Skip to content
RecessionIndicatorEconomic Observatory
Observatory / Editorial policy

Economic observatory

Editorial policy

Last updated

Daily briefings and update dates

The daily recession briefing selects material changes from comparable saved source records: newer observations, same-period value changes and explicit withdrawals. It connects each change to the indicator’s economic meaning, qualifications and next checks. These briefings are automatically compiled; a named human review is shown only when one has actually taken place.

The selection prioritises recent observation periods. Rounding differences, older historical changes and differences caused by collecting the same source again remain in the full comparison ledger. A same-period value change enters a briefing only when a newer recorded publisher release or observation vintage supports it. These selection rules do not establish statistical significance.

A briefing is first published when it becomes available in this archive. Its update date advances when a substantive development is added or its content is corrected. The evidence day, publisher’s release date and period measured remain separately labelled. An unchanged source check does not create a new article or refresh its publication date. Backfilled evidence retains its original observation chronology and its actual, later briefing publication date. Selection corrections retain the original publication date, explain what changed and preserve links to the earlier evidence.

Regional and indicator pages show the latest substantive editorial or relevant briefing update. Historical snapshots stay fixed, and changes to unrelated indicators do not refresh an individual guide’s date. Subscribe to the briefing RSS feed.

Responsibility and review

Recession Indicator is maintained by a team of researchers. Correction requests can be submitted through the private correction form. The project distinguishes responsibility for maintaining the site from review of a particular result.

Software checks input availability, units and dates. AI-assisted research briefs refer to a completed snapshot. Automated checks can detect inconsistencies in supplied numbers while leaving judgement about interpretation and source support open. Human review is credited only where a publication identifies the person, scope and date of the review.

Evidence and uncertainty

A quantitative claim should identify its region, indicator, unit, observation period and source. Briefs distinguish observations used in calculations from additional source commentary. Each regional outlook should present supporting evidence, counter-evidence and coverage limits together.

Current activity, leading warnings and longer-term vulnerabilities answer different questions. Missing or withdrawn observations reduce coverage. The 0–100 concern score measures observed strain. Probability estimates require their own outcome, horizon and validation. See the methods and forecast studies.

Corrections and versions

Readers can flag an error through the private correction form. A published correction should identify the affected page and snapshot, original claim, corrected claim, supporting evidence, correction date and person responsible. Earlier evidence stays available for comparison.

The release-change ledger records changes between snapshots that may reflect a later observation or source revision. The corrections register records editorial errors.

Commercial relationships

Any sponsorship or commercial work affecting a publication should be disclosed with that publication. Source selection, scoring and corrections follow the published evidence and methods. The ownership and disclosures page records the information available about the project.