The Energy-Climate Complex: The Smart Grid: Separating Perception from Reality

The Energy-Climate Complex

LAWRENCE J. MAKOVICH

The Smart Grid Separating Perception from Reality

Instead of a disruptive technology poised to transform the power sector in a decade, expect a more evolutionary change toward a “smarter” grid, with more modest results.

There is a widespread expectation in the United States and around the world today that the smart grid is the next big thing, a disruptive technology poised to transform the electric power sector. The belief is that the use of smart meters and other devices and systems will allow consumers to manage their own electricity use to radically reduce energy costs. The implementation of a smart grid system will enable the widespread use of renewable energy sources, allow more-distributed electricity generation, and help reduce carbon emissions.

The reality, however, is more complex and sobering. The smart grid idea is more accurately characterized as an extension of innovations that have been ongoing for decades. Change will continue but will be incremental because the technology is still evolving and because most consumers do not want the more flexible and uncertain pricing schemes that would replace the predictable and stable pricing of today’s system. Indeed, it appears that most consumers, at least in the short term, will not benefit from moving to a smart grid system. Although a smart grid would probably help slow increases in electricity bills in the long run, it will not reduce them, because too many other factors will be pushing prices and power usage up in the years ahead.

The evidence from an IHS Cambridge Energy Research Associates study, which draws on the knowledge and experience of those closest to smart grid implementation, is that the smart grid “revolution” is off to a bumpy start and that there will be many more bumps in the road ahead. That road is still worth pursuing, but we will need to develop a more realistic understanding of how the electric power system in the United States is evolving. Instead of a demand-side–driven transformation of consumer behavior and the elimination of future capacity needs, expect a supply-side, engineering-driven application of smart grid technologies to improve network operation and reliability in the short term and to slow growth in generating capacity needs in the long run. In many respects, we already have a smart grid in the United States. In coming decades, we will be moving to a “smarter” grid. The pace will be gradual, but the eventual benefits will be real.

The smart grid narrative

In the United States and other developed countries, an appealing and optimistic vision of the future smart grid has gained credence, even though the move toward a smarter grid is likely to turn out quite differently. In the current narrative, the United States and others are currently crippled by a balkanized “dumb” grid with endemic cascading failures, a result of continued reliance on antiquated, century-old technology. The solution is the smart grid: a continental-scale network of power lines incorporating advanced meters, sensing, and communication and control technologies that are linked through universal standards and protocols. It will be coordinated with advanced two-way broadband communication technologies that feed data into complex optimization software systems, allowing control technologies to deliver a more secure, self-healing, higher-quality, and lower-cost power network.

Smart grid deployment, the story continues, will dramatically reshape power use. The smart grid will present consumers with real-time power prices and displays of information regarding power use by specific end uses. These price signals and information streams will empower consumers to have more control over their power consumption. Consequently, the smart grid will alter consumer decisions either directly through behavioral changes or indirectly through preprogrammed smart appliances and control applications. As a result, market failures will be fixed and much of the low-hanging fruit of the efficiency gap will be harvested. These efficiency gains will provide enough savings to drive monthly power bills lower. In addition, the gains in reducing peak power demand will be more than enough to offset the baseline growth in power in the future. Consequently, the smart grid will eliminate the need to build conventional power plants in the years ahead.

The smart grid will also enable a transformation in power supply, the narrative says. Indeed, eventually the smart grid will allow renewable sources such as wind and solar to supplant traditional sources. The use of small-scale, distributed-generation resources will lead to a significant decarbonization of future power production. “Smart systems may well be mankind’s best hope for dealing with pressing environmental problems, notably global warming,” said the Economist in a November 6, 2010, special report.

The smart grid narrative also envisions a rapid increase in electric vehicles, which will generate power or act as batteries in the grid. In time, there will no longer be a need to build conventional power plants to deal with peak power periods because of the new distributed, small-scale power generation.

Finally, according to the current narrative, the pace of smart grid investment, including widespread installation of smart meters, demonstrates that smart grid technology is reliable, economical, and gaining enough momentum that the smart grid will be ubiquitous in power systems within a decade.

The above story about the smart grid has been repeated so often by industry leaders, technologists, and the media that is has taken on a life of its own. It is appealing because it reflects optimism that a disruptive technology can transform the power sector by solving problems that otherwise appear difficult and expensive to address with current technology, and that it can do so without downsides. But this vision is also too good to be true. In reality, forcing a technological transformation of the power sector through the deployment of smart grid technologies along with real-time power prices appears to be not only a formidable task but also not a very likely outcome any time soon.

Killer app?

Dynamic or real-time pricing, the ability to price electricity based on moment-to-moment changes in production costs, is expected to be the killer app of an emerging smart grid. The reality is that although some consumers can benefit from smart grid capabilities and dynamic pricing schemes, the majority cannot.

Real-time pricing is not a new idea. Economists have long considered the ability to use real-time prices that reflect the marginal cost of electricity at different times of the day as a more economically efficient way to price electricity. The Public Utility Regulatory Policy Act of 1978 encouraged utilities to use time-of-use–based rates to price electricity. Congress, in the Energy Policy Act of 2005, encouraged state regulators and utilities to shift from fixed rates to time-varied electric rates in order to increase energy efficiency and demand response.

But most consumers focus on their pocketbook rather than the theoretical basis of this supposedly more efficient pricing system. After all, the prospect of real-time pricing involves higher and more unpredictable prices; on an hour-to-hour basis, the marginal cost of electricity is hard to predict and can change by a factor of 100 during any given day. Research clearly indicates that most consumers far prefer the stable and predictable power pricing schemes they currently have.

Real-time power prices are usually higher than traditional rates during peak periods and lower during off-peak periods. But most consumers use more electricity during peak periods than during off-peak periods. Thus, unless they can shift enough of their power use, typical consumers face a higher bill with a move to real-time pricing. Most consumers, according to research, doubt they can do this and expect that real-time pricing will increase their bills.

Policy designed to support smart grid investments should avoid setting unrealistic expectations, especially the belief that smart grid programs will reduce power bills.

The residential consumers who are more supportive of dynamic pricing tend to be higher-income people with bigger homes who have more space to heat and to cool and more electric appliances. They are more likely to find an adequate payoff from investing in systems to manage this consumption across time and against dynamic prices. Pilot studies show that electric-intensive nonindustrial consumers respond favorably to enabling technologies such as programmable thermostats, price-alert mechanisms, or direct-load controls. In contrast, consumers with smaller homes and fewer electric appliances generally have less flexibility in shifting their power use. It is not surprising that consumer participation rates in dynamic pricing programs have usually been extremely low.

Participation in almost all dynamic pricing programs in the United States has been voluntary. Currently, time-of-use rates are offered by more than half of investor-owned utilities. Many of these programs have been offered for years, and in some cases decades. The average participation rate in such programs is estimated at 1%.

Participation in programs in Illinois is typical. Commonwealth Edison ran a residential real-time pricing pilot program from 2003 to 2006, and for the past four years has made it available to all of its residential consumers. A neighboring utility, Ameren, has a similar program. As of September 2010, Ameren Illinois and Commonwealth Edison each had about 10,000 participating customers, representing 1% and 0.3% respectively of their eligible consumers. In the eastern United States, Baltimore Gas and Electric made time-of-use rates available to residential consumers for several years, but only 6% of residential consumers opted to participate.

Arizona provides an example of how the characteristics of the customer base affect the outcomes. Consumers there tend to be more electric-intensive because of above-average cooling loads. In addition, the nature of these loads provides greater-than-average flexibility in the time pattern of electric use and thus a higher-than-average probability that shifting power use could lower a consumer power bill. The Salt River Project and Arizona Public Service (APS) have about half of their customers on a dynamic pricing scheme. APS offers four time-of-use rates to customers. A 2010 analysis of two of the rates indicated that customers saved 21% on their electricity bills as compared to being on a flat rate.

The same economic logic that helps to understand the Arizona versus Illinois results also applies to nonresidential consumers. Some industrial and commercial consumers find that power bills make up a large percentage of their operating costs. They also have the flexibility to alter their consumption pattern and can thus benefit from dynamic pricing schemes. Still, it appears that only a minority of nonresidential consumers can benefit from dynamic pricing. For example, although Georgia Power runs one of the most successful real-time pricing programs in the country, it has signed up only 20% of its largest commercial and industrial customers.

Even for large nonresidential consumers, switching to real-time pricing does not guarantee lower prices. Indeed, many face higher power bills, according to research by Severin Borenstein in a September 2005 National Bureau of Economic Research working paper. In a four-year study of 1,142 large industrial and commercial customers in Northern California, Borenstein found that holding all else constant, about 55% would see their bills rise under real-time pricing. He estimated that most customers would see their bills rise or fall by less than 10%, with more variability in their monthly payments.

A majority of power customers are not clamoring for access to dynamic pricing. So what explains the enthusiasm expressed by many who have participated in smart grid pilot projects? First and foremost is the fact that the programs have been voluntary. As a result, participants are self-selected members of a small set of the population who are inclined to try a new technology because they like experimenting with innovations. But self-selection bias can make pilot-project results unreliable as an indicator of how the larger population is likely to react to the new technology. It is risky to assume that if other consumers were to learn about these programs or were required to participate, they would end up loving them too. Mandatory participation could also lead to a backlash and derail any significant implementation of the technology.

Indeed, a bit of a backlash has already occurred. Many smart grid initiatives are going forward without any dynamic pricing schemes and those that do use dynamic prices employ highly muted price signals. Currently, there are no real-time pricing mandates for small customers (residential or small commercial) anywhere in the United States. This outcome of the regulatory process aligns with lessons from the past. The Maine Public Utility Commission mandated time-of-use rates for large-use residential consumers during the late 1980s, and the state of Washington mandated such rates for 300,000 residential consumers served by Puget Sound Energy in 2001. But in both cases most consumers were not able to shift enough usage to lower their electric bills, and the programs were eliminated within two years. In addition, these consumer preferences often translate into laws and regulations. California passed a law prohibiting dynamic pricing for residential customers, and New York imposed restrictions on the use of such pricing.

Many states, however, have recognized that some residential customers have the flexibility in power use to benefit from dynamic pricing and have required utilities to install a smart meter at the customer’s request. As expected, only a minority of consumers have requested the meters. Also as expected, these consumers are primarily large industrial firms. However, even for larger consumers, the offerings typically involve dampened price signals that fall far short of real dynamic pricing.

In addition to lackluster consumer demand, there have also been bumps on the supply side, as utilities have struggled to install the equipment and systems needed to make the smart grid work. There have been notable examples of technology problems and cost overruns, indicating that smart grid technologies and their optimal technical configurations are not yet proven and fully commercially available.

  • In Boulder, Colorado, Xcel Energy’s costs to implement a smart grid program have soared from an estimated $15.2 million in 2008 to $42.1 million in February 2010.
  • In Texas, Oncor Electric Delivery Company installed smart meters that later turned out not to comply with the standards set by the Public Utilities Commission of Texas. Oncor was subsequently allowed to recover $686 million from customers to install meters incorporating the new standards, as well as recover the $93 million cost of obsolete smart meters that were never installed.
  • In California, the communication system included in the original smart meter deployment at Pacific Gas and Electric Company (PG&E) turned out to be incompatible with the communication and control needs of the evolving smart grid applications. PG&E was allowed to increase prices to recover almost $1 billion of associated costs. In addition, in November of 2009, PG&E was forced to temporarily stop deploying smart meters in Bakersfield, California—part of its $2.2 billion, 10-million smart meter deployment program—because of consumer complaints and lawsuits concerning perceptions of billing errors. Although these perceptions turned out to be wrong, the backlash illustrates the problem of attempting to roll out the smart grid program at the same time that power prices were increasing.
  • In Maryland, the public service commission refused Pepco’s request to implement one form of dynamic pricing, even on an opt-in basis, because it considered the risk too great that customers would opt into the system with the expectation of lower bills only to find that, at least initially, the new rate would result in higher bills.
  • Also in Maryland, after consumer advocates challenged the cost/benefit analysis of Baltimore Gas and Electric’s (BG&E’s) smart grid initiative, the company’s request for rate recovery of the $835 million cost of its smart grid meter deployment plan was initially denied. The state Public Service Commission (PUC) ruled against BG&E even though the company had received a $136 million grant from the U.S. Department of Energy to help fund the project. The PUC found that, “The Proposal asks BG&E’s ratepayers to take significant financial and technological risks and adapt to categorical changes in rate design, all in exchange for savings that are largely indirect, highly contingent and a long way off.” In rejecting the proposal, the PUC also noted that the cost estimate did not include the approximately $100 million in not-yet-depreciated value of existing meters that would be retired before the end of their useful lives.

As the above examples make clear, the direct benefits of smart grid investments have not yet proven certain or significant enough to fully offset the costs of implementation. The implication is clear: The United States is not moving to a rapid full-scale deployment of smart grid technologies and systems anytime soon. Future implementation is likely to be phased in by customer segments and be geographically uneven and far from complete in one decade.

One way to manage expectations is to stop using the term smart grid because it implies a disruptive technology investment and instead portray the evolution toward a smarter grid as just business-as-usual grid automation and modernization.

A more realistic outlook

A more realistic vision of the future begins with the recognition that the smart grid is an incremental technology trend well under way rather than a disruptive technology that will transform the power sector in the next decade. The evolution toward a smarter grid has been taking place for several decades, as the power sector has incorporated available and emerging monitoring, automation, and control and communications technologies into the grid in varying degrees. These developments have already produced tangible gains: reduced costs for metering and for service connections and disconnections, as well as improved detection and isolation of problems during power outages and faster restoration of power. These gains in security and reliability have thus far reinforced the traditional grid and large central station power system rather than created economic forces pushing toward a distributed supply structure. As a result of these changes, it is inaccurate to think of the U.S. system as having a dumb grid that technology is poised to transform into a smart grid. Instead, smart technologies are already adding another layer of visibility to the condition and operation of the grid and also adding another layer of reliability by enhancing the capabilities needed to predict potential instabilities in the system. In short, the evolution to a smarter grid is helping to maintain and improve the high levels of reliability to which consumers have become accustomed.

The evolving smart grid will allow more experiments with various dynamic pricing schemes, but they should be experiments, and they must be gradually introduced or face a possible backlash from consumers, who mostly cannot benefit from dynamic pricing and value the stable and predictable prices of the current system. As dynamic pricing schemes evolve in the years ahead, they will mostly be used by larger, electric-intensive consumers who have the capability and the money to invest in and manage the new systems.

Investment in smart grid technologies in the years ahead will depend to some degree on the political tolerance for increases in power prices, because developing a smarter grid is not likely to reduce bills, for two reasons: First, the percentage increase in prices will probably not be offset by a larger reduction in electricity use enabled by the smart grid. Second, smart grid implementation is occurring during a period of rising real power prices. Even if smart grid savings could offset costs, there are other factors that are continuing to push prices up. As a result, the case for smart grid investments will involve a different expectation: that although power prices are increasing, prices are going to be lower than they otherwise would have been but for the smart grid investments. This is a harder argument to demonstrate and thus a weaker driver for smart grid investment than the straightforward guarantee of a lower power bill.

The evolution of smart grid technologies could allow the introduction of meaningful numbers of electric vehicles, but this process, too, will be slow. The big hope is that electric vehicles can act as roving batteries to the grid, thus reducing the need for new system capacity. But this outcome is unlikely anytime soon, because current electric batteries are technically not well suited to power system storage and their prices are extremely high. Still, effective coordination of smart grid policy and policy support for electric vehicles could help accelerate smart grid development.

Smart grid implementation is also not likely to reduce energy use enough to provide meaningful greenhouse gas emissions reductions. The reason is that the primary link between the smart grid and greenhouse gas emissions is not within the power sector—enabling renewable power or reducing demand—but rather outside the power sector by enabling the use of electric vehicles, something that adds rather than detracts from power usage.

Finally, the pace of smart grid implementation will probably be slowed by consumer privacy and cybersecurity concerns. Many privacy advocates are concerned that smart grid data could provide a detailed profile of consumer behavior.

Policy implications

Policy designed to support smart grid investments should avoid setting unrealistic expectations, especially the belief that smart grid programs will reduce power bills. The long-run success of smart grid policies hinges on delivering what has been promised. Policies that fail to meet expectations will lead to disappointment, a search for a scapegoat, and a political backlash that will impede progress in the years ahead. One way to manage expectations is to stop using the term smart grid because it implies a disruptive technology investment. It would be wiser and more accurate to speak of the evolution toward a smarter grid as just business-as-usual grid automation and modernization.

The smarter grid rollout should start first with consumers that meet the profile of those most likely to benefit from smart grid programs: electric-intensive consumers with significant flexibility in their use of power over time. Because customer characteristics, particularly the flexibility to cost-effectively shift power use, are so varied from one place to the next, we can expect the implementation of smart grid capabilities to be geographically uneven.

The pace of implementation, especially of dynamic pricing schemes, should be phased in based on the political tolerance of consumers for power price increases. The move to real-time prices should begin with mildly time-differentiated prices that move gradually toward real-time price signals over the long run. Education of consumers will be necessary, but policymakers must recognize the limits of education in divorcing consumer preferences from underlying pocketbook issues.

A significant role remains for smart grid pilot projects to manage the technology risk associated with the evolving smart grid, although policymakers need to recognize the limits on generalizing the results of these projects. The focus for pilot programs should expand from testing dynamic pricing schemes to experimenting with new applications for smart grid capabilities.

In sum, by resetting our expectations and taking modest, gradual steps forward, we can eventually move toward a more robust, smarter power grid in the United States.


Lawrence J. Makovich () is vice president and senior advisor, Global Power Group, at IHS Cambridge Energy Research Associates (IHS CERA) in Cambridge, MA, and directs the firm’s research efforts in the power sector.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>