America’s energy distribution: the top 1% of homes consume 4 times more electricity than average (and why it matters)
The novelist William Gibson is widely credited for noting that “the future is here, it’s just not evenly distributed.” That was back in the 1990s.
Since then, changes in how things are distributed in our society have been cause for both hope and concern.
On the one hand, improvements in semiconductor technology for things like computer chips and solar panels have made ideas like “distributed computing” and “distributed energy” a practical reality. On the other hand, recent economic turmoil has brought America’s uneven distribution of wealth and income under scrutiny.
With both of those threads in mind, we thought it would be interesting to consider how evenly energy consumption is distributed, and what that means for the viability of energy efficiency as a distributed resource.
So we’ve drawn upon our vast dataset of US electricity consumption to examine some key distribution-related questions:
- Among American households, what’s the breakdown between high, medium, and low electricity users?
- What drives differences in electricity consumption across households?
- Is electricity consumption as unequally distributed as the nation’s income?
- What does the nation’s electricity distribution suggest about how to go about saving energy at scale?
Starting with 25.8 million homes for which we had 2011 electricity usage information, we narrowed the dataset to 8.57 million American homes that we confirmed have natural-gas heating systems (the prevalent heating fuel across the US). In this way, we could make an apples-to-apples statistical comparison of different homes’ energy use.
We discovered that the top 1% of homes consume a full four times more electricity than average. Still, residential electricity usage ends up being much more evenly distributed than income. We’ll draw upon a couple key economic principles to explain why that is, and describe how the shape of the nation’s electricity distribution is a critical consideration in any large-scale effort to make America more energy efficient.
The Top 1% of households (by usage) consume 4% of residential electricity
Put another way, for every unit of electricity the average American home consumes, the top 1% of homes (by usage) are consuming four units. And the heaviest 10% percent of users are responsible for nearly a quarter of all residential electricity use.
What else does this mean?
- The top 1% of households (by usage) spend approximately $4,000 per year on electricity, while the average household’s yearly electric bill is around $1,000.
- Supplying electricity to each household in the top 1% entails greenhouse gas pollution from power plants equivalent to driving 5 gasoline-powered cars for a year. In comparison, the average household’s electric usage contributes pollution equivalent to 1.25 cars.
- 1 day of combined residential electricity usage across the top 1% of US households (comprising approximately 3.1 million people) is roughly equal to 1 year of total electricity consumption in the African country of Sierra Leone (a nation of 5.5 million people).
What’s driving the disparity in how Americans consume electricity?
Mega-Homes Mean Mega Usage (Usually)
Generally, the prime suspect in the search for causes of high energy use is large home size (i.e. square footage).
To investigate, we evaluated the relationship between electricity consumption and home size across more than 4.3 million residences in our dataset. We wanted to see how electricity usage of a mega home (i.e. among the largest 1% of homes) compares to that of an average-sized home in our dataset (approximately 1,600 square feet).
Our findings reveal a marked difference: an average mega-home uses 2.5 times more electricity each year than a typical home.
In this light, the chart below displays a predictable correlation: the larger the home, the higher the electric bill.
But while a general correlation between home size and energy consumption makes intuitive sense (e.g. more space to cool, more rooms with TVs), a deeper examination of the data reveals a complication: there can be substantial variation in electricity use among homes that have the same square footage.
To see how this is true, take a look at the line graph below: it shows that among households of the same square footage, it is not uncommon for energy usage to vary by as much as six times. This wide degree of variation suggests that while home size can serve as a rough predictor for usage, other factors – such as income, occupancy, climate, construction features, and especially behavior – are also important drivers.
Electricity usage is much more equally distributed than income
We’ve seen that the top 1% of electricity users consume 4% of the nation’s residential electricity. How should we view this in relation to income distribution in the US, where the top 1% of households take home nearly 20% of national income?
We shouldn’t be super surprised that household electricity consumption is more equally distributed than income. That’s because as a family’s income increases, their electricity consumption is likely to grow less than proportionally.
The principle at work here is a straightforward concept from Economics 101, called “diminishing marginal utility.” Basically, as we obtain more of a good, we value each additional unit less. For example, there’s a big difference between having no fridge and one fridge in a home. But there is much less incremental value of going from four fridges to five fridges. In other words, people’s demand for electricity has its limits, even as their income may grow.
Another reason that the distribution of electricity is more equal than income is that, although wealthier Americans are likely to live in larger homes, they are also more able and likely to invest in energy-efficiency improvements like insulation and triple-pane windows.
A nifty way to compare the distribution of electricity and income is to use a statistical measure called a Gini coefficient (named after the Italian sociologist who created it), which is a number that ranges between zero and one. A Gini coefficient of 1 indicates a totally unequal situation (e.g. a single household using all the electricity in the country), whereas lower values (i.e. closer to 0) represent a more equal distribution of resources.
For instance, the Gini coefficient for land ownership in the Middle-Eastern country Qatar is equal to 0.9, as the Emir of Qatar owns almost all the land in that country. By contrast, land ownership in Norway has a Gini coefficient of 0.18, reflecting greater equality in the distribution of land there.
The Gini coefficient for income in the US is around 0.47. Based on our dataset, the Gini for residential electricity consumption is 0.34 (see Methodology), further suggesting that it is much more equally distributed than income. This finding is consistent with other studies that have statistically examined the distribution of energy and utilities.
Why understanding energy distribution is important for realizing efficiency opportunities
We’ve seen that, at least when compared to income, residential electricity consumption in the US is relatively evenly distributed. The top 1% of US homes, although they use 4 times more electricity than average, only account for a sliver of overall national consumption.
Therein lies an important implication for how to go about reducing residential energy consumption: large-scale energy efficiency efforts (e.g. cutting energy waste in half by 2030) can’t exclusively focus on the very highest users, for the simple reason that such homes are in limited supply (e.g. only 4% of homes).
Instead, saving energy at scale requires a broad-based approach that works well for homes across the usage spectrum. And such approaches do exist, as evidenced by Opower’s own behavioral efficiency programs – which have enabled millions of households to save energy, regardless of their geographic location, home size, income segment, age, and initial level of consumption.
Energy efficiency initiatives that successfully reach large swathes of the population are likely to do more than save a lot of energy: they may also provide certain groups — such as seniors and low-income families — with much-needed relief from burdensome energy costs. For example, recent statistics show that elderly and needy American families routinely see 19-26% of their paycheck go toward utility bills, compared to just 4% for the median American household. This suggests that effective broad-based energy efficiency programs like Opower’s can be beneficial along multiple dimensions — environmental, social, and monetary.
While there are differences in how American homes use energy, there are often similarities in their ability and reasons to use less. To explore which savings opportunities are most relevant to you, your utility’s website or the new EnergySavers portal from the US Department of Energy are great places to start. Because the future is here…and it’s full of potential for energy efficiency.
Special thanks to David Moore, Jon Margolick, Chris Corcoran, Katie Dewitt, Jillian Cairns, Efrat Levush, Ashley Sudney, Tyler Curtis, and Arhan Gunel.
Follow @OpowerOutlier on Twitter
Appendix: Questions for the curious reader to consider
The foregoing analysis has left a few questions lingering in our heads:
1. Are we underestimating energy usage inequality by not accounting for multiple-home ownership?
We evaluated the distribution of electricity by treating every household as a distinct energy-consuming unit. Given that some families consume electricity across multiple homes, our analysis may be understating inequality. In some cases, multiple-home ownership may be nontrivial. The 2010 Census found that 3.5% of homes nationwide are for seasonal, recreational, or occasional use. And in some states, that fraction exceeds 10%.
2. Are the heaviest electricity users necessarily less energy efficient?
No. From an environmental and energy-efficiency perspective, high electricity consumption may not always be a bad thing. For example, a large multi-generational family living under one roof is likely to face a high electricity bill, but compared to a scenario where all family members live in separate homes, per capita energy consumption may be quite low.
Similarly, if you own a plug-in electric car, your annual electricity consumption will increase significantly. Consider an all-electric Nissan Leaf vehicle driven 12,000 miles per year: at a fuel economy of 34 kilowatt-hours (kWh) per 100 miles, it will require 4,080 kWh of electric charging— increasing an average home’s annual electricity usage by 40-50%. But, relative to driving an average gasoline-powered car, your environmental impact and overall energy costs will decrease.
3. How does the distribution of residential electricity compare to other measures of energy inequality?
To fully assess Americans’ relative energy and carbon footprints, it’s necessary to look beyond household electricity usage. A peek at the transportation sector suggests that Americans’ energy usage in the air and on the roads may be more unequal than in their homes.
For example, market research from the airline industry suggests that 17 million Americans (less than 6% of us) account for 58% of all flights taken by Americans. And the energy-related carbon emissions from flying are disproportionately large: the global warming pollution from one round-trip flight between San Francisco and New York (for a single customer) is equivalent to ~1 month of an average home’s electricity use. A recent New York Times analysis suggests that if you take five long flights a year, they may well account for three-quarters of the total pollution you create.
Day-to-day, the energy consumed for getting around town may exhibit a similarly unequal distribution, especially through a suburban versus urban lens: the EPA has calculated that the transportation energy use of a household in a typical suburban area is more than double that of a household in a transit-accessible area.
For the purposes of comparability, we narrowed our dataset to 8.57 million homes that have natural-gas heating systems. Analyzing gas-heat homes helps reduce the effect of exogenous/climate-related factors on our analysis. For example, Minnesota’s winter is much colder than northern California’s winter, but our analysis is to a significant degree insulated from this variation because the homes we considered do their heating with natural gas rather than electricity. This approach is especially important because heating represents a large fraction (42%) of home energy use.
Estimates for annual electricity costs are based on the average 2011 US retail electricity rate of $0.118/kWh. Note that average annual US household consumption is estimated at 11,496 kWh. Our average value (8,548 kWh) is lower largely because we have intentionally restricted our dataset to gas-heat households.
To compute a Gini coefficient for residential electricity consumption, we divided the 8.57 million households in our dataset into 100 groups of equal size, to determine percentiles of consumption. We computed each percentile group’s share of total electricity consumption, and then determined the cumulative share of consumption up to each percentile level. This data series allows for the construction of a Lorenz curve equation, L(X), which we integrated between 0 and 1 using a Riemann-sum approach across the 100 subintervals.
The resulting Gini coefficient (G) for residential electricity consumption, G = 1- 2 ∫ L(X) dx, was 0.34. This result parallels the 0.37 Gini coefficient for US residential electricity consumption computed by Jacobson, Milman, and Kammen (2004). It makes sense that our Gini coefficient is slightly lower (i.e. reflecting a more even distribution) than the literature’s existing estimate, as our analysis controls for heating type while Jacobson et al. appears not to.
Households in our dataset for this analysis are distributed across 23 states. Although geographically diverse, this dataset is not a perfectly representative sample of American households. However, we are confident that it is the largest dataset ever analyzed for the purpose of examining the distribution of US residential energy consumption, and that our analytical results are validly indicative of a national phenomenon.
Sierra Leone’s national electricity consumption was 111,600 MWh in 2009. The average annual electricity usage of a top-1% household in our data set is 33,654 kWh. Extrapolative multiplication of this usage by 1,147,614 households (i.e. 1% of the US’ 114,761,359 households) yields 38,621,802 MWh/year, or (dividing by 365) 105,813 MWh per day. This amount of electricity is approximately equivalent to Sierra Leone’s total annual consumption across all sectors.
The greenhouse gas pollution from one round-trip flight between San Francisco and New York is 675 kg CO2, according to the International Civil Aviation Organization’s carbon emissions calculator. 675 kg CO2 is tantamount to approximately 1 month of an average home’s electricity use, according to the EPA’s Greenhouse Gas Equivalencies Calculator.
Data Privacy: All data analyzed here are anonymous and treated in strict adherence to Opower’s Data Principles.
Author’s note: The analysis and commentary presented above solely reflect the views of the author(s) and do not reflect the views of Opower’s utility partners.