Tariffs After the Ruling

A court decision struck down one layer of a four-layer tariff schedule. This report rebuilds the stack, explains why merchandise trade volumes still slow sharply in 2026, and sets out what firms should model instead of an average rate.

A revised schedule

The United States average effective tariff rate reached 11.0 percent in early 2026, the highest level since 1943. On 20 February the Supreme Court declared the tariffs imposed under the International Emergency Economic Powers Act unconstitutional, removing one of several stacked layers.

Since then the most common question we have been asked reduces to: by how much did my tariff fall. For most firms the honest answer is that the question is wrong, because they were never facing the average.

Why the average rate misleads

A shipment's total duty is the sum of independent layers. There is a most favoured nation base rate, set by product category and applied equally to WTO members regardless of origin. On top sit Section 232 rates, Section 301 rates and any global surcharge, each with its own legal basis, product scope and origin scope. A single shipment from a highly exposed origin can face all of them simultaneously.

Removing one layer does not scale the others down. It changes the relative cost of sourcing a specific product from a specific origin, which is a composition effect rather than a level effect. Two firms importing the same headline category from different countries have experienced this ruling completely differently.

An average tariff rate is a weighted summary of a schedule no individual importer faces. Incidence has to be computed at the level of the tariff code and the origin.

The preference channel compounds this. Where a trade agreement provides a lower rate and goods qualify on rules of origin, the headline rates are not the operative number. That claim rate is administratively contingent: it depends on documentation, origin determination and the cost of compliance relative to the duty saved. The effective rate can therefore move with no policy change at all.

Why volumes still slow

The intuitive inference from a falling effective rate is that the trade outlook should improve. It has not. The WTO's March 2026 baseline has merchandise trade volume growth slowing to 1.9 percent in 2026 from 4.6 percent in 2025, recovering to 2.6 percent in 2027. Commercial services eases to 4.8 percent from 5.3 percent, then accelerates to 5.1 percent. Goods and services together grow 2.7 percent in 2026 against 4.7 percent in 2025.

The 2025 number was borrowed

Trade held up far better in 2025 than the tariff arithmetic implied, and the reasons are all timing or scope. New US tariffs were suspended until August, retaliation was limited, exemptions were numerous, and there was substantial frontloading of imports ahead of expected increases. Separately, a surge in demand for AI-enabling goods offset much of the drag.

Frontloading borrows from the future at a one-to-one rate. The WTO's own framing is that the tariff impact shifted into 2026 as accumulated inventories are drawn down. The sharp drop in North American imports originally projected for 2025 is now expected to materialise in 2026 instead.

Uncertainty is a separate tax

A tariff that might be reimposed is not equivalent to no tariff. UNCTAD estimates foreign direct investment in tariff-exposed and global value chain intensive sectors, including textiles, electronics and machinery, fell by 25 percent in 2025. Investment responds to the variance of policy as much as to its level, and a schedule that has been through a court is more uncertain, not less. A shortening of value chains reduces the trade intensity of growth, which is persistent rather than cyclical.

Redirection of flows

The adjustment has been substantially geographic. South-South trade grew 8 percent year on year in value terms in the first half of 2025, against 6 percent for world trade overall, with a rough estimate suggesting South-South trade excluding China grew around 9 percent. Increased trade among emerging economies, a measured response to tariff changes, and the AI goods cycle together eased the setbacks of 2025.

This is the substitution margin working, and it argues against reading slower aggregate volumes as a collapse in trade. It is a reallocation with a transition cost, and the cost falls on whoever holds fixed assets pointed at the old pattern.

Recommended modelling approach

1. The stack for the specific codes and origins in the book, not a portfolio average. This is the single highest-value change most firms can make to their tariff modelling.
2. The preference claim rate as a variable, with a sensitivity for administrative tightening. It moves the effective rate without any policy change.
3. The sunk cost of adjustments already made. Firms that re-routed, re-sourced or relocated do not unwind when a rate falls. That sets the switching threshold for going back.
4. A separate risk premium for policy variance, applied to any commitment with a payback longer than the political cycle. The FDI data suggests firms are already applying one.
5. AI-enabling goods as a distinct series. This category has repeatedly offset tariff drag in the aggregate. If the technology cycle cools, the underlying weakness becomes visible immediately.

Source register