Harvard Law Review Harvard Law Review Harvard Law Review

Harvard Law Review Forum

Maryland v. King Per Se Unreasonableness, the Golden Rule, and the Future of DNA Databases

Responding to Erin Murphy, License, Registration, Cheek Swab: DNA Testing and the Divided Court, 127 Harv. L. Rev. 161 (2013)

In License, Registration, Cheek Swab: DNA Testing and the Divided Court,1 Professor Erin Murphy deftly summarizes and situates the Supreme Court’s opinion in Maryland v. King.2 As she observes, the case can be read narrowly or broadly.3 Murphy reads the case broadly, suggesting that King is “a watershed moment”4 that portends “a new Fourth Amendment in town”5 and that its “reimagination of the idea of ‘identity’”6 “arguably invite[s] a new era of genetic identification.”7 Here, I offer a less dramatic view of the doctrinal significance of King and the limits of the majority’s identification theory. I also offer a precept for officials seeking to expand or improve DNA databases in this new era.

I. The Same Old, Same Old Conflicted Fourth Amendment Doctrine

In some twenty opinions, the Supreme Court has stated that the Fourth Amendment generally makes searches conducted without a warrant and probable cause unreasonable per se unless they fall within a categorical exception.8 Typically, the per-se-unreasonable-with-exceptions (PSUWE) rule suffices to invalidate warrantless searches with no further analysis of the totality of the circumstances. As shown in Figure 1, however, there are situations in which the Court balances individual and state interests to ascertain reasonableness.9

Figure 1: Situations in which Balancing Determines Fourth Amendment Reasonableness

When the Court does balance interests in Fourth Amendment cases, it almost always stays within the PSUWE framework. First, to decide whether to recognize a new categorical exception and how to define its boundaries,15 the Court must balance.16 Second, within the PSUWE framework, program-specific balancing occurs in the sprawling category of special-needs and administrative-search cases. The justification for balancing within this exception is that the State has special interests beyond the acquisition of evidence for criminal prosecutions. When evidence production is the sole objective, the balance already is conclusively presumed to favor a warrant and probable cause (or an applicable exception).17 When more is at stake — when, for example, the government randomly tests all weapons-wielding drug enforcement agents to ensure that they are not using drugs18 — this presumption is less applicable.

The balancing in King, however, does not fit into either of these two PSUWE categories — at least not as currently conceived. The special-needs exception requires that the “primary purpose” of the program be something other than acquiring evidence or contraband.19 As Murphy indicates (and every Justice realized), the primary purpose of Maryland’s law was evidence production. Neither was the Court fashioning a new exception for certain kinds of biometric data.20 Instead, the majority stepped outside the PSUWE framework, which it described as defeasible. This was not the first time the Court had done so,21 and it did not swing the door “wide open”22 — at least not wide open to a world in which every search will be judged for reasonableness on an ad hoc basis.23 The Court could have discarded the PSUWE framework entirely. Instead, it maintained that King resembled some special-needs balancing cases and then included King in the small set of cases in which it balances to ascertain reasonableness even though the primary purpose of the search is the production of evidence or investigative leads.24

To be sure, this set is not well defined. The majority’s elliptical description of the circumstances in which the per se rule gives way to direct balancing is reminiscent of Ludwig Wittgenstein’s theory of “family resemblances.”25 Like Wittgenstein, who deemed it unnecessary to articulate any common denominator for a generic term like “games,” Justice Kennedy, without articulating any essential factors, pointed to a family of searches eligible for direct balancing. The opinion suggests that the most recognizable family trait is that “the search involves no discretion that could properly be limited by the ‘interpo[lation of] a neutral magistrate between the citizen and the law enforcement officer.’”26 As with a policy that always requires inventory searches of arrestees,27 a magistrate has nothing to decide with regard to mandatory DNA sampling. This rationale for dispensing with warrants applies in various special-needs cases, like the inventory search (even though it is preceded by a discretionary decision to make an arrest). It also applies to the non-PSUWE case of Illinois v. Lidster,28 in which a highway checkpoint for solving a hit-and-run case applied to all drivers, rendering pointless a magistrate’s judgment of whether any given driver had done anything wrong or had pertinent knowledge. But the other two direct balancing cases, Samson and Knights, are members of the family that do not share this feature, leaving the boundaries of the non-PSUWE cases obscure.

Nonetheless, insofar as “a new Fourth Amendment in town”29 goes, the fact remains that the majority chose to step outside — but not to discard — the PSUWE framework. This is not an approach to totality balancing that advocates of the direct recourse to reasonableness will appreciate. It is not a declaration that “case-by-case assessments under a test as elastic as ‘reasonableness’”30 are the new normal. If I am right that this part of King and the “tendrils”31 in other cases decided this past Term are not quite so exceptional, it may not be necessary for casebook writers to start over.32 The King analysis does not dethrone the PSUWE framework or fundamentally destabilize that regime.

The case is doctrinally disappointing, I would argue, because it is a lost opportunity to devise a candid categorical exception for certain biometric data or to revisit the “primary purpose” limitation on the special needs exception.33 First, depending on the “scientific and statutory safeguards” noted in King,34 a categorical exception for noninvasive biometric data should include fingerprints and might also encompass DNA profiles.35 Second, as an alternative to creating a categorical exception, the Court could have reconsidered whether denying special-needs balancing to secondary-special-purpose searches is consistent with the rationale for engaging in balancing. The rationale, one would think, is that combining the standard law enforcement objective of evidence gathering with other functions makes inapposite the presumption that warrantless searches are per se unreasonable. This rationale applies to secondary-special-purpose searches just as it does to primary-purpose ones. As such, the King Court could have dropped the problematic primary-purpose limitation and addressed the Maryland program as a special-needs case within the PSUWE framework.

II. Defining Identity and the Future of DNA Databases

The other prong of Murphy’s penetrating paper that I would like to discuss is the concern with the use of the word “identification.” The majority’s undifferentiated use of the term incensed the dissent,36 and it led Murphy to insist that “almost none of the state or federal statutory regimes authorizing DNA collection” would prevent the government from testing DNA samples for a “violence gene” or “pedophile gene” and then “incarcerating people based on a probabilistic predisposition to violence or pedophilia.”37

In ordinary discourse, “identify” can have three meanings. The first is authenticating a person’s identity via some mark or token — a name, a password, or a biometric characteristic. Used in this manner and to provide a permanent record of an individual’s identity, DNA is similar to photographs, body measurements, and fingerprints. The second meaning is associating a known individual with a location, as when a witness to a robbery picks a photograph of an individual from a set of mugshots or when a trawl through a fingerprint database of known prints produces a hit to a latent print from a crime scene. Both these types of identification, which I will call authentication-identification and association-identification, respectively, are commonplace,38 historically accepted,39 and not deeply invasive of legitimate privacy interests.40 Third, there is identification in the sense of describing someone’s character, health, politics, thoughts, and so on. Jane Doe, one could say, has been identified as a communist sympathizer, as a carrier of a disease, or as prone to violence. In contrast to the use of a mark or token in authentication or association-identification, this trait-identification does not necessarily rely on a set of features that serve to distinguish one individual from all (or nearly all) other people, and it has much more serious implications for personal privacy. As such, why would one assume that a legislator’s vote for a computer searchable database of DNA profiles — or a Justice’s opinion upholding the database as constitutional for authentication-identification and association-identification — authorizes or endorses trait-identification?

Whereas the majority opinion blurs the line between the first two meanings of “identification,” Murphy elides all three when she asserts that laws adopted solely to establish and fund collection of DNA from convicted offenders (and later, arrestees) for authentication and crime-scene matching reach well beyond these contemplated uses. Her argument that these laws permit testing people for “the ‘pedophile’ gene” or “the ‘violence gene’”41 seems to be that the statutory term “identification” reaches to its farthest linguistic limits in the absence of an express limitation.42 This mode of interpretation overlooks the purpose, intent, and original understanding of the statutes that paved the way for DNA “identification.”43

The same unwarranted assumption infects Murphy’s analysis of the Court’s use of the word “identification” for Fourth Amendment purposes. In a remarkable sleight of hand, she transforms the Court’s explicit caution about using DNA samples in ways that would be especially invasive of privacy into an approbation of many such uses. She suggests that when “King says simply that, should police conduct testing for ‘predisposition for a particular disease or other hereditary factors not relevant to identity, that case would present additional privacy concerns not present here,’”44 the opinion actually is giving a green light to future genetic testing that will incarcerate people for having the wrong genes. But this truncated sentence45 follows several others that indicate that the Court is using “identity” and “identification” only to denote authentication-identification and association-identification via the CODIS loci. The Court’s point was that “law enforcement officers analyze DNA for the sole purpose of generating a unique identifying number against which future samples may be matched.”46 Testing for behavioral traits or other phenotypes would be “beyond identification.”47 In contrast, the balancing in King did not include the individual interests implicated by genetic tests for mental or physical traits. It therefore seems extravagant to maintain that by accepting the two token-of-identity uses of DNA, the Court implicitly approved of genetic testing for “family ties . . . [,] asocial behavior or addiction, and . . . violence.”48

Nevertheless, Murphy may be making a more subtle observation about the implications of the majority’s “identification” rationale. The majority reasoned that DNA profiling strictly for authentication-identification and association-identification advances important state interests in pretrial decisionmaking and supervision of arrestees. But why limit DNA analysis to the current, relatively innocuous loci if direct testing for putative violence genes or the like also would serve these interests? Why not allow such direct trait-identification? “After all, law enforcement needs to know just whom it is dealing with.”49

The answer lies in King’s balancing test. If the Court were faced with troubling trait-identification uses, it easily could distinguish King: “that case would present additional privacy concerns not present here.”50 Even indulging the very dubious assumption that some of this trait-identification by genes would supply valid predictive data for pretrial release decisions or for supervising prisoners, such trait-identification would implicate more robust claims of genetic privacy. When meaningful privacy interests are on the scale, the outcome of direct balancing is not so obvious.

Thus far, I have suggested that the resort to direct balancing in King was unnecessary and unfortunate, but, contrary to the dissenting justices, it did not cross a line that had never been crossed before. Although King adds to the size of the small family of direct balancing cases, these cases remain the exceptions to the normal PSUWE framework. As for the semantics of “identification,” I have maintained that King does not change the statutory meaning of the term, which refers to DNA profiles as tokens of identity that distinguish one individual from another, and that the majority’s balancing does not commit it to upholding all manner of alleged “identification” no matter how invasive of privacy it may be.

III. Beyond Doctrine and Words

Being a constitutional case, King does not address the question of whether involvement in the criminal justice system should be the triggering event. In that regard, it has been said that there is no principled distinction between taking DNA from arrestees and the rest of us51 and that a population-wide database would obviate or mitigate many of Murphy’s understandable concerns — disproportionate inclusion of minorities, pretextual arrests to acquire DNA, kinship trawling, unregulated local databases, and so on.

Although universality is too costly at present and might never be politically feasible, the idea has an immediate application — as a thought-experiment. Those who have called for serious reflection on a universal database tend to accept the premise that Murphy disparages — namely, that “DNA sampling [and profiling] is [not] that big of a deal.”52 I share this view, at least for a properly designed and administered database system (perhaps one that does not retain DNA samples). I hope this is not a “general nonchalance about government genetic testing,”53 but rather is the result of a more careful inquiry into the interests that DNA profiling realistically threatens.54 Still, I could be wrong. To the extent that well informed observers can differ in their risk assessments, a simple precept should guide the expansion of DNA databases. Legislators and database administrators should do unto others as they would to themselves. They should not adopt or operate any DNA identification system unless they would be willing to include their own DNA in it. It is too easy to approve of taking DNA from arrestees (or any group) when we presume that we will not be one of “them.”55 But if “we” truly conceive of ourselves as the recipients of this treatment, then we are more likely to arrive at a system with sufficient safeguards. And if the benefits of such a system are substantial, it could be foolish to foresake them.


1. Erin Murphy, The Supreme Court, 2012 Term — Comment: License, Registration, Cheek Swab: DNA Testing and the Divided Court, 127 Harv. L. Rev. 161 (2013).

2. 133 S. Ct. 1958 (2013).

3. Murphy, supra note 1, at 173.

4. Id. at 161.

5. Id. at 161.

6. Id. at 177.

7. Id. at 179.

8. For recent examples, see David H. Kaye, A Fourth Amendment Theory for Arrestee DNA and Other Biometric Databases, 15 U. Pa. J. Const. L. 1095, 1102—03, 1117—19 (2013). If Justice Scalia is now part of a “nascent alliance” of four justices who “believe in the warrant requirement,” Murphy, supra note 1, at 187, he has radically altered his views. It is more likely that his position in King stems from an affinity for bright-line rules and that, King notwithstanding, he has yet to foresake his position that there is no “general rule” requiring warrants — only a rule that they are required when the pre-Constitution common law required them and when “changes in the surrounding legal rules . . . make a warrant indispensable to reasonableness where it once was not.” California v. Acevedo, 500 U.S. 565, 583 (1991) (Scalia, J., concurring in the judgment).

9. The Figure comes from David H. Kaye, Why So Contrived? The Fourth Amendment and DNA Databases After Maryland v. King, 104 J. Crim. L. & Crimonology (forthcoming May 2014). It does not include cases on whether excessive force is used to conduct a search or seizure. E.g., Tennessee v. Garner, 471 U.S. 1 (1985) (balancing to conclude that shooting a fleeing suspect was unreasonable).

10. E.g., Terry v. Ohio, 392 U.S. 1 (1968).

11. E.g., Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602 (1989).

12. Samson v. California, 547 U.S. 843 (2006); United States v. Knights, 534 U.S. 112 (2001).

13. Illinois v. Lidster, 540 U.S. 419 (2004).

14. Maryland v. King, 133 S. Ct. 1958 (2013).

15. A boundary-defining case that attracts Murphy’s attention is Bailey v. United States, 133 S. Ct. 1031 (2013). Bailey concerns a per se rule for seizures of the person, which ordinarily require probable cause but not a warrant. One exception to this per se rule permits “officers executing a search warrant ‘to detain the occupants of the premises while a proper search is conducted,’” id. at 1037 (quoting Michigan v. Summers, 452 U.S. 692, 705 (1981)), even though the detention is “without probable cause to arrest for a crime,” id., and without “particular suspicion that an individual is involved in criminal activity or poses a specific danger to the officers,” id. at 1037—38. According to Murphy, “Justices Thomas, Breyer and Alito in dissent advocated a general balancing approach.” Murphy, supra note 1, at 186. Yet the dissent agreed “that the question involves drawing a line of demarcation granting a categorical form of detention authority.” Bailey, 133 S. Ct. at 1046 (Breyer, J., dissenting). The dissenting Justices did not challenge the statement of three other Justices that “the ‘general rule’ is ‘that Fourth Amendment seizures are “reasonable” only if based on probable cause.’” Id. at 1044 (Scalia, J., concurring) (quoting Dunaway v. New York, 442 U.S. 200, 213 (1979)). Neither did they doubt the majority’s nearly identical observation. Id. at 1037 (majority opinion). The disagreement was over the question of how broadly to define the categorical exception for detention incident to a search. The majority wanted a seemingly bright-line rule: the police can detain a person only within “the immediate vicinity of the premises to be searched.” Id. at 1041. The dissent wanted the exception to extend beyond the immediate vicinity to encompass persons “in the process of leaving the premises,” id. at 1049 (Breyer, J., dissenting) (emphasis omitted) (quoting United States v. Bailey, 652 F.3d 197, 206 (2d Cir. 2011)), but detained “as soon as ‘reasonably practicable,’” id. at 1046. The dissent considered the expanded category to be “based on realistic considerations” related to the underlying reasons for having the exception and thus superior to the majority’s categorical rule of “indeterminate geography.” Id. at 1049. However, none of the opinions in Bailey evinces a movement to across-the-board ad hoc balancing for seizures of the person incident to the execution of a search warrant, let alone for searches of people or places.

16. See Kaye, supra note 8, at 1103; David H. Kaye, On the “Considered Analysis” of Collecting DNA Before Conviction, 60 UCLA L. Rev. Discourse 104, 113—20 (2013); Murphy, supra note 1, at 186.

17. See Kaye, supra note 8, at 1112.

18. Nat’l Treasury Emps. Union v. Von Raab, 489 U.S. 656 (1989).

19. The Court has found searches to have been unreasonable because of this limitation on the special-needs exception in only two cases. See Ferguson v. City of Charleston, 532 U.S. 67, 81 (2001); City of Indianapolis v. Edmond, 531 U.S. 32, 40—42 (2000).

20. E.g., Kaye, supra note 8 (recommending an exception for collecting and using biometric data under certain conditions). The State noted this possibility in its petition for certiorari, but the parties studiously ignored it after that.

21. The dissent conveniently acted as if, contrary to Figure 1, there never had been a case that resorted to balancing to uphold a search or seizure primarily intended to produce evidence for investigation or prosecution of a crime.

22. Murphy, supra note 1, at 184.

23. In Bailey v. United States, the case that Murphy sees as “strikingly” evincing the movement to overturn the PSUWE framework, id. at 186, Justice Kennedy — and every other Justice — agreed that direct balancing is not generally available. See supra note 15.

24. As shown in Figure 1, those cases are Samson, Knights, Lidster, and now King.

25. Ludwig Wittgenstein, Philosophical Investigations 32 (P. M. S. Hacker & Joachim Schulte eds., G. E. M. Anscombe trans., Wiley-Blackwell rev. 4th ed. 2009) (1953).

26. Maryland v. King, 133 S. Ct. 1958, 1969 (2013) (alteration in original) (quoting Nat’l Treasury Emps. Union v. Von Raab, 489 U.S. 656, 667 (1989)). When this condition exists, “[t]he need for a warrant is perhaps least.” Id.

27. Illinois v. Lafayette, 462 U.S. 640, 647 (1983) (standardized inventory procedures are appropriate).

28. 540 U.S. 419 (2004).

29. Murphy, supra note 1, at 163.

30. Id. at 186.

31. Id. at 185.

32. Id. at 187 (“[I]f freestanding interest balancing is the new touchstone of Fourth Amendment inquiry, then casebooks and class discussions deserve overhaul.”). This is not to say that the questions Murphy then raises do not belong in casebooks and class discussion. Quite the contrary.

33. Another disappointing aspect of the King Court’s analysis of routine DNA collection before conviction is the Court’s failure to consider the dominant function of arrestee DNA sampling as a possibly weighty consideration in its own right. See Kaye, supra note 9 (explaining how this departs from a true totality-of-the-circumstances inquiry and imbues the opinion with an air of unreality).

34. Maryland v. King, 133 S. Ct. 1958, 1980 (2013).

35. Kaye, supra note 8. Whether a particular program for collecting and using DNA from arrestees before any conviction falls within such an exception or is sustainable under a modified special-needs balancing is a matter on which I am mildly agnostic. Much depends on the privacy safeguards in the system and the marginal value of the information in criminal investigations. Although I have long maintained that the value is limited, e.g., David H. Kaye, The Constitutionality of DNA Sampling on Arrest, 10 Cornell J.L. & Pub. Pol’y 455 (2001), I am not prepared to say that it is a complete myth. In any event, that is an empirical question of more complexity than can be discussed here.

36. King, 133 S. Ct. at 1985 (Scalia, J., dissenting) (excoriating the majority opinion as falling below the level expected of a “minimally competent speaker of English”).

37. Murphy, supra note 1, at 180.

38. Professor Murphy knows “intuitively” that fingerprint databases are not routinely searched for matches to latent prints. Murphy, supra note 1, at 163. Only they are, every day. See, e.g., Kaye, supra note 8, at 1099, 1131 (noting 50,000 “hits” in 2005); Kathryn Wexler, Pairing Prints Rarely Key to Solving Crime, St. Petersburg Times (Dec. 19, 1999), http://www.sptimes.com/News/121999/news_pf/TampaBay/Pairing_prints_rarely.shtml (“Most of the 2,500 cases a year [Florida Department of Law Enforcement] experts run through AFIS are ‘cold’ cases in which there are no leads other than the print. About 25 percent of those eventually are matched with suspects whose prints are in the database because of arrests for unrelated crimes.”). These fingerprint “cold hits,” like DNA hits, sometimes stretch back to crimes committed decades ago. See, e.g., State v. Watson, 827 N.W.2d 507 (Neb. 2013); Latent Hit of the Year 2012, Fed. Bureau Investigation, http://www.fbi.gov/news/videos/latent-hit-of-the-year-2012 (last visited Nov. 4, 2013); FBI’S Integrated Automated Fingerprint Identification System (IAFIS) Helps Solve 38 Year-Old Cold Case, Chesapeake Examiner, Spring 2011, at 14, available at http://www.cbdiai.org/Articles/anon_sp11.pdf.

39. The earliest case identifying the source of a latent print by perusing a collection of known prints occurred in 1902. Kaye, supra note 8, at 1121.

40. See id. at 1134—39; David H. Kaye, The Genealogy Detectives: A Constitutional Analysis of “Familial Searching,” 50 Am. Crim. L. Rev. 109 (2013).

41. Murphy, supra note 1, at 180.

42. Erin Murphy, A Tale of Two Sciences, 110 Mich. L. Rev. 909, 915 (2012) (book review) (reasoning that “identification purpose” could include “behavioral genetics experimentation” as long as “nothing in the law as written . . . clearly proscribes it”).

43. The sponsors and supporters of these laws, after all, sold them as association-identification systems analogous to fingerprint databases. To infer the intent of earlier state and federal legislatures from the outer boundaries of what five Justices, writing decades later, might have found to be constitutional is unprecedented and unpersuasive. Cf. United States v. Price, 361 U.S. 304, 313 (1960) (“[T]he views of a subsequent Congress form a hazardous basis for inferring the intent of an earlier one.”).

44. Murphy, supra note 1, at 181.

45. The full sentence is: “If in the future police analyze samples to determine, for instance, an arrestee’s predisposition for a particular disease or other hereditary factors not relevant to identity, that case would present additional privacy concerns not present here.” Maryland v. King, 133 S. Ct. 1958, 1979 (2013). If “for instance” applies to the entire appositive phrase that follows it, then trait-identification that is irrelevant to pretrial decisionmaking is just one example of a case that could result in a different outcome. Other forms of trait-identification also could constitute unreasonable searches even though they might have some slight value in pretrial decisionmaking.

46. Id.

47. Quoting an article in the forensic science literature, the Court emphasized that the identifying numbers “are not at present revealing information beyond identification.” Id. (quoting Sara H. Katsanis & Jennifer K. Wagner, Technical Note, Characterization of the Standard and Recommended CODIS Markers, 58 J. Forensic Sci. S169, S171 (2013)) (internal quotation marks omitted). The phrase “beyond identification” refers to inferences about phenotypes (observable traits) arising from either a causal mechanism or linkage disequilibrium. See Katsanis & Wagner, supra.

48. Murphy, supra note 1, at 180.

49. Id.

50. King, 133 S. Ct. at 1979 (emphasis added).

51. Richard Lempert, Maryland v. King: An Unfortunate Supreme Court Decision on the Collection of DNA Samples, Brookings Up Front (June 6, 2013, 11:38 AM), http://www.brookings.edu/blogs/up-front/posts/2013/06/06-maryland-king-supreme-court-dna-samples-lempert.

52. Murphy, supra note 1, at 175.

53. Id.

54. See Kaye, supra note 8; Kaye, supra note 40.

55. Cf. Lempert, supra note 51 (“We should judge the result in King not with Alonzo King in mind but with ourselves, our friends and neighbors standing in his place.”).