Methodology

The complete rulebook, rendered from the versioned source. This is the document a settlement reference is held to: series definitions, formulas, calendar, first-print rule, revision policy, fallbacks, governance.

ETRX Rulebook

US Effective Tariff Rate Index · version 1.0 · effective 2026-08 (draft until first live print)

1. Purpose and benchmark statement

ETRX measures the realized effective duty rate on US imports for consumption — calculated duties as a share of customs value — by country of origin and HS chapter, monthly. It is designed to be usable as a settlement reference: fully rules-based, computed solely from official published US government data, with a published calendar, an immutable first-print convention, and a written fallback policy. No discretion enters the calculation.

2. Series definitions

All series use the same formula (§3) over different slices of US imports for consumption.

SeriesOrigin (Census CTY_CODE)HS chaptersNote
ETRX-USall countries (-)01–97settlement headline; excludes special-provision chapters 98/99
ETRX-US-INCLall countries (-)01–99companion; published, not intended for settlement
ETRX-STEELall countries (-)72+73iron and steel + articles thereof
ETRX-PHARMAall countries (-)30pharmaceutical products
ETRX-APPARELall countries (-)61+62knit + woven apparel
ETRX-CN-85China (5700)85electrical machinery and electronics
ETRX-CN-84China (5700)84machinery and mechanical appliances
ETRX-CN-95China (5700)95toys, games, sports equipment
ETRX-EU-87European Union (0003)87vehicles
ETRX-MX-87Mexico (2010)87vehicles
ETRX-CA-76Canada (1220)76aluminum and articles thereof
ETRX-CA-44Canada (1220)44wood and articles thereof
ETRX-CA-87Canada (1220)87vehicles

New series may be added under §11; existing series definitions are never changed in place.

Settlement classification. ETRX-US, ETRX-CN-85, ETRX-CN-84, ETRX-STEEL, ETRX-APPAREL, ETRX-CA-76 and ETRX-EU-87 are settlement-eligible: they cover slices with substantial monthly customs value and are the series offered as references for financial products. ETRX-US-INCL is informational (published for reconciliation, not for settlement). ETRX-CN-95, ETRX-PHARMA, ETRX-MX-87, ETRX-CA-44 and ETRX-CA-87 are informational pending review: they are published on the same rules but cover narrower slices whose concentration and month-to-month behaviour are assessed before any product may reference them. Classification changes are methodology changes under §11.

3. Formulas

For a slice S (origin × chapters) in statistical month M:

Rates are stored rounded to six decimal places (one hundredth of a basis point). Aggregates sum the underlying dollar values before dividing; they are value-weighted, never averages of rates.

Why calculated duty, not collected cash. Calculated duty is computed by Census from the duty rates applicable at entry. It is fixed at entry and is never restated by later refund litigation or collections timing. The February 2026 Supreme Court ruling that voided the 2025 IEEPA tariffs triggered a refund process exceeding $100B; cash-collections measures will be distorted by those flows for years, while the calculated-duty series records what importers actually owed at entry, cleanly, through the discontinuity.

Scope of "calculated duty." Census computes calculated duty from the tariff-schedule rates applicable at entry, including the additional duties imposed through HTS Chapter 99 provisions (Sections 232 and 301, IEEPA and successor actions). Antidumping and countervailing duties (AD/CVD) are assessed separately by CBP and are not included. ETRX therefore measures the tariff-schedule burden; series on goods with large AD/CVD orders (e.g., Canadian softwood lumber within ETRX-CA-44) understate the total trade-remedy burden, by design and by disclosure.

Why the headline excludes chapters 98/99. HS chapters 98 and 99 are special-provision classifications (US goods returned, repairs, temporary imports), largely duty-free, carrying material customs value. Including them dilutes the measured duty intensity of commercial trade. The settlement headline is chapters 01–97; the inclusive rate is published beside it for reconciliation with totals-based external measures.

4. Data source and provenance

5. Publication calendar

Census publishes monthly trade data with the FT-900 release, typically 35–40 days after the statistical month, at 8:30 a.m. ET on the scheduled date. ETRX publishes no later than 12:00 noon ET on the day of the release (the engine's scheduled run lands well before that; the print is normally available within two hours of the release). Where a print is not published by that time, the delay and its reason are stated on the site. The engine probes daily, so catch-up releases (several months at once, off-schedule dates) print on detection, under the same noon-ET commitment for the day detected. "Business day" means a day on which US Federal Reserve Banks are open. The site footer and latest.json always carry the next expected release date from the official Census schedule.

6. First-print settlement rule

The first value ETRX publishes for a (series, statistical month) is that month's settlement value, permanently. It is recorded in an append-only ledger with its vintage date and is never edited, deleted, or restated.

Backfill caveat (honesty rule): values for months before the first live print were built retroactively from current Census data and therefore embed all Census revisions to date. They are stamped backfill in the ledger and are suitable as history, not as evidence of real-time settlement behavior. True first-print vintages begin with the first live print.

7. Revision policy

8. Fallbacks

9. HS composition changes

HS chapter definitions are stable at the 2-digit level; chapter-level series are robust to sub-heading reclassification within a chapter. Two structural notes are on record:

10. Country and grouping changes

11. Methodology-change governance

12. IOSCO alignment statement

ETRX is administered with the IOSCO Principles for Financial Benchmarks (July 2013) as its reference standard: the methodology is public and rules-based; input data is official government statistics that the administrator cannot influence; the calculation is deterministic and reproducible by any third party from public data; conflicts of interest are structurally minimal (the administrator holds no positions referencing the index); and this rulebook, the calendar, the fallback policy, and the audit trail are published. A formal IOSCO self-assessment will accompany any use of the index as a settlement reference in a listed or cleared product.

13. License, contact, disclaimer