The Import Gate and K-beauty Counterfeits: How 97,000 Fakes Were Blocked
Korea gates commercial cosmetics imports through a registrant. An importer registers as a responsible distributor under Article 3 of the Cosmetics Act, files a clearance report before customs release, and tests each batch before sale. On 17 September two agencies announced that roughly 97,000 counterfeit units had moved through Coupang and Naver over fourteen months. None of those controls failed. None of them applied.
What did the two agencies actually announce?
The trademark special judicial police of the Intellectual Property Administration and the Central Investigation Division for Harmful Products at the Ministry of Food and Drug Safety referred a distribution broker to prosecutors. The announcement covers a fourteen-month window running from October 2024 to November 2025, during which the broker sold 61 counterfeit product types, roughly 83,000 units, for 4.5 billion won (about $3.25 million). A further 87 types, roughly 14,000 units with a stated market value of 1.3 billion won (about $940,000), were seized unsold at a logistics warehouse in Cheongju, North Chungcheong. All won figures in this article convert at 1,383.3 per US dollar, the Seoul close on 18 September 2026.
The mechanics are plain. The counterfeits were produced by a manufacturer in Shenzhen, brought into Korea as international express cargo and overseas parcels, consolidated at the Cheongju warehouse, listed on Coupang and Naver as authentic, and delivered to buyers by domestic courier. The agencies stated that this was the first case in which their two units ran a joint investigation to map a counterfeit distribution route, after receiving a tip that fakes were circulating on the platforms.
One figure in the announcement does not reconcile. The headline aggregate market value is given as 5.4 billion won, while the two component figures released alongside it — 4.5 billion sold and 1.3 billion seized — sum to 5.8 billion. This article uses the components.
Why does the list of copied brands matter more than the total?
Twenty-eight of the 61 types were cosmetics. Two of the names are the ones a counterfeit table always carries: Estée Lauder and SK-II, prestige products with the price gap that has funded counterfeiting for decades. The other three are not. Celladix, Kahi and Beauty of Joseon are Korean brands, and none of them sells at a prestige price.
That combination is the substantive finding. Copying a prestige serum is an arbitrage on margin — the counterfeiter captures the gap between a cheap formula and an expensive label. Copying a mid-priced Korean essence captures almost none of that gap. What it captures instead is recognition and velocity: a label that converts at scale on an open marketplace, in a category where consumers buy on reputation and repurchase quickly. Once counterfeiting economics run on turnover rather than unit margin, the protected zone is no longer defined by price.
The same announcement covered counterfeit health functional foods, including copies of a Denps probiotic line and a Korea Eundan multivitamin, and counterfeit Brita filter cartridges whose stated country of origin had been changed from China to Germany. The relabeling on the filter cartridges is the same operation as the label on the cosmetics, performed on a different product category.
Beauty of Joseon is an export brand. Counterfeits of it were manufactured in Shenzhen and sold to Korean consumers inside Korea. The counterfeiting of K-beauty has stopped being a problem that happens to Korean brands in foreign markets.
What did the laboratory find inside the products?
Twelve of the seized functional cosmetics and health functional foods were analysed. In the cosmetics, the functional ingredients tied to the claims printed on the packaging were not detected. In the supplements, the marker substances used to verify the declared ingredients were not detected either. The agencies stated that the products carry no assurance of the claimed function and none of safety.
A functional cosmetic, under Article 2 of the Cosmetics Act, is one that makes one of eleven defined claims — whitening, wrinkle improvement, ultraviolet protection, hair loss relief, acne care and others. Article 4 requires that such a product pass a review or, for formulations matching an already approved specification, a report, supported by efficacy data and human application data. The claim is the regulated object.
In an ordinary consumer good, a missing ingredient is a quality defect. In a functional cosmetic it is the entire product. The state did not approve a jar; it approved a sentence, on the condition that a specific molecule sat behind it. The laboratory result says the sentence shipped without the molecule.
The physical differences investigators recorded were narrow: packaging material, printed content, container shape and size, and the formulation's texture and colour. Those are the tells a buyer would have to catch on a product photograph, before purchase, against a genuine reference they do not have in hand.
How did 97,000 units reach a market with a licensed import gate?
A declared commercial cosmetics import in Korea runs through a fixed sequence. The importer must hold a responsible distributor registration — the licence that makes a company answerable for the quality and safety of what it puts into the market. Before customs release it files a standard clearance report through an approved electronic provider, attaching the registration certificate, a manufacturing certificate and a free-sale certificate from the origin market. After release, each production batch is quality tested, and only batches that pass are permitted to be sold.
The broker entered none of that sequence. International express cargo and the overseas parcel channel are built around consignments destined for personal use, cleared at item level and at low value, on the assumption that the recipient is the end consumer. Korea does exclude functional cosmetics from list clearance, the simplified regime that releases low-value parcels on a manifest alone, which pushes them to a normal import declaration. That exclusion is the closest thing to a product-level check on the parcel path, and it is a clearance formality, not a quality one.
This is the structural point. Every control in Korea's cosmetics regime attaches to a registered entity. Registration, clearance reporting and batch testing all presuppose that somebody has put their name on the goods. A parcel has a recipient, not a registrant. A marketplace listing has a seller account, not a registrant. The chain ran for fourteen months in the space between those two facts.
The penalty structure records the same asymmetry. Operating as a cosmetics responsible distributor without registration carries up to three years' imprisonment or a fine of up to 30 million won under Article 36 of the Cosmetics Act. Trademark infringement carries up to seven years or up to 100 million won under Article 230 of the Trademark Act. The heavier sanction attaches to the borrowed name, not to the absent ingredient.
What does a pre-market approval control once the label is printed offshore?
Korea's functional cosmetics system is a pre-market claim approval regime, and a strict one. A dossier goes in, efficacy and human application data are assessed, and a right to make a specific claim comes out. The regime is well suited to the market it was designed for, in which the party making the claim is a registered domestic entity that can be inspected, audited and deregistered.
Printing is the weak link, and printing has been offshore for years. Once packaging carrying an approved claim can be reproduced by any printer with a photograph of the original, the approval controls a document rather than a product. The enforcement surface of the entire regime is the registrant, and the counterfeit supply chain is constructed precisely to have no registrant.
That is why this case surfaced through a tip and a trademark investigation rather than through cosmetics surveillance. Post-market inspection samples products at registered sellers. Batch testing tests batches that were declared. Neither instrument looks at 83,000 units that were never declared to exist.
How large is this against what customs stops at the border?
The Korea Customs Service detected 117,005 counterfeit items bearing Korean brands at the clearance stage in calendar 2025, announced in January 2026. Cosmetics were the largest single category at 36%, ahead of toys and stationery at 33%. Goods shipped from China accounted for 97.7% of the total.
A special enforcement drive run from 4 May to 30 June 2026 gives the second lens. Clearance-stage seizures in that window reached 452,927 items, against roughly 177,000 in the same period of 2025. Foreign brands accounted for 448,183 of those and Korean brands for 4,744. By entry channel, general cargo carried 315,216 items, express and postal consignments 127,468, and traveller baggage 10,243. The investigative arm of the same drive booked 14 cases covering 3.65 million items with a stated market value of 153 billion won (about $111 million). Across all of 2025, the value of intellectual property infringing goods seized reached 278.9 billion won (about $202 million), up 64% from 170.5 billion won the year before.
Set against those totals, 97,000 units is rounding error. That is the reading that matters. The border apparatus is calibrated to intercept consolidated volume, and it does — millions of units, hundreds of billions of won. This shipment never presented as volume. It arrived in parcels sized to look like somebody's skincare order, and it took a tip and fourteen months to assemble into a case.
Where does liability sit when the sale happens on a marketplace?
The listings ran through open-marketplace seller accounts on Coupang and Naver. In that arrangement the platform is an intermediary — it hosts the listing and processes the transaction, and the seller of record is the merchant account. The consumer's contract is with that account. The brand owner's remedy is a trademark action against whoever controls it. The product regulator's remedy requires a registrant, and there is none.
The result is that the party holding the most complete transaction record — listing history, shipment volumes, return rates, buyer complaints, the repeat SKUs moving out of a single Cheongju address — carries the lightest product-law exposure of anyone in the chain. Eighty-three thousand units of one merchant's counterfeit inventory is a pattern visible in platform data long before it is visible to a customs scanner or a laboratory.
None of this is an argument that the platforms broke a rule. It is an observation about where the information sits relative to where the obligations sit, and the two are in different places.
What does the case change for the industry?
For manufacturers and contract developers, the copy target has moved down the price ladder. A formulation's commercial defensibility has rested on the assumption that mid-priced products are not worth faking. Twenty-eight cosmetics types including three mid-market Korean brands is evidence against that assumption. Formulation complexity is no longer a moat when the counterfeit contains no formulation at all.
For brands and buyers, the exposure has changed direction. A brand that scaled through many channels at once now carries a counterfeit problem in its domestic market as well as its export markets, and authentication at the point of sale is brand-side work. No other party in the chain is positioned to perform it, and no other party bears the reputational loss when a consumer applies an unregulated product from Shenzhen believing it to be a Korean one.
For regulators and investors, the finding is about architecture rather than effort. Korea's cosmetics regime governs entities, and it governs them competently. The parcel channel and the marketplace listing are transactions that produce no entity. Enforcement capacity added to the existing regime does not reach them, because the regime has nothing to attach to. Closing the gap means putting an obligation on a transaction rather than on a licence, which is a legislative decision and not an inspection schedule.
The counterfeits did not defeat Korea's cosmetics controls. They travelled past the point where those controls stand.
Comments