A holistic comparison of European climate policies according to the data, not opinions

The question of how best to address climate change is fraught with uncertainty. Instead of relying on often conflicting expert opinions or incomparable assessments of isolated policies, we employ a data driven approach to holistically evaluate all climate policies at the same time. We also explicitly account for model uncertainty, which is ignored in most policy analysis. We test how effectively 23 different types of climate policies reduced emissions across 16 European countries between 2012 and 2023. We find that carbon taxes were the most effective tool to reduce economy-wide emissions, but sector-level results show significant heterogeneity. Regulations combined with taxes were more effective in the energy sector, subsidies were more effective in agriculture, subsidies complemented taxes in the industrial sector, and regulatory policies reduced emissions most in waste management. Policy mixes appear to be more effective than individual, isolated policies. We use Bayesian Model Averaging (BMA) as a natural method for dealing with model uncertainty.如何最优应对气候变化这一议题，始终充满诸多不确定性。相较于采信时常相互矛盾的专家观点，或是对单项孤立政策开展缺乏可比性的评估，我们采用数据驱动的研究方法，同步对所有气候政策展开整体性评估。此外，我们还明确考量了模型不确定性——这一要素在多数政策分析中往往遭到忽视。我们针对2012年至2023年间16个欧洲国家的23种不同类型气候政策的减排成效展开了检验。研究发现，碳税是降低全域经济碳排放的最有效工具，但分部门的减排结果则呈现出显著的异质性。在能源部门，监管措施与税收政策协同实施的效果更佳；农业部门的减排以补贴手段最为有效；工业部门中，补贴可作为税收政策的有效补充；而在废弃物管理领域，监管政策的减排成效最为显著。整体而言，政策组合相较于单项孤立的政策，往往能实现更优的减排效果。我们采用贝叶斯模型平均（Bayesian Model Averaging）作为处理模型不确定性的天然适配方法。