Why the Pan Am 103 conviction survived two appeals

The operative fact in the Pan Am 103 Lockerbie bombing case is not that the conviction was referred back twice, and not that later reporting exposed real evidentiary weaknesses. It is that on January 15, 2021, a five-judge Scottish appeal court refused the posthumous appeal of Abdelbaset al-Megrahi. The court held that the verdict remained one a reasonable jury could have reached and that the undisclosed material identified in the appeal would not have altered the outcome. [1]

That refusal is a hard place to start, but it is the safest one. Megrahi’s conviction has never been overturned. His first appeal was refused in 2002; he was released in 2009 on compassionate grounds and died in 2012; the Scottish Criminal Cases Review Commission referred the case again in 2020; and the resulting posthumous appeal was rejected in 2021. [2]

Starting there does not clean the record. It prevents a different error: treating every proven defect as if it necessarily became legal unsafety. The record contains informant credibility problems, eyewitness identification weaknesses, reward-payment allegations, disclosure disputes, and chain-of-custody challenges. The appellate answer was not that those issues were imaginary. It was that, measured against the appellate tests before the court, they did not undo the conviction.

What the 2020 referral actually put back before the court

The 2020 SCCRC referral matters because it was not a press release, a campaign paper, or a television reconstruction. It was an independent Scottish review body deciding that two grounds met the statutory threshold for sending the conviction back to appeal. The Commission considered six grounds and referred on two: unreasonable verdict and non-disclosure. [3]

Those labels should not be inflated. A referral is not an acquittal in waiting. It is a conclusion that a possible miscarriage issue should be considered by the appeal court. The significance is procedural but serious: a second independent referral meant that, nearly two decades after conviction, the court had to revisit whether the case could lawfully stand in light of the identified defects.

“Unreasonable verdict” directed attention to whether the trial court’s conclusion was one no reasonable jury or fact-finder could have reached on the evidence. “Non-disclosure” directed attention to whether withheld material, if available to the defense, would have had a realistic capacity to affect the verdict. Those are related questions, but they are not the same question. One asks whether the evidentiary result was beyond lawful reach. The other asks whether the defense was deprived of material that mattered enough to undermine confidence in the result.

The 2021 court answered both against Megrahi. It accepted the appellate task as a review of safety, not a fresh trial on every contested item. Its summary states that the trial court had been entitled to convict and that the undisclosed documents did not create a miscarriage of justice. [1]

Why that refusal remains striking

The refusal is striking because the underlying case was never a simple single-witness prosecution. The conviction depended on a converging evidentiary picture: the timer theory, the purchase of clothing said to have been packed with the bomb, and the reconstruction of the baggage route. Each was contestable. None becomes irrelevant merely because the conviction survived.

The bombing itself needs only brief orientation here. Pan Am Flight 103 was destroyed over Lockerbie on December 21, 1988. Megrahi was later convicted; his co-accused, Lamin Khalifah Fhimah, was acquitted; the first appeal failed in 2002; Megrahi was released on compassionate grounds in 2009; and the 2020 SCCRC referral led to the 2021 refusal. [2]

The current U.S. prosecution of Abu Agila Mohammad Mas’ud Kheir Al-Marimi is a separate live docket, not the subject of this article. Risk Digest maintains that parallel record at Where the Pan Am 103 Prosecution Stands Before Trial. The point here is narrower: how the only completed conviction survived post-conviction review despite a record that gives evidence managers plenty to dislike.

The evidence-integrity map

A useful way to read the case is not as a mystery board but as a verification file. Four categories keep recurring: informant incentives, eyewitness reliability, disclosure control, and chain-of-custody discipline. They are not equal in legal consequence, but they are the places where a trial record either earns future resilience or leaves appellate counsel explaining why defects did not matter enough.

Informants: Giaka and the cost of late credibility collapse

The most direct lesson for any litigator handling intelligence-adjacent evidence is the treatment of Abdul Majid Giaka, a former Libyan intelligence officer and CIA informant. Reporting based on the reopened Lockerbie files records that the trial judges found his evidence “at best grossly exaggerated and at worst simply untrue.” [4]

That sentence is not a small impeachment point. It is the kind of finding that should make a case team reopen every assumption about source handling: what the witness was paid, what benefits he expected, what he had already told handlers, what had changed between earlier accounts and trial testimony, and whether any intelligence filtering had turned credibility material into something harder for the defense to see.

Yet Giaka’s collapse did not collapse the conviction. That is the procedural coexistence that matters. A witness can be badly damaged, even judicially discredited, without becoming the whole case. If the remaining evidence is treated as sufficient, the appellate question moves from “was this witness reliable?” to “did the conviction depend on this witness in a way that makes the verdict unsafe?” Those are very different working questions.

Eyewitness identification: Gauci, clothing, and reward-payment material

Tony Gauci’s identification evidence carried a different kind of risk. Gauci, a Maltese shopkeeper, was linked to the clothing said to have been packed with the bomb. Eyewitness identification is already fragile when time, exposure, suggestion, and repetition enter the file. In this case, the later argument was not simply that a witness might have been mistaken; it was that the defense should have had fuller visibility into material bearing on motivation and reliability.

The reward-payment issue has to be stated carefully. The Guardian reported during the 2020 appeal that Gauci and his brother were paid $3 million by the United States after Gauci testified, and that the appeal argued the court should have been told a witness wanted payment. That is appellate-advocacy reporting, not a freestanding judicial finding that the payment bought false testimony. [5]

The distinction matters. A reward-payment record may be powerful impeachment material without proving perjury. It may affect how a fact-finder assesses interest, expectation, pressure, and consistency. But the appellate court still had to decide whether the non-disclosed material, in the context of the whole case, would have made a material difference. In 2021, it held that it would not. [1]

Disclosure: the missing-document problem is rarely just a missing-document problem

The SCCRC’s second accepted ground was non-disclosure. The public reporting identifies undisclosed police and CIA-related material as part of the appeal context, and the court’s own summary frames the issue as whether undisclosed documents would have affected the verdict. [1][3]

For trial teams, the dangerous part is not only the legal remedy. It is the operational uncertainty. A document that is logged poorly, reviewed late, siloed as intelligence material, or described too generally can become an appellate exhibit years later. At that point, counsel is no longer deciding whether disclosure would help the trial run fairly; counsel is arguing about whether the absence of the document was important enough to undo a conviction.

That is a bad posture in any case. In a mass-casualty terrorism prosecution with multiple agencies, foreign jurisdictions, intelligence sources, and physical-evidence trails, it is worse. The absence of a clean disclosure log does not merely inconvenience reviewers. It hands later advocates a ready-made argument that no one can reliably tell what was known, by whom, and when.

Timer and physical evidence: chain of custody as appellate memory

The Mebo MST-13 timer path has long been one of the most contested parts of the Lockerbie record. The timer fragment history does not need to be retried here to make the evidence-integrity point. Physical evidence survives appellate scrutiny best when every transfer, examination, packaging step, comparison, and expert assumption can be reconstructed without relying on institutional confidence.

Chain-of-custody disputes rarely arrive alone. They interact with expert interpretation, witness memory, disclosure scope, and alternative explanations. A fragment is not just a fragment once lawyers are asking who first saw it, who labeled it, who compared it, what notes were made, what photographs existed, and whether any earlier descriptions conflict with the final theory.

The 2021 appellate refusal means those disputes did not legally unseat the conviction. It does not mean the record is a model for evidence handling. Appellate survival and evidentiary cleanliness are different claims, and only the first was finally resolved against Megrahi in that appeal. [1]

What the appeal court did with the SCCRC’s two grounds

The cleanest way to avoid overstating the case is to put the SCCRC referral and the 2021 answer side by side.

That table is austere, but it is where the legal risk sits. The SCCRC did not merely wave at “concerns.” It referred two grounds. The appeal court did not merely say the concerns were unattractive. It refused the appeal after applying the appellate safety standard. [1][3]

There is a temptation, especially in public accounts of Lockerbie, to let the existence of the referral do more work than it can carry. A referral shows that a statutory review body saw enough potential miscarriage risk to send the case back. It does not predetermine what the court must do. Conversely, the refusal does not erase the verification failures that made the referral possible.

The Lord Advocate later described the conviction as having been “upheld twice in the appeal court” when announcing that fuselage material had been transferred to U.S. custody as evidence for the separate American trial. [6] That formulation is procedurally important, but it should not be misread as an answer to every evidentiary complaint. It states the appellate status of the conviction.

The practical risk lesson

For litigators and knowledge-management teams, the useful lesson is not that defects are harmless if the case is large enough. That would be the wrong lesson. The lesson is that appellate review may tolerate defects that would have been far easier to confront before and during trial.

A trial team that waits until appeal to explain an informant’s incentives is already working from a damaged file. A team that cannot quickly show what reward discussions existed, what was disclosed, and who reviewed the material has turned impeachment management into a post-conviction hazard. A team that treats physical-evidence logging as clerical work is assuming that no later court will need to reconstruct the life of the exhibit under hostile scrutiny.

The same discipline appears in ordinary litigation, not only terrorism prosecutions. The site’s docket-verification pattern at Verify Hearn Docket draws the same line between a claim and the record that supports it. Lockerbie extends that habit from docket posture to evidence integrity: witness-reliability checks, informant-incentive documentation, disclosure logs, and chain-of-custody records are not housekeeping details. They are the record a future court will have to live with.

References

1. Megrahi decision summary, Judiciary of Scotland, January 15, 2021.
2. Timeline: Lockerbie bombing, BBC News.
3. Lockerbie bombing: Megrahi appeal bid allowed, BBC News, March 11, 2020.
4. What I learned when I reopened the Lockerbie bombing files, BBC News, July 2026.
5. Lockerbie court should have been told witness wanted payment, The Guardian, November 25, 2020.
6. Fuselage of Lockerbie plane transferred to US as evidence for trial, Crown Office and Procurator Fiscal Service, December 2024.