Recent supply shortages have intensified concerns about supply chain bottlenecks, prompting policymakers to adopt industrial policies to promote reshoring and friendshoring. Nowhere are these concerns more evident than in rare earth element (REE) markets, where supply chains are highly concentrated across stages, particularly refining. This paper quantifies the costs of reducing supply chain concentration using a dynamic global trade model calibrated with detailed market, industry and geological data. In the model, upstream (mining) and downstream (refining) production is slow-moving and determined by investment in capacity (CAPEX). We evaluate how different policies can raise U.S. downstream REE self-sufficiency from 10 to 25% by 2035, comparing CAPEX subsidies and price floors under unilateral and coordination action among importers. The fiscal cost of downstream unilateral CAPEX subsidies reaches 141% of the annual U.S. REE market (about $1.2 billion) by 2035, and falls to 96% under coordinated action. Downstream price floor policies are equally effective, but more fiscally inefficient. Despite their size relative to the market, absolute fiscal costs are contained, suggesting that de-risking REE supply chains is fiscally feasible.