Public good games with peer-pressure and information diffusion in multiplex networks. Models and Experiments

The success of sustainable technology as well as policies cannot prescind from the uptake these are going to have amongst the general public. The selfish human nature makes ideal policies hardly implementable when an immediate return is not evident for those to whom it is addressed. Very often, policies include punishments for those not following them, yet the effectiveness of this deterrent, as of any advantage the policy brings on its adoption is only in part correlated to its magnitude. Instead, policymakers have been increasingly turning towards behavioural nudges by changing perceived social norms about technology adoption. Interestingly, the adoption of a new technology or behaviour appears linked to the benefit that individuals perceive about their peers.

Adherence to sustainable policies, behaviours and using sustainable technologies often requires a degree of individual sacrifice for a greater common good, which eventually reflects onto the individuals. This sacrifice could be financial (e.g. purchasing more expensive but lower carbon-associated items), related to personal comfort (accepting to travel more slowly by public transport as opposed to private car) or to anything which affects selfish optimal behaviours.

When deciding whether to provide the public good, individuals are embedded in two types of social networks. The first network defines which people interact and benefit from the provided public good. The second network defines the flow of information, indicating the agents whose actions are observed. Since the two networks are often distinct, this project will study how decisions are affected by the structure of each network.

The theoretical study of the two types of networks will rely on behavioural game theory and complex adaptive systems approach. The interaction network determines the set of participants who can contribute to the public good, and also the participants who will benefit from it. Purely self-interested agents would never contribute anything, yet it is known that many people are conditional cooperators, which essentially transforms the public goods game into a coordination game. The project will study how the willingness to contribute depends on the interaction network structure, distribution of player types and other game parameters.

Instead, the information network determines what people learn from observing the actions and payoffs of others. If groups that overcome the collective action problem earn more than those that freeride, agents may learn that attempting to adopt the more efficient technologies might pay off in the long run, opting to initial such adoption in their own group. The project will run simulations using various types of payoff-based learning models to make predictions about how the information network structure affects the evolution of cooperation. The modelling may use an individual-based mean field approach where the state variable maps to the level of cooperation the players strategically decide upon.

Predictions from behavioural game theory and the simulations of learning models will be tested in incentivized laboratory experiments, where participants will play the public good games in a network. In the repeated game, participants will have an opportunity to update their strategies based on their own payoffs and on the information about the actions and payoffs of the subset of players that they observe. Results from these experiments will be used to tune the model parameters. In addition, methods from complex adaptive systems will be used to simulate the path of choices for a given social network to further extend the analytic model. Simulations will also allow us to study the evolution of choices under a broader set of circumstance, which would be unfeasible in the laboratory setting (e.g. large group sizes and more complex network types).

Informal enquiries are encouraged and should be addressed to the supervisor.

Funding Notes

Applicants are required to hold either an undergraduate honours degree (2:1 or 1st) or MSc (merit or distinction) in mathematics or engineering, with knowledge of game theory and networks. If the degree is not from an English-speaking country, the applicant needs an overall IELTS grade of 6.5 with a minimum of 6.0 in each component (or equivalent).

3.5 years scholarship. This funding opportunity is open to both home and international applicants. Successful candidates will receive financial support covering tuition fees at the domestic rate and a stipend for a duration of 3.5 years, irrespective of their residency status. It is important to note that international students will be responsible for paying the difference between home and international tuition fees.