In January and February 2026, China's apparel exports to the US fell 57.65% year-on-year - from $2.77 billion to $1.17 billion. In the same window, Bangladesh overtook China in US apparel exports for the first time ever. The switch took less than 90 days. This is not a gradual reallocation playing out over a planning cycle - it is one of the fastest visible shifts in the apparel sourcing map in recent years, and most brands are still reacting to it rather than planning around it.
The January-February data reflects the tariff regime then in force, when China's effective apparel tariff into the US ran roughly 34% (MFN 16.5% + Section 301 7.5% + Section 122 10%) against roughly 10% for Vietnam and near-zero for Bangladesh under Everything But Arms. The US tariff regime has since changed again: Section 122 expired on July 24, 2026 and was replaced by new Section 301 measures, so the exact landed-duty advantage today needs to be calculated against the current HTS classification rather than the January-February stack. What hasn't changed is the direction: that gap - not a change in China's manufacturing capability - is what moved the volume, and it went three places: Bangladesh and Vietnam for basics and volume wovens, India for compliance-heavy and certified categories, and Mexico for brands that can use USMCA's 0% duty on qualifying garments. The mistake brands are making is treating this as "which country has the cheapest tariff this quarter" instead of building a sourcing base that survives the next policy change too - and there will be a next one.
The numbers behind the shift
The tariff stack that drove the January-February shift compounded three separate policy layers: the standard MFN duty (typically 16.5% on most apparel HS codes), Section 301 tariffs (7.5% on the relevant categories), and the Section 122 tariff that was in force at the time (10%). Combined, brands were paying total duties in the 41.5-61% range depending on category, against roughly 10% for Vietnam-origin product and effectively 0% for Bangladesh under the Everything But Arms scheme.
That regime has already moved on. Section 122 expired July 24, 2026 and was replaced by new Section 301 measures - China is now subject to a 12.5% Section 301 tariff under the new structure, with the actual landed-duty burden on any given shipment depending on the specific HTS line and how the current layers interact. We're not going to pretend tariffs stand still long enough to quote one number as permanent - if you're modelling a specific programme, verify the current stack against your product's HTS classification rather than reusing this or any other article's snapshot.
What the January-February gap illustrates regardless of the exact current number: even a large, temporary tariff differential is enough to move volume within 90 days once a secondary sourcing base already exists. On a $10 FOB jacket, a 30-percentage-point tariff gap is $3.00 per unit before freight, financing, or anything else is even considered - at scale, that reorders sourcing decisions faster than any factory-capability argument could, whatever the exact stack is this quarter.
Where the volume actually went
Bangladesh and Vietnam picked up the basics and volume wovens. Bangladesh's near-zero EU and increasingly favourable US tariff position, combined with CMT rates that already undercut most competitors, made it the fastest destination for price-sensitive categories. Vietnam reclaimed the #1 export position from Bangladesh in 2025 ($39.64B vs $38.82B) and grew a further $22.2B in H1 2026, helped by its own FTA network (EVFTA, UKVFTA, CPTPP) stacking on top of the China tariff gap.
India picked up compliance-heavy and certified categories. Brands that need GOTS, OEKO-TEX, or SA8000-certified supply chains - or that are already managing EUDR documentation requirements - found India's deeper certification infrastructure a more natural landing spot than a pure cost-driven move to Bangladesh or Vietnam would have been.
Mexico picked up a smaller but real share of brands that could restructure their supply chain around USMCA's 0% duty on qualifying garments - mainly for shorter-run, faster-turnaround programmes where nearshoring's lead-time advantage matters as much as the duty saving.
Why Bangladesh moved faster than anyone expected
A 90-day shift of this scale should not be operationally possible - factory qualification, compliance audits, and sample approval normally take longer than that on their own. What made it possible was that Bangladesh's factory base was already integrated into many of these brands' supply chains as a secondary source, running smaller allocations alongside China. The tariff shock didn't require building new relationships from zero; it required reallocating volume within relationships that already existed. Brands with a single-country China dependency and no secondary base moved far slower, and are still moving.
That is the operational lesson underneath the trade-policy story: diversification isn't a hedge you build after the shock. It only works if it is already in place when the shock happens.
The risk in chasing this quarter's cheapest tariff
Bangladesh's near-zero EU tariff position is not guaranteed to hold. Bangladesh's Least Developed Country status graduation puts its Everything But Arms access at risk - EU RMG exports could decline more than 43% without a new preferential trade agreement to replace it. A brand that moved its entire basics programme to Bangladesh purely on today's tariff math, with no diversification, is one trade-policy decision away from facing this exact same disruption again, from the opposite direction.
The same caution applies to any single-country reallocation, including a move that looks smart today. Trade policy has moved faster than sourcing programmes can rebuild three times in the past decade. A programme built around one country's current tariff advantage is a programme built to be disrupted again.
What this actually means for how you source
The practical response isn't "find the new cheapest country" - it's building a programme that doesn't have to be rebuilt every time policy moves. That means running production across multiple countries deliberately, not accidentally: basics and volume wovens where CMT and duty are the primary variable, certified and compliance-heavy categories where certification depth matters more than tariff rate, and enough operational flexibility to shift allocation between them without renegotiating from zero each time. See our full India vs Bangladesh vs Vietnam breakdown for how we think about splitting a programme by category rather than picking one country and hoping the policy environment holds still.
If your current programme is concentrated in one country because of this year's tariff numbers, talk to our sourcing team about what a diversified base across India, Bangladesh, and Vietnam actually looks like operationally - not just on a spreadsheet.
FAQ
Is China still viable for apparel sourcing in 2026?
For specific categories, yes - China retains capability advantages in complex construction and fast sampling that other markets haven't fully replicated. But with an effective tariff well above competing markets - the exact stack has shifted more than once in 2026, so check the current HTS classification rather than assuming last quarter's number - it's no longer the default for price-sensitive or high-volume categories, and most brands are now using it selectively rather than as their primary base.
Will Bangladesh's tariff advantage last?
Not guaranteed. Bangladesh's Least Developed Country graduation puts its Everything But Arms EU access at risk, with estimates suggesting EU RMG exports could decline over 43% without a replacement preferential trade agreement. Any sourcing decision based purely on today's Bangladesh tariff position should account for that risk.
Should I move my entire programme to whichever country has the lowest tariff right now?
No - that's the mistake this shift exposed in single-country China programmes. A sourcing base built around one country's current tariff advantage is exposed to the next policy change the same way. Diversifying across 2-3 countries by category is more resilient than chasing this quarter's cheapest option.
How quickly can a brand actually shift production out of China?
Bangladesh's 90-day shift was possible because many brands already had a secondary Bangladesh allocation running alongside China - the tariff shock reallocated existing relationships rather than building new ones from zero. A brand with no existing secondary-country relationship should expect a much longer runway, typically several months for factory qualification and first production.
Is Mexico a realistic alternative to Asian sourcing?
For specific programmes, yes - USMCA's 0% duty on qualifying garments and much shorter lead times make it attractive for fast-turnaround or shorter-run product. It's not a like-for-like replacement for Asian sourcing's category breadth and cost structure at scale, but it's a genuine option worth evaluating for the right product mix.