Can Energy Efficiency Help Avoid Blackouts?
Energy efficient air conditioners save electricity exactly when the grid is most strained.
Summer is around the corner in the Northern Hemisphere and air conditioning will soon be ramping up. Most U.S. electricity markets are summer peaking which means that the highest levels of electricity demand occur in the summer. Particularly with growing amounts of renewable energy on the grid, these hot summer days can be a serious challenge.
Take California, for example. Extreme temperatures last summer throughout the western United States led electricity demand to outstrip supply, leading to rolling outages on August 14th and 15th. For the summer of 2021, CAISO “continues to see potential challenges in meeting demand during extreme heat waves” and grid managers across the west are bracing for a “white knuckle summer”.
Today I want to ask whether energy-efficiency can help. Energy Institute alum Judson Boomhower and I published a paper last year called “Do Energy Efficiency Investments Deliver at the Right Time?” We find that Southern California Edison’s air conditioner rebate program saves a lot of energy. But, more importantly, we find that the program saves energy exactly at the right time, when savings are most valuable and the electric grid is stressed.
Event Study
Our research was made possible because of rich household-level hourly data on electricity consumption generously shared by SCE (though they had no role in the study). During our sample period, 6,000 households received a rebate for installing a new energy-efficient central air conditioner. We compare the change in their usage on hot days to changes for similar non-participating customers — a difference-in-differences analysis — to measure energy savings.
The figure below shows estimates from a standard “event study”. We use hourly data from July and August, and control for weather and other variables. This figure allows us to see how electricity consumption changed on average when a household installed a new energy-efficient air conditioner through the program.
Electricity consumption is essentially flat during the three years leading up to air conditioner replacement. Then, in the year that a household installs a new air conditioner — normalized to zero in the figure above — electricity consumption falls by 0.2 kilowatt hours per hour. Electricity consumption is then again essentially flat during the subsequent three years after replacement.
Bottom line: When a household installs a new air conditioner, their electricity consumption during July and August decreases by 0.2 kilowatt hours per hour. A typical LED lightbulb uses 8 watts, so this is equivalent to turning off 25 LEDs. These households use an average of 1.5 kilowatt hours per hour during July and August, so this is approximately a 13% decrease in total household electricity consumption.
Larger Savings on Hot Days
These are significant energy savings. But what’s especially exciting is that the savings are largest on hot days. The figure below plots estimated electricity savings against daily mean temperatures.
On cold and mild days there are essentially no energy savings. This makes sense because households don’t use air conditioning on those days. However, between 70 and 100 degrees, there is a steep, approximately linear relationship between temperature and energy savings. As expected, air conditioners work harder on hot days, so there are greater savings from a more energy-efficient unit. On the very hottest days, energy savings exceed 0.4 kilowatt hours per hour.
This is enough energy savings to matter. Replacing 500,000 air conditioners, for example, could reduce California’s peak demand by 200MW (i.e. 500,000 * 0.4 kW). In terms of balancing the grid, this has approximately the same impact as 200MW of battery storage or a 200MW natural gas peaker plant.
Implications for the Grid
Thus we find that energy savings from energy-efficient air conditioners are particularly large on hot days. This is, of course, exactly when the grid is stressed and energy savings are most valuable. We show in the paper that wholesale electricity prices and capacity prices are significantly higher during summer months.
Overall, we find that the “timing premium” for air conditioning in California is 40%. That is, we find that energy-efficient air conditioning is 40% more valuable than an investment that saves the same amount of energy in all hours of the year. In contrast, we find that timing premiums for energy-efficient lighting, clothes washers, and refrigerators are all near 0%, so these investments do not disproportionately deliver during times when energy is most valuable.
Do it With Demand
Which brings us back to the hot summer days around the corner. The California Independent System Operator is working hard to prepare the California market for this summer’s peak demand conditions. There are plans to increase supply-side capacity, bring battery storage projects online, and improve communication protocols to ensure that last summer’s outages aren’t repeated.
My colleagues Severin, Meredith, and Catherine have all written about the potential during peak conditions for the demand-side of the market to do more (here, here, and here). Severin asks “Why Don’t We Do it With Demand?” and last summer even took to twitter to beseech Californians to use less air conditioning.
I agree with them. How could I not? But in addition to more demand response, we should also be optimizing our energy-efficiency programs. California electric utilities spend $600+ million annually on energy efficiency programs. We need to make sure we are prioritizing investments that create the most economic value and help prevent blackouts.
Keep up with Energy Institute blogs, research, and events on Twitter @energyathaas.
Suggested citation: Davis, Lucas. “Can Energy Efficiency Help Avoid Blackouts?” Energy Institute Blog, UC Berkeley, May 24, 2021, https://energyathaas.wordpress.com/2021/05/24/can-energy-efficiency-help-avoid-blackouts/
For more details see Boomhower, Judson and Lucas Davis, “Do Energy Efficiency Investments Deliver at the Right Time?” American Economic Journal: Applied Economics, 2020, 12(1), 115-139.
Categories
Lucas Davis View All
Lucas Davis is the Jeffrey A. Jacobs Distinguished Professor in Business and Technology at the Haas School of Business at the University of California, Berkeley. He is a Faculty Affiliate at the Energy Institute at Haas, a coeditor at the American Economic Journal: Economic Policy, and a Research Associate at the National Bureau of Economic Research. He received a BA from Amherst College and a PhD in Economics from the University of Wisconsin. His research focuses on energy and environmental markets, and in particular, on electricity and natural gas regulation, pricing in competitive and non-competitive markets, and the economic and business impacts of environmental policy.
Energy Institute Blog 
While increasing the efficiency (EER not SEER) of air conditioners is certainly a good way to bring peak power down in a win-win way, I think adding distributed battery storage is even better. If you are home with that half ton of Tesla battery (in the car or Powerwall), you can get rid of a lot more peak power than by increasing the sweat factor. You could even get rid of all grid power between 3-9 on a hot, sunny day.
Max. Neither is EER the proper metric for determining the efficiency of an air conditioner for dry or wet climates. Sensible EER is a step toward the proper metric for dry climates and then it still leaves out the energy stored in the moisture on the coil. It is past time to fix this problem for dry climates and for wet ones.
In California a few very special contractors are installing air conditioners at higher airflows per ton to boost the Sensible EER. For others there is the Western Cooling Control ™ that harvests the energy on the coil. However designing and rating air conditioners differently for hot and wet climates is necessary to move the industry.
P.S. I have not heard anyone suggesting increasing the sweat factor.
John,
I’m not arguing that anything is the perfect metric, but wouldn’t you agree that EER is better than SEER when sorting out performance at peak in CA? Of course if one has undersized the system, then it is just running flat out and efficiency only impacts comfort not power.
P.S. If my house running at 78F dry bulb at peak as Sev would suggest, it is definately increasing the sweat factor and I’d hear about it.
I’m sorry, but someone has to say the following. The units of measurement used in the following paragraph are bananas:
Bottom line: When a household installs a new air conditioner, their electricity consumption during July and August decreases by 0.2 kilowatt hours per hour. A typical LED lightbulb uses 8 watts, so this is equivalent to turning off 25 LEDs. These households use an average of 1.5 kilowatt hours per hour during July and August, so this is approximately a 13% decrease in total household electricity consumption.
First of all, the watt is a rate. One watt is one joule per second. A joule is a unit of work (like, loosely speaking, a foot-pound). The watt (like horsepower) is the time rate of work, a.k.a. power. In fact, one horsepower (550 foot-pounds per second) is equivalent to 746 watts. So, from the paragraph above, 0.2 kilowatt hours per hour is 200 joules/sec*hours/hours. This makes no sense. Instead, you could say that a household that typically uses power at a rate of 1500 watts reduces its power consumption by 200 watts, a savings of 13.3 percent. Or you could say that a typical household in July and August that consumes 129.6 megajoules of electrical energy in a day consumes 17.28 megajoules less, a savings of 13.3 percent.
The units of measure matter because most scientific studies of climate change and world energy output use watts and joules. According to Wikipedia, in 2008, the world used electrical energy at an average rate of 2.31 terawatts (2.31*10^12).
Don’t be sorry Michael, it needed to be said. But “kilowatthours per hour” is more than a silly, redundant way of saying kilowatts. It’s the sign of someone with a poor understanding of the difference between power and energy – two distinct, scalar, physical quantities. Understanding these fundamentals should be required of anyone who claims to understand electricity pricing, consumption, or efficiency.
And “doing it with demand”? That’s an equally-bananas way of shifting blame for insufficient system reliability to customers. “It’s not that CAISO doesn’t have enough reliable electricity to serve customers – it’s that customers use too much electricity at the wrong times! Instead of keeping our dispatchable nuclear and gas plants open, let’s charge more for electricity when renewables aren’t available, and spend even more of customers’ money on air-conditioner rebate programs. They’ll never notice the cost has been added to their electricity rates.”
“My colleagues Severin, Meredith, and Catherine have all written about the potential during peak conditions for the demand-side of the market to do more. Severin asks ‘Why Don’t We Do it With Demand?’” and last summer even took to twitter to beseech Californians to use less air conditioning. I agree with them. How could I not?”
Here’s how, Lucas: imagine someone who isn’t as well off as you are, a family that already has the thermostat set to 85°. Should they turn it up to 95º? Should the kids wait until 10 PM to do their homework? Then they can open the windows, and their computer won’t be using expensive prime-time electricity…
For all the talk about social justice around here, it gets very quiet when someone brings up how the skyrocketing price of California electricity might be affecting those least able to afford it.
Excellent and well written blog. I would just add a statement after your note, in order to avoid any misinterpretations:
“In contrast, we find that timing premiums for energy-efficient lighting, clothes washers, and refrigerators are all near 0%, so these investments do not disproportionately deliver during times when energy is most valuable.”
I would add something like:
“Nevertheless, these investments are valuable throughout the year, no matter what time, since these energy efficiency investments reduce load at all hours.”
Ed
Nice work!
It seems to me that energy efficiency could help in other ways. A well insulated building or water heater, as examples, can be more flexible about when they draw power, in addition to using less during peaking extremes. And as the grid changes, and the most challenging days and times change, efficiency would still maintain its flexibility value. All of this, of course, integrates well with Severin Borenstein’s “Do it with Demand.”
Thank you, Lucas, for the excellent blog, and the comprehensive accompanying paper.
Indeed, not all not all saved kWh’s and reduced kW’s are of equal value. The example of energy efficient air-conditioning is particularly compelling, as you show. Not only are these measures creating the most valuable savings on the days which need these savings the most, but they are producing savings during the time of day (3-9 pm) where they are needed the most. Of course, HVAC in California is a great end-use to prove out the time sensitive value of energy efficiency savings. Ultimately the load curve of a particular end-use (and the drivers behind it) will drive how to properly incentivize energy efficiency for when it is most needed and valuable.
It is head scratching that California policy makers seem to continue to shy away from demand side management as a primary solution for extreme weather events and their impact on energy supply and demand and that so few energy efficiency programs have time-sensitive incentive designs. With the past two years of snowpack being 50-60% of normal in California, it is very likely that there will be a 40-50% reduction in available hydropower—-which comprised 16% of total California electricity supplies in 2016-2020. This summer could be even more brutal than the summer of 2020.
Thanks for the thought provoking article and the underlying research. I hope it will motivate us to infuse time-sensitivity into energy efficiency programs.
Great analysis! A further observation in the context of the “duck curve” is that by pre-cooling well-insulated homes, you may be able to further reduce the coincident peak of AC consumption. See https://rossbaldick.com/theres-more-than-one-way-to-deal-with-a-duck-curve/
Personally, I can’t believe that there is not much more concern about near-future grid demand, particularly with the growing numbers of a household’s biggest appliance: the electric car. Tesla’s most popular model, the Model 3, has a 500 kg battery pack, that’s 6-7 times more than the usual single occupant (the driver) is weighing. Energy professionals should pay way more attention to bringing an EV’s weight down. Vehicle mass not displaced = kWh not needed and more range out of 1 kWh.
This is great information that we need to hammer home to utilities, State politicians, Utility commissions, and Federal regulators.
HOWEVER what needs to be added is that air conditioners can save way more than the 0.2 kw or even the 0.4 kW if they are designed and installed to take advantage of the climate. Hot Dry Air Conditioners have been proven to save more peak energy than the existing “high efficiency” air conditioners that focus make no distinction between climates that are humid and those that are dry. This is a problem that utilities, and the Federal government have punted on. SEER is not a proper measure of how air conditioners will perform in dry or damp climates.
This is excellent work and much appreciated. A follow-on blog article should help identify communities and utilities that are implementing programs that focus on conservation and response to Flex Alerts either through upgrade programs (high efficiency HVAC), smart thermostats and the emerging class of smart electrical panels that provide granular or appliance level load shedding (SPAN.IO for example). My view is that relentless local efforts will have to step up and meet the challenges of California’s “duck” curve. Thank you again for your excellent research and insights. We have the data. We have the technology. Do we have the policy and leadership?