Value of information under climate targets: an application of cost-risk analysis

The optimal policy that balances the cost of mitigation with the damages from climate change can be assessed by examining the interactions between the socioeconomic system and the climate system. Traditionally, Cost-Benefit Analysis (CBA) is used for this problem but requires an approximate economic evaluation of the totality of climate damages which, in turn, calls for extensive impact assessments and ethical debates. Until such damage functions are agreed upon, another operational framework is essential for policy analysis. To fill the gap, climate (e.g. temperature) targets have been used in conjunction with Cost-Effectiveness Analysis (CEA). However, this analysis breaks down if a violation of the target is inevitable. This becomes especially salient when considering uncertainty and learning under temperature targets, given the presently infinite tail of climate sensitivity. A remedy was proposed that trades off the costs for mitigating climate change against the risk of exceeding climate targets: Cost-Risk Analysis (CRA). The implicitly defined preference order contained in the formulation of a climate target is absorbed into CRA by a calibration process, whereas it is an explicit constraint in CEA. The UNFCCC’s climate negotiations and many other organizations refer to climate targets in discussions, such as a 66% probability of keeping global warming below 2°C. Such a target includes an implicit assessment of the associated risk. This thesis explores, for the first time, the consequences of using such a target to calibrate CRA. We apply CRA to the climate problem including uncertainty and future learning and derive optimal mitigation paths. We calculate the value of future learning about the temperature response to be around 1/5 of total costs of climate protection (these costs include a monetized “risk” measure). The framework of CRA is based on expected utility maximization augmented by a risk-related term implemented by a risk metric which is subtracted from utility. A risk metric based on the probability of violating a temperature target leads to maximum emissions if it is learned that the temperature response is strong. We propose that such behavior is not in line with the preferences of the community supporting climate targets. Therefore, the thesis explores a risk metric based on the concept of degree years. We develop a method of attributing the value of information and the cost of climate protection to their respective sources. A source can either be a change in consumption (i.e. economy related) or a risk-related utility change. Furthermore, an attribution to different time steps and states of the world is also presented. We find that about 2/3 of the value of information originates from the economy before 2050. The economic value of information offsets around 1/3 of the economic cost of climate protection (i.e. the cost of mitigation).