中文摘要：

考虑到气候变化及其经济影响的巨大不确定性,贝叶斯方法把参数看作随机变量,可以提高预测的精度.本文假设气候变化分别服从瘦尾指数分布和厚尾帕累托分布,选择伽玛分布为先验分布,并给出了气候变化的贝叶斯预测预报分布.此外,选择有界的CRRA效用函数,推导了气候变化的后验期望边际效用函数,研究发现,不论气候变化服从厚尾分布还是瘦尾分布,后验期望边际效用都是有界的,因此本文建议采取渐进式减排行动的“气候政策斜坡”.最后,通过模拟分析发现,当样本容量比较大时,先验分布的选择对贝叶斯后验边际效用函数的影响比较小,随着样本容量的减少,先验分布对贝叶斯后验期望边际效用函数的影响越来越大.

英文摘要：

There are numerous uncertainties in climate change and its economic impacts. Under Bayesian framework, the parameter is a random variable, which considers the structure uncertainty of climate change and improving the prediction accuracy. In this paper, suppose that the probability distributions for temperature change are the fat-tailed pareto distribution and the thin-tailed exponential distribution respectively. Combining data with gamma prior information, we derive the posterior predictive distribution for temperature change. Moreover, I argue that unbounded marginal utility makes little sense, and once we put a bound on marginal utility, expected marginal utility will be finite even if the distribution for outcomes is fat-tailed. Therefore, we support policies that would result in gradual GHG abatement. Finally, it is found that when the sample size is large, the prior affect to the posterior expected marginal utility is unclear. But the result is opposed as the sample is decrease.