Factories as Tariff Shields: Building Inside the Wall
The standard answer to import duties is to build where you sell — but a factory is a decades-long bet placed against politics that can change in a single budget.
Why the wall invites a ladder
Every tariff carries an implicit instruction: if you want to sell here without the penalty, build here. Chinese carmakers have heard that instruction in market after market, and the response has become a recognisable pattern. When the European Union added duties on Chinese-built electric cars, BYD was already building a plant at Szeged in Hungary, inside the EU's own borders. Cars assembled there are European production for tariff purposes, whatever badge they wear or wherever the parent company sits.
Turkey shows the logic in a sharper form. Turkey is not an EU member, but it sits in a customs union with the EU, which means Turkish-built cars can enter the European market without attracting the EV duties applied to cars shipped from China. BYD announced a Turkish plant against that backdrop. The arrangement also suited Ankara: Turkey had itself imposed extra tariffs on Chinese cars, while offering an exemption to makers willing to invest locally. The tariff, in other words, was designed less to keep Chinese cars out than to convert imports into factories, jobs and tax revenue. That is a transaction more and more governments are trying to strike.
Assembly lines on three continents
The same calculation is running elsewhere. In Brazil, BYD took over a former Ford plant at Camaçari, a site with an existing workforce and supplier base. Brazil is phasing its EV import tariff back up in steps, reaching 35% by 2026, and the timetable functions as a deadline: import while you can, but be building locally before the rate peaks. Thailand went further and wrote the trade into policy. Under its EV3.5 programme, the right to import cars is tied to local production quotas, so a maker's import privileges depend on how much it manufactures in the country. Chinese brands built Thai plants not simply to dodge a duty but because the subsidy and import system required it.
At the lighter end of the scale sits kit assembly. Leapmotor cars are assembled in Poland through the brand's partnership with Stellantis, using an existing European plant rather than a greenfield site. Assembly from imported kits can satisfy local-content rules and reduce exposure to duties at a fraction of the cost of a full factory, though critics argue it delivers fewer local jobs and less supplier investment than genuine manufacturing. Governments writing the rules care about that distinction, and the line between "made here" and "screwed together here" is becoming a live policy question.
Mexico is the counterexample. Geographically, it was the obvious North American answer: an established car industry, a trade agreement with the United States, and lower costs. But U.S. pressure made Chinese plant plans there stall, because a factory that exists to route around American tariffs invites exactly the political response it was built to avoid.
The bet that cannot be hedged
A factory is the least reversible decision a carmaker makes. Plants take years to plan, build and ramp up; tariff regimes can be rewritten in a single legislative session. A company that commits capital today is betting that the rules rewarding local production will still be in place when the line starts running — and that demand will be there to fill it. If tariffs fall, the local plant may be an expensive way to build cars that imports could have supplied more cheaply. If tariffs rise or local-content rules tighten, the plant may not qualify as local enough.
There is a spectrum hidden inside the word "factory." At one end is the full manufacturing plant: stamping, welding, painting and assembly, with a growing local supplier base — the expensive end, and the one governments reward most. At the other is the assembly hall, where kits of parts made in China are bolted together locally. Between them sit hybrids: assembly with a rising share of local content, written into the agreement as a schedule the maker must meet to keep its privileges. Thailand's quotas and Turkey's exemption both push makers up that spectrum. A maker choosing the cheap end holds a weaker shield: if a government later decides kit assembly does not count as local production, the protection the plant was built to secure can evaporate without a single duty rate changing.
There is also the question of what the factory buys politically. A plant employs local workers, signs local suppliers and gives regional politicians a reason to defend the company against the next round of measures. Jobs are a form of cover that no lobbying budget quite matches. That is why the factory announcements tend to be loud: the audience is not only customers but the governments setting the rules.
The risk runs both ways, though. Capacity built ahead of demand sits idle, and a half-empty plant is a visible reminder of overreach. Chinese makers are building on several continents at once, in markets where their sales are still small. Whether those plants look like foresight or excess in ten years depends on politics they do not control — which is, of course, the reason they are building them in the first place.
