What did you set out to find in this study?
Louisiana and Puerto Rico have been hit especially hard by hurricanes in the past decade or so, and there is a lot of evidence that hurricanes are only going to get stronger and, sadly, more devastating. It is vitally important to assess how utility companies are meeting these challenges.
In our study, we asked several important questions:
- How efficient are utility companies in restoring customers’ power, given the need to do so on a limited budget?
- How can we compare events involving very different hurricanes?
Although utilities keep excellent records on these questions, these records are often confidential or proprietary, so it can be hard for utility consumers to understand how well their utilities are doing. In addition, a way to compare publicly available data on the efficiency of electric power restoration across multiple utilities and hurricanes could also help state policymakers in hurricane-prone regions make important decisions.
How did you find out the answers?
Research has known for a long time that higher winds over more land typically mean longer restoration times. Our analysis considered the variability in wind speed in different events, thus allowing any one event to be compared with any other event.
We applied a well-known approach called data envelopment analysis, or DEA, to measure the efficiency of restoring electric power to customers following a hurricane. We considered hurricane wind speed, customer-days without power, and costs of restoring power. Our data covered multiple utilities, including five utilities for Hurricane Katrina in Louisiana and one utility for Hurricane Maria in Puerto Rico. This model allowed us to estimate utilities’ performance in restoring power after moderate to extreme hurricanes and to compare performance of different utilities and events.
What did you find out?
We found that the peak number of customers (as a percentage of all customers of the utility) was meaningful in calculating efficiency of electric power restoration. In fact, the greater the proportion of customers without service, the greater the inefficiency observed. This means that, as more of a region is without power, power restoration is less efficient. This also means that efficiency decreases in more-rural areas, highlighting a potential inequity in the system.
Next, we considered storms for which there was a great deal of controversy about the time frame in which power was restored. For example, several utilities weathering Hurricanes Katrina and Maria were vilified for taking a long time to recover. When we considered the severity of the hurricanes, our results indicated that, for the utility hit by Hurricane Maria, restoration was weakly efficient. That is, even though restoration of power took months, the utility was more efficient in restoring power than several other utilities in our database. This shows the importance of considering the hurricane itself when understanding how utilities perform.
We also found that the efficiency of electric power restoration differs between utilities but is largely the same for a given utility hit by multiple hurricanes. The implication of this result is that efficiency may be within a utility’s control, as opposed to mostly affected by external characteristics such as regional infrastructure and environmental characteristics. Our results suggest that models should consider more of these firm-specific characteristics, such as their supply chains and logistics, relationships with suppliers and prices, and labor resources, to better predict outages and their durations.
Were there any surprises in your findings?
We were a bit surprised with the findings for Hurricane Maria. People on the island of Puerto Rico were without power for months. Given that, typically, after a hurricane, power is restored much faster than that, we would have thought that the utility in Puerto Rico were being inefficient. Surprisingly, we found that, when we accounted for the high wind speeds in the major of the utility’s coverage area, as well as the cost to recover, this utility actually had weakly efficient power restoration. We had similar findings for Hurricane Katrina.
What should electricity providers consider doing to improve power delivery during a storm event?
Clearly, electric utility providers are already burdened with significant hardships when a category 4 or 5 hurricane makes landfall. Many utilities are already efficient in restoring electric power to customers. Electricity providers could continue to improve by learning from each other: what efficient utilities are doing well and the struggles that other utilities face. Sharing best practices has the potential to get the lights back on faster, for years to come.