Peter Navarro Revealed How Beijing Laundered $230 Billion in Exports
Consider a homeowner who installs a good deadbolt on his front door. A month later, he checks his records and finds that no one has forced the front door since the day the lock was installed. He also finds that his television is gone. The lock worked exactly as designed, which is why the thief stopped using the door. What the homeowner needs is not a better lock but a survey of his windows.
In 2018, President Trump used Section 301 to put a real tariff rate on Chinese goods, a price an importer had to actually pay at the border rather than a figure riddled with carve-outs. Chinese exporters responded, as any rational party would, to the price by looking for a cheaper route. They found one, and the route ran through Mexico, Vietnam, Malaysia, and a few dozen other jurisdictions willing to sell a stitch, a stamp, and a certificate of origin.
On Thursday, the Office of Trade and Manufacturing Policy, Peter Navarro’s shop, published a 25-page report called
. The language was blunt enough that even the Associated Press could not soften it into a story about supply chain diversification. The report describes a global architecture built for one purpose, a lattice of free trade zones, bonded warehouses, processing corridors, and re-export centers. It says plainly that the rerouting “made it look like US imports from China had dropped” while Beijing kept the factories, the workers, the tooling, and the profits. Navarro’s own summary was less clinical. For years, he said, the great transshipment scam has let communist China launder its exports.
Goods imported from China fell from $538.5 billion in 2018 to $308.7 billion in 2025, a decline of roughly $230 billion. Over the same period, imports from Mexico, Vietnam, India, and Malaysia rose by about $405 billion. The four substitute countries absorbed the Chinese decline and then some. They exceeded it by a wide margin, and they did so in exactly the product categories where the tariff differential was largest, the goods on which the 2018 rate had opened the widest gap between a Chinese invoice and an invoice from somewhere else. Nearshoring is a real phenomenon, and some of this is genuinely that. Nearshoring has a signature, though, and this does not match it.
For seven years, every quarter that Chinese import figures fell, people treated the drop as proof that decoupling was proceeding, that the tariffs had either worked or become unnecessary, and that it was time to discuss winding them down. The falling number was treated as a scoreboard rather than as a symptom, a quarterly verdict on whether the 2018 policy had done its job. A headline import figure measures where a shipment was last handled, not where it was made. The customs form records the last port, the last warehouse, the last jurisdiction willing to issue the paper, and it is silent on the factory that did the work. The decline in recorded Chinese imports shows that Chinese production learned to travel under an assumed name. Navarro’s report earns its keep by explaining what that statistic actually measures.
When production truly relocates, value relocates with it. A factory in Monterrey that actually builds a wiring harness buys Mexican steel, hires Mexican engineers, pays Mexican electricity bills, and the domestic content of that harness climbs year over year. The plant has a payroll that appears in Mexican labor statistics, a power bill that appears on a Mexican grid, and a purchasing file that fills with Mexican invoices. Over time, the harness becomes a Mexican article in the ordinary sense, because the work that makes it a harness is done there. Trade lawyers have a term for the threshold, substantial transformation, which means the article that leaves the second country has to be a different article than the one that entered it, different in name, character, and use. A shirt that arrives as a finished shirt and leaves as a finished shirt with a new label has been relabeled, and it fails the test on every count. The name is still shirt, the character is still a finished garment, and the use is still to be worn. The whole legal question of transshipment reduces to that distinction, and the data lets us test it.
Commerce already ran the test in solar and got an answer. Its investigators found five companies performing what the department called minor processing of Chinese cells in Cambodia, Malaysia, Thailand, and Vietnam, assembly operations that existed to change a customs form rather than to build a product. The cells were Chinese, and the work in those four countries was thin enough that Commerce named it minor, the kind of handling that leaves name, character, and use untouched. One of the firms, Qingdao Sunsong, said the quiet part in its own filing. Its Thai assembly line existed, in the company’s words, to ease the cost pressure of US tariffs, and the company estimated that the Thai operation added between 4% and 8% of the finished value. Take the midpoint, and you are looking at a 6% contribution, a sliver of labor and overhead laid over a Chinese cell so that the finished panel could travel under a Thai name. A company that adds 6% of value to a Chinese product and then presents it as a Thai product is buying a passport.
USTR’s 2026 report on transport equipment found that US content in Mexican transport equipment fell from 23.1% to 18.3%, while Chinese content in that same Mexican output rose from 4.5% to 7.1%. American content lost 4.8 points. Chinese content increased by 2.6 points within the same Mexican export stream over the same window. Mexico’s factories are running, and Mexico’s share of the finished vehicle is real, but the marginal input is increasingly Chinese and decreasingly American. The North American agreement was written to give preferential access to North American production. What it is delivering, at the margin, is preferential access to Chinese production wearing a North American label.
Customs and Border Protection, operating under this administration, uncovered more than $400 million in unpaid duties in 2025, including cases involving 23 importers and Chinese shell entities routing goods through Indonesia, Korea, and Vietnam. Each case has an importer, a shell, a third-country stamp, a duty that was never paid, and a file that now has a dollar figure attached. Those are seizures, penalties, and case files, the kind of record that survives a press cycle. The model numbers, when you look at them, are sobering in their own right. The OTMP report puts the central estimate of illegal transshipment at roughly $75 billion a year, with a plausible range running much higher, and it estimates the annual hit to federal revenue at $19 billion to $26 billion, with 450,000 American jobs displaced and $113 billion to $150 billion in lost annual output. More than 40 jurisdictions appear in the network.
A tariff is a price attached to a behavior. Its purpose is to change the relative attractiveness of two sourcing decisions, and revenue is a secondary effect. The importer is supposed to feel the difference between landing a Chinese good and landing a good that was actually made somewhere else, and that difference is the whole instrument. An unenforced tariff actively rewards the party willing to commit fraud and punishes the party unwilling to do so. The American firm that reshored a plant, and the honest Vietnamese manufacturer who actually built a factory, are both now competing against a Chinese producer paying a Vietnamese fee for paperwork. One paid to move production. One paid to build it. The third paid a clerk. The dishonest route becomes the cheapest route, and the policy inverts itself into a subsidy for cheating.
This administration’s approach is diffrent in kind from the white papers that preceded it. Executive Order 14411 puts the obligation where the leverage is, on the importer of record, through accountability rules and bonding requirements that make the party bringing the goods in bear the cost of a false origin claim. If the origin is false, the bond is the first payment to move, and the importer is the one who posted it. The anti-transshipment clauses now being written into trade frameworks do the same work internationally, and the one year retroactive lookback is the piece that actually bites, because it changes the arithmetic of the gamble, an importer who fakes an origin certificate today can no longer count on the fact that a shipment cleared last March is a shipment nobody will ever revisit, and when you add CBP’s new detection system, the one Navarro has taken to calling Detective Border, you get pattern recognition applied across millions of entries rather than the spot checks a human workforce could manage. Last March was safe because the calendar had turned. Under a lookback, last March is still live, and the duty that should have been paid is still collectible. A rule nobody audits is a suggestion, and the lookback is what turns the audit from a warning into a bill.
Xi Jinping arrives on September 24. The temptation at any such meeting is to trade a headline number for a handshake, to announce a rate and let the enforcement details go to a working group that meets in the spring. That would be the one outcome Beijing genuinely wants, because a 50% rate on Chinese goods is worth nothing to us and costs them nothing if a factory in Guanajuato can void it with a stamp and a seperate carton. Origin discipline belongs in the text of whatever comes out of that meeting, with verification rights, penalties that scale, and a lookback that survives the news cycle. Put those terms on the page, and the rate announced in the communique is still in effect in October.
If you enjoy my work, please subscribe https://x.com/amuse/creator-subscriptions/subscribe
Alexander Muse is a Fellow at the John Milton Freedom Foundation and publishes daily political analysis at amuseonx.com. Primary sources cited in this piece are linked inline; campaign finance figures are drawn from FEC filings, polling data from publicly released crosstabs, and legal claims from filed pleadings. Corrections are posted to the original URL with a dated changelog. Readers who identify errors are invited to contact the author directly. Each op-ed edited for grammar and clarity using Ai in a sponsored partnership with Grammarly. Data provided in a sponsored partnership with Polymarket.
Peter is amazing! Very smart man 😊