The impact of adopting an energy information system on household energy consumption: a dynamic difference-in-differences approach

This study analyzes effects of adopting an energy information system on energy use based on the advanced metering infrastructure (AMI) in South Korea. Adopting the AMI was randomly assigned by the Korean Government. However, the adoption timing varies across household, making it difficult to apply an ordinary difference-in-difference (DID) approach to measure the impact, given the variation in adoption timing across households. We use a dynamic DID that is suitable for estimating a causal effect of a policy whose treatment timing differs depending upon the individual. Results indicate that AMI adoption led to a reduction in household electricity consumption, peaking within 18 months after adoption with a slight rebound effect. Our findings suggest that providing information on energy consumption and cost in the household sector can significantly reduce electricity use by as much as one quarter of average monthly consumption. Reducing electricity use in households can be achieved in several ways, including monetary incentives as well as personal and social feedback provided by adopting the AMI, implying the importance of a dissemination policy.