Amarjot Singh v. Strafford County Department of Corrections, Superintendent, et al.

Opinion

UNITED STATES DISTRICT COURT FOR THE DISTRICT OF NEW HAMPSHIRE

Amarjot Singh

v. Case No. 1:25-cv-268-SE-TSM Opinion No. 2026 DNH 092 Strafford County Department of Corrections, Superintendent, et al.

O R D E R

Amarjot Singh, a native and citizen of India currently detained at Strafford County Department of Corrections, filed a petition for a writ of habeas corpus under 28 U.S.C. § 2241, contesting his detention and requesting, inter alia, a bond hearing. Granting the petition in part, the court ordered the respondents “to provide the petitioner with a bond hearing before an [Immigration Judge (IJ)] as soon as practicable.” Doc. no. 34. The IJ denied bond, finding that Singh was a danger to the community by clear and convincing evidence. In response, Singh filed a motion for review of the bond denial and a second amended petition arguing that the bond hearing did not comport with the Due Process Clause of the Fifth Amendment. The respondents object. For the reasons explained below, the court grants the motion and the second amended petition, in part.

Background At Singh’s January 22, 2026 bond hearing, the Department of Homeland Security (DHS) argued that he was both a danger to the community and a flight risk. DHS first pointed to a police report from Concord, New Hampshire. DHS argued that the report implicated Singh as the driver for an individual carrying out “a large scale sophisticated phone scam that was targeting elderly individuals.” Doc. no. 40-2 at 7. DHS noted that Concord police arrested Singh’s passenger, but did not charge Singh as the driver, and argued that Singh had been paid cash to drive the scammer “at least two to three times.” Id. at 9. DHS further contended that although Singh claimed that he did not know anything about the scam, he was complicit because of his

role as a driver, which rendered him a danger to property. Finally, DHS also mentioned a dismissed criminal case against Singh in Brookline, Massachusetts related to a similar scam carried out by the same passenger. Arguing that he was not a danger, Singh emphasized that he was merely a driver and lacked knowledge of his passenger’s fraudulent conduct. He noted that because of his limited involvement as a driver and his consistent denial of any knowledge of the fraud, the Brookline District Court found that there was insufficient evidence against him to constitute probable cause and dismissed the charge of larceny. Singh further argued that he was not a flight risk because he had a pending motion to reopen his removal proceedings, welcomed a child in 2025, and held a stable job. He submitted several letters attesting to his good moral character.

The IJ denied bond, finding that Singh was a danger to the community by clear and convincing evidence. The IJ declined to provide oral reasoning at the hearing. Because Singh planned to appeal the decision, the IJ subsequently issued a written decision explaining her reasoning. The IJ’s written decision, dated February 13, 2026, justified her conclusion by stating that Singh “was charged for [sic] conspiracy to commit theft by deception and by [sic] the Concord, New Hampshire police department.” Doc. no. 42-1 at 4. The IJ further reasoned that Singh had “admitted to having been paid to drive the other individual and pick up similar packages on at least three prior occasions.” Id. The IJ “felt [a financial scheme targeting the elderly] was a very serious offense and clearly put others at risk,” and concluded that “it was clear evidence that [Singh] presents a danger to others.” As such, the IJ found that Singh was a danger to the community by clear and convincing evidence. Id. The immigration court record contains documents related to overlapping investigations by the Brookline and Concord police departments. Although the alleged criminal events in

Brookline occurred first, in late June 2025, Singh and his passenger first encountered law enforcement on July 2, 2025, in Concord. On July 2, 2025, police stopped Singh’s vehicle outside a residence in Concord. A complaining witness had worked with the police to lure a suspected scammer to her home, and the suspect arrived as a passenger in Singh’s vehicle. The police interviewed Singh at the scene. He informed them that he had previously driven the same passenger “two or three times.” Doc. no. 42-1 at 32. Singh told police that he asked his passenger what he was picking up, but the passenger told him it was none of his business. Although the passenger paid Singh more than his Uber rate for the rides, Singh said that he “believed [the passenger] when he told him he wasn’t doing anything illegal.” Id. The police officer noted that Singh “appeared surprised” when the

officer told him that the passenger was “scamming vulnerable elderly people.” Id. After investigating, the Concord police did not charge Singh with any crime but charged his passenger with conspiracy to commit theft by deception. Homeland Security Investigations detained Singh for administrative immigration purposes following the police interview. The respondents have continuously detained him since. The Concord police later shared evidence of the July 2, 2025 incident with the Brookline police. The previous day, a complaining witness had reported to the Brookline police a similar scam that occurred in Brookline on June 27, 2025. Based on the Concord investigation, Brookline identified Singh and his passenger as suspects. The Brookline police reports list the crime as larceny by false pretenses and mention Singh’s role as a driver. A corresponding Brookline District Court docket reflects that a criminal complaint charged Singh with larceny over $1200 in September 2025. Singh moved to dismiss the charge based on lack of probable cause, arguing that he was merely a taxi driver. The Brookline court granted Singh’s motion to

dismiss in November 2025.

Analysis I. Exhaustion At the threshold, the respondents argue that the court should decline to entertain the relief sought because Singh failed to exhaust his administrative remedies and no exception to the exhaustion requirement applies. The respondents point out that although Singh administratively appealed the IJ’s denial of bond, that appeal is still pending before the Board of Immigration Appeals (BIA). Singh counters that awaiting the BIA’s resolution would be futile and that habeas relief is necessary to prevent the injury of continued detention. Many courts have recognized that

an exception to the common-law exhaustion requirement arises if denying immediate judicial resolution of a claim would work an irreparable harm, such as a loss of liberty. See, e.g., Makuraza v. Wesling, No. 1:26-CV-122-JL-AJ, 2026 WL 1068066, at *2 (D.N.H. Apr. 20, 2026); Miti v. Moniz, No. 26-11327-BEM, 2026 WL 884639, at *5 (D. Mass. Mar. 31, 2026). Courts have further observed that the lengthy processing times at the BIA make the potential injury particularly acute. See Makuraza, 2026 WL 1068066, at *2. For these reasons, the court will not require exhaustion. II. Review of Bond Denial The parties agree, as the IJ recognized, that DHS bore the burden of proving at the bond hearing that Singh posed a danger to the community by clear and convincing evidence, or that he posed a flight risk by a preponderance of the evidence. See Hernandez-Lara v. Lyons, 10 F.4th

19, 41 (1st Cir. 2021) (establishing the government’s burdens of proof at bond hearings granted pursuant to § 1226(a)); Destino v. FCI Berlin, Warden, No. 1:25-CV-374-SE-AJ, 2025 WL 4010424, at *9 (D.N.H. Dec. 24, 2025) (reasoning that a “logical extension” of Hernandez- Lara’s holding was to require, on due process grounds, a bond hearing at which the government was held to its burden). The IJ ultimately found that the government carried its burden of proving by clear convincing evidence that Singh poses a danger to the community. Singh challenges that decision on two grounds. First, he argues that the IJ’s failure to articulate any basis for detention contemporaneously with her decision at his bond hearing violates due process and precludes meaningful judicial review. Second, Singh argues that the record cannot, as a matter of law, establish dangerousness by clear and convincing evidence. The respondents, for their part,

emphasize the deference owed to the IJ’s decision and argue that the IJ applied the correct standard of proof and took a reasonable view of the evidence. Because the court grants relief on the basis of Singh’s second argument, that the record before the IJ would compel any reasonable factfinder to conclude that the government has not met its burden of showing that Singh is a danger to the community by clear and convincing evidence, it does not address Singh’s argument that the IJ was required to provide oral reasoning contemporaneously with her decision. The court finds that the standard of review applicable here is as set out by another court in this district in Gupta v. Brackett, No. 1:26-CV-207-PB-TSM, 2026 WL 1734839, at *2 (D.N.H. June 16, 2026). As the court there recognized, although 8 U.S.C. § 1226(e) “strips the Court of the power to review an IJ’s discretionary decision to detain a person subject to removal proceedings, it does not restrict the Court’s authority to review claims that the person is being detained in violation of his constitutional rights” or other questions of law contained within the

decision to detain. Id. Whether a noncitizen is an unacceptable danger to the community or a flight risk is a mixed question of law and fact and is thus subject to judicial review notwithstanding § 1226(e). Id. The court is not, however, “free to second guess an IJ’s discretionary decisions.” Id. Rather, “if the IJ properly allocates the burden of proof and correctly resolves any pure issues of law that bear upon her decision, her conclusion that [a noncitizen] poses either a danger to the community or an unacceptable risk of flight cannot be overturned as long as it is supported by substantial evidence.” Id. (citing Mayancela v. Bondi, 136 F.4th 1, 10 (1st Cir. 2025)). Under the substantial evidence standard, the court will uphold the IJ’s decision if it is “supported by reasonable, substantial, and probative evidence on the record considered as a

whole.” Acosta v. Blanche, 176 F.4th 58, 77 (1st Cir. 2026). “Substantial evidence means more than a mere scintilla. It means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Hernandez v. Blanche, No. 24-1482, 2026 WL 1959422, at *10 (1st Cir. July 7, 2026) (cleaned up). Accordingly, this court may overturn the IJ’s decision only if the record would “compel a reasonable factfinder to reach a contrary determination.” Gupta, 2026 WL 1734839, at *2. To answer that question, the court must review the entire record, not merely the evidence that supports the IJ’s conclusions. Jabri v. Holder, 675 F.3d 20, 24 (1st Cir. 2012). According to the IJ’s written decision explaining her reasoning, her dangerousness determination rested on the following: [Singh] was charged for [sic] conspiracy to commit theft by deception and by [sic] the Concord, New Hampshire police department, and that [sic] he admitted to having been paid to drive the other individual and pick up similar packages on at least three prior occasions. The scam was a financial scheme that targeted the elderly. The Court felt this was a very serious offense and clearly put others at risk, further the Court found it was clear evidence that [Singh] presents a danger to others.

Doc. no. 42-1 at 4 (citations omitted). Although the IJ was not required to “discuss ad nauseum every piece of evidence,” she still had an obligation to “give[] reasoned consideration to the evidence as a whole, ma[k]e supportable findings, and adequately explain[] her reasoning.” Gupta, 2026 WL 1734839, at *3 (quoting Bazile v. Garland, 76 F.4th 5, 16 (1st Cir. 2023)). The IJ failed to satisfy this requirement, instead premising her decision on findings that were either erroneous or irrelevant, without adequate explanation. Beginning with the erroneous findings, Singh argues that “one of the only concrete facts the [IJ’s] memorandum identified in support of the dangerous finding” was based on a material misreading of the record. Doc. no. 45 at 2-3. Indeed, the IJ’s order confusingly states that Singh “was charged for [sic] conspiracy to commit theft by deception and by [sic] the Concord, New Hampshire police department.” Doc. no. 42-1 at 4. To the contrary, the Concord police report states that Concord charged Singh’s passenger with conspiracy to commit theft by deception but did not charge Singh with any crime. Thus, the IJ’s order appears to attribute to Singh a charge brought only against his passenger. Even assuming that the IJ meant to identify both Singh’s encounter with the Concord police and the Brookline charge of larceny does not cure the issue because there is zero evidence that Singh was ever “charged . . . by the Concord” police or “charged [with] conspiracy to commit theft by deception.” Thus, the foremost fact on which the IJ relied to support her conclusion of dangerousness is not supported by substantial evidence. Further, even if the IJ merely misstated the charge and charging jurisdiction and meant to rely on the fact that Brookline initially charged Singh with larceny, the fact of that charge is not

substantial evidence of dangerousness on this record. The fact of a criminal charge may in some cases support a finding of dangerousness because a charge reasonably implies that law enforcement found at least probable cause that the defendant committed a crime or because the information contained in related police reports constitutes clear and convincing evidence of dangerousness. See Miti, 2026 WL 884639, at *7. And, to be sure, a court’s dismissal of a criminal charge “does not automatically cancel out [the] evidentiary value” of the arrest, charge, or underlying conduct. Id. at *6. But the Brookline District Court dismissed the charge against Singh because he demonstrated that there was no probable cause to support it. Dismissal on the merits for lack of probable cause demonstrates more than a dismissal under other circumstances, e.g., dismissal in the exercise of prosecutorial discretion. Without more, such as the presence of

additional evidence not available to the criminal court, dismissal on the merits for lack of probable cause destroys the probative value of the arrest, charge, and underlying conduct and their ability to act as “such relevant evidence as a reasonable mind might accept as adequate to support [the] conclusion” that Singh poses a danger to the community by clear and convincing evidence. See Hernandez, 2026 WL 1959422, at *10. As a matter of law, clear and convincing evidence is a significantly higher evidentiary burden than probable cause. Compare Cruzan by Cruzan v. Dir., Missouri Dep’t of Health, 497 U.S. 261, 285 n.11 (1990) (defining the clear and convincing evidence standard) with Maryland v. Pringle, 540 U.S. 366, 370-71 (2003) (probable cause). “In detention cases, applying [this] heightened standard of proof reflects the value society places on individual liberty, and avoids the risk associated with the preponderance standard of increasing the number of individuals erroneously committed.” Hernandez-Lara, 10 F.4th at 39-40 (cleaned up). Yet, the IJ found that the same constellation of facts that the Brookline court found legally insufficient to establish

probable cause of criminality nonetheless amounted to clear and convincing evidence of dangerousness. The court acknowledges that it is not aware of any legal principle that required the IJ to adopt the reasoning of the Brookline District Court. But the IJ’s decision is irreconcilable with the outcome of Singh’s criminal case. The record reflects that the Brookline police obtained a warrant for Singh based on evidence the Concord police gave to them and that the Brookline police report attached the Concord police report. Doc. no. 42-1 at 136. Accordingly, the records before the IJ and the Brookline court each included the facts of both the Concord and Brookline investigations and the related police reports. There is no indication that the IJ had before her any evidence probative of Singh’s culpability that was not before the Brookline District Court. Under these circumstances, the arrest, charge, and underlying conduct

are at best not substantial evidence supporting the IJ’s finding and at worst not probative of Singh’s dangerousness. At a minimum, the IJ was required to explain how she bridged the wide gap between the Brookline District Court’s finding of insufficient evidence to support probable cause and her own finding of clear and convincing evidence of dangerousness. Instead, the IJ did not grapple with the Brookline court’s dismissal let alone explain how Singh’s conduct could be insufficient to support probable cause and yet nonetheless clear and convincing evidence of dangerousness despite the conduct’s consistency with Singh’s argument that he was an innocent driver. Accordingly, on this record, no reasonable IJ could have found that Singh was involved in “a very serious offense and clearly put others at risk” on the basis that he was charged in connection with his passenger’s scheme. Doc. no. 42-1 at 4. Importantly, the IJ’s reasoning turned on the dangerousness of the fraud scheme, not the dangerousness of Singh’s participation in, or proximity to, that crime. In other words, based on

the erroneous premise that Singh’s conduct supported a criminal charge, the IJ assumed that Singh was criminally culpable in the scheme. Yet as explained above, the record shows that Singh’s conduct did not sustain any criminal charges. In addition to her erroneous reliance on the criminal charge, the IJ reasoned that Singh “admitted to having been paid to drive the other individual and pick up similar packages on at least three prior occasions.”1 Doc. no. 42-1 at 4. Just as the charging history fails to impute the passenger’s dangerousness to Singh, the evidence that Singh had driven the passenger previously is consistent with Singh’s position that he was merely a hired taxi driver, not an accomplice to his passenger’s crimes. A reasonable factfinder could not, reviewing the record as whole, discern something nefarious about the fact that Singh had driven the passenger multiple times, and his

role as driver is not substantial evidence that he is a danger to the community by clear and convincing evidence. In sum, the IJ’s conclusion that Singh was dangerous rested on facts that were either unsupported by substantial evidence or not probative of the issue of Singh’s dangerousness. As

1 The IJ’s finding that Singh admitted to “at least three prior occasions” is likewise not supported by substantial evidence. See doc. no. 42-1 at 4. The police report cited by the IJ states that Singh told the officer that he had “driven [the same passenger] about two or three times.” Doc. no. 42-1 at 32. This may seem a small difference, but the court has an obligation to examine the IJ’s findings under the substantial evidence standard and the record would compel a reasonable factfinder to conclude that Singh did not admit to driving the other individual “on at least three prior occasions.” such, the IJ’s conclusion that Singh is a danger to the community cannot be sustained on its own reasoning. Despite the flaws in the IJ’s reasoning, the respondents argue that the IJ’s finding of dangerousness can nonetheless be sustained on an alternative ground not mentioned by the IJ.

Specifically, the respondents contend that the “IJ’s written memorandum reflects that she gave great weight to the information contained in the [Concord] police report, which included information that Petitioner, at a minimum, strongly suspected that the individual he was driving was engaged in unlawful activity.” Doc. no. 44 at 8 (emphasis in original). In other words, notwithstanding the conclusion that the evidence upon which the IJ relied is not substantial evidence to support the IJ’s finding that Singh is dangerous, the record nonetheless supports that Singh is dangerous on the basis that he “turned a blind eye” to his passenger’s criminal conduct. Id. at 9. The record, though, does not contain substantial evidence that Singh did so. The police interview on which the respondents rely does not provide substantial evidence that Singh was willfully blind to his passenger’s criminal conduct. Singh told the Concord police

that he had driven the passenger a few times after meeting him while driving for Uber. He admitted that he got paid more than his Uber rate to drive the passenger but denied knowing what the passenger picked up on each ride. Singh asked what the passenger was doing, but the passenger told him, “you are a driver, so that’s not your business[,] . . . it’s part of my business.” Doc. no. 42-1 at 32. The police asked Singh if the passenger’s practice of travelling long distances for a brief interaction made him suspicious that the passenger was breaking the law. Singh told the police that he had asked the passenger directly, but the passenger denied any wrongdoing, and Singh “believed [him] when he told him that he wasn’t doing anything illegal.” Id. Tellingly, the interviewing officer noted that Singh “appeared surprised” when the police told him that his passenger “was scamming vulnerable elderly people into taking their gold.” Id. This is not substantial evidence that Singh was willfully blind to his passenger’s crimes. Willful blindness requires “evidence that red flags existed that the defendant consciously avoided

investigating.” United States v. Valbrun, 877 F.3d 440, 446 (1st Cir. 2017). To the extent that red flags existed, the only evidence in the record shows that Singh did not avoid investigating but rather inquired about his passenger’s activities and believed him when he denied wrongdoing. Certainly, a factfinder could consider Singh’s denials of knowledge in his police interview self- serving, but those statements are corroborated by the interviewing officer’s own impression that Singh “appeared surprised” when informed of his passenger’s fraud. Further, the outcome of the Concord investigation indicates that the investigating officers concluded that Singh was not criminally culpable. The officers not only declined to charge Singh but also noted in the affidavit in support of the passenger’s arrest that Singh “was determined to be a taxi driver,” implying that they ruled him out as an accomplice. Doc. no. 42-1 at 40; see Miti, 2026 WL 884639, at *7

(reversing the IJ’s denial of bond based on a police report that resulted in a dismissed criminal charge in part because “the actions and statements of the individuals who first responded to the alleged assault [did] not support a strong inference that those actors perceived [the petitioner] as a dangerous person”). The Concord police’s conclusion was reaffirmed by the Brookline District Court, which, after considering the same evidence that DHS presented to the IJ, dismissed the charge against Singh on the basis that there was insufficient evidence to establish probable cause that he was more than a taxi driver. After considering the entire record, the court is compelled to conclude that it does not contain “such relevant evidence as a reasonable mind might accept as adequate to support” the IJ’s conclusion that Singh presents a danger to the community by clear and convincing evidence. Hernandez, 2026 WL 1959422, at *10. The circumstantial evidence in the record that Singh was culpable in the fraud scheme is insufficient to constitute substantial evidence in support of the IJ’s finding. That same circumstantial evidence was not enough to establish even probable cause

in the eyes of the Brookline District Court, and no reasonable factfinder viewing the evidence in any alternative way could reach the high clear and convincing evidence threshold.

III. Remedy Because a reasonable adjudicator would be constrained to conclude that the record cannot establish by clear and convincing evidence that Singh is a danger to the community, it is inappropriate to order the IJ to revisit her dangerousness finding. See Timbigamba, 2026 WL 1470320, at *11-12. The IJ did not, however, make any finding as to flight risk that the court can review at this time under the substantial evidence standard. Consequently, the court concludes that the IJ must decide in the first instance whether Singh is a flight risk by a preponderance of

the evidence. The court orders the respondents to provide a new bond hearing for that limited purpose.

Conclusion For the foregoing reasons, the court finds that the IJ’s decision that Singh is a danger to the community by clear and convincing evidence is not “supported by reasonable, substantial, and probative evidence on the record considered as a whole.” Acosta, 176 F.4th at 77. Therefore, the court grants, in part, the motion for review of the bond denial and the second amended petition. The respondents shall provide the petitioner with a bond hearing before an IJ as soon as practicable at which the respondents will bear the burden of proving that the petitioner is a flight risk by a preponderance of the evidence. If the respondents fail to prove risk of flight by a preponderance of the evidence, the respondents shall release the petitioner under reasonable conditions of supervision. If the respondents fail to provide a bond hearing, the court will set appropriate terms and conditions of the petitioner’s release during the pendency of his removal proceedings. The respondents shall file a status report within seven days.

SO ORDERED.

Samantha D. Elliott United States District Judge July 17, 2026 ce: Counsel of Record