This thesis examines how trade and industrial policies are used to influence the location of production and reallocate risk along global value chains (GVCs). I develop a novel framework of ‘precautionary’ (risk-based) policies over GVCs by adapting two models of public-private 'wedges' from the economics literature. The first is an insurance (or resilience) model – where GVCs are deemed fragile to disruptions and insurance markets are incomplete. The second is a bargaining (or power) model – whereby governments manipulate GVCs to shift the balance of bargaining power with rival states. This paper (Chapter 4) analyses the implications of information constraints under both models. Governments devising GVC policy – to build resilience, or alleviate trade dependencies – are constrained by information gaps and uncertainty. For supply chain resilience, information about network structure is crucial for evaluating risk, but typically incomplete, because knowledge about supplier relationships is piecemeal distributed among firms and across jurisdictions. Existing models of optimal GVC policy assume complete visibility over cross-border networks. Economic resilience is more feasible through broad macroeconomic rules and post hoc responses. For governments intervening in supply chains for strategic (bargaining) purposes, informational constraints exist but are less binding. Governments have better knowledge than firms as to the strategic (e.g. military) spillovers of products – although these judgments too can be speculative. Using an extensive database of trade and industrial policy, I document new findings around subsidy use since 2009 – that countries target subsidies at sectors that are moderately upstream, with high market concentration, and where geopolitical ‘friends’ are already subsidising – and relate these findings to imperfect information under both the bargaining and insurance models of GVC policy.