AI clustering splits the world’s digital economies into two speeds

A latest cross-country analysis published in the Journal of Theoretical and Applied Electronic Commerce Research (JTAER) reports a sharp, data-driven divide in the global digital economy and a roadmap for narrowing it. Using artificial intelligence (AI) to cluster 152 countries across 47 e-commerce and digital-readiness indicators, the authors find a stable split between a larger group of less-developed digital economies and a smaller cohort of more-developed ones, with clear levers for “cluster migration” through policy and industry action.

The study, “Using Artificial Intelligence to Determine the Impact of E-Commerce on the Digital Economy,” builds a fused indicator matrix spanning ICT infrastructure, payments, trade and logistics, legal frameworks, financing, e-commerce readiness, and multi-year UNCTAD B2C index values and ranks. Roughly 30% of the dataset comprises missing values, a structural reality the authors treat not as noise but as signal, deploying a multiple-imputation and consensus-clustering pipeline to separate meaningful patterns from sparsity.

Can AI-driven clustering reveal how e-commerce shapes the digital economy?

The authors assemble a country-by-indicator matrix for 152 economies and 47 variables, then apply a two-stage AI approach: first, clusterMI-based multiple imputation to handle missing data in a way that respects plausible cluster structures; second, k-means consensus clustering (with hierarchical methods used for confirmation) to identify robust groupings. This method returns a remarkably stable two-cluster world. One cluster aggregates less-developed digital economies; the other aggregates more-developed ones. Critically, the variables that most strongly separate countries are e-commerce readiness measures and UNCTAD B2C e-commerce index ranks and values, underscoring the central role of e-commerce capability in the broader digital economy.

A striking methodological result emerges: even when the authors replace observed numerical entries with uniform random values, the pattern of missingness alone is sufficient to reproduce the broad developed vs. less-developed split. That finding reframes data scarcity as a macro signal, exposing structural weaknesses in digital measurement, governance, and reporting that correlate with real capability gaps. The authors argue that resolving those gaps, through better data, stronger legal frameworks, and more complete coverage of e-commerce indicators, should be treated as a policy priority on par with infrastructure or financing.

Within the higher-performing cluster, the study proceeds to a four-way sub-clustering, revealing distinct profiles among advanced and emerging digital economies. This finer segmentation highlights where countries align on readiness, payments penetration, logistics competence, and rule-of-law protections that underpin trust and scale in e-commerce. The resulting map creates a practical scaffold for pairing partners across clusters and sub-clusters to accelerate capability transfer.

Do countries with similar structures integrate better via e-commerce indicators?

The analysis shows that countries with comparable economic and institutional structures tend to sort into the same digital-economy sub-clusters when the model emphasizes e-commerce-centric indicators. In practice, that means policy and business alliances work best when they target like-for-like complementarities: mature economies with deep logistics and legal capacity can help emerging peers plug financing, governance, or skills gaps; ambitious middle-income countries can scale faster by aligning with sub-clusters that mirror their strengths while counterbalancing their weaknesses.

Indicator families tell the story. Payments, trade and logistics, and legal frameworks are the biggest differentiators between the two headline clusters, while broadband/service and e-commerce skills show smaller, but still material, gaps. This pattern suggests that many developing markets have already built meaningful formal technical capability, yet lag on institutional and ecosystem enablers needed to commercialize those capabilities at scale. The data also reveal unexpected outliers: some large tech-savvy markets fall into the lower cluster due to uneven coverage and fragmentation, indicating that pockets of excellence do not automatically raise national performance on e-commerce indices without parallel progress in data completeness and institutional depth.

The authors detail how sub-cluster contrasts illuminate practical integration pathways. Among advanced economies, one sub-cluster concentrates countries with long industrial traditions and strong business-to-business competencies, while another groups smaller or emerging players with improving readiness but uneven financing or legal depth. For policymakers and firms, the implication is direct: choose cooperation partners whose indicator profiles complement your constraints, and tie collaboration to measurable improvements in payments access, logistics reliability, consumer protection, cybersecurity, and data privacy enforcement - the very variables the model finds most predictive of digital performance.

Can clustering accelerate digital transformation and guide policy?

The authors argue that AI-guided clustering is not just descriptive; it is operational. By quantifying how indicator sets move countries across cluster boundaries, the framework provides actionable levers for governments and industry. The study recommends that lagging clusters prioritize fast-lane skills programs in IT/AI aligned to e-commerce applications, expand rules-based legal protections around transactions and data, and mobilize financing channels that lower the cost of digital adoption for firms. Meanwhile, developed sub-clusters are urged to adopt inclusive cooperation policies that translate institutional strength and know-how into scalable partnerships, enabling faster capability lift for partners and opening new markets for advanced providers.

A key policy message is that digital transformation depends as much on institutional reliability as on bandwidth. Payments penetration, delivery timeliness, customs procedures, consumer-protection enforcement, and cybercrime deterrence map strongly to the higher-performing group. Because these are multi-actor systems, involving regulators, courts, banks, logistics operators, and platforms, progress requires coordinated action rather than siloed technology upgrades. The authors frame this as a blueprint for targeted cluster migration: focus on the indicators with the highest separating power to move national profiles from the lower to the higher digital cluster.

The paper also takes a clear-eyed view of limitations. Missing values are pervasive across global e-commerce datasets; many indicators are not consistently reported over time or across countries. The authors mitigate this with stability-oriented multiple imputation and consensus clustering, but they stress the need for richer primary data and expanded indicator coverage to sharpen policy precision. They point to future extensions, topological data analysis, constrained or semi-supervised clustering, and variable-relevance detectors, to better identify “margin countries” most likely to switch clusters with incremental reforms.

The findings carry operational implications for business. Retailers, marketplaces, and payment providers seeking expansion can use the cluster map to sequence market entry, prioritizing countries where institutional readiness and e-commerce indices predict smoother scaling. Logistics firms can align service footprints with sub-clusters showing rising readiness but weak delivery competence, offering reliability as the growth catalyst. Fintechs and banks can target collateral, working-capital, and inclusion bottlenecks that the financing indicators expose. For each sector, the path to growth runs through the same policy levers the model elevates: payments, logistics, legal assurance, and skills.

The research concludes with a practical agenda: promote open, high-quality data to reduce reliance on imputation; invest in open-access AI tools that help governments and firms model policy trade-offs; and structure international alliances around measurable e-commerce outcomes. The message is consistent with the study’s empirical core: e-commerce capability is a leading edge of the digital economy, and AI-based clustering can show exactly where to push to lift national performance - fast, measurably, and at scale.