The Tariff Is Not the Price Tag

The economic argument over tariffs often proceeds as if a customs duty were a remote-control button: press it in Washington, jobs materialise in Ohio, and somebody else pays. The available evidence suggests something less theatrical. A model of full US-China trade cessation puts the long-run consumption loss from an orderly break at 0.59%, but a disruption compressed into months at about 2.4%—roughly one to two years of normal consumption growth. The meaningful variable is not merely whether trade is cut; it is whether firms can find, qualify, finance and physically receive substitutes before inventories, contracts and production schedules run out. The Cost of Decoupling with China

That distinction is the supply chain in its natural habitat: not a slogan, but a timetable. Under the gradual scenario, prices adjust and resources shift; under the sudden one, capital is fixed and the economy gets less time to improvise. The model also finds that closing third-country routes for Chinese inputs raises the projected long-run consumption loss from 0.59% to 0.73%. Routing a component through another country is therefore not proof that dependency vanished; it may simply mean the dependency acquired a longer itinerary. The Cost of Decoupling with China

Kevin Warsh’s first speech as Fed chair supplies the awkward macroeconomic backdrop. He said inflation remains too high, judged broad financial conditions hard to call restrictive, and described consumer spending as having risen more than 2% over four quarters despite oil prices, tariffs and other shocks. He offered no rate path, because apparently even the trail map has been placed in a locked filing cabinet. But the key point is serious: when trade policy changes the cost and availability of inputs, the Fed inherits an inflation problem whose first invoice may arrive long after the announcement. Warsh’s Jackson Hole assessment

The reassuring headline from the decoupling model—less than a 1% long-run consumption loss under an orderly baseline—deserves to be read in the same voice one uses for a favourable weather forecast that mentions hail. China’s share of US imports fell from just over 20% at its 2018 peak to around 7% in 2025, while Chinese-import spending fell to about 1.6% of personal consumption expenditures. That reduced direct exposure helps explain why the estimated aggregate loss is lower than it would have been a decade earlier. The Cost of Decoupling with China

But an aggregate is a national average, not a household budget. The authors note that lower-income households devote a larger share of spending to imports from China, and that China supplies machinery, equipment and parts as well as consumer goods. The first effect concentrates pressure at the checkout; the second travels through factory costs, investment decisions, productive capacity and, ultimately, real wages. A tariff can thus look manageable in a macro chart while being highly legible to the family replacing a broken appliance or to the small manufacturer repricing a production run. The Cost of Decoupling with China

The historical tariff record in the analysis is equally unromantic. The average US tariff rate on Chinese goods rose from roughly 5% in 2017 to 25% in 2020, alongside a decline in Chinese imports relative to consumption from 4% to around 1.5%. That demonstrates adaptation, not costlessness. Importers can switch suppliers, alter product specifications, squeeze margins, redesign logistics or pass costs onward; consumers and workers receive the residue in different proportions and at different times. Economics calls this substitution. Everyone else calls it discovering that the identical-looking replacement has a six-month lead time. The Cost of Decoupling with China

The model’s central condition is worth underlining: a trade break is more manageable when it is gradual, predictable and mediated by prices. Prices are efficient messengers of scarcity, but they are not industrial policy with a forklift certificate. They can induce buyers to economise and firms to seek alternatives; they cannot instantly create downstream processing capacity for strategically important materials, certify a new supplier, or make specialised machinery appear beside a loading dock. The analysis flags rare earths as a particular risk because China dominates downstream processing, while the cost of building productive capacity elsewhere remains unknown. The Cost of Decoupling with China

Nor does the model claim to have captured every way a rupture could go wrong. It excludes a possible financial crisis or recession, does not fully model input-containing goods routed through third countries, and assumes prices can adjust rather than being frozen by controls or rationing. It also excludes tariff revenue—estimated at roughly 0.2% of consumption expenditures in 2024—from its welfare calculation. These omissions do not invalidate the estimate; they tell us what it is: a disciplined scenario, not a warranty card for geopolitical improvisation. The Cost of Decoupling with China

For the Fed, this is why “tariffs raise prices” and “tariffs do not matter” are equally inadequate bumper stickers. Warsh says the central bank needs confidence that underlying inflation is moving to target at sufficient speed, even as he sees a robust economy, stable labour market and strong capital spending—more than half of this year’s capex growth, by his estimate, tied to the AI build-out. Monetary policy can lean against economy-wide demand; it cannot decide whether a missing input is a transitory annoyance or the beginning of a costly new production geography. The tariff is collected at the border. Its economic afterlife is distributed across contracts, margins, investment plans and shelves, where it becomes much harder to campaign against. Warsh’s Jackson Hole assessment