Using game theory for environmental modelling

International environmental agreements are an important framework for countries to make a commitment to sustainability and bring pressing issues to the forefront of public attention. However, the targets and agreements made at international conferences such as COP27 often fail to meet expectations in the long-term as individual governments or nations renege on their commitments.

Climate finance – in which developed nations offer financial incentives to developing nations to change an aspect of their behaviour – is a real point of contention, with both sides often failing to adhere to their promises. A precedent of broken treaties erodes trust between nations and has serious consequences for future international collaboration, potentially limiting our ability to mitigate the most serious impacts of climate change.

Understanding the different motivations and pressures experienced by each party is vital to create an effective agreement that is fair and attainable for all sides. Supported by the Open University (OU)’s Open Societal Challenges programme, computer scientist Dhouha Kbaier, her Postdoctoral Research Associate, Ian Kenny and their team of multidisciplinary collaborators are working on a strategy to visualise and balance the often-conflicting issues surrounding these agreements. This project is in collaboration with researchers from the University of Bath, Richard Fairchild and Neal Hinvest.

"We want to use mathematical modelling and game-theory to understand the economic, behavioural, and psychological factors affecting the incentives for governments and nations to form international agreements, and to adhere to those agreements over the long-term when there are individual incentives to renege or free-ride," Kbaier explains. "The model will analyse the conditions for large, effective, and stable international environmental agreements to form, and for international governments to adhere to the targets. This should provide clear policy implications for the future."

The team's model will be based on game-theory which considers how different strategic interactions between ‘players’ lead to different possible outcomes. In many scenarios, collaboration or investment comes at a short-term cost but yields a better result overall. 'Players’ must therefore balance this cost against the immediate gain of an easier every-man-for-himself strategy.

For example, recent research carried out by OU researchers in conjunction with Exeter, indicated that the cost of continuing to invest in fossil fuel infrastructure is disincentivised by the reduced cost of investing in renewables, thereby negating the advantage of a free-ride strategy. Kenny is therefore looking at the potential of adjusting a game-theoretic model using machine learning to recognise this move away from the free-ride model.

The parallels with international politics are clear: a united global strategy to tackle the myriad of environmental problems would be the ideal solution, but this must be balanced against the individual pressures on governments from their own citizens. "Our model considers the economic (including transfers from developed nations) and behavioural (warm glow) incentives to form an agreement, but also each nation's internal pressures from its own citizens and super-influencers," says Kbaier. "These factors, as well as the influence of individuals such as Greta Thunberg or Donald Trump, change over time, affecting public attitudes which in turn have an effect on the governmental implementation of policies."

In the first part of the project, the team will gather data on the public perception of different influencers and how these individuals change people’s attitudes to climate and government responsibility over time. These data will then be fed into a game-theory model of possible strategic interactions between democratic countries forming an agreement, creating a dynamic model to reflect the changing incentives and pressures of each party. The understanding gained from this model will therefore provide a clearer picture of the ideal circumstances to support a successful agreement and the potential barriers which could derail negotiations.

"The vision is to develop a deeper understanding of how governmental and international bodies are influenced with respect to these agreements," says Kbaier. "Ultimately, we want these models to help us develop stronger and more stable international environmental agreements, offering greater hope for the planet's climate and survival."