How the score is built.

Every metric is normalized to [1, 100] before it enters aggregation. Two regimes apply, depending on the shape of the underlying series — the bounds table below shows which each metric uses.

Log-space min-max is used for unbounded or wide-range metrics — volumes, counts, capacity — where values span orders of magnitude. The log transform makes normalization unit-invariant and appropriate for quantities that grow exponentially rather than linearly. Linear min-max is used for naturally bounded metrics such as ratios and percentages, where the raw value already lives on a finite scale.

All bounds are published and versioned. The served methodology is v1.3; each metric also carries its own boundsVersion in the table below — the amendment that last set that metric's normalization basis, which may leave min/max unchanged (e.g. ERC-8004 is at 1.2 for the dual-cap basis while most metrics remain at 1.1). This is the replicability commitment — anyone with the raw data can reproduce the normalization step.

The bounds are an editorial choice: tight bounds make a metric swing more. Ours are published, versioned, and varied in the sensitivity analysis.

The “launch N₀” column shows where each metric sits at launch on its 1–100 scale. These positions are a deliberate snapshot calibrated against verified anchor values — not an accident, and not an implication that the metric is “good” or “bad.” Some metrics launch low (nascent) and some high (mature) by design; the bounds are set to capture realistic future growth. Per-metric specifics — floor, ceiling, cadence, and source — are in the table.

Legal rail = LRRS (see LRRS section). No scored metrics on the Legal rail.

Every metric must pass six tests before it becomes an MEI input. These criteria were applied strictly during the v1.0 redesign — only a fraction of the metrics considered during development qualified; the rest were excluded as named gaps or contextual signals rather than index inputs.

The Legal Rail Readiness Score (LRRS) measures how far the world's legal frameworks have caught up with the machine economy — GDP-weighted across jurisdictions and five legal categories. It is reported as an integer from 0 to 100; the underlying float feeds the MEI's Legal component directly.

The model uses five categories, each with a rank-sum weight: legal identity (3/10), stablecoin frameworks (3/10), regulatory sandboxes (1/6), machine-economy free zones (1/6), and multilateral instruments (1/15). These are not milestone points — the LRRS is explicitly a coverage model, not a tally.

• Legal identity — whether a jurisdiction legally recognizes a machine-readable identity for autonomous agents: the capacity to be identified and to bear obligations within its legal system.weight 3/10
• Stablecoin — frameworks that address fiat-referenced tokens (stablecoins) as such and govern either their issuance or services in them — offering, exchange, custody, transfer or payment. They are the settlement layer behind most machine payments. MiCA is the leading example in force; the US GENIUS Act counts from the day it takes effect, 18 January 2027.weight 3/10
• Sandbox — government-run regimes that formally admit participants to test AI or machine-economy systems for a bounded period, with legal effect and a formal exit. A general financial or sectoral sandbox counts once its regulator has admitted a participant whose system under test uses AI.weight 1/6
• Free zone — machine-economy free zones: special jurisdictions with bespoke legal regimes purpose-built for autonomous agents to hold assets, contract, and transact.weight 1/6
• Multilateral — cross-border instruments (treaties, model laws, international standards) that coordinate machine-economy legal treatment across jurisdictions rather than within a single one.weight 1/15

Coverage is built from per-jurisdiction × per-category cells. Each cell has one of three values. 0 — no instrument of the jurisdiction's own is in force, including a law that has passed but whose operative provisions have not yet commenced. 0.5 — the jurisdiction's own instrument is in force, but operation is not yet evidenced. 1.0 — official operating evidence exists and is current: licences issued, participants admitted, live register entries. A sandbox stays at 1.0 while a participant's test is live or its newest operating record is less than 12 months old, then falls back to 0.5 while the regime stands. A jurisdiction's coverage in a category is its cell value; category coverage is the GDP-weighted sum across jurisdictions; the LRRS is the weighted sum across categories.

5 November 2026: how four sandbox and stablecoin rules are read (Rulings #21–#24). A testing regime's formal exit is a record its rules require at the ordinary end of a test: a report by or to the authority on the participant or its cohort, proof of the exit actions, or a decision on the test's outcome. A test differs from a type approval by its test purpose, its test term and that exit. A regime's level is the level of the instrument that sets its admission conditions, test term, legal effect and exit, so national rules administered by local authorities are national. A stablecoin provision counts only if it governs issuance or services on conditions specific to such tokens; one that only caps, delays, monitors, taxes or bans dealings does not. These rulings clarify existing rules: no stored status changed, and the methodology version is unchanged.

Some EU-level instruments constitute an operating regime in each member state — MiCA is the leading example — and the model propagates them to all 27 members. EU instruments that only oblige member states to create a regime confer no status and propagate nothing; each member scores its own national instrument, or zero. Where both apply, a stronger national instrument can override: cell status is the maximum across national and propagated EU instruments.

GDP weighting uses IMF World Economic Outlook data, frozen per methodology version (current vintage: WEO-2026-04). The weighting is reproducible and does not drift between releases. The LRRS is hand-curated. Every instrument already in the table is re-verified monthly, so a change to one enters the score within 35 days; a full search for new instruments runs at least every 95 days, so a new instrument enters within 95 days of taking effect; dated events already recorded apply on the day. Each finding's lag is published, and history is re-dated to the legal events at the next methodology version. Every search is recorded in the coverage log.

The LRRS is decomposition-first: the category bars above show where readiness exists before the single headline number. The dashed ceiling line marks the combined GDP of the surveyed jurisdictions — the 25 largest economies, all 27 EU members, and Singapore, the UAE, Hong Kong and the Marshall Islands, 89.1% of world GDP — the practical upper bound for coverage in the LRRS universe. Jurisdictions outside it are unverified and scored 0, so the score is understated by at most 10.9 points (realistically far less).

See the per-jurisdiction status map on the Index page.

Some dimensions of the machine economy do not yet have a metric that passes all six inclusion criteria — or cross a disclosed construct boundary the current proxy cannot reach. Rather than filling those slots with weak proxies, the MEI and LRRS name them explicitly in one register. Each entry states what Tier-1 source would close it.

These are real parts of the machine economy the indices do not yet measure. Naming them is a credibility choice: the platform states what it cannot yet quantify rather than substituting unverified numbers.

The LRRS category bars in the LRRS section reflect live coverage — categories at zero indicate no jurisdiction in scope yet has an operational framework for that category.