What we publish, and what we will not

Every research output passes through four stages before publication. We do not publish work that has not cleared all four.

Stage 1 — Commissioning. The Programme Director scopes a question against the published programme agenda. Commissioning notes are saved on file but are not published.

Stage 2 — Drafting. Author or author team produces a complete first draft including a quantitative data appendix. No interview material is paraphrased; direct quotation requires written attribution or anonymisation with the source's consent.

Stage 3 — Standing-committee review. The draft is circulated to the programme's standing committee — typically four to seven members. Reviewers have ten working days to respond. Reviewer identities are published on the final paper.

Stage 4 — Editorial review. The Institute's editorial team performs sub-editorial review, fact-checking, and citation verification. Citations are spot-checked end-to-end on every paper.

The Institute is constituted as a non-profit foundation under Belgian law. We are financed by a diversified base of individual donors, philanthropic foundations, and subscriber revenue from The Long View.

We do not accept funding from sovereign states, single corporations, or political parties. We accept funding from philanthropic foundations and individual donors subject to the following caps and disclosures:

Donor caps. No single donor or foundation may represent more than 6% of the Institute's annual revenue. The Institute's auditor verifies this annually.

Disclosure. Donors who provided more than €25,000 in any year are listed on the annual report. Specific funding sources for any working paper are disclosed on the published version of that paper.

No conditional funding. The Institute does not accept funding conditional on publishing a specific conclusion, on consultation rights over editorial direction, or on access to embargoed material outside the press protocol.

Every resident and non-resident contributor maintains a current disclosure of paid affiliations, board seats, and securities holdings exceeding €10,000 in any single position.

Disclosures are reviewed by the Institute's editorial team before each publication. Where a contributor has a material conflict with the subject matter of a paper, they are recused from primary authorship and the conflict is noted on the published paper.

We do not consider general European citizenship, prior public service, or academic affiliation to constitute conflicts. We do consider current paid retainers, equity positions, and direct policy advocacy roles to be material conflicts requiring disclosure.

Subscribed press contacts receive embargoed access to forthcoming working papers 24 hours before public release. Embargo terms are simple: do not publish, quote, or attribute material under embargo, including to colleagues outside your masthead.

Breaking an embargo results in removal from the embargo list for six months on the first incident, and permanent removal on a second incident. We have not exercised either sanction to date.

Embargoed material is shared as a personalised PDF with the recipient's email watermarked into the metadata. We do not pursue leak forensics absent a substantive editorial-integrity reason to do so.

Errors of fact are corrected on the published version with a dated correction note at the top of the affected section. The original wording is preserved in a footnote so that subsequent readers can verify what was changed.

Substantive corrections — meaning corrections that change a numerical finding, a sourced quotation, or a policy recommendation — are flagged in the Sunday edition of The Long View following the correction.

Editorial improvements that do not change meaning (typos, punctuation, link rot) are made silently. No-meaning changes never affect a paper's headline citation or DOI.

All factual claims are sourced via footnote or inline citation. Quantitative claims link to a public data table that reproduces the calculation. Qualitative claims are attributed.

Interviews on background — meaning interviews where the source's identity is withheld from publication — are conducted under written ground rules signed by the interviewer and the source. The Institute's archive of background interview notes is held under encryption and is not shared internally or externally.

We do not republish material without permission. We accept citations of our work under the CC-BY-4.0 license, which requires attribution to the Institute and the named authors. Commercial reuse requires written permission.

We use language models and quantitative data tools in our research workflow. We do not allow language-model output to be published as the work of a named author without substantive human verification of every claim, citation, and quotation.

Where a language model has been used to draft a section that has been substantially preserved in the final version, the use is disclosed in the methodology appendix.

Quantitative analysis is performed in versioned scripts (R, Python) that are published alongside the paper or made available on request. We do not use language models to generate numerical results that are presented as primary findings.