Published on Data Blog

How is trade data collected?

This page in:

WITS banner

This blog is part of a series about WITS, the World Integrated Trade Solution, a collaborative trade data platform developed by the World Bank and other institutions. This is the second installment of the series—for further reading, here are the first, third, and fourth installments.

No one would argue against trade's crucial role when it comes to ending extreme poverty, promoting shared prosperity, and fueling economic growth.

Trade is a fundamental and pivotal condition for countries to create job opportunities, for entrepreneurs to flourish, and to create wealth for their citizens. Behind this lies the importance of measuring trade accurately, correctly and highlighting the importance of reliable and protected data sources when it comes to what, how, and where goods are traded; not only for the sake of having valid statistics, but also because it impacts the quality of macroeconomic policymaking.

What data gets collected on trade? 

Merchandise or goods trade data can be broadly classified into trade—exports and imports— and trade protection—tariffs and non-tariff measures.

Data on exports and imports are collected, curated, and published by the United Nations Statistical Division, and made available through the UN COMTRADE database. Likewise, the UNCTAD Trade Analysis Information System (TRAINS) has information on tariffs and non-tariff measures.

Also, there is the World Trade Organization's (WTO) Integrated Database (IDB) which features imports by commodity, partner country, and Most Favored Nation (MFN) applied, and where available, displaying preferential tariffs at the most detailed commodity level. Another dataset comes from WTO’s Consolidated Tariff Schedule Database (CTS), which highlights WTO’s Bound Tariffs. (If you want to know more on tariffs, read this blog on Types of Tariff.)

The World Bank and UNCTAD’s World Integrated Trade Solution (WITS) brings all these databases under a single platform and provides users with ability to query, derive, aggregate, and run simulations using the underlying data.  

Boubacar Tebely (top) and Yaya Banou (center) unloads sacks of onions in a farmers market in Bamako, Mali on November 3, 2013. Photo © Dominic Chavez/World Bank
Boubacar Tebely (top) and Yaya Banou (center) unloads sacks of onions in a farmers market in Bamako, Mali on November 3, 2013. Photo © Dominic Chavez/World Bank


Categorizing data

Trade involves the movement of goods across national boundaries. These commodities can be of different nature, ranging from animal feed to rocket engine parts, as countries export and import several types of products.

To record trade statistics, the World Customs Organization (WCO) built a hierarchical structure of products called the Harmonized System (HS). This structure is updated every 5 years to accommodate new products and remove obsolete ones. The HS system has around 5,300 cross-country comparable products, and each product is represented by a 6-digit code, also called a sub-heading. Six-digit codes are grouped under a 4-digit heading and 4-digit codes grouped under two-digit code Chapter. Country-wise, each nation has its own 8- or 10-digit code (National Tariff Line (NTL) level) under the 6-digit level.

Here is an example to illustrate this: two different countries export vanilla as a product, but one of them does it in the form of “neither crushed nor ground”, while the other exports it grounded or crushed. To differentiate both products, the system uses separate codes.

How is this categorization visualized, then? As shown below, Chapter 09 is labeled as “Coffee, Tea, Mate and Spices”, under which you have heading 0905 and two sub-headings 090510 and 090520. 


But, what was in place before the Harmonized System was enacted? The United Nations Statistics Division used the Standard Industrial Trade Classification (SITC)—which has data from 1962, covering all data until the current system was placed.

So, how does the user decide which nomenclature to use while querying on WITS? Choosing a nomenclature depends on how back in time the user wants to conduct their analysis: SITC is used for data starting all the way back in 1962, while the HS is chosen for datasets available from 1988. Users are also encouraged to decide on which nomenclature to choose by looking at the data available by country of interest. (Please note, tariff data is not available prior to 1988 in WITS.) 


Are there other ways to search for products?

There are other ways of classifying trade data, such as the Broad Economic Category (BEC), the International Standard Industry Classification (ISIC), the Global Trade Analysis Project (GTAP), and more.

How does the Harmonized System (HS) coexist with these other nomenclatures? Conversions can be done between HS and SITC codes to BEC and GTAP codes, which allow to convert products from the reported nomenclatures of HS and SITC to these derived nomenclatures.

For example, the chart below explains how ISIC’s Revision 3 Product Code 9214 is converted into four distinct products in HS 2012. 


In WITS, users can execute the Advanced Query feature and search for data across these nomenclatures while using standard or custom product grouping. Users can also create new product groups to generate customized groups.

Want to access WITS and use all its distinctive features on trade data? Click here



Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000