Opinion
20-CR-286 (ARR) (SJB)
09-10-2021
UNITED STATES OF AMERICA, Plaintiff, v. CLAUDE CELESTINE, Defendant.
OPINION & ORDER
Allyne R. Ross, United States District Judge
On August 6, 2020, a grand jury sitting in Central Islip returned an indictment charging defendant Claude Celestine with making false statements in a passport application, in violation of 18 U.S.C. § 1542, and social security fraud, in violation of 42 U.S.C. § 408(a)(7)(B). Presently before this court is Mr. Celestine's motion to dismiss the indictment against him based on substantial failures to comply with the Jury Selection and Service Act of 1968, 28 U.S.C. § 1861 et seq. (the “JSSA” or “Act”). For the reasons set forth below, Mr. Celestine's motion is DENIED.
BACKGROUND
Procedural History
I will discuss only the facts necessary to resolve the present motion. On August 6, 2020, a grand jury sitting in Central Islip indicted defendant Claude Celestine, charging him with making false statements in a passport application in violation of 18 U.S.C. § 1542, and social security fraud, in violation of 42 U.S.C. § 408(a)(7)(B). Indictment, ECF No. 4. On September 22, 2020, Mr. Celestine filed a motion requesting access to the records and papers related to the “constitution and implementation” of the jury wheels from which his grand jury was selected. See Def.'s Mot. Inspect Grand Jury Rs., ECF No. 14. I granted Mr. Celestine's request in part on November 5, 2020. Order, ECF No. 20. On May 12, 2021, I granted his request for disclosure of supplemental materials. Mr. Celestine filed the instant motion to dismiss his indictment on June 18, 2021. Def.'s Mot. Dismiss, ECF No. 39.
As an initial matter, I find Mr. Celestine's motion is timely made pursuant to the Jury Selection and Service Act. See 28 U.S.C. § 1867(a) (allowing a defendant to move to dismiss their indictment within seven days of discovering a defect in jury procedures or before voir dire examination begins, whichever is earlier). On May 17, 2021, five days after I granted Mr. Celestine's motion for disclosure of supplemental materials, Mr. Celestine moved with consent of the government for extension of time to file his instant motion. Consent Mot. Extension Time, ECF No. 37. I granted the parties' jointly proposed briefing schedule that same day.
Mr. Celestine has additionally complied with the Act's procedural requirements for bringing a motion to dismiss the indictment. Included with his motion is a sworn statement of facts by statistician Jeffrey Martin, “which if true, would constitute a substantial failure to comply with the provisions of [the Act]….” 28 U.S.C. § 1867(d).
Jury Selection and Service Act
The JSSA ensures the right of litigants in federal court to grand and petit juries “selected at random from a fair cross section of the community in the district or division wherein the court convenes.” 28 U.S.C. § 1861. To achieve its purpose, the Act requires federal districts to develop juror-selection procedures that comply with certain requirements, including that 1) all prospective jurors be selected from either voter registration lists or the lists of actual voters of the political subdivisions within the district or division, 2) the processes be designed to “ensure the random selection” of a fair cross section of persons residing in the district or division, and 3) each political subdivision within the district or division be “substantially proportionally represented” in the master jury wheel for that district or division. Id. §1863(b)(2), (3). The JSSA also authorizes districts to supplement their voter lists “where necessary to foster the policy and protect the rights secured” by the Act. Id. § 1863(b)(2).
A party whose rights under the JSSA are violated may move the court to dismiss their indictment and stay further proceedings until the failures are corrected. Id. §1867(a), (d); see also United States v. Rioux, 97 F.3d 648, 660 (2d Cir. 1996) (“The JSSA provides a method by which a defendant may challenge any conduct that substantially fails to comport with the JSSA.”).
The Eastern District's Jury Selection Process
The Eastern District of New York (the “District”) is a single district consisting of Richmond, Kings, Queens, Nassau, and Suffolk Counties. 28 U.S.C. § 112(c). Its grand and petit juries are convened pursuant to the District's October 30, 2006 Amended Jury Selection Plan (the “Jury Selection Plan”). Under the Jury Selection Plan, the Clerk of Court is charged with establishing a master jury wheel of “all persons randomly selected from the combined source lists of all the counties of the Eastern District.” Def.'s Mot. Dismiss, Ex. A § 5 (“Jury Selection Plan”), ECF No. 39-1. These “source lists” consist of “voter registration lists of all the counties within [the District] supplemented by lists for these counties from the New York State Department of Motor Vehicles.” Id. §§ 4, 5.
Under the Jury Selection Plan and in accord with the Act, names selected for the master jury wheel must proportionally represent the populations of the District's five counties. Id. § 4. By the September 1st following each presidential election and every two years afterward, the master jury wheel is emptied and refilled. Id. § 5.
From the master jury wheel, juror names are randomly selected and mailed a juror qualification form. Id. § 6. Those jurors who return the form and are neither excused, exempt, nor otherwise disqualified constitute the qualified jury wheel. Id. § 11. It is from this qualified wheel that jurors are randomly selected to sit on grand and petit juries. Id. § 13.
The grand jurors who returned Mr. Celestine's indictment in the instant case were randomly selected from a qualified jury wheel consisting of 66, 209 names. Gov.'s Mot. Opp'n Mot. Dismiss 3 (“Gov.'s Opp'n”), ECF No. 41. The relevant master jury wheel was created on April 25, 2017 and consisted of 531, 797 names from combined source lists. Gov.'s Opp'n, Ex. 1 ¶ 8 (“Siskin Rep.”), ECF No. 41-1.
LEGAL STANDARD AND DISCUSSION
I. Mr. Celestine's Objections to His Jury Selection Process Under the JSSA.
Mr. Celestine submits that his right to a grand jury “selected at random from a fair cross section of the community in [his] district, ” 28 U.S.C. § 1861, was violated on four separate grounds.
First, Mr. Celestine argues that the empaneled grand jury was not selected from a fair cross section of his community because Black and Latinx individuals were systematically underrepresented. Mr. Celestine attributes this underrepresentation to the District's exclusion of inactive voters from its voter registration lists and persons with non-driver's license identification cards from its supplemental Department of Motor Vehicles (“DMV”) lists. Def.'s Mot. Dismiss 11-13 (“Def.'s Mot.”), ECF No. 39. Second, Mr. Celestine challenges the exclusion of inactive voters as itself a substantial violation of the JSSA. Id. at 14. Third, Mr. Celestine alleges that the District did not appropriately deduplicate prospective juror names when it merged the voter registration and DMV lists, resulting in a 2017 compiled source list that was 36.95% larger than the District's jury-eligible pool-a difference of 2, 008, 918 names. Id. at 15. Finally, Mr. Celestine contends that the master jury wheel was not substantially proportional to the District's counties, as required under the Act, because Queens County was overrepresented by 2.55%. Id.; Def.'s Mot., Ex. B ¶¶ 81-87 (“Martin Decl.”), ECF No. 39-2.
As the Court did in United States v. Scott, 20-CR-332 (AT), 2021 WL 2643819, at *2 (S.D.N.Y. June 28, 2021), I use the term Latinx to refer to the community described by the parties and their experts as “Hispanic or Latino.”
II. Mr. Celestine's Fair Cross Section Claim.
A. Legal standard.
A violation of the JSSA is actionable only if it is substantial. “Mere ‘technical' violations” do not constitute substantial failures to comply with the Act. United States v. LaChance, 788 F.2d 856, 870 (2d Cir. 1986) (quoting United States v. Carmichael, 685 F.2d 903, 911 (4th Cir. 1982)); see also H.R. Rep. No. 90-1076, at 1805 (1968) (“[C]hallenges will lie only for substantial failure to comply with the statutory provisions. There is room for a doctrine of harmless error.”). Whether a violation is “technical” or “substantial” turns on the “nature and extent” of its impact on the venires used to create the grand jury. LaChance, 788 F.2d at 870.
The JSSA extends to federal juries the Sixth Amendment's requirement that a jury be drawn from a fair cross section of the community. See Id. at 864 (citing 28 U.S.C. § 1861). Alleged violations of the Act's fair cross section guarantee are thus analyzed under the same three-pronged test, first stated in Duren v. Missouri, 439 U.S. 357 (1979), used for evaluating Sixth Amendment challenges. LaChance, 788 F.3d at 864. To establish a prima facie case, a defendant must show,
While Mr. Celestine does not bring a Sixth Amendment challenge in the instant motion, had he, the analysis would be the same as that for his JSSA fair cross section claim.
(1) that the group alleged to be excluded is a “distinctive” group in the community;
(2) that the representation of this group in venires from which juries are selected is not fair and reasonable in relation to the number of such persons in the community; and
(3) that this underrepresentation is due to systematic exclusion of the group in the jury-selection process.Duren, 439 U.S. at 364. To prevail on a fair cross section claim, a defendant is not required to show that the jury selection system was motivated by discriminatory animus. Instead, she need only show that the processes used in the selection of jurors led to the systematic exclusion of a cognizable group. See United States v. Biaggi, 909 F.2d 662, 678 (2d Cir. 1990) (“[D]iscriminatory intent is not an element of a Sixth Amendment ‘fair cross-section' claim.”).
B. Threshold Determinations.
To conduct the Duren analysis in the present case, I must first determine the proper comparators and means of comparison for evaluating Mr. Celestine's claim. This requires assessing three threshold issues: (1) the relevant jury wheel to compare to the community, (2) the composition of the relevant community, and (3) the proper statistical method for measuring underrepresentation in the relevant jury wheel. See United States v. Barlow, 732 F.Supp.2d 1, 28 (E.D.N.Y. 2010), aff'd, 479 Fed.Appx. 372 (2d Cir. 2012).
1. The relevant jury wheel
Our Court of Appeals has stated that the relevant jury wheel “may be defined by: (1) the master list; (2) the qualified wheel; (3) the venires; or (4) a combination of the three.” Rioux, 97 F.3d at 657; see also United States v. Allen, No. 20-CR-366 (NSR), 2021 WL 431458, at *5 (S.D.N.Y. Feb. 8, 2021) (“The Second Circuit has not stated a preference for the use of one wheel over the other.”). Mr. Celestine and the government contest whether the master or qualified jury wheel is here proper. The decision of the relevant venire is not insignificant, as the percentages of Black and Latinx persons between the master and qualified jury wheels substantially differ. Whereas Black individuals made up 19.9% and Latinx individuals made up 16.18% of the master jury wheel based on the government's estimates, Siskin Rep. 13 tbl. 3, those numbers dropped to 15.65% Black and 14.07% Latinx in the qualified jury wheel, Martin Decl. ¶ 35.
While self-reported racial and ethnicity information exists for the qualified jury wheel, which is composed based on returned juror questionnaires, no such information exists for the master jury wheel, active voter registration lists, or DMV records. To arrive at racial and ethnicity information for the master jury wheel, the government's expert relied on geocoding, which estimates “the proportion of persons who are of a given race based on the racial mix of the area in which they live.” Siskin Rep. ¶ 17.
In deciding the proper wheel to use, other district courts in our circuit have “defined the relevant jury pool with reference to ‘the systematic defect identified by the defendant.'” United States v. Schulte, No. 17-CR-548 (PAC), 2021 WL 1146094, at *4 (S.D.N.Y. Mar. 24, 2021) (quoting United States v. Rioux, 930 F.Supp. 1558, 1565-66 (D. Conn. 1995), aff'd, 97 F.3d 648 (2d Cir. 1996)); see also Allen, 2021 WL 431458, at *5 (defining the relevant jury wheels based on the wheels impacted by the systematic selection issues); United States v. Scott, No. 20-CR-332 (AT), 2021 WL 2643819, at *5 (S.D.N.Y. June 28, 2021) (finding the master jury wheel the proper comparator for the defendant's claim of, inter alia, exclusion of inactive voters).
The government argues that examination of the master jury wheel is appropriate in this case because the “purported systematic defects identified by the defendant”-exclusion of inactive voters and holders of non-driver's license state identification cards-“all relate to the creation of the [source lists] and Master Wheel.” Gov.'s Opp'n 18. But Mr. Celestine urges consideration of the qualified jury wheel because it is the first venire to contain self-reported racial and ethnic data. Def.'s Reply Gov.'s Opp'n 2-3 (“Def.'s Reply”), ECF No. 43.
Mr. Celestine's arguments in favor of the qualified jury wheel are twofold. First, Mr. Celestine contends that “underrepresentation in the source list carries through to the qualified jury wheel, as well as every other step downstream in the jury selection process.” Id. at 2. I find this argument unavailing. Unlike the master jury wheel, which is comprised of persons randomly selected from the relevant source lists, the qualified jury wheel is made up of jurors who choose to return the questionnaire form. In other words, “there are non-random factors that influence the composition of the qualified wheel and the qualified wheel's composition can be influenced by external forces.” Allen, 2021 WL 431458, at *5. The master jury wheel, therefore, provides a better indication of underrepresentation caused by the jury selection process itself, rather than by various external forces.
Next, Mr. Celestine suggests that the qualified jury wheel is more reliable than the master jury wheel because it has “the benefit of self-reported racial and ethnic data.” Def.'s Reply 2. Mr. Celestine rejects the government's use of geocoding-a method of statistical analysis whereby a person's race or ethnicity is estimated based on their neighborhood-arguing that geocoding lacks the accuracy of self-reported demographic information. Id. In support of his contention, Mr. Celestine points to the imprecision of geocoding when used to estimate the racial and ethnic composition of the qualified jury wheel. In absolute terms, geocoding leads to an overestimation of the Black population by 1.71% and an overestimation of the Latinx population by 1.37%. Def.'s Reply, Ex. A ¶¶ 12-13 (“Martin Supp. Decl.”), ECF No. 43-1. Mr. Celestine posits that geocoding “would be expected to do the same in the master jury wheel, as well.” Def.'s Reply 2.
While it is true that the second prong of the Duren analysis requires a degree of precision that, as here, might favor the use of the qualified jury wheel, see United Sates v. Lawrence, No. 21-CR-127 (PGG), 2021 WL 3500838, at *9 (S.D.N.Y. Aug. 9, 2021) (finding the master wheel “less probative” than the qualified wheel because “reliable information concerning the demographics of the master wheel is not available, and techniques such as geocoding have limited utility”), the defect identified by Mr. Celestine occurred upstream, at the creation of the source lists and master jury wheel. Despite these competing interests, I find it unnecessary to resolve whether the master or qualified jury wheel is the relevant comparator because in either instance, the government prevails under Duren's second factor. I will therefore examine both the master and qualified jury wheels for purposes of my analysis.
2. The relevant community
Mr. Celestine and the government agree that the relevant community under the Duren analysis is the District's jury-eligible population-all citizens in the District eighteen years and older-but submit different years for measuring the community: 2019, when the grand jury in Mr. Celestine's case was selected, versus 2017, when the source lists and master jury wheel were created. The Duren analysis looks to whether the processes used to create jury panels substantially violate the JSSA's fair cross section guarantee. It follows, therefore, that the proper year for measuring the community should match the year in which the defect in process is alleged to have occurred. As I will analyze the second prong of the Duren test using both the qualified and the master jury wheels as comparators, discussed supra II.B.1, the proper community for the instant motion is the jury-eligible community in 2017, for comparison to the master jury wheel, and in 2019, for comparison to the qualified jury wheel.
According to the government's expert witness, the jury-eligible community in 2017, based on 2017 census data, was 20.19% Black and 16.44% Latinx. Siskin Rep. 8 tbl. 1. These numbers minimally changed in 2019, with the percentages of Black and Latinx persons in the community being 20.10% and 16.93% respectively. Id. ¶ 27.
3. The proper method for measuring underrepresentation
In comparing the jury wheel to the jury-eligible population, courts may use any one of several different statistical methodologies. See Rioux, 97 F.3d at 655 (noting that “[c]ourts have evolved several statistical approaches to assess compliance with the second prong of the Duren test”). Mr. Celestine undertakes his analysis under three models-statistical decision theory, comparative disparity, and absolute disparity. Def.'s Mtn. 8-9. The statistical decision theory “calculates probabilities and measures the likelihood that underrepresentation could have occurred by sheer chance.” Rioux, 97 F.3d at 655. The comparative disparity method “measures the diminished likelihood that members of an underrepresented group, when compared to the population as a whole, will be called for jury service.” Id. (quoting Ramseur v. Beyer, 983 F.2d 1215, 1231-32 (3d Cir.1992)). Finally, the absolute disparity method measures the difference between the group's representation in the jury-eligible population and in the relevant jury wheel. Id. So, for example, if Black individuals make up 10% of the jury-eligible group, but just 2% of the relevant jury wheel, the absolute disparity is 8%. See id.
In the Second Circuit, the absolute disparity approach is the “favored approach for Sixth Amendment claims.” Scott, 2021 WL 2643819, at *8 (citing Rioux, 97 F.3d at 655-56)); see also Biaggi, 909 F.2d at 678 (reaffirming court's use of the absolute numbers approach). While few courts use the statistical decision theory approach to calculate jury underrepresentation, compare Rioux, 97 F.3d at 655 (“[N]o court in the country has accepted SDT alone as determinative in Sixth Amendment challenges to jury selection systems.”), with State v. Lilly, 930 N.W.2d 293, 304 (Iowa 2019) (rejecting the absolute disparity and comparative disparity methods in favor of statistical decision theory), there are limited instances when the comparative method is favored over the absolute disparity approach, see e.g., Rioux, 97 F.3d at 656 (noting that the absolute disparity method may be improper where the minority population is a “tiny percentage” of the jury-eligible population, the jury selection procedures are not benign, and there is insufficient data to determine if the absolute numbers analysis yields a statistically significant underrepresentation). Whether those instances are present in Mr. Celestine's case is immaterial because the proper method does not appear to be in dispute between the parties. Mr. Celestine finds substantial underrepresentation of Black and Latinx jurors under each of the three statistical approaches but does not state a preference for any one method. Def.'s Mot. 9-10. The government conducts its analysis only under the absolute disparity theory. Gov.'s Opp'n 18. In line with courts in this circuit, I adopt the absolute disparity method to compare the qualified and master jury wheels to Black and Latinx voters in the District.
C. Analysis.
As an initial matter, the government argues that exclusion of inactive voters and non-driver's license holders from the source lists does not rise to the level of a JSSA violation, let alone a substantial violation under the Duren test. Gov.'s Opp'n 8-17. I disagree and find that exclusion of inactive voters contravenes the plain meaning of the Act. I decline to evaluate the government's contention that exclusion of non-driver's license holders comports with the JSSA for reasons discussed herein.
Under the JSSA, “voting registration lists” are defined as “the official records maintained by State or local election officials of persons registered to vote in either the most recent State or the most recent Federal general election.” 28 U.S.C. § 1869(c). The government argues exclusion of inactive voters complies with the JSSA because under New York law, inactive voters are not listed in polling-place ledgers and, consequently, can vote only by affidavit ballot on election day. Gov.'s Opp'n 10 (citing N.Y. Elec. Law § 8-302(e)(ii)). According to the government, inactive voters “are thus not necessarily included on lists of ‘persons registered to vote in either the most recent State or the most recent Federal general election, '” id. (quoting 28 U.S.C. § 1869(c)), as the term “voter registration lists” is defined under the Act.
In its opposition papers, the government implies that compliance by the Clerk of Court with procedures laid out in the District's Jury Selection Plan frustrates Mr. Celestine's fair cross section claim. See Gov.'s Opp'n 11 (“While the defendant argues that New York authorities (not the Clerk of Court) improperly categorize many eligible voters as inactive and while the question of whether the list used on election day or the fuller list of all registered voters is the ‘official record[]' is a difficult one, those are not issues the Clerk of the Court is responsible for resolving.”) (citation omitted). This misconstrues the JSSA. To satisfy his burden under this statute, Mr. Celestine “need not prove discriminatory intent on the part of those constructing or administering the jury selection process.” United States v. Jackman, 46 F.3d 1240, 1246 (2d Cir. 1995); cf. LaChance, 788 F.2d at 864 (noting that the defendant's statutory and Sixth Amendment claims are properly analyzed under the Duren test).
In support of its contention, the government cites a series of district court cases from this circuit finding that exclusion of inactive voters would, at most, be a “technical violation, ” if a violation at all. See Allen, 2021 WL 431458, at *10 (finding the exclusion of inactive voters was not a violation of the JSSA, but even if it were, it would be a technical violation); United States v. Middlebrooks, No. 21-CR-89 (RMB), 2021 WL 2402162, at *4 (S.D.N.Y. June 10, 2021) (adopting Allen's reasoning); Scott, 2021 WL 2643819, at *15-16 (holding that the exclusion of inactive voters was neither violative nor substantially violative of the JSSA). I decline to adopt such reasoning. While exclusion of inactive voters may be an insubstantial violation, and thus unactionable under the JSSA, an examination of the process whereby voters become inactive leads me to conclude that their exclusion contravenes the JSSA's plain meaning.
Pursuant to New York Election law, a voter becomes inactive when “any mail sent to the voter is returned as undeliverable or when a change of address is received through the National Change of Address System.” Common Cause/New York v. Brehm, 344 F.Supp.3d 542, 550 (S.D.N.Y. 2018) (citing N.Y. Elec. Law § 5-712(1), (5)). Voters placed on inactive status are immediately removed from the voter registration poll books, which are used at polling stations on election day. However, inactive voters remain on file with the Board of Elections (“BOE”). N.Y. Elec. Law § 5-213(2).
When a voter is put on inactive status in New York, the BOE contemporaneously sends her a notice asking for confirmation of her residence or for her to alert the Board of Elections of her changed address. Common Cause/New York, 344 F.Supp.3d at 550. Pursuant to the National Voter Registration Act (“NVRA”), any registered voter who does not return her confirmation notice must be kept on her state's list of eligible voters in federal elections “for a period covering two general elections for federal office.” Id. (quoting Husted v. A. Philip Randolph Inst., 138 S.Ct. 1833, 1839, (2018)) (emphasis added); see also 52 U.S.C. § 20507(d).
New York State maintains such a list. Under N.Y. Elec. Law § 5-602(1), after the last day of local registration and by the sixth day before the next general election in each year, the local boards of election are required to compile “a complete list of names and residence addresses of the registered voters for each election district over which the board has jurisdiction, ” including inactive voters. Each local board has discretion over the list's arrangement. It can indicate on the list “the registrants who are in inactive status, ” or it may “publish the names of the registrants in inactive status as a separate list.” N.Y. Elec. Law § 5-602(1). A copy of each such list(s) must be sent to the BOE, which compiles the respective county lists into a “single integrated list.” This integrated list constitutes the “official record of the registration of each voter” and is considered “the official list of voters for the state of New York.” N.Y. Elec. Law § 5-614(1), (2) (emphasis added). That inactive voters are included in the “official records … of persons registered to vote, ” 28 U.S.C. § 1869(c), was the position of the BOE itself in Common Cause/New York v. Brehm, 344 F.Supp.3d 542, 551. There, Defendant Commissioners of the New York State Board of
“In lieu of publishing such a registration list, such board of elections may publish a complete list of the names and residence addresses of all registered voters whose names do not appear in the annual enrollment lists published in such year by such board, in the same form as such enrollment lists, and a list of the registered voters whose names appear in such annual enrollment lists but who have been placed in inactive status or whose registrations have been cancelled since the publication of such annual enrollment lists.” N.Y. Elec. Law § 5-602(1).
Elections argued that exclusion of inactive voters from county poll books used on election day did not constitute removal from the official list of voters for the state of New York. See Defs.' Mem. Law in Supp. Mot. to Dismiss at 6-7, Common Cause/New York, 344 F.Supp.3d (No. 17-CV-6770 (AJN)). The court was persuaded by the defendants' argument, holding that removal of inactive voters from poll ledgers did not constitute removal from the “official list of eligible voters” in contravention of the NVRA. Common Cause/New York, 344 F.Supp.3d at 555.
To summarize, inactive voters are included in New York's official voter registration list and are published by county boards of election in their complete list(s) of names and residences of registered voters over which they have jurisdiction. However, the District does not include these voters in its source lists. I thus find that the District has not complied with its own Jury Selection Plan to select prospective jurors from “the official records maintained by State or local election officials of persons registered to vote in either the most recent State or the most recent Federal general election.” 28 U.S.C. § 1869(c). Because the District has elected to select prospective jurors from voter registration lists, rather than lists of actual voters, as alternatively provided by the JSSA, the exclusion of inactive voters from the source lists and master jury wheel violates the letter of the Act.
The government additionally contends that exclusion of non-driver's license holders from the District's supplemental DMV lists complies with the plain meaning of the JSSA. Gov.'s Opp'n 12-13. However, I need not reach this argument. Because the Duren analysis is required for any alleged violation, the exclusion of inactive voters from the jury wheels is itself sufficient to compel such analysis. I now turn to whether the District's violation of the Act was substantial.
I also need not address the parties' disagreement on how the District populated the master jury wheel-whether by random selection, as the government contends, Gov.'s Opp'n 12-13, or in proportion to each county's share of the voter registration lists, which, Mr. Celestine alleges, “blunt[s]” the “salutary effect” of supplementation by DMV records, Def.'s Mot. 13. This dispute is irrelevant to the analysis at hand.
Although I do not reach whether the exclusion of non-driver's license holders from the source lists violates the plain meaning of the law, my Duren analysis requires that I consider the composite source lists used by the District to assess whether the JSSA was substantially violated. In the District, these source lists are comprised of voter registration lists, consisting of only active voters, and DMV records, consisting of only driver's license holders. I therefore consider the DMV supplemental lists in the analysis that follows.
1. First Duren factor: distinctive group
Both Mr. Celestine and the government aver that Black and Latinx people are “distinctive” groups in the community. See Def.'s Mot. 8; Gov.'s Opp'n 17; see also United States v. Jackman, 46 F.3d 1240, 1246 (2d Cir. 1995) (“There is little question that both Blacks and Hispanics are ‘distinctive' groups in the community for purposes of this test.”).
2. Second Duren factor: significant underrepresentation
Under the second Duren prong, the representation of Black and Latinx people in the jury pool must be fair and reasonable in relation to the number of Black and Latinx jury-eligible persons in the District. Per my discussion in II.B.3, I will analyze the significance of the underrepresentation of Black and Latinx individuals in the master and qualified jury wheels using the absolute disparity approach.
In United States v. Biaggi, 909 F.2d 662, the Second Circuit found that absolute disparities of 3.6% for Black individuals and 4.7% for Latinx individuals pressed the outer limits of acceptable disparities under Duren but rejected the defendant's fair cross section claim because the underrepresentation had resulted from “benign” circumstances-the use of voter registration lists. Id. at 677-78. Since then, courts in this circuit have repeatedly rejected fair cross section claims where the absolute disparities fall near Biaggi's but the reasons for underrepresentation are benign in nature. See, e.g., Allen 2021 WL 431458, at *8 (holding Sixth Amendment fair cross section claim was not violated by absolute disparities of 3.69% and 3.64%); Scott, 2021 WL 2643819, at *10 (stating absolute disparities of 3.69% and 3.64% were “within the range found to be acceptable where the cause is ‘benign'” (quoting Rioux, 97 F.3d at 658)).
Here, the absolute disparities between Black and Latinx individuals in the jury wheel as compared to the jury-eligible community fall within the acceptable range under Biaggi regardless of the venire comparator used. According to the government, the absolute disparity between the master jury wheel and the jury-eligible community based on 2017 data, the year the master jury wheel was constructed, was 0.29% for Black individuals and 0.26% for Latinx individuals, see Gov.'s Opp'n 18, disparities far below those tolerated in Biaggi. Looking to the qualified jury wheel instead, the disparities jump significantly to 4.45% for Black persons and 2.86% for Latinx persons. Def.'s Mot. 10; Martin Decl. ¶ 35-36. While these latter disparities approach the outer limits articulated in Biaggi, Mr. Celestine's argument that the underrepresentation of Black and Latinx persons is due to less “benign” circumstances goes against the weight of law in this circuit. Mr. Celestine alleges that the exclusion of inactive voters and non-driver's license holders from the District's source lists and jury wheels impacts Black and Latinx people disproportionately because 1) based on the demographics of inactive voters in other jurisdictions, “there is reason to believe that inactive voters are more diverse…than active voters, ” Def.'s Mot. 12, and 2) New York State residents who have non-driver's license identification cards are “more likely to rely on public transportation, live in the district's urban areas, and be racially diverse than driver's license holders, ” id. at 13. While the District's exclusion of inactive voters and non-driver's license holders from its venires may have disparate impact on communities of color, courts have found that the use of “facially neutral means to construct a jury pool is benign, even if, because of demographic patterns or other ‘private sector influences,' these result in disproportionate representation.” Scott, 2021 WL 2643819, at * 9 (quoting United States v. Barnes, No. 94-CR-112 (AHN), 1996 WL 684388, at *6 (D. Conn. June 26, 1996)); see also Biaggi, 909 F.2d at 678; cf. United States v. Charles, No. 20-CR-419 (VM), 2021 WL 2457139, at *4 (S.D.N.Y. June 16, 2021) (explaining that while the four-year replenishment period of the district's master jury wheel may disproportionately impact younger, Black, and Latinx people, who move at higher rates, “moving tendencies and people's ages are external forces”). As here, where the exclusion of inactive voters and holders of non-driver's license photo identification cards from the jury pool is facially neutral and results in relatively small disparities, I find those disparities fair and reasonable. Mr. Celestine's fair cross section claim therefore fails under the second Duren factor.
Even assuming, arguendo, that the jury-eligible community in 2019, when Mr. Celestine's grand jury was selected, is the proper year against which to compare the master jury wheel, the absolute disparities fall substantially below Biaggi's at 0.20% for Black individuals and 0.75% for Latinx individuals. Gov.'s Opp'n 18 (citing Siskin Rep. 14 tbl. 4).
3. Third Duren factor: systematic exclusion
The third Duren prong is satisfied where “the underrepresentation is due to the system of jury selection itself, rather than external forces.” Rioux, 97 F.3d at 658. So, for example, “[t]he inability to serve juror questionnaires because they were returned as undeliverable is not due to the system itself, but to outside forces, such as demographic changes.” Id.
Because Mr. Celestine's fair cross section claim fails under the second Duren factor, I decline to examine whether the exclusion of inactive voters and non-driver's license holders from the jury venires relates to the system of jury selection itself. See, e.g., LaChance, 788 F.2d at 869 (declining to consider the third prong of Duren because defendant failed to satisfy the second); Charles, 2021 WL 2457139, at *4 (finding that because the defendant failed to satisfy Duren's second factor, the court did not need to address the third).
III. Mr. Celestine's Additional Challenges Under the Jury Selection and Service Act.
In addition to his fair cross section claim, Mr. Celestine challenges the District's grand jury selection procedures on the grounds that they violate the JSSA in three additional ways. Def.'s Mot. 14-16. He first challenges the exclusion of inactive voters as itself substantially violative of the JSSA. Id. at 14 Mr. Celestine next alleges that he was denied a randomly selected grand jury because the District did not adequately deduplicate prospective juror names when it merged the voter registration and DMV lists, resulting in a 2017 source list that, at 7, 445, 208 people, was 36.95% larger than the jury-eligible pool. See Id. at 15; Martin Decl. ¶¶ 74-80. Finally, Mr. Celestine argues that counties within the District were not substantially proportionally represented in the master jury wheel, as required under 28 U.S.C. § 1863(b)(3), because Queens County was overrepresented by 2.55%, at the expense of Kings, Nassau, and Suffolk Counties. Def.'s Mot. 15.
A. Legal standard.
In his dissent in United States v. Jackman, 46 F.3d 1240, Judge Walker posited that relief for substantial violations of the JSSA extends beyond the “substantial underrepresentation of a cognizable minority [required] to prevail under the Sixth Amendment.” Id. at 1250 (Walker, J., dissenting). Since then, courts in our and sister circuits have found that in addition to substantial violations of the fair cross section guarantee, the JSSA is contravened when “the alleged violation frustrates the Act's principles of random selection and objective determination of juror disqualification, exemptions and excuses.” Rioux, 930 F.Supp. at 1580-81; see also United States v. Bearden, 659 F.2d 590, 600-01 (5th Cir. 1981) (finding that the determination of substantial failure to comply requires weighing the alleged violation against the JSSA's dual goals of random selection and objective criteria for juror disqualification, excusal, exemption, and exclusion). Assessment of Mr. Celestine's three remaining claims in turn hinges on what is meant by “random selection” as used in the JSSA-whether it merely requires jurors to have “equal odds” of selection, as Mr. Celestine suggests, see Def.'s Mot. 15, or whether, as the government argues, it is evaluated vis-à-vis its impact on a distinctive group. Gov.'s Opp'n 24.
The JSSA's legislative history is instructive here. According to the Act's Senate Report, the method of random selection “virtually eliminates the possibility of impermissible discrimination and arbitrariness at all stages of the jury selection process, and thereby tends to ensure that the jury list will be drawn from a cross section of the community.” S. Rep. No. 90-891, at 16 (1967). Thus,
to the extent that the bill does provide for random selection, it does not insist upon randomness in the sense in which that term might be understood by statisticians…. No doubt [statistically random] systems enhance the likelihood of attaining the cross sectional goal of the bill. But…[i]t is sufficient for the purpose of this legislation if the plan adopts some system of selection that affords no room for impermissible discrimination against individuals or groups.Id. at 16 n.9. In accord with this objective, courts have found that jury selection procedures substantially violate the JSSA's random selection provision where they “result or have the potential to result in discrimination among cognizable groups of prospective jurors.” Bearden, 659 F.2d at 602; see United States v. Ramirez, 884 F.2d 1524, 1530 (1st Cir. 1989) (“There is no doubt that the use of volunteer jurors offends the random selection requirement of the Act. The test, however, for a substantial violation is whether it ‘either allowed discriminatory selection of jurors or otherwise prevented jury panels from consisting of fair cross sections of the community.'” (quoting Bearden, 659 F.2d at 602)); Scott, 2021 WL 2643819, at *15 (explaining that violations of the JSSA's random selection requirement are analyzed under the Duren test to determine if non-randomness led to impermissible discrimination).
Mr. Celestine's outstanding claims are properly analyzed, therefore, with respect to whether they permitted the discriminatory selection of grand jurors or prevented Mr. Celestine's grand jury panel from consisting of a fair cross section of his community.
B. Analysis.
1. Automatic exclusion of inactive voters
Mr. Celestine challenges the exclusion of inactive voters as violative of the JSSA because “[t]heir exclusion … violates the JSSA's plain text and meaning.” Def.'s Mot. 14. As discussed supra, Discussion II.C, I agree with Mr. Celestine's argument that exclusion of inactive voters violates the JSSA, which requires jury venires to be constructed using the list of actual voters within the district or, as is the practice in this District, from voter registration lists. However, I do not find this violation to constitute a “substantial failure to comply” with the JSSA. 28 U.S.C. § 1867(d).
According to Mr. Celestine, the voter registration lists used to fill the master and qualified jury wheels excluded 6.38% of registered voters (or 323, 753 eligible jurors) who were labeled “inactive” by the county boards of election. See Def.'s Mot. 11. Mr. Celestine argues this disparately impacted Black and Latinx community members because, based on the demographics of inactive voters in other jurisdictions, inactive voters in the District are likely more diverse. Def.'s Mot. 12; Martin Decl. ¶ 61. However, Mr. Celestine provides no District-specific evidence to this effect, citing only the demographics of statewide inactive voters in the 2016 presidential election to support his contention. Just as I found that Mr. Celestine's fair cross section claim could not succeed, so too must I find that exclusion of inactive voters fails as a substantial violation of the JSSA.
The government and its expert claim that inclusion of inactive voters would minimally impact the racial composition of the master jury wheel. See Gov't Opp'n 11-12; Siskin Rep. ¶ 22. Mr. Celestine challenges this finding, alleging that the list the government relied on in reaching this conclusion excluded inactive voters as well. Def.'s Reply 4 n.8. I decline to assess the racial composition of inactive voters absent District-specific evidence of their demographics.
2. Inadequate deduplication of merged voter registration and DMV lists
Under the District's Jury Selection Plan, the selection of juror names from the source lists, master jury wheel, and qualified jury wheel must “insure that the mathematical odds of any single name being picked are substantially equal, ” Jury Selection Plan § 4, in satisfaction of the JSSA's random selection requirement, see 18 U.S.C. § 1863(b)(3). Mr. Celestine alleges this provision was violated because the DMV and voter registration lists were not sufficiently deduplicated when merged to create the source list. According to Mr. Celestine, this resulted in a jury-eligible pool that was 36.95% larger than the jury eligible population, thereby increasing the likelihood of some names being placed in the master jury wheel.
While this may upset the JSSA's requirement of “random selection, ” absent a showing that the jury selection process permitted the discriminatory selection of jurors or prevented jury panels from consisting of fair cross sections of the community, Mr. Celestine's substantial violation claim fails. See Bearden, 659 F.2d at 602; Scott, 2021 WL 2643819, at *15. The exact number of duplicate entries cannot be ascertained as some driver's license holders may be inactive voters and some active voters may not be driver's license holders. Indeed, the government suggests that the effect of duplicates in the source lists may very well be “miniscule”-of the 373, 380 persons in Mr. Celestine's master jury wheel who were mailed juror qualification forms, just two were duplicates. Gov.'s Opp'n 23 n.10; Siskin Rep. ¶ 24. Irrespective of the actual number of duplicates in the merged source lists, I find this claim unavailing, as Mr. Celestine has adduced no evidence suggesting that the District's failure to adequately deduplicate juror names disfavored Black and Latinx individuals in the master jury wheel.
3. Impairment of Substantial Proportionality
The JSSA requires that “each county, parish, or similar political subdivision within the district or division [be] substantially proportionally represented in the master jury wheel for that judicial district, division, or combination of divisions.” 28 U.S.C. § 1863(b)(3). Proportional representation can be determined based on the number of actual voters in the last general election or the number of registered voters, if voter registration is uniformly required throughout the district or division. Id. While the District's Jury Selection Plan requires substantial representation in accordance with the Act, see Jury Selection Plan § 4, Mr. Celestine argues that his master jury wheel lacked proportionality because Queens County was overrepresented by 2.55% when compared to its share of the voter registration rolls, and King, Nassau, and Suffolk Counties were underrepresented. See Def.'s Mot. 15; Martin Decl. ¶¶ 82-86. As with his deduplication claim, Mr. Celestine does not allege that this overrepresentation impacted a cognizable group, and “[g]eographical groupings are not cognizable under the JSSA or the Sixth Amendment.” Scott, 2021 WL 2643819, at *16 n.13; cf United States v. Yonkers Contracting Co., 682 F.Supp. 757, 768 (S.D.N.Y. 1988). However, disproportionate representation claims cannot succeed absent a showing that the proration violated the JSSA's fair cross section requirement. See Scott, 2021 WL 2643819, at *15-16 (citing relevant case law). Because Mr. Celestine made no such showing, I find that the overrepresentation of Queens County and underrepresentation of King, Nassau, and Suffolk Counties did not substantially violate the JSSA.
CONCLUSION
Mr. Celestine has failed to make a prima facie case that his right to a grand jury randomly selected from a fair cross section of his community was violated under the Jury Selection and Service Act. Nor has he established that the exclusion of inactive voters, inadequate deduplication of prospective juror names in the merged source lists, and overrepresentation of Queens County in the master jury wheel amounted to substantial violations of the JSSA. Accordingly, Mr. Celestine's motion to dismiss the indictment is DENIED.
SO ORDERED.