Bangladesh’s development story is, by most measures, a qualified success and a cautionary one. A country that gained independence amid famine went on to become the world’s second-largest apparel exporter, reduce poverty rapidly, and sustain growth rates that many of its neighbours would envy. The engine of that transformation was narrow but powerful: cheap labour, preferential market access, and a garment industry that absorbed millions of workers into formal employment for the first time.

That engine is now sputtering. Bangladesh is scheduled to graduate from the UN’s Least Developed Country (LDC) category on 24 November 2026. The newly elected BNP government has formally applied for a three-year deferral to 2029, and the UN is reviewing the request. Whatever the outcome, the trajectory is clear. Preferential market access will narrow, compliance demands will rise, and the LDC cushion will thin. Post-graduation tariffs on garments could rise to 9–12 per cent in the EU alone, with projected export losses of up to seven billion dollars annually, against an export base in which garments account for more than 80 per cent of earnings. This comes as rising wages and competition from Vietnam, Ethiopia, and Cambodia erode the cost advantage that made garments work in the first place.

Eliminating anti-export bias: A floor, not a ceiling
Bangladesh’s garment concentration is not purely a market outcome. It is partly the product of government-created distortions. Import tariffs exceeding 25 per cent in many sectors, compounded by para-tariffs and regulatory barriers, impose a persistent anti-export bias: a hidden tax on every exporter that lacks a special carve-out. The garment industry has exactly that: duty-free inputs via bonded warehouses and back-to-back letters of credit. Footwear, light engineering, agro-processing, and non-garment textiles pay the full price.

The first-order policy response is straightforward: comprehensive tariff rationalisation, elimination of the discretionary Statutory Regulatory Orders that entrench garment privileges, extension of bonded warehouse access to all export-oriented sectors, and a market-determined exchange rate. For sectors adjacent to garments in terms of skills and supply-chain logic, removing the bias may be enough to restore their competitive advantage.

But this argument, while substantially correct, is incomplete. Removing a tax on exports clears the path; it does not build the vehicles to travel it. What Bangladesh ultimately lacks is diversified productive capability: the ability to design, manufacture, and market new products competitively. Pharmaceuticals, technical manufacturing, and knowledge-intensive services face obstacles that price signals alone cannot resolve: coordination failures, learning curves, and knowledge spillovers that markets chronically underprice. These are genuine market failures that create a legitimate role for selective industrial policy.

Bangladesh faces a trifecta of converging pressures: the retreat of globalisation and the narrowing of preferential market access; the rise of AI and the automation of service activities that it had hoped would anchor the next phase of growth; and the impending loss, upon LDC graduation, of the pharmaceutical competitive advantage that the TRIPS waiver has quietly sustained. These are not abstractions. They come with timelines, and those timelines are already short.

The consensus has shifted: Industrial policy is back
For three decades, developing countries were effectively coerced, through loan conditionalities and structural adjustment programmes, into abandoning industrial policy. The Washington Consensus was not a suggestion; it was an enforcement mechanism. Countries that deviated were penalised, while those that complied, with few exceptions, stagnated.

Now, in a remarkable institutional about-face, the World Bank’s Chief Economist, writing alongside the March 2026 report Industrial Policy for Development, has conceded that this orthodoxy has run its course, observing with disarming candour that the Bank’s earlier advice “has the practical value of a floppy disk today”. The IMF and UNCTAD have moved in broadly the same direction. Industrial policy, defined as deliberate government intervention to develop specific sectors or productive capabilities beyond what market forces alone would generate, was once dismissed as the refuge of statist romantics. It is now encouraged by some of the very institutions that spent three decades stigmatising it.

The new consensus is unambiguous: eliminating anti-trade bias is a floor, not a ceiling. The choice is not between liberalisation and industrial policy, but rather to pursue both simultaneously: remove distortions that penalise exporters while intervening selectively where markets alone cannot generate new capabilities. Drawing on the work of economists such as Ricard Hausmann and Dani Rodrik, a sector qualifies for support when it exhibits latent comparative advantage blocked by coordination failures, knowledge externalities, capital market imperfections, or the under-provision of public goods — failures that private investors cannot self-correct and whose resolution would generate economy-wide spillovers exceeding the private return. This is not a licence to subsidise every struggling industry. It is a practical filter that distinguishes productive intervention from rent redistribution. Applying this filter to Bangladesh’s economy, seven sectors emerge as plausible candidates for selective support (though identifying the right sub-sectors and instruments remains a matter of adaptation and experimentation rather than top-down prescription.)

Seven candidate sectors
Pharmaceuticals and APIs. Bangladesh meets 98 per cent of domestic pharmaceutical demand and exports to 166 countries. Yet it imports up to 85 per cent of its active pharmaceutical ingredients (APIs) at a cost of US$1.3 billion per year, a textbook coordination failure that no single firm has the incentive to resolve on its own. The foundation of the sector’s competitiveness is the WTO’s TRIPS pharmaceutical waiver, which exempts LDCs from patent obligations and has enabled royalty-free production of cancer drugs, antiretrovirals, and insulin. Critically, this waiver is tied to LDC status, not to a fixed date: upon graduation, Bangladesh will lose this protection when the TRIPS waiver expires in 2033. The consequences are material: insulin prices alone could rise by as much as eightfold. The API Park at Munshiganj was designed to reduce import dependence before that window closes but has stalled for years because of energy shortfalls, infrastructure gaps, and regulatory coordination failures. Once the API Park’s infrastructure and regulatory coordination problems are resolved, Bangladesh could use a time-limited programme of targeted infrastructure support and conditional credit to address this failure. The precise mix of interventions, however, should be treated as a hypothesis to be tested rather than a blueprint to be implemented.

Light Engineering. With a large number of enterprises and substantial domestic demand, the sector has scale but not export credibility. Fragmented production and the absence of shared testing and certification infrastructure — collective goods that no individual firm can provide — appear to be the binding constraints. Investment in high-quality testing and certification infrastructure, alongside standards development, is the natural starting point, with the specific sub-sectors and product categories refined through pilot programmes and market feedback.

ICT and Digital Services. Bangladesh has a large freelance ICT workforce and clear potential in digital services. Yet exports reached only US$724.6 million in FY2024–25 against a US$5 billion target, dwarfed by India’s US$224 billion. The failure appears to be an ecosystem issue rather than a talent issue: skills detached from industry demand, weak university-industry linkages, and absent coordination mechanisms. Promising niches include garment-industry compliance software, AI training data at scale, and fintech solutions for mobile-first, low-income economies — areas in which Bangladesh has a genuine contextual advantage — but which niches will ultimately yield export success can only be discovered through structured experimentation.

Leather and Footwear. Exports grew by 29 per cent to US$620 million in FY2024–25, accounting for 0.29 per cent of the global market, which totalled US$215 billion. Environmental certification appears to be the binding collective constraint: international buyers require effluent-treatment standards that no individual tannery can credibly meet on its own. Common effluent-treatment infrastructure and standards certification are first-rank instruments suited to addressing this failure, though the sequencing and cost-sharing arrangements will need to be worked out in consultation with industry.

Shipbuilding. Confirmed export orders worth US$200 million demonstrate real competitive capability in a scale-intensive, steep learning-curve industry. The binding constraint appears to be access to patient long-term capital that commercial banks cannot provide at the required tenors. Conditional directed credit — as a second-rank instrument, deployed once institutional monitoring capacity is established — is the plausible remedy, though the appropriate conditionality structure will require careful design and iteration.

Agribusiness and Food Processing. Cold-chain infrastructure is a public good that private firms systematically under-provide; food-safety certification requires collective regulatory coordination; and export-market credibility for new products is subject to a self-discovery externality — the pioneer bears the full cost of opening a market that all subsequent exporters share. Agro-processing parks, food-safety quality infrastructure, and market-access assistance are the natural first-rank interventions, with specific value chains to be identified through deliberate trial and error rather than central planning.

Green and Renewable Energy. Energy unreliability is a cross-cutting constraint on diversification across all other qualifying sectors. Public investment and well-designed incentives strengthen the case for intervention, but which renewable technologies and supply-chain segments Bangladesh can realistically develop competitiveness in will need to be discovered through sustained engagement with investors and developers.

Removing a tax on exports clears the path; it does not build the vehicles to travel it. What Bangladesh ultimately lacks is diversified productive capability: the ability to design, manufacture, and market new products competitively. Pharmaceuticals, technical manufacturing, and knowledge-intensive services face obstacles that price signals alone cannot resolve: coordination failures, learning curves, and knowledge spillovers that markets chronically underprice. These are genuine market failures that create a legitimate role for selective industrial policy.

Across all seven sectors, the causal diagnosis points in the same direction: latent comparative advantage blocked by market failures that private investors cannot self-correct. The sector-level analysis above should therefore be read as a map of promising terrain, not a set of firm prescriptions. Successful industrial policy is not a matter of governments picking winners from a list, but of building feedback mechanisms — pilots, performance reviews, and structured dialogue with firms — through which winning combinations of sectors, instruments, and institutions are discovered over time. First-rank instruments — quality infrastructure, coordination mechanisms, and regulatory reform — are the right starting point in most cases. Second-rank instruments — directed credit and conditional subsidies — should follow only where capital-market failures are demonstrable and monitoring capacity exists to enforce conditionality. The potential is not in question; the discipline required to discover and back it is.

The EDF: A warning that bears repeating
Before any of this is attempted, policymakers should be required to review the history of the Export Development Fund. Established in 1989 to provide subsidised foreign-currency loans for imported inputs, the EDF grew over three decades into a US$7 billion facility. Its mandate was export diversification. By that measure, it accomplished essentially nothing. Bangladesh’s dependence on garments is as entrenched today as it was when the fund was created.

No performance conditions were ever attached. No sunset clause existed. And the fund attracted, with the inevitability that anyone familiar with public choice theory would predict, the organised political attention of the garment lobby, the most powerful in Bangladesh’s political economy. When the IMF required the fund to be reduced from US$7 billion to US$2.2 billion, the industry campaigned to restore the original level. The EDF became not an instrument of transformation but a permanent entitlement. The lesson is not that industrial policy cannot work. It is that industrial policy without discipline, conditionality, and insulation from capture does not work. It is also a lesson that sits uneasily alongside any call for structured experimentation: experimentation requires the willingness to abandon what fails, which is precisely what Bangladesh’s political economy has historically refused to do.

Discipline, institutions, and skills
Korea’s industrial policy success in the 1960s and 1970s is routinely cited and almost as routinely misunderstood. What made it work was not the scale of subsidies but conditionality: chaebols received support on the explicit understanding that missing export targets meant losing it. The threat was credible, enforced, and produced firms that eventually competed without support. Bangladesh’s challenge is to replicate that discipline in an environment where almost everything in its political economy runs in the opposite direction.

This requires institutions designed to be capture-resistant rather than goodwill-dependent: a ring-fenced implementation agency with professionally recruited staff; a governing board with genuinely independent membership; performance conditions embedded in primary legislation rather than administrative guidance; and hard sunset clauses requiring demonstrated results for renewal. Small, transparent pilots should precede large bets so that failure, which will happen, remains cheap enough to survive politically.

Skills are the overlooked binding constraint. Industrial policy without human capital investment is a building without a roof. A training levy of one to two per cent of payroll on firms in target sectors, directed into industry-controlled funds, connects training to what employers actually need, a model that Malaysia has operated for thirty years. Bangladesh should also be more deliberate in leveraging its diaspora: a significant proportion of Bangladeshis hold middle and senior positions in Western pharmaceutical and technology firms. A structured return programme — salary supplements, tax relief, and fast-tracked approvals — would transfer capabilities that would otherwise take a generation to build.

The reckoning
Bangladesh faces a trifecta of converging pressures: the retreat of globalisation and the narrowing of preferential market access; the rise of AI and the automation of service activities that it had hoped would anchor the next phase of growth; and the impending loss, upon LDC graduation, of the pharmaceutical competitive advantage that the TRIPS waiver has quietly sustained. These are not abstractions. They come with timelines, and those timelines are already short.

Seven sectors offer plausible starting points for a more diversified export base. But the deeper challenge is not merely identifying candidates. It is building the diversified productive capability that Bangladesh still lacks — and the institutional capacity to support new activities selectively. The knowledge of what needs to be done is not the scarce resource. The discipline to do it honestly is.

Dr. M.G. Quibria is an economist and public affairs commentator writing on trade, development, governance, and democratic change in Bangladesh and beyond. He can be reached at [email protected]. The views expressed here are his own.

Send your articles for Slow Reads to [email protected]. Check out our submission guidelines for details.