Skip to content

How Much Electricity Consumption Is Too Little?

Energy efficiency world views collide.

There are two dueling, strongly held, views on the definition of energy efficiency. The idea of energy efficiency, at least to economists, is to overcome market failures that can lead to people consuming energy even when the full societal costs of the energy exceed their benefits. An alternative perspective also pervades policy circles. This perspective appears to be that people should just use less energy, period. To economists, this view is a perversion of the notion of energy efficiency. Energy efficiency should be about the efficient use of energy, not the non-use of energy.

One policy arena where these dueling views are colliding is electricity rate design. About a month ago I participated in a workshop at SMUD concerning a proposal to add a monthly fixed surcharge to homes that newly add rooftop solar. The logic behind the proposal was a familiar one to readers of this blog site: many fixed utility distribution costs are recovered in variable, per kWh rates, and solar homes avoid paying for those fixed costs when they generate their own electricity but stay connected to the system. For SMUD, this is a financial concern: how to equitably recover the fixed costs of their infrastructure? But there is a larger societal issue that gets overlooked when we focus too much on just the financial viability of a distribution utility. The larger question is: exactly what kind of behavior do we want to discourage, or encourage, from consumers when we set electricity prices, and why?

LightBulbDollarSign                                              Source:

The SMUD proposal was, not surprisingly, roundly criticized and opposed by solar trade groups. Somewhat frustrating, but not surprising, was the vocal opposition from and other environmental groups as well. My frustration stems from my belief that we have a much better chance at combating climate change if we direct our scarce resources away from rooftop solar toward more cost-effective solutions like grid-scale solar. What was surprising to me, however, was how the conversation turned to the wisdom, even the ethics, of SMUD’s general tariff structure, which has a higher monthly fixed charge, and lower variable prices, than most other California utilities.

Electricity Prices: How High is Too High?

The general tone of this part of the discussion was that it was socially irresponsible for SMUD to charge a lower variable price of electricity, because it would encourage people to use more electricity. The argument is often extended to support steeply rising increasing-block rate structures, such as exist in much of California, on the grounds that higher prices encourage conservation (i.e., discourage electricity use). This begs a question that I wish I had asked at the time, but didn’t. If lower electricity prices are “bad”, and by implication higher electricity prices “good”, then how high is too high?

Economists have a framework for answering this question. It is called marginal cost. Because we, as a society, are worried about climate change and other environmental costs, we should include those in marginal cost as well. That’s called social marginal cost (the cost of producing the electricity plus the external damages done by it). Ideally marginal prices would be set at social marginal cost, so that when a consumer turns on a light bulb, or charges their electric vehicle, the incremental amount they pay matches the incremental cost they impose on society.

In a previous blog, Severin Borenstein talked about work we have been doing estimating the social marginal cost of electricity around the US, and comparing it to the marginal ($/kWh) price faced by residential customers. These social costs reflect the marginal wholesale cost of electricity and researchers’ estimates of the environmental costs of generation. There is a striking diversity across the US in the relationship between marginal prices and social marginal cost, but one fact that stands out is that marginal prices in California are among the highest in the country even though our marginal cost of electricity is among the cheapest and cleanest in the country.


Energy Efficiency: Dueling Definitions

Again, the idea of energy efficiency—at least as an economic concept—is to overcome market failures that lead to people consuming energy even when the costs exceeded their benefits. There are two types of market failures, broadly speaking: either the energy price is “wrong” or the price is right but consumers don’t respond correctly to it.

The first failure is usually linked to externalities, like climate change, whose costs may not appear in the energy price, leading consumers to consume “too much” because the price, lacking the environmental cost, is “too low.” The second failure can be attributed to a myriad of institutional breakdowns, like landlords who don’t have an incentive to invest in efficiency for tenants, or behavioral factors such as consumers misunderstanding or not wanting to spend the time understanding their electricity prices.

But a corollary to the economic view of energy efficiency is that if true social costs are low, it’s OK to consume more. In fact, it’s a bad idea, even wasteful, to devote scarce resources to reducing consumption if the costs of those investments exceed the benefits provided. This is where electricity pricing comes into the picture. If we set electricity prices well above the costs of serving customers, we are encouraging consumers to take steps to reduce electricity consumption when the electricity cost savings outweigh the investment costs to the customer, but not to society. Rational consumers will reduce their electricity consumption (or install rooftop solar) based upon these price distortions. Indeed, this is exactly what my colleagues at UC Davis, Kevin Novan and Aaron Smith find in their 2016 paper, The Incentive to Over-invest in Energy Efficiency. They study air conditioner replacements in Sacramento and estimate that while the AC investments save about $11.50 per month in avoided social costs, they save the consumers who make the investments about $26.50 per month because of SMUD’s rate structure where marginal prices exceed marginal social cost.

Considering the fact that marginal electricity prices are more than double the marginal cost of energy (including externalities) in much of California, any behavioral reluctance on the part of consumers to invest in energy efficiency could actually improve rather than reduce total benefits. The customer’s cost-benefit test for saving money needs to be passed by a wide margin before energy efficiency makes economic sense in places like California. Unfortunately, as the above map illustrates, as a country, we are devoting funds to overcoming customer inertia in all the wrong places. Energy efficiency program expenditures are highest in states with high prices and clean electricity and low to non-existent in the states where electricity is dirty and more expensive.

Energy Efficiency Program Expenditures per Customer by State



Less is more, no matter what?

One can argue with the specific numbers, but the general principle of marginal cost pricing is pretty compelling. If consumers want to consume energy and are willing to pay the societal cost to provide it, their consumption creates a benefit that economists call welfare. If prices rise well above social marginal cost, then we are inefficiently discouraging the use of electricity. Yet there are some who are not persuaded. They appear to think people should use less energy, period, regardless of whether costs are low or costs are high.

The inconsistency in the “less is more, no matter what” view of energy efficiency is becoming more obvious as the grid gets cleaner and we are hoping to electrify other sectors, like transportation and home heating. The former trend means that the social marginal cost is getting cheaper, even while the total cost of providing electricity is getting more expensive (including fixed costs like renewable capacity, the transmission system, etc.). In fact, there are times and places where electricity is effectively costless. Do we really want to discourage consumption, even the charging of EVs, through high prices during times like these?

It is interesting that some opponents of rate structures like monthly fixed charges also support increased time-varying prices. Support for the latter implies a recognition that when costs are low it’s OK to encourage consumption. However, opposition to fixed charges when marginal prices are so far in excess of costs implies a rejection of the same principles of marginal cost pricing that would lead one to favor time varying prices.

The other area where the view of “less electricity is better” runs into trouble is when we consider what the alternatives to electricity consumption are. Those alternatives are increasingly gasoline or natural gas. If marginal electricity is clean and cheap, we want people to shift from gasoline to electricity to power transportation. But high electricity prices clearly undermine that transition.

So, what exactly are we trying to achieve with electricity prices? Once we deviate from the principle of marginal cost pricing, we risk making moral judgments about how other people perceive the benefits of consuming energy. Now I’m not against doing that. I quite enjoy judging other people, in fact. But it’s a wobbly foundation to base public policy upon.

As a policy community we need to come to some common understanding about what energy efficiency is and should be. This means recognizing that consuming energy is not, in and of itself, a bad thing. Many fantastic goods and services are made and enjoyed using energy. What is “bad” is wasting money and polluting the environment. Energy efficiency efforts should be focused on truly wasteful, inefficient consumption. When we place the marginal price of electricity excessively high, we are throwing out the good consumption with the bad and making the achievement of our ultimate goal of a prosperous, clean-energy society harder to reach.

Keep up with Energy Institute blogs, research, and events on Twitter @energyathaas.

Suggested citation: Bushnell, James. “How Much Electricity Consumption Is Too Little?” Energy Institute Blog, UC Berkeley, June 3, 2019,

26 thoughts on “How Much Electricity Consumption Is Too Little? Leave a comment

  1. “Now I’m not against doing that. I quite enjoy judging other people, in fact.”

    So you are sayin you are God??

  2. I’m also curious about the calculation of the marginal cost. Some questions I have: Does this include long range costs? To illustrate, I believe I read that a pound of carbon dioxide emitted today will remain in the atmosphere for several hundreds of years. Also, does your analysis include the costs for global warming cited in the Fourth National Climate Assessment, published in 2018, prorated for contributions to emissions related to electricity generation? Finally, are marginal costs computed differently based on generation source, so that a kw of coal generated electricity has a higher marginal cost than pv generated? In some areas consumers can choose to buy “green energy”, is their marginal cost reflected in the analysis?


  3. I would also add to “Energy efficiency efforts should be focused on truly wasteful, inefficient consumption.” to include loads like motors and air conditioners that are not operating at unity power factor. I would also add that injection of reactive power is also not energy efficient and affects system capacity. Borrowing from a Sandia National Labs 2007 report here two exerpts, “The goal of the power grid is to distribute electricity from the source to the load with a power factor of 1.” and “In this example, the clear tradeoff is between a physical capacitor bank and a power flow control device. The capacitor is simple, reliable, and fixed capacity. The control device is flexible, adaptable, and presently more expensive and possibly less reliable. Also, there are a couple of major issues with this approach that should be addressed in future studies: 1. Does it actually get closer to a power factor of 1 since one is expending exergy to run power electronics?”

    If we begin to use almost exergy free power electronics, we can eliminate truly wasteful, inefficient electricity consumption using zero reactive power from the electrical grid and the losses associated with transporting reactive power.

  4. “My frustration stems from my belief that we have a much better chance at combating climate change if we direct our scarce resources away from rooftop solar toward more cost-effective solutions like grid-scale solar. ”

    While this statement isn’t central to your post, I can’t leave it unchallenged. The fact is then when we add in true marginal transmission costs and acknowledge that scale economies compressing for solar, rooftop (at least for commercial/multi family size) approaches parity with utility-scale. When we add in the avoided local environmental impacts, diversification of sources including potential localized storage, and most importantly the devolvement of political power from large utility companies, the benefits which aren’t immediately calculable in a spreadsheet weigh toward distributed resources, not large plants (which got us into this mess in the first place).

    As for your measure of “marginal cost” I’m unclear on what you used. If you followed the PCIA exit fee proceedings, you know that there is not a consensus in California on this metric. As I’ve posted here numerous times, the correct metric must acknowledge that key market and regulatory failures cause short run and long run marginal costs to diverge and they will never converge. (CAISO market prices have not produced sufficient revenues to fund a new generation plant since 2001–I think think that should count as “long run”.) If you use a correct definition of marginal costs, what does this graphic look like?

    • I agree with much of this mcubedecon, but I think the decentral bv central can be argued on narrow consideration of costs only as well – see my post.

    • @mcubedecon: what values are you using for

      1) “true marginal transmission costs”
      2) “scale economies compressing for [rooftop] solar” (which don’t compress more for utility-scale solar)
      3) “avoided local environmental impacts” (which aren’t avoided more efficiently by a solar farm)
      4) “diversification of sources including potential localized storage”
      5) “the devolvement of political power from large utility companies”

      I understand the benefits of #4-5 might not be immediately calculable in a spreadsheet, but without individual benefits/costs it’s impossible to arrive at an aggregate. A casual reader might even believe they’re imaginary.

      • I have posted these in response to other blogs here, and I don’t have the time to dig these up again in response to a comment. But I’ll summarize quickly — marginal transmission costs appear to be about $80/MWH based on CAISO analysis; I conducted the cost of generation studies for the CEC from 2001-2015 and analyzed the compressing scale economies; I worked on the DRECP for the CEC as well as solar siting cases around the state and have cataloged the environmental impacts of solar farms in desert locations.

        I’m not conducting an alternative analysis because I don’t have funding from an outside source to do so. I’m bringing to bear the data and analyses that I have done in the past to point out the flaws in the analysis presented here.

        • Ah, I see. And readers have the time to dig them up for you?
          Doesn’t work that way. Without references, your assertions are worth the opinion of “someone else on the internet” – not much.

          • I provided my sources – CAISO reports (and I’ll add testimony filed by Tom Beach in SCE’s 2018 GRC Phase II on transmission costs); the CEC Cost of Generation Report (and I’ll add analysis by the Institute for Local Self Reliance); the joint DRECP studies conducted by state and federal agencies. I am not providing a full blown cost analysis, because I’m not being paid to spend the many hours to conduct that analysis. I’m pointing out the shortcomings in the analysis presented in this blog post, and I’ve given sources from the authors to appropriately revise their analysis.

            You have failed to provide supporting documentation for your assertions in post here, so you have little standing to try to make this claim.

          • @mcubedecon, links please. Among thousands of “CAISO reports” and “state and federal agencies”, no one has the time to try to figure out to which one(s) you’re referring.

            In particular, I’m interested how you’re determining benefits/costs for “diversification of sources including potential localized storage” and “the devolvement of political power from large utility companies” – they sound like ideological mumbo-jumbo.

            If you’d like a reference to any of the assertions I’ve made all you have to do is ask.

          • You haven’t offered any links for your assertions presented on this blog. This is not a legal or regulatory proceeding and I am not being paid to satisfy your needs for information. I’ve given you the sources and you can go find them. (Start with Tom Beach’s testimony to find the CAISO data.) I’m not your servant and you cannot demand more of my resources without paying me. I’ve presented my counterpoints with references (more than what you’ve presented in the past) and that is sufficient until we are in a forum where we face decision makers.

            As to the last two benefits, you will need to familiarize yourself with those topics yourself. There is substantial information on the benefits of source diversification. Just the very basis of the N-1 reliability standard illustrates how relying on 10 100MW plants is preferable to relying on one 1000MW plant. Unless you show that you actually have interest in learning about those rather than just trying to set me up to shoot them down, I”m not going to bother on a blog to do your work for you.

            On the political benefits of resource centralization, substantial research exists on how concentrated economic and political power go together. That principle is the very motivation of the Sherman Anit-trust Act. And that President Peevey was forced to retire from the CPUC highlights how the CPUC has been captured by the IOUs and unable to resist that concentrated political power.

          • @mcubedecon, I guess I can take from your rather long-winded explanation you’re unable to support your criticism with references – I suspected as much. That’s you’re job, not mine or anyone else’s. The idea someone else should pay you to support your own position, and that you’re entitled to be my teacher are – rich. I’ve never seen such unmitigated arrogance on a discussion board before, and given your unfamiliarity with scientific discourse and the history of U.S. electricity it would be a waste of money anyway.

            Here’s something about which I can teach you – The Sherman Antitrust Act. It had nothing to do with “resource centralization” but eliminating anti-competitive behavior. The original electrical utilities, natural monopolies which replaced the original “distributed energy resources” of the 1880s, were allowed to do so for reasons which had nothing to do with eliminating competition. They made public electricity available to everyone in society (unlike solar shingles and Tesla Powerwalls) by eliminating multiple, duplicative distribution networks. They standardized the plugs, sockets, voltage, and alternating current Americans use today, at a time when “distributed energy resources” with differing standards were responsible for disastrous fires on a daily basis. When they were successully regulated under the Public Utility Holding Company Act of 1935 (PUHCA), it marked the start of 70 years of reliable, low-cost electricity in America that became a standard for the world.

            From the start utilities, like any companies, chafed at being under the watchful eye of the Securities and Exchange Commission, and in the 1980s were able to secure exemptions from PUHCA. Finally, in 2005, it was repealed by the petroleum industry in league with wind and solar entrepreneurs. Since the Glorious Distributed Energy Revolution the price of electricity has risen faster than ever in history (16% in the last decade), and Americans have among the highest per-capita carbon emissions in the world. Kind of what you’d expect when a well-regulated natural monopoly is privatized, isn’t it – when free-market capitalism is entrusted with providing a social need (electricity) and solving a social problem (climate change)?

            If renewables evangelists search for a source of the failures in U.S. energy and environmental policy, they need only look in the mirror.

          • I have supplied references, rather specific, just not the weblinks that you so desire to ease your own workload. This is a blog, not a legal or regulatory proceeding. We are not building an evidentiary record that will be used by a decision making authority. I have met my burden of proof for this forum. You are not the judge of what is sufficient–the others in the audience can judge that. I will not kowtow to your unreasonable personal demands.

            You completely missed my point on the Sherman Anti trust act–it was put in place to blunt the political power of the well established
            railroad, steel and oil companies. Electric utilities were just a fledgling industry that wasn’t important at the time. It had nothing to do with weighing against resource centralization.

            But then Sam Insull figured out that he could protect the monopoly status of his Edison companies by setting up state commissions, and he played the game well. The undoing of the regulatory framework began in the 1980s with the formation of utility holding companies. The 2005 action was just an acknowledgement of the changed circumstances.

            As for rising electricity rates, you’re wrong. From 1973 to 1982, rates rose 55% in constant dollars (much worse nominal). That event was the key evidence that the existing regulatory framework had failed and that utilities could act with no constraints on passing along costs. At the core of that was the massive failure in nuclear cost overruns. The result was the series of deregulation steps since 1978.

            And the most resistant to addressing the social consequences of the climate change threat have been the remaining regulated utilities across the nation. We would have made even less progress than we have if we hadn’t had some degree of deregulation as the utilities would have acted to protect their status quo assets based on the evidence of the lack of action in those states where conventional regulation remains.

  5. Thank you for an interesting article. I do think however that the aversion to distributed rooftop solar relative to centrally-produced large scale solar might be considered in further. I suspect the relative economics of these two applications depends on many factors. In Australia, rooftop solar is installed at $1 per watt, about the same as ground-based large scale, roughly, and yields from rooftop solar are not meaningfully different from fixed axis ground-based solar. Rooftop solar operating costs are absorbed by the household and rooftop does not incur the many other operational costs associated with large scale ground-based solar. So the author’s concerns about allocative inefficiency in California are not relevant in Australia. It is also somewhat surprising that rooftop solar is so much dearer than ground based in California. Is this really the case and if so, why?

  6. Thanks for daring to elevate economics above political correctness. Arrow asserted that “it is natural in academic life to lean against the wind,” but what is natural is not common.
    In the Philippines, electricity prices are far higher than their ASEAN neighbors, Thailand, Malaysia, and Indonesia, and their consumption per capita less than half. This results not only in foregone consumer benefits but in “premature deindustrialization” (Rodrik) more radical than in other developing countries. Since growth externalities tend to be concentrated in manufacturing, this means that dynamic welfare losses are particularly high.
    Ref: “The Role of Power Prices in Structural Transformation”

  7. “There is striking diversity across the US in the relationship between marginal prices and social marginal cost, but one fact that stands out is that marginal prices in California are among the highest in the country even though our marginal cost of electricity is among the cheapest and cleanest in the country.”

    James, I’m curious how you’re calculating the marginal cost of electricity. Currently our state’s largest utility service providers supply 43% of California electricity by burning natural gas, the cost for which is billed to consumers. For Sempra, Edison, and PG&E, fuel is provided by gas subsidiaries to electricity subsidiaries of the same holding company. With this self-dealing arrangement, the cost of gas-fired electricity is extraordinarily low – in fact, the holding company is earning a profit on the gas it’s “selling itself” (each of the three own fuel subsidiaries which operate outside the jurisdiction of California’s PUC).

    Either way renewables, for the foreseeable future, will continue to depend upon natural gas for backup when they’re not available. Renewables + gas is a package deal – and that it has somehow escaped a social cost of carbon is largely responsible for the premature shutdown of California’s two carbon-free nuclear plants. If wasting money and polluting the environment is a bad thing, California should be following the examples of Illinois, New York, New Jersey, and Ohio, and extending a zero-emission credit (ZEC) to support Diablo Canyon’s continued operation. Unlike RECs, which operate on the untenable premise emissions can be erased by simply buying an equivalent amount of clean energy, ZECs permit generators to add a marginal credit for avoided carbon to the retail price of clean electricity, based upon a social cost of $42/ton CO2e. Protecting the environment and keeping prices low doesn’t get more efficient than that.

    • PG&E has already found that relicensing Diablo Canyon would have cost $100 to $120/MWH in its application for retirement. (Read the testimony and filings in A.16-08-006.) Given the multiple bankruptcies resulting from failed nuclear construction projects around the world in the last couple of years, nuclear power still needs to answer its big questions: how to control costs and how to dispose of its fuel. Those still haven’t been answered although we’ve been hearing promises for decades from the industry. Time to move on.

%d bloggers like this: