Executive finding
The diversification deficit is a scale deficit, and the policy wall taxes the market that would build scale
Chapter 18 of 60 in the Bangladesh 2036 research base. Contents of the series.
The diversification deficit is a scale deficit, and the policy wall taxes the market that would build scale
Bangladesh exports 2,690 products and outside ready made garments not one has crossed 2 billion USD a year: footwear sold 1.65 billion USD, home textiles 1.12 billion, the jute complex 0.67 billion, leather and leather goods 0.67 billion, and every other line under 0.6 billion in 2024 [BACI 2024], against a garment line of 50.90 billion USD whose own record chapter 17 carries [BACI 2024]. The claim this chapter defends is that the binding constraint is scale, not variety, and that the policy regime, a 13.59 percent applied MFN average tariff [WB WITS 2023] wrapped around a bonded warehouse enclave, taxes exactly the domestic market depth that would let a footwear or light engineering line grow to garment scale. The stagnation is measurable against 2017: non-RMG exports of 7.91 billion USD in 2024 stand 6.6 percent above their 2017 level of 7.42 billion, while garments grew 45.5 percent from 34.97 billion over the same seven years [BACI 2024]. The decade question is whether the zone model, around 450 operational industries exporting 8.22 billion USD in FY25, 17.03 percent of national exports, with 533,527 workers at end June 2025 [BEPZA 2025], becomes the template for the whole manufacturing economy or stays an archipelago, and chapter 03 poses the same question to the balance of payments as its unreformed adjustment leg.
The record: 2,690 products sold, none outside garments above 2 billion USD, and a non-RMG basket no bigger than 2017's
The product record at HS6 resolution says the variety exists. The stored trade file carries 2,690 exported products in 2024, against 2,769 in 2022 and 2,774 in 2023 [BACI 2024], and the Herfindahl index of export concentration sits at 0.0459 [BACI 2024], low by the arithmetic of product counts because the file counts everything. The economic complexity index tells the value story the counts cannot: minus 1.37 with rank 207 in 2021, minus 1.64 with rank 215 in 2022, minus 1.68 with rank 216 in 2023, and minus 1.74 with rank 218 in 2024 [BACI 2024]. Chapter 02 carries the same index as a graduation mark; this chapter reads it as a manufacturing reading: the country makes and exports fewer capability-intensive goods relative to its income than almost any peer at its export scale.
The non-RMG lines each carry their own decade story. Footwear exported 1.65 billion USD in 2024, recovering from 1.44 billion in 2023 but below the 1.76 billion of 2022 [BACI 2024]. Home textiles fell from 1.68 billion in 2022 to 1.12 billion in 2024 [BACI 2024], the pandemic boom unwinding. The jute complex, raw fibre plus yarn plus woven fabrics plus sacks, fell from 1.01 billion USD in 2022 to 0.67 billion in 2024, a 33.6 percent fall in two years, with jute yarn at 0.39 billion, raw fibre 0.15 billion, woven fabrics 0.08 billion and sacks and bags 0.06 billion in 2024 [BACI 2024, HS lines 5303, 5307, 5310 and 630510]. Leather and leather goods exported 0.67 billion USD, from 0.69 billion in 2022 [BACI 2024], the Savar tannery estate's compliance story unresolved. Headgear, a quietly successful niche built on the same bonded warehouse rules as garments, exported 0.58 billion USD in 2024, down from its 0.75 billion peak in 2022 [BACI 2024].
The lines the diversification conversation names as next engines are an order of magnitude smaller. In 2024 plastics exported 0.19 billion USD, toys 0.13 billion, bicycles 0.10 billion after a 2022 peak of 0.19 billion, furniture 0.10 billion, chemicals 0.06 billion, electronics assembly 0.06 billion after falling from 0.09 billion in 2022, ceramics and glass 0.05 billion, processed food 0.07 billion, and light engineering machinery 0.03 billion after falling from 0.10 billion in 2022 [BACI 2024]. Ship exports, the scorecard of two decades of shipbuilding announcements, went from 0.04 billion USD in 2020 to zero in 2024 [BACI 2024]. The aggregate is the indictment: non-RMG exports peaked at 9.57 billion USD in 2022 and fell 17.3 percent to 7.91 billion by 2024, 13.4 percent of the total [BACI 2024].
The zone economy scales; the tariff-protected economy stalls
The enclave model is the one part of manufacturing that scales. The export processing zones and the BEPZA economic zone exported 8.22 billion USD in FY25, 17.03 percent of national exports, up 23.8 percent from 6.64 billion in FY21 and recovering from 7.07 billion in FY24, with around 450 operational industries and 533,527 workers at end June 2025, 181,367 of them in Chattogram EPZ [BEPZA 2025]; about 49 percent of EPZ exports come from non-RMG sectors, electronics, footwear, camera lenses and protective equipment among them [BEPZA 2025], a diversification share the national basket does not approach. The zonebuilt footprint the geographic series records, Chittagong EPZ, Dhaka EPZ, Adamjee and the newer Mirsarai economic zone [BEPZA 2023], is the model's physical form. The economic zone programme beyond the EPZs, the BEZA estates, carries no stored occupancy or export series, and the outline's occupancy counts could not be confirmed here, BEZA publications being the resolving source.
The hi-tech park programme is the state's second zone instrument, and its record is the gap between announcement and occupancy in its purest form. The chapter 13 record carries the counts, 39 parks under construction with 5 at commercial operation and 120 companies allocated space as of the stored reading [ICT Division 2020], and no stored series tracks their exports, which the ICT Division annual reports would resolve. The parks matter here because electronics assembly, at 0.06 billion USD of exports and falling [BACI 2024], is exactly the line the parks were built to host, and because their utility model, subsidised power and rent inside a bonded perimeter, is the same gradient logic the EPZs run. If world market terms inside the fence cannot fill the parks, the fence is not the binding constraint.
The domestic market is the diversification target the export numbers hide. Per capita GDP stood at 2,687 USD in FY22 and fell to 2,625 USD in FY24 as the taka depreciated [BBS NA 2024], a consumer market of 173.6 million people in 2024 [UN WPP 2024, medium variant] whose manufactures, from processed food to furniture to electronics, are supplied in part by imports. The tariff wall makes that market profitable for incumbents and unattractive for entrants at the same time: the applied MFN simple average of 13.59 percent in 2023 [WB WITS 2023] sits on 6,877 dutiable tariff lines, only 4.65 percent of the 7,212 lines enter duty free, and the peak rate is 25 percent [WB WITS 2023]. Inside the wall, the bonded warehouse regime gives export manufacturers duty-free inputs, which is why garments, headgear and footwear scale: they live inside the regime. Firms selling into the domestic market pay the wall, which is why the light engineering and electronics lines that would serve it stay at 0.03 and 0.06 billion USD [BACI 2024]. The credit side completes the cost structure: banks carried 5.78 trillion BDT of industry advances of 14.46 trillion BDT total at end-2023 [BB Econ 2024], weighted average interest rates on advances reached 11.52 percent in June 2024 [BB AR 2024], and chapter 06's related-party allocation record names where within industry the credit sat.
Mechanism: four gradients hold the scale cap, and the jute record shows the cap is institutional
The mechanism that kept diversification stalled has four parts. First, the incentive gradient: export enclaves get bonded warehouses, cash incentives and EPZ utilities; domestic market manufacturing gets the 13.59 percent wall that raises its inputs and blunts its discipline [WB WITS 2023]. The cash incentive schedule is measured: rates of 1 to 20 percent against exportation of 43 products and services, with the Export Promotion Bureau issuing 382 recommendation letters for light engineering products, 86 for jute particle board and 38 for PET bottle flake in FY23 alone [EPB 2023]; the jute sector's rates are 12 percent on hessian and sacking and 7 percent on jute yarn [BTTC 2024]. The schedule is LDC-era compensation structured as export subsidy, the legal exposure the trade commission's own record argues in India's countervailing duty proceedings [BTTC 2024], and its annual budget cost could not be confirmed, the Finance Division budget documents being the resolving source. Every line that escaped the gradient, garments, caps, footwear, did so by exporting; every line serving the domestic market stayed small.
Second, the finance gradient: the banking system's credit went to working capital for the established export houses and to the related-party networks chapter 06 names, while the SME tier chapter 22 measures lives on refinance lines whose share of the book that chapter marks as not established. A new line needs patient capital; the system's pricing at 11.52 percent [BB AR 2024] prices out exactly the firms that would build it.
Third, the input gradient: the electronics assembly and light engineering lines import their components at the same landed cost their competitors pay, so their advantage rests on wages, and the wage gap against Vietnam and Cambodia has narrowed in exactly the skill bands these lines need. The stored record shows electronics exports falling from 0.092 billion USD in 2022 to 0.061 billion in 2024 [BACI 2024], and the chapter treats the mobile phone assembly announcements, which could not be confirmed here, as claims awaiting the trade data.
Fourth, the compliance gradient: leather's 0.67 billion USD [BACI 2024] is the ceiling the Savar estate's effluent treatment story imposes. The Leather and Leather Goods Development Policy 2019 records the relocation from Hazaribagh as done and hangs the decade on central effluent treatment plant capacity, audit and certification actions [Leather Policy 2019]; footwear's fall from 1.76 to 1.65 billion USD tracks the unresolved compliance. Buyers pay for certified supply, and certification is an input the policy regime has not delivered.
The employment arithmetic makes the stakes concrete. The Economic Census 2013, the latest held, counted 5.70 million persons engaged in manufacturing of 16.32 million persons engaged economy wide, 34.9 percent [BBS EconCensus 2013]; the EPZs alone employed 533,527 persons at end June 2025 [BEPZA 2025]. Where the next formal jobs must come from is not ambiguous: 44.7 percent of employment in agriculture [ILO 2025] against 11.30 percent of GDP in agriculture [BBS NA 2024] means the workers leaving farms need employers at exactly the scale the non-RMG lines have not reached, and the demographic window chapter 08 dates does not wait for a second garment miracle to arrive on its own.
The jute record is the mechanism's natural experiment, and it indicts borders and ownership, not the fibre. Domestic production rose from 8,432 thousand bales in 2022 to 9,581 thousand bales in 2024 [BBS Agri 2024] while complex export value fell 33.6 percent, so the constraint sits in policy, not agronomy. India's anti-dumping duty on Bangladeshi jute products, imposed in 2017 and carried through a sunset review in December 2022, closed the largest neighbouring market [BTTC 2024]; all state mills under the Bangladesh Jute Mills Corporation have been shut since July 2020 [BTTC 2024], the SOE account chapter 41 carries; and the private base, 220 mills, 700 producers of 282 designated diversified jute products and a mandatory jute packaging rule covering 19 products [BIDA 2023], sells on. On the Export Promotion Bureau's fiscal-year basis, jute and jute goods earned 912.25 million USD in FY23, 708.08 million from goods and 204.17 million from raw fibre [BIDA 2023], against the trade file's calendar year 0.75 billion for 2023 [BACI 2024], a basis difference, not a dispute. The fibre's biodegradability is exactly what the buyer compliance regimes of chapter 58 would pay for; what failed is the institutional layer between the fibre and the buyer, a border duty here and a bankrupt state mill layer at home.
The decade ahead: four decisions decide whether the archipelago becomes an economy
The forces the FY36 scenarios assume are four. First, the tariff path: the IMF programme's tariff reform anchors and the customs modernisation chapter 23 carries point at a falling wall, and the 13.59 percent MFN average [WB WITS 2023] is the number to watch; every point of reduction moves the domestic market lines' economics. Second, the zone model's extension: whether the BEZA estates and the 100 economic zone target, not established (BEZA's own publications state it, none in the registry), reach the occupancy and export series the EPZs show is the decade's industrial policy test, and no occupancy series is stored, so this could not be confirmed, with BEZA publications as the resolving source. Third, the graduation effect: chapter 02's calendar removes the LDC cover under the 1 to 20 percent cash incentive schedule, and chapter 44's market access agenda prices what survives; the lines whose advantage is real keep their buyers. Fourth, the energy and compliance costs chapters 10, 26 and 58 price: gas-dependent ceramics and glass face the fuel security question chapter 26 opens, whose 2022 to 2023 rationing episodes could not be confirmed in measured volumes, on that chapter's own mark, and buyer due diligence extends from garments to footwear and home textiles.
No stored dataset carries line-by-line export projections to 2036, and the chapter treats any such figure as not established until the World Bank's manufacturing sector work or the Policy Research Institute's diversification studies publish one. The decision points: the tariff decision, the zone utility and governance decision, the tannery compliance decision, and the incentive conversion decision, whether the 1 to 20 percent schedule becomes productivity support or lapses with LDC status.
Three risks print in monthly series; the upside is the tariff unification dividend
Risks. First, enclave stagnation: the around 450 operational industries and 8.22 billion USD of FY25 zone exports [BEPZA 2025] grow while the domestic base hollows behind the wall; the revealing indicator is the non-RMG export total against the complexity index, which fell every year from 2021 to 2024 [BACI 2024]. Second, the incentive cliff: removing the 1 to 20 percent schedule at graduation without replacing it with productivity support would hit exactly the lines the diversification conversation names [EPB 2023]; the revealing indicator is the incentive budget's cost, not confirmed here, and the export response in the following year. Third, the energy allocation failure: if the gas rationing chapter 26 records lands on the ceramics and glass lines, the 0.05 billion USD base [BACI 2024] goes the way of ships, which fell from 0.04 billion USD in 2020 to zero in 2024 [BACI 2024]; the revealing indicator is the industry gas allocation series chapter 26 tracks.
Upside. First, the tariff unification dividend: a fall in the 13.59 percent average [WB WITS 2023] toward the regional norm raises the domestic market's contestability and the export lines' input diversity simultaneously. Second, the footwear leather cluster: 2.32 billion USD combined across HS chapters 41, 42 and 64 [BACI 2024] with a compliance fix at Savar [Leather Policy 2019] is the nearest candidate to a second garment-scale industry. Third, the zone multiplier: EPZ export growth from 6.64 billion USD in FY21 to 8.22 billion in FY25 [BEPZA 2025] applied to the wider economic zone programme is the growth mechanism the manufacturing scenario of chapter 15 assumes.
What to watch: five indicators whose thresholds mark the diversification regime
- Non-RMG merchandise exports. Current value 7.91 billion USD in 2024, 13.4 percent of the total, the difference between the 58.81 billion total and the 50.90 billion garment line [BACI 2024]. Threshold: sustained growth above 10 percent a year off the 7.42 billion 2017 base marks the diversification regime; a share below 12 percent of total exports marks deepening concentration.
- Economic complexity index. Current value minus 1.74, rank 218, in 2024 [BACI 2024]. Threshold: a rise above minus 1.5 by around 2030, the target chapter 02 sets, marks the first sustained capability gain; a fall below minus 1.9 marks the stall.
- Applied MFN average tariff. Current value 13.59 percent in 2023 [WB WITS 2023]. Threshold: a fall below 10 percent marks the tariff unification regime; a rise confirms the wall holding.
- EPZ exports and employment. Current value 8.22 billion USD, around 450 operational industries and 533,527 workers in FY25 [BEPZA 2025]. Threshold: zone exports sustained above 10 billion USD with operational industries above 550 marks the enclave to economy turn; a fall back below the FY22 level of 8.66 billion USD [BEPZA 2025] marks the enclave stalling with the rest.
- Leather cluster exports. Current value 2.32 billion USD of HS chapters 41, 42 and 64 in 2024 [BACI 2024]. Threshold: a rise above 4 billion USD marks the second cluster regime; a fall below 2 billion marks the compliance ceiling binding.
Sources used
[BACI 2024] CEPII bilateral trade at HS6 via the trade/bd_hs6_trade, bd_hhi and bd_eci parquets, series: bd_hs6_trade, bd_eci, bd_hhi with num_products, product chapters 41, 42, 63, 64, 65, 39, 85, 94, 84, 89, 95, 28, 29, 69, 70, 03, 16, HS lines 5303, 5307, 5310 and 630510, and the hhi and num_products columns. [BEPZA 2023] Bangladesh Export Processing Zones Authority zone footprint via the lake gis series, series: gis_infrastructure_epz_built_area_km2_*. [BEPZA 2025] Bangladesh Export Processing Zones Authority Annual Report 2024-25 via ocr_text/bepza, EPZ export table FY21 to FY25, 17.03 percent share of national exports, around 450 operational industries, zone-wise employment at end June 2025 summing to 533,527, non-RMG share of EPZ exports. [WB WITS 2023] World Bank World Integrated Trade Solution applied tariff statistics via bdpolicy.db, series: bd_tariff_mfn_simple_avg, bd_tariff_duty_free_share, bd_tariff_peak_rate. [BB Econ 2024] Bangladesh Bank Monthly Economic Trends via bdpolicy.db, series: bb_bank_advances_industry, bb_bank_advances_total. [BB AR 2024] Bangladesh Bank Annual Report 2023-24 via ocr_text/bb/annual_report, weighted average interest rate on advances, June 2024. [BBS NA 2024] Bangladesh Bureau of Statistics national accounts via bdpolicy.db, series: bbs_per_capita_gdp_usd, bbs_gdp_share_agriculture_pct. [UN WPP 2024] UN World Population Prospects 2024 medium variant via the demography/un_wpp2024_bd parquet. [ILO 2025] ILOSTAT modelled estimates via rplumber.ilo.org, series: ilo_emp_temp_sex_eco_nb_a. [BBS EconCensus 2013] BBS Economic Census 2013 via bdpolicy.db, series: ec2013_manufacturing_employment_nos, ec2013_total_persons_engaged_nos. [BBS Agri 2024] BBS agricultural statistics via bdpolicy.db, series: bbs_jute_prod_k_bales. [EPB 2023] Export Promotion Bureau annual report 2022-23 via ocr_text/epb_deep, cash incentive provision of 1 to 20 percent across 43 products and services and FY23 recommendation letter counts. [BTTC 2024] Bangladesh Trade and Tariff Commission annual report 2023-24 via ocr_text/btc, jute goods anti-dumping record: India duty from 2017 with December 2022 sunset review, jute cash incentive rates of 12 and 7 percent, BJMC mill shutdown of July 2020. [BIDA 2023] Bangladesh Investment Development Authority diversified jute products sector profile, June 2023, via ocr_text/bida_deep, FY23 jute export values on the EPB fiscal-year basis, mill and producer counts, mandatory jute packaging coverage. [Leather Policy 2019] Leather and Leather Goods Development Policy 2019, Ministry of Industries gazette via ocr_text/policies, Savar Tannery Industry Estate relocation record and central effluent treatment plant action matrix. [ICT Division 2020] ICT Division annual report 2019-20 via ocr_text/govtwin_min_ict, hi-tech park construction and operation counts, as carried in chapter 13.
Verified line by line against primary sources: 47 claims checked, 7 corrected.
Previous: 17 Ready-made garments
Next: 19 Pharmaceuticals
Cite / Reproduce
BDPolicyLab Research. (2026). 18 Manufacturing diversification. Bangladesh Policy Laboratory. https://bdpolicylab.com/publications/18-manufacturing-diversification
Method and source
Source: Primary sources cited at point of use in the publicationAs of 6 Sep 2026