Services can be supplied through cross-border trade, consumption abroad, commercial presence, and the presence of a natural person. The choice of delivery mode and the extent to which service traders can switch between modes (‘substitution elasticity’) can significantly affect how firms respond to trade barriers. This column analyses UK services trade following the Brexit referendum. UK firms circumvented expected future barriers by switching to modes less affected by Brexit, such as commercial presence sales. However, this came with lower UK employment growth as services exporters prioritised employment growth at their new foreign affiliates rather than at home.
The last decade has provided an opportunity to study the economic effects of trade disruptions after more than half a century of post-WWII trade liberalisation. Events like the US–China trade war and Brexit have reshaped global trade patterns, offering case studies on the consequences of new trade barriers (Amiti et al. 2019, Fajgelbaum et al. 2024, Fajgelbaum and Khandelwal 2022, Freeman et al. 2022). However, existing research primarily focuses on goods trade, often overlooking services trade. This is despite the fact that services trade constitutes a third of global trade and services represent a substantial share of GDP and employment in OECD countries (Baldwin 2022). This omission is significant, as the trade of goods and services is closely linked, and excluding services can underestimate the broader impact of trade restrictions on economic welfare (Cernat 2024).
Complexity of services trade and WTO modes
Unlike goods, services trade is more complex, as services can be supplied through four different modes as defined by the WTO: cross-border trade (Mode 1), consumption abroad (Mode 2), commercial presence (Mode 3), and the presence of a natural person (Mode 4). Firms face different challenges and opportunities depending on which mode they use. For instance, intellectual property services are primarily delivered cross-border, while health and education services often use all four modes (Figure 1).
Figure 1 Modes of services supply, 2017
Notes: Share of the mode of services world export supply, by type of services in 2017. The shares are computed using the total services trade flows and modes of supply for the countries included in the TISMOS dataset.
Source: TISMOS (WTO).
The choice of delivery mode can significantly affect how firms respond to trade barriers. Take a software company: it can either electronically transmit software to foreign clients (Mode 1) or establish a local office abroad (Mode 3). If trade barriers make cross-border trade more difficult, software companies may simply switch to delivering software via foreign affiliates, leaving total services trade broadly unaffected. However, if switching modes is more difficult, trade restrictions could have more severe effects on overall services trade. On the other hand, if switching to foreign affiliate sales is difficult, domestic employment will be less affected because increasing foreign affiliate sales is likely associated with cutting back on domestic jobs. Thus, the extent to which service traders are able to switch between supply modes shapes how firms and countries experience the impact of trade restrictions.
UK services trade before and after Brexit
In a recent paper (Breinlich and Magli 2024), we analyse services trade in the UK, the world’s second-largest services trader, to understand how firms adapted to trade barriers caused by the UK’s exit from the EU (Brexit). We construct a dataset covering UK services trade from 2009 to 2019, enabling us to examine firms’ trade flows and foreign direct investment (FDI) in services by type, sector, and partner country. We find striking variations in how UK firms combine different supply modes. For instance, 75% of health services exports are delivered through commercial presence (Mode 3). In contrast, almost all engineering services are traded cross-border (Mode 1) (Figure 2).
Figure 2 Services exports by supply mode, 2017
Notes: Mode of export supply by type of service. Mode 1 is defined as the exports of services surveyed in the ITIS, while Mode 3 is defined as the profits of the foreign affiliates of UK firms surveyed in the AFDI dataset. Firm-level information is aggregated at the services type and year level, and shares are computed for each year. E.g. share of Mode 1 of advertising is computed as the total exports in advertising surveyed in the ITIS dataset, divided by the total exports in advertising surveyed in the ITIS dataset plus the total profits of foreign affiliates selling advertising owned by UK firms surveyed in the AFDI dataset. The figure refers to the year 2017.
Source: Own computation using information from ITIS and AFDI.
We then study how UK firms adapted their service delivery methods in response to the expected trade barriers following the 2016 Brexit referendum. Although no actual changes occurred until 2021, existing research documents that firms began adjusting their operations soon after the vote due to uncertainty and anticipation of future restrictions (Born et al. 2019, Bloom et al. 2019, Graziano et al. 2021).
We show that UK firms increasingly relied on affiliate sales for EU exports after the Brexit vote, indicating that they substituted away from cross-border trade (Mode 1) in response to anticipated barriers. This shift toward affiliate-based trade was particularly pronounced in the ICT, wholesale, and retail sectors.
However, we do not observe the same shift on the import side, even though European exporters probably also anticipated higher trade barriers. Instead, the share of affiliate imports in total imports declined after Brexit, driven by a reduction of sales of foreign affiliates in the UK. While surprising, this finding aligns with previous research by Breinlich et al. 2020, which showed a significant drop in EU-UK FDI flows post-referendum.
Implications for UK services exports and employment
Having established that UK services exporters switched delivery of services to commercial presence (Mode 3) after the Brexit referendum, we then investigate the implications of this switch for UK services exports and employment. To do so, we first measure how easy it is to switch delivery modes for different service types. We indeed find substantial differences across service types in terms of this ease of switching, which we label ‘substitution elasticity’ (Figure 3).
Figure 3 Substitution elasticities by type of service
Notes: Dependent variable: share of exports via commercial presence on total exports. The unit of analysis is firm-service-country of destination through the years. Each point indicates the regression coefficient of the triple interaction term Brexitt × EUc × Services, where Brexit is a dummy variable taking value 1 if the year is post-2016; EU is a dummy variable taking value 1 if the country of destination is part of the EU; and Service is a dummy variable taking value 1 for each of service listed. The lines indicate the 95% confidence interval. Each regression also includes 2-digit sector-year, firm-year, service, and country-fixed effects.
Source: Own computation.
When we investigate how differences in substitution elasticities across services affected UK exports and employment after 2016, we discover two interesting patterns. On the one hand, firms exporting services with higher substitution elasticities experienced smaller drops in overall exports. On the other, these firms also saw lower growth in domestic employment. We interpret these patterns as showing that UK firms successfully circumvented future expected barriers by switching to modes less affected by Brexit, such as Mode 3 sales. However, this successful adaptation came at the expense of lower UK employment growth because services exporters now prioritised employment growth at their new foreign affiliates, rather than at home.
Broader implications for trade policy
We draw several broader lessons from our findings. First, the ability of firms to switch how they deliver services – whether by exporting directly or establishing offices in foreign countries – makes it harder to predict the full impact of trade restrictions. While switching modes can help firms continue exporting despite trade barriers, it often results in fewer jobs in the home country. This means that policymakers should approach services trade liberalisation initiatives differently depending on their objectives.
If the goal is to increase services exports, they should focus on reducing barriers for service types where it’s harder for companies to switch delivery modes. These services will benefit most from easier trade conditions because they cannot easily circumvent barriers by switching modes. On the other hand, if the goal is to protect domestic jobs, it’s essential to prevent other countries from imposing trade barriers on service sectors with higher substitution elasticities. In these sectors, companies can quickly shift operations overseas, often leading to job losses at home.
Conclusion
Understanding how firms adapt to barriers is critical for businesses and policymakers as trade disruptions become more common. Our research provides new evidence on how UK firms adjusted their services trade modes in response to Brexit, with important implications for trade policy and employment. By calling attention to mode substitution’s role in mitigating the effects of trade barriers, our findings highlight the importance of considering the full complexity of services trade when designing trade policies.
Source : VOXeu
