PRECEDENTIAL UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT _______________ No. 22-1970 _______________ UNITED STATES OF AMERICA v. JOSEPH R. JOHNSON, JR., Appellant _______________ On Appeal from the United States District Court for the Eastern District of Pennsylvania (D.C. No. 2:19-cv-00367-001) District Judge: Honorable Harvey Bartle, III _______________ Argued: April 10, 2024 Before: CHAGARES, Chief Judge, PORTER and SCIRICA, Circuit Judges (Filed: August 21, 2024) Tadhg Dooley David R. Roth [Argued] WIGGIN & DANA One Century Tower Church Street New Haven, CT 06510 Lydia Laramore [Argued] Evan Lisman [Argued] Laila Robbins YALE LAW SCHOOL Wall Street New Haven, CT 06511 Counsel for Court-Appointed Amici Curiae Linwood C. Wright, Jr. [Argued] OFFICE OF UNITED STATES ATTORNEY Chestnut Street Suite 1250 Philadelphia, PA 19106 Counsel for Appellee _______________ OPINION OF THE COURT _______________ PORTER, Circuit Judge.

Before we vacated Joseph Johnson’s criminal conviction and directed his acquittal, he spent fifteen months in federal prison. He now seeks compensation as a “person unjustly con-

victed of an offense against the United States and imprisoned.” 28 U.S.C. § 1495. But the District Court found that Johnson could not prove that “he did not by misconduct or neglect cause or bring about his own prosecution[,]” which is a requirement for receiving compensation under § 1495. 28 U.S.C. § 2513(a)(2).

We will affirm. Johnson committed “misconduct” by using a lawyer’s signature without her consent to file an exhibit in federal court, which was a but-for “cause” of the government’s decision to “prosecut[e]” him. If he had not filed the exhibit, the government would not have prosecuted him. He therefore cannot satisfy the requirements for compensation under § 2513(a).

I. FACTS AND PROCEDURAL BACKGROUND A plaintiff sued Bill Cosby for sexual assault in the United States District Court for the Eastern District of Pennsylvania.

Johnson, who was not involved with the case, filed an exhibit using a copy of the plaintiff’s lawyer’s signature without her consent. The Clerk’s Office added the exhibit to the docket.

The exhibit contained a document accusing the plaintiff of underreporting her taxable income. The plaintiff’s lawyer rec- ognized the exhibit as fraudulent, and the presiding judge (the “Judge”) quickly struck it from the docket upon the lawyer’s request.

The government prosecuted Johnson for making a false statement under 18 U.S.C. § 1001 and aggravated identity theft

under 18 U.S.C. § 1028A.1 To convict Johnson for making a false statement under § 1001, the government was required to prove: “(1) that [Johnson] made a statement or representation; (2) that the statement or representation was false; (3) that the false statement was made knowingly and willfully; (4) that the statement or representation was material; and (5) that the state- ment or representation was made in a matter within the juris- diction of the federal government.” United States v. Moyer, 674 F.3d 192, 213 (3d Cir. 2012) (emphasis added). To convict Johnson for aggravated identity theft under § 1028A, the gov- ernment was required to prove that Johnson made a false state- ment under § 1001. So for both counts, the government was required to prove the five elements articulated in Moyer, including the materiality of Johnson’s false statement.

A jury convicted Johnson on both counts, but we over- turned his conviction on direct appeal. See United States v. Johnson, 19 F.4th 248, 252 (3d Cir. 2021). On materiality, the government was required to prove that Johnson’s false state- ment—using the lawyer’s signature without her consent—was “of a type capable of influencing a reasonable decisionmaker.”

United States v. McBane, 433 F.3d 344, 351 (3d Cir. 2005). At trial, the government’s theory was that “the Judge alone was the governmental decisionmaker.” Johnson, 19 F.4th at 261.

But the government “failed to identify a single decision entrusted to the Judge . . . that could have been influenced by” Johnson’s false statement. Id. at 258. For example, the govern- ment did not show “that the Judge would need to make a cred- ibility determination as to [the plaintiff], to which the [false

Each count of the indictment also charged Johnson with aid- ing and abetting the commission of the primary offense under 18 U.S.C. § 2.

statement] arguably could have been relevant.” Id. at 257 n.9.

Because the government failed to prove that Johnson’s false statement was material, we vacated his conviction and directed his acquittal. Id. at 263–64.

Before we directed his acquittal, Johnson spent more than fifteen months in prison. After his release, he sought compen- sation from the government under 28 U.S.C. § 1495, for which he is required to obtain a “certificate” of his innocence, 28 U.S.C. § 2513(b). He petitioned for a certificate under his orig- inal criminal docket number. The District Court denied his petition. It found that Johnson had not proved that “he did not by misconduct or neglect cause or bring about his own prose- cution[,]” which is a requirement for obtaining a certificate of innocence. § 2513(a)(2).

Johnson appealed. We appointed David R. Roth and Tadhg Dooley as Amici Curiae to submit briefs regarding Johnson’s entitlement to a certificate of innocence.2 II. JURISDICTION AND STANDARD OF REVIEW The District Court had jurisdiction over Johnson’s criminal case under 18 U.S.C. § 3231. Johnson was permitted to petition for a certificate of innocence in the District Court—“the court” where “the requisite facts” for a certificate “are alleged to appear”—under 28 U.S.C. § 2513(b). See Abu-Shawish v. United States, 898 F.3d 726, 736 (7th Cir. 2018). We have Amici were assisted on their briefs by student members of the Yale Law School Advanced Appellate Litigation Project, two of whom presented oral argument. Amici and their students discharged their duties admirably. We thank them for their excellent oral and written advocacy.

appellate jurisdiction because the District Court’s denial of Johnson’s petition was a “final decision[]” under 28 U.S.C. § 1291. Cf. United States v. Rodriguez, 855 F.3d 526, 531 (3d Cir. 2017) (recognizing that, in the context of “sentencing judgments,” district court decisions are “final” if “they close . . . criminal cases once again” (quoting United States v. Jones, 846 F.3d 366, 369 (D.C. Cir. 2017))).

Several courts have stated that a district court’s denial of a certificate of innocence is reviewed for abuse of discretion.

See, e.g., United States v. Davis, 16 F.4th 1192, 1193 (5th Cir. 2021). This would differ from our typical standard of review in civil appeals, for which “we review a district court’s findings of fact for clear error and its conclusions of law de novo.”

McCutcheon v. Am.’s Servicing Co., 560 F.3d 143, 147 (3d Cir. 2009); see also Abu-Shawish, 898 F.3d at 731 (describing a § 2513 petition as “a new civil case embedded within a closed criminal case”). But to resolve Johnson’s appeal, we need not decide which standard of review is generally applicable to § 2513 appeals. This appeal turns on the correct interpretation of § 2513(a)(2), which is a pure question of law. The District Court necessarily abused its discretion if it interpreted § 2513(a)(2) incorrectly, so we review its interpretation de novo. See In re Bayer AG, 146 F.3d 188, 191 (3d Cir. 1998) (“Where the district court misinterpreted or misapplied the law . . . our review is plenary.”).

III. DISCUSSION To obtain a “certificate” of his innocence, Johnson must “allege and prove” the following: (1) His conviction has been reversed or set aside on the ground that he is not guilty of the offense

of which he was convicted, or on new trial or rehearing he was found not guilty of such offense, as appears from the record or certificate of the court setting aside or reversing such con- viction, or that he has been pardoned upon the stated ground of innocence and unjust conviction and (2) He did not commit any of the acts charged or his acts, deeds, or omissions in connection with such charge constituted no offense against the United States, or any State, Territory or the District of Columbia, and he did not by miscon- duct or neglect cause or bring about his own prosecution.

§ 2513(a). Courts break these provisions into three elements, each of which is required to obtain a certificate of innocence.

See, e.g., United States v. Moon, 31 F.4th 259, 262 (4th Cir. 2022). Only the third requirement is contested in this appeal.

But to understand how the third requirement works, it is nec- essary to canvas all three.

First, under § 2513(a)(1), a petitioner must show that his conviction was reversed based on his innocence, or that he was acquitted in any new trial or rehearing, or that he was pardoned for his innocence. Essentially, this element requires the peti- tioner to show that his conviction was vacated based on his innocence of the charged offense, not for reasons “unrelated to his culpability.” Betts v. United States, 10 F.3d 1278, 1284 (7th Cir. 1993).

Second, under the first clause of § 2513(a)(2), a petitioner must show that “[h]e did not commit any of the acts charged or

[that] his acts, deeds, or omissions in connection with such charge” did not constitute any crimes. Because this require- ment is disjunctive, it may be satisfied in two independent ways. The first option is satisfied “only in cases of mistaken identity or the like, when the petitioner simply did none of the acts charged in an indictment.” Amicus Br. 28. The second option is satisfied if the petitioner’s conduct does not satisfy the elements of any crime, regardless of whether it was charged. See United States v. Racing Servs., Inc., 580 F.3d 710, 712–13 (8th Cir. 2009).

Finally, under the second clause of § 2513(a)(2), Johnson must show that “he did not by misconduct or neglect cause or bring about his own prosecution.”

A. Johnson Caused His Own Prosecution By Miscon- duct.

To interpret § 2513(a)’s third requirement, we begin with the text. See Ross v. Blake, 578 U.S. 632, 638 (2016) (“Statu- tory interpretation . . . begins with the text.”). Under a straight- forward reading of the text, Johnson must prove a negative to satisfy the third requirement. He must show that he did not commit “misconduct or neglect” that “cause[d] or br[ought] about” the government’s decision to “prosecut[e]” him.

§ 2513(a)(2). Our analysis begins and ends with the ordinary meaning of “cause or bring about.”

1. Factual Causation Differs From Proximate Cau- sation.

The phrase “cause or bring about” refers to a causal rela- tionship between a petitioner’s “misconduct or neglect” and his “prosecution.” Id. To interpret statutory language that refers to

causation, courts consider the standards for causal relation- ships in other legal contexts, such as tort law and criminal law.

See Burrage v. United States, 571 U.S. 204, 210 (2014).

Broadly speaking, tort law and criminal law distinguish between two concepts of causation: “actual” or “factual” cau- sation, and “legal” or “proximate” causation. See id. (“The law has long considered causation a hybrid concept, consisting of two constituent parts: actual cause and legal cause.”); H.L.A.

Hart & Tony Honoré, Causation in the Law 110 (2d ed. 1985) (describing the law’s “bifurcation of causal questions”).

Factual causation entails “an ordinary, matter-of-fact inquiry into the existence . . . of a causal relation as laypeople would view it.” Paroline v. United States, 572 U.S. 434, 444 (2014) (ellipsis in original) (internal quotation marks and quoted source omitted). Often, courts equate factual causation with “but-for” causation. See Univ. of Tex. Sw. Med. Ctr. v. Nassar, 570 U.S. 338, 346–47 (2013) (“In the usual course,” factual causation “requires the plaintiff to show that the harm would not have occurred in the absence of—that is, but for— the defendant’s conduct.” (internal quotation marks and quoted source omitted)). But not always. In the tort context, “[i]f multiple acts occur, each of which . . . alone would have . . . cause[d] . . . [a] physical harm at the same time in the absence of the other act(s), each act is regarded as a factual cause of the harm.” Restatement (Third) of Torts: Liab. for Phys. & Emot.

Harm § 27 (Am. L. Inst. 2010). In such circumstances, an act may factually cause an injury even if the injury would have occurred absent the act. Id. § 27 cmt. a. More broadly, courts have considered alternatives to but-for causation as standards for factual causation, including whether an act is a “contributing,” “substantial,” or “sole” factor in producing an injury. See James A. Macleod, Ordinary Causation: A Study in

Experimental Statutory Interpretation, 94 Ind. L.J. 957, 974– (2019) (canvasing these alternatives).

In contrast, “the phrase ‘proximate cause’ is shorthand for the policy-based judgment that not all factual causes contrib- uting to an injury should be legally cognizable causes.” CSX Transp., Inc. v. McBride, 564 U.S. 685, 701 (2011). Because an injury may have infinite factual causes, “courts and legisla- tures” use proximate cause principles to “place limits on the chain of causation that may support recovery on any particular claim.” Id. The phrase “proximate cause” is “notoriously con- fusing” because there is no “consensus on any one definition.”

Id. (listing various “[c]ommon-law formulations” for limiting liability to a subset of factual causes). Regardless, courts some- times read statutory causal language as incorporating proxi- mate causation principles. See, e.g., Staub v. Proctor Hosp., 562 U.S. 411, 419–20 (2011); see also Sandra F. Sperino, Statutory Proximate Cause, 88 Notre Dame L. Rev. 1199, 1218 n.79 (2013) (collecting cases).

With this clarification, we must interpret § 2513(a)(2)’s causal language in two steps. First, we must determine the cor- rect standard for factual causation under § 2513(a)(2)—the ordinary meaning of “caus[ing] or bring[ing] about [one’s] own prosecution.” Second, we must determine whether § 2513(a)(2) incorporates proximate causation principles. At both steps, we must be attentive to statutory context. “When a statutory provision includes an undefined causation require- ment, we look to context to decide whether the statute demands only but-for cause as opposed to proximate cause or sole cause.” Husted v. A. Philip Randolph Inst., 584 U.S. 756, 769 (2018).

2. Factual Causation Under § 2513(a)(2) Equals But- For Causation.

Amici did not distinguish between factual and proximate causation in their briefs. But they seem to argue that Johnson’s false statement did not factually cause the government to pros- ecute him, based on their “commonsense interpretation” of § 2513(a)(2)’s “ordinary meaning.” Amicus Br. 35; see Paroline, 572 U.S. at 444; Macleod, supra, at 982 (“Many commentators . . . treat[] the concept of factual causation in law as a matter of ‘common sense.’”). Amici note that “[o]nly the Government can begin a prosecution, and it can do so only when Government attorneys have a good-faith belief [that] each element of the charged offense is met.” Amicus Br. 3.

They argue that Johnson’s false statement did not cause gov- ernment attorneys to have a good-faith belief that each element of §§ 1001 and 1028A was satisfied because “the Government identified nothing Johnson ever did that somehow caused it to erroneously conclude his [false] statement was material.” Id. at 59. Thus, Amici conclude that Johnson satisfies the third requirement of § 2513(a). Id. at 38 (“[A] person ‘by miscon- duct or neglect cause[s]’ the federal government to initiate a criminal prosecution only if the person intentionally or negli- gently makes the Government believe that each element of the offense is satisfied.” (emphasis added)).3 Amici’s interpreta-

Amici describe § 2513(a)’s third requirement as embedding an “estoppel principle” because a petitioner fails to satisfy it only if his deceitful conduct is responsible for the govern- ment’s good-faith belief that each element of the charged crime is satisfied. Amicus Br. 36; see Estoppel, Black’s Law Dictionary (9th ed. 2009) (“An affirmative defense alleging good-faith reliance on a misleading representation.”).

tion finds support in Betts, where the Seventh Circuit held that a petitioner fails § 2513(a)’s third requirement only if he “act[s] or fail[s] to act in such a way as to mislead the authori- ties into thinking he . . . committed an offense.” 10 F.3d at 1285.

Amici’s argument fails because it does not reflect the ordi- nary meaning of factually “caus[ing]” a “prosecution.”

§ 2513(a)(2). In recent years, the Supreme Court has consist- ently interpreted statutory causal language as denoting but-for causation. See Nassar, 570 U.S. at 346–47; Burrage, 571 U.S. at 210–11 (defining the “ordinary meaning” and “traditional understanding” of factual causation as but-for causation); Comcast Corp. v. Nat’l Ass’n of Afr. Am.-Owned Media, 589 U.S. 327, 332 (2020) (equating factual causation under the Civil Rights Act of 1866 with but-for causation because the “ancient and simple ‘but for’ common law causation test . . . supplies the default or background rule against which Congress is normally presumed to have legislated when creating its own new causes of action” (internal quotation marks and quoted source omitted)). The Court’s consistency in equating factual causation with but-for causation has led one commentator to describe it as a canon of interpretation. See Sandra F. Sperino, The Causation Canon, 108 Iowa L. Rev. 703, 704 (2023) (“When a statute uses any language that might relate to factual cause, the Court will assume that Congress meant to require the plaintiff to establish ‘but-for’ cause.”).

Amici’s argument to the contrary depends on a different form of factual causation than the “ancient and simple” but-for.

Comcast, 589 U.S. at 332. Amici interpret the third require- ment as embedding a standard that approximates sole causa- tion. Under their theory, a petitioner fails § 2513(a)’s third requirement only if his misconduct was the sole cause of the

government’s belief that each element of the charged crime is satisfied. That is, if the government mistakenly believes that an element of the charged crime is satisfied for reasons other than the petitioner’s misconduct—like the government’s independ- ent misinterpretation of law—the third requirement is satis- fied.4 The Supreme Court sometimes interprets statutory causal language as incorporating a different kind of factual causation than but-for, but only if statutory context requires it. In Husted, the Court interpreted causal language in the National Voter Registration Act (“NVRA”). 584 U.S. at 768. Under the NVRA, state policies “shall not result in the removal of the name of any person from the official list of voters registered to vote . . . by reason of the person’s failure to vote[.]” 52 U.S.C. § 20507(b)(2) (emphasis added). The Court interpreted this provision as “forbid[ding] the use of nonvoting as the sole cri- terion for removing a registrant,” not merely as a but-for cause of removal. Husted, 584 U.S. at 768. It chose this interpretation because of the NVRA’s statutory context. A separate provision of the NVRA allows “removal if a registrant did not send back a return card and also failed to vote,” so nonvoting was a per- missible reason for removal if accompanied by another permis- sible reason. Id. And Congress clarified § 20507(b)(2) in a

Technically, Amici’s theory does not require sole causation because the government must exercise its discretion to prose- cute, even if it reasonably believes that each element of a crime is satisfied. See United States v. Nixon, 418 U.S. 683, 693 (1974) (referring to the government’s “absolute discretion to decide whether to prosecute a case”). But regardless of its char- acterization, Amici’s theory depends on a standard for factual causation that is stricter than but-for causation.

third provision, stating that “no registrant may be removed solely by reason of a failure to vote.” § 21083(a)(4)(A) (emphasis added). Combining these provisions, the Court con- cluded that the causal language in § 20507(b)(2)—“by reason of”—requires sole causation, not mere but-for causation. See Husted, 584 U.S. at 769.

Amici do not point to any similar contextual evidence in § 2513 to support their strict theory of factual causation.5 Instead, adopting their theory would require us to rewrite the statute. See Jeffrey S. Gutman, Are Federal Exonerees Paid?: Lessons for the Drafting and Interpretation of Wrongful Conviction Compensation Statutes, 69 Clev. St. L. Rev. 219, (2021) (“The statute does not by its terms qualify the term ‘prosecution’ with words like ‘fair,’ ‘just,’ ‘proper,’ or ‘law- As we previously mentioned, Amici’s theory coheres with the Seventh Circuit’s decision in Betts v. United States, 10 F.3d 1278 (1993). But we do not find Betts textually persuasive. The Seventh Circuit considered the possibility that § 2513(a)(2)’s causal language denotes but-for causation: “In a moral sense, perhaps, a person who engages in conduct that a prosecutor . . . mistakenly believes to constitute a crim[e] . . . might be said to have ‘brought about’ his own prosecution, on the theory that he would not have been charged had he comported himself in a more upstanding fashion.” Betts, 10 F.3d at 1285. It rejected this interpretation purely on policy grounds, hesitating to “require courts to assess the virtue of a petitioner’s behavior even when it does not amount to a criminal offense.” Id. But such “policy concerns cannot trump the best interpretation of the statutory text,” so we decline to follow Betts. Patel v. Garland, 596 U.S. 328, 346 (2022).

ful.’”). The statute requires Johnson to prove that his miscon- duct did not cause his prosecution, full stop—not that his mis- conduct did not cause the government to prosecute him fairly, reasonably, or lawfully. Absent contextual evidence favoring the latter interpretation, we interpret § 2513(a)(2) as equating factual causation with but-for causation, consistent with Supreme Court precedent regarding factual causation.6 We thus align ourselves with the Fourth Circuit, in conflict with the Seventh. Compare Moon, 31 F.4th at 266 (considering whether a petitioner’s “misconduct . . . was a but-for cause of his conviction”), with Betts, 10 F.3d at 1285.

If anything, the ordinary meaning of causal language is more permissive than but-for causation, not stricter like Amici’s interpretation of § 2513(a)(2). See James A. Macleod, Ordinary Causation: A Study in Experimental Statutory Interpretation, 94 Ind. L.J. 957, 1006 (2019) (conducting a sur- vey and concluding that “a clear majority” of respondents deemed statutory causal language satisfied even “absent but- for causation”); Sandra F. Sperino, The Causation Canon, 108 Iowa L. Rev. 703, 707 (2023) (critiquing the Supreme Court’s tendency to equate factual causation with but-for causation because “courts [sometimes] relax the standard for proving factual cause from ‘but-for’ to a looser ‘substantial factor’ standard”). As a result, departing from Supreme Court prece- dent regarding factual causation would favor the government, not Johnson. But because the government interprets § 2513(a)(2) as denoting but-for causation, we have not been asked to consider whether a looser standard is appropriate. See Oral Arg. Tr. 18:16–17 (“I think the best reading is the plain reading, and that’s . . . but for.”).

3. Section 2513(a)(2) Does Not Incorporate Proxi- mate Causation Principles.

Amici did not mention proximate causation in their briefs, but they contended at oral argument that § 2513(a)(2) incorpo- rates proximate causation principles. See Oral Arg. Tr. 5:22– (“If this Court wanted to consider elementary principles of tort causation in the common law, we believe our interpretation fits with a pproximate [sic] cause standard.”). Amici described the government’s mistake regarding the materiality of Johnson’s false statement as a “superseding cause of the result- ing injury . . . that cuts off the causal chain from” Johnson’s misconduct. Id. 6:1–3. Under Amici’s view, Johnson satisfies § 2513(a)’s third requirement because his misconduct did not proximately cause his prosecution, regardless of whether his misconduct factually caused his prosecution.

We disagree because § 2513’s “context” does not suggest that it incorporates proximate causation principles. Husted, 584 U.S. at 769. Courts typically apply proximate causation princi- ples to statutes that condition remedies on plaintiffs showing that defendants caused their injuries by unlawful conduct. See, e.g., Apple Inc. v. Pepper, 587 U.S. 273, 279 (2019) (applying “principles of proximate cause” to limit recovery under the Clayton Act for injuries caused by antitrust violations); Lexmark Int’l, Inc. v. Static Control Components, Inc., 572 U.S. 118, 132 (2014) (“incorporat[ing] a requirement of prox- imate causation” into a private cause of action for Lanham Act violations); Holmes v. Sec. Inv. Prot. Corp., 503 U.S. 258, 267–68 (1992) (limiting civil recovery for Racketeer Influence and Corrupt Organizations Act (“RICO”) violations to proxi- mately injured plaintiffs). The “premise” underlying these cases is that “when Congress creates a federal tort” or cause of action, “it adopts the background of general tort law[,]” includ-

ing proximate causation principles. Staub, 562 U.S. at 417. For similar reasons, courts apply principles of proximate causation to suits brought for constitutional violations under Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388 (1971). See Egervary v. Young, 366 F.3d 238, 246 (3d Cir. 2004) (proclaiming as a “self-evident principle that . . . tort law causation must govern” Bivens claims because they are analogous to “any tort case”).

The “premise” underlying these cases does not justify incorporating proximate cause principles into § 2513, which does not create a cause of action mirroring common-law tort remedies. Unlike a common-law tort, § 2513 does not require Johnson to prove that the government negligently or intention- ally caused his injury—his allegedly “unjust” prosecution and imprisonment. Instead, it requires Johnson to “allege and prove,” § 2513(a), three “requisite facts” about his own con- viction, acts, and misconduct or neglect, § 2513(b), regardless of the government’s negligence or misconduct regarding his prosecution. If a petitioner proves these facts about himself, his prosecution and imprisonment are deemed “unjust” and he is entitled to recover damages; his entitlement to recovery does not depend on proof of the government’s unlawful conduct.

Because § 2513 creates a system for recovery that differs sig- nificantly from common-law tort remedies, we see “little rea- son . . . to hark back to stock, judge-made proximate-cause for- mulations.” CSX, 564 U.S. at 702–03. We also hesitate to read proximate causation into § 2513(a)(2) because Congress “has written the words ‘proximate cause’ into a number of statutes.”

Id. at 702 & n.11 (collecting examples). Congress knows how to create a tort-like remedy and how to expressly require prox- imate cause showings, but it did neither in § 2513. See Sanofi Aventis U.S. LLC v. HHS, 58 F.4th 696, 704 (3d Cir. 2023)

(considering the implications of Congress’s failure to use lan- guage that it “knew how to” use).

Section 2513(a)(2) also is disanalogous to the tort law con- cept of contributory negligence, although it bears a superficial similarity thereto. “Contributory negligence is conduct on the part of the plaintiff which falls below the standard to which he should conform for his own protection, and which is a legally contributing cause . . . in bringing about the plaintiff’s harm.”

Restatement (Second) of Torts § 463 (Am. L. Inst. 1965). “At common law, of course, a plaintiff’s contributory negligence operated as an absolute bar to relief.” Norfolk S. Ry. Co. v. Sorrell, 549 U.S. 158, 166 (2007). And “the causation stand- ards for negligence and contributory negligence were the same,” including proximate causation. Id. Section 2513(a)(2) is superficially similar to contributory negligence because it bars recovery if a petitioner’s “misconduct or neglect” caused his injury. But the two differ significantly because “[t]he bur- den of establishing the plaintiff’s contributory negligence rests upon the defendant.” Restatement (Second) of Torts § 477.

That is, contributory negligence is an affirmative defense that must be raised and proved by the defendant. See, e.g., Saporito v. Holland-Am. Lines, 284 F.2d 761, 765 (3d Cir. 1960). Sec- tion 2513(a)(2) is disanalogous because it places the burden squarely on Johnson to prove that his “misconduct or neglect” did not cause his prosecution; it does not make Johnson’s con- tributory “misconduct or neglect” an affirmative defense that must be raised and proved by the government. See United States v. Grubbs, 773 F.3d 726, 732 (6th Cir. 2014) (“The per- son seeking the certificate bears the burden of proof.”). As a result, Johnson’s petition for a certificate of innocence is not “akin to a ‘tort action,’” so proximate causation principles are

inapposite. Bank of Am. Corp. v. City of Miami, 581 U.S. 189, 201 (2017).

Finally, § 2513(a)(2)’s use of the doublet “cause or bring about” also cuts against reading proximate cause principles into the provision. Ordinarily, “terms connected by a disjunc- tive [‘or’ should] be given separate meanings.” Reiter v. Sonotone Corp., 442 U.S. 330, 339 (1979). But “cause” and “bring about” have essentially the same meaning. This sug- gests that § 2513(a)(2) uses repetitive causal language “ex abundanti cautela ([out of an] abundance of caution),” a canon of statutory interpretation “which teaches that Congress may on occasion repeat language in order to emphasize it.” Marx v. Gen. Revenue Corp., 668 F.3d 1174, 1183 (10th Cir. 2011); see King v. Burwell, 576 U.S. 473, 502 (2015) (Scalia, J., dissenting) (“Lawmakers sometimes repeat themselves . . . out of a desire to add emphasis[.]”). By repeating causal terms and connecting them with a disjunctive “or,” § 2513(a)(2) empha- sizes the breadth of Johnson’s burden regarding causation. Cf. Frias-Camilo v. Att’y Gen., 826 F.3d 699, 703 (3d Cir. 2016) (recognizing that disjunctive language typically has the effect of “broaden[ing]” statutory scope).7 Amici’s interpretation— that Johnson must prove only that he did not proximately cause his prosecution, not that he did not factually cause his own prosecution—conflicts with Congress’s emphasis because it lightens Johnson’s burden regarding causation.

Congress’s emphasis is compounded by its use of “less legal- istic [causal] language” like the phrase “bring about,” which suggests that Johnson’s burden extends beyond “judge-made proximate-cause formulations” to factual causation more broadly. CSX Transp., Inc. v. McBride, 564 U.S. 685, 702 (2011).

4. Johnson’s Misconduct Was A But-For Cause Of His Prosecution.

We therefore conclude that § 2513(a)(2) requires Johnson to prove that his misconduct did not factually cause his own prosecution, with factual causation equaling but-for causation.

The District Court correctly concluded that Johnson cannot prove this negative.

First, he cannot show that he did not commit “misconduct or neglect” because he used a lawyer’s signature without her consent to file an exhibit in federal court. That false statement falls squarely within the ordinary meaning of “misconduct.”

See Misconduct, Black’s Law Dictionary (9th ed. 2009) (“A dereliction of duty; unlawful or improper behavior.” (emphasis added)); see also Johnson, 19 F.4th at 264 (describing “Johnson’s actions” as “malicious”). As Amici conceded at oral argument, “nobody here is denying that Mr. Johnson engaged in misconduct.” Oral Arg. Tr. 23:6.

Second, Johnson cannot show that his misconduct did not “cause or bring about” the government’s decision to “prose- cut[e]” him. § 2513(a)(2). His misconduct was a but-for cause of his prosecution. If Johnson had not used the lawyer’s signa- ture to file the exhibit, the government would not have prose- cuted him. See But-for cause, Black’s Law Dictionary (9th ed. 2009) (“[A] cause without which the event could not have occurred.”). Put differently, the government would not have prosecuted him “but for” his misconduct, so his misconduct factually “cause[d] . . . his own prosecution.” § 2513(a)(2). As a result, Johnson cannot satisfy the third requirement for obtaining a certificate of innocence.

B. Johnson And Amici’s Counterarguments Fail.

Johnson and Amici advance several additional text- and policy-based arguments to combat our reading of § 2513(a)(2), but none is successful.

1. “Misconduct Or Neglect” Includes Charged Mis- conduct Or Neglect.

First, Johnson argues that under § 2513(a)(2), “misconduct or neglect” does not cover the conduct for which he was crim- inally charged. He argues that his false statement does not fall within the meaning of “misconduct” in § 2513(a)(2) because that provision separately refers to “the acts charged” and “acts, deeds, or omissions in connection with such charges.” The lat- ter phrases cover Johnson’s false statement because that was the “act[]” for which he was “charged.” Because § 2513(a)(2) uses different phrases than “misconduct or neglect” to refer to charged conduct, Johnson reasons that “misconduct or neglect” must not refer to charged conduct. See United States v. Graham, 608 F.3d 164, 180 (4th Cir. 2010) (Gregory, J., dissenting) (“It must follow that to give meaning to all words in the statute, one cannot ‘cause’ one’s own prosecution by engaging in the very conduct which was found to be non- criminal in the first part of the inquiry.”). Examples of miscon- duct other than charged conduct would include “an attempt to flee, a false confession, the removal of evidence, or an attempt to induce a witness or an expert to give false testimony or opin- ion, or an analogous attempt to suppress such testimony or opinion.” Betts, 10 F.3d at 1285 (quoting United States v. Keegan, 71 F. Supp. 623, 638 (S.D.N.Y. 1947)).

This argument fails because the phrase “misconduct or neglect” sweeps broadly. Without a modifier like “separate” or

“other,” the phrase necessarily covers all misconduct, includ- ing misconduct that was the basis for criminal charges. In read- ing “misconduct or neglect” to exclude charged misconduct, relying on Betts, Johnson is inserting a modifier into § 2513(a)(2) that does not exist. See Graham, 608 F.3d at 175 (“To make its argument then, the dissent must (and does) insert a modifier—‘other,’ ‘additional,’ ‘subsequent,’ or ‘separate’— before ‘misconduct’ in the second clause of § 2513(a)(2).”); United States v. Valle, 467 F. Supp. 3d 194, 204 (S.D.N.Y. 2020) (“Betts reads into the statute a restriction that simply is not there.”). We lack the power to insert such a modifier, so we reject Johnson’s interpretation. See Baker Botts L.L.P. v. ASARCO LLC, 576 U.S. 121, 134 (2015) (“[W]e . . . lack the authority to rewrite the statute.”).

In any event, an interpretation of “misconduct” that includes charged misconduct gives independent meaning to each phrase in § 2513(a)(2). In some cases, “misconduct or neglect” may cover “the acts charged” and the “acts . . . in con- nection” with the charges. Id. But in other cases, the phrases may not overlap. Consider a petitioner who is convicted of forging a check but whose conviction is vacated because the check was later found to be legitimate. There, the “acts . . . in connection” with the charges might not be “misconduct or neglect.” Id. But the petitioner may have committed other “misconduct” unrelated to the check’s legitimacy that “br[ought] about his own prosecution,” such as witness tam- pering. Id. Thus, the meanings of the phrases do not completely overlap, insofar as they may refer to different acts in some cases. We need not adopt Johnson’s strained interpretation to “give effect . . . to every clause and word of [the] statute.”

Advoc. Health Care Network v. Stapleton, 581 U.S. 468, 478 (2017) (internal quotation marks and quoted source omitted).

2. The First And Second Elements Of § 2513 Are Not Superfluous.

Second, Amici argue that our interpretation of § 2513(a)’s third requirement renders the first two requirements superflu- ous. They describe our reading of the third requirement as “subsum[ing]” the first and second: any failure to satisfy the first or second requirement entails failure of the third. Amicus Br. 42. For example, if a petitioner has not been acquitted for innocence, failing the first requirement, his misconduct neces- sarily was a but-for cause of his prosecution, failing the third requirement. And if a petitioner’s acts constituted a crime that was not charged, failing the second requirement, he likewise fails the third. Because our reading of the third requirement does not give the first and second independent meanings, according to Amici, it should be rejected.

But Amici are wrong. Our interpretation gives independent meaning to each of § 2513(a)’s requirements. A petitioner may satisfy the third requirement but not the first: if he is innocent of the charged crime and has not committed any misconduct or neglect in connection with his prosecution, but he has not been officially acquitted or pardoned. A petitioner may satisfy the third requirement but not the second: if one of his charged acts was criminal, but he shows that a corrupt prosecutor was plan- ning to frame and prosecute him regardless of his misconduct, such that his misconduct was not a but-for cause of his prose- cution. And, of course, a petitioner may satisfy the first and second requirements but not the third: Johnson fits this profile.

Thus, our interpretation of § 2513(a)’s third requirement does not subsume the first and second. Contrary to Amici’s argument, a petitioner’s failure of the first or second does not entail failure of the third—and failure of the third does not

entail failure of the first or second, as Johnson’s own petition demonstrates.

3. Our Interpretation Does Not Produce An Absurd Result Here.

Third, Amici argue that our reading of § 2513(a)(2) would lead to absurd results. Suppose that a man commits a minor traffic violation, and a police officer initiates a traffic stop.

During the stop, the officer mistakenly identifies the man as a bank robber. The man is eventually prosecuted for and wrong- fully convicted of bank robbery. Amici suggest that this man would fail the third requirement under our reading: his traffic violation was misconduct that was a but-for cause of the traffic stop, which was a but-for cause of the misidentification, which was a but-for cause of the prosecution. By the transitive prop- erty, his misconduct was a but-for cause of his prosecution, so he fails the third requirement. Amici argue that this is an absurd result, such that our reading of the third requirement should be rejected.

Even if our interpretation of the third requirement leads to unpalatable results in a small fraction of cases, we will not reject it on that basis. We must resolve Johnson’s appeal, not a hypothetical bank robbery petition, and our interpretation of § 2513(a)(2) does not produce absurd results here. See United States v. Moreno, 727 F.3d 255, 259 (3d Cir. 2013) (“[W]hen the statute’s language is plain, the sole function of the courts— at least where the disposition required by the text is not absurd—is to enforce it according to its terms.” (alteration in original) (quoting Lamie v. U.S. Trustee, 540 U.S. 526, 534 (2004))). Johnson is not like a man whose minor traffic viola- tion indirectly leads to his prosecution; his false statement “dis- rupted the administration of justice, interfered with the orderly

work of the federal courts,” and directly led to his prosecution, such that it is not absurd to deny him a remedy. Johnson, 19 F.4th at 263; see Riccio v. Sentry Credit, Inc., 954 F.3d 582, 588 (3d Cir. 2020) (recognizing that a “result cannot be absurd” if “Congress could have any conceivable justification for” it). In a future case, a court may determine that applying our reading of § 2513(a)(2) requires an absurd “disposition.”

Moreno, 727 F.3d at 259. But such a case is not before us, so we need not address that potentiality.

4. Section 2513’s Statutory History Does Not Sup- port Johnson.

Fourth, Amici argue that the statutory history of § 2513(a)(2) supports their strict interpretation of its causal language. The first version of the third requirement, enacted in 1938, barred recovery if a petitioner “either intentionally, or by willful misconduct, or negligence, contributed to bring about his arrest or conviction.” Act of May 24, 1938, Pub. L. No. 75- 539, § 2, 52 Stat. 438, 438. Amici highlight two differences between this language and the current version of the third requirement. First, the old language referred to the petitioner’s “arrest or conviction,” whereas § 2513(a) refers to his “prose- cution.” The government must have a good-faith belief that each element of a crime is satisfied before initiating prosecu- tion, but it needs only probable cause to make an arrest. So Amici argue that this change supports their interpretation of § 2513(a)(2), denying relief only if the petitioner’s misconduct causes the government’s good-faith belief that each element of the charged crime is satisfied. Second, the old provision required only that the petitioner’s misconduct “contributed” to his arrest, whereas § 2513(a)(2) requires that it “cause or bring about” his prosecution. According to Amici, this strengthening of the causal language supports their strict interpretation.

Neither change in the statutory language supports Amici’s reading of § 2513(a)(2) over ours. First, our reading of § 2513(a)(2) accommodates Congress’s use of “prosecution” instead of “arrest.” Because Johnson’s misconduct was a but- for cause of his prosecution, he fails the third requirement, regardless of the relationship between his misconduct and his arrest. Second, our reading of § 2513(a)(2) accommodates Congress’s use of “cause or bring about” instead of “contrib- uted.” Congress’s use of “contributed” in the old provision suggests that it required less than but-for causation. See Macleod, supra, at 974–75 (describing “contributing factor” causation as an “alternative” and more “permissive” standard for factual causation than but-for). The change from “contrib- uted” to “cause or bring about” is consistent with a change to traditional but-for causation. It does not imply that § 2513(a)(2) requires a standard stricter than but-for causation or incorporates proximate causation principles, so Amici’s statutory history argument fails.

* * * Finally, Amici appeal to legislative history to support their understanding of § 2513(a)(2). Because we conclude that the text is clear, we need not consider this evidence. See S.H. ex rel. Durrell v. Lower Merion Sch. Dist., 729 F.3d 248, 259 (3d Cir. 2013) (“Legislative history has never been permitted to override the plain meaning of a statute.”).8

Even if we considered Amici’s evidence, it does not support Johnson. Amici note that § 2513 had its origins in a 1912 bill supported by Edwin Borchard. See Edwin M. Borchard, State Indemnity for Errors of Criminal Justice, S. Doc. No. 62-974,

at 32 (1912) (requiring “the claimant [to] show that he has not, by his acts or failure to act, either intentionally or by willful misconduct or negligence, contributed to bring about his arrest or conviction”). Borchard observed that a “limitation almost uniformly expressed in [similar] [European] statutes is that the claimant shall not have intentionally or by gross negligence caused his detention.” Id. at 17. And he associated this limita- tion with misconduct like “an attempt to flee, a false confes- sion, the removal of evidence, or an attempt to induce a witness or an expert to give false testimony or opinion.” Id. at 18. These examples are consistent with Amici’s “estoppel” theory of § 2513(a)’s third requirement, so Amici conclude that the leg- islative history cuts decisively in Johnson’s favor. But Amici’s description of Borchard’s report is misleading. Borchard listed examples like flee attempts and false confessions because “[t]he statutes of some of the countries, such as Germany, Hungary, Norway, and Sweden, specifically mention [those] limitations” on relief. Id. at 17–18. The Norwegian statute, for example, explicitly barred relief “for detention pending exam- ination which has occurred because the accused has attempted to flee or has so acted that the conclusion had to be drawn that he has sought to remove traces of the deed, or induce others to bear false witness, or to suppress their testimony.” Id. at 25.

But Borchard recognized that some countries did not explicitly bar relief in those circumstances. Id. at 18 (“France expressly declines to specify any limitations on the right, leaving it to the judge to determine what acts . . . shall constitute a sufficient objection to the payment of an indemnity.”). And he concluded his report by noting that Congress could structure such limita- tions as it pleased. Id. at 21 (“[W]ithin what limits and under what conditions the indemnity shall be awarded, are matters

IV. CONCLUSION Johnson cannot prove that “he did not by misconduct or neglect cause or bring about his own prosecution” because his false statement in federal court was a but-for “cause” of the government’s decision to “prosecut[e]” him. § 2513(a)(2).

Johnson and Amici’s arguments to the contrary fail. The phrase “misconduct or neglect” is unqualified, so it covers charged misconduct like Johnson’s false statement. Our interpretation gives independent meaning to each of § 2513(a)’s three requirements, does not produce an absurd result in this case, and coheres with § 2513(a)’s statutory history. We will there- fore affirm the District Court’s order denying Johnson’s peti- tion for a certificate of innocence.

which legislatures can work out with little difficulty.”). Thus, Borchard never stated that a limitation like § 2513(a)’s third requirement covers only false confessions and other “estoppel” misconduct. False confessions were merely among a few “example[s]” of misconduct for which European countries had specifically chosen to withhold relief. Id. at 18.