Companies announce changes to where they manufacture, and the trade figures move years later. The delay reflects what relocating production actually requires.
Qualification is the binding step
A new supplier must demonstrate that its output meets specification consistently, which involves sample runs, inspections, certification and often regulatory approval.
For regulated goods such as medical devices, aerospace parts or pharmaceuticals, that approval process alone can take years and cannot be compressed by spending more.
Until qualification completes, the old supplier continues shipping, which is why announcements and trade statistics diverge for extended periods.
Clusters cannot be relocated piece by piece
Manufacturing depends on nearby suppliers of components, tooling, packaging and maintenance, and on a workforce with the relevant skills already present.
Moving a final assembly plant without that surrounding network raises costs, because inputs must now be imported and specialist repairs require flying people in.
Building the network takes a decade or more, which is why relocation tends to go to places that already host a related industry.
Contracts and tooling create sunk costs
Supply agreements run for years, and the moulds, jigs and testing equipment for a product are usually installed at a specific factory and expensive to duplicate.
Writing off that investment is a real cost that must be weighed against tariff or risk savings, and the calculation often favours waiting for the next product cycle.
Firms consequently shift at model changeovers rather than mid-production, which ties the pace of relocation to product development cycles.
Diversification is more common than relocation
Rather than move, firms frequently add a second source elsewhere while retaining the original, accepting higher unit costs in exchange for reduced concentration risk.
This shows up in the data as a gradual change in shares rather than as a departure, and it is far less visible than an announced move.
It also preserves the option of reversing course cheaply, which matters when the policy conditions prompting the change may not survive the next election in either country.
Intermediate goods obscure the picture
Final assembly can move while components continue to originate where they always did, so the country of export changes without the underlying dependency changing much.
Trade statistics record the last country of substantial transformation, which means headline bilateral figures can shift considerably with little change in who actually makes what.
Analysts therefore look at value-added measures rather than gross flows, since those track where production genuinely occurs.