A paper released this week by Peter Navarro, Counsellor to the President for Trade and Manufacturing, accuses more than forty countries of serving as a shadow network through which Chinese exports reach the United States while avoiding tariffs. Titled The Great Transshipment Scam, it names India alongside Mexico, Canada, member states of the European Union, Japan and South Korea, and describes goods being relabelled, repackaged, lightly processed or rerouted so that their origin is obscured.
The charge against India is unusually specific. The paper says the production belt running from Pune through Gujarat to Chennai takes in Chinese pumps and compressors, which are then relabelled and shipped onward to the United States. Navarro put it in one line that is meant to travel, saying that a Chinese pump leaving Pune as Indian is a pump not machined in Cincinnati, Dayton or Columbus.
Before any of that can be assessed, look at the number the paper is built on. It puts the annual value of illegally transshipped goods at somewhere between 40 billion and 303 billion dollars, the range depending on which methodology and definition are used. That is a factor of more than seven. A range that wide is not a measurement of a thing. It is an admission that the thing has not been measured, and that what it turns out to be worth depends almost entirely on where a line is drawn.
The line in question is a real and old problem in customs law rather than a rhetorical one. Origin is determined by whether goods underwent substantial transformation in the country that ships them, and that test has always been a judgement rather than a fact. Sticking an Indian label on a finished Chinese pump is fraud, and nobody sensible defends it. Importing Chinese castings and machining, assembling and testing them in Pune is manufacturing, and it is what every industrialising economy has done, including China itself for two decades. Between those two poles sits a long stretch of activity that can be described either way depending on the intent of the describer. The gap between 40 and 303 billion dollars is precisely the width of that stretch.
This matters for India more than for most of the others named, because the accusation lands exactly where its industrial policy is aimed. The whole point of the incentive schemes of the past few years has been to attract assembly first and hope that value addition deepens afterwards, since no country has ever begun by making everything. A policy that succeeds in its early phase looks, from a customs desk in another country, a great deal like the thing the paper is describing. That is uncomfortable, and it is not the same as being guilty of it.
The rhetorical move in the Cincinnati line is worth naming too. It assumes that the alternative to a pump assembled in Pune is a pump machined in Ohio. That is a claim about a counterfactual, not a finding. On the industry's own economics the more likely alternative is a pump assembled in Vietnam, Mexico or Thailand, which is what happened the last time the cost of routing through one country rose. Tariffs move where assembly happens far more reliably than they move whether it happens in the United States.
The enforcement half of the paper is where this becomes concrete for Indian exporters. Washington proposes an artificial intelligence screening system called Detective Border, combining shipment data, routing histories, product classifications, ownership relationships, production capacity indicators, anomaly detection and computer vision, to sort high risk consignments from legitimate trade, backed by interdiction, duty collection, penalties and exclusion of suspect goods.
Production capacity indicators is the phrase to watch. The underlying logic is that a plant which could not plausibly have produced what it claims to have produced is probably passing on someone else's output, and as a detection heuristic that is perfectly reasonable. It also describes, with some accuracy, a genuinely new factory scaling quickly from a low base, which is what a successful industrial policy generates by the hundred. The exporters most likely to be flagged by such a system are the ones growing fastest, which in India at the moment is the same list as the ones the government has been subsidising.
No response has come from the Indian government so far. The useful one would not be a denial of the general phenomenon, which is real, but a demand that the definition be fixed before the enforcement machinery is switched on. A screening system pointed at a category whose value cannot be pinned down within a factor of seven will produce a great many decisions that nobody can check.