The IPCC Does Not Carry Out Any Research. The Document Everyone Quotes Is Agreed Line by Line With Member Governments in the Room.

by Arjun Mehta · September 8, 2026

Science & Technology Writer, The Globe Ledger

When an Intergovernmental Panel on Climate Change report lands, the coverage reduces to a formula: scientists said. The formula is half right and half a misdescription of a procedure, and the procedure is written down, published, and more interesting than the shorthand.

Start with what the Panel is. Its own factsheet, revised in January 2024, states plainly: “It does not conduct its own scientific research.” The same document records that “The IPCC was set up in 1988 by the World Meteorological Organization (WMO) and United Nations Environment Programme (UNEP),” that “It currently has 195 members,” and that “IPCC assessments are written by hundreds of leading scientists who volunteer their time and expertise.” No laboratory, no observing network, no research budget for original science. The IPCC is an assessment body whose product is a reading of everyone else's work.

The Principles Governing IPCC Work — approved at Vienna on 1 October 1998 and amended at sessions in 2003, 2006, 2012 and 2013 — put the mandate in one sentence: to assess “on a comprehensive, objective, open and transparent basis the scientific, technical and socio-economic information relevant to understanding the scientific basis of risk of human-induced climate change, its potential impacts and options for adaptation and mitigation.” The same document sets the boundary: “IPCC reports should be neutral with respect to policy, although they may need to deal objectively with scientific, technical and socio-economic factors relevant to the application of particular policies.” The factsheet renders the same rule in five words: the assessments are “policy-relevant but not policy prescriptive.”

What gets assessed is published literature, at a scale most readers never see quantified. The Panel's factsheet on literature states that “The Sixth Assessment Report cited over 66 000 sources.” It also records that the assessment reaches beyond journals — “Sources other than scientific journals also provide crucial information for a comprehensive assessment,” including reports from governments, industry, research institutions, international organisations and conference proceedings — and that authors using such material “have a special responsibility to ensure its quality and validity.” There is a hard stop on the intake: literature “has to be accepted for publication approximately one to three months before completion of the final draft.”

Now the part that surprises people. The Panel's procedures use three different verbs for three different things, and it defines all three. Appendix A to the Principles Governing IPCC Work — adopted at the Fifteenth Session in San José in April 1999 and, in the version consulted here, amended through the Panel's Thirty-Seventh Session in Batumi, 14 to 18 October 2013 — says that approval “signifies that the material has been subject to detailed, line by line discussion and agreement.” It defines adoption as a “process of endorsement section by section (and not line by line) used for the longer report of the Synthesis Report.” And it says acceptance “signifies that the material has not been subject to line by line discussion and agreement, but nevertheless presents a comprehensive, objective and balanced view.”

Apply those definitions to an assessment report and the architecture becomes clear. The Summary for Policymakers — the short document that produces the headlines, the graphs and the political arguments — is approved. The underlying chapters, the thousands of pages containing the actual assessment, are accepted. The distinction is not decorative. It means the summary has been through a sentence-by-sentence negotiation with member governments, and the report beneath it has not.

The Panel's factsheet on approval, issued in July 2021, spells out what happens in the room: “Approval signifies that the material has been subject to detailed, line-by-line discussion, leading to agreement among the participating IPCC member countries, in consultation with the scientists responsible for drafting the report.” Delegations propose changes to sentences. Authors sit at the front. Nothing moves to the next line until the line in front of them is agreed.

Two guard rails stop that becoming a vote on the science. Appendix A requires that “Coordinating Lead Authors should be consulted in order to ensure that the Summary for Policymakers is fully consistent with the findings in the main report.” The factsheet describes the same constraint from the other direction: the presence of the assessment's authors “ensures that any changes to the SPM are consistent with the underlying report and are scientifically robust.” A government can argue that a sentence is unclear, unbalanced or unsupported. It cannot introduce a finding that the accepted chapters do not carry, because the authors who wrote those chapters are the ones being asked whether the new wording still tracks them.

The factsheet also states the process's purpose in terms of the text rather than the politics: approval “strengthens the SPM by ensuring that SPM statements are as direct, clear and unambiguous as possible in summarizing the material contained in the corresponding Working Group Assessment Report or Special Report.”

There is a further step that is routinely collapsed in reporting. A Working Group approves its own Summary for Policymakers; for that summary to become an official IPCC report it must then be accepted at a Session of the Panel. And the sequencing is one-directional, for a reason the factsheet gives in the same breath: “Because the Working Group approval process is open to all governments, Working Group approval of an SPM means that the Panel cannot change it.” What the Panel does instead, in the factsheet's words, is “review the Report at a Session, note any substantial disagreements, and formally accept it.” The plenary of 195 members can receive the text and record dissent. It cannot rewrite what the Working Group settled.

Long before any of that, the drafts have been through a review process whose volume is the least-cited statistic in climate coverage. Every interested expert may comment on the first order draft. The second order draft goes to experts and governments simultaneously. The final draft goes to governments for a last written round on the summary. For the Sixth Assessment Report, the Panel's review factsheet as revised in July 2021 records 23,462 comments on Working Group I's first order draft and 51,387 on its second; 16,348 and 40,293 for Working Group II; 21,703 and 32,665 for Working Group III, the second-order figures covering experts and governments together. That is roughly 186,000 comments on the three Working Group contributions alone. The same table leaves the Synthesis Report's figure, and the overall total, marked to be determined — so the full count for the assessment is higher than that, by an amount the factsheet does not state.

Those comments are not filed and forgotten. Review Editors, the factsheet states, “ensure that all substantive comments received during review are given appropriate consideration by the author teams,” and “All comments submitted through this review process are addressed by authors, and a written response is provided to each comment on the First Order and Second Order Drafts for later publication.” The audit trail is the point: an assessment nobody can check is not an assessment.

The Panel's decision rule matches the design. The Principles provide that “In taking decisions, and approving, adopting and accepting reports, the Panel, its Working Groups and any Task Forces shall use all best endeavours to reach consensus,” and that “Participation in the work of the IPCC is open to all WMO and UN Member countries.” This is a body of governments that assesses science, not a scientific academy that advises governments, and the line-by-line step is precisely where that identity is asserted. The same states that will later negotiate over the findings have already signed off, word by word, on the sentences summarising them.

That is what makes the misdescription worth correcting in both directions. Anyone claiming the Summary for Policymakers is a purely scientific document is ignoring the procedure. Anyone claiming it is politically written is ignoring the constraint that every approved sentence must remain consistent with an underlying report the governments did not draft and cannot amend — a report written by volunteer authors and assessed against tens of thousands of published sources.

The vocabulary is doing exact work, and it is public. Approved means fought over line by line. Adopted means endorsed section by section. Accepted means comprehensive, objective and balanced, and never negotiated at all. Three words, three procedures, one appendix — and a distinction that decides how much weight any given IPCC page can carry.