“Grave Doubt” in Coram Nobis: Johnson v. United States and the Second Circuit’s Refined Harmless-Error Framework after Ciminelli

Introduction

Johnson v. United States, No. 24-1221 (2d Cir. July 17, 2025), is the Second Circuit’s first substantive encounter with the Supreme Court’s repudiation of the “right-to-control” theory of fraud in Ciminelli v. United States, 598 U.S. 306 (2023). The case raises two intersecting questions of first impression:

What harmless-error standard governs a petition for a writ of coram nobis when a general verdict may rest on an invalid legal theory (a “Yates error”)?
How robust must the proof of fiduciary duty and misappropriation be when the Government pivots to a misappropriation-of-confidential-information theory after Ciminelli?

Although the panel (Judges Calabresi, Nathan, and Kahn) declined to choose definitively between the Chapman or Kotteakos harmless-error tests, it announced a pragmatic rule: if the court harbors “grave doubt” that the jury actually relied on the lawful theory, the conviction cannot stand—even on a coram nobis petition.

Summary of the Judgment

Mark Johnson, formerly global head of HSBC’s FX trading desk, was convicted in 2017 of wire fraud and conspiracy. The Government presented two theories:

Right-to-control – later declared invalid in Ciminelli.
Misappropriation – based on an alleged fiduciary duty to Cairn Energy.

After serving his sentence, Johnson sought coram nobis relief. The district court found the erroneous instruction harmless, reasoning that the same jury would have convicted under the misappropriation theory. The Second Circuit reversed:

The misappropriation case was “comparatively weak.”
The panel had “grave doubt” whether the jury actually relied on it.
Because that doubt persists under either harmless-error standard, the writ issues: the conviction is vacated and the petition granted.

Analysis

Precedents Cited and Their Influence

Ciminelli v. United States, 598 U.S. 306 (2023): declared the right-to-control doctrine invalid; supplied the trigger for Johnson’s petition.
Yates v. United States, 354 U.S. 298 (1957) & Griffin v. United States, 502 U.S. 46 (1991): establish that a general verdict resting on alternative theories—one invalid—requires harmless-error review.
Hedgpeth v. Pulido, 555 U.S. 57 (2008): confirms that Yates errors are not structural but subject to harmless-error analysis.
Kotteakos v. United States, 328 U.S. 750 (1946) vs. Chapman v. California, 386 U.S. 18 (1967): competing tests for harmlessness; panel declines to choose.
Percoco v. United States, 598 U.S. 319 (2023): warns against “lightly implying” fiduciary duties; pivotal in the panel’s skepticism toward the Government’s fiduciary-duty theory.
United States v. Chestman, 947 F.2d 551 (2d Cir. 1991) (en banc) & United States v. Skelly, 442 F.3d 94 (2d Cir. 2006): define fiduciary or “trust and confidence” relationships for misappropriation liability.
Brecht v. Abrahamson, 507 U.S. 619 (1993) & O’Neal v. McAninch, 513 U.S. 432 (1995): describe the “substantial and injurious effect” and “virtual equipoise” approaches that the panel adapts into its “grave doubt” language.

Legal Reasoning

Existence of a Yates Error. With Ciminelli voiding right-to-control, the jury instructions necessarily contained an invalid legal theory. Because the verdict was general, the record was infected.
Standard of Harmless-Error Review. The panel canvassed Chapman (reasonable-doubt) and Kotteakos (substantial-injurious-effect) without selecting one, holding the Government loses even under the more lenient Kotteakos test. The operative formulation became “grave doubt”—language borrowed from O’Neal.
Assessment of the Misappropriation Evidence.
- Fiduciary relationship: Contractual documents (Mandate Letter & ISDA) expressly disclaimed any advisory or fiduciary role; NDA not enough to override.
- Misappropriation of confidential information: FX dealers must pre-hedge; Government failed to delineate how Johnson’s trading crossed the line from legitimate pre-hedging to prohibited front-running.
The panel described the Government’s proof as “decidedly weak,” making it improbable that the jury would have agreed unanimously on that theory absent the right-to-control avenue.
Application to Coram Nobis. The court recognised coram nobis as the “extraordinary remedy” of last resort but emphasised that substantive retroactive decisions—like Ciminelli—must be enforced to prevent punishment for non-criminal conduct.

Impact on Future Litigation

Harmless-Error in Post-Conviction Relief. The “grave doubt” formulation gives district courts and litigants a workable yardstick: if evidence on the valid theory is not overwhelming and the invalid theory is compelling, relief is likely even years after conviction.
Fiduciary-Duty Threshold in Misappropriation Cases. Prosecutors must now reckon with extensive contractual disclaimers; mere “trust” or industry custom will rarely suffice.
FX Market Prosecutions. By scrutinising pre-hedging versus front-running, the opinion hints that the foreign exchange market’s structure complicates traditional misappropriation analysis. Future indictments will require clearer proof that defendants exceeded hedging norms.
Resurgence of Coram Nobis. Johnson shows that federal courts are willing to employ the writ aggressively when a substantive rule change (here, Ciminelli) casts doubt on convictions, especially for non-citizen defendants facing collateral consequences.

Complex Concepts Simplified

Coram nobis: A rare common-law writ letting federal courts revisit criminal convictions after sentence is served, where no other remedy is available.
Yates error: Occurs when a jury returns a general verdict after being instructed on multiple theories, one of which is later deemed invalid.
Right-to-control fraud: The (now-rejected) notion that depriving a victim of valuable economic decision-making information is fraud even if no money or property is taken.
Misappropriation theory: Fraud committed by stealing and exploiting confidential information obtained through a duty of trust or confidence.
Harmless-error standards:
- Chapman: Government must prove the error harmless beyond a reasonable doubt.
- Kotteakos: Court asks whether the error had a substantial and injurious effect on the verdict.
- Grave doubt: When the court cannot say the valid theory more likely than not secured the conviction, the judgment cannot stand.
Pre-hedging vs. Front-running: In FX, banks legitimately accumulate currency in advance (pre-hedging); trading on confidential client information to profit for themselves without authorisation crosses into front-running/misappropriation.

Conclusion

Johnson v. United States cements two critical propositions in federal criminal law:

In post-Ciminelli practice, harmless-error analysis in coram nobis employs a “grave-doubt” lens; if the lawful theory is not plainly decisive, the conviction is vacated.
Misappropriation liability demands clear fiduciary and misuse evidence; contractual disclaimers and market norms loom large.

Practitioners should treat Johnson as a potent precedent: substantive Supreme Court reversals of fraud doctrines will ripple through final convictions, and the Second Circuit is prepared to grant extraordinary relief when uncertainty lingers.