GTA Coverage & Scope

Every record is written from an official source and cleared by three people

Global Trade Alert works as a public newsroom for trade and industrial policy. Analysts watch what governments themselves publish, verify each change against the text that announces it, and enter it into a record anyone can read. More than 97% of entries cite an official government source.

Every automated lead is checked by an analyst before it becomes a record

Automated collection sweeps official gazettes, ministry sites and regulatory registers for activity that falls inside GTA's scope. Everything it surfaces is a lead. An analyst reads each one against the source document and writes the entry by hand.

Analysts also monitor on their own account. Where official publication is slow, scattered or hard to parse, they subscribe to specialist legal and trade services, including boutique law firms with local practice, and use them to catch changes an automated sweep would miss.

The team is assembled for language and institutional knowledge

GTA's analysts work remotely across time zones, recruited for the languages and the institutional knowledge the coverage requires rather than for one location. Reading a ministry notice in the original, and knowing which agency in that jurisdiction actually holds the power to issue a given instrument, is what makes coverage of 63 jurisdictions possible.

Three desks see every record before it publishes

GTA applies the six-eyes principle. No entry reaches the public database on one person's judgement.

1. Analyst

   Reads the official source and writes the first draft of the entry directly into the database: what the government did, to which products, from when, and in which direction it moves market access.

2. Senior analyst

   Reviews the draft against the same source. Scope, instrument type, products, dates and direction of change are each checked. The draft goes back for correction until both analysts agree the write-up is right.

3. Publication desk

   Takes the agreed entry and publishes it to the live dataset.

GTA publishes on a daily cycle

Between 20 and 30 new and updated records reach the public database on every business day, drawn from across the 63 jurisdictions on the standing list. Coverage begins in November 2008 and has been maintained continuously since.

The record carries no commentary

A GTA entry states what a government did and links to the document that says so. It does not argue whether the measure was wise, and it does not report what commentators made of it. Working from official texts and legal instruments is what lets the same standard apply to Washington, Brussels and Beijing alike, and it is why any reader can follow the source link and check the entry against it.

Sixty-three jurisdictions carry a standing analyst assignment

Global Trade Alert monitors 63 jurisdictions on a standing basis. Each one has a named analyst assigned to it, and a jurisdiction without an assigned analyst is not monitored. Together they account for the bulk of world trade and for most of the trade and industrial policy activity that moves it.

Coverage runs at two intensities. Thirty-six jurisdictions sit on the priority tier, where policy activity is heaviest and analyst attention is concentrated. The remaining 27 are monitored on a regular cycle.

Every G20 member state sits on the priority tier, as does the European Union. So do the six largest economies of Southeast Asia.

EU-level measures bind all 27 member states, so a measure adopted in Brussels enters the record whichever member state it lands in. The member states listed above are those whose own national measures GTA records as well.

Two regions are covered thinly

In South America the standing list runs to Argentina, Brazil, Chile and Peru. Colombia and the smaller economies of the continent are not on it.

In Africa six economies are covered: South Africa, Nigeria, Kenya and Ghana, plus Egypt and Morocco on the Mediterranean. The rest of sub-Saharan Africa is not.

Seven instrument families, recorded in one database

GTA records the full range of instruments a government can use to change the terms on which foreigners trade and invest. One search returns a tariff change, an export licensing requirement, a subsidy and a local-content rule on the same product, where a user would otherwise need four separate sources.

Instruments follow UNCTAD's MAST classification for non-tariff measures wherever it applies. Tariffs, foreign direct investment and labour migration sit outside that classification and carry GTA's own groupings.

Five classes of measure fall outside the database

Every record names who acted, what they did, and how big the change was

A record starts with a state act: the government announcement itself, dated, sourced, and titled with the implementing jurisdiction named up front.

A state act contains one or more interventions, each carrying a single policy instrument. Most announcements produce one intervention. An announcement produces several when it bundles distinct instruments, applies at different intensities to different trading partners or different firms, or moves market access in both directions at once.

Every record links to the source it was built from. Where no official statement exists, a record needs at least three reputable, independent sources in agreement before it is entered. For where those sources come from and who checks each entry, see How We Monitor.

Every record carries a reading on which way market access moved

GTA applies one test before a measure enters the database at all. Does the change alter market access conditions differently for foreign firms than for domestic ones? GTA calls this the relative treatment test. A measure that moves conditions equally for everyone fails it and is not recorded, which is why broad tax reform and general infrastructure spending do not appear.

Measures that pass the test carry the GTA evaluation, a traffic light on the direction of the change.

Green records group as liberalising. Amber and red group as restrictive.

Amber is deliberately narrow. It covers the cases where the evidence supports discrimination but leaves real ambiguity: a trade-defence case still under investigation, or a change whose before and after cannot be compared because the units of account differ. Most records resolve to green or red.

The evaluation is what turns a list of policy actions into a directional signal. A user reading the record learns which way each government action moved market access for the firms that trade with it.

Five datasets cover what GTA leaves out

GTA is not the right source for everything. Pairing it with the right dataset for what it excludes gets you a complete picture faster than making GTA do a job it was not built for.

For technical and sanitary notifications, use WTO ePing or I-TIP. ePing carries the sanitary and technical-barrier notifications GTA excludes, updated daily from a platform covering notifications back to 1995; I-TIP covers WTO members' formally notified measures. Neither judges direction of change; both are the notification record GTA does not aim to duplicate.

For exchange-rate regimes and capital-account rules, use the IMF's AREAER, its Annual Report on Exchange Arrangements and Exchange Restrictions. It is the authoritative annual record across the IMF's entire membership, staff-verified, and it covers ground GTA does not: monetary and exchange-rate policy rather than trade and investment measures.

For the underlying trade-flow numbers, use ITC Trade Map or Trade Data Monitor. Trade Map covers 220-plus countries and territories with the statistics themselves, not a policy reading of them. It is the layer beneath a GTA finding when you need the trade values a measure actually affects.

For the negotiated content of trade agreements, use the World Bank Deep Trade Agreements database. It codes what governments agreed together across some 400 preferential agreements; GTA records what they then do unilaterally, which is close to the mirror image.

For country-risk scores and forecasts, use Oxford Economics. Its risk and macro-forecasting products are built to be forward-looking. They answer "how risky is this market"; GTA answers "what did this government just do, to which product, from which trading partner".

Use GTA for the measures themselves, and reach for the dataset above that matches the question GTA was not built to answer.