Today, more than 70 carbon pricing schemes have been implemented around the globe, but their contributions to emissions reductions remains a subject of heated debate in science and policy. Here we assess the effectiveness of carbon pricing in reducing emissions using a rigorous, machine-learning assisted systematic review and meta-analysis. Based on 483 effect sizes extracted from 80 causal ex-post evaluations across 21 carbon pricing schemes, we find that introducing a carbon price has yielded immediate and substantial emission reductions for at least 17 of these policies, despite the low level of prices in most instances. Statistically significant emissions reductions range between –5% to –21% across the schemes (–4% to –15% after correcting for publication bias). Our study highlights critical evidence gaps with regard to dozens of unevaluated carbon pricing schemes and the price elasticity of emissions reductions. More rigorous synthesis of carbon pricing and other climate policies is required across a range of outcomes to advance our understanding of “what works” and accelerate learning on climate solutions in science and policy.

如今，全球已经实施了 70 多个碳定价方案，但它们对减排的贡献仍然是科学和政策领域激烈争论的话题。在这里，我们使用严格的机器学习辅助系统审查和荟萃分析来评估碳定价在减少排放方面的有效性。根据从 21 个碳定价方案的 80 项因果事后评估中提取的 483 个效应大小，我们发现，尽管大多数情况下价格水平较低，但引入碳价格至少为其中 17 项政策带来了立即和实质性的减排。整个计划的统计显着减排量在 –5% 至 –21% 之间（校正出版偏差后为 –4% 至 –15%）。我们的研究强调了数十个未经评估的碳定价方案和减排价格弹性的关键证据差距。需要对一系列成果进行更严格的碳定价和其他气候政策综合，以增进我们对“什么有效”的理解，并加速对科学和政策中的气候解决方案的学习。