(Machine) Learning from Schools about Energy Efficiency

Wednesdays@NICO | 12:00-1:00 PM, October 19, 2016 | Chambers Hall, Lower Level

Mar Reguant - Assistant Professor, Department of Economics, Weinberg College of Arts & Sciences, Northwestern University.



We study the impacts of energy efficiency investments at public K-12 schools in California. Our empirical setting offers two advantages. First, schools provide a rare laboratory to analyze energy efficiency as there are thousands of them, all pursuing very similar economic activities but exposed to different outdoor temperatures and with different existing infrastructures. Second, we make use of high frequency metering data—electricity consumption every fifteen minutes—to develop several approaches to estimating counterfactual energy consumption absent the energy efficiency investments. In particular, we use difference-in-difference approaches with rich sets of fixed effects. We also implement a novel machine-learning approach to predict counterfactual energy consumption at treated schools, and validate the approach with non-treated schools. Using both approaches, we find that the energy efficiency projects in our sample reduce electricity consumption between 2 to 4% on average, which can result in substantial savings to schools. We compare the estimates of the actual energy savings generated by measures to ex ante engineering estimates of savings, and, in ongoing work, compare the costs of installing measures to our estimates of the value of energy saved to come up with measure-specific cost-benefit metrics.


Mar Reguant is an Assistant Professor in the Department of Economics, Northwestern University.  Her research deals with the economics of energy, with an emphasis on electricity and the pollution associated with electricity generation.


Add this event to your calendar via Plan It Purple