Pay to Publish, Publish to Promote: The Political Economy of Article Processing Charges, Predatory Journals, and the Computational Turn in Global Science

Abstract Scientific publishing has shifted from a reader-pays subscription model to an author-pays open-access economy built on article processing charges (APCs). Global APC payments to six large publishers alone are estimated at US$8.35 billion for 2019–2023, nearly tripling from $910.3 million in 2019 to $2.54 billion in 2023, while flagship APCs now exceed US$12,000. This transition interacts with two under-examined phenomena in lower- and middle-income countries: promotion-driven publishing in predatory venues, an ecosystem of 11,000–19,000 + listed journals whose estimated 420,000 articles (2014) average 2.6 citations, 60% remaining uncited; and a retreat from experimental work toward computational-only studies, decoupling publication metrics from application. A third workaround completes the picture: transnational co-authorship, in which unfunded authors borrow funded Gulf or Asian co-authors to access APC-funded mega-journals such as MDPI. We integrate these dynamics into a single pay-promote-compute framework and propose five measurable indicators, including an Experimental Validation Rate. Applying the protocol to Algeria, we extracted 116,299 works (2015–2025) from OpenAlex: computational-only articles (10.3%) reach twice India's measured share (5.2%), the 2022–2024 retraction rate is 0.99 per 1,000 publications, predatory exposure reaches 3.71% (about twice India's 1.73%), and the strict Experimental Validation Rate stands at 70%. In Algeria, our focal case, the 1998–2018 Arab-region record already showed the highest retraction proportion of any Arab country, with plagiarism the dominant cause. Cross-country benchmarking against India, Turkey, and Egypt, a lenient-versus-strict validation analysis, and policy recommendations for the Global South complete the study.