A shop floor in February is a record of decisions made the previous spring. The buyer chose a shape, a fabric weight, and a quantity somewhere between eight and twelve months before you saw the rail, a mill was booked, a factory took the order, and a container crossed an ocean. By the time a trend is visible in a window it has already been paid for twice over, and none of that spending shows up in the coverage that told you the trend was coming.
Some of it does show up in public records, though, and for free. Every garment entering the United States is declared at the border with a product code, a country of origin, a quantity, and a value. The government publishes the aggregate monthly. It is the least glamorous document in fashion and one of the few that reports what companies actually committed money to, rather than what anyone predicted.
Every shipment arrives with a number attached
The office that assembles this for apparel is OTEXA, the Office of Textiles and Apparel, inside the International Trade Administration at the Department of Commerce. It republishes Census Bureau import figures for textiles, apparel, footwear, leather, and travel goods, broken out by country, by three-digit product category, and down to individual tariff lines. The monthly import tool reports each selection three ways: in dollars, in units, and in square metre equivalents.
Nothing about it requires a subscription, a login, or a trade association membership, which separates it from almost every other dataset in this industry. The interactive reports let you stack several categories or several countries in one view and export the result. The static reports are quicker if you know the one category you care about.
Why the three-digit categories outlived their quotas
The odd part is the numbering. Apparel gets sorted into categories like 338 and 339, men's and boys' and then women's and girls' cotton knit shirts and blouses, or 347 and 348, the cotton trouser and shorts pair. Those numbers are older than most of the brands using them, and they exist because of a quota system that no longer operates.
Under the Multifibre Arrangement, importing countries could cap quantities of specific garment types from specific suppliers, so somebody had to define what counted as a cotton knit shirt. The WTO Agreement on Textiles and Clothing took effect on 1 January 1995 with a ten-year schedule to dismantle that machinery, and it required the remaining quotas to be gone by 1 January 2005, when the agreement itself terminated. The quotas went. The category numbers stayed, because a reporting vocabulary that everyone already knew was too convenient to throw away.
Square metres beat dollars for reading volume
The three units answer three different questions, and mixing them up is how people misread the file.
| Unit | What it answers | Where it misleads |
|---|---|---|
| Dollars | How much money moved | Rises when prices or freight rise, with no extra garments |
| Units | How many pieces landed | Counts a vest and a coat as one each |
| Square metre equivalents | How much cloth landed | Says nothing about garment type inside a category |
Dollars are the number the trade press quotes and the weakest of the three for reading a trend, because a rise can be entirely price. Units count pieces without weighting them, so a season of small garments inflates the count. Square metre equivalents convert everything to fabric area, which is the closest available proxy for how much of a category is genuinely being made. If a wide trouser is replacing a narrow one, area moves and unit counts barely budge.
One category, three seasons, one conclusion
The method is unremarkable and takes ten minutes. Pick the category that matches the garment you are curious about. Pull thirty-six months of it. Then compare each month against the same month a year earlier, never against the month before, because apparel imports are violently seasonal and a January to February comparison tells you about the calendar rather than about demand.
Read the square metre column first for direction, then the unit column to see whether piece counts moved with it, then dollars last to see what happened to price. Three consecutive years of the same month rising in area is a category being fed. One spike is noise.
Country columns add a second layer. A category holding steady in total while its supplying countries reshuffle is a sourcing story, not a fashion one, and those two get conflated constantly in trade coverage. Between 2024 and 2025, US apparel imports shifted heavily from China toward Vietnam under changed tariff conditions, which changed almost nothing about what garments arrived and a great deal about where they were sewn.
What a category cannot tell you
This data has one large blind spot, and pretending otherwise leads people to overclaim. A category has no colour, no silhouette, no hemline, and no brand. Category 348 counts a cotton trouser whether it is a barrel leg or a cigarette cut, and a customs code has no opinion on which one is having a moment. What the file confirms is that a category is being supplied at a certain volume, not which shape inside it will fill the rail.
Timing has limits too. These are landed goods, so the record lags the buying decision by many months and leads the shop floor by weeks rather than seasons. It is early against the window display and late against the order book. Tariff changes distort it further, since importers front-load shipments ahead of a duty increase and a resulting jump reflects a customs calendar rather than any appetite for the clothes.
A ten-minute habit each month
Choose one category tied to something you buy repeatedly and follow only that. Cotton trousers, cotton knit tops, or wool outerwear are the useful ones for most wardrobes, because they cover the garments where a genuine volume shift changes what is available in your size at a price you would pay.
Write down the square metre figure for the current month and the same month last year, and keep the note in the same place each time. Two years of that gives you a personal series nobody is selling you, and it takes less effort than reading a forecast.
Then use it the way it deserves to be used, which is as a check on hype rather than a shopping list. A category with three years of falling volume is a category where the good version is getting harder to find and worth buying when you see it. A category climbing steeply is one where next spring brings more choice, lower prices, and heavier discounting, which is an argument for waiting. Neither reading tells you what will look good on you, and that part stays where it always was.