Characterizing and predicting household adaptations to electric power outages

Abstract This paper advances understanding of the implementation of household adaptations in response to electric power outages—who undertakes which ones and under what circumstances. Specifically, using household survey data from New York state and North Carolina, USA following a 2022 winter storm, we apply the Household Adaptations to Service Interruption (HASI) typology for the first time. We also use revealed and stated preference data to fit mixed logit models that predict the probability a household implements an adaptation as a function of the HASI categories and adaptation attributes. For the first time, the models include generalized versions that can be applied to any adaptation type. Results suggest the hierarchical categories and adaptation attributes (e.g., expensive) in the HASI typology distinguish among adaptations in a way that relates to how frequently they are implemented, as was the typology’s intention. In particular, adaptations that require relocation (e.g., going to a hotel or public shelter) are the least likely to be implemented. Those that work by reducing or delaying consumption, or by providing alternative ways to accomplish specific uses (e.g., candles for light) generally are more likely to be implemented the more favorable attributes they have (e.g., does not require time/effort, meets needs).