About this index
Think tanks publish a great deal of serious work and most of it never reaches anyone outside the field. Where it is indexed well, it sits behind a library subscription. This is the free, public version: 5,666 publications from 600 institutions, organised so you can see who is arguing with whom.
What we hold
Titles, authors, institutions, dates, links, document types, topics and regions. Facts about publications. We never host or republish anyone's work. Every record here is a pointer to the institution's own page, and every link goes to them.
Where an institution's own terms permit it we also show their published abstract or a short attributed quotation. Until someone here has actually read those terms, a source sits at the most restrictive setting and we show neither. Right now 0 of 604 sources have been through that review, which is why most records here are a title and a link.
How records are selected
We read each institution's own feed or sitemap — the files they publish for machines — and never scrape past a robots.txt rule. Where an institution blocks automated readers, we leave it alone and ask rather than work around it.
Three things are then filtered out:
- Press clippings. “X quoted in the Financial Times” is a mention of a researcher, not a publication, and the thing on the other end belongs to a newspaper.
- Anything hosted elsewhere. Institutions list their staff's journal articles and op-eds in their own feeds. Those are real work, but we would be sending you to a paywall under the institution's name.
- Anything outside the three topics. These are whole-institution feeds, so most of what arrives is out of scope. When the signal is weak the record waits in a review queue instead of being published — a smaller index is better than one with the wrong things confidently filed.
Dates, and the ones marked “c.”
Many institutions publish no date in the page metadata. Where that happens we fall back to the lastmod in their own sitemap, which is when the page last changed rather than when the work came out. Those dates are written c. 8 May 2026, the way a scholar marks an approximate one. Roughly half of the index is marked this way, and we would rather say so than present a modification date as a publication date.
Yes, No and Both ways
On a question page a title occasionally carries a coloured marker. It appears only where the title itself states a position in a construction we can point at — “the case for”, “the myth of”, “the trade-offs of” — and hovering it shows the exact phrase.
Almost no titles qualify, and that is deliberate. We hold titles and links, not articles, so we cannot read a paper and decide what it argues. An earlier and looser version of this read “Triple Threat: America, China and Russia on the Path to Conflict?” as a yes, because the words “path to” appear in it. A confident colour is a strong claim, and a wrong one puts a position in an institution's mouth that it never took. So the marker is rare by design.
What this does not do
No subscriptions, no paywall, no account. No comments, no forum. We do not host documents and we do not summarise them in our own words. If you want the work, the link goes to the people who did it.
Coverage
Worldwide in ambition. In this first version we cover three topics: security and defence, energy and climate, and international economics and trade.
Corrections
If you work at one of these institutions and something here is wrong, or you would rather not be indexed, write to [email protected] and it will be changed. Each institution's page shows exactly which of their files we read.