Geopolitical tensions have reshaped international trade and investment flows in recent years. This column uses data on the worldwide affiliates of European multinationals over 2010-2020 to study how these firms are restructuring their foreign networks. Foreign network restructuring has transitioned from expansion to contraction over this period, with an increasing share of episodes involving nearshoring, friendshoring, or both. Crucially, network contractions are not associated with higher home-country activity, contrary to expectations. By contrast, network expansions are shown to boost parent activity and domestic networks.
Since the late 2000s, the nature of globalisation has changed under the influence of multiple shocks and crises. In recent years, it has been geopolitical tensions that have reshaped international trade and investment flows. This has been investigated in a growing number of contributions to the literature: Gopinath et al. (2025) identify changes in trade and investment linkages along geopolitical lines that are consistent with a division of the world economy into blocs. For trade, Alfaro and Chor (2023) show that beneath relatively stable aggregate trade patterns, US sourcing has been progressively reallocated away from China towards other partners, and for foreign investment, Aiyar et al. (2024) find that geopolitical alignment has become an increasingly important determinant of bilateral foreign direct investment (FDI) flows. Against this background, this column documents how multinational firms are restructuring their foreign networks with a focus on network contraction and what this means for economic activity in their home countries.
Much of the evidence on the changing geography of globalisation is based on bilateral trade and investment flows. However, this aggregate perspective overlooks how multinational firms organise their activities through networks of foreign and domestic affiliates. Looking inside the foreign affiliate networks of individual multinational firms allows us to gain additional insights.
Recent research has indeed started to look inside multinational firms. Doan et al. (2026), for example, show that Japanese multinationals respond to geopolitical risks associated with exposure to China by diversifying production towards the Association of Southeast Asian Nations (ASEAN) economies. Looking more broadly at the worldwide affiliate networks of European multinationals, our recent research (Merlevede and Michel 2026) examines how these networks restructure, whether this restructuring changes their geopolitical and geographical exposure, and what happens to activity in their home country.
Our analysis is based on a large firm-level dataset that tracks the worldwide affiliates of European multinational networks. For the years 2010–2020, the data cover close to 100,000 unique networks. Because restructuring often unfolds over several years, we group consecutive changes in the number of foreign affiliates into restructuring episodes. We identify more than 36,000 such episodes ending between 2010 and 2019 and classify them as expansions, contractions, or reshufflings according to the overall change in the number of foreign affiliates.
The nature of foreign network restructuring changed markedly during the 2010s. Figure 1 shows that during the first part of the decade, most restructuring episodes were expansions, i.e. the number of foreign affiliates in the network increased. Between 2010 and 2016, expansions accounted for 68% of restructuring episodes on average, compared with 25% for contractions. From 2017 onwards, however, the number of expansion episodes fell substantially while contractions increased. By 2019, contractions accounted for 66% of restructuring episodes and expansions for only 26%.
Figure 1 Foreign network restructuring has shifted from expansion towards contraction
These results confirm the change in globalisation identified with aggregate data on trade and foreign direct investment (FDI): by the end of the 2010s, contraction had become the dominant form of foreign network restructuring.
The shift from expansion towards contraction tells us how multinational network restructuring has changed, but not whether the geographical and geopolitical composition of these networks has been adjusted. We therefore examine how the networks’ geographical distance and geopolitical exposure have changed during restructuring episodes. A decline in the former is consistent with nearshoring, while a decline in the latter is consistent with friendshoring.
The results indicate that in the first half of the decade, less than 30% of restructuring episodes led to a reduction in geographical distance, in geopolitical exposure, or in both. From 2016 onwards, this share increased substantially, reaching almost 50% in 2019. The increase is not simply driven by Brexit: it remains when UK multinationals are excluded. Moreover, expansion episodes also increasingly involve friendshoring, indicating that the shift towards geopolitically more aligned locations is not restricted to firms dropping foreign affiliates. These patterns complement more aggregate evidence of an increasing geopolitical reorientation of international economic activity: Bosone et al. (2024) document a growing role for geopolitical alignment in international trade, while Grover and Vézina (2025) find that foreign investment has become increasingly sensitive to geopolitical distance and provide evidence of friendshoring.
A widespread concern voiced during the decades of fast-growing globalisation and expanding foreign affiliate networks was that firms investing abroad might substitute foreign employment and production for activity at home. This was, however, not supported by empirical evidence, which instead revealed a complementarity of foreign investment and domestic activity. Desai et al. (2009), for example, found that the foreign expansion of US manufacturing multinationals over 1982–2004 was associated with increased rather than reduced activity at home.
By the same token, the contraction of foreign affiliate networks is now associated with hopes for an increase in the activity of multinationals in their home countries. The growing importance of foreign network contraction during the 2010s allows us to investigate whether network contraction abroad has indeed gone hand in hand with increased activity at home. For this purpose, we look at patterns in value added, employment, and other measures of activity for multinational parents and their domestic affiliates before and after foreign contraction episodes (intensive margin). In addition, we examine whether multinationals that drop foreign affiliates expand at home by adding domestic affiliates (extensive margin).
Figure 2 reports the results with a distinction between networks that continue to be multinational after the contraction episode and networks that dissolve altogether, i.e. cease to be multinational. For parents, value added and employment tend to decrease in the wake of a foreign contraction episode. This decline is substantially stronger for dissolving networks and, with few exceptions, statistically significant only for this category. For existing domestic affiliates, value added and employment also decline, both before and after the foreign contraction episode. Regarding the number of domestic affiliates, we find no significant change before, during, or after a foreign contraction episode. According to these results and contrary to popular hopes, multinationals reducing their number of foreign affiliates do not offset this contraction at home, neither by expanding the activity of the parent and its existing domestic affiliates nor by adding new domestic affiliates to the network. Foreign network contraction is rather associated with a reduction in home-country activity, in particular for domestic affiliates that belong to contracting networks.1
Figure 2 Foreign contraction is not associated with domestic expansion
In addition, these patterns do not differ systematically when restructuring involves friendshoring or nearshoring: reducing geopolitical exposure or geographical distance is not associated with stronger home-country activity.
In the underlying paper (Merlevede and Michel 2026), we find a different pattern for multinationals expanding their foreign networks: parent value added and employment increase around foreign expansion episodes, and the number of domestic affiliates also rises before and during expansion. These findings are consistent with earlier evidence that foreign expansion comes with a rise in home-country activity (Desai et al. 2009).
Our work reveals several distinct policy-relevant trends of European multinational network restructuring during the 2010s. First, while most restructuring episodes were expansions in the first half of the decade, contractions became largely dominant in the second half. Moreover, an increasing share of restructuring episodes involved nearshoring, friendshoring, or both. Notably, friendshoring became more common not only in contracting networks but also in expanding networks. Finally, there is no evidence that foreign contraction is associated with increased home-country value added or employment, neither in general nor in cases where restructuring involves friendshoring or nearshoring. Foreign expansion, by contrast, goes hand in hand with growing parent activity and domestic networks.
In sum, these results indicate that contraction and friendshoring have become the main trends of foreign network restructuring of European multinationals. Crucially, they have, however, not been accompanied by an increase in home-country activity. This distinction carries significant policy implications. Rising concerns about economic security and geopolitical exposure have motivated governments to encourage a reorientation towards more resilient and geopolitically secure production networks. Our evidence suggests that multinational firms began adjusting their foreign networks along these dimensions already in the late 2010s. But expectations about the domestic benefits of this reorientation should be tempered: neither reducing foreign activity nor shifting it towards closer and more geopolitically aligned locations has so far been associated with production and employment increases at home.
Source : VOXeu
US net migration turned negative in 2025 following a sharp tightening of immigration policies. This…
The year 2025 was unprecedented in the frequency, complexity, and volatility of US trade policy…
Central banks can move policy rates, but transmission depends on households acting on them. This…
Rising bond yields are likely down to stronger investment demand, more supply of long-duration debt…
Falling agricultural prices have been found to generally increase armed civil conflict in producing regions,…
The White House Faith Office is expanding federal support for religious organisations. Can government policy…