Every country page on shortage.life carries a single headline number — the stress index — and that number does a great deal of work. It must be comparable across forty-seven countries and nine commodities, stable enough to track week to week, and transparent enough that a journalist or a machine reader can reconstruct it. This page documents how it is built, and what the v3.2 build changed.
The index is a weighted composite of five inputs. Each input is normalised to a 0–1 scale against the full panel, multiplied by its weight, and summed. The result is itself a 0–1 figure: closer to one means more exposed, closer to zero means more resilient. It is a screening signal, not a probability and not a forecast.
The five inputs.
The four weights — 0.30, 0.25, 0.20, 0.15 — front-load the structural inputs. Event pressure carries only 0.10 on purpose: a flare of headlines should nudge the index, not dominate it.
The index is a screening signal, not a probability — built to be comparable, stable, and reconstructable.
Spare capacity, defined.
The OPEC+ spare-capacity series that feeds the energy dashboards uses a deliberately strict definition: production that can be brought online within thirty days and sustained for at least ninety. This is consistent with the IEA, the OPEC Secretariat and Argus, and excludes the larger nameplate buffer that press coverage often quotes. The gap between the two grows under stress, when maintenance backlogs and field-decline rates start to bite.
Data cadence.
The index is only as fresh as its slowest input. Each source has its own publication rhythm; we ingest on that rhythm and timestamp every observation with its date of observation, its date of publication, and our date of fetch.
Frequent questions.
What does a stress index of 0.80 actually mean?
It is a unitless composite on a 0–1 scale, not a probability. A reading above 0.75 places a country in the upper decile of the panel — high import dependence, thin autonomy, or concentrated suppliers, usually two of the three at once. It is a screening signal, not a forecast.
Why did the weights change in v3.2?
We added fiscal headroom as a fifth input and removed a redundant trade-balance term that correlated too tightly with import dependence. The four-weight scheme above is the result. Every historical series was recomputed so the back-history stays consistent.
How is spare capacity defined?
As production that can be brought online within thirty days and sustained for at least ninety. This is the definition used by the IEA, the OPEC Secretariat and Argus — and it is deliberately stricter than the nameplate buffer often quoted in the press.
Can I cite these numbers?
The live platform carries a DEMO watermark and is not for citation. The production dataset is published under CC-BY-4.0 with a required backlink — see the terms.
Methodology v3.2 supersedes v3.1 (2026-02). Versioned changes are logged with every build. Questions to editorial@shortage.life.