Why ‘decoupling’ from China is harder than it sounds

Decoupling from China has become one of the most repeated phrases in debates about the global economy. The idea sounds clean: untangle supply chains, reduce dependence and build resilience elsewhere. In practice, the slogan collides with an awkward reality. The connections built over decades are deep, specialised and expensive to replace, and wishing them away does not make it so.

This is not an argument that reducing concentration is impossible or undesirable. It is an argument that the task is slower, costlier and more partial than confident talk of decoupling implies, and that pretending otherwise leads to disappointment.

What people mean by decoupling

The word covers a range of ideas. For some, it means moving final assembly out of one country. For others, it means reducing reliance on a single source for critical inputs, or building parallel systems in sensitive areas such as technology. These are very different ambitions, and lumping them together under one slogan obscures how difficult each can be.

It helps to separate diversification, spreading risk across more suppliers and locations, from separation, trying to sever ties almost entirely. The first is pragmatic and already under way. The second runs into formidable obstacles.

Supply chains are deeper than they look

A finished product is rarely made in one place. It is the end point of a long chain of components, materials and processes, often spanning many borders. A single device may contain parts from dozens of suppliers, each with its own upstream dependencies. To understand how these webs formed, our history of Asia-Pacific economic integration is a useful companion.

Because of this depth, moving one stage of production does not move the whole chain. A factory may relocate final assembly while still importing most of its components from where it always did. The visible flag on the box changes; the underlying dependence often does not.

The power of scale and clustering

Large manufacturing hubs enjoy advantages that are hard to reproduce quickly. When many suppliers, skilled workers and specialist firms cluster together, they create what economists call agglomeration benefits.

• A deep pool of workers trained in specific skills.
• Dense networks of nearby suppliers who can respond fast.
• Infrastructure, logistics and services built around the industry.
• Accumulated know-how that lives in people and routines, not manuals.

Replicating such an ecosystem elsewhere takes years and heavy investment. A new location may offer lower wages, but it often lacks the surrounding web that makes a mature hub so efficient.

Scale is not easily matched

Some manufacturing capacity exists at a scale that few other places can match in the near term. When a single region produces an enormous share of a given product, shifting even part of that volume requires building vast new capacity. That capacity does not appear on demand; it must be financed, constructed, staffed and refined.

The hidden cost of duplication

Building parallel capacity is expensive, and someone pays. Duplicated factories, higher-cost suppliers and less efficient logistics tend to raise prices. Resilience, in other words, is not free, and the costs often land on consumers and firms in ways that are easy to overlook when the goal is framed as strategic necessity.

Relocation can create new dependencies

Even successful diversification can be deceptive. Production that moves to a new country may still rely heavily on inputs, machinery and materials from the original source. The result can be a longer, more complex chain that is harder to see but not necessarily more independent.

This is why headline figures about shifting trade flows can mislead. A drop in direct imports from one country may be partly offset by that country’s inputs arriving indirectly, routed through a third location where final assembly now takes place.

What is realistic

None of this means firms and governments are powerless. Sensible steps are available and are being taken.

• Diversifying suppliers so no single source is irreplaceable.
• Building buffer stocks of critical inputs.
• Developing alternative capacity in genuinely strategic areas over time.
• Mapping supply chains to understand hidden dependencies.

These measures reduce vulnerability without pretending that deep, efficient trade relationships can be switched off overnight. They accept trade-offs rather than denying them.

An honest framing

In my view, the most useful way to think about this is to drop the language of decoupling in favour of something more modest and more accurate: reducing concentration where it matters most, while accepting continued interdependence elsewhere. The region’s economies are woven together by scale, specialisation and decades of investment. Those ties can be managed and rebalanced, but they cannot be cheaply cut.

The honest conclusion is that resilience and efficiency pull in different directions, and every choice involves a cost. Recognising that tension is not defeatism. It is the starting point for policy that works in the real world rather than in a slogan, and it sets realistic expectations for how long rebalancing truly takes.

Frequently asked questions

Does decoupling mean cutting all ties with one economy?

Usually not. In practice it ranges from reducing reliance on a single source to, more ambitiously, trying to separate entirely, which is far harder than it sounds.

Why can’t supply chains move quickly?

Products depend on long chains of specialised components and clustered suppliers, so relocating one stage rarely moves the underlying dependence, which takes years to rebuild elsewhere.

Is diversification the same as decoupling?

No. Diversification spreads risk across more suppliers and locations and is realistic, whereas full separation faces formidable obstacles of scale and cost.

What are realistic steps to reduce risk?

Diversifying suppliers, holding buffer stocks of critical inputs, mapping hidden dependencies and building alternative capacity in genuinely strategic areas over time.