Is the Smart Grid Really a Smart Idea?
A smart electrical system can bring social benefits, but smaller customers may pay too high a price. A more modest plan, guided by government policy, would be better.
It is hard to quarrel with the idea that it is good to be smart. That presumably is why the proponents of some radical changes in the design of the U.S. electrical system came up with the name “smart grid.” The Obama administration has signed on. So have members of Congress from both parties and state utility regulators all over the country. Propelled by promises of greater energy efficiency and reduced greenhouse gas emissions, the smart grid is on a roll.
The smart grid has the potential to bring the United States a more stable, economical, and environmentally friendly electrical system. Unfortunately, it is far from the unalloyed plus portrayed to the public. The cost will be high: Although the economic stimulus program approved by Congress last year included $4.5 billion to help create the smart grid, the full build-out will cost at least a couple of hundred billion dollars more. The potential savings will justify the cost only if the smart grid brings sweeping changes in the way consumers use and pay for electricity. But these changes have the potential to saddle them with unnecessarily high prices, force them to bear unnecessary risks, and make their local utility company an uninvited participant in the intimate details of their everyday lives. These potential changes deserve a thorough airing before the United States commits to such large investments in the name of smartness.
The meaning of dumbness
The nation’s electrical system is an extremely complex network. At its base are generating plants, ranging from windmills to nuclear reactors. Some of these generators, such as big coal-fired plants and nuclear power stations, produce electricity at a steady rate, day in and day out. This baseload power, constantly available, is usually the cheapest type of electricity. Other types of generators, such as wind, solar, and hydroelectric plants, function every day, but their output varies with environmental conditions. A third type of generating resource is a peaking plant, which can be switched on at the times of highest demand. In the United States, peakers usually burn natural gas, but sometimes diesel oil, heating oil, or other fuels. Peakers are often expensive sources of electricity. They are usually uneconomically small and spend most of their time unused, representing wasted investment capital. Some, particularly gas-fired plants, are relatively clean power sources; others, such as those fueled by oil, can be highly polluting.
These generating facilities may be owned by the regulated utilities that distribute power to individual customers, by separate companies (merchant generators) that produce electricity but do not sell it directly to end users, or by industrial plants (cogenerators) that make electricity as a byproduct of producing hot water or steam. Regardless of where they obtain their electricity, utilities devote considerable effort to forecasting how much they will need. A utility knows its customers’ consumption patterns, the weather forecast, and special factors that may have short-term effects on electricity use. The utility’s planners use these data to estimate how many kilowatts of electricity its customers will want each hour of the day for several days ahead.
Once it has forecast demand, the utility looks for the cheapest way to obtain the power it expects to need. There are nearly 18,000 generating plants nationwide, owned by hundreds of different entities. The utility will probably obtain most of its electricity from its own baseload plants or from a merchant generator with which it has a long-term contract. If its expected needs are higher than its baseload supply plus the power it expects to produce from intermittent sources, the utility has several choices. It can switch on some of its peaker plants. It can go into the wholesale market, a well-established arrangement in which utilities bid for electricity offered by sellers with power to spare. It can also mount a public relations campaign to encourage energy conservation. At the other extreme, if the utility expects demand for electricity to be low, its generating plants may produce excess power that it can try to sell in the wholesale market. The baseload plants will run in any event, because a temporary shutdown is too costly and time-consuming.
Benefits of a smart grid
At its core, the smart grid is about capital utilization. A large amount of capacity is seldom or never called into service: U.S. generating plants can make about 16% more electricity than they expect to need even on the hottest summer day. This reserve margin has actually increased slightly in recent years. Regulators require utilities to keep excess capacity because demand is so difficult to forecast accurately and because a lack of power can have large economic and social costs. But this cushion comes at a cost to ratepayers, who must pay for plants and transmission lines that rarely operate. The smart grid is intended to help shave off the demand peaks and fill in the valleys, allowing more efficient use of extremely expensive generation and transmission assets and reducing the need for new plants.
The term “smart grid” incorporates a number of different concepts; not all utilities’ smart-grid programs are identical. But certain key features are common to all visions of the smart grid.
- The grid has two-way integrated communications, both between a utility and the consumer and along the transmission and distribution grid. Most electric meters used in the United States today allow no communication whatsoever. The somewhat smarter meters now being installed in many areas permit radio communication so the utility can read the meter from outside the property but do little else beside recording the number of kilowatt-hours consumed. True smart meters would allow constant two-way communication between utility and customer. Along the transmission grid and local distribution lines, two-way communication could inform a central office of changes in the flow of electricity at any metered location.
- The grid incorporates sensing and measurement technologies that can monitor equipment health, grid integrity and congestion, and energy theft. Many utilities now are unaware of power outages until a customer calls in. Once the grid has two-way communications capability, sensing technology could provide instant notification of a power outage. It could detect that a transmission line is becoming overloaded. It could also sense that a particular circuit is carrying more electricity than is being measured by meters, helping to identify customers who have tapped into the line illegally or circumvented their meters.
- The grid has advanced transmission and storage components that make the use of generation and transmission infrastructure more efficient. Technologies within these categories include such things as high-temperature superconducting cable; distributed generation, which means small-scale generation close to consumption locations; electricity storage, so that power generated by intermittent sources at times of low demand could be saved and transmitted to customers when demand is high; and transformers capable of remote monitoring.
- The grid has diagnostic and control devices and software that can identify and propose precise solutions to specific grid disruptions or outages. These devices and software would make for a big improvement over the current situation, in which it may take considerable time for the utility to identify the location of a problem on the grid and understand the cause.
- The grid includes interfaces and decision-support tools that can help better manage the electrical system. The smart grid will provide system managers with far more operating data than have been available in the past, and software tools will apply that information in decisionmaking.
These various communications and diagnostic devices will make electricity distribution more efficient and reliable. They will improve utilities’ ability to spot and resolve problems. The advanced systems also will make it easier for utilities to incorporate electricity from intermittent sources, such as wind generators and rooftop photovoltaic solar installations, into their supplies, because they will be better able to predict when and where those alternative sources will be available. Although there are many questions about how the individual components will work and whether they will work together, a system with these components would have substantial social benefits.
Need for demand management
These social benefits, however, offer relatively meager financial benefits to electric utilities. Smart systems by themselves do not address the most expensive problem utilities face: the need to install and maintain large amounts of generation and transmission capacity that are used only a few days, and perhaps only a few hours, each year. If utilities are to obtain significant financial benefit from the smart grid, they need to make use of the smart grid’s capabilities to manage demand. Demand management is an integral part of the smart-grid concept, but it has the potential to give rise to very considerable difficulties between utilities and their customers.
The logic of demand management is simple. Today, almost all residential customers and most nonresidential customers pay for power according to a pre-established price schedule. The prices may vary by season, or the price per kilowatt-hour may be different for the first 500 kilowatt-hours used in a month than for the next 500. Retail electricity prices, however, generally do not change over the course of a day, and they do not vary with supply and demand in the electricity market. If a utility’s published rate schedule calls for households to pay $0.10 per kilowatt-hour, that will be the price whether the electric system is awash in excess capacity or is so overloaded that a blackout threatens.
Whereas households and small businesses buy power from their utilities at firm prices, the utilities themselves are constantly buying and selling in the wholesale electricity market, where utilities and generators trade with one another. In the wholesale market, the price is set only for short periods, often as little as 15 minutes. In other words, utilities face what is known as dynamic pricing, which changes frequently depending on supply and demand, whereas the prices they charge their customers are static. In the extreme case, when power is in extremely short supply, a utility may pay more to purchase electricity in the wholesale market than it receives for selling that electricity to retail customers. The end users of electricity receive no price signal to cut back on consumption.
The smart grid gives utilities an opportunity to alter this situation. The savings are potentially large: the Electric Power Research Institute, a utility-industry organization, estimated in 2009 that demand management alone can shave 4.6% off of peak summertime demand by 2020 and 7% by 2030. Over the next decade, the projected savings from demand management are greater than those from increased efficiency in electricity consumption.
Demand management requires that electric meters at each home, farm, factory, or office be replaced by far more advanced devices that incorporate communications and data-processing capabilities. These smart meters could inform customers of price changes in advance or in real time, enabling a utility to change prices as often as regulators permit. The meter might even be linked to a display screen in the customer’s kitchen or office. The customer could be informed of an upcoming change in the price of electricity and could then choose to schedule electricity-using activities at a time when the price will be low.
The smart meter is only the first stage in demand management. Over time, the meter could be linked to a new generation of “intelligent” appliances or business equipment. These devices could be set to operate only under price conditions selected by the customer, or they could be controlled by the utility itself by means of the smart meter. A family might program its electric dryer to run only when electricity costs are low, and to shut off should the price rise above $0.12 cents per kilowatt-hour. Utilities could develop new businesses monitoring individual electrical devices; the smart meter might observe, for example, that a supermarket’s freezer is consuming more electricity than it was designed to use and might contact the store to suggest checking whether a seal is broken. Another new line of business might involve managing consumption: A household or business might ask its utility to hold the electric bill below $100 per month, and the utility would turn lights, appliances, and cooling and heating systems down or off at high-price times to keep the bill below the specified level.
These possibilities have created an unusual amount of excitement in the staid electric industry. Future relationships between utilities and their customers may be very different from what they are today. But when people who have spent their careers in regulated monopolies begin using such phrases as “rich transactive environment” and “consumer enablement” to describe the future, it is worth asking why.
The smart grid, linked to smart meters, will facilitate far more complex pricing systems than household and small-business electricity users are accustomed to. Some observers have described the aim of the smart grid as providing variable pricing. But this description is inaccurate, or at least incomplete. Variable pricing implies different prices at different times of the day or year. The pricing structure permitted by the smart grid is quite different. It may involve not just variable pricing but real-time dynamic pricing, allowing retail electricity prices to vary constantly according to supply and demand conditions in the wholesale power market at any given moment.
Real-time dynamic pricing can lead to high price volatility as market conditions change. There might be advance notice: The utility could notify customers that the present price of $0.12 will treble from 2 p.m. to 4 p.m. because of anticipated high demand. But it is just as possible that a customer currently paying a low rate per kilowatt-hour could experience a sudden price jump due to unexpected events, such as an explosion at a generating plant or a lightning strike that disrupts a transmission line. Prices would be fixed only for very brief periods, perhaps 15 minutes at a time.
Real-time dynamic pricing has the potential to save money for consumers who are able to shift their electricity consumption to times when rates are low. But it could harm small users of electricity in three different ways. First, household and small-business customers will have to devote time and effort to keeping abreast of the price situation or risk paying significantly more to operate electrical equipment than they anticipated. Second, consumers who are unable to shift their consumption in response to price changes could face extremely high electricity bills. This increase would result because bills would be based not on the utility’s average cost of electricity, but on the marginal cost, which is usually higher and which is the only cost that fully reflects supply and demand conditions in the wholesale market. Third, the risk of changes in wholesale electricity prices would effectively be transferred from producer to consumer.
Each of these factors represents a major change from the customary treatment of small electricity users. At present, electricity consumers’ information costs are minimal: a price per kilowatt-hour appears on the monthly bill, and no further price information is required. A sudden spike in the price utilities pay for power will not increase their bills. The risk of changes in the wholesale electricity price is borne by the utility, not by the residential and small-business customer. Under dynamic pricing, all of that would change, to the customer’s disadvantage.
Some utility customers can handle these risks. A large factory or a big office building can purchase hedges in the futures market, plan ahead to close certain operations when electricity prices are high, or sign long-term fixed-price supply contracts. Households and small businesses can do none of these things; a pizzeria cannot turn off the oven for a few hours just because the cost of electricity has soared. And consider the price risks facing the household that turns on its dishwasher before going out for the evening and later learns that during its absence, the price of electricity rose from $0.12 cents per kilowatt-hour to $0.65, making the cost to wash the dishes far higher than anticipated. Small electricity users would have few realistic ways to control their costs under this scenario, aside from programming their smart meters to turn appliances off whenever the price gets above a certain level. The most realistic alternative would be for the customer to hand control of electricity consumption to the electric company, which would use smart-grid technology to reach into a house or business at times of high demand and reduce consumption directly by turning electrical devices down or off.
For utility engineers who regard the smart grid as another tool for optimizing the efficiency of the electrical system, utility control over users’ power consumption would be the most desirable outcome. The reason is that real-time dynamic pricing, by itself, may not fully achieve the goal of flattening demand peaks. From the utilities’ point of view, it is not enough to know that higher prices will curb demand. Optimal asset use requires a high degree of certainty about the extent to which demand will change in response to a change in the price of electricity; if the demand response is not highly predictable, utilities will still need to maintain copious spare capacity, and their basic financial goal in building the smart grid will not be accomplished. Only direct utility control of power consumption will provide enough certainty about peak demand for utilities to radically reduce their reserve generating capacity.
The idea that a utility should be able to turn off part of a customer’s power consumption is not new. Utilities do this now by offering what is known as interruptible power to some large users. The user gets a special low rate, in return for which the utility can disconnect it whenever the utility faces a lack of power, a process known as load shedding. The smart meter will make it technically possible for a household to purchase power on an interruptible basis. It will even allow the utility to micromanage load shedding by first turning off the electric water heater, then the clothes dryer, then turning up the thermostat by two degrees. This could occur in the order determined by the customer. But as houses and businesses fill up with smart equipment, it also is possible that state regulators could give a utility the authority to override thermostats or turn appliances off without customer consent. The virtue of this approach, from a utility’s point of view, is that it would know precisely how much load it could shed at each of a million homes and businesses. With that knowledge, it could safely reduce its reserve generating capacity and avoid purchasing power at high prices, thereby lowering its costs by more than enough to pay for smartening up the grid.
Half a loaf
What seems optimal to utility engineers, though, may not seem ideal to electricity consumers. Households and small businesses will want to control their electricity consumption to maximize their own welfare, not to optimize a utility’s cost function. For the utility industry, this raises the prospect that the smart grid will not adequately modulate demand and hence will not lead to the reduction in reserve generating capacity that is necessary to make the investments in smart technology financially worthwhile.
There is perhaps a better way. Right now, an important subset of utility customers—namely, large commercial and industrial users of electricity—is eager to have the smart grid. These customers are equipped to handle the risks of dynamic pricing, and they are sophisticated enough to figure out how to maximize their profits by turning off various sources of power demand when rates rise. Some of these industrial and commercial customers already purchase electricity on an interruptible basis in order to take advantage of the cost savings that come with it. Many of them would sign up for real-time dynamic pricing immediately if it were possible to do so. They want the smart grid now.
These big power users probably account for about a fourth to a third of all electricity consumption. Finding ways to smooth their demand is low-hanging fruit. It can be picked without installing smart meters at 140 million locations around the country. The potential for smoothing demand is not as large as in the household and small-business sectors, but the hurdles to adoption are much lower. In this area, unlike the household sector, smart-grid investments entail low political risk and a reasonable probability of a decent return.
And what of residential and small-business customers? State utility regulators need to begin thinking of ways to protect the interests of those customers as the smart grid is gradually put into place. There are many ways to expand the use of price signals in the electricity market while stopping short of dynamic real-time pricing. For example, customers might receive a simple price schedule that takes advantage of the smart meter’s ability to change prices during the course of the day. A scheme with one price per kilowatt-hour from 7 a.m. to 2 p.m., a much higher price until 8 p.m., and then a very low price until the morning would be easy for customers to remember, and they would soon get used to cooling the house in the morning and dialing back the air conditioner in the afternoon. Confronting customers with a price change every 15 minutes, in contrast, will not encourage this simple rule-of-thumb behavior to conserve electricity. Dynamic pricing at the retail level is a recipe not for customer enablement but for customer confusion and paralysis.
One of the biggest advantages of this modest approach is that it does not shift price risk to the consumer. That is as it should be. Utilities have the knowledge, skills, and financial capability to protect against the risk that a spell of hot weather will cause wholesale power prices to spike; that an unscheduled outage at a generating plant will lead to a temporary shortage of electricity; and that, on occasion, the wholesale price of power may even exceed the retail price. Individual households and small businesses, on the other hand, have absolutely no means of protecting against such price risks. These are risks that utilities should bear.
A similarly simple pricing model could be used to achieve one of the greatest potential benefits of the smart grid: the integration of electric vehicles into the nation’s vehicle fleet. Electric vehicles have the potential to strain the electrical system under the flat-rate pricing schemes widely used today, because drivers have no incentive not to recharge their batteries at times when power demand is already high. A dynamic pricing structure would leave drivers unsure whether they should recharge their batteries when they park for several hours at a charging point in an office complex or a shopping center. With an easily remembered static pricing rule, on the other hand, drivers would be aware that daytime recharging is costly and nighttime recharging cheap, and would be likely to develop the simple habit of recharging overnight whenever possible.
Most important of all, using the smart grid to enable households and small businesses to make sensible and comprehensible choices about their own electricity consumption will avert the backlash that is likely to come if the public sees the smart grid as a means of manipulating behavior and forcing people to do things they do not wish to do. Regrettably, customer sensitivity is often weak in the utility industry, where most customers have no choice but to deal with their local electric monopoly. In the case of the smart grid, though, customer sensitivity is paramount. A virulent public reaction against dynamic pricing could impede the adoption of smart-grid technologies, delaying the many public benefits that the smart grid can bring.
Both federal and state governments have roles to play in protecting consumers during the transition to the smart grid. Although most of the money to build the smart grid will be provided up by utilities themselves, federal grants and coordination efforts are vital. As they support the smart grid, the Department of Energy and the Federal Energy Regulatory Commission need to make clear that dynamic real-time pricing for small electricity users is not part of the project. State utility commissions need to encourage experimentation with other pricing models, including the rule-of-thumb approach described above, that will promote conservation and reduce peak consumption while at the same time protecting household and small-business users. Equally important, state regulators need to make sure that, as large power users shift to real-time dynamic pricing, the utilities’ fixed costs for reserve generation and transmission are not shifted onto small customers covered by static pricing schemes.
These policies may mean slowing the pace of deployment, because a smart grid without dynamic pricing will produce a smaller return on utilities’ investments. But even though a smart grid without dynamic real-time pricing does not offer the perfect price signals that engineers and economists find so alluring, it will bring many social benefits while protecting the interests of electricity users and reducing potential political backlash. Although dynamic pricing may be the theoretical ideal, in the case of the smart grid, the best may be the enemy of the good.
Marc Levinson ([email protected]) is senior fellow for international business at the Council on Foreign Relations in New York.