As the international community continues to fall short on reducing emissions to avoid disastrous impacts of climate change, some scientists have called for more research into solar geoengineering (SGE) as a potential temporary fix. Others, however, have adamantly rejected the notion of considering SGE in climate policy discussions. One prominent concern with considering SGE technologies to help manage climate change is the so-called “free driver” conjecture. The prediction is that among countries with different preferences for the level of SGE, the country that prefers the most will deploy levels higher than the global optimum. This paper tests the free-driver hypothesis experimentally under different conditions and institutions. We find that aggregate deployment of SGE is inefficiently high in all settings, but slightly less so when players are heterogeneous in endowments or when aggregate deployment is determined by a best-shot technology. Despite persistent inefficiencies in SGE deployment, free-driver behavior, on average, is less extreme than the theoretical predictions.

由于国际社会在减少排放以避免气候变化灾难性影响方面仍然存在不足，一些科学家呼吁对太阳能地球工程（SGE）进行更多研究，作为潜在的临时解决方案。然而，其他人坚决拒绝在气候政策讨论中考虑 SGE 的想法。考虑使用 SGE 技术来帮助管理气候变化的一个突出问题是所谓的“自由司机”猜想。预测是，在对 SGE 水平有不同偏好的国家中，最偏好的国家将部署高于全球最优水平的国家。本文在不同条件和制度下对自由驾驶假说进行了实验检验。我们发现，SGE 的聚合部署在所有设置中效率都很高，但当参与者的禀赋不同或聚合部署由最佳技术决定时，效率会稍低一些。尽管 SGE 部署效率持续低下，但平均而言，自由驾驶行为并没有理论预测那么极端。