by Tanya Bodell, Energyzt
(This is Part 2 of a series. Click here to read Part 1.)
There is a growing sense in the industry that something is awry. Some of the most efficient baseload power plants in the fleet are not dispatching during off-peak hours. Coal plants are extending outages with nitrogen blankets to protect their plants during layup in the shoulder months. And most markets have experienced negative prices during off-peak hours, in effect requiring power plants to pay someone to take the energy they produce. What in the world is happening to the power sector?
Start with the economics of natural gas power plants, which have operational flexibility to respond to market price signals on an intraday basis. If electricity prices are higher than the marginal cost of the natural gas power plant, the plant operates; if lower, the plant does not operate.
Spark spreads, the difference between the price of electricity and the price of gas converted to the same units using an assumed heat rate, continue to reflect the economics of natural gas plants. If the spark spread is positive, gas plants are in the money; if negative, they are out of the money.
For purposes of the following analysis, assume the heat rate of an average baseload natural gas generating unit and calculate the spark spread as follows:
Spark Spread = Price of Electricity — Price of Gas * 7,000 kWh/Btu
The price of gas impacts the spark spread directly and indirectly. The direct impact comes from the conversion of gas into a power price using the heat rate. As gas prices go down, the cost of gas goes down and the marginal cost of producing power decreases, thereby increasing the spark spread.
Lower gas prices, however, also impact electricity prices because natural gas power plants generally are setting the price for power. As gas prices fall, electricity prices decrease, increasing the spark spread.
These countervailing forces mean the relationship between gas prices and spark spreads is not obvious.
As shown in Figure 1, spark spreads have varied in the Northeastern markets, with periods of negative values during off-peak hours and peaks during summer when electricity demand is highest.
Dark spreads are the coal plant equivalent of spark spreads and represent the difference between the price of electricity and the price of coal converted to a dollar-per-megawatt-hour metric using an assumed heat rate.
If the dark spread is positive, coal plants are in the money; if negative, they are out of the money. For purposes of the following analysis, assume the heat rate of an average coal plant and calculate the dark spread as follows:
Dark Spread = Price of Electricity — Price of Coal * 10,000 kWh/Btu
Unlike gas plants, which have significant operational flexibility, coal plants generally cannot turn on and off quickly, limiting their ability to respond to intraday dark spreads. These operational limits are a function of the technology, the efficiency hit associated with lower generation output and the higher cost of fuel oil required to start the boiler.
Such operational constraints did not matter when the cost of producing electricity from coal was well below the cost of producing electricity from gas and dark spreads were consistently positive.
As dark spreads have declined, however, the short-term operational constraints are limiting the ability of coal plants to respond to their new position in the supply curve. Many coal plants in Northeastern markets have lost money during off-peak hours to realize positive energy margins from the on-peak dark spreads. Barely covering their intraday marginal costs of production, thin energy margins are not covering the fixed costs associated with these plants.
Figure 2 illustrates the average monthly dark spreads in New York during the past decade. Dark spreads peaked during the high gas and electricity prices in 2006, again in 2008 and have declined since.
The actual situation is hidden by the average across the entire market. In some zones, dark spreads have been negative during on- and off-peak hours. Yet it does not take a negative dark spread – which only measures the marginal fuel cost – to imply negative margins.
Coal plants have many other marginal costs of production such as emissions credits and ash removal that must be incorporated into any plant-specific measure of energy margins.
In addition, coal plants must cover fixed costs that, depending on the capacity factor, can range from an equivalent of $10 per megawatt-hour to $30 per megawatt-hour.
Thus, any dark spread less than $10 per megawatt-hour implies negative energy margins, and a dark spread less than $20 per megawatt-hour implies potentially unsustainable operations over the longer term.
A new term is required to understand the new economics of U.S. coal-fired generation. During the past few years, coal prices have increased and natural gas prices have decreased.
Increasing coal prices are being driven by higher marginal costs of production and increased global demand.
Decreasing gas prices are driven by the production of gas from new shale and combination plays. The difference between the costs to produce electricity using natural gas vs. the cost to produce electricity using coal has flipped: Coal-fired generation is more expensive than combined cycles. This relationship can be measured using a new term:
Bed Spread = Dark Spread — Spark Spread
= PElectricity — PCoal x 10,000 – (PElectricity — PGas x 7,000)
= PGas x 7,000 – PCoal x 10,000
Figure 3 shows the bed spread for each of the Northeastern markets. As illustrated, the relative cost of coal to natural gas has increased significantly; coal is on par with or more expensive than natural gas as a baseload unit fuel source during all months except the winter peaking season when gas prices spike.
The result is that coal plants are at the money as opposed to their operationally and economically preferred historic position of being in the money with respect to the electricity supply curve.
If gas prices continue to decline and coal prices maintain or increase, some coal plants will be more expensive than natural gas combined-cycle peaks. This situation can be measured by how much the cost of coal converted to electricity varies from the cost of natural gas converted to electricity using a peaker plant. First, we need to define two new terms:
Head Spread = On-peak Price of Electricity — Off-peak Price of Electricity
Bedhead Spread = PGas x 11,000 – PCoal x 10,000
The difference between on-peak and off-peak prices is the head spread. This term is important for demand response units, batteries and other peak load-shifting resources.
The larger the difference between electricity prices during peak and off-peak hours, the greater the incentive to shift load away from peak to off-peak hours.
Historically, the average monthly head spread has varied by market and season. Characteristics of the market’s generation fleet and imports define the steepness of the supply curve and therefore the difference between peak and off-peak prices.
Seasonality defines shifts in the supply and demand curves, thereby impacting peak and off-peak prices.
In the Northeast markets, head spreads generally reflect the regional hub gas price times a heat rate ranging from 2,000 to 4,000 British thermal units per kilowatt-hour (i.e., the difference in heat rates between a natural gas peak plant and baseload combined-cycle units for those markets).
As gas prices decline, the head spread compresses. As low marginal cost renewable resource generation enters the supply curve, the different heat rates of the units setting the price in peak vs. off-peak hours converge, further compressing the head spread.
As peak demand declines because of a recession, increased energy efficiency or increased demand response, the head spread similarly declines. As shown in Figure 4, head spreads have fallen since 2008 because of all these factors.
As coal prices have increased and dark spreads have compressed, coal plants increasingly are setting the price during peak hours. The bedhead spread, the difference between the marginal cost of a natural gas peaker plant and a baseload coal plant, has declined and is below $20 per megawatt-hour during all but the winter months.
If gas prices continue to decline, coal plants might become more expensive to operate than peak plants when all of the variable costs of production such as emissions credits and ash removal are considered.
Existing natural gas-fired peaker plants could become a cost-effective source of new baseload generation.
To show how extreme the economics of electricity production has become, it is not outside the realm of possibility that certain coal plants also will be setting prices during off-peak hours. As mentioned, coal plants do not have operational flexibility to turn on and off quickly. As coal plants have become more expensive to operate during off-peak hours, bidding behavior has incorporated the marginal cost of startup and the opportunity cost of missing peak prices into off-peak bids, sometimes resulting in negative dark spreads. This leads the final definition.
A coal plant might be willing to accept negative dark spreads during off-peak hours to stay operational, forego startup costs and realize positive upside during on-peak hours.
If anticipated on-peak dark spreads are not sufficient to cover the negative energy margins, the plant should shut down.
A deadhead spread provides an indication of whether a coal plant should continue to operate in the near term.
To approximate the economic decision on whether to shut down and install a nitrogen blanket or go into wet layup, calculate the deadhead spread as the weighted average of the dark spreads over daily, weekly or monthly prices as follows:
Deadhead Spread = TPeak * Dark Spread Peak + TOff-peak * Dark Spread Off-peak
= [(TPeak /TTotal)] * PPeak + [(TOff-peak /TTotal)] * POff-peak — (PCoal x 10,000)
The decision to shut down a plant temporarily must be made before actual shutdown. For this reason, the deadhead spread could be calculated using month-ahead on-peak and off-peak strips traded on the futures market.
Such an approach, however, would not consider the option value associated with continuing to operate the plant, given price volatility and the corresponding potential for positive energy margins within the month.
Such optionality would need to be built into the decision process, creating a tendency to continue operating even if monthly futures markets indicate a negative deadhead spread ahead. In addition, start-up costs and other re-commissioning expenses would need to be considered as part of the economic analysis.
Last, the specific plant operating characteristics and other variable costs associated with the plant would need to be incorporated, making the operating decision a plant-specific calculation. That said, a deadhead spread using historic prices provides an indication of economic conditions (see Figure 5).
The deadhead spread can provide a standardized metric for what futures prices imply for coal plants in general. In combination with a fundamental price projection, this equation can be used to provide a high-level test of the implication of alternative scenarios on coal plant operations.
In making such decisions, however, realize that such decisions are plant- and location-specific. In addition, changes in the value of assets that are at the money fluctuate dramatically with the smallest change in conditions (i.e., in trading parlance, the gamma is large). A temporary decision to shut down should be re-evaluated regularly with diligent attention paid to market conditions and changes to operating costs.
Energy market conditions during the past few years have had dramatic implications for wholesale electricity markets. These changes are perhaps best illustrated by the relative change in position between natural gas and coal-fired generating units.
Although dark spreads and spark spreads have been compressed, a new set of metrics better illustrates the upheavals being experienced by generators in the power sector.
Bed spreads have become negative in some markets; implying coal is more expensive than natural gas for purposes of generating energy.
Head spreads are declining, limiting the potential upside for units on the margin during off-peak hours.
Bed head spreads are declining, implying the potential for coal plants to either set the price during on-peak hours or be displaced by natural gas peaker plants. Natural gas peak plants could start to look like baseload generating units if gas prices continue to decline.
Deadhead spreads indicate coal plants might be better positioned as seasonal peak plants, shutting down during shoulder months and operating only when the higher prices during seasonal peaks justify their operation.
The volatility of the underlying fuel prices makes such prognostications perilous. Yet plant owners and operators must make economic decisions.