What the World Learns From a North American Trade War

Trade agreements are commitments about future behaviour. Their value to any prospective signatory is therefore not determined by their text but by observed conduct when honouring them becomes inconvenient.

The agreement governing North American trade is among the most comprehensively negotiated instruments of its kind. It contains consultation procedures, panel formation rules and binding determination mechanisms. When disputes among its parties are pursued instead through unilateral measures, the observation available to every other government in the world is that a modern, thoroughly drafted agreement did not constrain behaviour at the moment constraint was the point.

That observation is then priced. A country negotiating a bilateral agreement elsewhere concedes less, demands stronger enforcement language, or seeks shorter review cycles, because the expected value of the counterparty's commitment has fallen. None of this appears in trade statistics. It appears in the terms of agreements signed several years later, and by then it is attributed to other causes.

The cost is diffuse, delayed and borne by no identifiable constituency, which is exactly the profile of a cost that gets incurred repeatedly.

Where continental suppliers become more expensive, buyers look elsewhere, and European and Asian exporters capture a share of what moves. The immediate effect is smaller than headline substitution estimates imply, because qualifying a new supplier is a process of quality certification, capacity verification and compliance review that commonly runs several quarters, and buyers are reluctant to begin it for a barrier they expect to be temporary.

The important property is what happens afterwards. Once qualification is complete the switching cost has been paid, and there is little commercial reason to pay it again in reverse when the barrier lifts. A share of trade displaced during a dispute therefore never returns, and the longer the dispute runs the larger that permanent share becomes. This is the mechanism by which a temporary policy produces a permanent reallocation, and it operates quietly, one supplier decision at a time.

Blocs pursuing alternative settlement arrangements and independent commodity pricing mechanisms face real technical constraints. Liquidity, convertibility, legal enforceability and the depth of secondary markets are not problems that a North American trade dispute solves.

What the dispute supplies is the argument. The case for diversifying away from the incumbent system is not principally about efficiency, since the incumbent system is more efficient. It is about predictability, and specifically about whether the rules governing cross border commerce depend on the domestic political cycle of a single participant. Every episode in which they visibly do makes that case easier to state, and a case that can be demonstrated by pointing requires no persuasion.

This is the channel with the longest tail. Settlement infrastructure, once built and used, tends to persist beyond the conditions that motivated it, because the institutional investment has already been made.

The fourth channel is the least visible and possibly the largest. Investors allocating capital to production capacity require a return commensurate with perceived predictability. Persistent uncertainty about the terms on which goods cross an internal continental border raises the hurdle rate for any project whose economics depend on those crossings.

The effect appears as projects that are never announced. A plant that would have been built in Ontario or Michigan or Nuevo Leon is instead built elsewhere, or not built at all, and there is no press release recording the decision. Because the cost takes the form of an absence it never enters the debate, and because it applies to the continental bloc as a whole it falls on Mexico as well, including on the Mexican investment that trade diversion was supposed to attract.

Much of the supply chain reorganisation of recent years rested on the premise that relocating production toward politically aligned jurisdictions purchases reliability. A tariff dispute between two of the closest allies in the international system tests that premise directly, and the test does not go well.

Firms that consolidated production around alignment now face the question of what alignment actually guaranteed. The rational response is broader diversification rather than deeper concentration, which means more jurisdictions, smaller facilities, higher unit costs and more duplicated capacity. That is a real efficiency loss distributed across the entire global manufacturing base, and it is caused not by tariffs but by the demonstration that alignment does not predict predictability.

The exchange rate absorbs part of any sustained trade friction. The more trade dependent currency tends to weaken, which offsets some of the competitiveness loss from tariffs while raising the domestic cost of imported inputs.

For central banks the result is genuinely difficult, because the same measure operates in both directions. Weaker investment and softer external demand are disinflationary. Higher input costs and a weaker currency are inflationary. The two effects arrive on different timetables, with the cost side typically leading and the demand side following, so the policy question is one of sequencing rather than of direction. This is the channel through which a trade dispute reaches households in countries entirely uninvolved in it, by way of the rate decisions taken in response.

The credibility cost described in the first channel is recoverable, though not by announcement.

It is recovered by the visible use of the mechanisms that exist. A panel convened, a determination issued, a determination accepted by the losing party without political reopening. That sequence, performed once and observed internationally, is worth more than any quantity of communiqué language, because it is evidence rather than assertion.

The same logic applies to the risk premium. Capital returns when policy risk falls, and policy risk falls when firms observe rules being applied consistently across a change of administration. Both of these are achievable within the existing framework. Neither is achievable quickly, because credibility is rebuilt on the timetable of demonstrated behaviour rather than on the timetable of a news cycle.

The global cost of a North American trade dispute is mostly not paid in North America, and mostly not paid in tariffs. It is paid in the discount other states apply to treaty commitments, in supply chains that reorganise once and do not reorganise back, in the strengthened case for alternatives to the incumbent settlement architecture, and in the projects that quietly locate somewhere else.

Three of those four costs are difficult to reverse. That asymmetry, rather than any calculation about the rates themselves, is the strongest available argument for settling the dispute inside the institutions built to settle it.