Why has the impact of Brexit on UK–EU goods trade varied so widely across products and trading partners?

Nearly six years after the EU-UK Trade and Cooperation Agreement (TCA) replaced Single Market and Customs Union membership for the UK, the relationship is being reconsidered. The summit postponed in July is now expected before the end of the year, with agreements on Sanitary and Phytosanitary measures (SPS) and on linking the two carbon markets the most likely outcomes. The Prime Minister has also said he wants to set out longer-term options, which I have written about in a recent The Productivity Institute blog. This piece is narrower: It is about what should come after the summit, and where the next round of cooperation would do most good.

What reaches the agenda will depend on what both sides find politically possible, and that will not always follow the evidence. But within whatever room politics allows, five years of product-level data can now tell us a good deal about where cooperation would pay, and the answer is more specific than the debate usually allows.

That the TCA reduced bilateral UK–EU goods trade is no longer seriously disputed, as several independent analyses have confirmed. What is striking is that the shock was not uniform. Trade in some product categories experienced near-zero disruption; others suffered severe losses. Understanding that variation is what turns a general case for regulatory cooperation into a specific one.

Our research

Since the TCA was established, my colleagues Oleksandr Shepotylo, Yujie Shi and I have been tracking its effects, and our latest study is among the first to examine the role of bilateral supply chain structures. Using monthly bilateral trade data from 100 countries between January 2017 and October 2025, we estimate a 53.8% decline in UK export product varieties to the EU and a 31.5% decline in import varieties, while export values fell by an estimated 16.5% and import values by 23.1%.

One finding shapes much of what follows: Within UK–EU trade, the size of the TCA’s effect varied roughly ten times more across products than across EU partner countries. Geography continues to matter: distance and market size still shape how much the UK trades with each member state. Yet, what changed after 2021 depended far more on what was being traded than on where it was going.

Production networks

Where UK and EU firms were tightly integrated through shared production networks, trade has held up: in pharmaceuticals, chemicals and automotive components. Firms in these sectors carry sunk costs in bilateral relationships: certified production lines, specialised logistics, workforces calibrated to European standards. Absorbing the new regulatory burden was more rational than severing the relationship.

The gap between the variety and value estimates supports this explanation. Although export varieties fell by more than half, export values fell by roughly a sixth, consistent with embedded relationships maintaining their volume while marginal product lines exited.

Where production networks were not as embedded, in sectors characterised by high consumer-goods orientation, perishability and low product differentiation, product lines exited and have largely not been replaced.

This resilience should not be read as stability. Each year the commercial logic of maintaining cross-border production under two regulatory regimes weakens, as firms qualify alternative suppliers or restructure their value chains. The relationships that survived the initial shock are thinning, and the window in which they can be preserved is narrowing.

Policy implications

The sanitary and phytosanitary agreement (now under negotiation) covering the trade and movement of plants, animals and their products, is expected to remove routine border checks and export health certificates on both sides once in force. Implementation is likely to take a year or more, and the final legal scope, exemptions and transitional arrangements are still to be settled. For agri-food it will matter a great deal. Our data shows how directly these frictions bite: estimated import variety losses deepened from 12% under initial light-touch border controls to 30% under full enforcement, tracking the phasing-in of physical checks step by step.

But the dominant export friction falls outside of the agreement’s scope. Technical barriers to trade, operating primarily through conformity assessment and dual certification, impose per-product-line costs that bear hardest precisely where UK and EU production networks are most deeply intertwined: pharmaceuticals, chemicals, medical devices and beverages. The sectors that survived are the sectors still paying.

There is a revealing piece of evidence here. The UK's decision to shelve mandatory UKCA marking and continue recognising CE marking indefinitely means that, for most manufactured products, the UK already operates inside the EU's conformity assessment orbit. Divergence was attempted and then abandoned because it proved too costly. What remains is the cost of duplication without the benefit of difference.

Next steps

The EU already has mutual recognition agreements on conformity assessment with several partners, including the US, Canada, Japan and Switzerland. The EU turned down such an agreement with the UK during the TCA negotiations, and a Swiss-style relationship as a whole has never been popular in Brussels. A sectoral agreement, confined to the sectors where both sides depend on each other, is a narrower path. There is now firm evidence. Our work on mutual recognition of conformity assessment, with Oleksandr Shepotylo of Aston University and Lin Zhang of TPI, estimates that recognising each other’s testing and certification increases trade by around 10% on average, and by close to 28% in the sectors where the effect is significant. Industry has been asking for this. Ahead of the 2025 summit, a coalition including Make UK, the Chartered Institute of Export and International Trade, LightingEurope and the Confederation of Swedish Enterprise wrote jointly in support of an EU–UK mutual recognition agreement on conformity assessment. The constituency for this exists on both sides of the Channel, and the evidence now gives a clearer sense of where it would pay most.

The conformity assessment debate tends to overlook a separate barrier. Rules of origin, and in particular how far inputs from elsewhere can count towards origin thresholds, raise the cost of entry for firms not already established in cross-Channel trade. That may help explain a striking absence in our data: almost no new bilateral trade relationships have formed in five years. Origin rules are written into the TCA, so any change has to be negotiated. The UK cannot simply do this on its own, and trying it on the import side alone would raise questions at the WTO and would not be welcomed by exporters. Still, it is a smaller ask than a new regulatory framework, and it could be taken forward alongside the SPS talks.

A practical sequence for the next phase could therefore start with origin rules, negotiated alongside SPS, and use the momentum from SPS to open the conversation on technical barriers to trade. Negotiating effort could then concentrate on the mutually dependent sectors where both sides have the strongest economic incentive to cooperate. Sector-specific conformity assessment agreements take time to negotiate, but they carry the largest potential return.

Need for action

Forecasts made around the time of the referendum put the long-run cost of Brexit at around 4% of GDP per head. Bloom and co-authors now estimate it at 6–8% by 2025. On trade, five years of data show that the damage is neither uniform nor temporary: it is structural, and it is concentrated in identifiable sectors.

That is not an easy conclusion, but it is a more useful one than the aggregate figures allow, because a loss that can be located is a loss that negotiation can be aimed at, especially if the loss is mutual. The rescheduled summit will not lack for agenda items. The question is not whether regulatory cooperation is worth having; it plainly is, but whether negotiation effort will be directed where the evidence says the return is greatest, or where agreement is simply easiest to reach.