Energy systems in the United States are under increasing stress. Challenges tied to aging infrastructure and rising energy demand are being compounded by increasingly frequent and severe storms, floods, and wildfires. Yet despite growing exposure to climate-driven outages, utilities, insurers, investors, local governments, and communities still lack clear, consistent data on the economic and social impacts of power disruptions. This gap makes it harder to evaluate and justify investments in resilient solutions and technologies that could reduce energy waste, protect households and businesses, and avoid future losses.

This article reviews the methodologies and tools currently used to estimate outage costs — focusing on extreme weather-related losses — and highlights opportunities to strengthen the methodologies needed to support smarter planning and capital allocation toward more resilient power systems.

Power outages caused by extreme weather events cost the US economy billions of dollars every year and endanger the lives and livelihoods of the people affected by them. NOAA’s Billion-Dollar Disasters data shows that, between 2019 and 2023, the average direct cost of billion-dollar disasters was $120 billion per year. While costs include physical damage and some time-element losses such as business interruption, they exclude many broader or long-run economic losses. When electricity systems fail, it often disrupts services when they are needed most. Businesses close, transportation systems and supply chains stall, refrigerated goods and medicines spoil, and residents’ health, safety, and wellbeing suffer.

These direct and indirect costs accumulate quickly and hinder broader recovery. In some cases, these indirect impacts can exceed physical asset damages. For example, research shows that surveyed business interruption losses from Hurricanes Sandy and Harvey were 800–900 percent higher than property damages. If we conservatively assume that additional business interruption and related effects adds even 30–50 percent to the direct totals in the NOAA study — which are far below the observed ratios in Sandy and Harvey — that implies an additional $35–60 billion per year in US disaster losses that are not captured today. Without measuring these losses more systematically, decision-makers lack a full picture of the economic stakes of extreme weather, and lack the information needed to effectively manage risk and finance resilience.

These risks will only grow as electrification and rising demand from data centers and advanced manufacturing push grids toward new capacity and reliability limits while more severe weather is straining the grid. As a result, investments in resilient systems are increasingly important risk management strategies — from hardened substations and microgrids to backup generation, advanced sensors, and smart infrastructure that can isolate faults and keep critical loads online. These resilience technologies offer a multitude of benefits to the grid system, but more accurate data on outage-related losses can better inform return on investment and tip the scale in favor of these solutions.

The benefits of resilience investments remain difficult to quantify with current data and metrics. For example, existing metrics in the insurance industry tend to focus on physical damage to insured assets, while many outage-related losses — such as from business disruption, lost inventory, and supply chain impacts — fall outside and are inconsistently measured. One of the main reasons for this is the lack of standardized post-event data collection. Many methodologies and models rely on survey data, but the inputs, sample populations, and outputs vary significantly across regions, events, and user groups (government agencies, insurers, utilities, etc.). This inconsistency makes it hard to account for impacts, replicate or extrapolate metrics, and assess likely future costs. As a result, decision-makers cannot fully value and incentivize preventative actions that could safeguard communities and save lives.

One available metric, the Value of Lost Load (VoLL), estimates the monetary value of the incremental energy not supplied, typically calculated from surveying customers’ willingness to pay to avoid outages or restore power. For insurers, VoLL can serve as an input to exposure pricing and catastrophe loss modeling. However, this approach has several limitations:

• Reliance on survey responses introduces methodological biases. VoLL results are highly time- and region-specific, and extrapolating VoLL figures can mask inequalities and variation in important socioeconomic conditions, cognitive biases, and contextual factors such as income, customer class, business type, and regional conditions.
• VoLL does not capture system-wide effects of power outages, such as supply chain disruptions for businesses outside the outage zone.
• VoLL is static and backward-looking, and does not account for non-linear compounding losses over time — such as food spoilage and restocking costs — which typically increase after critical windows (e.g., four to eight hours for temperature-sensitive items).

Our research identified two tools that use the VoLL concept as their foundation to create more comprehensive and realistic assessments of outage-related losses:

1. The Interruption Cost Estimate (ICE) model, developed by the Lawrence Berkeley National Laboratory (LBNL), provides a survey data-driven platform to quantify utility customer-level outage costs by customer class such as residential, commercial and industrial, duration, and operating context. It is well suited for assessing localized or short duration outages.
2. The Power Outage Economics Tool (POET), developed by LBNL and University of Southern California, models longer-duration and widespread outages using behavioral survey data coupled with a computable general equilibrium (CGE) model to simulate economy-wide “disequilibrium” effects such as evacuation, supply-chain disruption, and recovery costs.

The Table below covers the pros and cons of these different tools.

While the ICE model is useful for assessing short-duration and localized outages across the United States, POET offers macro-level insight into longer-duration disruptions but requires additional regional calibration for geographical flexibility. Pairing the two models could provide a more comprehensive input for insurers and investors seeking to model and price risk, allocate capital, and understand the value of resilience technologies. However, both models rely heavily on survey-based historical data that may not represent future hazard conditions or evolving grid behavior, pointing to a need for more forward-looking and consistent approaches to quantifying outage losses and supporting resilience investment decisions.

Strengthening methods for quantifying power system losses will make results more actionable for utilities, insurers, and investors. Opportunities include:

1. Use forward-looking scenarios that reflect evolving technologies and hazards.

Current methods rely primarily on historical outage data and survey responses, often in response to regulatory constraints on pricing model inputs. However, future conditions — including load growth and changing hazard frequency — will look different, so traditional approaches may overstate risk in some places while underestimating it in others. Scenario-based methods could better incorporate technology adoption, policy changes, shifting load profiles, and how resilience investments affect outage the probability, duration, cost, customer class, and distribution of outages.

2. Improve spatial and sectoral granularity.

Losses are unevenly distributed across industries and communities. High-resolution outage maps, hazard overlays, and county-level economic parameters can identify hot spots and guide investors and communities toward targeted grid-hardening technology investments. For insurers, this could support better risk segmentation and portfolio diversification.

3. Model compounding losses across the full outage-duration curve

Economic losses from outages increase non-linearly. As outages stretch on, business impacts grow more severe. Mapping marginal cost curves using empirical and behavioral data could improve catastrophe models and improve return on investment (ROI) analysis for resilience technologies that shorten outages.

4. Integrate income and equity considerations into willingness-to-pay metrics.

Traditional VoLL surveys treat customer classes uniformly, overlooking how income, wealth, energy burden, and community resilience influence outage experience and willingness to pay for reliability. Including socioeconomic and demographic indicators alongside adaptive response capacities (such as generator access, evacuation likelihood, community resource centers etc.) can help link social vulnerability and economic loss. Measuring relative loss as a share of income or revenue and introducing income-weighted VoLL distributions could help underwriters or policymakers identify inequities and design more targeted and affordable resilience solutions.

While methods and tools like VoLL, ICE and POET help quantify the cost of power outages, today’s rapidly changing climate and grid landscape require more forward-looking, granular, non-linear approaches that reflect how outages affect households and communities differently. Improving outage-loss methodologies will help insurers, regulators, communities, and utilities to more accurately value resilience solutions — and direct investment where it delivers the greatest benefit. Better data will not eliminate climate risk, but it can help communities prepare for extreme weather, reduce exposure to prolonged outages, and build a power system capable of withstanding future shocks.

RMI is working to close market gaps around investment in resilience technologies. If you’d like to learn more about what we’re doing in this space, please get in touch via info@climatealignment.org.