World Trade Organization

What You're Getting

WTO Stats covers international merchandise trade, commercial services trade, tariff data, and trade policy indicators for 170+ WTO members. Key datasets: merchandise trade values and volumes, trade in commercial services by sector, tariff profiles (MFN and preferential rates), and notifications. Data is annual; some quarterly. Organized by reporter country, partner country, product group (broad — not HS-level), and indicator.

Ingestion Strategy

WTO Stats API is the primary programmatic access point. No authentication required. The API returns JSON with data organized by indicator, reporter, and time period. For bulk downloads, the WTO Stats website offers Excel/CSV downloads per dataset. The API is best for targeted queries; for full dataset loads, use the bulk download.

import requests

# Merchandise exports for top economies
url = "https://api.wto.org/timeseries/v1/data"
params = {
    "i": "WT_MTE_AX",  # Merchandise trade exports
    "r": "USA,CHN,DEU,JPN",
    "p": "000",  # All partners (world)
    "pc": "TO",  # All products
    "ps": "2015,2016,2017,2018,2019,2020,2021,2022",
    "fmt": "json",
    "mode": "full",
    "max": 10000
}
headers = {"Ocp-Apim-Subscription-Key": "your_key"}  # optional for higher limits
r = requests.get(url, params=params, headers=headers)

Rate Limits & API Details

• Unauthenticated: 50 requests/hour; registered API key: 500 requests/hour
• API key registration: api.wto.org (free)
• Bulk download: stats.wto.org — Excel/CSV per dataset, no automation support
• API docs: docs.wto.org

Schema Stability

Indicator codes (e.g. WT_MTE_AX) are stable. Product classification is broad — WTO uses its own commodity groupings, not full HS. Country codes use ISO 3 + WTO-specific codes. Data revisions occur annually; previous years' data is restated.

Data Quality Gotchas

• Broad product groups only: WTO data is not HS-6 level — use UN Comtrade for detailed product-level trade
• Reporter asymmetries: import and export values between the same pair of countries often don't mirror — document the asymmetry in your model
• Services trade: methodologically different from goods trade; don't aggregate the two
• Missing years: low-income countries often have 2–3 year lags in data

Tool Compatibility

• Python: direct API calls with requests; no dedicated package
• R: direct API calls
• Bulk Excel/CSV works for dbt seeds or warehouse loads
• Not supported by major Airbyte connectors

Benchmarking Approach

Global merchandise exports total should be ~$24T for 2022. WTO members' reported total should match within 10–15% of mirrored imports (asymmetry is normal). Alert if latest year data is more than 24 months old for a major economy. Track count of reporters with complete data for the most recent year — should be 140+.