The Open Access Paradox: Quantifying Returns from University-Publisher Agreements

Status: Working Paper
Coauthors: Eddie Ning (UBC), Qiyuan Wang (Hong Kong Polytech)

Abstract

nderstanding the impact of paid open access publishing on scholarly visibility is crucial because open access often comes at a steep price tag for researchers, universities, and funders, making it important to evaluate the value of these investments. This paper provides a large-scale causal analysis of the impact of paid open access in subscription-based journals on citations, leveraging a natural experiment created by the staggered implementation of transformative agreements across the U.S.. Using an instrumental variables approach, we find transformative agreements significantly boost hybrid open access adoption by around 26 percentage points, but paying for open access yields a modest citation increase averaging 14%. Notably, citation gains are heterogeneous across publishers and research domains, driven partly by differential accessibility through PubMed Central (PMC), a publisher-mediated repository of published articles relevant to biomedical research. Increased hybrid open access also reduces use of self-archiving repositories, suggesting a substitution between paid and free open access channels. After controlling for a paper’s presence on PMC and self-archiving repositories, we observe a much smaller impact of paid open access on citations with less heterogeneity across publishers and domains. These results illustrate the limited effectiveness of paid open access and emphasize the importance of discipline-specific repository infrastructure, providing important policy implications for institutions evaluating transformative agreements.