Eduardo Gutiérrez | Banco de España
Aitor Lacuesta | Banco de España
César Martín-Machuca | Banco de España


Policy uncertainty , Brexit , trade diversion

JEL Codes:

F02 , F13 , F14 , F15 , F61 , F68

During the long process of negotiation after the 2016 Brexit referendum, there was a high uncertainty about the final shape of bilateral trade relations between the European Union (EU) and the United Kingdom (UK). This uncertainty provoked a significant trade diversion in the case of Spanish exporters highly exposed to the UK (above 10%), while it was more limited in the case of imports. Regarding the destination market, trade diversion appears to be more intense towards the EU, where exporters and importers maintain stable trade relationships.


The unexpected vote of the United Kingdom (UK) electorate to leave the European Union (EU) initiated a negotiation period in which uncertainty about trade relations between these regions was very high. Although the two regions finally reached an agreement on December 2020 whereby no bilateral tariffs have been established, the uncertainty surrounding the long process of negotiations was very high. In response, Spanish firms could have partially replaced the British market with domestic flows as well as alternative partners. In our recent paper (Gutiérrez et al., 2021), we explore the effect of uncertainty on trade and the capacity of firms to substitute markets using Brexit as a quasi-natural experiment.

In order to so, first, we explore how firm participation in the British market changed under the renegotiation of an existing agreement. We implement a difference in difference strategy to estimate the impact of uncertainty on the intensive margin of bilateral trade with the UK. Uncertainty is captured through future potential losses, which are proxied with potential tariffs in absence of a trade deal and the firm level dependence on the UK. In a second step, we analyze trade diversion patterns exploiting the interaction of uncertainty indicators as instrument for changes in trade with the British market. This market substitution might reveal firms’ contingency plans in response to potential significant losses provoked by Brexit.

Crucially, our results point to an almost full trade diversion in the case of those firms more exposed to that particular market (above 10%), while the response is heterogeneous for Spanish firms with a low share of British bilateral flows over total trade. Thus, this note focuses on the group of firms highly exposed to the UK, whose exports might have been severely hit since the Referendum.

Data and framework

We employ three different data sources to assess trade diversion patterns. First, we use information of monthly declared exports and imports with the UK, EU27 and Rest of the World per operator exposed to the UK between 2015 and 2018 which is provided by the customs agency. Second, an annual firm level database including their sector of activity, turnover, and number of employees coming from the Central Business Register (National Statistics Institute – INE). Lastly, in order to incorporate potential trade barriers, we construct sectoral tariffs by exploiting Most Favored Nation (MFN) tariffs at the HS2 level from the World Trade Organization (WTO).

We end up with a database including about 35,656 exporters and 40,394 importers, which account for around 90% of exports and 85% of imports with the UK containing sectoral MFN tariffs, and firm level sector of activity, employment, turnover, and sales and acquisitions with the UK/EU/RoW. Information for 75 sectors is provided at the NACE-2, 3 and 4-digit level with more details on particular sectors that have large trading flows.

With respect to the framework of the analysis, the effect of uncertainty on trade with the UK is assessed. Firm level uncertainty is measured as the interaction between its relative trade exposure to the UK in the past and potential sectoral tariffs after Brexit which may affect the firm. We expect a negative response of trade flows with the UK to potential tariffs, especially in those firms where sales or purchases were more concentrated in the affected market.

Then, we quantify trade diversion patterns using the predictions of the first stage above. On the one hand, those firms which would face higher tariffs in case of hard Brexit, due to their main sectoral activity, might have reduced their trade with the UK more intensely and replaced, at least partially, the British market with other alternative destinations. Additionally, a higher firm’s relative exposure to the UK will increase the risks associated with Brexit, and thus could lead in turn to a higher trade diversion.

Trade destruction and diversion patterns

Given the higher potential trade barriers between the EU and the UK, their bilateral trade could have declined as a consequence of higher uncertainty. Indeed, our estimations suggest that the growth rate of sales to the UK for the average exporter decreases by 2.6 pp in response to a 1% tariff, and in the case of importers this reduction reaches 3.6 pp. These results allow obtaining estimates of Brexit uncertainty impact which will be used as instrument in order to quantify trade diversion patterns. In fact, there is a positive correlation between trade reductions among markets, as expected given that the ex-ante productivity level of firms determines simultaneously the level of sales in each market (Melitz 2003). However, when we isolate an exogenous variation of a particular market, we find a negative and statistically significant effect of trade growth in the UK on trade growth in the rest of the world.

In terms of the quantitative relevance of our results, following the computations in Almunia et al. (2021), for a firm with an initial export share to the UK of 35% (firm level average share for those exporting more than 10% to the UK), a drop of 100 euros in their British sales would lead to an increase in exports between 92 and 151 euros elsewhere (see Figure 1). This means a close to full diversion of exports. In terms of imports, the diversion appears to be slightly more limited. In particular, for every 100 euros lost in British purchases, the average firm in the sample with a British import share of 44% would be able to divert between 64 and 100 euros. With respect to the markets replacing the UK, the area that has benefited the most is the EU as expected. Indeed, EU is Spain´s main market (accounting for around 60% of Spanish goods exports in 2019), a significant percentage of firms has stable trade relationships with the EU and it has a similar level of development and demand preferences as the UK.

To sum up, the uncertainty generated by trade potential expected losses from Brexit could have affected bilateral trade flows between the UK and Spain. Indeed, during the long negotiation period there was a high uncertainty about the final shape of the trade relationship between the UK and the EU, with potential significant losses in the case of hard Brexit, which would have implied bilateral tariffs. This situation could have dramatically reduced trade links between Spain and the United Kingdom, enforcing firms to substitute their sales/acquisitions in the United Kingdom with the domestic market or other countries. Our paper shows that those firms with the highest exposure to the UK, and therefore the highest potential losses, have reduced their trade with the UK market and diverted trade to other markets. This diversion has been more pronounced for exports than for imports.




Almunia, M., Antràs, P., Lopez-Rodriguez, D., & Morales, E. (2021). Venting out: Exports during a domestic slump. American Economic Review, 111(11), 3611-62.

Gutiérrez Chacón, E., Lacuesta, A., & Martín, C. (2021). Brexit: Trade diversion due to trade policy uncertainty. Banco de España Working Paper No. 2140.

Melitz, M. J. (2003). The impact of trade on intra‐industry reallocations and aggregate industry productivity. econometrica, 71(6), 1695-1725.

  • 1.

    The opinions in this Policy Brief are solely the authors’ and do not necessarily represent those of Banco de España or the Eurosystem. 

About the authors

Eduardo Gutiérrez

Eduardo Gutiérrez is a Research Economist at Banco de España (DG Economics, Statistics and Research). He holds a Master’s degree in Economics from the Nova School of Business and Economics, and is currently a Ph.D. candidate in Economics at the University Complutense of Madrid. His research interests include international economics and economic geography.

Aitor Lacuesta

Aitor Lacuesta is head of the Structural Analysis Division of the Bank of Spain (DG Economics, Statistics and Research). This division develops tax and employment micro-simulation models to analyse the impact of different policies (such as VAT, personal income tax, pensions, MLW, Minimum Interprofessional Salary, Social Security contributions and unemployment benefits). It is also responsible for conducting research related with the impact of shocks and regulations on the labour market and companies in Spain to be published in leading national and international scientific journals, such as SERIEs, IZA Journal of Migration and the Journal of Human Resources He holds a PhD in Economics from the University of Chicago. He has taught at Deusto Business School and the Instituto de Empresa Business School. In the course of his career, he has also worked at the World Trade Organisation and in the cabinet of the Secretary of State for the Economy José Manuel Campa in 2010-2011.

César Martín-Machuca

César Martín-Machuca is a Research Economist at Banco de España (DG Economics, Statistics and Research). He holds a Master in Economics and Finance from CEMFI. His research interests are mostly related to international trade and policy.

More on these topics