How the index is built

The Intelrift Geoeconomic Index (IGE) measures, country by country and every quarter, the structural geoeconomic risk of an economy on a 0–100 scale where higher means more risk. It is an exposure index, not a forecast: it describes how much damage a route closure, a tightening of sanctions, a swing in commodity prices, an institutional crisis or an episode of sovereign stress would do to an economy.

Coded sanctions regimes

Tier per jurisdiction and authority; the country takes the maximum. 100 comprehensive embargo · 60 sectoral regime · 30 targeted regime.

Each input is oriented so that higher means more risk, winsorised at the 5th and 95th percentiles of the countries with data in the edition and scaled min–max to 0–100 within those bounds. Fixed-scale inputs (the sanctions regime tier) are not rescaled. An input that is identical for every country carries no information and is dropped for that edition.

p5–p95

A sub-index is the mean of its available inputs provided its required inputs are present and its minimum is met. The index is the weighted mean of available sub-indices with renormalised weights and requires at least 3 of 5; otherwise the country is shown without coverage.

Minimum coverage: 3 of 5 sub-indices.

One edition per calendar quarter. The data cut-off is the last day of data considered; annual indicators take the latest period not later than the cut-off and sanctions lists the most recent snapshot at computation time. Published editions are never recomputed; a method correction is released as a new version and applied from the next edition onwards. Besides the quarterly editions, the index publishes yearly retrospectives (2019 onwards, 31 December cut-off) computed with the current method so the historical series exists from day one: they use the current snapshot of the sanctions lists and carry no newsroom events, and are flagged as such in the archive and the API (kind = retrospective). The first quarterly edition's change is measured against the latest retrospective.

• Chokepoint exposure is estimated from World Bank partner shares and a documented routing matrix; it is not measured from AIS transit data, whose licence does not allow commercial use. Transit dependence of landlocked countries enters as a separate signal (UN LLDC list).
• Commodity concentration is computed over the five World Bank product groups, not tariff lines, so it understates concentration within each group; natural resource rents and terms-of-trade volatility complement it. UNCTAD's concentration index awaits licence confirmation.
• Sovereign stress does not include bond spreads or CDS, which have no open-licence source in v1 (EMBI indices are J.P. Morgan property); institutional fragility has no election calendar either (Freedom House and similar sources require a commercial licence).
• Reserve-adequacy ratios (months of imports, broad money or short-term debt over reserves) are not applied to reserve-currency issuers (dollar, euro, yen, sterling, renminbi, Swiss franc): they cannot run out of the currency the rest of the world holds, and the euro area keeps its reserves at the ECB. Their sovereign stress is measured with the remaining inputs.
• Own inputs (newsroom events, incidents) depend on Intelrift's news coverage and may under-represent countries with little press in the languages we ingest.
• The sanctions regime table (OFAC, EU, UN, UK) is hand-coded from the official programme pages and carries a review date; designation and export-control counts come straight from the primary lists.

Every country and every edition carries a SHA-256 hash computed over its raw inputs, sources, periods, winsorisation bounds and weights. The public API returns the full breakdown per country; download the CSV, recompute with the published method and compare the hash.

GET /api/v1/geo-index · GET /api/v1/geo-index?format=csv · GET /api/v1/geo-index/countries/{code} · lib/geo-index