In short

A Bangladesh garment export order passes from a brand or retailer, often through its own Dhaka office or an agent, to a factory that may subcontract part of it. Published lead times for sea shipments mostly run from about 60 to 134 days, but sources define lead time differently, and none measures how many orders go through intermediaries.

Want a costing for your own style? Send us your enquiry

This chapter is part of a research guide to the Bangladesh garment trade, which has the literature review, the method and the full bibliography.

This chapter follows one garment export order from the first enquiry to the ship. It asks two questions. Who stands between a foreign buyer and the sewing line in Bangladesh, and what does the research say about each of them? And how long does each step take, according to the sources that have tried to measure it?

The short answers are less tidy than the buyer guides suggest. The theory of how such orders are organised is well developed and has been applied to Bangladesh many times. The numbers are weaker. I found three published estimates of how much of the trade passes through intermediaries, from 72% of buyers sourcing directly to 80% of orders going through agents, and none of them is a measurement. The lead-time figures that many reports repeat (90 to 120 days) trace back to a handful of surveys, most of them more than ten years old, which do not define lead time the same way.

How the industry began, including the first use of bonded warehouses and back-to-back letters of credit, is told in CMB’s cited history of the industry, so I do not repeat it here. Prices and order changes are the subject of chapter 2, and trade finance of chapter 3; this chapter touches them only where an order cannot be followed without them.

The theory: buyer-driven chains and value chain governance

Gereffi’s buyer-driven commodity chain

A common starting point for the scholarship on how clothing is bought is Gary Gereffi’s 1994 chapter on “buyer-driven” global commodity chains. Gereffi described a commodity chain by three things: its input-output structure, its spread across territory, and its governance, meaning the power relationships that decide how money, materials and people are allocated along it (Gereffi 1994, 96–97). He then contrasted two kinds of governance. In producer-driven chains (cars, computers, aircraft), large manufacturers control production. In buyer-driven chains, “large retailers, brand-named merchandisers, and trading companies” set up decentralised production networks in exporting countries, and production is carried out by independent factories making finished goods to the buyers’ specifications (Gereffi 1994, 97).

Two details of the 1994 chapter matter for Bangladesh and are often forgotten. First, Gereffi’s diagram of a buyer-driven chain does not show a single link from buyer to factory. Brand-named companies have a direct line to the factories, but retailers reach them through “overseas buyers”, and “traders” stand between both kinds of buyer and the factories. He studied these intermediaries through interviews with managers of overseas buying offices and trading companies (Gereffi 1994, 96, 98). Intermediaries were part of the model from the start. Second, he described “triangle manufacturing”, in which a buyer places an order with a manufacturer it knows in Hong Kong or Taiwan, which shifts some or all of the work to affiliated factories in lower-wage countries. The established manufacturer becomes a middleman. Gereffi expected that as order volumes in the new countries grew, buyers would be pushed to bypass these intermediaries and deal with the factories directly (Gereffi 1994, 114). Whether that has happened in Bangladesh is one of the open questions of this chapter.

Gereffi’s 1999 article, which I could read only in abstract, linked trade networks to upgrading: it sets out how organisational learning happens in trade networks and the usual paths from assembly to original equipment manufacturing (OEM) and own-brand manufacturing (OBM) (Gereffi 1999). A 2003 UNIDO report by Gereffi and Memedovic restates the framework and gives short definitions. Assembly is industrial subcontracting, in which a sewing plant receives imported inputs to assemble. OEM is commercial subcontracting, in which the supplier makes a product to the buyer’s design, sold under the buyer’s brand. The report treats OEM, “specification contracting” and “full-package supply” as broadly the same thing (Gereffi and Memedovic 2003, 1). It also describes the lead firms of buyer-driven chains as “manufacturers without factories” (Gereffi and Memedovic 2003, 3).

Five ways a value chain can be governed

In 2005 Gereffi, Humphrey and Sturgeon replaced the two-way contrast with five types of governance: market, modular, relational, captive and hierarchy. Which type appears depends on three variables: how complex the information passed between buyer and supplier is, whether it can be codified (written down in a specification), and how capable the suppliers are (Gereffi, Humphrey, and Sturgeon 2005, 78, 83–84). In captive chains, small suppliers depend on much larger buyers and face high costs of switching. In relational chains, buyers and suppliers depend on each other and exchange knowledge that is hard to write down (Gereffi, Humphrey, and Sturgeon 2005, 84).

Their apparel case is the one most relevant here. They argue that the industry moved from captive chains, in which the supplier assembles cut fabric to detailed instructions, to relational chains, because full-package production requires the supplier to interpret designs, make samples, source inputs, monitor quality, meet the buyer’s price and deliver on time (Gereffi, Humphrey, and Sturgeon 2005, 91–92). Overseas buying offices and frequent travel supported that exchange. They also note that the quota system “prompted the rise of value-chain intermediaries” such as Li & Fung, and predicted that the end of quotas in 2005 could concentrate production and “undermine the position of intermediary firms” (Gereffi, Humphrey, and Sturgeon 2005, 92).

How the theory has been applied to Bangladesh

Three World Bank and Duke studies apply this framework to Bangladesh directly. Fernandez-Stark, Frederick and Gereffi place Bangladesh as having upgraded from assembly to the full-package or OEM stage. They describe three types of manufacturer: integrated firms that import cotton and do the rest themselves, firms that import yarn, and firms that import fabric and only cut and sew. Most knit factories belong to the first two and most woven factories to the third (Fernandez-Stark, Frederick, and Gereffi 2011, 91, 103). Citing a 2005 World Bank study, they say that two-thirds of apparel firms were in cut, make and trim (CMT) production by 2000, and they link the later growth of domestic knit fabric to the European Union’s two-stage rule of origin (Fernandez-Stark, Frederick, and Gereffi 2011, 104–105).

Staritz’s comparative study of low-income exporters describes the sourcing side. Intermediaries such as importers, exporters, agents and trading houses coordinate production, input sourcing and logistics, and increasingly design and product development. Large buyers have moved towards direct sourcing, with their buying offices in the producing country handling supplier selection, negotiation, monitoring and quality control, while smaller buyers still rely much more on intermediaries (Staritz 2011, 28–30). Lopez-Acevedo and Robertson, writing for the World Bank in 2016, found that buyers choose suppliers mainly on cost, quality, lead time and reliability, full-package services, and compliance. Bangladesh competed on cost but was placed by buyers in the weakest group, with India and Pakistan, on lead time and reliability (Lopez-Acevedo and Robertson 2016, 50, 54–56).

A shorter World Bank paper, catalogued in 2019 but undated, whose Bangladesh section has no figure later than 2013, puts the argument most sharply. It places Bangladesh in a “simple scheduled-response” segment: basic products with few inputs and simple designs, made in large batches with long lead times, “whose production can be managed through intermediaries without the need for direct contact between buyer and manufacturer” (World Bank 2019, 4). In my reading this is close to what Gereffi, Humphrey and Sturgeon would call a modular or even market link for basic goods, sitting alongside the relational links that large factories have with their biggest buyers. The studies above describe both kinds of link in the same country, and I did not find a study that measures how much of Bangladesh’s trade falls under each.

A newer lens: relational and spot buyers

The strongest recent evidence comes from economics rather than sociology. Cajal-Grossi, Macchiavello and Noguera, using customs records from six countries and order-level data from Bangladeshi exporters, distinguish buyers with a “relational” sourcing strategy from those with a “spot” strategy. A buyer’s strategy is consistent across products and countries, which suggests it reflects the buyer’s own organisation, and exporters earn higher markups on otherwise identical orders made for relational buyers (Cajal-Grossi, Macchiavello, and Noguera 2023, 2391, 2410). Their illustration is stark: across the main woven categories, H&M traded with 157 sellers and placed an average of 847 shipments with each, while Kik Textilen traded with 206 sellers and placed an average of 26 with each (Cajal-Grossi, Macchiavello, and Noguera 2023, 2409). For the study of intermediaries the paper has one further finding, which I return to below: in Bangladesh, sourcing intermediaries “rarely take ownership of the good” and so may not appear in customs data at all (Cajal-Grossi, Macchiavello, and Noguera 2023, 2409).

Who takes part in an order

I group the firms the literature names into seven kinds. In practice one firm can play several of these parts, and the same word is used for different things, which is a trap for anyone counting them.

Brands and retailers

The lead firms are retailers (mass merchants and specialty chains, many selling private labels), brand marketers that own a brand but no factories, and brand manufacturers that own both (Fernandez-Stark, Frederick, and Gereffi 2011, 80). They decide what is made, to what specification, at what price and when. How their purchasing practices affect factories is the subject of chapter 2.

Buyers’ own offices in Bangladesh

Many buyers have their own sourcing offices in Dhaka and Chattogram (Chittagong in the older sources). McKinsey’s 2011 survey of chief purchasing officers found that buyers had been strengthening their supply base “by shifting toward direct sourcing and opening their own local offices in Dhaka and Chittagong” (Berg et al. 2011, 6). The same report sets out five “archetypes” of sourcing set-up, from using an agent, through sourcing directly from headquarters, to a small, traditional or full-fledged local office. Each takes on a different share of the work: product and technical design, supplier screening, quotation and negotiation, order follow-up, sample approval, in-line inspection, compliance and shipping (Berg et al. 2011, 21). A buyer’s office in Dhaka is still not always the decision maker. Kader and Akter report that local buying offices often have to send fit, pre-production and wash samples to headquarters by air for approval, which takes 10 to 15 days where a local decision could take 2 to 3 (Kader and Akter 2014, 275).

International sourcing agents and trading houses

The best-known intermediary in the literature is Li & Fung of Hong Kong. Gereffi and Frederick call it the pioneer of the “agent-sourcing model”, which is most popular with buyers that need smaller volumes, or with large buyers that need small quantities of some items. They give agents’ commission as 4–8% of the wholesale price (Gereffi and Frederick 2010, 17). Staritz records Li & Fung becoming the primary purchasing agent for Walmart and for Liz Claiborne (Staritz 2011, 30), and the 2019 World Bank paper names it as the most prominent intermediary working in Bangladesh (World Bank 2019, 7).

Local buying houses

“Buying house” is the Bangladeshi term for a local firm that does an agent’s work for one or more foreign buyers. The NYU Stern study of 2014 describes what such agents do: identify and select suppliers, negotiate the terms of the order and see it delivered to the freight forwarder, and sometimes also market research, raw material sourcing, quality control, compliance, pre-shipment inspection, customs clearance and logistics. They are usually paid a commission on the free on board price, and sometimes a retainer (Labowitz and Baumann-Pauly 2014, 20). The authors counted about 1,000 agents listed in the BGMEA members’ guide, against 4,417 factories (Labowitz and Baumann-Pauly 2014, 20). One small agent they interviewed worked with about 40 factories and chose among them by the buyer’s market segment (Labowitz and Baumann-Pauly 2014, 21). For practical advice on working with one, see CMB’s buying-house guide.

The word is not used consistently. The 2019 World Bank paper says that representatives of retailers from markets such as South Africa “have also set up buying houses in Bangladesh” (World Bank 2019, 3), which is a buyer’s own office under another name, and Abir’s case study describes an order received from “a renowned buying office located in Germany” (Abir 2020, 244). A count of “buying houses” can therefore include independent agents, buyers’ own offices, and offices abroad.

Factories: CMT, full package and vertical knit

The factory may work on a CMT basis, cutting and sewing fabric that the buyer has bought and supplied, or on a full-package (often called FOB) basis, buying and financing its own fabric and trims (Fernandez-Stark, Frederick, and Gereffi 2011, 85–86). The Bangladeshi usage blurs this. Both Fernandez-Stark and co-authors and the 2019 World Bank paper call woven factories that import their own fabric “CMT factories” (Fernandez-Stark, Frederick, and Gereffi 2011, 103; World Bank 2019, 7), although in the strict sense such a factory is buying its own inputs. Knit factories are more often vertical: the 2019 World Bank paper reports that domestic suppliers provide around 90% of knitted fabric and 80% of accessories (World Bank 2019, 7). Knitwear is the subject of chapter 5, and the difference between the price bases is set out for buyers in CMB’s guide to FOB, CMT and full package.

Subcontractors

Behind the factory that holds the order there may be others. Labowitz and Baumann-Pauly called this “indirect sourcing”: “the routine practice of subcontracting, often through purchasing agents and in a manner that is not transparent to buyers or regulators” (Labowitz and Baumann-Pauly 2014, 17). Their interviews describe large groups that take orders beyond their own capacity and subcontract the simpler work, and smaller factories that take subcontracts to fill empty lines (Labowitz and Baumann-Pauly 2014, 18, 25). In a 2015 follow-up they estimated more than 7,000 factories producing for export, against earlier estimates of 4,000–4,500, and found that 32% of 479 factories surveyed in two sub-districts of Dhaka were informal subcontractors (Labowitz and Baumann-Pauly 2015, 4). These figures are disputed, and I return to the dispute below.

Hybrid firms that both make and source

Several strands of the literature describe firms that are both maker and intermediary. Gereffi’s triangle manufacturers are factories turned middlemen (Gereffi 1994, 114; Gereffi and Memedovic 2003, 13–14). Gereffi and Frederick list “full-package ‘manufacturers’ becoming intermediaries”: firms that set up networks of other suppliers and offer “full-package services along with production capabilities” (Gereffi and Frederick 2010, 17). Staritz describes transnational producers whose plants receive orders, inputs and specifications from a headquarters that deals with the buyer, so that buyers call the arrangement “direct sourcing” while the plant itself has no direct link to the buyer (Staritz 2011, 30). Appelbaum, in an article I read only in abstract, argues that the growth of giant East Asian contractors may partly offset the imbalance of power between big buyers and their suppliers (Appelbaum 2008). In Bangladesh the “mother factory” that subcontracts part of an order is, in my reading, a domestic version of the same thing, and Labowitz and Baumann-Pauly note that the same factory can be a mother factory on one order and a subcontractor on the next (Labowitz and Baumann-Pauly 2014, 25). I found no study that counts hybrid firms in Bangladesh or compares their orders with those of pure factories or pure agents.

How much of the trade passes through intermediaries

This is the question a reader most often asks, and the published answers do not agree. The table below lists every estimate I found that I could open and read at the page. They count different things: some count buyers, some count orders, some count firms, and one counts factories by whether they import their own inputs. None of them is drawn from a complete record of orders.

Published estimates of intermediation and indirect sourcing in Bangladesh’s garment export trade. Each figure is as reported by its source; they measure different things and should not be compared directly.
What is countedFigureBasis of the estimateSource
Surveyed European and US buyers with activities in Bangladesh that source directly72%Interview survey of chief purchasing officers, 2011Berg et al. 2011, 6
Surveyed Bangladeshi suppliers that focus on working directly with international buyers69%Telephone survey of more than 100 suppliers, 2011Berg et al. 2011, 3, 6
Orders run through agents80%Estimate by the then BGMEA president, reported by the authors without its basis; they add that it is very difficult to checkLabowitz and Baumann-Pauly 2014, 20
Agents listed in the BGMEA members’ guide, against factories listedAbout 1,000 agents; 4,417 factoriesCount from the members’ guideLabowitz and Baumann-Pauly 2014, 20
Firms in Bangladesh that work with intermediariesAround 85%Interviews and “market intelligence”; the paper warns that breadth was put before rigourWorld Bank 2019, 1, 7
Ratio of intermediary to direct-buyer transactionsAbout 85:15As above, given in the Sri Lanka section for comparisonWorld Bank 2019, 17
Margin lost by a manufacturer selling through an intermediaryAround 7–9%One manufacturer’s estimateWorld Bank 2019, 7
Agents’ commission, worldwide4–8% of the wholesale priceStated without a source; not specific to BangladeshGereffi and Frederick 2010, 17
Factories on the trade associations’ lists of direct exporters (those that import inputs duty free against an export order)3,200Authors’ total of the two associations’ lists, against more than 7,000 factories producing for exportLabowitz and Baumann-Pauly 2015, 4, 20
Informal subcontractors among factories surveyed in two sub-districts of Dhaka32% of 479Field survey, June 2015Labowitz and Baumann-Pauly 2015, 4

Download this table (CSV)

Read side by side, the figures are less contradictory than they first look. McKinsey’s 72% counts buyers, and its sample was weighted towards large ones: the purchasing officers it interviewed accounted for 66% of Bangladesh’s apparel exports to Europe and the United States (Berg et al. 2011, 3). A survey of large buyers, many with their own offices, and a trade in which many smaller buyers use agents could produce both a high share of surveyed buyers sourcing directly and a high share of orders, counted one by one, passing through agents. The two supplier figures are harder to square, because they count the same kind of firm: McKinsey’s 69% of suppliers focusing on direct work and the World Bank’s 85% of firms working with intermediaries can both hold only if many factories use both channels. The World Bank paper does note that some manufacturers use intermediaries in certain segments and not in others (World Bank 2019, 7). But this reconciliation is my reading, not a finding, because no source measures the share of export value or of orders that goes through an intermediary.

The reason no one has measured it is worth stating plainly. Cajal-Grossi, Macchiavello and Noguera note that in Bangladesh sourcing intermediaries rarely take ownership of the goods, so they may not appear in the customs records that economists use (Cajal-Grossi, Macchiavello, and Noguera 2023, 2409). A buying house that arranges an order, follows it and inspects it, but never buys or sells the garments, is unlikely to appear as a party in those records. The gap comes from how the trade is organised, not simply from a lack of effort.

What intermediaries add, and what they cost

The case for intermediaries in the literature rests on information. The 2019 World Bank paper puts Li & Fung’s advantage in its “superior information about and capacity to manage supplier performance in a highly fragmented market”, and records intermediaries taking on compliance monitoring for buyers and advising factories on finance for remediation (World Bank 2019, 7–8). Gereffi and Frederick list scale, buying power, flexibility and the spreading of risk across suppliers (Gereffi and Frederick 2010, 17). Labowitz and Baumann-Pauly add that for small brands an agent is “an almost necessary entrée” into manufacturing (Labowitz and Baumann-Pauly 2014, 20).

The case against is cost and opacity. The World Bank paper reports that established Bangladeshi firms prefer to deal directly with buyers to keep the margin and choose their own suppliers (World Bank 2019, 7). Labowitz and Baumann-Pauly argue that agents give buyers a simpler supply chain at the price of less transparency and control, and let buyers expand or shrink their supplier base at will, which leaves factories without long-term relationships and can lead owners to forgo investment in safety, working conditions and productivity (Labowitz and Baumann-Pauly 2014, 20). One agent described agents as matchmakers between buyers and factories, a role that includes judging whether the buyer wants a “compliant” or a “non-compliant” factory (Labowitz and Baumann-Pauly 2014, 20).

The size of the hidden tier is the most contested number in this literature. Labowitz and Baumann-Pauly’s 2015 count of more than 7,000 export factories was challenged by Anner and Bair, who found closed and duplicate records in the database and estimated that its 7,165 factories were inflated “by at least two thousand” (Anner and Bair 2016, 1). They also pointed out that, on the Stern team’s own average size for unregistered factories, such factories would employ less than 2% of workers producing for export (Anner and Bair 2016, 2). The NYU Stern centre’s later report kept the claim in a softer form: there is “no consensus on the precise number of subcontractors”, which it put in the thousands, “perhaps as many as 3,000” (Barrett, Baumann-Pauly, and Gu 2018, 2), and it gave two reasons for subcontracting: buyers changing orders on tight deadlines, and manufacturers pushing lower-margin work to cheaper factories (Barrett, Baumann-Pauly, and Gu 2018, 13). Both sides agree that subcontracting exists and that buyers often cannot see it. They disagree about how much of the workforce it involves, and that disagreement has not been settled by new data.

The steps of one order, from enquiry to shipment

The order of the steps is broadly agreed. Their names vary from firm to firm, and their length varies with the product, the fabric source, the buyer and the season. What follows is the sequence as the published sources describe it.

  1. Enquiry and tech pack. The buyer, or its office or agent, sends a design and specification, usually called a tech pack. Its accuracy matters for everything after it. Suppliers rating buyers in the Better Buying survey of 2025 reported a decline in tech pack accuracy compared with the previous year (Cascale 2025, 15); that survey covers 1,360 suppliers rating 35 buyers and, its authors warn, is not representative of whole regions (Cascale 2025, 4–5).
  2. Costing and quotation. The factory or agent prices the style. In the 2019 World Bank paper, raw materials (mostly fabric and accessories) make up around 60–70% of a typical apparel firm’s costs and labour around 20% (World Bank 2019, 4). How prices are then negotiated is the subject of chapter 2.
  3. Development samples. The factory makes samples for approval: in Abir’s case study of a knit order these were a lab dip (a dyed swatch to approve the shade), a styling sample, a print strike-off, a size set, a photo sample, a pre-production sample and an original carton sample (Abir 2020, 250). Kader and Akter’s typical chart names a fit sample and a pre-production sample, with a wash step for denim (Kader and Akter 2014, 274).
  4. Order confirmation. The buyer issues the order. In the World Bank’s 2012 schedule for an order of US$500,000, the order arrives 30 days after first contact with the buyer (World Bank 2012, 24).
  5. Payment arrangement. The letter of credit is “the most common mode of payment” for both imports and exports of the industry (World Bank 2012, 23). A factory without the cash to buy fabric opens a back-to-back letter of credit for its inputs against the buyer’s credit, which is the subject of chapter 3.
  6. Fabric and trims booking and in-housing. Fabric is ordered only after the buyer places the order, because a factory cannot afford to stock fabric to every buyer’s specification (Staritz 2011, 121). In Kader and Akter’s typical chart, in-housing the materials is the longest single step (Kader and Akter 2014, 274).
  7. Bulk production. Cutting, sewing, finishing and packing.
  8. Inspection. McKinsey’s archetypes list in-line inspection, sample approval and compliance among the tasks that an agent, a buyer’s office or the buyer’s headquarters may take on (Berg et al. 2011, 21). None of the sources I read gives a separate duration for the final inspection.
  9. Shipment. In 2012 almost all garment exports went by sea on free on board terms, mostly to retail chains, and most export cargo was taken to an off-dock container yard rather than the port, with what the report calls “the nominated forwarder” arranging the move to the vessel (World Bank 2012, 22). Air freight is used, one source notes, mainly when a factory is late (World Bank 2019, 5).
  10. Documents and payment. The factory presents its export documents, with the utilization declaration for its duty-free inputs, for customs clearance (World Bank 2012, 22), and its shipping documents to the bank for payment under the credit (the subject of chapter 3).

How long each step takes

Only a few sources give durations step by step. The table brings them together. The rows are not additive, because many steps overlap: fabric is booked while samples are still being approved, and several sample rounds run side by side (Abir 2020, 249–50; Kader and Akter 2014, 274).

Durations of the steps of a garment export order from Bangladesh, as given by published sources. Days unless stated. The sources measure different orders in different years; where they differ, both figures are shown.
StepDurationWhat was measuredSource
Agree the production design with the buyer3–13 weeksConsultation with exportersWorld Bank 2012, 23
Development and fit sampleApproval days 5–13; fit sample days 13–27Typical 90-day order, authors’ chartKader and Akter 2014, 274
Sample approval through the buyer’s headquarters10–15 (2–3 if approved locally)Authors’ estimate; their method was a survey of 15 factories and interviewsKader and Akter 2014, 273, 275
One sample approval, usual time8 (styling sample); 9 (metal button)One knit order of 73,890 pieces, 2015Abir 2020, 246, 248
Arrange financing3–4Consultation with exportersWorld Bank 2012, 23
Open a back-to-back letter of credit4–6Paper’s table of typical stage times; source not statedWorld Bank 2019, 5
Order inputs, domestic7–10Consultation with exportersWorld Bank 2012, 23
Order inputs, imported14–35Consultation with exportersWorld Bank 2012, 23
Raw material production, then shipping of raw materials15, then 21Paper’s table of typical stage times; source not statedWorld Bank 2019, 5
Woven fabric, by origin7 (Bangladesh); up to 15 (India); up to 30 (China)Report’s statement, 2011; basis not givenBerg et al. 2011, 15
Fabric booking; in-housing of materialsDays 9–34; days 30–69Typical 90-day order, authors’ chartKader and Akter 2014, 274
Fabric sourcing31One knit order, 2015Abir 2020, 250–51
Clear inputs at the port and deliver to the bonded warehouse2–3Consultation with exportersWorld Bank 2012, 23
Clearance at port5Paper’s table of typical stage times; source not statedWorld Bank 2019, 5
Produce garments3–6 weeksConsultation with exportersWorld Bank 2012, 23
Produce garments20Paper’s table of typical stage times; source not statedWorld Bank 2019, 5
Bulk production15–20; days 70–86 in the typical chartAuthors’ survey and chartKader and Akter 2014, 274, 276
Clear export cargo; deliver to port and load3–4; 3Consultation with exportersWorld Bank 2012, 23
Feeder vessel to mother vessel in Singapore7–10Authors’ account of the routeKader and Akter 2014, 277
Transport to the foreign port26–35Consultation with exportersWorld Bank 2012, 23
Shipping to the importer25Paper’s table of typical stage times; source not statedWorld Bank 2019, 5
Clear cargo and deliver to the buyer’s distribution centre3–10Consultation with exportersWorld Bank 2012, 23

Download this table (CSV)

Two things stand out. The production step itself is short: 15 to 20 days in Kader and Akter, 20 in the 2019 World Bank paper, three to six weeks in the 2012 one. Most of the time goes before it, into samples, approvals and above all fabric. Kader and Akter put the in-housing of materials at up to 40 of their 90 days (Kader and Akter 2014, 274–75). Staritz makes the same point from the other side: sourcing fabric locally saves at least the shipping time from East Asia, which she puts at about 30 days on average from mill to factory (Staritz 2011, 121).

Total lead time: what the sources say

Total lead times for garment export orders from Bangladesh in published sources. Days. The definitions of lead time differ between sources, as the third column shows.
EstimateDaysBasis and definitionSource
Woven clothing, 200890–120Gherzi Textile Organization, as cited by Rasiah (2009), as cited by Staritz; definition not givenStaritz 2011, 121
Knit clothing, 200860–80As aboveStaritz 2011, 121
Average lead time for 45% of surveyed suppliersAround 90Supplier survey, September 2011Berg et al. 2011, 16
Average lead time for 25% of surveyed suppliers120As aboveBerg et al. 2011, 16
Order of US$500,000, inputs mostly imported, by sea: first contact to delivery of final shipment134World Bank staff calculations with exporters; order received on day 30, production begins on day 60World Bank 2012, 24
Same order with better supply chain management113As aboveWorld Bank 2012, 24
Order of US$200,000, local inputs, by air50As aboveWorld Bank 2012, 24
Knit composite factories; knit factories; woven factories70–90; 90–110; 120Authors’ questionnaire and survey of 15 factories; the paper defines lead time both from order receipt and from receipt of raw materialsKader and Akter 2014, 268, 270, 273, 275
Early 1990s; after a reduction the authors date to 2007120–150; 90–100Stated without a sourceKader and Akter 2014, 273
Firms interviewed60–120 (as low as 40 for an influential buyer)Interviews; the paper’s stage-by-stage table totals 90World Bank 2019, 5
One knit order of 73,890 pieces: confirmed 5 March 2015, shipped in full 4 September 2015208 as reported; 183 elapsed between the two datesTotal of the periods in the authors’ chart, each from submission to approval; the authors note that activities ran in parallel. The 183 days are my countAbir 2020, 246, 249–51

Download this table (CSV)

Why the lead-time figures do not agree

The first reason is definition. Lopez-Acevedo and Robertson note that country lead-time data are hard to obtain “because the definition often differs”: lead time is usually measured from order placement to receipt by the buyer, but can also mean the time from receipt of raw materials to shipment (Lopez-Acevedo and Robertson 2016, 54–55). In their own country comparison, a note defines it as the time from the arrival of fabric ready to cut to garments packed and ready to ship (Lopez-Acevedo and Robertson 2016, 56, 71). The ILO’s global supplier survey found the same problem in practice: some buyers count production time from order placement, others from sample approval (ILO 2017, 6). Kader and Akter define lead time as order receipt to shipment in their abstract, and from the receipt of raw materials to goods ready to ship two pages later (Kader and Akter 2014, 268, 270); the World Bank’s 2012 schedule runs from first contact with the buyer to delivery of the final shipment (World Bank 2012, 24). These are not the same interval, and a difference of 30 or 40 days between two sources can come entirely from where the clock starts.

The second reason is that the figures travel. The 90–120 days for woven garments appears in Staritz’s 2011 table, which cites Rasiah (2009), who cites the Gherzi Textile Organization (Staritz 2011, 121); I could not open either of the earlier sources. The figure of 120–150 days in the early 1990s appears without a source in Kader and Akter (Kader and Akter 2014, 273), and again in Khan in 2021, citing a 2004 work (Khan 2021). Khan’s sentence on the feeder vessel to Singapore (“It takes 7 to 10 days for the feeder vessel to reach mother vessel, unload and load again”) is word for word the one in Kader and Akter seven years earlier (Kader and Akter 2014, 277; Khan 2021). A reader who meets the same number in five papers should not take it as five measurements.

The third reason is that single cases are not typical. Abir’s order took six months, from confirmation on 5 March 2015 to full shipment on 4 September 2015, because of slow approvals, repeated sample rejections and imported trims, and was 45 days late (Abir 2020, 246, 248–49). The paper itself is not fully consistent: it gives the planned shipment date as 30 July 2015 in one place and 20 July in another (Abir 2020, 246, 249), and its total of 208 days is the sum of the periods in its chart, each measured from submission to approval and many of them overlapping, not the elapsed time, which I count as 183 days from the two dates it gives. Such cases are useful for showing where time is lost. They are not estimates of the average.

Finally, almost none of the stage-by-stage data is recent. The newest step-by-step table I found is in the 2019 World Bank paper, which gives no date or source for it and whose Bangladesh section has no figure later than 2013 (World Bank 2019, 4–5). Nothing I could find measures lead times by step after the COVID-19 pandemic or the political changes of 2024.

Shipping terms: who carries what

Incoterms in brief

The Incoterms rules are published by the International Chamber of Commerce (ICC) and revised about every ten years; the current edition, Incoterms 2020, took effect on 1 January 2020 (UNCITRAL 2020, 3). The ICC’s 2026 work plan includes beginning the revision that will lead to an Incoterms 2030 (ICC 2026, 5). The 2020 edition has eleven rules: seven for any mode of transport (EXW, FCA, CPT, CIP, DAP, DPU and DDP) and four for sea and inland waterway only (FAS, FOB, CFR and CIF) (UNCITRAL 2020, 4). Each rule says where the seller delivers, at what point the risk of loss or damage passes to the buyer, and who pays which costs. Under FOB, for example, the seller’s delivery is complete once the goods are loaded on the ship the buyer has chosen, at the agreed port; from then on, loss or damage is the buyer’s risk and further costs are the buyer’s. Under FCA, delivery happens when the seller hands the goods to the carrier or other party the buyer has named, at an agreed place, and the risk moves at that point (UNCITRAL 2020, 5–6). I paraphrase the ICC’s summary here; the full rules are the ICC’s publication and should be read in the original.

In 2012, the World Bank reported that “almost all garment exports” from Bangladesh were shipped by sea on FOB terms, mainly to retail chains, and that less than 5% went on cost and freight terms, mostly to traders and small importers. Imports of yarn and fabric, by contrast, usually came on cost and freight terms, and land shipments from India usually ex works (World Bank 2012, 22).

One point in that account deserves a legal scholar’s attention. The same report says that most exports were transferred to an off-dock container yard rather than the port, and that “the nominated forwarder” arranged the movement from there to the vessel; it does not say who nominated the forwarder (World Bank 2012, 22). Its own footnote defines FOB loosely, as delivery “to the port of departure” at the seller’s cost and risk (World Bank 2012, 22). Under the FOB rule as the ICC summarises it, the seller carries the risk until the goods are on board; under FCA, risk passes when the seller hands the goods to the carrier or other person the buyer has named, at the agreed place (UNCITRAL 2020, 5–6). In my reading, if the forwarder at the container yard is the buyer’s nominee, the handover described looks more like the FCA rule than the FOB rule, and a contract written on FOB terms would leave the factory carrying risk for goods it no longer holds. That is an inference from two short texts, not a finding: the report does not discuss who carries the risk at the container yard, and I have not read the full rules or any contracts. I did not find a study of which rule Bangladeshi garment contracts actually use since 2012, or of how losses in the container yard have been settled.

“FOB” and “CMT” as ways of pricing an order

In the garment trade, FOB has a second meaning. It names a price basis and a business model as well as a delivery rule. Fernandez-Stark, Frederick and Gereffi use “OEM/Full Package/FOB” for the model in which the manufacturer buys and finances its own materials, and note that “FOB” is a common industry name for this kind of contract manufacturer even though it is technically a delivery term (Fernandez-Stark, Frederick, and Gereffi 2011, 85–86). The NYU Stern report quotes a 2014 government incentive based on the “freight on board value” of exports (Labowitz and Baumann-Pauly 2014, 18), which shows how loosely the letters are used even in an official announcement.

Who carries what follows from the price basis. Under CMT, the buyer buys the fabric and supplies it to the factory with detailed specifications, so the buyer carries the cost and risk of the materials and the factory sells its labour and overhead (Fernandez-Stark, Frederick, and Gereffi 2011, 85). Under full package, the factory must buy and finance fabric and trims, organise their delivery and, in the industry’s sense of FOB, deliver the finished goods at an agreed price (Fernandez-Stark, Frederick, and Gereffi 2011, 85–86). Labowitz and Baumann-Pauly point out that this is why direct orders are hard for small firms: a factory taking a direct order assumes responsibility for everything from buying materials to transport, which needs capital, international relationships and the ability to absorb delays in materials (Labowitz and Baumann-Pauly 2014, 19). CMB’s guide to FOB, CMT and full package sets out the practical differences for a buyer, and its guide to who pays when an order goes wrong covers the commercial side.

Prices, payment and finance, in brief

Three topics run through every order and are treated in their own chapters. The first is purchasing practice. In the ILO’s 2016 survey of 1,454 suppliers in 87 countries, only 17% said their orders had enough lead time, and textile and garment manufacturers were among those with the most inadequate lead times (ILO 2017, 1, 6). Better Buying’s 2025 survey found that 84.6% of suppliers received a forecast of their buyers’ seasonal plans, down from 88.3%, and that suppliers in South Asia were among those receiving forecasts least far in advance (Cascale 2025, 13). Kader and Akter make the link for Bangladesh: without early word of buyers’ plans, factories have no chance to plan their fabric and trims in advance (Kader and Akter 2014, 275). Prices, cancellations and changes are in chapter 2.

The second is finance. Labowitz and Baumann-Pauly were told by a factory owner that “Bank[s] give financing on the basis of sales contracts”, and they conclude that this is one reason the biggest factory groups take the most orders and subcontract much of the sewing (Labowitz and Baumann-Pauly 2014, 18). The mechanics of back-to-back credit and bonded warehouses are in chapter 3, and the terms buyers usually see are in CMB’s guide to payment terms.

The third is measurement. How many factories there are, how much they export and how the official figures compare with importers’ records is the subject of chapter 4.

What the research does not yet tell us

  1. The share of orders that pass through intermediaries. The published figures (72% of buyers direct, 80% of orders through agents, 85% of firms using intermediaries) count different things, rest on interviews or one person’s estimate, and were published between 2011 and 2019. Because local intermediaries rarely own the goods, customs data may not show them (Cajal-Grossi, Macchiavello, and Noguera 2023, 2409). Bank records of who applies for and who benefits from each export letter of credit, or a survey of factories’ order books, might. This matters for any argument about who sets prices and who is responsible for working conditions.
  2. What buying houses add. The only figures for an intermediary’s margin in Bangladesh are one manufacturer’s estimate of 7–9% (World Bank 2019, 7) and a worldwide commission range of 4–8% (Gereffi and Frederick 2010, 17). I found no study that compares orders placed through local buying houses with direct orders on price, lead time, quality or compliance.
  3. Hybrid firms. The theory describes manufacturers that also act as intermediaries, but I found no count of such firms in Bangladesh and no study of how their own production and their sourced orders differ.
  4. Current lead times, step by step. The step-level data come from consultations around 2012, an undated table in a paper catalogued in 2019, and small surveys of 15 factories or single orders. A standard definition (where the clock starts and stops) and a new survey would make the numbers comparable across countries and over time. Inspection and documentation are not timed separately in any source I read.
  5. Which Incoterms rule is used, and what happens when goods are lost before loading. The only account of shipping terms I found is from 2012, and it raises a question about risk in the off-dock container yard that I could not answer from published sources.
  6. How big the subcontracting tier is. The Stern estimates and Anner and Bair’s critique disagree on the number of export factories by thousands and on the share of workers in informal subcontractors (Labowitz and Baumann-Pauly 2015, 4; Anner and Bair 2016, 1–2). Neither side has published a new count since 2018 that I could find.
  7. Relational and spot buyers in practice. Cajal-Grossi, Macchiavello and Noguera show that buyers differ systematically (Cajal-Grossi, Macchiavello, and Noguera 2023, 2391). How a buyer’s strategy shapes the steps of an order (the number of sample rounds, the lead time offered, the use of agents) has not, as far as I can tell, been studied directly.

How I researched this chapter

I carried out this review with the help of AI research tools, which searched for, read and summarised the sources and drafted the text to my brief; a separate AI check then reopened the sources and tested every figure, quotation and citation. I set the questions and the rules, and I am responsible for what is published here.

This chapter is a desk review of published sources and data. I did no interviews, fieldwork or archive visits. I read the sources on 22 November 2026 and cite only what I read on the page cited.

I read in full, or in full for the sections cited: Gereffi and Memedovic (2003); Gereffi, Humphrey and Sturgeon (2005); Gereffi and Frederick (2010); the apparel chapter of Fernandez-Stark, Frederick and Gereffi (2011); the chapters of Staritz (2011) and Lopez-Acevedo and Robertson (2016) on sourcing and lead times; the World Bank reports of 2012 and 2019; the McKinsey report of 2011; the three NYU Stern reports (2014, 2015 and 2018) and Anner and Bair’s critique; Cajal-Grossi, Macchiavello and Noguera (2023); the ILO issue brief (2017); the Better Buying report (2025); the two European Scientific Journal papers; and the United Nations document that reproduces the ICC’s summary of Incoterms 2020. Gereffi’s 1994 chapter is held as a scanned image; I read pages 95–99 and 114–115. For Gereffi (1999) and Appelbaum (2008) I read only the abstracts, and I summarise nothing from them beyond what the abstracts say. Khan (2021) I read on the publisher’s web page, which has no page numbers. The 2019 World Bank paper has no printed page numbers, so I give the PDF page.

What I could not verify: the Gherzi lead-time figures, which I saw only as cited by Staritz from Rasiah (2009); the 2005 World Bank finding that two-thirds of firms were in CMT by 2000, which I saw only as cited by Fernandez-Stark and co-authors; the BGMEA president’s estimate of 80%, which Labowitz and Baumann-Pauly report without saying how it was made; and the full Incoterms 2020 text, which the ICC publishes and which I did not read.

On my own position: I co-founded CMB Sourcing Ltd. The firm makes knitwear and caps in its own factory, which is the factory of our sister company, Collective Studio Ltd., where I am Head of Growth; for other products it works as a buying house, sourcing from partner factories. I have a commercial interest in how intermediaries and hybrid firms are described, which is one reason I have set out the evidence against them as fully as the evidence for them. Nothing in this chapter draws on CMB’s own orders.

Each table on this page can be downloaded as a CSV file under CC BY 4.0. The guide’s method and full bibliography are on the guide’s main page, and other datasets are listed on CMB’s data page.

References

Chicago author-date style. Web pages were read on 22 November 2026 unless the entry says otherwise.

Sources

Checked . Rules and figures change, so confirm anything that affects your pricing.

  1. Gereffi (1994), The Organization of Buyer-Driven Global Commodity Chains (opens in a new tab)
  2. Gereffi, Humphrey and Sturgeon (2005), The Governance of Global Value Chains (opens in a new tab)
  3. Fernandez-Stark, Frederick and Gereffi (2011), The Apparel Global Value Chain (opens in a new tab)
  4. Staritz (2011), Making the Cut? World Bank Study (opens in a new tab)
  5. World Bank (2012), Consolidating and Accelerating Exports in Bangladesh (opens in a new tab)
  6. World Bank (2019), Apparel GVC Analysis: Bangladesh, Sri Lanka and Turkey (opens in a new tab)
  7. Labowitz and Baumann-Pauly (2014), Business as Usual Is Not an Option, NYU Stern (opens in a new tab)
  8. Cajal-Grossi, Macchiavello and Noguera (2023), Quarterly Journal of Economics (opens in a new tab)
  9. Berg et al. (2011), McKinsey survey of buyers and suppliers in Bangladesh (opens in a new tab)
  10. UNCITRAL (2020), A/CN.9/1028, with the ICC summary of Incoterms 2020 (opens in a new tab)

All guides