Answers are grounded in this paper and cite their sources.
Electric power disruptions impose substantial economic and societal costs, particularly in islanded power systems exposed to extreme weather, geographic isolation, and limited opportunities for grid support. Existing outage cost estimates are largely derived from mainland U.S. data and short-duration interruptions, limiting their applicability to prolonged outages in islanded systems. In addition, conventional approaches implicitly assume customer coping strategies, such as geographic relocation and spatial substitution, that are often unavailable in island environments, where constrained import capability, limited reserve sharing, and slower restoration may amplify the impacts of extended disruptions. This study presents the first empirical assessment of customer interruption costs in Puerto Rico for outages lasting one to 30 days. Using large-scale survey data and non-linear econometric modeling, we estimate duration-dependent customer damage functions for residential, non-residential, and public-sector customers. Results show that outage costs increase non-linearly with duration. Expressed as average daily costs, impacts initially decline as customers rely on available mitigation measures, but increase again during extended outages as mitigation resources become increasingly constrained and cumulative losses grow. Territory-wide economic impacts are estimated at approximately $1 billion, $5 billion, and $29 billion for 1-, 14-, and 30-day outages, respectively. The findings demonstrate that linear valuation approaches, including FEMA's Value of Unit Service assumptions, may substantially misrepresent the economic consequences of prolonged outages. While the estimated customer damage functions are specific to Puerto Rico, the methodological framework provides a foundation for improving outage valuation and resilience planning in other islanded systems and, with appropriate adaptation, interconnected mainland systems.
No takes yet. Share an insight, caveat, or question.
Baik et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: