Artificial intelligence analyzed 1,500 policies to reduce emissions. These were successful.

Researchers used machine learning to analyze around 1,500 climate policies and identify those that dramatically reduced carbon dioxide emissions. Your study, which today inSciencepublished found that policies that combine several instruments are more effective at reducing emissions than stand-alone measures ¹.

The analysis identified 63 interventions in 35 countries that resulted in significant emissions reductions, reducing them by an average of 19%. Most reductions were linked to two or more policies. Together, the 63 policies reduced emissions by 0.6 to 1.8 gigatons (Gt) of carbon dioxide equivalent.

Getting the right mix of policies is more important than using many different policies, says Annika Stechemesser, co-author and researcher at the Potsdam Institute for Climate Impact Research in Germany. For example, in the UK, the phase-out of coal-fired power plants worked because it was used in combination with pricing mechanisms such as a floor price for carbon, while in Norway the ban on internal combustion engines was most effective when combined with a pricing incentive that made electric cars cheaper.

“To my knowledge, it is a unique study that provides such a global assessment,” says Jan Minx, an environmental economist at the Mercator Research Institute for Global Commons and Climate Change in Berlin.

Path to emissions reductions

As part of the analysis, Stechemesser and her colleagues used a database of 1,500 climate policies implemented between 1998 and 2022 in 41 countries, including the three largest greenhouse gas emitters worldwide: China, the United States and India. The policies fell into 48 categories, from emissions trading schemes to fossil fuel subsidy reform.

“Previous assessments have typically focused on a limited number of prominent policies in selected countries and overlooked the hundreds of other measures,” says Stechemesser.

The authors combined machine learning with a statistical analytical approach to identify large emissions reductions in four high-emitting sectors - buildings, power, industry and transport. They compared the results with the policies in the database to assess which policies and combinations of policies led to the largest reductions in emissions.

“This is a pretty clever method,” says Zheng Saina, who has analyzed global climate policies at Southeast University in Nanjing, China. The conventional way would have been to review the many policies and select the important ones, but this approach is subjective and laborious, she adds. "The authors instead used machine learning to detect large changes in emissions. This is more objective."

Right mix

The results showed that certain policy combinations worked better in certain sectors and economies. In terms of reducing emissions associated with electricity generation, pricing measures such as energy taxes have been particularly effective in highly developed countries, but less so in low- and middle-income countries.

In the construction sector, policy mixes that phase out and ban emissions-generating activities doubled reductions compared to implementing these policies individually.

Taxation was the only policy that achieved almost equal or greater emissions reductions in all four sectors as a stand-alone policy, as opposed to a policy mix.

Minx says the study's AI-powered approach allowed researchers for the first time to assess the effectiveness of a large number of climate policies from a global inventory of emissions covering different countries and sectors.

For other researchers, the paper is alarming. “This study warns countries around the world that their climate policies have had very limited impact so far,” said Xu Chi, an ecologist at Nanjing University. “Existing policies need to be reviewed and changes made,” Xu added.

The annual emissions in the world By 2030, emissions are expected to be 15 Gt carbon dioxide equivalents higher than would be needed to keep global warming to less than 2°C above pre-industrial levels, according to the United Nations.