Estimating intensive and extensive margin adaptation to climate change from consumption data∗

The widespread rollout of air conditioning has resulted in massive positive impacts on human health in both developing and developed countries over the past century. A warming world will further increase demand for cooling in climates both through increased operation of existing air conditioners and the adoption of new ones. This paper provides an empirical approach to separately quantifying the intensive and extensive margin response without having to make explicit assumptions about the current day or future penetration of air conditioners. Exploiting a panel dataset containing close to one billion electricity bills for California, we estimate the temperature response of residential consumption at the zip code level. Using the most recent set of IPCC climate models, we simulate future consumption by letting households vary consumption along their current response function as well as letting the response function change using an econometrically estimated second stage. We show that overall increase in residential electricity consumption is modest, yet failing to account for extensive margin adjustment leads to drastic underestimate of future impacts.