Running to Stand Still: Why LDCs are Falling Behind in Digital Trade?

Least Developed Countries are being left behind in the fastest-growing segment of world trade. While digitally deliverable services now account for over half of global services exports, LDCs capture a vanishing share of that growth, and their position is worsening even as their exports rise in absolute terms. Between 2015 and 2023, their digital-services exports grew 43% in dollar terms, yet their global market share fell over the same period. That divergence is the puzzle this analysis takes up: not why a digital divide exists, which is intuitive enough, but why it keeps widening even as LDCs’ own exports grow. Is this simply a matter of infrastructure catching up over time, or does the digital services economy, with its compounding returns, concentrated capital, and now-fragmenting trade rules, structurally reward early movers in ways that make the gap self-reinforcing rather than self-correcting?

Global trade is undergoing a structural shift from goods to services, and within services, from physically delivered to digitally deliverable modes. According to UN Trade and Development’s (UNCTAD) September 2026 Global Trade Update, services now account for 71% of global intermediate inputs and 27% of total global exports, having grown roughly 6.7% annually over the past decade and 8.3% in 2025 alone, outpacing goods trade. Within that expansion, digitally deliverable services, telecommunications, computer services, financial services, and other outputs transmitted over networks rather than shipped, have grown even faster, at 7.1% annually, and now constitute 56% of global services exports.

This would ordinarily read as a development opportunity: digital services carry near-zero marginal transport costs, theoretically allowing distant, low-infrastructure economies to compete on skill and cost rather than proximity to markets. The empirical record contradicts this optimism. Least Developed Countries (LDCs), the 44 lowest-income, most structurally vulnerable economies as classified by the UN, captured just 0.6% of global services exports in 2025, and within the digitally deliverable segment specifically, their share was only 0.16%, the lowest level since records began.

Digitally deliverable services make up just 16% of LDCs’ own services exports, versus 61% in developed economies. The gap is not closing; it is the widest it has ever been measured.

The central analytical question this raises is not why a divide exists, but why it is widening even as the market grows and even as LDCs’ absolute exports rise. Between 2015 and 2023, LDCs’ digitally deliverable services exports grew 43% in absolute dollar terms, yet their global market share fell over the same period. This is not a story of stagnation; it is a story of relative velocity, LDCs exports are growing more slowly than the market as a whole, and positive absolute growth can coexist with a continuing decline in market share.

One plausible mechanism behind this divergence is a difference in export composition: developed economies are more heavily represented in high-value-added, IP-intensive segments, proprietary software, data analytics, cloud architecture, AI-related services, that command greater pricing power, while LDC exports remain concentrated in more labour-intensive, lower-complexity activities such as entry-level business-process outsourcing and routine IT support. This should be treated as a testable hypothesis rather than an established finding: the observed fact is that LDC digital-services exports have grown while their global share declined; the proposed mechanism is that differences in composition and value added may contribute to that divergence. Demonstrating it conclusively would require country- and category-level evidence on export volumes, prices, and value added over time, evidence not currently assembled at that granularity. That distinction should discipline the rest of the analysis: the right policy question is not simply how to get LDCs online, a threshold problem already largely understood, but how to get them onto a comparable growth trajectory; a harder, dynamic problem.

Treating the digital divide as a single barrier obscures more than it reveals. The evidence supports separating it into four analytically distinct channels, connectivity, payments and foreign-exchange infrastructure, regulatory and data governance, and skills/AI capacity, each operating through different economic mechanisms and requiring a different policy response. These channels are complementary rather than independent: connectivity, for example, is likely necessary for digital-services exports but clearly not sufficient on its own.

First, the connectivity challenge in LDCs is not only about network access but about whether people can use the connectivity available. ITU’s Facts and Figures 2025 estimates that only 34% of the LDC population uses the internet, against roughly 74% globally.

Earlier ITU analysis helps explain what sits behind this divide: in 2022, 17 percentage points of the LDC population faced a pure access gap, no broadband coverage, while a much larger 47 points faced a usage gap despite living within coverage, driven by affordability, digital skills, device access, and the perceived relevance of online services. Affordability remains particularly important: mobile broadband costs a substantially larger share of income in LDCs than globally, especially for lower-income households. Quality compounds the problem, only around 4% of people in low-income countries have 5G coverage, versus 84% in high-income countries, and for latency-sensitive digital services such as video-based customer support or cloud-based freelance work, this quality gap can translate into a competitiveness disadvantage even for those who are technically connected.

Second, even where connectivity exists, payments infrastructure presents a severe cloud-payment paradox. The World Bank’s 2025 Global Findex reports that 75% of adults in low- and middle-income economies now hold a financial account, an 80% increase since 2011, driven substantially by mobile money. But the binding constraint has shifted from ownership to activity: only 42% of adults in low- and middle-income economies made a digital merchant payment in 2024, with South Asia at 15% and Sub-Saharan Africa at 20%. Beneath this sits a sharper, firm-level constraint that consumer statistics don’t capture: scaling a digital-services exporter requires outbound foreign-currency payments for cloud hosting, software licenses, and developer APIs, and in a number of LDCs, foreign-exchange controls and illiquidity make this difficult. Firm-level examples from African technology markets, cloud providers adopting mobile-money-based payment channels where conventional international transfers are costly or constrained by foreign-exchange shortages, illustrate this asymmetry, though they should not be read as representative of all LDC exporters. Regional payment-rail integration, such as Africa’s Pan-African Payment and Settlement System, is a more tractable near-term intervention than physical infrastructure, which requires decade-scale capital.

Third, beyond physical and financial infrastructure sits a less-discussed non-tariff barrier: regulatory compliance and cross-border data governance. As advanced economies enforce stringent data-privacy regimes such as the EU’s GDPR, alongside cross-border data-transfer rules and emerging AI safety standards, vendors exporting into these markets must demonstrate compliance. The mechanism need not be discriminatory regulation: common compliance requirements still have unequal effects, since their fixed costs represent a far larger share of revenue for a small LDC firm than for an established multinational, creating a disproportionately high cost of entry for exactly the exporters LDCs need to succeed.

Finally, skills, AI readiness, and price-deflation dynamics threaten to entrench this disparity. UNCTAD’s Frontier Technology Readiness Index, a composite of ICT deployment, skills, R&D, industrial capacity, and access to finance, shows the skills gap persists independent of income level: India ranks 3^rd globally in R&D and 10th in industrial capacity, yet only 99^th in ICT deployment and 113th in skills. Bangladesh, one of the few LDCs with genuine digital-export traction, ranks 112^th of 170 on the same index, though it has climbed from 121^st in 2022. The AI dimension sharpens this from a skills problem into a concentration problem: in 2022, just 100 companies accounted for around 40% of global business-funded R&D, evidence of broad corporate-innovation concentration, though not by itself proof that AI concentration specifically is driving weak LDC digital-services exports.

Modern digital services and AI platforms also exhibit strong increasing returns to scale, high fixed costs in research, compute, and data against near-zero marginal distribution costs, letting capital-rich firms amortize investment across global user bases and reduce average costs as they scale. Reinforced by network effects and the agglomeration of talent and capital, these forces can concentrate digital production in established hubs and raise structural barriers to entry for LDC firms even absent any formal trade restriction. A plausible, forward-looking risk, best treated as a scenario rather than a demonstrated explanation for the current decline, is that AI may simultaneously lower the skill threshold to enter categories like basic coding or routine customer support while reducing the market price of those same standardized services, leaving LDCs at risk of scaling into categories whose revenue per unit erodes faster than volume can grow.

The risk of conflicting international trade rules is no longer purely prospective. At the WTO’s 14^th Ministerial Conference in Yaoundé, Cameroon, in late March 2026, members failed to renew the moratorium on customs duties on electronic transmissions in place since 1998, and it lapsed on March 30, 2026. Two days earlier, on March 28, a group of WTO members (reported variously as 66 or 67) co-convened by Australia, Japan, and Singapore adopted Interim Arrangements for the WTO Agreement on Electronic Commerce, committing to duty-free treatment among participants pending full incorporation into WTO law. The United States did not join. The landscape is more fragmented than a simple participant/non-participant split, however: on May 8, a separate group of 19 members, including the United States, Japan, and Singapore, pledged their own continued duty-free treatment after the broader deal stalled over opposition from Brazil and Turkey. And while most LDCs remain outside the main Electronic Commerce Agreement, it would be inaccurate to say virtually all are excluded: Benin, Burkina Faso, The Gambia, and Lao PDR participate in the agreement.

The distributional consequences for LDCs are genuinely ambiguous rather than straightforwardly negative. The concern is that digital trade now risks fragmenting into groups operating under different tariff commitments and legal arrangements, raising compliance complexity for exporters least equipped to absorb it. But some developing and middle-income economies, including Indonesia and Brazil, had resisted the underlying agreement, arguing it constrained tariff revenue and policy space for domestic digital industrialization, an argument that carries more force for net digital-services importers with ambitions to build domestic digital industries than for LDCs, which are overwhelmingly net importers without near-term export capacity, and thus stand to gain relatively little from new tariff authority while potentially facing higher costs on the imported digital inputs their emerging sectors depend on. This suggests, rather than definitively establishes, that the lapse could impose relatively greater adjustment costs on LDCs than on more technologically advanced developing economies, an important disaggregation the developing-versus-developed binary in much commentary misses. A partial offset is emerging through preferential agreements: 55% of those signed between 2000 and 2025 contain e-commerce or digital-trade provisions, though whether LDCs are included on favorable terms remains an open empirical question. The expiry of the multilateral moratorium is, in effect, a natural experiment: how digital-input costs and export participation evolve across countries subject to different post-2026 arrangements could help identify how much trade-rule certainty matters for digital-services competitiveness.

UNCTAD’s Least Developed Countries Report 2025 finds that LDC digitally deliverable services exports are concentrated among a small group (Bangladesh, Ethiopia, Senegal, Nepal, Cambodia, and Uganda) within narrow categories, chiefly telecommunications, computer services, and professional consulting. Bangladesh is a useful illustrative case: a mid-tier and improving Frontier Technology Readiness ranking, combined with a critical mass of trained IT-enabled-services workers, has produced export outcomes disproportionate to its connectivity or income level. Calling it the definitive clearest positive deviant would require specifying a formal benchmark; without one, it is safer to treat Bangladesh as illustrative rather than uniquely strongest. Its experience is suggestive rather than conclusive evidence that good enough infrastructure paired with a targeted skills pipeline may partially compensate for broader infrastructure constraints, a stronger test would compare it against LDCs with similar connectivity gains but weaker export outcomes, isolating skills from confounds like diaspora links, language, and labour costs.

Barriers also operate on both sides of the border. Abroad, restrictive visa regimes and non-recognition of professional qualifications constrain LDC service suppliers’ ability to deliver services requiring temporary movement of people (Mode 4 trade); since some higher-value digital services depend on initial face-to-face relationship-building, this can indirectly constrain the digital trade that follows. At home, weak regulatory frameworks and limited access to finance constrain firm-level scaling, more tractable in the short run than infrastructure gaps, since visa facilitation and mutual recognition agreements require negotiation rather than capital.

The evidence resists a single-lever narrative and points to a layered, time-sequenced intervention: reducing foreign-exchange and payment frictions, regulatory-compliance support, and mobility agreements in the short term; targeted skills pipelines, financing access, and sector-specific export strategies in the medium term; and sustained investment in broadband quality and domestic innovation ecosystems over the longer run. This sequencing reflects a distinction between necessary and sufficient conditions: connectivity is necessary for participation in digitally delivered trade but unlikely alone to generate competitive exporters, just as skills development has limited impact if firms cannot receive international payments or meet foreign regulatory requirements.

Taken together, LDCs are caught in a compounding-returns market where connectivity, payments, regulatory alignment, and trade-rule access must improve in complementary ways to prevent the export-share gap from mechanically widening even under positive absolute growth. Bangladesh’s experience suggests a well-targeted skills strategy can generate disproportionate gains even before broader infrastructure constraints are resolved, but it is one data point, not a template. The central implication is not simply more digitalization, but the deliberate sequencing of an export ecosystem in which complementary constraints are removed in an economically sensible order: tractable institutional and financial reforms first, while longer-term investments in connectivity, skills, and innovation capacity accumulate.

International Telecommunication Union. “Global Internet Usage Remains Uneven, ITU Reports.” Press release, November 17, 2025. https://www.www.itu.int/en/mediacentre/Pages/PR-2025-11-17-Facts-and-Figures.aspx

International Telecommunication Union, “Statistics Update,” December 2025, https://www.itu.int/en/ITU-D/Statistics/Pages/StatisticsUpdate/December2025.aspx

UNCTAD. Global Trade Update: Services Are Reshaping Global Trade. Geneva: United Nations Conference on Trade and Development, September 2026. https://unctad.org/publication/global-trade-update-september-2026-services-are-reshaping-global-trade

UNCTAD. The Least Developed Countries Report 2025. Geneva: United Nations Conference on Trade and Development, 2025. https://unctad.org/publication/least-developed-countries-report-2025

World Bank. Global Findex 2025 Overview. Policy Research Talk Presentation, September 2025. https://thedocs.worldbank.org/en/doc/a4aaae5a017a5bea7922686af0ed3636-0050022025/original/PolicyResearchTalk-GlobalFindex-September2025.pdf

World Trade Organization. “Joint Statement Initiative on E-Commerce.” Accessed September 21, 2026. https://www.wto.org/english/tratop_e/ecom_e/joint_statement_e.htm

World Trade Organization. “MC14 Briefing Notes: Work Programme on Electronic Commerce.” World Trade Organization. Accessed September 20, 2026. https://www.wto.org/english/thewto_e/minist_e/mc14_e/briefing_notes_e/ecommerce_e.htm