HS 4202.92 vs 4202.32: How to Classify Canvas, Leather and Synthetic Bags

These two lines are close enough in the schedule that they get confused, and far enough apart in practice that the confusion costs money. The difference is not canvas against leather against synthetic — all three can appear in either line. The difference is what the article is, decided first, and then what the surface is, decided second.

A canvas coin pouch and a canvas tote are made of the same cloth and sit in different subheadings, because one is carried inside a bag and the other is the bag. That is the whole test, and it is worth stating plainly at the beginning because every other argument about these lines comes back to it.

What Does "Normally Carried in the Pocket or Handbag" Actually Mean?

It is a test about the article, not about how one particular customer happens to use it. The phrase has been read for decades in rulings and it points at small personal articles that accompany a person inside a bag or a pocket: wallets, purses, card holders, key cases, spectacle cases, cigarette cases and tobacco pouches. Size is evidence, not the rule, and the shape and the function carry more weight than the dimensions.

Presentation is evidence here, and it is evidence the importer controls. A pouch shipped with a hanging loop and a barcode card reads as a retail article with its own identity; the same pouch packed six to a polybag with a care label that says "organiser" reads as an accessory to a bag. Neither reading is decisive on its own, but they are the kind of detail an officer reaches for when the product sits on the boundary, and they cost nothing to get right at the specification stage.

The failure mode is the in-between product. A flat zip pouch made of canvas can be sold as a pencil case, a cosmetics pouch, a cable organiser or a wallet, and the classification follows what it is presented as and how it is made rather than what the catalogue calls it. Where a product is genuinely built to sit inside another bag, 4202.32 is a defensible line; where it is built to be carried on its own with a strap or a handle, it is not.

Where Does the Outer Surface Come In?

Second, and only second. Both 4202.32 and 4202.92 carry the digit 2 because both are the plastics-or-textile surface branch of their group. A leather version of the same pouch is 4202.31, and a leather version of the same travelling bag is 4202.91. The material never moves an article between 4202.32 and 4202.92; it only moves it around inside its own group.

It is also worth separating the surface from the structure. A bag can have a textile face bonded to a foam core and a foil lining, and only the face decides the digit. Buyers who write "polyester and foam" on a specification leave the question open, because the foam and the foil are not surfaces in the sense the heading uses. State the face material, then state what is behind it, in that order, and the classification stops being arguable.

That ordering matters because buyers tend to classify by what they can see, which is the fabric. Working in the right order — article, then surface — removes most of the ambiguity, and it also produces a record that a customs officer can follow, because it follows the structure of the heading.

What Does the Choice Cost on a Real Order?

The two lines carry different national rates, and the gap between them is what makes the argument worth having. Take a 500-piece order at FOB Xiamen USD 3.60 entered in the United States, so the transaction value is USD 1,800.00 and the duty base is the same on either line. Put an illustrative 8% on one line and an illustrative 17.6% on the other, and add a Section 301 rate of 7.5% in both cases, because that layer follows origin rather than the subheading.

At 8% the duty is USD 144.00; at 17.6% it is USD 316.80. The difference is USD 172.80 on the order, or USD 0.35 per unit. Everything else is identical: MPF falls to its floor of USD 32.71 on both, HMF is USD 2.25 on both, and freight, insurance, entry and delivery do not move at all.

How Do Leather and Synthetic Shells Change the Reading?

They change the final digit and nothing else. Full-grain, corrected-grain and split leather, and composition leather, take the digit 1. Sheeting of plastics and textile materials take the digit 2. Anything else — metal, wood, a rigid moulded shell, bare cellular rubber — takes the digit 9, and may take the article out of the heading altogether if it fails the materials condition.

Synthetic is the word that causes trouble, because it can mean a plastic sheeting or a man-made textile, and both land on the digit 2 anyway. Where it matters is further down, at the national level, where the break is by fibre and by whether the surface is coated. Write the construction, not the category: a PU-coated 600D polyester shell is a description a broker can work from, and "synthetic" is not.

When Does an Article Fall Out of Both Lines?

When the exterior is not leather, not plastics sheeting and not textile. A bare neoprene shell, a moulded EVA case and an aluminium flight case all fail the surface condition on their face, and each has to be tested against its own chapter before any 4202 line is claimed. This is the boundary that produces the most rulings, because the products look like bags and are not described by the heading materials.

The second exit is function. An article that exists to protect a specific piece of equipment can be argued into a heading for that equipment, and an article that is essentially packaging can be argued into a packaging heading. Both arguments exist, both have published rulings behind them, and neither can be settled by looking at the product photograph.

The practical consequence of both exits is that a buyer should not assume a 4202 line just because the article is soft and has a zip. Softness is not a test, and neither is having a zip. The tests are the list in the heading and the materials condition, and both are answered from a construction statement rather than from a photograph of the finished product.

How Should the Decision Be Documented?

In the tech pack, in one paragraph, written before the sample is made. Record the article description in the words of the heading, the surface construction, the subheading that follows, and the date the nomenclature was read. Where the product sits near a boundary, record the competing line too and why it was rejected.

The paragraph should also carry the sample reference. A classification tied to a photographed, numbered sample is a classification that can be defended when the production run is questioned; one tied to a description alone depends on somebody remembering what the first article looked like. Where the programme repeats, keep the sample through the life of the style.

That last sentence is the one people skip, and it is the one that matters. A classification file that only records the winning line looks like a choice. A file that records both candidates and the reasoning between them looks like an analysis, and an analysis is what an officer accepts when the entry is reviewed eighteen months later.

What Does This Mean for a Mixed Catalogue?

It means one product family can span both subheadings, and the split has to be managed line by line. A programme that runs insulated delivery bags, bottle carriers and drawstring bags is all 4202.92. The moment it adds a flat organiser pouch sold to live inside a tote, part of it becomes 4202.32, and that part needs its own entry line, its own rate and its own evidence.

The administrative answer is a classification table in the master spec, one row per style, maintained in the same review as the tech pack. It costs an hour a quarter and it removes the scramble that happens when a broker asks for the code on a style nobody has classified. The HS code finder keeps the same table in the format the entry needs.

How Does This Feed the Landed Cost Model?

A mixed catalogue cannot use one blended duty rate. The rate has to be attached per style, because the mix changes the weighted average every time the order composition changes, and a blended figure hides the one style that is carrying a rate three points above the rest.

Build the model per style and aggregate at the end. Each row carries units, FOB unit price, the transaction value, the rate for that line, the fee layer and the freight allocation, and the total is divided by the total units. That is the structure behind the landed cost calculator, and it is also the structure an auditor can follow backwards from the entry to the purchase order.

There is a further reason to keep that table per style rather than per programme: a style can move. A pouch that grows a wrist strap, or a tote that shrinks into an accessory sized to live inside another bag, can cross the boundary between the groups, and the only way to notice is to have the old line written down beside the new construction and compare them on the day the change is made.

Frequently Asked Questions

People Also Ask