Collective action and self-governance in the digital commons : an institutional analysis of China's Baidu Baike

Abstract

From irrigation systems and fisheries to communal pastures and groundwater basins, how commons stay efficient and sustainable has been a central question in the governance literature. As online encyclopedias and other digital platforms have grown to shape the way millions share information as a resource across the world, research about the way these platforms qua informational commons are structurally constituted, and how users respond to incentives defined by platform architecture, has subsequently expanded. Although digital technologies make it easier for users to contribute without active coordination, valuable assets such as an extended article on a controversial historical event or a program implementing a new machine learning algorithm involve complex tasks and can only be realized by users partnering with one another in dedicated groups. In an analysis of China’s largest online encyclopedia, I focus on three missing links in the literature. The first link is concerned with perception, specifically the disconnect between standard theory and the values and beliefs on the ground. Content creators on Baidu Baike form groups based on shared interests in different topics and establish their own rules to address their specific organizational needs within each group, such as specifying the threshold for the admission of members on probation and determining how rewards should be distributed among members. While current theoretical propositions would expect perceptions to converge, I conduct in-depth interviews with users and find significant heterogeneity in their perceptions of what rules are supposed to do. Importantly, their values and beliefs often stray from established propositions. The second link has to do with problem definition. This has to do specifically with the puzzle of rule diversity across groups despite environmental stability and limited variability in problem attributes. I find that even if groups face the same challenges, their problem definition and subsequently rule choice still diverge considerably. For example, when confronted with falling asset quality, some groups define it as a social trap dilemma and opt for stricter access rules, while others define as a resource dilemma and focus on mentoring. The diversity in problem definition, a phenomenon typically overlooked in the literature, drives institutional diversity. The third link connects group outcomes to specific types of rules. I adopt Ostrom’s rule typology to quantify the rule choices by the user groups. My analysis reveals an imbalance in how different types of rules impact collective action. Certain rule types significantly influence article quantity and quality, while others show no discernible effect. Groups whose rules are typologically more diverse are also more productive and tend to build better articles, whereas frequent rule changes do not appear to lower performance. My findings show that self-governed groups exhibit remarkable efficiency and sustainability in the digital commons context, but rule configuration accounts for important variation in resource outcomes not well understood in the current literature. My project brings new comparative and methodological value to the existing discussions by drawing these empirical insights from a major non-English information commons and by offering a quantitative application of the IAD rule typology to it.