Preference Utilisation and Customs Data: The Missing Pieces of the FTA Puzzle

Even though free trade agreements (FTAs) have proliferated for several decades, researchers still don’t fully understand their effects. In the wake of the recent wave of mega-regional trade deals, many governments are now interested in evaluating the effectiveness of their existing commitments and the benefits they have brought for business. It is, therefore, more important than ever to understand how FTAs are applied and to pinpoint any issues in their design or implementation.


Unfortunately, policymakers have only a limited understanding of FTA utilisation and often rely on incomplete data and estimates. As the political momentum for greater transparency on trade agreements is increasing, there is a pressing need to create a centrally-held, WTO-driven and standardised database of utilisation data for trade agreements. This would allow for a more systematic review of how current trade agreements have been implemented and are applied.


ECIPE Policy Brief by Dr Anna Jerzewska and Nicolas Köhler, Trade Policy Advisor at International Trade Intelligence and PhD Candidate at the University of Cambridge. Published July 2019. Available (HERE).





Even after decades of academic and policy analysis of free trade agreements (FTAs), their static and dynamic effects are still not fully understood. The recent wave of mega-regionals and a surge of protectionist tendencies have revived questions about the consequences of economic integration, both for the countries that enter such agreements as well as or the multilateral trading system.


It may seem obvious that the economic benefits of an FTA cannot be achieved if the agreement is not fully implemented and utilised, i.e. when goods continue to be imported under the most favourite nation (MFN) rates rather than under preferential tariffs. And yet, the utilisation of FTAs remains poorly understood.


Renewed interest in the topic follows the rapid proliferation of these agreements over the last three decades. Since the establishment of the World Trade Organisation (WTO) in 1995, more than 294 trade deals were notified to the WTO, a significant increase over the 41 that existed during the entire General Agreement on Tariffs and Trade (GATT) era. This peaked in 2009 when 35 new agreements were notified, but the process has notably slowed down during the last couple of years. Only 13 new deals were notified in 2017 and a mere six in 2018. At the same time, in the wake of the global financial crisis, the world witnessed an increase in protectionist measures. The Global Trade Alert project found 151 new discriminatory interventions in 2010, which almost doubled to 314 by 2018 (1).


After years of signing new trade agreements, many governments are now interested in evaluating the effectiveness of their existing FTA commitments together with the benefits they have brought for business. Policymakers are trying to assess the real-world impact of FTAs that are already in place and explore ways to maximise their economic benefits. The negotiations of mega-regional agreements and the difficulties in completing them have further catalysed the debate and generated considerable public interest in the topic. In this context, it may be more important than ever to understand how FTAs are utilised and pinpoint any issues in their design or implementation.


Unfortunately, policymakers have only a limited understanding of FTA utilisation and often rely on incomplete data and estimates. To remedy these shortcomings, the topic needs to be investigated further. Systematic analysis of preference utilisation, in particular, would be an important step forward towards understanding the tangible economic benefits that FTAs can deliver. Both governments and businesses would significantly benefit from better utilisation rate data and analysis in order to achieve deeper economic integration and the accompanying welfare benefits.


Measuring FTA utilisation and the need for better data


Researchers have developed various indicators to measure the performance of FTAs (2). Perhaps most importantly, utilisation rates indicate the percentage of trade occurring under FTA preferences versus MFN tariffs for countries that are parties to a trade agreement. In other words, utilisation rates measure how much trade takes place under an FTA. But a key obstacle to measuring utilisation (and therefore the impact of FTAs) is the availability of data. This refers both to the governments’ ability to collect reliable customs utilisation data as well as the quality and usefulness of collected data for measuring the effects of FTAs.


In theory, the utilisation of trade preferences could be relatively straightforward to quantify: it is the ratio of the value of preferential imports over the value of eligible imports under FTA schemes. But such analysis requires knowledge of the volume of trade under preferential tariffs coupled with the amount of goods eligible for preferential treatment.


This indicator is particularly useful in assessing the quality of FTA design as it helps to identify underutilised tariff lines. Together with other metrics, it can inform policymakers about aspects that need to be improved to increase preferential trade flows. This includes restrictive origin provisions, and the amount of administrative burden or asymmetry of information, which can be particularly problematic for SMEs as they lack sufficient trade expertise and resources.


But herein lies the first problem. The analysis of underutilised tariff lines requires the ability to identify goods that fulfil FTA rules of origin but are nevertheless imported under MFN tariffs. This can only be done by the companies themselves, based on an in-depth understanding of their supply chains and processing. As that information is held at firm level and rarely published, any analysis using this ratio is based on estimates and proxies.


This data limitation is unlikely to be addressed in the medium term, if ever. Even in the context of the recent wave of interest in how blockchain technology could allow companies to track parts and components throughout the supply chain, this data will likely still be considered commercially sensitive supply chain information.


The second-best approach is to take the volume of imports at a tariff level (for example at the 6-digit level of the Harmonised System) over a period of time and assess what percentage of these imports were declared for preferential treatment. This would allow measuring the proportion of trade under preference that is nevertheless imported under MFN rates. But the method is far from perfect, as it does not enable distinguishing between a number of potentially valid reasons why companies might import under MFN tariffs. This might be due to restrictive Rules of Origin, supply chains and the inability to fulfil direct shipment requirements, a low margin of preference, a lack of knowledge or other factors. Therefore, while the metric helps to determine utilisation rates and identify potential underutilisation of certain tariff lines, it is not enough to comprehensively inform policymakers about the reasons behind such state of affairs. Nonetheless, identifying the underutilised tariff rates would be a significant step towards understanding the impact of current FTAs on trade flows and determining areas where the implementation of an FTA might require additional efforts.


Policymakers could use this information to further explore pockets of underutilisation by other methods, such as direct industry engagement and surveys. Studies on utilisation rates could, therefore, be viewed as building blocks for a more targeted analysis.


This approach also relies on the availability of sufficient data, in particular, customs data. Import data is collected by customs authorities. Whenever a good is imported into a country, a customs declaration is filled and submitted. Each customs declaration includes comprehensive information about the product, in addition to whether it was imported under an MFN tariff or was subject to any customs duty discounts, such as preferential tariffs under an FTA. Export data is also collected but, given that exports do not have fiscal implications in most jurisdictions, this tends to be less comprehensive and reliable.


Most national customs authorities, therefore, have the data required for FTA utilisation rate analysis – data collected on all imports and over time. But this data is not necessarily shared with other governmental departments, FTA partner countries and the general public. In order to enable systematic, comprehensive and coherent FTA utilisation rate analysis the following steps would need to take place, ideally at the multilateral level to avoid duplicating efforts and inefficiencies.


1. Creation of a centrally held database


Some countries have begun to collect and publish customs import data on the value and volume of goods imported under preference. For example, the European Union, Japan and the United States currently publish import data for both non-reciprocal and reciprocal preferential arrangements. This, however, is not standard practice amongst WTO members.


As such, conducting any analysis requires relying on the customs authorities of FTA partner countries to provide data on preferential imports on a case-by-case basis. This could be avoided by creating a comprehensive, transparent and centrally held database.


The European Commission has recently conducted utilisation rate studies of the EU’s trade agreements (3). However, since there is no centralised database of preference utilisation data, the Commission had to request this information from partner governments – a time-consuming effort, which was also limited in terms of the period it covered. In the end, the Commission was only able to obtain incomplete time-series data from 18 partner countries, which were the destination of only 25% of EU exports. This may, therefore, lead to significantly biased results.


Other attempts have also been made. The Comprehensive Economic and Trade Agreement (CETA) between the EU and Canada includes a clause that mandates the sharing of data on preference utilisation (Art. 30.3). The clause builds on a ‘gentleman’s agreement’ between the EU and South Korean authorities on regular data sharing on preference utilisation under the EU-Korea FTA. Although such bilateral information exchanges enhance the understanding of FTAs by participating authorities, they are significantly more costly and inefficient than a centralised collection of utilisation data that would improve transparency on a wider scale.


2. Development of a comparable time-series dataset and consistent methodology


In the absence of standardised time-series data and common methodologies, the evidence on preference utilisation, perhaps unsurprisingly, encompasses a wide range of results. A centrally held database where customs data is published in the same format and on the same level would, over time, lead to the development of comparable time-series datasets.


Use of time-series data could lead to a significant improvement of insights drawn from FTA utilisation studies. To date, studies on reciprocal trade preferences rarely include time-series data. One of the most comprehensive cross-country analysis on the share of preferential trade relied on a single year of data from 2008, at which point major FTA preferences were still being phased in and many observers held out hope for a successful conclusion of the Doha Round (4).


Once a more comprehensive time-series dataset is created, a consistent methodology could be applied to other utilisation rates studies.


The way forward


As the political momentum for greater transparency on regional trade agreements is growing, the functional need for a central platform has increased. At the July 2017 Hamburg Summit, G20 leaders ‘reaffirm[ed] the importance of transparency for predictable and mutually beneficial trade relations’ and ‘note[d] the importance of bilateral, regional and plurilateral agreements being open, transparent, inclusive and WTO-consistent, and commit[ted] to working to ensure they complement the multilateral trade agreements’.


A centralised approach of collecting and analysing utilisation data would be in line with the spirit of the WTO’s Transparency Mechanism on FTAs, demands from the Parliamentary Conference on the World Trade Organization to increase transparency on trade agreements, and the G20 commitments for open data and greater transparency in the public sector, including in the area of customs.


Recent initiatives for enhanced procedures to enhance transparency and strengthen notification requirements in the WTO Council for Trade in Goods could similarly present a window of opportunity to move this important agenda forward (5). The utilisation of tariff preferences for the Least Developed Countries has also recently been the focus of the WTO’s Committee on Rules of Origin (6).


Discussions at the WTO level have so far not included reciprocal trade agreements and focus on the preferences granted under non-reciprocal Generalized Systems of Preferences for Least Developed Countries. However, since some of the major sponsors of the push for greater transparency at the WTO – the European Union, Japan and the United States – already publish preferential import data, there could be political scope for them to encourage other members to follow in their footsteps.


The creation of a centrally-held, WTO-driven and standardised database of utilisation data for reciprocal trade agreements would finally enable the evaluation of the impact of FTAs on businesses and the multilateral trading system. A more systematic review of how current FTAs have been implemented and are utilised would, over time, allow WTO members to learn about best practices elsewhere. This could impact prospective FTA negotiations and contribute to the WTO’s mission to achieve a more open, predictable and transparent trading system.


Policy recommendations


• Improve the comprehensive collection and availability of preferential import data from national customs authorities of WTO members
• Create a centralised database of utilisation rates for FTAs that are notified to the WTO
• Identify pockets of FTA underutilisation to improve diagnosis of causes and consequences





(1), accessed 15 May 2019.
(2) Hamanaka, Shintaro (2013): On the Use of FTAs: A Review of Research Methodologies, ADB Working Paper Series on Regional Economic Integration, No 113.

(3), accessed 15 May 2019, see also the National Board of Trade Sweden’s analysis based on this data:, accessed 15 May 2019.

(4), accessed 15 May 2019.
(5) WTO document JOB/GC/204/Rev.1
(6), accessed 15 May 2019.

Other Posts You May Like