Cambridge Finds Friendshoring Accidentally Makes EV Supply Chains More Global
A Cambridge study models Country+1, Friendshoring and Reshoring on the global EV supply network and finds these localizing policies mostly increase interdependence.
Takeaways
- Cambridge team models Country+1, Friendshoring and Reshoring on an 18,833-firm EV supply graph.
- Friendshoring counterintuitively increases global links and network density across allied countries.
- Reshoring is largely infeasible for EVs: 800+ non-substitutable products, mostly mining and smelting.
- Country+1 adds redundancy cheaply, with only 12 to 22 non-substitutable products per cluster.
- European cluster is the most vulnerable overall; Asian cluster most resilient to Friendshoring.
- Code available on request; pseudocode for all three rewiring algorithms is in the appendix.
Trade policies designed to shrink supply chains and reduce foreign dependence may end up doing the opposite. A team from the University of Cambridge and The Alan Turing Institute modeled three widely discussed interventions on the global electric vehicle supply network, and the structural math does not match the political rhetoric.
The paper treats the EV supply chain as a directed graph of firms and products, then simulates what happens when governments try to rewire it under Country+1 (add a supplier outside a risky country), Friendshoring (source only from politically aligned countries), and Reshoring (bring production home). Two of the three policies push the network toward more globalisation, and the third is largely infeasible in the EV sector.
Why topology beats spreadsheets
Prior work on trade policy has mostly looked at costs and cash flows from the perspective of a single focal firm. Effects on network topology, robustness, and the propagation of disruptions through interconnected nodes have not been systematically analyzed, leaving a gap in understanding the broader ramifications of these interventions. That gap matters because supply networks have specific structural properties (heavy-tailed degree distributions, modular communities, negative degree assortativity) that determine how disruptions spread.
In plain terms: a graph with a few giant hubs is robust to random failures but fragile when those hubs are hit. Strong modular clusters can contain shocks locally. Policy changes reshape both properties, and the authors argue you cannot judge a trade intervention without measuring what it does to the graph.