http://www.lse.ac.uk/GranthamInstitute/publication/tipping-points-and-loss-aversion-in-international-environmental-agreements/

We study the impact of loss-aversion and the threat of catastrophic damages, which we jointly call threshold concerns, on international environmental agreements. We aim to understand whether a threshold for dangerous climate change is an effective coordination device for countries to overcome the free-riding problem, so that they abate emissions sufficiently to avoid disaster. We focus on loss-averse countries negotiating under the threat of either high environmental damages (loss domain) or low damages (gain domain). Under symmetry, when countries display identical degrees of threshold concern, we show that such beliefs have a positive effect on reducing the emission levels of both signatories to the treaty and non-signatories, leading to weakly larger coalitions of signatories. We then introduce asymmetry, by allowing countries to differ in the degree of concern about the threat of disaster. We show that stable coalitions are mostly formed by the countries with higher threshold concerns. When enough countries have no threshold concern, coalition size may diminish, regardless of whether the other countries have mild or strong threshold concerns.