Reliability
How reliable is electricity here?
Most utilities report how often their customers lose power and for how long, once a year. These numbers are useful, but they are averages with rules attached. Here is how to read them.
What we know and don't know
Power Grid Watch has not published its own reliability figures yet. The official yearly data exists, and we link to it below. When we checked it, some official summary figures did not agree with each other. We will not choose one quietly or smooth over the differences. We plan to publish reliability figures only when we can show where each number comes from and explain any disagreement.
The three main measures
SAIDI
System Average Interruption Duration Index. The average number of minutes without power per customer over a year. Higher means more time in the dark.
SAIFI
System Average Interruption Frequency Index. The average number of interruptions per customer over a year. Higher means outages happen more often.
CAIDI
Customer Average Interruption Duration Index. When an outage does happen, about how long it lasts on average. It is SAIDI divided by SAIFI.
A "customer" here means an account or meter, not a person. These measures usually count only sustained interruptions. Brief flickers are left out. Under the widely used IEEE method, that means interruptions of five minutes or less; other methods can draw the line differently.
Read together, SAIFI tells you how often a typical customer lost power, and SAIDI tells you how much time they spent without it in total. But "typical" hides a lot, as the next sections explain.
Major event days: included or excluded
A big hurricane or ice storm can cause more outage minutes in a few days than the rest of the year combined. To separate everyday performance from these rare events, utilities often report two versions:
- All events included. Everything that happened, storms and all. This is closer to what customers actually lived through.
- Major event days excluded. Days that crossed a statistical threshold are taken out. This shows day-to-day performance and is steadier from year to year.
Both are legitimate. They answer different questions. When you see a reliability number, check which version it is. Comparing one utility's "excluded" figure with another's "included" figure gives a misleading result. Some official data also has a version with "loss of supply" removed, which leaves out interruptions caused by problems on the larger transmission system.
An average is not your experience
A utility's average covers everyone it serves. Your own experience can be very different.
- A home at the end of a long rural line through the woods may lose power far more often than a home downtown served by the same utility.
- One bad storm in your area can make your year much worse than the utility's average, or a quiet year can make it better.
- Last year's average is not a forecast. It does not tell you how many outages you will have next year.
Use these numbers to understand a utility's general track record over several years, not to predict what will happen at your address.
Why fair comparisons are hard
It is tempting to rank utilities from best to worst. Be careful. Two utilities can only be compared fairly when their numbers were made the same way. Things to check:
- Method. Utilities report using different standards. Many follow the IEEE method, and others use their own or their state's definitions. Results under different methods are not directly comparable.
- Major events. Both figures should include major events, or both should exclude them.
- Year. Compare the same year or the same span of years. One storm season can swing the numbers.
- Area served. A dense city system and a large rural cooperative face very different conditions.
- Changes over time. Mergers, new reporting rules and better outage tracking can shift numbers without any real change in service.
Not every utility reports
Reliability reporting is not universal. Some smaller utilities, including some city-owned utilities and cooperatives, do not report these measures to the federal government, or report only some of them. A state or national figure covers only the utilities that reported, using the methods they used. A missing figure means "not reported." It does not mean zero outages.
Your state utility regulator may publish its own reliability reports, sometimes with different definitions. Look for it on your state page.
Where to find the official data
The U.S. Energy Information Administration (EIA) collects yearly reliability data from utilities on its Form EIA-861.
- EIA-861 annual electric power industry dataFederalDownloadable utility-level files, including a reliability file with SAIDI and SAIFI by reporting method, with and without major events. Detailed reliability data starts in 2013.
- EIA Electric Power Annual reliability tableFederalEIA's published summary of reliability measures by state. Read the table notes on method and major events before comparing.
Want to know how outages, grid strain and weather alerts differ? See Understand the grid. For what this site will publish and when, see Sources and coverage.