Attributing Increases in Green Patenting to Disruptive Climate Policies

While the role of innovation in the green transition is widely acknowledged, ex-post evaluations of policy instruments and policy mixes are limited to a few headline policy instruments. In this paper, we use patent family data from 22 countries for the years 2000 to 2019, covering 95% of total green patenting activity, and combine a differences-in-differences model with a machine learning-based variable selection algorithm to detect structural breaks in the patent counts time series. We attribute positive and negative structural breaks to policies from the recently published OECD climate mitigation policy database and find that policy mixes that include pricing instruments are most effective in stimulating energy, solar and wind patenting. The results are robust to alternative specifications using standard synthetic control methods.