The Cost of Decoupling: What US-China Economic Fragmentation Really Costs the World

World trade is splitting into two blocs. The United States and China, the world's two largest economies, are steadily reducing how much they depend on each other — a process economists call ‘decoupling.’ Tariffs, export controls, and restrictions on technology and investment have replaced decades of ever-deepening trade ties.

The bill for this shift is coming into focus. The World Economic Forum estimates a full split between Western and Chinese-aligned trading blocs could remove up to $6.9 trillion from global GDP over time. Even short of a full break, ongoing fragmentation is already shaving an estimated $213–307 billion off world output and adding roughly 0.2–0.3 percentage points to global inflation, as companies pay more to reroute supply chains through ‘friendlier’ countries.

Businesses are responding by diversifying: shifting factories to Vietnam, India, and Mexico (‘friend-shoring’), stockpiling critical inputs, and duplicating supply chains rather than relying on single sources. This insurance against geopolitical risk is expensive, and someone has to pay for it — largely consumers, through higher prices, and shareholders, through lower margins. The question for 2026 is no longer whether decoupling is happening, but how fast, and who absorbs the cost.

Decoupling was once a fringe idea; today it is explicit US and increasingly Chinese industrial policy. Washington has layered Section 301 tariffs from the first trade war era with more targeted export controls on advanced semiconductors, chipmaking equipment, and AI-related technology, framed explicitly as national security measures rather than pure trade policy. Beijing has responded with its own restrictions on rare earth and critical mineral exports — inputs the West depends on for everything from electric vehicle batteries to fighter jets — and has accelerated its own domestic substitution programme for foreign technology.

The economic logic driving this is only partly about efficiency anymore. For much of the post-Cold War era, companies built ‘just-in-time’ supply chains optimised purely for cost: a single factory in a single country supplying the world. The pandemic exposed how fragile that model is, and geopolitical risk — sharpened further by the Iran conflict and Strait of Hormuz disruptions rattling energy and shipping routes in 2026 — has pushed firms toward ‘just-in-case’ redundancy: multiple suppliers, geographically diversified, even if that means paying more.

The World Economic Forum's modelling puts numbers on both the gradual and severe versions of this shift. In a full fragmentation scenario, growth could fall by as much as 6.4 percentage points in the worst-affected economies, with inflation rising by up to 6.1 percentage points — effects concentrated in economies most exposed to either US or Chinese supply chains, with limited ability to diversify. Even the milder, current trajectory is not free: an estimated $213-307 billion has already been shaved off global GDP, alongside a modest but persistent inflationary drag.

Not every economy loses. ‘Friend-shoring’ — the relocation of manufacturing to politically aligned or neutral third countries — has been a clear net positive for Vietnam, India, and Mexico, all of which have captured meaningful shares of electronics assembly, textiles, and light manufacturing that once went to China. India's Production Linked Incentive (PLI) schemes and ‘China+1’ positioning are a direct beneficiary of this reallocation. Mexico's exports to the US have surged partly on genuine nearshoring and partly on Chinese firms using Mexican assembly to bypass tariffs altogether — a form of trans-shipment that complicates any simple ‘decoupling is working’ narrative.

That complication matters. Aggregate US-China bilateral trade has indeed fallen from its peak, but China's total exports have kept rising, increasingly routed through third countries. Some economists argue this means measured decoupling overstates true separation: the two economies remain deeply intertwined, just with more intermediaries and higher transaction costs layered in. Others counter that even this rerouting represents genuine economic cost — the extra shipping, tariff-avoidance structuring, and duplicated infrastructure are exactly the kind of friction the WEF estimates are trying to capture.

For 2026, the IMF's April World Economic Outlook cut China's growth forecast to 4.4%, citing trade fragmentation alongside the drag from Middle East conflict on global energy markets — a reminder that decoupling costs do not stay contained to the two countries directly involved. In an integrated global economy, fragmentation anywhere raises costs everywhere, even for bystander economies with no direct stake in the US-China relationship.

Economists analysing decoupling generally invoke two related but distinct frameworks: trade diversion versus trade creation, and the reversal of what trade economist Richard Baldwin termed the ‘second unbundling’ — the geographic splitting of production stages across borders that defined globalisation from the 1990s onward. Standard trade theory predicts that tariffs and barriers between two large economies redirect commerce toward third countries (trade diversion) rather than eliminating it outright, which is consistent with the Vietnam-Mexico-India reallocation observed since 2018. The efficiency loss comes from abandoning the lowest-cost global supply configuration in favour of a second-best, politically constrained one.

Quantifying this loss requires computable general equilibrium (CGE) modelling, the methodology behind most of the headline figures now circulating. These models simulate an economy (or the whole world) as a system of interlinked markets and estimate how tariffs, non-tariff barriers, and investment restrictions ripple through wages, prices, and output. The World Economic Forum's $6.9 trillion full-decoupling estimate and the IMF's more conservative $213-307 billion current-fragmentation figure both derive from variants of this approach, differing mainly in their assumptions about how completely trade blocs separate and how quickly firms can substitute alternative suppliers. The wide range between these numbers is itself informative: CGE models are highly sensitive to elasticity assumptions (how easily buyers switch suppliers) and to whether technology transfer effects are included, since restricting Chinese access to advanced semiconductors carries growth costs beyond simple trade-volume effects.

Sectoral analysis sharpens the picture considerably. Semiconductors sit at the centre of the conflict: the US CHIPS and Science Act subsidises domestic and allied fabrication capacity while export controls restrict China's access to sub-14nm lithography equipment, aiming to constrain Chinese progress in AI and military applications. China's countermove — restricting exports of gallium, germanium, and rare earth elements critical to chip manufacturing, EV batteries, and defence hardware — demonstrates that supply chain leverage runs in both directions. Roughly 70% of global rare earth processing capacity remains concentrated in China, meaning Western ‘de-risking’ in this sector requires years of mine development and refining capacity that does not yet exist at scale elsewhere.

The electric vehicle and battery supply chain illustrates a parallel dynamic. Chinese firms dominate lithium-ion battery production and lithium/cobalt/graphite refining; US and EU policy has responded with local-content requirements that effectively exclude Chinese-linked supply chains from full subsidy benefits. The result is duplicated capital investment — batteries and refining capacity built twice, once for each bloc — which is precisely the kind of redundant, cost-raising investment that CGE models attempt to capture as decoupling's dead-weight loss.

A financial and monetary dimension compounds the trade story. Export controls and tariffs have been paired with an expanding US sanctions regime, prompting some countries — chiefly China and Russia — to accelerate efforts to reduce dependence on the dollar-denominated financial system for trade settlement, a theme covered in EconoLens's earlier de-dollarisation analysis. Reduced dollar centrality is not yet a first-order macro effect, but it represents a slow-moving structural response to the same underlying fragmentation pressure, with implications for US Treasury demand and global liquidity over a longer horizon than the trade effects discussed above.

There is genuine debate among economists over whether ‘decoupling’ is the right description of what is actually happening, or whether ‘de-risking’ — the term favoured by the European Union and, more recently, by US Treasury officials — better captures a more selective, sector-specific separation that leaves the bulk of US-China commerce intact while ring-fencing strategically sensitive technologies. Aggregate bilateral trade data support a nuanced reading: direct US-China trade has fallen from its 2018 peak as a share of each country's total trade, yet China's global export share has continued rising, increasingly channelled through Vietnam, Mexico, and other intermediary economies whose imports from China and exports to the US have both grown in tandem — a pattern consistent with transshipment and final-assembly relocation rather than a genuine reduction in Chinese manufacturing's role in US-bound goods.

From a growth-theory perspective, the long-run cost of fragmentation operates primarily through the total factor productivity channel rather than through capital accumulation or labour supply. Trade openness has historically been associated with productivity gains from specialisation, economies of scale, and technology diffusion; reversing that openness, even partially, works against these channels. This is why the WEF's severe scenario shows growth effects (up to -6.4 percentage points) that dwarf what a simple tariff-revenue or trade-volume calculation would suggest — the model is capturing lost productivity growth compounding over multiple years, not merely a one-off price shock.

The events of 2026 add a further complication that most pre-existing decoupling models did not anticipate: the Iran conflict and the effective closure of the Strait of Hormuz from March 2026 onward, disrupting roughly a fifth of global oil flows and forcing the IMF to cut China's growth forecast to 4.4% in its April World Economic Outlook. This illustrates a broader point applicable well beyond the US-China relationship: in a fragmenting world with less spare capacity and fewer redundant supply routes — precisely the outcome decoupling is producing — external shocks of any kind propagate faster and hit harder, because the shock-absorbing flexibility of a more integrated system has been deliberately dismantled in the name of resilience. Whether this trade-off between geopolitical security and macroeconomic shock resilience is worth its measured cost remains the central, unresolved question for policymakers navigating the rest of the decade.