508 F.3d 941
9th Cir.
2007

Lead Opinion

McKEOWN, Circuit Judge:

In 2004 we held that the DNA Analysis Backlog Elimination Act of 2000 “satisfies the requirements of the Fourth Amendment” with respect to individuals on supervised release. United States v. Kincade, 379 F.3d 813, 839 (9th Cir.2004) (en banc). The 2000 Act required collection of DNA samples from individuals in custody and on probation, parole, or supervised release who had . been convicted of “qualifying Federal offenses,” then defined as certain violent crimes. 42 U.S.C. § 14135a (2000). Congress amended the Act in 2004 to expand the qualifying offenses to all felonies. Joining every other circuit to consider the 2004 Act, we hold that the amended statute passes constitutional muster with respect to a convicted felon on supervised release.1

1. Statutory and Regulatory Background

In 2000, Congress enacted the DNA Analysis Backlog Elimination Act (the “DNA Act” or the “Act”), which required DNA samples to be collected from individuals in custody and those on probation, parole, or supervised release after being convicted of “qualifying Federal offenses.” 42 U.S.C. § 14135a. The DNA Act originally defined “qualifying Federal offenses” as the following: (A) murder, voluntary manslaughter, or other offense relating to homicide, (B) an offense relating to sexual abuse, to sexual exploitation or other abuse of children, or to transportation for illegal sexual activity, (C) an offense relating to peonage and slavery, (D) kidnaping, (E) an offense involving robbery or burglary, (F) any violation of 18 U.S.C. § 1153 involving murder, manslaughter, kidnap-ing, maiming, a felony offense relating to sexual abuse, incest, arson, burglary, or robbery, (G) any attempt or conspiracy to commit any of the above offenses. See DNA Analysis Backlog Elimination Act, Pub.L. No. 106-546, § 3, 114 Stat. 2726, 2729-30 (2000). In 2001, the USA PATRIOT Act added to § 14135a “[a]ny offense listed in section 2232b(g)(5)(B) of Title 18 [acts of terrorism transcending national boundaries],” “[a]ny crime of violence (as defined in section 16 of Title 18, United States Code),” and “[a]ny attempt or conspiracy to commit any of the above of~ fenses” to the list of qualifying offenses. See Pub.L. No. 107-56, § 503, 115 Stat. 272, 364 (2001). Together, these qualifying offenses are generally characterized as violent crimes.

Congress passed the Justice for All Act in 2004, which further amended the DNA Act by expanding the definition of “qualifying Federal offenses” as follows:

(d) Qualifying Federal offenses
The offenses that shall be treated for purposes of this section as qualifying Federal offenses are the following offenses, as determined by the Attorney General:
(1) Any felony.
(2) Any offense under chapter 109A of Title 18 [sexual abuse crimes].
(3) Any crime of violence (as that term is defined in section 16 of Title 18).2
(4) Any attempt or conspiracy to commit any of the offenses in paragraphs (1) through (3).

Pub.L. No. 108-405, § 203(b), 118 Stat. 2260, 2270 (2004).3 Rather than specifying certain crimes, the amendment included all felonies, all crimes of violence, and all sexual abuse crimes under Chapter 109A of Title 18.

The Attorney General has authority to promulgate regulations to carry out the statute. See 42 U.S.C. § 14135a(e); 28 C.F.R. § 28.2; DNA Sample Collection From Federal Offenders Under the Justice for All Act of 2004, 70 Fed.Reg. 4,763-01 (Jan. 31, 2005) (“DNA Sample Collection”). In response to the 2004 changes, the Attorney General revised 28 C.F.R. § 28.2, the regulation that identifies qualifying federal offenses for the purposes of DNA sample collection, to track the new language of § 14135a(d).

As under the original DNA Act, probation offices collect DNA samples from individuals on probation, parole, or supervised release who have been convicted of a qualifying federal offense, 42 U.S.C. § 14135a(a)(2), and the samples are furnished to the Director of the Federal Bureau of Investigation (the “FBI”), “who ... carriles] out a DNA analysis on each such DNA sample and include[s] the results in CODIS,” id. § 14135a(b).4 CODIS is the FBI’s Combined DNA Index System — a centrally-managed database linking DNA profiles culled from federal, state, and territorial DNA collection programs, as well as profiles drawn from crime-scene evidence, unidentified remains, and genetic samples voluntarily provided by relatives of missing persons. See Kincade, 379 F.3d at 819 (plurality opinion).

The 2000 Act also provided privacy protection standards, which remain in place after the 2004 amendment. Each act of unauthorized collection, use, or disclosure of a DNA sample is a separate crime, and “[a] person who knowingly discloses a sample or result described in subsection (a) in any manner to any person not authorized to receive it, or obtains or uses, without authorization, such sample or result, shall be fined not more than $250,000, or imprisoned for a period of not more than one year.” 42 U.S.C. § 14135e(c). Subsection (a) provides that in general, “any sample collected under, or any result of any analysis carried out under, section 14135, 14135a, or 14135b of this title may be used only for a purpose specified in such section.” Id. § 14135e(a).

II. Kriesel’s Claims

In March 1999, Thomas Edward Kriesel, Jr. pleaded guilty to one count of conspiracy to commit the crime of possession with intent to distribute methamphetamine in violation of 21 U.S.C. §§ 841(a)(1) and 846. He was sentenced to thirty months of imprisonment and three years of supervised release. At the time of judgment, the terms of Kriesel’s supervised release included this standard condition: “You shall answer truthfully all inquiries by the probation officer and follow the instructions of the probation officer.” Kriesel was also advised that he “shall submit his person, residence, place of employment or vehicle to a search upon request by the U.S. Probation Office.”

When he was initially scheduled for DNA testing, Kriesel informed the probation officer that he was opposed in principle to the government’s collection and permanent storage of his DNA. In August 2005, the Probation Department petitioned the district court to revoke Kriesel’s supervision because he failed to report for DNA testing. Because Kriesel’s conviction for conspiracy to distribute methamphetamine is a felony, it is a “qualifying Federal offense” under the DNA Backlog Elimination Act as amended in 2004. 42 U.S.C. § 14135a(d) (2004). At the hearing on the petition to revoke supervised release, Kriesel’s counsel argued that the Attorney General promulgated the regulation gov erning DNA collection in violation of the notice and comment provisions of the Administrative Procedure Act (“APA”), 5 U.S.C. § 553, and that the DNA Act as amended in 2004 violated the Fourth Amendment’s prohibition on unreasonable searches and seizures. See U.S. Const., amend. IV.

The district court rejected these arguments and upheld both the validity of the regulation and the constitutionality of the Act. The district court also granted a stay of its order pending appeal.

III. APA Challenge

Kriesel first contends that the Attorney General was required to follow the notice and comment procedures in 5 U.S.C. § 553 because in revising 28 C.F.R. § 28.2, the Attorney General promulgated a substantive or legislative rule. The APA provides that administrative rules must be adopted through the rulemaking process, which includes notice and an opportunity for public comment. 5 U.S.C. §§ 551(4), (5); 553. Although not referenced in the statute, the courts have denominated such rules as “legislative rules.” See Richard J. Pierce, Jr., Distinguishing Legislative Rules from, Interpretive Rules, 52 Admin. L. Rev. 547, 549 (2000). In contrast, the APA specifically exempts “interpretive rules” from the rulemaking process. 5 U.S.C. § 553(b)(3)(A).

The Attorney General’s regulation issued in response to the 2004 amendment is a classic interpretive rule: it is a rule “issued by an agency to advise the public of the agency’s construction of the statutes and rules which it administers.” Chrysler Corp. v. Brown, 441 U.S. 281, 302 n. 31, 99 S.Ct. 1705, 60 L.Ed.2d 208 (1979) (quoting United States Department of Justice, Attorney General’s Manual on the Administrative Procedure Act (1947)); see also Hemp Indus. Ass’n v. Drug Enforcement Administration, 333 F.3d 1082, 1087 (9th Cir.2003) (“In general terms, interpretive rules merely explain, but do not add to, the substantive law that already exists in the form of a statute or legislative rule.”).

The 2004 amendment dictates the basis for the regulatory revision. The statute provides in relevant part: “[t]he offenses that shall be treated for purposes of this section as qualifying Federal offenses are the following offenses, as determined by the Attorney General: (1) Any felony.” Justice for All Act, § 203(b).

The Attorney General issued regulations that mirror the statute, by designating “any' felony” as a qualifying offense. 28 C.F.R. § 28.2(b)(1). The regulations define “felony” in accord with federal law as an “offense that would be classified as a felony under 18 U.S.C. § 3559(a) or that is specifically classified by a letter grade as a felony.” Id. § 28.2(a).

Kriesel’s contention, which is not easy to divine, is apparently that by designating “all” felonies, rather than some, the Attorney General was legislating rather than interpreting the statute. This argument is difficult to countenance as the Attorney General simply defined felony in accord with an existing federal standard and adopted the “any felony” designation directly from the statutory language. See H.R.Rep. No. 108-711 (2004), as reprinted in 2005 U.S.C.C.A.N. 2274, 2284, 2004 WL 2348416 (stating that the 2004 amendment meant to authorize collection of DNA samples from “all felons convicted of Federal crimes and qualifying military offenses.”). In promulgating the regulation, the Attorney General specifically recognized that “[t]he notion of a ‘felony’ is a standard, familiar concept in Federal criminal law, and this rule simply refers to existing statutory provisions for its definition,” and that § 28.2 simply “defines ‘felony’ as it is ordinarily understood — i.e., as referring to offenses for which the maximum authorized term of imprisonment exceeds one year.” DNA Sample Collection, 70 Fed. Reg. 4,764, 4,766 (citing 18 U.S.C. § 3559(a)).

As the Attorney General explained, he understood the 2004 amendment itself to “authorize[ ] DNA sample collection from all Federal offenders convicted of felonies.” DNA Sample Collection, 70 Fed.Reg. 4,766. We earlier observed that “penalizing the agency” for explaining the bad news about the DNA Act, “by labeling the explanation ‘substantive,’ would be killing the messenger. The regulation impose[s] no other substantive legal duties ... other than what the statute already imposed.”5 Alcaraz v. Block, 746 F.2d 593, 613-14 (9th Cir.1984). Nothing in these regulations supports a claim that the rules are legislative and thus merit a full-blown rulemak-ing process.

IV. Constitutional Challenge

Every circuit to' consider a Fourth Amendment challenge to the 2004 Act has reached the same conclusion: collecting DNA from nonviolent felons as authorized by the Act does not violate the Fourth Amendment.6 The majority of circuits adopt a “totality of the circumstances” framework. United States v. Weikert, 504 F.3d 1, 9 (1st Cir.2007); Banks v. United States, 490 F.3d 1178, 1183 (10th Cir.2007); United States v. Kraklio, 451 F.3d 922, 924 (8th Cir.2006); United States v. Castillo-Lagos, 147 Fed.Appx. 71 (11th Cir.2005).7 In contrast, the Second and Seventh Circuits rely on the “special needs test.” United States v. Amerson, 483 F.3d 73, 78 (2d Cir.2007); United States v. Hook, 471 F.3d 766, 772-74 (7th Cir.2006). The Sixth Circuit has upheld the 2004 Act under both tests. United States v. Conley, 453 F.3d 674, 677-81 (6th Cir.2006).8

In light of Samson v. California, we continue to ground our analysis in the totality of circumstances test. 547 U.S. 843, 126 S.Ct. 2193, 165 L.Ed.2d 250 (2006). Before Samson, no Supreme Court case had addressed this issue. See Kincade, 379 F.3d at 832 n. 27 (plurality opinion). In Samson, however, the Court applied the totality of the circumstances test in upholding a California law provid ing that, as a condition for release, every prisoner eligible for state parole must agree to be subject to a search or seizure by a parole officer with or without a warrant, and with or without cause. 126 S.Ct. at 2197, 2199 n. 3, 2202.

Taking our cue from Samson, we reaffirm that “the touchstone of the Fourth Amendment is reasonableness,” id. at 2201 n. 4, and adopt the “general Fourth Amendment approach,” which “ex-amin[es] the totality of the circumstances to determine whether a search is reasonable.” Id. at 2197 (quoting United States v. Knights, 534 U.S. 112, 118, 122 S.Ct. 587, 151 L.Ed.2d 497 (2001)) (internal quotation marks omitted). “Whether a search is reasonable ‘is determined by assessing, on the one hand, the degree to which it intrudes upon an individual’s privacy, and on the other, the degree to which it is needed for the promotion of legitimate governmental interests.’ ” Id. (quoting Knights, 534 U.S. at 118-19, 122 S.Ct. 587).9

A. Kriesel’s Privacy Interest

In assessing the nature of Kriesel’s privacy interest, “we begin our resolution of the issue by taking note of the well-established principle that parolees and other conditional releasees are not entitled to the full panoply of rights and protections possessed by the general public.” Kincade, 379 F.3d at 833 (plurality opinion). Rather, the Supreme Court has often recognized, as it did yet again in Samson, that parolees “have severely diminished expectations of privacy,” and that the Fourth Amendment does not necessarily prohibit a police officer from conducting a suspicionless search of a parolee. 126 S.Ct. at 2199, 2202; see also Kincade, 379 F.3d at 834 (plurality opinion) (“[C]ondi-tional releasees enjoy severely constricted expectations of privacy relative to the general citizenry.”)

As a direct consequence of Kriesel’s status as a supervised releasee, he has a diminished expectation of privacy in his own identity specifically, and tracking his identity is the primary consequence of DNA collection. The DNA analyzed by the FBI consists primarily of “junk DNA” — “non-genic stretches of DNA not presently recognized as being responsible for trait coding” that were purposefully selected because “they are not associated with any known physical or medical characteristics.” Kincade, 379 F.3d at 818 (plurality opinion) (internal quotation marks omitted). But see id. n. 6 (“Recent studies have begun to question the notion that junk DNA does not contain useful genetic programming material.”). And, based on Kriesel’s status as a qualified offender on supervised release, he can claim only the most limited expectation of privacy, if any, in his identity given that he was lawfully convicted of a predicate offense. See id. at 842 n. 3 (Gould, J., concurring) (noting that it is permissible to maintain identifying fingerprints of felons even after they have been released because fingerprints reveal only identity, but declining to endorse the practice of retaining DNA in the CODIS database after a felon has paid his debt to society).

In assessing the nature of the privacy intrusion, we are mindful of the caution that DNA often reveals more than identity, and that with advances in technology, junk DNA may reveal far more extensive genetic information. Judge Gould observed in his concurrence in Kineade, “un like fingerprints, DNA stores and reveals massive amounts of personal, private data about that individual, and the advance of science promises to make stored DNA only-more revealing over time. Like DNA, a fingerprint identifies a person, but unlike DNA, a fingerprint says nothing about the person’s health, their propensity for particular disease, their race and gender characteristics, and perhaps even their propensity for certain conduct.” Id. n. 3; see also Amerson, 483 F.3d at 85 (recognizing “the vast amount of sensitive information that can be mined from a person’s DNA and the very strong privacy interests that all individuals have in this information”) (citing Kincade, 379 F.3d at 843 (Reinhardt, J., dissenting)).

The concerns about DNA samples being used beyond identification purposes are real and legitimate. Nevertheless, those concerns are mitigated by the Act’s privacy protections, which provide criminal penalties for the unauthorized use of DNA samples. They are also outweighed by the competing notion that supervised releasees have little to no privacy interest in their identities. See 42 U.S.C. § 14135e(c) (“A person who knowingly discloses a[DNA] sample or result ... in any manner to any person not authorized to receive it, or obtains or uses, without authorization, such sample or result, shall be fined not more than $250,000, or imprisoned for a period of not more than one year. Each instance of disclosure, obtaining, or use shall constitute a separate offense under this subsection.”).10

The physical drawing of blood also implicates Kriesel’s interest in bodily integrity, “a cherished value of our society.” Schmerber v. California, 384 U.S. 757, 772, 86 S.Ct. 1826, 16 L.Ed.2d 908 (1966). A blood draw “is inherently more intrusive than a purely external search such as fingerprinting.” See Weikert, 504 F.3d at 12. Nevertheless, the Supreme Court has held that the intrusion caused by a blood test itself “is not significant, since such ‘tests are a commonplace in these days of periodic physical examinations and experience with them teaches that the quantity of blood extracted is minimal, and that for most people, the procedure involves virtually no risk, trauma, or pain.’ ” Skinner v. Ry. Labor Executives’ Ass’n, 489 U.S. 602, 625, 109 S.Ct. 1402, 103 L.Ed.2d 639 (1989) (quoting Schmerber, 384 U.S. at 771, 86 S.Ct. 1826). “[I]t is [also] a commonly accepted condition of probation [or supervised release] that the probationer [or re-leasee] submit to drug tests, with the result that, often, the requisite DNA sample can be collected without any incremental intrusion.” Amerson, 483 F.3d at 85. Consequently, the additional privacy implications of a blood test collecting DNA, as opposed to a cheek swab or other mechanism, do not significantly alter our analysis. Cf. id. at 84-85 (distinguishing between the physical taking of DNA samples and the “more serious invasion” of “analysis and maintenance” of an individual’s information). Based on all these considerations, we conclude that as a convicted felon who currently continues to serve his term of supervised release, Kriesel has a diminished privacy interest in the collection of his DNA for identification purposes.

We emphasize that our ruling today does not cover DNA collection from arres-tees or non-citizens detained in the custody of the United States, who are required to submit to DNA collection by the 2006 ver sion of the DNA Act. See 42 U.S.C. § 14135a(a)(1)(A) (2006). Nor do we have before us “a petitioner who has fully paid his or her debt to society, who has completely served his or her term, and who has left the penal system.... Once those previously on supervised release have wholly cleared their debt to society, the question may be raised, ‘Should the CO-DIS entry be erased?’ ” Kincade, 379 F.3d at 841 (Gould, J., concurring in the judgment). We do not answer these questions. See also Green, 354 F.3d at 679-81 (Easterbrook, J., concurring) (noting that “[fjelons whose terms have expired” form a different category of individuals than supervised releasees for the purposes of a Fourth Amendment inquiry). Rather, our decision is confined to the precise circumstances before us.

B. Government’s Interests

In Kincade, the plurality identified as “undeniably compelling” and “monumental” three governmental interests justifying DNA collection from violent felons: First, “[b]y establishing a means of identification that can be used to link conditional releasees to crimes committed while they are at large,” compulsory DNA profiling serves society’s interest in ensuring that releasees comply with the conditions of their release. 379 F.3d at 838 (plurality opinion). Second, the deterrent effect of such profiling fosters society’s interest in reducing recidivism. Id. Finally, collecting the DNA of offenders contributes to the solution of past crimes. Id. In elaborating on those interests, the analysis did not distinguish as a practical matter between violent and nonviolent felons. Here the government advances the same arguments with respect to nonviolent felons, and Kin-cade’s rationale applies with equal force.

The governmental interest in identifying releasees and linking them to crimes committed while “at large” is significant. While DNA evidence is often central to the investigations of violent crimes such as murder or sexual assault, see Roe v. Marcotte, 193 F.3d 72, 82 (2d Cir.1999), it can be useful in solving nonviolent crimes as well. As the Tenth Circuit recently noted, “[i]t is important to realize ... that DNA can be extracted from hair, saliva, and numerous other parts of our bodies that even a non-violent criminal could leave behind on a piece of inculpatory evidence.” Banks, 490 F.3d at 1190; see also Kincade, 379 F.3d at 838 n. 37; Amerson, 483 F.3d at 88 n. 15 (“[Tjhere are also indications that DNA can be, and is increasingly, being used to solve nonviolent crimes.”). Although fingerprint evidence might often be sufficient to identify a past offender, DNA collection provides another means for the government to meet its significant need to identify offenders who continue to serve a term of supervised release. Cf. Banks, 490 F.3d at 1190.

With respect to the deterrent effect, “[t]he Supreme Court has repeatedly recognized that rates of re-arrest among parolees and probationers are significantly higher than the general crime rate.” Banks, 490 F.3d at 1189; see also Samson, 126 S.Ct. at 2200 (collecting cases describing the state’s interests in reducing recidivism and noting that “parolees ... are more likely to commit future criminal offenses”) (internal quotation marks and citation omitted). “[T]he very assumption of the institution of probation” or supervised release is that the probationer is “more likely ... to violate the law.” Griffin v. Wisconsin, 483 U.S. 868, 880, 107 S.Ct. 3164, 97 L.Ed.2d 709 (1987).

Kriesel counters that nonviolent offenders have lower recidivism rates than violent offenders. This argument is not persuasive in Kriesel’s case. Indeed, he is already a recidivist, as he violated the terms of his release when he tested positive for controlled substances. While it is true that “recidivism rates vary with factors like the offender’s age and type of conviction,” nonetheless, “the high likelihood that non-violent offenders will re-offend — and therefore, as the Supreme Court recognized in Knights, be put in the position to conceal their crimes and identities from the authorities — underscores the Government’s interest in obtaining the most accurate identification information it can from these individuals.” See Banks, 490 F.3d at 1191.11

Finally, the Kincade plurality also explained that “by contributing to the solution of past crimes, DNA profiling of qualified federal offenders helps bring closure to countless victims of crime who long have languished in the knowledge that perpetrators remain at large.” 379 F.3d at 839. Although the weight of this rationale may vary when considering individual types of crimes, nonviolent crimes can also cause significant and lasting damage to innocent individuals. For example, a family defrauded of its life savings, home, and financial stability by an unscrupulous con artist running a telemarketing or mail fraud scheme is just as much a victim as a family that is the victim of a burglary, and just as DNA samples can clear suspects of crimes, they can also contribute to the resolution of past crimes. The government may certainly credit those concerns in pursuing ways to solve open investigations.

In sum, we agree in principle with the other circuits that have considered the issue, and hold that in the case before us, requiring Kriesel to comply with the 2004 amendment to the DNA Act is constitutional because the government’s significant interests in identifying supervised releas-ees, preventing recidivism, and solving past crimes outweigh the diminished privacy interests that may be advanced by a convicted felon currently serving a term of supervised release.

AFFIRMED.

Notes

. In so doing, we acknowledge that in Kin-cade and elsewhere, much ink has been spilled over this sensitive and contentious issue, and emphasize that we confine our discussion to resolving the constitutionality of the 2004 amendment, as applied to individuals like Kriesel. Cf. 379 F.3d at 837-38.

. 18 U.S.C. § 16 defines “crime of violence” as “(a) an offense that has as an element the use, attempted use, or threatened use of physical force against the person or property of another, or (b) any other offense that is a felony and that, by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.” All crimes of violence, whether felonies or misdemeanors, are covered by the statute and implementing regulation.

.The Act has since been amended to authorize DNA collection "from individuals who are arrested, facing charges, or convicted [of qualifying felonies,] or from non-United States persons who are detained under the authority of the United States.” See Violence Against Women and Department of Justice Reauthorization Act of 2005, Pub.L. No. 109-162, § 1004(a), 119 Stat. 2960, 3085 (2006). We do not address the constitutionality of the 2006 version of the Act, which greatly expanded the applicability of the statute beyond convicted individuals.

. Kincade provides a fuller description of the mechanics of the DNA Act and CODIS. 379 F.3d at 817-20 (plurality opinion).

. By contrast, the Attorney General specifically acknowledged his discretion in “mak[mg] judgments in determining which particular offenses constitute 'crimes of violence' as defined in 18 U.S.C. [§ ] 16 — but these judgments were already made, following public notice and the receipt of comments, in the version of 28 C.F.R. [§ ] 28.2 that was published on December 29, 2003, and went into effect on January 28, 2004 [68 Fed.Reg. 74,-855]. The revised regulation does not change these determinations.” DNA Sample Collection, 70 Fed.Reg. 4,766.

. Whether a search is unreasonable under the Fourth Amendment is a question of law reviewed de novo. United States v. Stafford, 416 F.3d 1068, 1073 (9th Cir.2005). "The compulsory extraction of blood for DNA profiling unquestionably implicates the right to personal security embodied in the Fourth Amendment, and thus constitutes a 'search' within the meaning of the Constitution.” Kincade, 379 F.3d at 821 n. 15 (plurality opinion).

. Prior to 2007, the Eleventh Circuit allowed citation to unpublished dispositions as persuasive authority. See 11th Cir. R. 36-2, Robert Timothy Reagan, Federal Judicial Center, Citing Unpublished Federal Appellate Opinions Issued Before 2007 (2007), http://www. uscourts. gov/rules/Unpub_Opinions .pdf.

. The federal and state courts have also upheld a variety of other statutes authorizing DNA collection from all convicted felons. See, e.g., Green v. Berge, 354 F.3d 675 (7th Cir.2004) (upholding Wisc. Stat. Ann. § 165.77 (West 1999)); Jones v. Murray, 962 F.2d 302 (4th Cir.1992) (upholding Va.Code Ann. § 19.2-310.2 (1990)); Doles v. State, 994 P.2d 315 (Wyo.1999) (upholding Wyo. Stat. Ann. § 7-19-403 (1997)).

. For Fourth Amendment purposes, our cases do not distinguish among parolees, probationers, and those on supervised release. Kin-cade, 379 F.3d at 817 n. 2 (citing United States v. Harper, 928 F.2d 894, 896 n. 1 (9th Cir.1991)).

. "Should the uses to which 'junk DNA’ can be put be shown in the future to be significantly greater than the record before us today suggests, a reconsideration of the reasonableness balance struck would be necessary,” even with respect to individuals in Kriesel's exact position. Amerson, 483 F.3d at 85 n. 13.

. To be sure, the rate of recidivism for nonviolent offenders is less than for violent offenders. Nonetheless, the rate remains significant. For example, the recidivism rate for drug offenders is 21.2%. U.S. Sentencing Commission, Measuring Recidivism: The Criminal History and Computation of the Federal Sentencing Guidelines 30 (May 2004), available at http://www.ussc.gov/research. htm.



Dissenting Opinion

BETTY B. FLETCHER, Circuit Judge,

dissenting:

The majority holds, with an air of shrugging inevitability, that without a warrant, without probable cause, indeed without any suspicion whatsoever, the federal government may seize and repeatedly search the DNA of all federal felons on supervised release, regardless of their offense or their likelihood to re-offend. They sanction the inclusion of that DNA in a massive and permanent computer database, the sole purpose of which is to aid generalized criminal investigation. This offends not only the Fourth Amendment but our precedents. I respectfully dissent.1

I. Factual Background and Statutory Framework

In March of 1999, Kriesel was convicted of a non-violent drug offense, conspiracy to possess methamphetamine with intent to distribute. He was sentenced to thirty months imprisonment and placed on supervised release. After his release in February 2008, Kriesel failed three urinalyses, testing positive once for morphine and twice for marijuana. In light of his steady employment and established ties to the community, however, the court nonetheless allowed Kriesel to remain on supervised release. Since early 2005, he has consistently passed drug and alcohol tests and has remained fully employed. No specific requirement that Kriesel submit to DNA testing was included in the terms of his release. It did include language that he must “follow the instructions of the probation officer” and that he “submit his person ... to a search upon request by the U.S. Probation Office.” When requested by the probation officer to submit to DNA testing, he refused based on his objection on principle to invasion of his privacy interest without cause. This resulted in revocation of his probation, stayed, however, pending appeal.

Non-violent drug offenders like Kriesel fall within a category of federal offenders that, according to government-conducted studies, have one of the lowest rates of recidivism. U.S. Sentencing Commission, Measuring Recidivism: The Criminal History and Computation of the Federal Sentencing Guidelines 13 (May 2004), available at http://www.ussc.gov/research. htm (Commission Report) (drug trafficking offenders within the group of offenders that “are overall the least likely to recidi-vate”); U.S. Dept. of Justice, Bureau of Justice Statistics, Offenders Returning to Federal Prison, 1986-97 1, 3 (Sept.2000) (persons convicted of drug offenses were the least likely to return to prison, with lower recidivism rate than property and public-order offenses). The government does not dispute the accuracy of these studies, nor did the government come forward with any contrary evidence to suggest that nonviolent drug offenders present a high risk of recidivism.

The statute the court approves today is a revision, and significant extension, of the predecessor version of the 2000 DNA Act. Under the 2000 DNA Act, a limited number of crimes were categorized as a “qualifying ... offense,” and only supervised releasees convicted of those qualifying offenses were required to submit to DNA sampling as a condition of their release. See DNA Analysis Backlog Elimination Act, Pub.L. No. 106-546, § 3, 114 Stat. 2726, 2729-30 (2000) (hereinafter “2000 DNA Act”). The DNA obtained from those samples are placed within “CODIS,” the FBI’s Combined DNA Index System. CODIS is a database that acts as a clearinghouse for DNA information taken from state and federal DNA collection programs, as well as crime scenes. The offenses enumerated in the 2000 DNA Act were primarily violent crimes and crimes related to illegal sexual activity. Id.

The Justice for All Act in 2004, however, amended the DNA Act so that “any felony” now serves as a qualifying offense. Pub.L. No. 108-405, § 203(b), 118 Stat. 2260, 2270 (2004) (hereinafter “2004 DNA Act”). As before, that DNA is filed in the CODIS database. Notably, although the 2004 amendment brought the entire universe of non-violent federal felonies within the Act’s purview, the House Report that accompanied the amendment does not suggest that DNA evidence has any utility in solving non-violent crimes. See H.R. Rep. 108-711 (2004), reprinted in 2005 U.S.C.C.A.N. 2274. To the contrary, the clearest statement on this point is that, “When used to its full potential, DNA evidence will help solve and may even prevent some of the most serious violent crime.” 2005 U.S.C.C.A.N. at 2277 (emphasis added).

In order to obtain his DNA, blood will be extracted from Kriesel while he is on supervised release, but the 2004 DNA Act contains no provision that requires the destruction or return of that biological sample (or the DNA analysis derived from it) once his period of supervised release ends. Indeed, the statute provides only for the destruction of the DNA analysis of a person included in CODIS in very limited circumstances. The burden to remove DNA from CODIS that was collected as a result of a conviction for a qualifying offense falls on the felon. Only by providing a “certified copy of a final court order establishing that such conviction has been overturned” will a person’s DNA analysis be expunged. 42 U.S.C. § 14132(d)(l)(A)(i). Thus, once Kriesel’s DNA is placed within CODIS, it will remain there permanently and can be continually accessed and searched so long as the search is conducted by Federal, State or local “criminal justice agencies for law enforcement identification purposes.... ” 42 U.S.C. § 14132(b)(3)(A). Simply put, once they have his DNA, police at any level of government with a general criminal investigative interest in Kriesel can tap into that DNA without any consent, suspicion, or warrant, long after his period of supervised release ends.

Although it is easy enough to conceive of the hypothetical risks to civil liberties invited by approving this kind of regime, here, there is no need to speculate. The most recent version of the DNA Act permits extraction of DNA from “individuals who are arrested, facing charges, or convicted [of qualifying felonies] or from non-United States persons who are detained under the authority of the United States.” 42 U.S.C. § 14135a(a)(l)(A) (2006).2

II. Discussion

In our fractured opinion in United States v. Kincade, a plurality of the court reasoned that the 2000 DNA Act was properly analyzed under “a traditional assessment of reasonableness gauged by the totality of the circumstances.” 379 F.3d 813, 831 (9th Cir.2004) (en banc). That view did not, however, command a majority of the court. Judge Gould separately concurred in Kincade, arguing that the DNA Act’s regime of suspicionless searches was subject to a “special needs” analysis. Id. at 840. Thus, although in Kincade the court ultimately upheld the constitutionality of the 2000 DNA Act, the opinion failed to produce any cohesive view on the appropriate analytical construct.

How analytically to approach the constitutionality was effectively answered by the Supreme Court in Samson v. California, 547 U.S. 843, 126 S.Ct. 2193, 165 L.Ed.2d 250 (2006). In Samson, the Supreme Court held that the totality-of-the-cireum-stances test was the proper mode of analysis to examine a California statute (Cal.Penal Code § 3067(a) (West 2000)) that permits suspicionless and warrantless searches of California parolees while they remain on parole. Id. at 2197. Samson, however, merely clarifies that “totality of the circumstances” is the proper analytical construct when confronted with a statute that authorizes warrantless and suspicion-less searches of parolees. Samson does not dictate, or give any support to, the outcome reached by the majority in this case.

Samson re-stated the now familiar totality-of-the-circumstances test used to determine whether or not a. search is “reasonable” by “assessing, on one hand, the degree to which it intrudes upon an individual’s privacy, and on the other, the degree to which it is needed for the promotion of legitimate governmental interests.” Id. (quoting United States v. Knights, 534 U.S. 112, 118-119, 122 S.Ct. 587, 151 L.Ed.2d 497 (2001)). In Samson, the California statute under consideration survived Fourth Amendment scrutiny based on two premises. First, Samson, by virtue of his status as a parolee, had a “severely diminished expectation ] of privacy.” Id. at 2199. Second, the Court acknowledged California’s legitimate interest in “supervising parolees” as a means to deter recidivism present in the parolee population. Id. at 2200. But, as the Court repeatedly emphasized, the governmental interest was inextricably linked to the need to supervise a parolee while he remained on parole. See id. at 2200-01 (“Thus, most parolees require intense supervision ... individualized suspicion would undermine the State’s ability to effectively supervise parolees ... the incentive-to-conceal concern justified an ‘intensive’ system for supervising probationers in Griffin. ... That concern applies with even greater force to a system of supervising parolees ... [individualized suspicion required by other State parole systems of little relevance] to our determination whether California’s supervisory system is drawn to meet its needs and is reasonable.”) (emphases added). Thus, Samson does not lend constitutional support to a statute that sanctions warrant-less and suspicionless searches untethered from an immediate supervisory need. See United States v. Weikert, 504 F.3d 1, 19 (1st Cir.2007) (Stahl, J., dissenting) (“[T]he Supreme Court [in Samson ] has now identified three limited circumstances in which a suspicionless search will survive Fourth Amendment review: (1) programmatic searches; (2) special needs searches; and (3) searches conducted as part of a state’s conditional release program. This last category is limited by the Court’s language to a state search program that is genuinely designed to improve the monitoring and reintegration of conditional releasees.”). With this framework and guidance from the Court in mind, I turn to the 2004 DNA Act.

A. Totality-of-the-Circumstances and Kincade

The most apposite precedents in this Circuit to pass on the constitutionality of the 2004 DNA Act are Samson and the plurality decision from Kincade. As noted, three years ago the far more limited 2000 DNA Act was upheld by a plurality in Kincade, relying on a totality-of-the-eir-cumstances analysis. 379 F.3d at 838-40. The plurality, however, was careful to note that their decision was limited to the version of the Act before the court, a version that targeted violent criminals. Distinguishing state programs that collected information from “non-violent drug offenders,” the plurality emphasized that “it is therefore particularly important to observe that we deal here solely with the legality of requiring compulsory DNA profiling of qualified federal offenders on conditional release.” Id. at 819 n. 9. The plurality made clear that their decision therefore did not concern “the authority of the federal government ... to pass less narrowly tailored legislation.” Id. (emphasis added). Thus, while Kincade provides some guidance, by its own terms it does not purport to decide whether or not the “less narrowly tailored legislation” before us today passes Fourth Amendment scrutiny. See id.

Although none are controlling, the majority places great weight on recent decisions from other circuits that have upheld the 2004 DNA Act against a Fourth Amendment challenge. Maj. Op. at 945-47, 950. A closer look at those cases reveals that the precise issues before us today have been treated with more breadth than depth. Two of those decisions relied on a “special needs” test which, in light of Samson, is the improper analytical method for analysis of the 2004 DNA Act. See United States v. Amerson, 483 F.3d 73, 78 (2d Cir.2007); United States v. Hook, 471 F.3d 766, 772-74 (7th Cir.2006).3

The decisions of the Eighth and Eleventh Circuits contain little to no analysis of the interests at stake when the government subjects all non-violent felony offenders to compulsory DNA sampling. In United States v. Kraklio, 451 F.3d 922, 924-25 (8th Cir.2006), the court’s discussion is devoted almost entirely to choosing between a special needs or totality-of-the-circumstances analysis. Once the totality-of-the circumstances test is chosen, Kraklio summarily concludes with a one-sentence determination that the 2004 DNA Act is constitutional under that standard. Id. at 925. Similarly, in an unpublished opinion the Eleventh Circuit upholds the 2004 DNA Act based on circuit precedent concerning a staterbased DNA collection statute, without engaging in any analysis whatsoever as to the government’s interests in permanently maintaining the DNA of non-violent federal felons. United States v. Castillo-Lagos, 147 Fed.Appx. 71, 75 (11th Cir.2005). In Weikert, the First Circuit engages in a more substantial totality-of-the-eircumstances analysis than the other decisions, but it does not consider how lowered recidivism rates among non-violent offenders would affect the government’s interests under that analysis. 504 F.3d at 12-13.

In United States v. Conley, in order to justify the 2004 DNA Act under a special needs analysis, the Sixth Circuit discussed with approval government data that suggested that white-collar crime had recidivism rates “in certain groups,” close to recidivism rates for firearms and robbery offenses. 453 F.3d 674, 679 (6th Cir.2006).4 Conley, however, is silent on recidivism rates for non-violent drug offenders like Kriesel, and the court in Conley does not cite to any recidivism argument advanced by the government when trying to justify the 2004 DNA Act under the totality of the circumstances. Id. at 680-81.

Only Banks v. United States, 490 F.3d 1178 (10th Cir.2007), genuinely addresses whether the government’s interests in maintaining a permanent DNA collection of felons on supervised release are diminished in the case of non-violent offenders. The Tenth Circuit in Banks concedes that, “[t]o be sure, DNA might prove less valuable in solving non-violent crimes than violent crimes, making the Government’s interest in testing more compelling with respect to felons convicted of violent crimes.” 490 F.3d at 1189-90. The court goes on to explain away this distinction, however, by characterizing a supervised releasee’s privacy interests as virtually non-existent and contending that “the ef fectiveness of the Government’s plan need not be high where the objective is significant and the privacy intrusion is minimal.” Id. at 1190.

I have two problems with this analysis. First, Banks places such little weight on a releasee’s privacy interest as to make it a meaningless consideration. Second, the argument that a “significant” government objective is sufficient even if the statute under consideration does not actually promote that objective is sophistry — and is shockingly wrong. Knights instructs that evaluating the reasonableness of a search for Fourth Amendment purposes should be grounded in the “degree to which it is needed for the promotion of legitimate governmental interests.” 534 U.S. at 118—119, 122 S.Ct. 587 (emphasis added) (quoting Wyoming v. Houghton, 526 U.S. 295, 300, 119 S.Ct. 1297, 143 L.Ed.2d 408 (1999)). It simply cannot be correct that a statute authorizing warrantless searches can pass Fourth Amendment scrutiny with nothing more than a hypothetical interest advanced by the government but not actually promoted by the statute.

The Tenth Circuit in Banks also cited statistical data collected from state prisons in 1994 indicating that “non-violent offenders have higher recidivism rates than the general population.” Id. at 1191 (citing U.S. Dept, of Justice, Bureau of Justice Statistics, Profile of Nonviolent Offenders Exiting State Prison (2004), http: //www. ojp.usdoj.gov/bjs/pub/pdfipnoesp.pdf).5 Working with these statistics, the court in Banks goes on to argue without authority or empirical evidence that non-violent offenders who recidivate are “committing crimes that DNA might solve (for example, violent offenses, drug offenses, and property offenses.).” Id. This unsupported leap is remarkable, given that just two pages earlier the court concedes that “DNA might prove less valuable in solving nonviolent crimes[.]” Id. at 1189.6

Because of the substantial erosion of Fourth Amendment protection posed by the 2004 DNA Act, we should not uncritically adopt decisions from other Circuits where controlling law in this Circuit does not dictate the same outcome. Thus, with the limitations from Kincade and Samson as a guide, we should consider Kriesel’s privacy interests as affected by the 2004 DNA Act on the one hand, balanced against the interests advanced by the government on the other.

B. Kriesel’s Privacy Interests

It is true that conditional releasees like Kriesel have diminished privacy expectations. Samson, 126 S.Ct. at 2199; Kincade, 379 F.3d at 833. But as the plurality observed in Kincade, diminished does not mean extinguished. 379 F.3d at 835 (“Let us be clear: Our holding in no way intimates that conditional releasees’ diminished expectations of privacy serve to extinguish their ability to invoke the protections of the Fourth Amendment’s guarantee against unreasonable searches and seizures. Where a given search or class of searches cannot satisfy the traditional totality-of-the-circumstances test, a conditional releasee may lay claim to constitutional relief — just like any other citizen.”). Thus, while Kriesel cannot claim the level of protection afforded ordinary citizens under the Fourth Amendment, his privacy interests cannot be treated as weightless in the reasonableness balance.

In considering that privacy interest, I also cannot overlook that the search here is not limited to the initial extraction of a biological sample from Kriesel, and with it, his DNA. Rather, the warrantless “search” permitted by the 2004 DNA Act extends to repeated searches of his DNA whenever the government has some minimal investigative interest. Kincade, 379 F.3d at 873 (Kozinski, J., dissenting) (“[I]t is important to recognize that the Fourth Amendment intrusion here is not primarily the taking of the blood, but seizure of the DNA fingerprint and its inclusion in a searchable database.”). Thus, I look to the interests advanced by the government, mindful of the fact that the Act permits this ongoing search of Kriesel’s DNA for his lifetime.7

C. Government’s Interests

In Kincade, the plurality pointed to three interests it deemed “monumental” in balancing the government’s interests under the 2000 DNA Act. 379 F.3d at 839. First, it relied on the use of DNA to provide “a means of identification that can be used to link conditional releasees to crimes committed while they are at large,” in order to ensure that a releasee complies with the terms of his or her release. Id. at 838. Second, DNA profiling was claimed to provide a deterrent effect that “fosters society’s enormous interest in reducing recidivism.” Id. at 839. Last, the plurality pointed to the use of DNA to solve past crime. Id. As the majority notes, these same interests are advanced by the government once again to defend the 2004 DNA Act.

The interest in ensuring compliance with terms of release, while obviously legitimate, must be viewed in light of the Court’s decision in Samson. As Samson explained, warrantless and suspicionless searches of parolees are sometimes justified under a totality-of-the-circumstances analysis provided that they are based on a legitimate supervisory need. 126 S.Ct. at 2200-01. But, unlike the California statute under consideration in Samson, the 2004 DNA Act permits those searches to continue once the period of supervised release has ended, enabling searches divorced from the government’s supervisory interests. On its face, the 2004 DNA Act allows only a releasee’s DNA profile to be removed if his or her underlying conviction is later overturned. 42 U.S.C. § 14132(d)(1)(A)(i). Thus, while ensuring compliance with the terms of supervised release stands as a credible government objective, the 2004 DNA Act is far broader than necessary to achieve that objective.

The majority disingenuously refuses to confront the fact that the 2004 DNA Act clearly permits the retention of Kriesel’s DNA once his term of supervised release is over. Apparently, it defines “the precise circumstances before us” as the rights of a parolee while on parole, with no consideration of the fact that his DNA will be retained in CODIS and be searchable for the rest of Kriesel’s lifetime. Maj. Op. at 949; see also Kincade, 379 F.3d at 841 (Gould, J., concurring in the judgment) (“Once those previously on supervised release have wholly cleared their debt to society, the question may be raised, ‘Should the CODIS entry be erased?’”). But we have previously found statutes not “limited by appropriate regulations so as to preclude general searches,” unconstitutionally overbroad without waiting for the discrete violation of the Fourth Amendment enabled by that statute. Rush v. Obledo, 756 F.2d 713, 723 (9th Cir.1985) (invalidating portion of California statute that allowed warrantless searches of family day care centers without appropriate narrowing regulations). Without reaching whether the 2004 DNA Act is independently invalid under the Fourth Amendment for that reason, the fact that a DNA entry is permanently lodged in CODIS at the very least detracts from the weight afforded the government’s claim that the statute is truly designed as a supervisory tool.

Next, the deterrent effect advanced by the government is seriously undermined here because Kriesel has offered unrebut-ted data demonstrating that rates of recidivism are among the lowest for non-violent drug offenders. Measuring Recidivism: The Criminal History and Computation of the Federal Sentencing Guidelines at 13; Offenders Returning to Federal Prison, 1986-97 at 1 and 3. The majority excuses the government’s failure to rebut Kriesel’s evidence of low recidivism rates by asserting that he “is already a recidivist, as he violated the terms of his release when he tested positive for controlled substances.” Op. at 949-50. Ironically, the authorities had all the tools they needed to detect the recidivism without resort to DNA, nor was his conduct a crime except as it violated parole.

The majority reasons that a statute permitting suspicionless extraction of DNA from all persons within a population that has low rates of recidivism is permissible because there is a basis to conclude that Kriesel himself will re-offend. Thus, the majority allows the government to escape its failure to justify a program that requires participants to submit their DNA without any suspicion, because it has suspicion in this particular case.8 This syllogism does not help the government do what it must — make some showing that extracting DNA from non-violent drug offenders discourages those persons from re-offending.

Finally, the majority reasons that creating a DNA profile of non-violent felons like Kriesel will “contribute] to the solution of past crimes.” Maj. Op. at 950. I agree in principle that a DNA program with some demonstrable effect on solving crime within the population profiled may satisfy the Fourth Amendment. Rise v. Oregon, 59 F.3d 1556, 1561 (9th Cir.1995) (“The defendants produced uncontroverted evidence documenting the high rates of recidivism among certain types of murderers and sexual offenders. Moreover, investigations of murders and sexual offenses are more likely to yield the types of evidence from which DNA information can be derived, such as blood, semen, saliva, and hair evidence, than property crimes or other offenses committed without substantial personal contact.”). But the totality-of-the-cireumstances test is not so standardless that the court may invent speculative justifications when the government offers none of its own. The majority’s suggestion that non-violent crimes have victims is true as far as it goes, but it does not go far enough. Kincade exhorted the closure brought to victims of violent crimes, but that is only half of the analysis. It justified the 2000 DNA Act on the grounds that the statute actually contributes “to the solution of past crimes.” 379 F.3d at 839. Thus, it is not enough for the majority or the government to simply argue that nonviolent crimes have victims; rather, there must be some .basis to believe that DNA profiling actually aids those victims. The government placed nothing before the court to speak to that aspect of the inquiry.

Because the government interests articulated in Kincade and re-constituted here are not sufficiently weighty to overcome Kriesel’s privacy interests, I would hold that the warrantless searches permitted by the 2004 DNA Act are unreasonable, and that the Act fails to survive review under a totality-of-the-circumstances test.

III. Conclusion

When the 2000 DNA Act narrowly survived Fourth Amendment review in this court just three years ago, we were told to take solace in the “limited nature of[the] holding.” Kincade, 379 F.3d at 835. Yet, by invoking the analysis from Kincade, the majority approves, without flinching, a statute that effects a far broader and far less justified erosion of the Fourth Amendment, extending Kincade without acknowledging it does so. However well-intentioned it may be, I find cold comfort in the majority’s assurance that its decision today is “confined to the precise circumstances before us.” Maj. Op. at 949.

I do not question the efficacy of the government’s methods. An ever-expanding and unerasable electronic index of DNA profiles, monitored by the government’s unblinking digital eye, may no doubt prove to be an effective law enforcement tool. But our compact with the government requires constitutional means, not just effective ends. Once expediency infects the Fourth Amendment analysis, as it has with the majority’s blessing of the “significant” crime-solving purposes of DNA profiling, there is no limiting principle beyond what the government says it needs. The line should be drawn far short of where the majority puts it. I dissent.

. I concur in Part III of the majority's opinion and analysis of Kriesel’s APA challenge. For reasons explained below, however, that portion of the opinion merely holds that the Attorney General did not violate the APA’s procedural requirements in promulgating regulations to implement an unconstitutional statute.

. There is a parallel provision of the act that allows for an arrestee's DNA to be expunged only if the Attorney General receives "a final court order establishing that [] charge[s] [have] been dismissed or [have] resulted in an acquittal or that no charge was filed within the applicable time period." 42 U.S.C. § 14132(d)(1)(A)(ii).

. The Supreme Court has considered and rejected the contention that the government’s ever-present generalized interest in criminal law enforcement qualifies as a “special need.” Indianapolis v. Edmond, 531 U.S. 32, 41-42, 121 S.Ct. 447, 148 L.Ed.2d 333 (2000). Thus, even if it were the proper method of analysis (and after Samson it is not), the 2004 DNA Act would fail to survive Fourth Amendment review under that rubric. Kincade, 379 F.3d at 854-57 (Reinhardt, J., dissenting) (analyzing 2000 DNA Act under special needs).

. The appellant in Conley had been convicted of bank fraud, committed while she was on probation for a similar fraud offense. 453 F.3d at 674-75.

. The statistics relied on in Banks are undermined by the government data regarding federal offenders placed in the record by Kriesel here, particularly drug offenders.

. The study of state offenders relied on by the court in Banks does not indicate whether the "violent offenses” are.of a type where DNA evidence could be of assistance. Profile of Nonviolent Offenders Exiting State Prison at 4. The other two categories, "property offenses” and "drug offenses,” are not defined or characterized by the study as violent offenses.

. The penalties for unauthorized disclosure of DNA held in CODIS do nothing to address this, as disclosure of Kriesel’s DNA for general criminal investigation is a “permissive use[]” under the Act. See 42 U.S.C. § 14135e(b); 42 U.S.C. § 14132(b)(3)(A) (disclosure permitted “to criminal justice agencies for law enforcement identification purposes”).

. Moreover, as noted Kriesel’s prior drug lapses were detected and addressed without the need for his DNA profile — which the government does not yet have. There is nothing in the record to suggest that Kriesel would have been deterred from those lapses if the government had possessed his DNA.