Human cooperation for the sustainable management of public goods under institutional punishmentand corruption

Human society faces significant challenges in managing public goods and commonresources such as soil, water, energy, and urban spaces. These are underscored bythe Horizon Europe Missions, which emphasise sustainable development,environmental protection, and social cohesion. The effective management of collective resources is particularly challenging due tothe tension between optimizing collective outcomes and individual incentives.Achieving optimal collective outcomes requires persistent and effortfulengagement from participants, whereas individuals often have incentives to avoidsuch effort. Promoting cooperative interactions and behaviours among members ofsociety is then the key to achieving sustainable public good creation andmanagement. One of the most successful approaches to addressing these problems is theirframing as social dilemmas, especially the public goods and common resourcegames [11,15,16]. There, the conflict between public goods and private incentivesleads to the so-called tragedy of the commons, involving the depletion of commonresources and inefficient provision of public goods [12,13]. While many mechanisms to enable cooperation have been studied in the literature,including reputation, direct and indirect reciprocity, incentives, and punishments [14], several others are known to disrupt the ability of a group to effectivelymanage common resources: misinformation, deceptive communication, weakinstitutions, and corruption [5]. Finally, some mechanisms produce less clear-cuteffects on cooperation, such as the possibility of an individual to opt-out (ormigrating) to avoid an unwanted interaction [2]. The overarching goal of this proposal is to advance the understanding of thesemechanisms (exit and corruption) in a context characterized by weak institutionswith limited ability to punish individuals. To this end, we will formulate novelvariations of public goods or common resource games that more closely mirrorreal-world social situations, suitable for the study of corruption and/or exitstrategies. We will then assess the impact of the aforementioned mechanisms oncooperation in these social dilemmas through behavioural economics experimentsand computational models involving cognitively-rich agents capable of learningnovel strategies (i.e. via multi-agent reinforcement learning). The computationalmodels will inform the definitions of the precise setups and behavioural patterns tobe investigated through experiments. The latter, in turn, will be used to improvethe behavioural modelling of computational agents. In conclusion, this project represents a comprehensive approach to understandingand addressing the challenges of managing public goods and common resources,contributing to the EU's mission of promoting sustainable development for thebenefit of all by helping policymakers decide which factors to consider whenattempting to incentivise sustainable cooperation.