The reputational economics of open inputs science

This thesis is concerned with the problem of how best to organise the institutions of scientific research. In particular, it is focused on the problems raised by contemporary debates over 'open inputs science': the broad sharing of the intermediate inputs of scientific production on the open web. The problem motivating the thesis can be stated as follows: what institutional structures can incentivise efficient scientific production, via the broad cooperative sharing of intermediate inputs, subject to the constraint that the credibility of scientific output must be maintained or improved? The thesis draws on the resources of new institutional economics, and uses the tools of agent-based modelling, to develop a series of illustrative models of scientific institutions. I use these models to make a multi-stage argument about the structure of science as a global institution, and the implications of this institutional structure for 'open inputs science'. At the most abstract level, the thesis analyses two distinct categories of scientific institution. First, the thesis analyses traditional 'Mertonian' science, in which scientists compete within a reputational economy for priority of publication. Within this reputational economy, the publication of scientific findings is incentivised, but the broad sharing of the intermediate inputs of scientific production is disincentivised. Second, the thesis analyses the institutional structures and proposals associated with 'open inputs science'. These 'open inputs' proposals seek to incentivise both open publication of research findings and the widespread sharing of intermediate inputs like data, code, research protocols, etc. Open inputs advocates and practitioners therefore face the challenge of transforming the incentive structures of traditional, Mertonian science into an institutional system that can encourage broader open input sharing. The thesis analyses these categories of institutional structure along two dimensions. First, I am interested in cooperation within scientific production. I build on Dasgupta and David's analysis of science as a prisoner's dilemma, to explore the institutional scenarios in which different forms of cooperative behaviour are likely to spread across a community of practising scientists. I argue that a scientific community structure in which many small-scale subcommunities are weakly bound together within a larger network can foster the spread of cooperative practices. Second, I am interested in the epistemic outcomes associated with different scientific institutions. I draw on the social epistemology literature, and use agent-based opinion dynamics modelling, to study the impact of specific institutional changes on scientific epistemic outcomes. I argue that there is good reason to believe that open inputs science has a significant positive epistemic impact on scientific consensus-formation. Finally, alongside these formal modelling dimensions of the thesis, I also draw on and contribute to the qualitative literature on science as commons. In particular, I apply the concept of the 'innovation commons' to the study of scientific institutions, and I argue that policymakers interested in achieving open inputs science might be well served by a greater focus on the small-scale innovation commons as a locus of institutional transformation. Taken all together, these elements of the thesis provide a complex synthetic account of the interrelation between different subcomponents of science as a global institution, and the implications of this institutional structure for the project of 'open inputs science'.