Uneven Returns from Digitalization and AI: Economic Mechanisms and Policy Implications
Razvan Vlaicu
No 14833, IDB Publications (Working Papers) from Inter-American Development Bank
Abstract:
Digitalization and artificial intelligence (AI) are increasingly viewed as generalpurpose technologies with the potential to reshape economic activity, raise productivity, and improve public services. Yet their returns are highly uneven across firms, workers, and countries. This paper develops an economic framework for understanding these differences, focusing on four properties of digital and AI technologies: strong complementarities among inputs, the need for workflow and organizational redesign, economies of scale, and economies of scope. These features help explain why technological adoption alone may generate limited gains, why productivity effects often emerge with a lag, and why benefits tend to concentrate in firms and locations with stronger capabilities and infrastructure. The paper applies this framework to Latin America and the Caribbean and discusses policy priorities for expanding digitalization, strengthening digital government, addressing information and coordination failures, and broadening access to the productivity gains associated with AI.
Keywords: digitalization; artificial intelligence; general-purpose technologies; Productivity; economic returns (search for similar items in EconPapers)
JEL-codes: D24 O33 O38 O54 (search for similar items in EconPapers)
Date: 2026-10
References: Add references at CitEc
Citations:
Downloads: (external link)
https://publications.iadb.org/publications/english ... icy-implications.pdf (application/pdf)
Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
HTML/Text
Persistent link: https://EconPapers.repec.org/RePEc:idb:brikps:14833
DOI: 10.18235/0014588
Access Statistics for this paper
More papers in IDB Publications (Working Papers) from Inter-American Development Bank Contact information at EDIRC.
Bibliographic data for series maintained by Felipe Herrera Library ().