"Come make your product in America," Trump told business leaders at Davos, framing tariffs as a friendly nudge: build here, skip the duties.
A year later, the July trade deficit widened by $17.4 billion to $88.6 billion — the widest since March 2025. The gap the tariffs were supposed to close blew out by nearly a quarter in a single month.
And the foreign capital he was courting? It's walking away. German companies just cut their US investment to a three-year low.
Both halves of the tariff logic — shrink the deficit, pull production onshore — failed in the same month. But here's the part Washington keeps missing: the tariffs aren't the main culprit. The whipsaw around them is.
Last week this column tracked the IEEPA-to-Section-122 legal collapse — the daisy chain of authorities the administration reaches for each time a court knocks one down. This week we pair that thread with hard research on what the resulting uncertainty actually costs, and with a deficit print that mocks the policy's own stated purpose.
Start with the FDI research, because it reframes everything. Jonathan Ostry and colleagues examined bilateral investment flows from over 200 source countries into 35 OECD economies across nearly four decades, isolating two distinct shocks: trade restrictions and trade-policy uncertainty.
The point of separating them is that they are not the same animal, and they do not behave the same way.
Restrictions have ambiguous effects. Yes, tariffs raise trade costs and deter efficiency-seeking, export-oriented FDI that depends on moving intermediate goods across borders. But they can also pull in tariff-jumping investment — firms building inside the wall to preserve market access. That's exactly the behavior seen last year when Chinese firms expanded North American production to dodge US duties. It's also what Trump was selling at Davos.
Uncertainty has no such offsetting upside. It is pure deadweight loss.
The mechanism is the option value of waiting. When you're contemplating a plant — a large, sunk, irreversible commitment — the rational move under a fog of policy is to sit on your hands until the fog clears. Every credible threat of a rule change raises the value of delay.
The magnitudes are the story. A one-standard-deviation uncertainty shock cuts inward FDI by roughly 50% in the first year, measured against the sample mean. The comparable tariff shock cuts it about 15%.
That's roughly three times more potent. The unpredictability does triple the damage of the restriction itself.
A one-standard-deviation uncertainty shock cuts inward FDI by roughly 50% in the first year; the comparable tariff shock cuts it about 15%. Roughly three times more potent.
Which means Washington is paying an enormous capital-formation tax — and getting nothing for it. A committed protectionist can at least point to tariff-jumping FDI and reshored production as intended effects of a tariff wall. Nobody gets anything from the fog. It is, in Ostry's words, an own goal.
The damage is worst precisely where it should hurt most: close allies. The research finds the FDI hit is amplified when host and source countries are geoeconomically and geopolitically close — deeply integrated production partners, aligned UN voting records.
So the countries most likely to build in America — arguably close partners like Canada and Germany — may be among those most repelled by the volatility. Germany cutting US investment to a three-year low isn't noise. It's the model working as predicted.
The greenfield data confirm it in aggregate. Global FDI rose 14% in 2025, but productive investment grew only 5%, greenfield announcements fell 16%, and in tariff-exposed, value-chain-intensive sectors, new project numbers dropped by a quarter.
A quarter. That is capacity that would have been built and now won't be — or will be built somewhere else, in a corridor that offers something the US no longer does: predictability.
Now to the engine of that uncertainty, because the crucial claim is that it's structural, not incidental. This isn't policy that happens to be volatile. It's volatility manufactured by the legal scaffolding.
Follow the sequence. The Supreme Court's February ruling reportedly blocked the use of the 1977 IEEPA to impose sweeping tariffs. Trump pivoted to Section 122 of the 1974 Trade Act. When that expired in late July, he pivoted again — to Section 301, slapping 10% to 12.5% tariffs on more than 60 countries.
Three legal workarounds in under a year. Each pivot resets the clock, reopens the litigation risk, and tells every investment committee on earth that the rules governing their sunk capital have a shelf life measured in months.
The latest move takes it further. The new Canada tariffs, per PIIE, "invoke a zombie authority found in the notorious and discredited Smoot-Hawley Act, which deepened the Great Depression." Their verdict: "The courts should reject these tariffs as patently illegal."
Reaching into Smoot-Hawley — the statute synonymous with trade-policy catastrophe — to justify tariffs on America's closest ally is not a stable high-tariff wall. It's a legal experiment whose own architects can't be sure will survive review.
And that's the deeper point. Even the maximalist case for tariffs assumes durability. A predictable 30% wall lets firms recalculate and, in some cases, jump it. A wall that might be 30% or 12.5% or struck down next quarter, under a statute that might be IEEPA or Section 122 or Section 301 or Smoot-Hawley, cannot be planned against at all.
The option value of waiting stays permanently elevated. Investors keep sitting on the sidelines because the policy never resolves.
Which brings us back to the $88.6 billion. The number will be misread — the July spike owes much to a surge in AI and semiconductor imports, capital goods hitting a record $140.3 billion as the computing buildout pulls in foreign hardware. That's an economy investing, not collapsing, and it's a separate story.
But the political reflex won't parse that. The gap widened, so the answer will be more tariffs, more restrictions, another authority-swap — feeding the exact uncertainty that's already repelling three times more investment than the duties themselves.
That's the trap. The deficit doesn't shrink, so the policy intensifies; the intensification raises uncertainty; the uncertainty deepens the capital drought that would eventually help close the gap.
Washington has built a machine that fails on both of its own terms — and mistakes each failure for a reason to run it harder.
The fix costs nothing. Pick a policy, any policy, and hold it still. Even a committed protectionist would be better served by a stable high wall than by this fog, because the fog is the part with no upside at all.
Instead, the administration is reaching for Smoot-Hawley. The unpredictability isn't a bug in the tariff regime. At this point it is the regime.