Skip to content

Why Aren’t We Talking About Net Energy Metering for LEDs?

The fights over net energy metering have gotten loud and heated. For those of you who have missed the drama, here, in a nutshell, is what “net metering” means. Say I install enough solar panels on my roof to provide about half of my electricity over the course of a year. On a sunny afternoon, if I’ve turned off my Tivo and my refrigerator and dryer aren’t running, my system might be generating more electricity than my house is consuming.

Net metering means that my utility will credit me for the “extra” power my system generates at times like this and charge me based on the difference between my total consumption and my total solar production, i.e. my net consumption. I will be selling back to the grid during the sunny afternoons when my own consumption is low. (See Severin’s previous post on this.)

Will you neighbors admire your solar panels?
Will you neighbors admire your solar panels?

So, how does this apply to LEDs? From a purely technical perspective, it doesn’t. An LED would never bring a house below zero consumption so that it’s selling back to the grid. But the zero consumption threshold isn’t what all the fights are about. If my consumption without the solar panels puts me on the fourth tier, where my rates are as high as 36 cents per kWh, my solar system helps me avoid paying for some pretty expensive electricity. Even if I never go below zero, a solar system will keep me down on the lower tiers, paying only 13 cents per kWh. (On a side note, the solar installers who have approached me understand this and size their systems to avoid the expensive power but not the cheaper stuff.)

But, PG&E has fixed costs, which won’t disappear just because I install solar – it still has to run distribution wires to my house and pay for my meter. It’s paying property taxes and financing costs for its power plants no matter how much electricity they produce.

The basic problem is that utilities are collecting fixed costs – which by definition do not vary as a function of how many kWh customers consume – on a volumetric basis. So, every time someone installs  solar panels, the remaining ratepayers have to pay slightly more to cover those costs. Yes, my neighbors (and all of PG&E’s residential customers) would have to pay for my meter and my “share” of the distribution lines running down our street if I got a large solar system.

The opponents of net metering argue that it provides unfair incentives for people to install solar, which leave the rest of the users on the hook to cover the utilities’ fixed costs. They cite the “death spiral,” meaning that rates get higher for non-solar customers, which induces more of them to switch to solar. The more sophisticated opponents note that solar installations are generally on rich people’s houses, so net metering is regressively subsidizing the rich.

This brings us to the question posed in the title. Why isn’t anyone complaining (at least very loudly) about unfair cost shifting or the death spiral when I buy an LED bulb? Just like a solar system, my LED bulb will help me avoid the 36 cent power, and, given that some of that is collecting fixed costs, my neighbors will be left paying a tiny bit more – if not tomorrow, after PG&E’s next rate case.

My first guess was that we’re not talking about an energy efficiency death spiral since we’re still talking about the energy efficiency gap, which implies that customers are not investing in seemingly cost-effective energy efficiency measures. In other words, customers have been leaving the proverbial $20 bills on the sidewalk and bypassing energy efficiency opportunities, like replacing their incandescent bulbs with LEDs, even though the switch could save them money.

But, I did some rough calculations and concluded that annual energy savings from LEDs could be on par with or even larger than distributed solar, especially when you bring in commercial lighting. Very roughly, California added approximately 1,000 MW of solar in 2013. At an estimated capacity factor of 17%, that’s roughly 1,500 GWh of annual solar production. This report estimates nearly 1,000 GWh of annual savings from light bulb standards in California, climbing to 11,000 GWh by 2018.

Lots of potential?
Lots of potential?

And, we’re talking about potentially large amounts of money shifted across customers, as a large share of the typical utility’s costs is fixed. The California Public Utilities Commission put out a study on net energy metering that calculated that the typical solar customer paid bills that were 54% higher than the utility’s incremental cost of serving them before they installed solar and 12% less than the incremental cost of serving them after they installed solar. This is suggestive of just how much fixed costs are collected on a volumetric basis, particularly on the higher tiers.

So, I predict many more years of heated discussions about rate restructuring. I’d guess that we will look back with amusement at the contentious debates about whether to add a monthly fixed fee to PG&E rates, because the typical customer will be paying a much larger share of their bill as a fixed charge.



Catherine Wolfram View All

Catherine Wolfram is Associate Dean for Academic Affairs and the Cora Jane Flood Professor of Business Administration at the Haas School of Business, University of California, Berkeley. ​She is the Program Director of the National Bureau of Economic Research's Environment and Energy Economics Program, Faculty Director of The E2e Project, a research organization focused on energy efficiency and a research affiliate at the Energy Institute at Haas. She is also an affiliated faculty member of in the Agriculture and Resource Economics department and the Energy and Resources Group at Berkeley.

Wolfram has published extensively on the economics of energy markets. Her work has analyzed rural electrification programs in the developing world, energy efficiency programs in the US, the effects of environmental regulation on energy markets and the impact of privatization and restructuring in the US and UK. She is currently implementing several randomized controlled trials to evaluate energy programs in the U.S., Ghana, and Kenya.

She received a PhD in Economics from MIT in 1996 and an AB from Harvard in 1989. Before joining the faculty at UC Berkeley, she was an Assistant Professor of Economics at Harvard.

48 thoughts on “Why Aren’t We Talking About Net Energy Metering for LEDs? Leave a comment

  1. If the daily “net metering” demand charges become a reality without factoring in daily capacity needed, would those who demand less capacity for running relatively smaller equipment not feel cheated by people who demand more capacity for running heavier equipment?

    If the answer will be yes, then the daily “net metering” demand charges also should not be flat. They should consider load factor also.

  2. Is there any discussion at the CPUC about using a demand charge for residential customers with net metering instead of the fixed fee that seems to be the popular and simpler way of late?

    • The technology is now available to have daily demand charges which avoid the monthly “ratchet” problem.

  3. Electric loads are generally thought of as being inelastic in respect to price. Load control programs will often move some of that inelastic loads to other time periods, but still, think inelastic loads. PV and other distributed generators are commercial transactions. Their installation (and sometimes their operation) responds to prices.

    Utility load research suggests that annual load is a good indicator of the customer’s peak demand, and thus the need for local wires. This is generally in the range of 15-20% annual load factor. Further away from the customer, the demand is diversified, such that wires closer to the substation, and the substation itself, might experience a 30-40% annual load factor. At the generator, residential loads might be so diversified that they have a 50% annual load factor.

    Those are “load factors,” not net energy factors.

    On a net energy basis, all bets are off as far as “load factors.” An “overbuilt” PV system that on net pumps electricity to the grid each year, would have a negative load factor using conventional definitions of load factor, metered energy divided by peak demand. Thus, I believe that utilities need to adopt the old Burbank model of residential demand charges, but set the demand period over shorter time intervals that are available with AMI. I need to follow up on Robert Borlick’s comment about BG&E thinking about adopting residential demand charges, such as I have suggested in these comments and as I discuss on my personal blog.

    Oh, in regard to my nit picking with Robert Borlick, we live about 20 miles apart in Maryland, which both is south of the Mason Dixon line (which is the southern border of Pennsylvania, generally with Maryland but also with Delaware) and has been part of PJM for close to 80 years (the “M” in PJM stands for Maryland), as I remember two minor points of history.

  4. I think we should back up on the premise of this article, and ask “why is the top tier 36 cents per kWh while the bottom tier is 13 cents?” We know that cost is NOT driving that difference. One argument was that it drove energy conservation but we now know after at least one excellent study run out of the Energy Institute and published in the American Economics Review (Do Consumers Respond to Marginal or Average Price? Evidence from Nonlinear Electricity Pricing: that this does not create additional incentives. So we’re left with the only other argument that the rate difference is intended to create a cross subsidy from higher users (thought to also be higher income) to lower users (thought to be lower income). Using less energy to avoid subsidies rather than costs has a much different implication. The solution is to come up with a different method to create those subsidies. But that’s for a different time.

    And in terms of fixed costs, customers are NOT avoiding 23 cents per kWh in costs–more like 7 or 8 cents. And the transmission and distribution network costs are probably avoidable, particularly when looking at a time horizon greater than 5 years (this short horizon is what the IOUs use in California’s rate cases.) The real answer is to figure out what the true long run costs are for distribution investment in this situation, and to force local distribution planners to use more than log paper for forecasting loads (which in fact they still are doing in PG&E’s and SCE’s cases–I’ve asked.)

    Solar panels are probably almost as reliable as LEDs in reducing loads. Panels have lives in terms of decades. We don’t yet know the lives of LEDs (and especially after the CFL debacle it could be very short.) We don’t know if customers will replace LEDs with LEDs or something else, especially if home residency changes every 7 years or less. Note that panels lose output when the sun goes behind a cloud–in the West those are days when temperatures are already cooler and the clouds reduce air conditioner load. I hope that California utilities are not so naive as to assume loss of solar output on the very hottest peak days when there are no clouds!

    One solution might be to put utilities at greater risk for recovering distribution related investment. Why should ratepayers bear all of the risk if shareholders are getting a premium on their return above the cost of debt? If shareholders want to shift all of this cost risk, then give them the return on corporate debt.

    • Richard,

      The AER article claiming that utility customers do not respond to marginal prices is at odds with what is happening in California, i.e., that these same customers are installing solar panels on their roofs in order to avoid paying the top tiered rates. If you look at SolarCity’s advertisements you will discover that they urge electricity consumers to install the amount of solar capacity that just gets them off the top two pricing tiers. Price elasticity is not limited to doing without but also to substituting equivalent products.

      A utility’s fixed costs typically amount to about $30 to $50 per month (including generation capacity costs). About $10 of that represents the cost of metering, billing and that portion of the distribution system (e.g. the line drop to the customer’s house and a portion of the the last step-down transformer) that are invariant with energy usage. Some portion of the remaining costs are avoidable but it is difficult to quantify how much. If load growth were flat few of the costs would be avoidable because there would be little need to expand the system. The average customer served by one of the California IOUs consumes about 600 KWh per month so about 5 to 8 cents per KWh represents the fixed costs of the distribution system (confirming your claim).

      I seriously doubt that any major utility forecasts demand with log paper. That technique went out in the early 1970s.

      However, I agree with your comment that distribution utilities who have revenue adjustment clauses that remove the volumetric risk should only earn returns equal to the interest rates on corporate debt. Aside from regulatory risk (i.e., the unpredictability of what the retail regulators will do) volumetric risk is the only significant risk that a “wires” company assumes.

      • The study looks across ALL customers. Because tiered pricing must be revenue neutral, those customers who have marginal prices below the average cost consume more than enough to offset the 20% of residential customers in the top tier. The study even finds a slight DISincentive for energy efficiency. You can’t look at individual customer responses–you need to look across the entire customer base to see why tiered pricing doesn’t work. (BTW, I’ve seen this looking at CARE customers, who’s lower income should lead to lower usage. Yet their average usage is almost the same as for all customers.)

        I was being flip about “log paper” but PG&E and SCE distribution planners are simply using % trends to forecast local loads at the substation level. The sum of those forecasts exceed the system wide forecast developed by the utilities, and far exceed the adopted CEC forecasts.

        Recent CPUC decisions have largely removed even regulatory risk.

  5. We need to distinguish between the various parts of the electric system: Generation; Transmission; and, Distribution. We have ways to create markets for Generation and Transmission, most of which is no longer quite subject to state regulation. And the ISOs extend over a much greater area than Robert Borlick said.

    Much of the Net Energy Metering issue, such as storage, can be handled by pricing Generation more finely, perhaps every minute. When the sun goes away, NEMs would pay the higher cost of energy during those periods. When the sun is hot, NEMs would be paid the lower cost of energy during those periods. Transmission goes along with this generation issue.

    But distribution is still cost of service regulated. Decoupling doesn’t stop the death spiral issue, just automates it. The answer is demand charges over very short billing intervals, such as the one minute periods mentioned in regard to generation and storage.

    • Well Mark, you got me.

      Yes, ISOs exist well to the West of the Mississippi, such as my client, the Midcontinent ISO and also the Southwest Power Pool. I didn’t mention them because their utility members are predominately vertically-integrated, thus still heavily regulated at the state level.

      Also, you are right re the problems with Net Metering going away if retail tariffs are dynamic and more granulated, which is simply saying that they should be more cost-reflective. However, pricing by the minute would be infeasible because the ISO software calculates a economic dispatch solution (and associated energy prices) only once every 5 minutes. Perhaps as computers become more powerful 1 minute pricing will be possible.

      Lastly, Baltimore Gas and Electric is exploring the potential for including demand charges in the tariffs of residential customers who have smart meters.

  6. One of reasons one exports and imports power is because one’s rooftop solar does not have the capacity to run these high power users at once. Unless one is in a very remote residential area where one owns a solar farm that generates enough to power the heavy users when one needs them, one has to depend on the grid for adequate and capable power to run the heavy power users. The grid is therefore indispensable. On the other hand, one has to store what one generates to use one needs it in future, without involving the grid.
    Like the Zucchini, one has to buy a refrigerator and run it to enable one eat a somewhat fresh Zucchini in future. One has to buy smart inverters, capacitor banks etc. to store what one generates and make it capable of running the heavy power users when one needs them in future. To avoid this, and for EE reasons, one makes the grid store the incapable power one produces and return a capable and adequate power when one needs it. The grid is therefore kind of indispensable to running the heavy power users. It follows that utilities could have netbacked one’s price, for selling power to them, to the cost of providing one’s own storage and/or capacitors banks.
    If this line of argument is acceptable, then it is better to pay 12% incremental price to keep the grid sustainable and running than to resist that, crush the utilities out and store one’s own power.
    The LEDs do not even have a choice. The utilities are indispensable to the use of LEDs. But the cost that the LEDs are passing on to Donkey (who bears the heaviest shifted costs) is just a tiny bit. If the LED users also think of that tiny bit crushing the Donkeys out of consumption, the LED users would then be forced to bear the load the Donkeys have been bearing to keep the utilities in business. So why not pay that tiny bit to avoid taking over the Donkey’s load since the grid indispensable to both the “rich” rooftop solar user and the LED user? At least, at the time when one thinks the grid is indispensable to both.
    This is not to suggest that the utilities should not find alternative ways of getting the fixed costs reduced. The utilities should also be held to plough back part their profits to reduce the fixed costs that otherwise would be shifted across consumers among other alternatives.
    The donkeys should also not be left to remain carrying the huge shifted costs their inefficient consumption creates. That would continue to retain the energy inefficiency gap. But to reduce the efficiency gap the Donkeys create, one has to assume that the Donkeys are willing to participate in energy efficiency projects but cannot afford it and/or do not have the technical know-how to realize that buying the LED will save the $20 down the drain. There is a need to introduce interventionist policies like supporting the producers of the incandescent to phase it out and replace it with LEDs production. If there are no incandescent bulbs, the Donkeys will buy the LEDs and there will be much less shifted costs for the donkeys and EE gap goes much down as well.

  7. Really liked this article and discussion. One of the things I’ve thought about over time, though, is whether new demand will be created by things like electric cars. So ultimately it’s necessary (or at least helpful) for people to install energy systems like solar panels so that the utility companies don’t need to build more plants. So ultimately folks would continue to pay some sort of fee to have the infrastructure in place, but maybe that’s worth moving away from other forms of energy.

    I’m not saying this example is how it should go, but is just one example of how complex the topic is.

  8. I’m not super familiar with the utility structure in New England, but I do know that the industry restructuring that occurred there in the 1990s created “electric distribution utilities.” I’m sure someone on this blog is familiar with the model and can tell us whether they also sell electricity. I believe that all happens through a pool. So, how do the New England utilities address this? Or other regions/states/countries? I agree it’s important and will be around for a long time.

    • Its simple, Wall Street shell companies own the power plants that regulators forced to be divested by the IOUs and thus control the wholesale spot price, which results in everyone else paying inflated prices, with utilities held ransom in the middle as regulated entities and take the brunt of the public outcry as the bad guys while the real culprits are the politicians (paid off by Wall Street and special interest groups) that made a mess of the energy business in the first place.

      • Glenn, that way too cynical – and untrue.

        The movement to divest generation and create competitive wholesale markets began in California in the early 1990s. It was a reaction to the excess costs incurred from nuclear plant cost overruns and cancellations as well as the excess payments made under standard offer QF contracts. Recovering those costs drove electricity prices sky high and resulted in industrial customers leaving (or threatening to leave) California. This resulted in the creation of the California Independent System Operator and the Power Exchange. Many East Coast and Midwest states followed suit, resulting in the adoption of competitive wholesale markets East of the Mississippi and North of the Mason-Dixon line.

        The basic concept underlying this restructuring is sound; it is to replace the discretion of politically motivated state regulators with competitive markets. Those markets are still regulated at the federal level but only to ensure they are truly competitive and do not allow the generators to exercise market power. These markets are not perfect and there are instances where they have been abused. But overall I think they work far better than the mess created by the regulatory process in the 1980s, which resulted in a huge oversupply of high cost base load generation plants.

        Yeah, Wall Street and the generators are greedy and more interested in making money that doing what is in the public interest. What else is new? But guess who paid for all the excess IPP capacity built shortly after the markets were restructured? The investors and the creditors of companies like Calpine – not electricity customers. Wall Street took a big hit. That sounds good to me.

      • hello Glenn, In Texas, Power generators were de-coupled from the distributors. The generators sell into a wholesale market, which is then re-packaged and sold retail to individual consumers. Consequently, the cost of electricity varies more or less directly with the price of natural gas. I have been paying about nine cents per kwh for a couple of years now plus four or five sents distribution charge. That’s thirteen or fourteen cents a kwh total. I would not characterize that as inflated prices. What do you pay in your local community?

        • Michael,

          I’m surprised you pay that much. I thought most Texans paid about 10 cents per KWh including distribution charges. In the Washington DC – Baltimore area we pay about 9 cents per KWh for energy, generation capacity and transmission service plus and another 5.5 cents for distribution services.

    • Michael Thomas: About the same here, or less. EXCEPT we also get stuck with an Energy Efficiency Program up-charge subsidy (income redistribution plan) per meter that goes into a giant pool that ends up in a utility run black hole / overseen by our public service commission. In essence some folks pay for other folks benefits, like putting CFLs in closets, buying new furnaces for poor folks and paying to round up / dispose of folks old broken refrigerators. As far as I know, the Texas example is a unique situation. Isn’t your territory also your own RTO (ERCOT)? Isn’t your’e generation and distribution more or less all within your borders and with your own local fuels? That makes you unlike the other 49 states? We in Michigan have not yet followed California style government meddling and many other deregulated economic disasters, I hope we never do (although some statehouse ill-informed meddlers backed by Wall Street are trying to take us down that path).

      • You are correct, that most of Texas is on a separate grid(ERCOT) from the rest of the country. The exception is El Paso which is not part of ERCOT.

        The big issue with ERCOT these days is the federal wind subsidy of 2.5 cents. This causes significant distortions in the wholesale power market, expecially during the low demand period in the spring and fall, when wind producers ask very low or negative prices into the wholesale market and forcing prices way, way down. This is nice for the electric consumer but really bad for baseload generators. The big power companies are threatening not to build any more baseload power in Texas unless they get their own subsidy to compete with big wind in the spring and fall months.

        If they get what they want, then Texans will be assigned a fee to subsidize baseload power. It would make more sense to end the subsidy to wind but that is a federal subsidy. Alternatively, Texas could tax wind producers 2.5 cents a kwh(countering the subsidy) and give it to baseload producers and that would normalize the market. But that would be too fair. This is Texas, after all, and the guy with the biggest lobbiest wins.

  9. Great article!

    I just want to point out that the way utilities recover their fixed costs through rates that their regulators must approve. The three California investor-owned utilities charge those insane four-tiered rates, not because they want to but because they are required to by law. Consumers, particularly those picking up the tab for rooftop solar, should be placing blame for this insanity on their legislators and the CPUC – not on their utilities. Consumers with rooftop solar are merely exploiting the loophole that Sacramento has created and, in the process, are shifting more of the cost burden on those less affluent than themselves. Talk about unintended consequences!

    Oh, by the way, the energy efficiency advocates are doing the same thing and are just as guilty. Energy efficiency is not a sacred cow; we can over-invest in it just as we can over-invest in rooftop solar. While I’m on a roll, let me also say that California’s “loading order” is nonsense, as it conflicts with basic economic theory.

    • Robert,

      I certainly am not trying to advocate for any one result. However, the loading order in its working form places new resources on a cost basis. The cheapest unit of electricity is the one you do not need to generate, hence why EE and Demand Response are first. The rest of the order (renewables, distributed generation such as CHP and then fossil) is designed to reflect priorities where the market does not recognize all of the externalities. We want fuel diversity, renewables, duel use of fuel and avoidance of new major transmission lines before we want conventional fossil. I’m not certain how that’s not a market response, and certainly do not see that as nonsense.


      • “Market response” is where willing seller and willing buyer engage in voluntary exchange. What is occurring is not voluntary but coerced by the state via legislation. Not quite the same, regardless of your use of the royal “We”.

      • Michael,

        The way the “loading order” is portrayed is to “adopt as much EE and DR as possible, then go to renewables…then natural gas…etc.” This statement is indeed nonsense because the respective costs of the various resources available are not ordered by resource type in that simplistic manner, i.e, there are EE and DR measures that are far more expensive than generation from say a gas-fired CT. If the simplistic ordering you suggest were true, then California should implement the “loading order” by taking all the generators off line and be done with it! That would effectively max out EE/DR. It is comforting to hear that California realizes that is not what it wants to achieve and implements the “loading order” in a more rational manner (hopefully).

        Another point; as any competent economist knows, the optimal (i.e, cost-beneficial) solution is achieved when the marginal benefit derived from each resource per dollar of cost is the same. To get to that point it doesn’t matter where you start. Certainly you can start with EE and DR, or you can start with the dirtiest coal-fired resource. Clearly, the methodology underlying the “loading order” is an iterative, not a unilaterally progressive process. At least if it’s done right. Granted, it may make sense to start with EE/DR if that approach provides more rapid convergence to the optimal solution. But I have never heard that argument advanced.

        Finally, you made an earlier comment that EE resources must be “cost-effective.” That offers little comfort because the cost-effectiveness tests in the CEC’s Standard Practice Manual were flawed from the time it was published back in the 1980s. The flaw is that the loss of consumers’ surplus produced by an EE or DR measure is not quantified and counted as a legitimate cost. I pointed this out way back in 1987 and since then others, (e.g., Ahmad Faruqui) have also. Has this been fixed?

        The punitive four-tiered rate structure that the CPUC has imposed on the Investor-owned utilities by itself virtually guarantees that California is over-investing in EE while forcing electricity consumers to involuntarily pick up the tab. How can that be cost-effective?

  10. If the fixed infrastructure required to serve a customer is proportional to their peak use, then the LED retrofit is more deserving than PV since it is likely to reduce the capacity necessary to serve that customer, whereas PV might not at all, as residential peak demand comes nowhere near midday. And in a future where PV is already generating a significant amount of midday energy, on the margin, the effect can be much greater, where PV saves the system maybe nothing in capacity value but the lighting retrofit could be very beneficial indeed.

%d bloggers like this: