A new machine learning analysis has revealed the most effective climate policies out of 1,500 implemented worldwide over the last two decades. Some of the success stories – numbered at about 63 – involve rarely studied policies and unappreciated policy combinations.

“Our results provide a clear yet sobering perspective on the policy effort necessary for closing the remaining emissions gap of 23 billion tons carbon dioxide (CO₂) by 2023,” write the authors. To achieve the Paris Agreement’s climate targets, it is essential to know which climate policies work effectively at scale. However, despite the implementation of thousands of climate policies worldwide over the past two decades, there is little consensus on which policies are most effective. This emphasizes the need for a detailed global assessment.

Here, Annika Stechemesser and colleagues provide insights from an evaluation of 1,500 climate policies that were implemented across 41 countries between 1998 and 2022. They used data from the climate policy database from the Organization for Economic Cooperation and Development (OECD). They also used a machine learning-based extension of the standard difference-in-differences (DID) approach – a statistical technique used to estimate the causal effect of an intervention (such as a climate policy) by comparing changes in outcomes over time between a group exposed to the intervention and a group that is not.

Stechemesser et al. evaluated the impact of each policy on emissions across different regions and time periods. Out of the 1,500 climate policies, the authors identified only 63 policy interventions that reduced total emissions between 0.6 billion and 1.8 billion tons of CO₂.

According to the findings, combining various policy instruments is often more effective than using single measures, contradicting claims that policy mixes might be redundant. Effective policies typically involve a mix of subsidies and regulations that address different market failures.

Pricing mechanisms excel in industry and electricity sectors, while a combination of incentives and regulations benefits buildings and transportation. In developing countries, pricing was less effective, indicating that initial regulatory and subsidy measures might be necessary. While scaling up successful policies could close the emissions gap, it will require significant additional effort, including a need for further research and improved data.

IMAGE CREDIT: (ENTER NAMES)