The emission benefits of European integration

Brexit exemplifies the complexity in reversing long-term processes of economic integration, especially since European Union (EU) membership goes far beyond trade and investment alone. The tide is turning on multilateralism elsewhere as trade agreements become increasingly bilateral. In the process of trade fragmentation, concerns over the greenhouse gas (GHG) emissions are unlikely to make it to the negotiations table. The EU Parliament for example, has focused on evaluating the economic consequences of Brexit for the rest of the EU (Emerson et al 2017).

In the November 2018 political declaration³ between the EU and the UK, the energy and climate related intentions focus mostly on nuclear. Cooperation in the electricity and gas markets and the EU emission trading scheme are also mentioned. Uncoupling the electricity markets would raise generation costs by 1.5% in 2030, with 60% of those costs incurred in the UK in the form of lost infrastructures and overcapacity (Geske et al 2018). Beyond energy itself, the impacts on carbon emissions from a potentially new trade regime between the UK, the EU and the Rest of the World (RoW) has received little to no attention.

On the climate front, the EU has been pursuing a burden sharing strategy, joining forces to reduce emissions. A hard Brexit would also change national targets in the EU. Such impacts have already been quantified with most of the burden also falling onto the UK (Babonneau et al 2018). While this potentially reduces the efforts by remaining EU Member States, the UK will unlikely be the only one worse off as a result of leaving. Although economic integration through trade has brought benefits to countries across the world, the environmental benefits remain unclear. China, for example, accessed the World Trade Organisation (WTO) in 2001, and since then its carbon terms of trade have worsened. In other words, the net emissions embodied in trade between China and the rest of the world have grown rapidly (Liu et al 2016). Other multilateral trade agreements, the North American Free Trade Agreement (NAFTA) or the Southern Common Market (MERCOSUR) and the lowering of trade barriers or tariffs, were shown to increase carbon emissions embodied in trade (Islam et al 2019).

Consumption based carbon accounting paves the way for more accurate assessment of emissions embodied in trade (e.g. Tukker et al 2018). With energy and climate policies focusing on territorial emissions, those embodied in trade result from economic structures. In the UK, offshoring energy intensive activities has been shown to contribute significantly to the decoupling of energy consumption and economic output (Hardt et al 2018). In the EU, part of the emissions embodied in trade are offset among Member States themselves, justifying a common climate policy (Moreau et al 2019). Hence, the EU might be remarkable as lower trade barriers or tariffs potentially reduce the overall emissions of its members.

Over the last decades the UK accumulated a large carbon debt as net imports of embodied emissions remain among the highest in the world (Matthews 2015, Baker 2018). Emission transfers to the UK reported in the Global Carbon Project dataset (Le Quéré et al 2018) have increased at an average rate of 4.58 Mt CO₂eq each year between 1990 and 2014. In 2004 alone the UK–China trade resulted in an additional 117 Mt CO₂ to global CO₂ emissions (or 0.4% of global emissions) compared with a scenario in which the same type and volume of goods were produced in the UK (Li and Hewitt 2008). Outsourcing of emissions has reached unprecedented levels up until the 2008 economic crisis (figure 1). The UK has been successful in shifting embodied emissions in imports onto the RoW (figure 1(A)) as imports from the rest of the EU and exports have stayed relatively constant (figure 1(B)) at about 0.07 and 0.2 GtCO₂eq, respectively. A change in trade patterns in the aftermath of Brexit will likely affect these proportions in the long-term, and by extent, the UK's consumption based emissions. In the absence of a preferential access to the EIM, part of the UK's current imports from the rest of the EU could end up originating in the RoW. We define the rest of the EU as EU27+ which comprises all current EU28 Member States minus the UK plus Norway and Switzerland, which have preferential access to the EIM as members of the European Free Trade Association. Concurrently, part of the exports from the UK to the EU27+ can either be supplied by the RoW or by the remaining members themselves. How this reconfiguration adds up to a new emission trade balance between the UK and the EU27+ is yet unexplored.

Zoom In Zoom Out Reset image size

To provide a first order estimate of the embodied carbon implications of the UK leaving the EU, a post-Brexit scenario was constructed as follows. The average monetary trade flows of the UK, over the period between 1990 and 1995, were adjusted for inflation and taken as representative of a post-Brexit trade pattern. We call this exercise the Brexit scenario. On the UK side, the loss of trade flows with the EU27+ was reallocated to the RoW. For the EU27+, offsetting the lost imports within its borders, by increasing production, is more likely than for the UK (Latorre et al 2018). Alternatively, the RoW can partially compensate for the imports lost from the UK (e.g. leveraging on the EU–JAPAN Economic Partnership Agreement, 1st February 2019). The corresponding changes in embodied emissions in trade were estimated based on average 2010–2014 emissions intensities.

We test the difference between the additional territorial emissions in an intra-EU27+ trade substitution following Brexit versus the embodied emissions if the global market becomes the source of the trade lost from the UK. Both of these sub scenarios are identical in the implementation of Brexit scenario, that is, imports lost from the EU27+ to the UK are shifted onto the RoW. No temporal dimension is attached to the trade scenarios proposed as it is impossible to understand how post-Brexit trade deals will unfold. Accordingly, our exercise should be understood as setting the bounds of what a pre-EIM trade pattern would entail, given current trade volumes and emission intensities. The post-Brexit reality will likely be within the range of the numbers provided.

The implications of emissions embodied in trade for the scenarios proposed were estimated with input–output analysis (Peters et al 2011). Monetary trade flows came from the Eora multi regional input–output (MRIO) database, particularly suitable for trade analysis with 189 countries (Lenzen et al 2013). Greenhouse gas emissions (CO₂, CH₄, N₂O, HFCs, PFCs and SF₆) were taken from PRIMAP (Gütschow et al 2016), and include fuel combustion emission from transportation. The additional emissions from potentially longer than current distances traveled by imports from the RoW to post-Brexit UK, were not accounted for. At the high level of aggregation of economic activities in Eora, 26 sectors, the uncertainties in emissions are limited. Otherwise the choice of MRIO database has little impact as their content was shown to converge (Moran and Wood 2014).

From 1990 to 2014 the difference in GHG emissions per unit of economic activity between the RoW and the EU27+ embodied in UK's imports steadily declined, figure 2(A). Within this time frame, the efficiency gains of the RoW outpaced that of the EU27+ for the same type of traded products (goods and services), albeit starting from a worse level. In 1995 importing products from the RoW to the UK resulted, on average, in an additional 0.93 kg CO₂eq/$ (95% interval ranging between 0.59 and 1.86 kg CO₂eq/$) compared to imports of the same products from the EU27+. The difference then declined and from 2010 to 2014 stabilized at approximately 0.35 kg CO₂eq/$ (cross sectoral average of imports from the RoW and EU27+ of 0.47 and 0.12 kg CO₂eq/$, respectively, in 2014). This broadly implies that if imports to the UK from the RoW compensate a hypothetical loss of imports from the EU27+ due to Brexit, higher embodied emissions will most likely fall onto the UK. In quantitative terms, the Brexit scenario would release an additional 0.215 GtCO₂eq (an increase in 38%) in UK's embodied emissions compared to the year 2014. Under the Brexit scenario, emissions imported from the EU27+ drop from 24% of total embodied emissions in 2014 to 8% while those imported from the RoW rise from 75% of total embodied emissions in 2014 to 92%. Trade volumes with the EU27+ would not only be smaller, the EU27+ efficiency is also higher than that of the RoW. Further improvements in the efficiency of the RoW in providing comparable products at declining levels of emissions will continue, but are unlikely to overcome current levels in the EU27+ as the marginal cost of efficiency measures increases. In the short- to medium-term, a differential of emissions intensities between the EU27+ and the RoW will remain.

Zoom In Zoom Out Reset image size

For the EU27+, the Brexit scenario would imply a rise in its import emissions of 0.128Gt CO₂eq referenced to 2014 in case imports lost from the UK are to be compensated by the RoW. Unlike the UK, the EU27+ has the scale to offset the import losses within its borders and take advantage of its fluid and relatively emission-efficient economy. If the EU27+ compensates part of the UK imports by increasing production domestically, its territorial emissions would increase by an estimated 0.012Gt GtCO₂eq. This highlights the importance of a Single Member State in keeping the EU market efficient in terms of CO₂ per $ of activity, an iconic result that in terms of emissions embodied in trade the sum of the EU market is, paradoxically and for the better, less than the sum of its individual parts.

Fair decarbonization pathways for countries need to encompass both territorial and traded emissions (Costa et al 2011, Peters et al 2015). Although trade agreements are seldom adopted to mitigate climate change, we demonstrate that in the UK case, a loss of trade with the EU following Brexit could result in significant additional GHG emissions embodied in trade. More importantly, the estimated growth of 38% of the UK's imported emissions, compounds what is already one of the worst carbon trade balance in the world. This can be further exacerbated by the fact that European decarbonization, though too slow to achieve ambitious climate targets, is outpacing that of the RoW. Average growth of wind, solar and other non-hydro renewable energy sources over the last 5 years in the EU is on track to supply more primary energy than coal by 2021 (Jackson et al 2018). Transportation and wholesale activities in the EU27+ are affected by changes in trade patterns more than others which could relocate elsewhere within the EU such as manufactured metal products and financial activities.

The post-Brexit trade pattern simulated here most likely represents an upper bound of UK–EU trade disruption. Changes in trade flows between the UK and the EU27+ in this study were found to be comparable with those simulating a hard Brexit with Computable General Equilibrium (CGE) (Rojas-Romagosa 2016, Latorre et al 2018) and gravity models (Oberhofer and Pfaffermayr 2018). Under our Brexit scenario, exports between the UK and the EU27+ drop by 21.4% and 4.4% relative to total exports in the period 2010–2015. Latorre et al (2018) reports a drop of 16.9% for the UK and 3.5% for the EU27 compared to a 2020 baseline by reverting to WTO tariffs.

In policy terms, the withdrawal agreement of 2018, has neither a clear objective to raise emission standards nor to implement border carbon adjustments to reduce trade related emissions (Mehling et al 2018). A related strategy would consist of systematically shifting trade flows to the lowest emission intensive exporters (de Boer et al 2019). However, as embodied emissions reach a plateau, re industrializing EU Member States at low emission intensities might prove most effective in the medium to long run (Wood et al 2019). For the sake of emission reduction, any divorce from the EU should carefully evaluate its carbon terms.

The authors would like to thank the five anonymous reviewer for helpful comments and suggestions on the initial version of this article. LC acknowledges funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 730459 (Project EU Calculator: Trade-offs and pathways towards sustainable and low-carbon European). VM acknowledges funding from the Swiss National Science Foundation, under project agreement number 173635.

The data that support the findings of this study are openly available on the Eora website at www.worldmrio.com. Data deflators were taken from the World Bank CPI webpage at data.worldbank.org/indicator/FP.CPI.TOTL.