason for the strike.  Batson v. Kentucky, 476 U.S. 79, 93-96 (1986).  See Commonwealth v. Soares, 377 Mass. 461, 486-488, cert. denied, 444 U.S. 881 (1979).  It is for the judge to determine whether the party seeking to exercise the peremptory strike has met its burden.

      [16]Even if sufficient data is theoretically available, practically speaking we cannot expect every defendant to have the resources to collect, analyze, and present existing data to make a prima facie case of discrimination.  See State v. Brown, 930 N.W.2d 840, 865 n.8 (Iowa 2018) (Cady, C.J., dissenting), quoting Jackson, Profiling the Police:  Flipping 20 Years of Whren on its Head, 85 UMKC L. Rev. 671, 680 (2017) ("On average, to take an equal protection claim to trial costs anywhere from $45,000 up to $125,000").

      [17]I agree that statistical evidence is an excellent form of evidence and that it can be a vital part of demonstrating that a stop was based on race, particularly when implicit bias is at play.  Nonetheless, there are lingering concerns about the availability and accessibility of necessary data. 

      In 2019 the Legislature enacted G. L. c. 90, § 63, which requires municipalities to report annually traffic stop data.  See St. 2019, c. 122, § 10.  I am not aware of any reports or data yet published pursuant to that statute.  The court notes that a bill currently under consideration "likely would enable defendants to access publicly available, department-wide data on the demographics of all traffic stops in the relevant municipality, and would provide a plethora of relevant data available to support (or weaken) equal protection claims."  Ante at    . It appears that neither the 2019 act nor the bill currently under consideration would guarantee defendants access to the officer-specific data that the defendant analyzed here.  Moreover, even assuming a showing of a department-wide, as opposed to an officer-specific, pattern of racial profiling would be a sufficient statistical showing, the shortcomings I raise regarding the one-time collection of similar data pursuant to St. 2000, c. 228, apply equally to the new data sets, which are also aggregated by municipality and therefore may obscure particular race-based stops or the actions of individual biased officers.  Given the experience after Lora, if past is prologue, we have no assurance that such information will be available in the future.

      [18]The experience with the Legislature's attempt to mandate data collection on racial profiling via St. 2000, c. 228, further illustrates the uncertainty of defendants' access to useful data for equal protection claims.  Pursuant to that act, in 2005 the Secretary of the Executive Office of Public Safety (Secretary) ordered the 249 police departments found by the Northeastern experts to have engaged in racial or gender profiling to collect additional data on all traffic stops for one year, including information on the identities of officers making such stops.  See St. 2000, c. 228, § 10.  See generally Boston Police Patrolmen's Ass'n, Inc. v. Police Dep't of Boston, 446 Mass. 46, 48-49 (2006).  In that case, the union for Boston police officers sought to enjoin the collection of officer identities in this second phase of data collection.  Id. at 47.  We held that, "[i]n order to fulfil the Act's objective of eliminating profiling by police officers," the Secretary could require the collection of data identifying the officer making the stop during this second phase.  See id. at 52-53.  However, by the time Lora was decided in 2008, "nearly one-half of the targeted police departments did not follow recommended guidelines and the State did not receive or review any data" collected pursuant to the second phase of data collection by departments found to have engaged in racial or gender profiling.  See Lora, 451 Mass. at 449 (Ireland, J., concurring).

      [19]Obviously, this approach would affect substantially more stops than those where, as here, the defendant alleges that the ulterior motive for the stop was racial discrimination.  As discussed, supra, art. 14's bedrock protection against unreasonable seizures applies to all pretextual stops, not only those based on race.  Just as we should not sanction as reasonable a traffic stop based explicitly on a driver's race, we also should not allow stops based on nothing more than, for example, the driver's hairstyle or the apparent expense of the car relative to the neighborhood in which it is traveling.  See Brown, 930 N.W.2d at 928 (Appel, J., dissenting), quoting United States v. Scopo, 19 F.3d 777, 786 (2d Cir.), cert. denied, 513 U.S. 877 (1994) (Newman, C.J., concurring).  Given the historical and ongoing connection between systemic law enforcement practices and racial discrimination, it is not at all surprising that an effective remedy for racial profiling entails a broader prohibition on the use of pretextual stops.

      [20]As I advocate for adding an option to proceed under art. 14 in addition to the court's solution, there is no need to compare which strategy is better -- I conclude that a defendant should have the option to choose either, or both.  The difference between my position and that of the court is not a disagreement regarding "the best legal analysis"; it is a question of which position is more comprehensive.

      I differ with the court, however, with regard to what is meant by a systemic solution.  The court is of course correct that improvements in data collection will further illustrate the existence of systemic racism in traffic stops.  But using new data to confirm the existence of an already undeniable systemic problem is not the same as changing the system of rules and practices that perpetuate racial discrimination in policing.  A focus on individual instances of race-based stops without addressing the rule that enables them cannot be considered a systemic approach, no matter how well-meaning it may be.

footnotes for concurring

      [1] Through a mobile computer, "[a]n officer inputs license plate numbers, and, within seconds, receives various information in reply, including the type of vehicle that is assigned to the plate, whether the vehicle is registered, whether the vehicle owner has an active license, and whether any warrants are outstanding."  See Commonwealth v. Muckle, 61 Mass. App. Ct. 678, 679 n.3 (2004).

      [2] It is well established that there is a "lesser expectation of privacy in a motor vehicle because its function is transportation and it seldom serves as one's residence or as the repository of personal effects."  Cardwell v. Lewis, 417 U.S. 583, 590 (1974).