Tariff Engineering: Legal Ways to Lower Duty Rates

Tariff engineering is the deliberate design of a product so it lands in a lower-duty customs classification. It is legal as long as the product genuinely matches that classification when it crosses the border. The payoff is real enough that major brands plan for it at the design stage. Converse ships Chuck Taylors with a thin layer of felt on the soles so they classify as slippers, which carry a 6% duty versus 20% for sneakers, according to U.S. International Trade Commission data reported by Modern Retail. Columbia Sportswear sews small zippered "nurse's pockets" below the waistline of women's shirts because that feature drops the shirts into a lower-duty apparel classification. None of this happens after the goods are made. The duty decision is baked into the product before a single unit leaves the factory.

That timing is the part importers sourcing from China most often miss. Tariff engineering is not a paperwork exercise you do at the port; it is a manufacturing decision you make with your supplier. Changing a sole material, repackaging a set, or routing assembly through a different country all require factory cooperation months before shipment. This is where a sourcing partner earns its keep: CN Ally coordinates the engineering change with the factory so the goods that actually ship match the classification you are claiming. Without that link between design intent and factory execution, the strategy stays on paper.

What is tariff engineering, and is it legal?

Yes, it is legal — provided the product you import genuinely is the thing you say it is. Tariff engineering refers to design, sourcing, and manufacturing decisions made so the finished good is classified at a lower tariff rate than it otherwise would have been. The legal boundary is the doctrine of "commercial reality": the engineered feature must be a genuine step in the manufacturing process, or the product must have a real commercial identity as imported. If the feature exists only on paper, or is stripped off the moment the goods clear customs, it crosses the line into evasion.

The foundational cases are summarized in the tariff engineering overview on Wikipedia, which also covers the commercial reality doctrine. In Merritt v. Welsh (1881), a sugar importer added molasses to highly refined sugar to darken its color. At the time, the tariff schedule taxed sugar by color under the Dutch standard: darker meant lower grade, lower duty. The Port of New York ran chemical tests, found the added molasses, and applied the higher rate anyway. The Supreme Court sided with the importer. Congress had written the test around color, not chemical analysis or intent, so color was the only legal basis for classification. Justice Matthews put it plainly: "Has not a manufacturer a right to make his goods as he pleases? If they are less marketable, it is his loss; if they are not less marketable, who has a right to complain?"

United States v. Citroen (1912) added the other half of the rule: goods are classified in their condition at importation, not by what they were or will become. An importer, in short, has the right to design around the tariff schedule. Customs lawyer John Foote of Kelley Drye & Warren told CNBC there is "nothing inherently illegal or even untoward about leveraging strategic design choices that result in creating different products that are subject to different tariff classification and duty rates."

Which methods do importers actually use?

There are five standard approaches. Each changes the product, its value declaration, or its origin story — not just the paperwork.

Method · What changes · Real example · Fits best

• Material and design changes: Swapping fabrics, coatings, or small features · Felt on Converse soles (slippers, 6%, vs. sneakers, 20%); Columbia's waistline pockets · Apparel, footwear, consumer goods
• Functional additions: A feature that shifts the heading · Sensor-equipped trash bins reclassified from metal products to mechanical devices · Electronics-adjacent goods
• Condition at importation: Shipping unassembled or unfinished · Furniture shipped flat-pack instead of assembled · Furniture, electronics, kits
• First sale valuation: Declaring the manufacturer's price, not the middleman's · $6 first-sale price vs. $9 trading-company price · Buyers using trading companies or agents
• Country-of-origin planning: Assembling or finishing in an FTA partner · Garments cut and sewn in a free-trade partner country · High-volume product lines

The scale of the opportunity comes from the Harmonized System itself: more than 5,000 product classification codes used by over 200 countries, where small differences in construction produce big rate changes. Apparel and footwear are the most responsive categories, because duty rates there often hinge on exactly the details a designer controls — fabric composition, waterproofing, zippers, pockets. Attorney Mario Torrico, who advises U.S. and foreign manufacturers on customs compliance, notes that "sectors where small design tweaks can shift classifications tend to have more options."

Recent examples show the playbook spreading. Walmart has worked with suppliers to swap aluminum, subject to 50% tariffs, for fiberglass, its CEO Doug McMillon said on a May earnings call. One apparel brand cut duties by more than 15% — about $2 per unit — by switching hoodie blends from synthetic to cotton, according to cross-border logistics startup Portless. Coca-Cola's CEO James Quincey said in February the company could shift some aluminum-can volume into plastic bottles as tariff costs rose. The common thread: each company changed the product, not the declaration.

How does the first sale rule lower duty costs?

The first sale rule lets an importer declare customs value based on the first sale in the supply chain — the manufacturer's price — instead of the marked-up price charged by a middleman or trading company. For importers who buy through intermediaries, which covers a large share of China sourcing, this is often the fastest duty saving available, and it requires no product change at all.

The arithmetic is straightforward. A logistics trade analysis (PSA BDP) walks through a typical case: a company buys units from a middleman at $9 each, while the manufacturer's original price is $6. Importing 30,000 units at the $9 second-sale price with a 25% Section 301 tariff means $67,500 in duties. Declaring at the $6 first-sale price brings the bill to $44,000 — a $22,500 reduction, or 33%. Applied across $100 million in annual purchases, the same math produces roughly $8.33 million in savings.

Qualifying takes real documentation, though. U.S. Customs and Border Protection permits first sale valuation only when each transaction in the chain is a bona fide, arm's-length sale; the goods were clearly destined for the United States at the time of the first sale; and the importer can produce a complete audit trail — contracts, invoices, payment records, and shipping documents connecting every party. Where related parties sit in the chain, the importer must show the price reflects fair market value rather than the relationship. This is a records exercise as much as a strategy exercise: if your sourcing flows through a trading company and you can document the factory-level sale, there is likely money on the table.

What does unbundling change at the border?

Classification follows the article in its condition as imported. Ship the same product in a different condition — components instead of a finished assembly, or an item and its accessory in separate shipments rather than as one retail set — and a different heading can apply, often at a lower rate.

A familiar illustration: furniture imported fully assembled can classify as finished furniture, while the identical furniture shipped flat-pack classifies as parts or unassembled furniture, frequently at lower rates. The same logic applies to retail sets. When items are packaged together for sale as a set, classification rules look to the component that gives the set its "essential character"; rearranging what ships together — or shipping the accessory separately — can change which heading wins.

There are limits worth respecting. General Rule of Interpretation 2(a) lets customs treat an incomplete or unassembled article as the finished product when it has the essential character of the complete article, so breaking a product into pieces purely on paper does not fool anyone. Unbundling holds up when it reflects genuine manufacturing or logistics logic — real subassemblies made at different factories, products genuinely sold without accessories, flat-pack shipping that actually saves freight cost. As with every method here, the change has to be true before it can be useful.

Where is the line between engineering and evasion?

The distinction is simple to state and expensive to get wrong. Engineering changes the product. Evasion changes the paperwork.

Legal: tariff engineering · Illegal: tariff evasion

• Real design or material change before import: Product unchanged; description falsified
• Feature has commercial use as imported: Feature added only to dodge duty, removed after import
• Declared value supported by real transactions: Undervalued invoices or double books
• Classification disclosed to customs: Misclassification to hide what the goods are
• Documented product-development decisions: No documentation; intent concealed

The Ford Transit Connect case is the cautionary tale. From 2009 to 2013, Ford imported the vans from Turkey with rear seats, seat belts, and glass windows so they would classify as passenger vehicles at a 2.5% duty, instead of the 25% "chicken tax" on cargo vehicles. After arrival in the United States, Ford removed the seats and belts and sometimes replaced the windows with metal panels, selling the vans as cargo vehicles. U.S. Customs challenged the practice; the courts went back and forth for years, and in 2019 the Court of Appeals for the Federal Circuit ruled against Ford. The Supreme Court declined to hear the case in 2020, and in March 2024 Ford agreed to pay $365 million to settle Justice Department allegations that it had misclassified the vans. The seats were real, but they were not a commercial reality: they existed to evade the duty and were removed immediately.

Contrast the Subaru BRAT from the 1970s, which added rear-facing "jump seats" in the pickup bed to qualify as a passenger vehicle — and sold the trucks with those seats in place. That is the textbook legal version: the feature stayed with the product.

Practitioners converge on one piece of advice. As attorney Mario Torrico puts it, "it is good practice to support any product change with thorough documentation of the product development and engineering decisions in the event Customs ever inquires about the modification." Design memos, bills of materials, supplier correspondence, and test reports that explain why the product is the way it is — written before anyone asks — are what separate engineering from evasion when customs comes looking.

How do you work with a customs broker on tariff engineering?

Bring the broker in at the design stage, not at the port. A good broker does four things for a tariff engineering project: confirm the candidate classification against the Harmonized Tariff Schedule, request a binding ruling from Customs when the answer matters, review first-sale documentation chains for gaps, and flag the moves that look like evasion before you commit to them. The binding ruling is the most valuable deliverable: Customs' written commitment to a classification, turning a judgment call into a defensible position.

A broker can only work with what you show them: the bill of materials, design specifications, material composition, invoices at each stage of the chain, and any origin analysis. The more precise the product description, the more precise the classification — vague inputs produce conservative answers, and conservative answers cost duty.

Set expectations honestly. Rulings take weeks, not days. Some products simply have no lower-duty alternative classification, and a broker who tells you that is saving you from a misclassification penalty. Tariff engineering is a percentage game played on your highest-duty SKUs first; running your top five products by annual duty spend past a broker is usually enough to tell you whether the exercise is worth scaling. Importers who already route freight and customs clearance through one partner have a head start, since that partner already holds the shipment data the analysis needs.

Frequently asked questions

Is tariff engineering legal?

Yes, when the product genuinely qualifies for the classification claimed. U.S. courts have affirmed since 1881 that importers may design products to minimize duty liability. It becomes evasion when the goods are misdescribed, undervalued, or temporarily altered — for example, adding a feature that is removed right after import, as in the Ford Transit Connect case that ended in a $365 million settlement.

Does tariff engineering work against Section 301 tariffs on Chinese goods?

It can reduce them, though it does not erase them. Classification-based duties still apply per heading, so moving to a lower-duty heading shrinks the base on which additional duties compound. The first sale rule is especially relevant for China sourcing: Section 301 duties are ad valorem, so declaring the lower factory-level price cuts the 301 bill proportionally.

How much can tariff engineering realistically save?

It depends on the spread between the two candidate classifications. Converse's slipper-vs-sneaker move trades a 20% rate for 6%. First-sale savings on a typical trading-company markup run around 30% of the duty bill in the worked example above. But some products have no meaningful alternative classification. Run the math per SKU before committing.

Can customs reclassify my goods after import?

Yes. Customs can audit entries and reassess classification, which is one reason importers must keep entry records for five years. This is why the documentation habit matters: design memos, bills of materials, and correspondence showing the engineering decision was genuine. A binding ruling from Customs, obtained before you ship in volume, is the strongest protection against a reclassification dispute.

Do I need a customs broker to do tariff engineering?

Not legally, for the design changes themselves — anyone can change a product. But classification calls, binding rulings, and first-sale documentation reviews are broker and customs-counsel territory. The cost of getting a classification wrong (penalties, back duties, interest) usually exceeds the cost of the advice.

A decision rule for your next product

Here is the rule worth taping to the wall: if the duty spread between two plausible classifications covers the cost of the design change plus documentation, do it. If the plan depends on removing the feature after import, do not. Tariff engineering rewards importers who think about classification early, in the factory, with records to prove the product is what they say it is.

Start with your top five SKUs by annual duty spend and ask your broker whether each has a genuine, lower-duty alternative classification. If the answer is yes, the next step is a design conversation with your supplier — materials, packaging, documentation — before the next production run. For sourcing teams working with Chinese factories, factory coordination is where the strategy succeeds or fails, since the engineering has to happen on the production line. Questions about applying any of these methods to your own products: hi@cnally.com.