SIJAL · DECISION FILE Record 6 · Final English
Record 6 · Final

Can an evidence-based review of Regulation (EU) 2021/821 reveal an overlooked, verifiable finding?

The question

Can an evidence-based review of Regulation (EU) 2021/821 reveal an overlooked, verifiable finding?

The verdict

No.

The Court answered the question in the negative.

Strongly supported

3 of 3 lenses · read from the Court record

What to do

Treat the low denial rate under Regulation (EU) 2021/821 as an insufficient basis for concluding that the regulation lacks oversight effectiveness, and instead direct analytical effort toward documenting the self-selection mechanism before drawing conclusions about regulatory opacity.

The weight of the judges

6 for12 against

12 of 18 judge votes against (67%)

  • Lens 1 2-4
  • Lens 2 2-4
  • Lens 3 2-4

Finalized 12 Oct 2026, 18:20 UTC

Why this verdict

The Court's prevailing findings establish that authorised dual-use exports rose from €38.5 billion in 2021 to €57.3 billion in 2022, with denials representing only 0.03-0.04% of total extra-EU exports, and that this macro-level growth directly supports the position that the regulation does not suppress legitimate exports in aggregate, meaning any claim of systemic suppression must first account for the possibility that exporters are self-selecting out of the application process by withdrawing or restructuring transactions before submission, thereby reducing the denial numerator without any failure of oversight.

Additionally, the Court record identifies a traceable, stakeholder-confirmed harm in the IT and research sectors, documented chilling effects on intrusion software controls (SIPRI 2017), which represents a more empirically grounded and actionable concern than contested cross-period comparisons or unverified suppression claims. The prevailing side carried all three goals in a 12-6 judge-vote distribution and a 3-0 goal tally, reflecting the consistent weight of these findings across the full record.

The qualification the Court preserved is that whether the self-selection mechanism materially suppresses the denial numerator remains unsubstantiated in the record, as no fact documents withdrawal rates, restructured transaction patterns, or empirical evidence of self-selection behaviour under Regulation (EU) 2021/821; and separately, whether Article 4 catch-all variation across Member States constitutes intended regulatory discretion or an unintended structural enforcement gap cannot be resolved from the available record.

The proponents' most substantiated factual anchor is the documented disparity between the scale of authorised dual-use trade (€57.3 billion in 2022) and the absolute number of denials (831 in 2022), combined with the documented absence of a prior annual report, the first was published January 2025, creating a measurable three-year implementation gap with no auditable baseline. This gap is grounded in the record and constitutes the strongest evidentiary basis for the proponents' core claim that the system contains overlooked structural measurement deficiencies.

Lens 1

Its opening argument

Regulation (EU) 2021/821 mandates transparency yet produces a structurally opaque system: only 568 denials were recorded against EUR 45.5 billion in applications, yet Article 4 implementation varies sharply across Member States, some legislating catch-all controls, others enacting nothing comparable. The first annual…
Read its full case

  • The causal link between EU dual-use export license denial rates and surveillance technology proliferation in importing countries is entirely absent from the record; no fact documents this relationship.

    Raised by 18 judges
  • No fact in the record measures the effect of EU 2021/821 compliance burdens on research collaboration or innovation, dimensions explicitly named in the debate goal but entirely absent from the retrieved evidence.

    Raised by 5 judges
  • No fact in the record provides SME-disaggregated data on dual-use export license applicant counts, participation rates, or dropout rates under Regulation (EU) 2021/821. Two claims, and 270 cannot be resolved without this data.

    Raised by 4 judges
  • The quantitative rate of pre-submission withdrawal of dual-use license applications, central to a claim's argument that low denial rates reflect effective pre-screening, is not documented in any the record fact.

    Raised by 3 judges
  • No benchmark denial-rate data from a comparable export control regime (e.g., Wassenaar Arrangement member states, US BIS) is present in the record, making it impossible to determine whether the 0.04% denial rate is anomalous or normal relative to peer regimes.

    Raised by 2 judges
  • The causal link between EU dual-use export license denial rates and surveillance technology proliferation in importing countries is completely unsupported by any fact in the record. No fact traces denied export licenses to downstream surveillance outcomes in destination countries.

    Raised by 1 judge

THE JUDGES

Three lenses. Six judges in each.

METHOD

Independent aggregated judgment, as in the Good Judgment Project (IARPA ACE programme)

1

Evaluate the evidence concerning the regulation's security and non-proliferation objectives, implementation, licensing decisions, and enforcement. Identify measurable outcomes, documented implementati

2-4

Won by the case against, majority.

2

Investigate whether compliance burdens are associated with measurable effects on legitimate exports, research collaboration, innovation, or competitiveness. Examine differences across company sizes, s

2-4

Won by the case against, majority.

3

Integrate evidence across official evaluations, regulatory reporting methodologies, licensing data, stakeholder consultations, and independent research. Search for overlooked relationships, contradict

2-4

Won by the case against, majority.

Decision record 6 Finalized 12 Oct 2026, 18:20 UTC Fingerprint 37f744a8f330631fab3bc653794eed77cdc96139df016a58fdf06669b90ffa31

Argument: cross-examination debate, with an assigned devil's advocate. Adjudication: independent aggregated judgment, as in the Good Judgment Project (IARPA ACE programme). Report: ICD 203 analytic standards; the recommendation is set out in Toulmin's structure of argument.

Appendix: the judges' rulings