It is under-appreciated that the driver of real time price volatility is capacity not energy. The current outcry for, energy only, real time pricing is another simplistic and on its own will not work long term. Dynamic pricing is an energy only strategy with only indirect ties to capacities value and cost. This outcry is a clear indication of the poor understanding of the difference between energy and capacity. The danger for customers is the potential magnitude of the price volatility.


The North American electric system is arguably the largest, most complex machine ever built. The capacity portion of that machine is poorly understood and confused with the electrical energy it transforms from other primary energy forms. The unanswered challenge for the energy sector is to come up with an equitable method to charge retail customers for their capacity demand. Currently, retail customers are not concerned with capacity usage; only their energy usage is relevant to charges incurred. Some have a fuel charge with a fixed charge per KWH on top of it, or just an all in fixed charge. None is based on the true determinant of capacity, the highest demand during highest load period of the year. The elimination of the disconnect between retail customers and the value of the capacity they consume is integral but ignored initial step in any long term stable minimally regulated market type design.

An informative overview of capacity market designs for PJM was prepared by The Brattle Group1. Keep in mind that air-conditioning load drives capacity needs. Reducing or controlling comfort cooling load is just as beneficial to the system as adding capacity. This is just one of many complicating factors.

In order to meet long-term capacity needs successfully, capacity planning or markets must do three things:

  1. Markets must extend far enough into the future to allow the lead-time required for the addition of resources.
  2. The incentive must be of adequate magnitude and duration that the risk/reward ratio is attractive to investors.
  3. Even though subjective to a degree, it must maintain an economic balance between capacity types.
The easiest way to get adequate capacity in the optimal configuration is to simply buy it, this way the controlling authority gets to determine when, where, quantity and technology type. There are several complex methods employed to reach the end of having adequate capacity. Typically, the general intent of these methods is to create some sort of a semi-free capacity market that still puts the steel on the ground when and where needed. No current alternative design does a better job than the traditional vertically integrated regulated utilities of bringing and keeping capacity in service as needed.

Another complication to insuring adequate capacity is retirements of existing capacity. This is also the most likely avenue for the gaming of market type systems. All market systems have a reliability provision to keep capacity unprofitable for energy generation but profitable as capacity from being retired. The owners are insured a profit for continued operation often in the form of a capacity payment or an equivalent. This creates an incentive for owners to retire even marginally profitable machines to get the lower risk of assured profits. The market flaw is these machines are profitable they provide a benefit greater than their cost, but do not generate adequate margin in the energy market to cover their costs. These assets need payment for the capacity they supply in order to be solvent. A viable market design should provide this innately any design that does not is flawed and will need a "out of market" or regulated solution to maintain adequate capacity.

There are several less than optimal solutions to capacity acquisition. The issues with capacity acquisition are many; for example, what may the best for investors, speculators and generators may not be best for customers. What is the best near-term solution may not be best long-term solution.

Capacity, like most things it is more complex than it appears on the surface. The relationship between energy and capacity cost is inversely related; as the overnight cost of capacity goes up the cost of the energy it produces go down. For example, simple cycle combustion turbine capacity is cheaper than combined cycle capacity but the combined cycle uses less fuel for each unit of energy output. To a large degree future predicted fuel prices sets the balance. Oil and gas prices pushed the US towards coal and nuclear after US oil production peaked combined with the oil embargo of the 1970's. After Three Mile Island, nuclear capacity became cost prohibitive. High natural gas prices and political pressure on coal made nuclear look good for short a while, but the collapse of natural gas prices and the prediction long-term low natural gas price has made gas fired the current apparent best choice.

However, it is still not this simple, building too much of what appears the obvious choice now, Gas fired generation, will punish the ratepayers if gas price predictions are wrong. Capacity choices must be well managed to hedge this risk. Right now, everyone in the US is in love with Natural gas. If only one of our current gas assumptions is wrong gas prices, could move north quickly and punish those with large proportion of natural gas in their portfolio.

The driver of underinvestment in capacity is a lack of certainty. Investing in generation in an energy only market is risky. It is weather speculative; it can be years between rewarding weather patterns. It requires one to predict the behavior of competitors, how much other capacity will competitors add. It necessitates predicting future prices and the amount of time prices will be at profitable levels. Obviously, it lacks the certainty needed for leveraged investment. There are many designs to correct this flaw all of which devise a controlled means to bring adequate capacity online to meet system needs. These schemes are of varying complexity and details. It may not be the optimal configuration but planned "out of market" systems are generally successful in assuring capacity reserves are adequate.

As consumers, we exercise greater control during mundane purchases than some areas regions do for capacity additions. Generally, we did not put her money on the counter with our minimum specifications, and then let the store clerk decide the details of product we will receive. We shop and select the goods that best fit our needs at what we consider a reasonable price. This is a lesson the energy sector should emulate. It is unlikely that by random chance that unmanaged capacity additions driven purely by economic incentive alone will be optimum. Without some sort of a methodical planned approach, the capacity purchasing entity fails to create a "person-hood" to act as the capacity consumer to protect the interests of final the capacity customers.

Charging retail customers for the Capacity consumed

It will take smart meters to accomplish this, which is why it is only now becoming practical, capacity charges are common practice for industrial customers already. For a residential customer it would be there usage during the highest system load periods of the year. The remainder of the bill would be the energy charge, which will vary, based on usage. The important concept is capacity is real it is a product of the machinery. The quantity of capacity a customer uses on the worst day exists for that customer on the all other days. This tenet is greatly under-appreciated. How many retail customers even realize this, it is likely very few. In fluid a society even the details of a mechanism to charge customers for capacity by the KW year would have its challenges.

The upside is, if that customer can reduce consumption for those critical times their capacity bill will be reduced. In the absence of a predictable charge for capacity, there is little incentive to reduce capacity consumption. Charging customers for capacity would create a market place with incentive to reduce capacity consumption and to utilize substitutes that are more economical if and when available. Potentially there are many of these, from energy storage in electric vehicle batteries to Ice Storage Air Conditioners etc.

Volatility of Dynamic Pricing is Inversely Related Reserve Capacity

In short order, utilities, generators, and Load Serving Entities (LSE) will line up to tout the advantages to customers of real time dynamic pricing. Why? Dynamic pricing shifts a good deal of risk away from suppliers to customers. It is unlikely that many of the regulating entities have the depth of understanding needed to appreciate this and are likely to join the bandwagon since it is new and involves advanced technologies. Sadly in the US the rallying cry is, if it's "smart" it's good.

Several researchers and reports extol the benefits of dynamic retail pricing to customers; however, I challenge anyone to find a single study that exposed the subjects to the prices necessary to insure adequate future capacity. Like the ERCOT energy only design, they will work well at first but will self-destruct as capacity reserves diminish over time. They will also work during a study to determine the benefits of dynamic pricing as long as capacity is adequate. All the studies I read have been done under the favorable conditions of a market with adequate capacity. The methodology may have been fine, but their predictive value is near zero, these studies are like children hunting Easter eggs on Easter Sunday, they found what they were hunting partially because they were hunting under the ideal conditions to find it. After implementation, it could take several years for the market design flaws to induce the price spikes of a magnitude and frequency to invalidate this previous research and demonstrate that energy only dynamic pricing is not an adequate market design and is not in long-term best interest of customers.

The ERCOT ISO took over in August of 2001 and is just now beginning to have serious issues that will necessitate protocol and market design changes. It is eleven-year success does not make the design less flawed it simply means it started with a lot of extra capacity that was further aided by wind rush in west Texas. The same is true of dynamic pricing it may work well for years or indefinitely if another mechanism is incorporated to insure adequate capacity. Any other, of many, capacity acquisitions mechanisms will allow retail price caps at levels that customers can and will tolerate. However, if dynamic pricing is the only capacity incentive it will eventually crush retail customers and breed longer and more frequent blackouts. Capacity addition will always lag behind capacity needs in a purely energy only market.

In a previous installment, the flaws of the ERCOT energy only market were noted in particular the Brattle Group report that asserts that even spikes of $9000/MWH, 200 times typical levels, would be inadequate to create incentive to build adequate capacity to provide desired ERCOT reserves.2 This report is informative; it is another excellent read for anyone interested in the energy market designs.

Real time pricing will not incent adequate capacity at reasonable prices. What it will do coupled with retail pricing is subject customers to price spikes of hundreds of times typical levels and move risks to the retail customers that are largely outside of their control. Risks from poor plant maintenance, poor planning, poor line maintenance, fuel prices, fuel shortages, hedging failures, and poor infrastructure will be transferred to the customer through dynamic pricing. It produces a counterproductive incentive for generators. Generators will benefit from poor performance by inducing higher prices for their product.

California's early market failures clearly demonstrated this flaw when Enron and others used strategies like the now infamous "Get Shorty," "Death Star," and the "Forney Perpetual Loop" to drive real time wholesale prices up to benefit generators and traders. The California regulators and their consultants misinterpreted the situation and made large-scale long-term purchases at prices so high that they defaulted on the ridiculous obligations and spent years battling it out in the legal system. The mistake was buying energy when they needed capacity. The water department obviously over stepped it expertise when it went in the power business. The overnight cost of new capacity at the time was around $10/MWH but California paid around a $100/MWH for energy and its associated capacity. Simple capacity plus fuel and Variable O&M deals (Tolling Agreements) would have saved all the pain. These are the easiest and lowest risk deals for both buyer and seller. The novice water department took a simplistic energy only approach based on heavily skewed forward pricing curves that made ridiculous prices look like bargains. Whether induced by manipulation or not they should have backstopped them against fundamentals like the overnight cost of generation capacity.

In conclusion: Capacity and Dynamic prices are interrelated. It is definitely possible to build a dynamic market place that will benefit retail customers and electrical energy suppliers. Will we do it? Unlikely, far too much of our expertise is theoretical, with the involved experts each looking at small piece of greater whole through the focus of their own specialty and then making general assumptions about the remainder that fits their own outlook and model. Some of the best examples are Secretary Chu and Chairman Wellinghoff both demonstrating a limited perspective and agenda far removed from a functional market place for the mutual benefit of all stakeholders.