One Category: The Most Manufactures-Weighted Export Basket of the Six

Executive Summary. Manufactures were 94.45% of Bangladesh's merchandise exports in 2018, the highest of six Asian economies and 1.28 times the peer median of 73.95%. That is usually read as industrial success. The composition inside it argues otherwise. High-technology goods were 0.54% of manufactured exports in the same year, the lowest of the six against a peer median of 9.04%, and ICT goods were 0.05% of total goods exports in 2015, also the lowest, about 1.5% of the peer median of 3.43%. Vietnam on those same dates ran 40.75% and 29.01%. The volume is small as well as narrow: merchandise exports were US$47.2 billion in 2024, fifth of the six, 15.7% of the peer median of US$300.7 billion, with Vietnam selling 8.6 times as much abroad. The second escape route is closed too. Trade in services was 5.07% of GDP in 2025, the lowest of the five countries reporting. Bangladesh has one export category doing nearly all the work, and these series show no second one being built.

The share that has no peer in the group

Start with the composition, not the total.

In 2018, manufactures were 94.45% of Bangladesh's merchandise exports, the highest of the six and 20.5 percentage points above the peer median of 73.95%. The rest of the group: Vietnam 83.19%, Thailand 74.86%, Pakistan 73.95%, India 69.27%, Indonesia 43.12%. Bangladesh is not marginally more industrial in its export mix. It sits at 1.28 times the median comparator, and Vietnam at 83.19% is as close as anyone in the group gets.

Source: World Bank World Development Indicators, manufactures exports as a share of merchandise exports (TX.VAL.MANF.ZS.UN), 2018.

One qualification belongs immediately beside that number, because it cuts both ways. A high manufactures share is partly a statement about what a country does not have to sell. Indonesia's 43.12% is low because a large part of its merchandise exports are primary commodities, not because its factories are weak. Bangladesh has no oil, no significant mineral export, and no large agricultural surplus moving across the border, so nearly everything that leaves the country as goods is manufactured by default. The 94.45% is real, and it is as much the absence of alternatives as the presence of industry. That reading makes the number worse, not better: a basket with no primary-commodity leg has nothing to fall back on when manufacturing demand turns.

What is inside the ninety-four percent

A manufactures share tells you what an export basket is not. It does not tell you what it is, and the second question is where this group separates.

High-technology exports were 0.54% of Bangladesh's manufactured exports in 2018, the lowest of the six. Vietnam 40.75%, Thailand 23.7%, India 9.04%, Indonesia 8.21%, Pakistan 2.17%. Against the peer median of 9.04%, Bangladesh runs about 6% of the median comparator's technology share, a gap of 8.5 percentage points.

The ICT series, three years earlier, says the same thing more bluntly. ICT goods were 0.05% of total goods exports in 2015, again the lowest of the six, against a peer median of 3.43% and Vietnam's 29.01%. In ratio terms Bangladesh's ICT share was about 1.5% of the median comparator's. Thailand was at 16.45%, Indonesia 3.43%, India 0.83%, Pakistan 0.24%.

The mechanism that matters here is not the level of technology for its own sake. It is that an export sector concentrated in one category and flat in technological composition earns its foreign exchange from a single demand cycle, a single buyer structure, and a single set of tariff schedules. There is no second product line whose orders rise when the first one's fall. Diversification shows up in composition before it shows up in totals: a new line large enough to move the technology share is still far too small to move a US$47.2 billion export account. On both composition indicators Bangladesh is at the bottom of the group.

One confounder belongs here before a critic supplies it. Both series classify gross exports by product list, and a product list credits the full value of an assembled device to whoever assembled it. Vietnam's 40.75% high-technology share therefore measures where final assembly happens at least as much as domestic technological capability. Bangladesh's 0.54% is a genuine gap, but its size in value-added terms is smaller than the gross figures imply, and this data has no value-added trade series to settle by how much.

Narrow is survivable at scale. This is not scale.

A concentrated basket is a manageable risk for an economy large enough in world markets to set terms, hold inventory, or absorb a bad year out of reserves. Bangladesh is not in that position.

Merchandise exports were US$47.2 billion in 2024, fifth of the six. India US$442.9 billion, Vietnam US$404.8 billion, Thailand US$300.7 billion, Indonesia US$264.7 billion, Pakistan US$32.3 billion. Only Pakistan sells less abroad. Measured against the peer median of US$300.7 billion, Bangladesh exports 15.7% of the median comparator's total, a shortfall of about US$253 billion. Vietnam alone sells 8.6 times as much.

Source: World Bank WDI, merchandise exports (current US$, TX.VAL.MRCH.CD.WT), 2024.

Put that beside the size of the country it has to support. Bangladesh had 173.6 million people and a US$450 billion economy in 2024, growing at 4.22%. An export account of US$47.2 billion is what funds imported energy, imported capital goods, and imported inputs for the export sector itself. Concentration risk and small scale compound: the same shock that would cost a diversified exporter one product line costs this basket most of its merchandise earnings, and there is no cushion of size underneath it.

Thirty-five years of growing without opening

The trajectory is not stagnation, and nothing above should be read as implying it.

Merchandise exports rose from US$1.67 billion in 1990 to US$47.2 billion in 2024, a gain of about US$45.6 billion, or 2,727%, taking exports to 28.27 times their 1990 level. Very few economies have multiplied their export earnings by that factor in a single working lifetime.

Source: World Bank WDI, merchandise exports (current US$, TX.VAL.MRCH.CD.WT), Bangladesh against comparators, 1990-2024.

What did not move is the share of the economy that trades at all. Trade was 18.97% of GDP in 1990 and 27.95% in 2025, a gain of 8.98 percentage points, or 1.47 times the 1990 level. That leaves Bangladesh fifth of the six on trade openness, at 60% of the peer median of 46.27%. India 46.27%, Indonesia 43.39%, Pakistan 27.19%, and at the far end Thailand 138.4% and Vietnam 190.3%, both of which move more across their borders than they produce at home.

Those two series together describe something specific. The export sector grew quickly in dollars while the domestic economy grew quickly enough that trade never became a much larger part of it. That is not the export-led model this comparison group is usually assumed to share. It is a large domestic economy with a single significant export category attached, and the ratio has barely moved in thirty-five years. Two cautions on reading the ratio: revisions to measured GDP move the denominator without any change in trade, and a low openness ratio is normal for populous economies where more of the market is at home. Neither caution changes the composition finding, which is measured inside exports and does not depend on GDP at all.

The services door is not open either

When product diversification stalls, the standard second route is services, which need no port, no container, and no quota.

Trade in services was 5.07% of GDP in 2025, the lowest of the five countries reporting. Vietnam has no 2025 observation in this series, so this comparison rests on five countries, not six. Pakistan 5.19%, Indonesia 7.3%, India 15.54%, Thailand 26.58%. Against the peer median of 11.42%, Bangladesh runs about 44%, a gap of 6.36 percentage points.

One clarification prevents a common misreading. This indicator covers trade in services in the balance of payments. Wages earned by workers abroad and sent home are classified as compensation of employees and personal transfers, not as services exports, so a large remittance inflow does not appear anywhere in this series and does not contradict the 5.07%. That distinction matters for the argument rather than against it: remittances stabilise the current account without diversifying the export basket, because they are not a product anyone is selling.

What these series do not name

The honest limits of this comparison are worth stating in full, because they are wider than usual.

These indicators measure a manufactures share, not a concentration index. They do not name the product. They do not name a single HS line, and nothing here reports what fraction of exports is one garment category rather than a dozen unrelated manufactures. They say nothing whatever about destination markets: no share going to the EU, none to the United States, no measure of buyer concentration. A claim that Bangladesh sells one product to one market is a hypothesis this data supports only halfway. What it does establish is that the basket is the most manufactures-weighted of the six and, inside that weighting, the least technologically upgraded of the six on two separate indicators.

The vintages differ, and by a lot. Composition is read at 2018 for manufactures and high technology and at 2015 for ICT goods, while volumes are 2024 and ratios 2025, those being the latest years with comparable coverage across the group. If the composition has shifted since 2018, this essay is describing a basket that has already changed, and nothing in the data rules that out.

Six countries is a comparison, not a sample. Every relationship here is an association among five comparators and Bangladesh, and none of them identifies a cause.

What would change this conclusion

A specific, checkable test. If the next comparable WDI vintage shows the manufactures share falling toward the peer median of 73.95% while merchandise exports keep rising, and the high-technology share rising above Pakistan's 2.17%, then a second export line is being built and the argument here is wrong. If the manufactures share is still near 94.45% and the high-technology share still under 1%, the basket will have gone another decade without splitting.

Three moves, with owners and signals.

1. Measure the concentration before trying to manage it. The central claim about Bangladeshi export risk, that too much rides on one product going to too few buyers, is currently argued from a manufactures share because nothing better is published on a regular basis. Owner: Export Promotion Bureau, with BBS. Success signal: a standing published series on export concentration by product line and by destination, so this claim can be tested directly rather than proxied.
2. Buy the second line, not the tenth factory in the first. Incentives that reward expansion within the dominant category raise volume without touching composition, which is the variable at issue. Owner: Ministry of Commerce. Success signal: high-technology exports above Pakistan's 2.17% of manufactured exports in the next WDI vintage, which would move Bangladesh off the bottom of the six without yet approaching the peer median of 9.04%.
3. Treat services as an export category rather than an import bill. Services earn foreign exchange without a factory or a shipping lane, and at 5.07% of GDP that route is barely being used. Owner: Ministry of Commerce with the ICT Division. Success signal: trade in services above Pakistan's 5.19% of GDP and holding across two consecutive vintages, the first move off the bottom of the five reporting.

The counterargument

The strongest objection is a serious one, and it is a theory of development rather than a quibble with the numbers. Poor countries do not diversify and then industrialise. They concentrate on the one thing they can make at world prices, build the ports, the power, the logistics and the labour force around it, and diversify later out of the earnings. On that reading a 94.45% manufactures share is not fragility, it is focus, and the 28.27-fold rise in merchandise exports since 1990 is the proof that the focus worked.

Three answers.

First, the sequencing story makes a prediction, and the composition indicators are exactly where the prediction would show up first. It has not shown up. High-technology exports at 0.54% of manufactured exports and ICT goods at 0.05% of total goods exports put Bangladesh at the bottom of the six on both counts. Vietnam, in the same group and on the same dates, was at 40.75% and 29.01%. The concentrate-then-diversify path is not hypothetical, a comparator in this very group has walked it, and the markers of the turn are not moving here.

Second, focus is a strategy at scale and an exposure at 15.7% of the peer median. US$47.2 billion of merchandise exports, fifth of the six, does not buy the market power to hold prices or the reserves to sit out a bad two years in the dominant category.

Third, the concession the objection earns, and it is a real one. Gross product classification overstates the technological distance between Bangladesh and an assembly hub, and the composition readings are older than the volume readings by several vintages. If the manufactures share has fallen materially since 2018, the central figure in this essay is stale. A current composition print is precisely what would settle that, and it is not in this data.

---

Data sources: World Bank World Development Indicators, retrieved from the BDPolicyLab data lake, 2026-08-10. Comparators are India, Pakistan, Indonesia, Thailand and Vietnam; "peer median" is the median of the comparators reporting in the stated year, excluding Bangladesh. Years differ by indicator because coverage does: manufactures share and high-technology exports are 2018, ICT goods exports 2015, merchandise exports 2024, trade openness and trade in services 2025. Trade in services reports for five countries in 2025, Vietnam having no observation, and every comparison drawn from it is stated on that basis.

Sources

1. World Bank WDI, Manufactures exports (% of merchandise exports), TX.VAL.MANF.ZS.UN: https://data.worldbank.org/indicator/TX.VAL.MANF.ZS.UN
2. World Bank WDI, Merchandise exports (current US$), TX.VAL.MRCH.CD.WT: https://data.worldbank.org/indicator/TX.VAL.MRCH.CD.WT
3. World Bank WDI, High-technology exports (% of manufactured exports), TX.VAL.TECH.MF.ZS: https://data.worldbank.org/indicator/TX.VAL.TECH.MF.ZS
4. World Bank WDI, ICT goods exports (% of total goods exports), TX.VAL.ICTG.ZS.UN: https://data.worldbank.org/indicator/TX.VAL.ICTG.ZS.UN
5. World Bank WDI, Trade in services (% of GDP), BG.GSR.NFSV.GD.ZS: https://data.worldbank.org/indicator/BG.GSR.NFSV.GD.ZS
6. World Bank WDI, Trade (% of GDP), NE.TRD.GNFS.ZS: https://data.worldbank.org/indicator/NE.TRD.GNFS.ZS