International trade and balance of payments statistics

OECD Balanced Trade Statistics


Despite multiple efforts to tackle them at national and international level, asymmetries in international trade data are still significant: the exports of country A to country B rarely match the imports of B from A, both for merchandise and for services trade. To enable a better understanding of global trade patterns, the OECD developed transparent methodologies to reconcile these asymmetries. Two analytical datasets are available: the OECD Balanced International Merchandise Trade dataset (BIMTS) and the OECD-WTO Balanced Trade in Services dataset (BaTIS).

Further work to reduce asymmetries in official statistics, including through bilateral and multilateral meetings, is under way in collaboration with national statistical offices and other international organisations. These efforts are crucial to improve the reliability of trade statistics at source and are preferable to any mechanical balancing procedure. The BIMTS and BaTIS datasets will incorporate the results of such efforts in their future updates.


OECD Balanced International Merchandise Trade dataset (BIMTS)

Why are asymmetries in merchandise trade so large?

Asymmetries in merchandise trade data can be large, as shown in Table 1. There are several factors behind the observed discrepancies. First of all, exports and imports are valued differently: exports are valued ‘free on board’ (FOB), while imports include the ‘costs of insurance and freight’ (CIF). However, this makes only marginal contributions to asymmetries, as the  CIF/FOB margins account for an average of around 5% of merchandise trade flows (Miao and Fortanier, 2017). Differences in customs regimes, confidentiality policies, time of recording, or product classifications also play a role. Another important source of discrepancy is the convention that merchandise trade statistics record imports by country of origin, while exports are recorded by country of last known destination. As global production chains become increasingly complex, and goods cross borders several times before reaching the final consumers, this convention means that the recorded exports and imports will not be symmetrical.

Table 1: Selected merchandise trade flows, 2018 (bn USD)

Selected merchandise trade flows, 2018 (bn USD)

Source: OECD BIMTS. Note: the asymmetry is calculated as the absolute value of the difference between reported exports and mirror imports, divided by the sum of the two flows. The adjusted exports and imports, which represent an intermediate output of BIMTS, are available upon request.

How can asymmetries be reconciled?

The OECD has developed a process to reconcile the observed asymmetries in merchandise trade. First, detailed bilateral statistics at HS 6-digit level are collected (and cleaned) and imports are converted to FOB prices to match the valuation of exports. Secondly, data are adjusted for several specific problems known to drive asymmetries, most notably unallocated and confidential trade, re-exports by Hong Kong (China) and product misclassifications. In the third step, the reconciled figures are calculated as a weighted average between the two flows, where ‘symmetry indices’ – which vary by reporter, partner and product – constitute the weights in the balancing. Finally, the data are converted to the 2017 version of the Harmonised System (HS 2017) to provide users with consistent time series across the entire time span. In addition, the figures are also presented according to Classification of Products by Activity (CPA) products to better align with National Accounts statistics, such as in national Supply-Use tables.


Where to find the data?

The BIMTS dataset covers 160+ reporters and partners for the period 2007-2018. 

Access the BIMTS dataset in HS 2017 (at 6, 4 and 2 digits – around 6,900 products and product groups)


Access the BIMTS dataset in CPA (at 2 digits - around 50 products and product groups)




OECD-WTO Balanced Trade in Services dataset (BaTIS)

What about trade in services?

Services trade flows are inherently more complex to record than merchandise trade. The physical nature of goods means they are relatively easy to measure when they cross borders, while the delivery of services is more difficult to observe – even more so when they are delivered in digital form. Hence, only about 60 economies publish (some) bilateral trade in services statistics. The scarcity of detailed information on trade flows by partner (and service category) therefore adds to the asymmetry problem, severely hampering the analytical and policy use of trade in services data. To mitigate these problems, OECD and WTO have developed a global dataset of coherent bilateral trade in services statistics by main service categories.

The latest edition of the OECD-WTO Balanced Trade in Services dataset (BaTIS) covers over 200 reporters and partners and the 12 main EBOPS 2010 service categories for the period 2005-2019. The first edition of BaTIS provides annual data for 1995-2012, covering 191 economies, broken down for the 11 main EBOPS 2002 service categories.


How is BaTIS built?

To develop the dataset, the OECD and WTO have leveraged all official statistics available at national level and supplemented them with estimations and adjustments to provide users with a complete matrix of exports and imports of services. Subsequently, the asymmetries between reported and mirror flows are reconciled by calculating a symmetry-index weighted average between the two, following a similar approach to the one developed for merchandise trade statistics. 


Table 2: Selected trade in services flows, 2018 (bn USD)

Selected trade in services flows, 2018 (bn USD)


Where to find the data?

Access the latest edition of the OECD-WTO BaTIS dataset and the accompanying methodological paper (PDF)

Access the first edition of the OECD-WTO BaTIS dataset and the accompanying methodological paper (PDF)


Further reading


Contact us


Related Documents