American Civil Liberties Union v. United States Department of Justice

United States Court of Appeals

FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued February 20, 2014 Decided May 9, 2014

No. 13-5064

AMERICAN CIVIL LIBERTIES UNION AND AMERICAN CIVIL

LIBERTIES UNION FOUNDATION,

APPELLANTS

v.

UNITED STATES DEPARTMENT OF JUSTICE,

APPELLEE

Appeal from the United States District Court

for the District of Columbia

(No. 1:08-cv-01157)

Arthur B. Spitzer argued the cause for appellants. With

him on the briefs were Catherine Crump and David L. Sobel.

John S. Koppel, Attorney, U.S. Department of Justice,

argued the cause for appellee. With him on the brief were

Stuart F. Delery, Assistant Attorney General, Ronald C.

Machen Jr., U.S. Attorney, and Leonard Schaitman, Attorney.

Before: TATEL, BROWN, and KAVANAUGH, Circuit

Judges.

Opinion for the Court filed by Circuit Judge TATEL.

2

Concurring opinion filed by Circuit Judge TATEL.

Dissenting opinion filed by Circuit Judge BROWN.

TATEL, Circuit Judge: Three years ago, in American Civil

Liberties Union v. U.S. Department of Justice, 655 F.3d 1

(D.C. Cir. 2011) (ACLU I), this court held that the Freedom of

Information Act required the Justice Department to disclose

case names and docket numbers for prosecutions in which the

government had obtained cellular phone tracking data without

a warrant and the defendant had ultimately been convicted.

The court left open the question whether the Department

would also have to disclose docket information for similar

prosecutions in which the defendant had been acquitted or had

the charges dismissed. Now squarely facing just that question,

we conclude that given the substantial privacy interest

individuals have in controlling information concerning

criminal charges for which they were not convicted, the

Department has properly withheld this information.

I.

In order to “open agency action to the light of public

scrutiny,” Department of the Air Force v. Rose, 425 U.S. 352,

361 (1976) (internal quotation marks omitted), FOIA requires

federal agencies, “upon request, to make ‘promptly available

to any person’ any ‘records’ so long as the request

‘reasonably describes such records,’” Assassination Archives

& Research Center v. CIA, 334 F.3d 55, 57 (D.C. Cir. 2003)

(quoting 5 U.S.C. § 552(a)(3)). This broad statutory mandate

is subject to certain enumerated exemptions. See 5 U.S.C.

§ 552(b)(1)–(9). At issue here is FOIA Exemption 7(C),

which provides that an agency may withhold “records or

information compiled for law enforcement purposes” if

disclosure “could reasonably be expected to constitute an

unwarranted invasion of personal privacy.” Id. § 552(b)(7).

3

Determining whether an invasion of privacy is “unwarranted”

within the meaning of Exemption 7(C) requires, as the

Supreme Court held in U.S. Department of Justice v.

Reporters Committee for Freedom of the Press, 489 U.S. 749,

776 (1989), “balanc[ing] the public interest in disclosure

against the interest [in privacy] Congress intended the

Exemption to protect.”

In Reporters Committee, the Supreme Court considered

the applicability of Exemption 7(C) to a request for an alleged

mob figure’s “rap sheet”—a document compiled by the FBI

that “contain[ed] certain descriptive information, such as date

of birth and physical characteristics, as well as a history of

arrests, charges, convictions, and incarcerations.” Id. at 752.

Holding that the disclosure of such rap sheets implicates a

substantial privacy interest, id. at 771, the Court rejected the

contention that any interest in avoiding disclosure

“approaches zero” simply because “events summarized in a

rap sheet have been previously disclosed to the public,” id. at

762–63. The Court explained that an individual’s interest in

privacy “encompass[es] the individual’s control of

information concerning his or her person,” id. at 763, even

though “the information may have been at one time public,”

id. at 767. Disclosure of a rap sheet, the Court found, was

particularly troubling because it would in one fell swoop

bring to light many facts about a person that might otherwise

be subject to little public scrutiny. See id. at 769–71; see also

id. at 764 (emphasizing the “distinction, in terms of personal

privacy, between scattered disclosure of the bits of

information contained in a rap sheet and revelation of the rap

sheet as a whole”). On the other side of the balance, the Court

found the public interest in disclosure to be fairly limited

because a rap sheet would reveal little about “the

Government’s activities.” Id. at 754. Thus, the Court held “as

a categorical matter” that granting a “third party’s request for

4

law enforcement records or information about a private

citizen” that “seeks no ‘official information’ about a

Government agency” would constitute an “‘unwarranted’”

invasion of privacy. Id. at 780.

The case now before us arose after the American Civil

Liberties Union learned that federal law enforcement agencies

were, without first securing a warrant, obtaining data from

cellular phone companies that could be used to track phone

users’ whereabouts. The ACLU filed FOIA requests with the

Drug Enforcement Administration and the Executive Office

for United States Attorneys, seeking, among other things,

records related to: “The case name, docket number, and court

of all criminal prosecutions, current or past, of individuals

who were tracked using mobile location data, where the

government did not first secure a warrant based on probable

cause for such data.” To compel production of these records,

the ACLU then sued the Department of Justice.

In response, the Department identified a large number of

prosecutions—the total count is currently 229—in which a

judge had, since September 2001, granted the government’s

application to obtain cell phone location data without making

a probable cause determination. The Department refused to

turn this list of cases over to the ACLU, claiming that the

information fell within FOIA Exemption 7(C).

The parties each moved for summary judgment. The

district court, then Judge Robertson, concluded that each of

the individuals who had been prosecuted in these cases had a

privacy interest in preventing disclosure of the requested

information. The court went on to draw a distinction that

neither party had directly advanced, according “a greater

privacy interest to persons who were acquitted, or whose

cases were dismissed or sealed (and remain under seal), and a

5

considerably lesser privacy interest to persons who were

convicted, or who entered public guilty pleas.” American

Civil Liberties Union v. U.S. Department of Justice, 698 F.

Supp. 2d 163, 166 (D.D.C. 2010). Determining that “the

public has a substantial interest in the subject of cell phone

tracking” that would be advanced by the requested disclosure,

the court held that “the public interest in ‘what the

government is up to’ outweighs the privacy interests of

persons who have been convicted of crimes or have entered

public guilty pleas; but . . . the privacy interests of persons

who have been acquitted, or whose cases have been sealed

and remain under seal, or whose charges have been dismissed,

outweigh the public interest in disclosure of their names and

case numbers.” Id. The district court therefore directed the

Department to disclose the requested information regarding

prosecutions in which the government had secured a

conviction but permitted it to withhold the information

regarding the remaining cases.

Both sides appealed, and this court affirmed in part. We

began our analysis by noting that, although the ACLU sought

only the case name, court, and docket number of these

prosecutions, courts “evaluating the privacy impact of the

release of information . . . have taken into consideration

potential derivative uses of that information.” ACLU I, 655

F.3d at 7. The derivative uses to be made with the requested

docket information were fairly substantial: with “little work,”

someone could “look up the underlying case files in the public

records of the courts,” id., and could even attempt to contact

the defendants, or their attorneys, directly, id. at 11–12.

Nevertheless, we concluded that, with respect to those

defendants who had ultimately been convicted, disclosure

“would compromise more than a de minimis privacy interest,

[but] it would not compromise much more.” Id. at 12. We

emphasized that, unlike in Reporters Committee, the

6

requested information pertained only to a single, relatively

recent prosecution, the details of which were already “readily

available to the public” and not at all “‘practical[ly]

obscure[].’” Id. at 9 (quoting Reporters Committee, 489 U.S.

at 762) (alteration in original). As for the public interest, we

determined that disclosure would have the significant benefit

of “shedding light on the scope and effectiveness of cell

phone tracking as a law enforcement tool,” helping to “inform

[the] ongoing public policy discussion” regarding the

propriety of warrantless cell phone tracking. Id. at 13. “[I]n

light of the strength of [this] public interest . . . and the

relative weakness of the privacy interests at stake,” we held

that the district court had correctly rejected the Department’s

contention that production of this docket information would

represent an “‘unwarranted’ invasion of privacy under

Exemption 7(C).” Id. at 16.

Significantly, however, we did not affirm the district

court’s holding that information regarding acquittals,

dismissals, or sealed cases could be withheld. We did observe

that the distinction the district court had drawn “makes some

intuitive sense, as both parties agree that the disclosure of

information regarding [such cases] raises greater privacy

concerns than the disclosure of information regarding public

convictions or public pleas.” Id. at 17. But, we continued,

“whether that is enough of a distinction to justify withholding

under Exemption 7(C) is a harder question.” Id. Because it

was unclear from the record whether there were any cases that

fell within this category, we opted to forgo resolving the

issue, instead vacating this portion of the district court’s

decision and “remand[ing] the case for th[e] court to

determine whether any of the docket numbers refer to cases in

which the defendants were acquitted, or to cases that were

dismissed or sealed.” Id.

7

Following our remand, the Department identified 214

prosecutions that had resulted in convictions or public guilty

pleas and released the docket information for these cases. This

left a total of fifteen prosecutions that were responsive to the

ACLU’s request and had ended in dismissals or acquittals, or

had been sealed. Because the ACLU did not challenge the

Department’s authority to withhold the information regarding

sealed cases, only six remain at issue—four of which were

resolved by dismissal and two that ended in acquittal.

American Civil Liberties Union v. U.S. Department of Justice,

923 F. Supp. 2d 310, 313 (D.D.C. 2013). Having established

that these six cases in fact existed, Judge Amy Berman

Jackson, to whom the case was assigned after Judge

Robertson’s retirement, again granted the Department’s

motion for summary judgment. Id. at 314.

The ACLU appeals, thus presenting us with the “harder

question” we were previously able to avoid. ACLU I, 655

F.3d at 17. Our review is de novo. “In the FOIA context this

requires that we ascertain whether the agency has sustained its

burden of demonstrating that the documents requested are . . .

exempt from disclosure under the FOIA.” Id. at 5 (internal

quotation marks omitted).

II.

As in our previous decision, we begin by assessing the

privacy interest at stake. The Department argues that

“prosecuted-but-not-convicted individuals are . . . in a similar

position to persons investigated or arrested but not

prosecuted,” and that this court has “accord[ed] a strong

privacy interest to such individuals.” Appellee’s Br. 19. The

ACLU argues that the privacy interests of defendants whose

prosecutions resulted in dismissals or acquittals are only

“marginally greater” than those of defendants who were

8

convicted—which, as we held in our prior decision, are nearly

de minimis. Appellants’ Br. 20. Each party overstates its case.

It is true, as the Department observes, that we have

regularly concluded that individuals have a “strong interest”

in avoiding disclosure of their involvement in “alleged

criminal activity.” Fitzgibbon v. CIA, 911 F.2d 755, 767 (D.C.

Cir. 1990) (internal quotation marks omitted); accord, e.g.,

People for the Ethical Treatment of Animals v. National

Institute of Health, No. 12-5183, slip op. at 9 (D.C. Cir. Mar.

14, 2014); Fund for Constitutional Government v. National

Archives & Records Service, 656 F.2d 856, 866 (D.C. Cir.

1981). Those decisions, however, dealt with individuals who

were either the subject of or involved in government

investigations of criminal activity but never charged with a

crime. See ACLU I, 655 F.3d at 7 n.8. Such individuals’

privacy interests are strong in part because disclosure would

“reveal[] to the public that the individual had been the subject

of an . . . investigation.” Baez v. U.S. Department of Justice,

647 F.2d 1328, 1338 (D.C. Cir. 1980) (emphasis added);

accord, e.g., Branch v. FBI, 658 F. Supp. 204, 209 (D.D.C.

1987). The privacy interest in preventing disclosure is

diminished, however, if the fact of someone’s involvement in

alleged criminal activity is already a matter of public record—

as will be the case when a defendant was indicted. See

Reporters Committee, 489 U.S. at 763; ACLU I, 655 F.3d at 7

& n.8. Thus, we disagree with the Department that those who

have been acquitted or had their cases dismissed and whose

involvement in alleged criminal activity has already been

publicly revealed are in the same situation as those who were

never charged in the first place.

We likewise disagree with the ACLU that the privacy

interests of defendants who were indicted but not convicted

are essentially indistinguishable from those of defendants who

9

were convicted. To be sure, many of the factors we

considered important in concluding that convicted defendants

have a relatively weak privacy interest are equally applicable

to those individuals whose interests we now consider here. In

particular, just as was true with respect to convicted

defendants, the requested docket information regarding

defendants who were charged but not convicted would

“disclose only information that has already been the subject of

a public proceeding,” is “available in public records,” ACLU

I, 655 F.3d at 8, and is likely readily accessible by the public

through “computerized government services like PACER” or

even a simple “Google search for that person’s name,” id. at

10. Indeed, as we have observed, this prior public exposure is

precisely what distinguishes the individuals whose interests

we consider in this case from those who have been

investigated but not charged. But the fact that information

about these individuals’ cases is a matter of public record

simply makes their privacy interests “fade,” not disappear

altogether. ACLU I, 655 F.3d at 9 (internal quotation marks

omitted); see Cox Broadcasting Corp. v. Cohn, 420 U.S. 469,

494–95 (1975) (“[T]he interests in privacy fade when the

information involved already appears on the public record.”);

see Reporters Committee, 489 U.S. at 767 (“[O]ur cases have

. . . recognized the privacy interest inherent in the

nondisclosure of certain information even where the

information may have been at one time public.”). Consistent

with our decision in ACLU I, we reject the dissent’s surrender

of any reasonable expectation of privacy to the Internet—a

surrender that would appear to result from a failure to

distinguish between the mere ability to access information and

the likelihood of actual public focus on that information. See

Dissenting Op. at 4–6; cf. United States v. Jones, 132 S. Ct.

945, 964 (2012) (Alito, J., concurring) (concluding that

individuals generally have a reasonable expectation in being

free of long-term GPS surveillance notwithstanding the ready

10

availability of this technology); id. at 957 (Sotomayor, J.,

concurring) (suggesting that “it may be necessary to

reconsider the premise that an individual has no reasonable

expectation of privacy in information voluntarily disclosed to

third parties,” as that “approach is ill suited to the digital

age”). And if individuals not convicted have a substantially

greater privacy interest than convicted individuals to start

with, then even after both interests are discounted due to prior

public revelation, the former interest will remain substantially

greater than the latter.

In our view, defendants whose prosecutions ended in

acquittal or dismissal have a much stronger privacy interest in

controlling information concerning those prosecutions than

defendants who were ultimately convicted. The presumption

of innocence stands as one of the most fundamental principles

of our system of criminal justice: defendants are considered

innocent unless and until the prosecution proves their guilt

beyond a reasonable doubt. See Coffin v. United States, 156

U.S. 432, 453 (1895) (“The principle that there is a

presumption of innocence in favor of the accused is the

undoubted law, axiomatic and elementary, and its

enforcement lies at the foundation of the administration of our

criminal law.”). Individuals who are charged with a crime and

ultimately prevail of course remain entitled to a version of this

presumption. In the eyes of the law, they are not guilty. Cf.

Herrera v. Collins, 506 U.S. 390, 399–400 (1993) (following

conviction, “the presumption of innocence disappears,” and

“[t]hus, in the eyes of the law, petitioner does not come before

the Court as one who is ‘innocent’”). Unfortunately, public

perceptions can be quite different. Aware of the heavy burden

of proof that the government must satisfy in a criminal

prosecution, many may well assume that individuals charged

with a crime likely committed that crime regardless of how

the case was ultimately resolved. “We all know,” ACLU

11

counsel candidly observed at oral argument, “there are some

guilty people who are not convicted.” Oral Arg. Rec. 27:03–

:08. Or as former Secretary of Labor Raymond Donovan

wondered after being acquitted of larceny and fraud, “Which

office do I go to to get my reputation back?” Selwyn Raab,

Donovan Cleared of Fraud Charges by Jury in Bronx, N.Y.

Times, May 26, 1987, at A1. Thus, if the right to privacy is, at

its essence, “the right to be let alone,” Olmstead v. United

States, 277 U.S. 438, 478 (1928) (Brandeis, J., dissenting),

those who are acquitted or whose charges are dismissed have

an especially strong interest in being let alone. Although the

fact that such defendants were accused of criminal conduct

may remain a matter of public record, they are entitled to

move on with their lives without having the public reminded

of their alleged but never proven transgressions.

This special interest in shielding those charged with but

not convicted of a crime is reflected in state laws that limit the

disclosure of criminal history summaries involving data other

than convictions. See, e.g., Conn. Gen. Stat. § 54-142n

(“Nonconviction information other than erased information

may be disclosed only to: (1) Criminal justice agencies . . . ;

(2) agencies and persons which require such information to

implement a statute or executive order that expressly refers to

criminal conduct; (3) agencies or persons authorized by a

court order, statute or decisional law to receive criminal

history record information.”); Haw. Rev. Stat. § 846-9

(providing that “[d]issemination of nonconviction data shall

be limited” to certain specified entities, but that “[t]hese

dissemination limitations do not apply to conviction data”);

see also Reporters Committee, 489 U.S. at 754 n.2 (observing

that “[i]n general, conviction data is far more available

outside the criminal justice system than is nonconviction

data,” and that in “47 states nonconviction data cannot be

disclosed at all for non-criminal justice purposes, or may be

12

disclosed only in narrowly defined circumstances” (internal

quotation marks omitted)). It is also reflected in statutes and

court decisions providing for the sealing of cases in which the

defendant was never convicted. See, e.g., N.Y. Crim. Proc.

Law § 160.50 (“Upon the termination of a criminal action or

proceeding against a person in favor of such person . . . ,

unless . . . the court . . . determines that the interests of justice

require otherwise . . . , the record of such action or proceeding

shall be sealed . . . .”); Ohio Rev. Code § 2953.52 (“Any

person, who is found not guilty of an offense by a jury or a

court or who is the defendant named in a dismissed

complaint, indictment, or information, may apply to the court

for an order to seal the person’s official records in the case.”);

see also John P. Sellers, III, Sealed with an Acquittal: When

Not Guilty Means Never Having to Say You Were Tried, 32

Cap. U. L. Rev. 1 (2003) (describing Ohio courts’ expansive

use of this power). Perhaps most important for our purposes,

it is an interest whose relative significance is reflected in our

prior decision in this case, in which we observed that the

privacy interests of defendants who have been convicted “are

weaker than for individuals who have been acquitted or

whose cases have been dismissed.” ACLU I, 655 F.3d at 7;

see also id. at 8 (emphasizing that, unlike in Reporters

Committee, the requested disclosure would “disclose only

information concerning a conviction or plea; it would not

disclose mere charges or arrests”). Indeed, even our dissenting

colleague appears to acknowledge the relative strength of this

interest. See Dissenting Op. at 2 (stating that the privacy

interests here are “marginally greater than they were in ACLU

I”).

Release of the docket information the ACLU seeks would

substantially infringe this privacy interest. It would create the

risk—perhaps small, see ACLU I, 655 F.3d at 10–11, but

nonetheless real—that renewed attention would be paid to the

13

individuals who were the subject of these prosecutions. While

this attention would have been warranted at the time of

indictment, now that these defendants have been acquitted or

had the relevant charges dismissed they have a significant and

justified interest in avoiding additional and unnecessary

publicity. For example, someone who had been acquitted of

accounting fraud after a full and fair trial, moved on with his

life, and started a family might be especially dismayed were

his neighbors, friends, and family to learn about his previous

prosecution due to the publicity associated with the release of

the requested information. Or what of a defendant charged

with producing child pornography whose case was dismissed

after the government identified the real perpetrator, yet is

nevertheless viewed with suspicion by those who learn of his

mere involvement in such a case? If, as the Supreme Court

put it in Reporters Committee, an “ordinary citizen” has a

privacy interest “in the aspects of his or her criminal history

that may have been wholly forgotten,” certainly that interest is

particularly great when the ordinary citizen was never actually

convicted but nonetheless might be presumed by the public to

have been guilty. 489 U.S. at 769. Release of this information

would also permit the ACLU or others to contact the

defendants in question in order to learn more about their

cases, something the ACLU has expressly told us it plans to

do. Though “relatively minimal,” ACLU I, 655 F.3d at 11,

such an intrusion may be especially undesirable for

individuals who are understandably trying to put their past

ensnarement in the criminal justice system behind them.

III.

Having concluded that defendants who were acquitted or

had their cases dismissed have a substantial privacy interest at

stake, we must now balance this interest against the public

interest in disclosure. Such balancing decisions, generally

speaking, are among the most challenging sorts of cases that

14

judges face. Indeed, the task brings to mind the rhetorical

question often attributed to Chief Justice Traynor of the

California Supreme Court: “Can you weigh a bushel of

horsefeathers against next Thursday?” Brainerd Currie, The

Disinterested Third State, 28 Law & Contemp. Probs. 754,

754 (1963); cf. also William Prosser, Res Ipsa Loquitur in

California, 37 Cal. L. Rev. 183, 225 (1949) (“A presumption

. . . can no more be balanced against evidence than ten pounds

of sugar can be weighed against half-past two in the

afternoon.”) (internal quotation marks omitted)). In this case,

however, the comparison is not so amorphous and the

balance, while close, is nonetheless clear.

The ACLU argues that because warrantless cellphone

tracking remains an issue of great public concern, the public

interest in disclosure is the same as it was the last time this

case was before us. According to the Department, however,

the public interest in the disclosure of these six prosecutions is

reduced by the prior disclosure of the 214 prosecutions that

resulted in convictions. In support, the Department relies on

Schrecker v. U.S. Department of Justice, 349 F.3d 657 (D.C.

Cir. 2003), in which we observed that a court’s “inquiry” into

the interest in disclosure “should focus not on the general

public interest in the subject matter of the FOIA request, but

rather on the incremental value of the specific information

being withheld.” Id. at 661 (emphasis added).

We have no need to wade into this debate. Even

assuming, as the ACLU contends, that the public interest in

the disclosure here equals that in ACLU I, that interest pales in

comparison to the substantial interests in privacy that are now

at stake. The line drawn by Judge Robertson between

prosecutions that result in convictions and those that result in

dismissals or acquittals is not just one that “makes some

intuitive sense,” as we put it in our prior opinion; it is, we

15

now hold, a distinction that is fully consistent with FOIA.

Given the fundamental interest individuals who have been

charged with but never convicted of a crime have in

preventing the repeated disclosure of the fact of their

prosecution, we have little hesitation in concluding that

release of the remaining information the ACLU seeks “could

reasonably be expected to constitute an unwarranted invasion

of personal privacy.” 5 U.S.C. § 552(b)(7)(C). Indeed, the

government, having brought the full force of its prosecutorial

power to bear against individuals it ultimately failed to prove

actually committed crimes, has a special responsibility—a

responsibility it is fulfilling here—to protect such individuals

from further public scrutiny.

IV.

One last issue demands our attention. The ACLU argues

that neither this court nor the district court could properly

conclude that the Exemption 7(C) balance tilts in favor of

withholding because the Department has failed to provide the

information necessary to make that determination. The ACLU

lists seventeen facts the Department has refused to provide—

facts relating to the specifics of the litigation in these six

cases, the particular defendants charged, and the degree to

which the cases received prior publicity. This information, it

claims, might either increase the public benefit that would

flow from disclosure of this particular docket information or

decrease the privacy interest at stake. To the extent they are

relevant at all, however, sixteen of the seventeen specifics the

ACLU contends the Department should have produced are

facts for which the burden of production actually lies with the

ACLU. See National Archives & Records Administration v.

Favish, 541 U.S. 157, 172 (2004) (“Where the privacy

concerns addressed by Exemption 7(C) are present, the

exemption requires the person requesting the information to

establish a sufficient reason for the disclosure.”); Afshar v.

16

Department of State, 702 F.2d 1125, 1130 (D.C. Cir. 1983)

(“[A] plaintiff asserting a claim of prior disclosure must bear

the initial burden of pointing to specific information in the

public domain that appears to duplicate that being withheld.”).

The one piece of information requested by the ACLU that

the government would have to produce is whether any of the

defendants have died. As we have held, not only is an

individual’s death “a relevant factor” in assessing the privacy

interests implicated by a disclosure involving that individual,

but in some circumstances the government must take “certain

basic steps to ascertain whether an individual [is] dead or

alive.” Schrecker v. U.S. Department of Justice, 254 F.3d 162,

166–67 (D.C. Cir. 2001). But although death may “diminish”

the relevant privacy interests, it “by no means extinguishes”

them because “one’s own and one’s relations’ interests in

privacy ordinarily extend beyond one’s death.” Id. at 166; see

also Swidler & Berlin v. United States, 524 U.S. 399, 407

(1998) (holding that the attorney-client privilege survives the

client’s death because “[c]lients may be concerned about

reputation, civil liability, or possible harm to friends or

family” and “[p]osthumous disclosure . . . may be as feared as

disclosure during the client’s lifetime”). Here, even assuming

any of the six individuals who were the subject of the

prosecutions at issue have died, the relevant privacy interests

remain substantial. Deceased defendants never convicted of a

crime retain a reputational interest in keeping information

concerning their prosecutions out of the public eye. They may

also have family members who themselves have a legitimate

interest in avoiding the increased scrutiny that could follow

from the release of the requested docket information. Cf.

Favish, 541 U.S. at 170 (“FOIA recognizes surviving family

members’ right to personal privacy with respect to their close

relative’s death-scene images.”). Given the substantial nature

of these interests, we conclude that withholding the requested

17

docket information would be justified under Exemption 7(C)

even if some or all of the underlying defendants were dead.

Accordingly, the district court properly granted the

Department’s motion for summary judgment notwithstanding

the Department’s apparent failure to investigate this issue.

V.

For the forgoing reasons, we affirm the district court’s

grant of summary judgment to the Department.

So ordered.

TATEL, Circuit Judge, concurring: The court’s opinion

assumes without deciding that the public interest in disclosure

of the docket information for these six prosecutions is just as

great as was the interest in disclosing the information for the

214 prosecutions the Justice Department was previously

ordered to release. See Majority Op. at 14. I write separately

to explain why I believe this prior disclosure has substantially

reduced the value of the remaining information the ACLU

continues to seek, thus further tilting the balance in favor of

withholding.

In evaluating the public benefit of disclosure under FOIA

Exemption 7(C), D.C. Circuit precedent requires that we

focus on the “incremental value” of the “specific information”

sought. Schrecker v. U.S. Department of Justice, 349 F.3d

657, 661 (D.C. Cir. 2003); accord, e.g., Bast v. U.S.

Department of Justice, 665 F.2d 1251, 1254 (D.C. Cir. 1981);

King v. U.S. Department of Justice, 830 F.2d 210, 234–35

(D.C. Cir. 1987). That is, instead of simply asking whether

there might be some general public interest in the subject

matter of the FOIA request, we ask whether and how the

information sought in a particular FOIA request will actually

cast light on the government’s activities. See, e.g., ACLU v.

U.S. Department of Justice, 655 F.3d 1, 12–16 (D.C. Cir.

2011) (ACLU I). Examining the incremental value of a given

disclosure follows from the basic purpose of the Exemption

7(C) balancing test: determining whether a particular record

or piece of information is worth the privacy costs of release

requires an assessment of the potential benefits that would

actually flow from release.

When assessing the “incremental value” of the

information sought, we of course apply the common sense

notion that the value of information depends on the mix of

data already publicly available—including that previously

released by the agency subject to the FOIA request. In U.S.

Department of State v. Ray, 502 U.S. 164 (1991), the

2

Supreme Court illustrated this approach in the process of

analyzing the benefit of disclosing information retained

pursuant to FOIA Exemption 6, a parallel exemption that

authorizes withholding records if disclosure would “constitute

a clearly unwarranted invasion of personal privacy.” 5 U.S.C.

§ 552(b)(6). In Ray, the State Department had released

documents relating to its efforts to monitor Haiti’s compliance

with its promise not to persecute certain refugees. In doing so,

however, the State Department redacted information

regarding the identity of the refugees—information that, as

the Court recognized, would, in a vacuum, have been helpful

to the FOIA requesters because it would have enabled them to

track down the refugees and ask them about any persecution.

See Ray, 502 U.S. at 171, 177; cf. id. at 179 (observing that

the Court’s resolution of the case allowed it to avoid deciding

whether derivative uses of disclosed information could

qualify as a public interest in disclosure); ACLU I, 655 F.3d at

15 (“this court takes derivative uses into account in evaluating

the impact of disclosure on the public interest”). The dissent

cannot simply misquote away the value of this information.

See Dissenting Op. at 9 (quoting from section of the Court’s

opinion describing the Eleventh Circuit’s conclusion “that the

redacted information would not, in and of itself, tell

respondents anything about Haiti’s treatment of the returnees

or this Government’s honesty, but . . . the indirect benefit of

giving respondents the means to locate the Haitian returnees

and to cross-examine them provided a public value that

required disclosure,” Ray, 502 U.S. at 170–71.) Nevertheless,

the Court rejected an effort to ascertain the refugees’

identities because the “public interest” had already been

“adequately served by disclosure of the redacted interview

summaries.” Ray, 502 U.S. at 178. It explained that the

released “documents reveal how many returnees were

interviewed, when the interviews took place, the contents of

individual interviews, and details about the status of the

3

interviewees.” Id. Thus, it concluded, “[t]he addition of the

redacted identifying information would not shed any

additional light on the Government’s conduct of its

obligation.” Id. (emphasis added). As the Court reiterated in

even clearer language: “[T]here is nothing in the record to

suggest that a second series of interviews with the already-

interviewed returnees would produce any relevant information

that is not set forth in the documents that have already been

produced.” Id. at 179; accord, e.g., Painting & Drywall Work

Preservation Fund v. Department of Housing & Urban

Development, 936 F.2d 1300, 1303 (D.C. Cir. 1991)

(concluding that the public interest in the disclosure of worker

records that would provide information on agency

enforcement efforts was minimal because interested parties

could obtain similar information through alternative means).

Perhaps, as the dissent suggests, the relevant privacy interest

in Ray was more substantial than here. See Dissenting Op. at

8. But FOIA requires us to balance privacy interests against

the benefits of disclosure, and the critical point for our

purposes is that in Ray the Court evaluated the latter by

examining the incremental effect of the information sought in

light of prior disclosures.

Consistent with the forgoing principles, and given the

unique way in which this case has evolved, I believe that the

public interest at issue here is less than it was when the case

was previously before us. Of course, there is little doubt that

“[t]he use of and justification for warrantless cell phone

tracking” continues to be a “topic of considerable public

interest.” ACLU I, 655 F.3d at 12; see, e.g., Kate Zernike,

Court Restricts Police Searches of Phone Data, N.Y. Times,

July 19, 2013, at A1 (describing differing positions taken by

courts on the legality and propriety of this investigatory

technique); Joe Palazzolo, Montana Requires Warrants for

Cell Phone Tracking, Wall St. J. L. Blog, June 21, 2013,

4

http://blogs.wsj.com/law/2013/06/21/montana-requires-

warrants-for-cell-phone-tracking (describing efforts by states

to require police to obtain a warrant in order to access cell

phone tracking information). The disclosure of these six cases

could also “shed[]” at least some additional “light on the

scope and effectiveness” of this practice. ACLU I, 655 F.3d

at 13.

But most of the benefit we anticipated from the release of

the requested docket information flowed from the fact that

access to a large sample of prosecutions would provide a basis

for the public to discern general trends regarding the

government’s use of cellphone tracking data and the means by

which the government obtains such data. For example, we

observed that disclosure would “provide information about

the kinds of crimes the government uses cell phone tracking

data to investigate,” the “standards the government uses to

justify warrantless tracking,” and “facts regarding the duration

of tracking and the quality of tracking data.” Id. at 13–14. As

a result of the district court’s and our own prior decisions,

however, the Department has already released docket

information for 214 prosecutions in which the government

obtained cell phone tracking data without a warrant, and those

214 cases presumably provide much of the necessary basis for

assessing when, how, and why the government utilizes this

particular investigative tool. Compare with id. at 14–15

(rejecting government’s argument that release of the

information was unnecessary due to the “extensive public

attention that this issue is already receiving,” because “much

of the information the plaintiffs seek to develop from the

FOIA disclosure . . . is not currently in the public domain”

(internal quotation marks omitted)). True, the six remaining

cases could contain some interesting anecdotal evidence: the

ACLU speculates that it is “more likely that a failed

prosecution involved” a “motion[] to suppress evidence

5

derived from cell phone tracking,” and that suppression

hearings are particularly likely to yield useful information.

Appellants’ Br. 33–34; see also ACLU I, 655 F.3d at 14

(describing information that could be derived from

suppression hearings). Even so, if the Department’s disclosure

of 214 prosecutions has failed to reveal the nature and extent

of the government’s practice of obtaining cell phone tracking

data without a warrant, the probability that disclosure of these

six remaining cases would yield significant benefits is

relatively low. To paraphrase the Supreme Court in Ray, there

is little to suggest that these six cases “would produce any

relevant information that is not set forth in the [214

prosecutions] that have already been produced.” 502 U.S.

at 179.

The ACLU argues that applying the “incremental value”

test in this fashion would give federal agencies license to

arbitrarily withhold portions of requested records—

presumably “the more important or embarrassing responsive

records”—on the ground that the public interest in disclosure

will be satiated by the records they choose to actually release.

Appellants’ Br. 35. Although I have no doubt that this court

would look with great suspicion on any attempt to manipulate

FOIA in this fashion, this case involves no such mischief. It is

well-established that federal agencies may disclose particular

records or portions of records responsive to a request without

disclosing all responsive records so long as they have some

legitimate FOIA-based reason for doing so. See 5 U.S.C.

§ 552(b) (“Any reasonably segregable portion of a record

shall be provided to any person requesting such record after

deletion of the portions which are exempt . . . .”);

Assassination Archives & Research Center v. CIA, 334 F.3d

55, 58 (D.C. Cir. 2003); see also, e.g., Ray, 502 U.S. at 178–

79; King, 830 F.2d at 234–35 (D.C. Cir. 1987) (permitting

government to release investigative report with portions

6

redacted in order to protect privacy interest of individuals

named). When, for example, an agency withholds certain

records that implicate greater privacy interests than those it

releases, and then evaluates the public benefit of releasing

these remaining records in light of the information already

released, it acts just as FOIA requires—efficiently trading off

privacy costs and disclosure benefits. That is almost exactly

what happened here. Having released all of the information

our prior decision required, the Department now resists

disclosure of a particular type of information that implicates

stronger privacy interests. That being so, I see no reason to

now disregard this prior disclosure. Just as we would certainly

take account of the existence of the docket information for

these 214 cases had it been uncovered and published by the

Washington Post, we may take account of it here even though

its release resulted from this litigation.

BROWN, Circuit Judge, dissenting: While I sympathize

with the court’s protective instincts, I subscribe to Lady

Macbeth’s drear insight: “What’s done cannot be undone.”

Redemption is still possible, but in the modern world, the

right to be left alone, once forfeited, is gone for good. An

individual who is indicted and tried has no privacy interest

that can protect the public record of prosecution from

disclosure—even if the ultimate outcome was acquittal or

dismissal. The residual privacy concerns we identified in

ACLU I are insufficient to meet the Exemption 7(C)

threshold. There we noted that the privacy right at common

law rested in large part on the “degree of dissemination,” and

that “interests in privacy fade” when the information is

already part of the public record and is readily available. Am.

Civil Liberties Union v. U.S. Dep’t of Justice (ACLU I), 655

F.3d 1, 9 (D.C. Cir. 2011). Because the privacy interest here

started small and the pace of technology continues to diminish

it, I respectfully dissent.

At the outset, I should note the court does get one thing

right. As a general matter, judges tasked with balancing

equally metaphysical concepts, like privacy and the public

interest, face what are among the most difficult and largely

standardless endeavors. See Reporters Comm. for Freedom of

Press v. U.S. Dep’t of Justice, 816 F.2d 730, 741 (D.C. Cir.

1987) (expressing doubt that there is any principled basis for

federal judges to make such ad hoc and idiosyncratic

determinations). Even so, I agree the balancing in this case is

relatively clear. The six disputed records already exist in the

public domain. Indeed, the court acknowledges the records

are accessible via a simple Google search or through PACER.

Furthermore, the court correctly determines the public interest

in disclosure is no more or less than it was in ACLU I. There,

we characterized the public interest as “significant.” ACLU I,

655 F.3d at 12. With one arm of the balance thus weighted,

the only question is whether the privacy interests of

unconvicted persons tip the scales against disclosure. The

2

court today holds that the contest, while close, is nevertheless

convincingly won by a supposedly more substantial privacy

interest. I am not persuaded. On balance, the permanence

and accessibility of the records render any privacy interests

only marginally greater than they were in ACLU I, thus

tipping the balance in favor of disclosure.

The majority’s privacy analysis rests on two pillars: the

presumption of innocence and the common law of

informational privacy. Both notions have shortcomings.

First, the presumption of innocence is an artifact of the

common law’s adversarial approach to the question of guilt.

What authority exists for the proposition that the presumption

of innocence affords indicted, but unconvicted, persons some

measure of informational privacy? The Supreme Court has

made it clear that the presumption of innocence applies only

to a criminal trial and, within the trial, only to the jury or other

trier of fact. See Bell v. Wolfish, 441 U.S. 520, 533 (1979)

(“The presumption of innocence is a doctrine that allocates

the burden of proof in criminal trials; it also may serve as an

admonishment to the jury to judge an accused’s guilt or

innocence solely on the evidence adduced at trial . . . .”

(emphasis added)). Any other, extra-trial reference to the

doctrine is both imprecise and impotent. If, as the Supreme

Court posits, the presumption of innocence is purely an

instrument for allocating the burden of proof at trial and

warning jurors against drawing untoward inferences, then

there is no basis for supposing the presumption of innocence

governs events beyond the trial itself. Contra Majority Op. at

10 (“Individuals who are charged with a crime and ultimately

prevail of course remain entitled to a version of this

presumption [of innocence].”). And why would a common

law presumption trump FOIA’s statutory mandate?

3

The court hypothesizes the plight of individuals who,

though never convicted, are viewed with suspicion when

others learn of their mere involvement in particularly ignoble

cases. See Majority Op. at 13. But even if true, persons who

are publicly indicted and tried can have no reasonable

expectation that the occurrence of these events will not be

publicly disclosed. Risk of disclosure inheres in the very

nature of these public proceedings. See Craig v. Harney, 331

U.S. 367, 374 (1947) (“A trial is a public event. What

transpires in the court room is public property.”). To be sure,

we previously discounted the small but nonetheless real risk

of renewed attention, dismissing such concerns as sheer

“speculation.” See ACLU I, 655 F.3d at 10–11 (“Such a list

[of publicly indicted persons] is surely less likely to draw

attention to a name than was the initial press coverage of an

indictment . . . .”).

Furthermore, what of the need for an informed citizenry

to hold public officials accountable? One “purpose of FOIA

is to permit the public to decide for itself whether government

action is proper.” Wash. Post Co. v. U.S. Dep’t of Health &

Human Servs., 690 F.2d 252, 264 (D.C. Cir. 1984) (emphasis

added); see also Richard A. Epstein, Privacy, Publication,

and the First Amendment: The Dangers of First Amendment

Exceptionalism, 52 STAN. L. REV. 1003, 1004, 1047 (2000)

(discussing as a counterweight to privacy goals, the social

ideal of full disclosure of information about others to allow

individuals to make “full and informed decisions on matters

of great importance”); Sadiq Reza, Privacy and the Criminal

Arrestee or Suspect: In Search of a Right, in Need of a Rule,

64 MD. L. REV. 755, 807 (2005) (“The government should

arguably inform the public about its suspicions regarding an

arrestee or suspect so that people may practice ‘informed

living,’ the right to exercise an informed choice about those

with whom they live and associate. That is, X should have

4

access to information that Y has been arrested for or

suspected of a crime so that X can decide intelligently

whether to socialize with Y, let her children play with Y’s

children, patronize Y’s business, or use Y’s professional

services, and so forth.”). From the point of view of the

wrongfully accused, this will be a continuing injustice, but a

person can be found not guilty and still not be innocent of the

crime charged. See Rigsbee v. United States, 204 F.2d 70,

72–73 (D.C. Cir. 1953) (holding that an acquittal differs from

innocence and that the former would be insufficient by itself

to obtain a certificate signifying the latter). The rough

balance courts must strike can never resolve such anomalies.

The Court’s reliance on common law informational

privacy doctrine is similarly unavailing. “[B]oth the common

law and the literal understandings of privacy encompass the

individual’s control of [personal] information.” Reporters

Comm., 489 U.S. at 763. The touchstone of informational

privacy—the right to be let alone—has long rested on the

degree to which an allegedly private fact has been

disseminated, and the extent to which the passage of time has

rendered it private. Id. Nevertheless, technological advances

seem to presage the death knell for this previously workable

standard. In today’s echo chamber of big data, metadata, and

the Internet, the once wholly forgotten memory of some

unsavory, minimally broadcast misdeed is resurrected for

global consumption. Against this backdrop, it seems fanciful

to believe that individuals who were publicly indicted but

never convicted (though in some cases publicly tried), retain

an objective, substantial privacy interest in controlling

information about these public facts.

The court says unconvicted persons are “entitled to move

on with their lives without having the public reminded of their

alleged but never proven transgressions.” Majority Op. at 11.

5

Alas, Google, unlike God, neither forgets nor forgives.1

Indeed, Google is not alone in its uncanny ability to keep the

proverbial score. It is true that most jurisdictions treat

aggregations of data confidentially, but they also insist on

transparency for records of individual cases. Courts, too,

have a penchant for reminding acquitted individuals of their

“alleged but never proven transgressions.” See, e.g., Dowling

v. United States, 493 U.S. 342, 354 (1990) (holding that

admission of evidence relating to a crime the defendant had

previously been acquitted of committing did not violate

double jeopardy or due process); United States v. One

Assortment of 89 Firearms, 465 U.S. 354, 356–66 (1984)

(holding that a gun owner’s acquittal on criminal charges

involving firearms did not preclude a subsequent in rem

forfeiture proceeding against those firearms under 18 U.S.C §

924(d)); United States v. Foster, 19 F.3d 1452, 1455 (D.C.

Cir. 1994) (noting that virtually every circuit permits

enhancement of sentence based on acquitted conduct).

The proposition that “an ‘ordinary citizen’ has a privacy

interest ‘in the aspects of his or her criminal history that may

have been wholly forgotten,’” Majority Op. at 13 (citing

Reporters Committee, 489 U.S. at 769), is thus inapt. Thanks

to the Internet (for better or worse), information that was once

scattered, localized, and forgotten with the passage of time is

now effectively permanent and searchable. And though one

might wish quietly to melt into the shadow of obscurity, the

inexorable march of time is simply no match for the

unflagging, unforgiving memory that is the World Wide Web.

1

There are exceptions, of course, but records

memorializing a public indictment and trial do not appear to

be one of them. See Removal Policies, GOOGLE,

https://support.google.com/websearch/answer/2744324#offen

siveimages (last visited Apr. 23, 2014).

6

Once a secret is disclosed online, neither the courts nor

society may unring the lingering echo of the bell. In this

respect, Reporters Committee is an anachronism. The aspects

of an “ordinary citizen[’s]” criminal history the Court thought

would be wholly forgotten were the data contained in rap

sheets, which were maintained in a localized computer

database. See Reporters Comm., 489 U.S. at 752, 771; see

also ACLU I, 655 F.3d at 8. Nowadays, bits and pieces of

data are aggregated and immortalized on public and private

systems, and the private systems have no purge schedules.

This is not to say the modern man has abdicated any

expectation of privacy in facts partially disclosed. As the

Supreme Court observed, “the fact that an event is not wholly

private does not mean that an individual has no interest in

limiting disclosure or dissemination of the information.”

Reporters Comm., 489 U.S. at 770. But there is a vast chasm

between facts disclosed to a discrete group that otherwise

treats the information as private and facts that are

unqualifiedly revealed and accessible to virtually everyone.

In my view, the case before us falls into the latter camp. Cf.

id. at 752 (“As a matter of executive policy, [DOJ] has

generally treated rap sheets as confidential and, with certain

exceptions, has restricted their use to governmental

purposes.”); Dep’t of Air Force v. Rose, 425 U.S. 352, 359–60

(noting that the Academy treated “all matters discussed” at

hearings for honor code violations as “confidential,” marked

case summaries “for official use only,” and instructed cadets

“not to read the case summary unless they have a need,

beyond mere curiosity, to know their contents”).

Considering the fissures in the two pillars supporting the

court’s privacy analysis, one would expect the privacy

interests to become less significant. At the very least, these

serious deficits ought give way to the court’s obligation to

7

“make a reasonable effort to account for the death of a person

on whose behalf the [agency] invokes Exemption 7(C).

Schrecker v. U.S. Dep’t of Justice, 349 F.3d 657, 662 (D.C.

Cir. 2003). This would include the deaths of family members.

After all, “death clearly matters, as the deceased by definition

cannot personally suffer the privacy-related injuries that may

plague the living.” Campbell v. U.S. Dep’t of Justice, 164

F.3d 20, 33 (D.C. Cir. 1998). I am not swayed by the

majority’s contention that reputational interests are enough to

carry the day. The posthumous reputational interest the

Supreme Court recognized in Swidler & Berlin v. United

States, 524 U.S. 399 (1998), is the one rooted in the venerable

attorney-client privilege, not informational privacy. See id.

(holding that the attorney-client privilege survives a client’s

death).

One last point warrants discussion. Judge Tatel’s

concurrence seeks to lend credence to DOJ’s invocation of the

incremental value test—a test allegedly of precedential value.

I am not so certain. First, the test is of dubious provenance.

In King v. U.S. Dep’t of Justice, 830 F.2d 210, 234 (D.C. Cir.

1987), the case cited by Schrecker as authority for its

statement about the incremental value test, the court never

actually used the words “incremental value.” Instead, the

court merely held that, because the appellant failed to

demonstrate how disclosing the redacted names was relevant

to the public interest, the privacy interests “outweighed any

public interest attending disclosure.” See id. at 234–35. That

is all. King did not hold that the incremental value of

information depends on the mix of data already publicly

available.

In the 200-plus FOIA cases since the Schrecker decision,

we have referenced the incremental value test only three

times. In each instance, we have understood it to mean

8

exactly the opposite of what the concurrence posits: “even if

the ‘absolute value’ of the requested information is small, it

must nevertheless be released if it adds any incremental value

of public interest.” Appellants’ Reply Br. at 15; see ACLU I,

655 F.3d at 15 (rejecting DOJ’s “incremental contribution”

argument because “[t]he fact that the public already has some

information does not mean that more will not advance the

public interest” (emphasis added)); Lardner v. U.S. Dep’t of

Justice, 398 F. App’x 609, 611 (D.C. Cir. 2010) (affirming

the district court’s decision to disclose the identities of denied

pardon and commutation applicants despite the previous

disclosure and existence of approved applicants’ identities on

the public record. Significantly, the court noted: “The

incremental value of the withheld information is not

speculative . . . .”); Consumers’ Checkbook Ctr. for the Study

of Servs. v. U.S. Dep’t of Health & Human Servs., 554 F.3d

1046, 1060 (D.C. Cir. 2009) (Rogers, J., concurring in part

and dissenting in part) (“[E]ven though the requested data will

only partially reveal physicians’ experience levels, the data

has ‘incremental value’ for ascertaining the quality of services

performed.”).

In any event, even assuming the court is bound by the

version of the incremental value test Judge Tatel espouses, the

cases cited in support of this test are all distinguishable for

one reason or another. In Ray, for example, the privacy

interests were more significant than those implicated here.

The information redacted from the disputed records was

obtained via interviews with requested Haitians who were

promised confidentiality. U.S. Dep’t of State v. Ray, 502 U.S.

164, 172 (1991). In other words, the parties agreed to treat

the information obtained as private. Furthermore, the Court’s

conclusion that “[t]he addition of the redacted identifying

information would not shed any additional light on the

Government’s conduct of its obligation,” id. at 178 (emphasis

9

added), is in accord with this court’s prior application of the

test. The redacted information was withheld precisely

because the Court recognized that, “in and of itself” it would

not “tell respondents anything about Haiti’s treatment of the

returnees or this Government’s honesty.” Id. at 170–71

(emphasis added).

Perhaps most importantly, however, Ray involved

redacted information, not wholly undisclosed records. The

difference is not merely academic. Judge Tatel’s version of

the incremental value test would make little sense where, as

here, a court is dealing with undisclosed records that are

substantively dissimilar to records previously disclosed.

Unlike Ray, where the redacted information was sought so

that interviews with Haitians could be conducted anew, 502

U.S. at 178–79, disclosing the records of unconvicted persons

would be neither duplicative nor speculative. It is reasonable

to believe the six files could contain new information

precisely because the records sought—unlike the Haitian

interviewees—are qualitatively different. In fact, Judge Tatel

agrees. Concurring Op. at 4 (“The disclosure of these six

cases could also shed at least some additional light on the

scope and effectiveness of [warrantless cell phone

tracking].”). Nothing more is required.

At bottom, the public interest in disclosure remains as

robust as it was in ACLU I. Conversely, in the Internet age,

privacy is no longer what it once was. Times have changed,

and so, too, must our expectations. I respectfully dissent.