Relative profits for some natural gas- and coal-fired generators in several Midwestern and Mid-Atlantic states may have decreased since 2016 because of higher natural gas and coal prices and lower wholesale electricity prices. A common measure of profitability for power plants within a region is the difference between their input fuel costs, such as the cost of coal or natural gas, and their wholesale power price.
For electric power generation fueled by natural gas, this difference is called the spark spread; for coal, the difference is called the dark spread. Spark spreads and dark spreads in the first part of 2017 were lower than the 2016 averages in the PJM Western hub, which covers electricity generation in parts of several Midwestern and Mid-Atlantic states.
Changes in spark spreads and dark spreads for a given electricity power market indicate the general operational competitiveness of coal-fired or natural gas-fired electric generators in meeting the market’s electricity demand. These spreads are calculated by comparing the day-ahead, wholesale power market price with the delivered input price of the fuel, and are adjusted for the energy content of the fuel and the relative conversion efficiency of power plants. These values can then be compared with wholesale power prices, which, in this example, are the average day-ahead prices at the PJM Western hub.
Delivered coal prices vary among coal supply regions based on the quality of coal, the transportation costs of shipping the coal, and other contract terms. Natural gas prices vary regionally and are calculated using spot market prices, which can be volatile on a day-to-day basis.
Coal and natural gas have different energy contents, and the power plants using these fuels have different heat rates, or energy conversion efficiencies. For this reason, spark and dark spreads are location-specific and reflect the characteristics of the fuels and the technical specifications of power plants in a given market.
For example, natural gas consumed in the electric power sector in the PJM region has an estimated heat content of 1,033 British thermal units (Btu) per cubic foot of natural gas. So far in 2017, the average price for natural gas in this area has averaged $2.54/million Btu, based on prices at the Texas Eastern Transmission Market Zone 3 (Tetco M3) trading hub, which generally reflects natural gas prices in Pennsylvania, New Jersey, and New York. Spot prices within PJM vary widely because of pipeline constraints transporting natural gas from production areas in the Appalachian region to different markets.
Natural gas combined-cycle plants in the PJM region are generally expected to produce one kilowatthour of electricity for every 7,300 Btu of natural gas. In the PJM region, combined-cycle plants are more commonly operated in direct competition with coal-fired generators. At the national level, average heat rates for all natural gas-fired generators have decreased over time (i.e., become more efficient) as more efficient natural gas power plants such as combined-cycle units have been installed and as older and less efficient units have been retired or converted to more efficient units.
Coal consumed in the PJM region has an estimated heat content of 22.5 million Btu/short ton, representing the consumption-weighted average heat content of various coal types. Coal prices in the region have averaged about $55/short ton so far this year. Adding in an assumed coal transportation cost of $17/short ton, the estimated delivered coal costs translate to about $3.20/million Btu, or about 34% higher than delivered natural gas costs. PJM-region coal plants are, on average, less efficient than natural gas combined-cycle units, requiring about 10,500 Btu of coal to produce one kilowatthour of electricity.
Although PJM-region spark and dark spreads appear to indicate that natural gas-fired units have been more profitable than coal-fired units recently, many factors affect these calculations, including the selection of representative fuel prices, generator heat rates, fuel delivery costs, and time of year considered.
Principal contributor: Augustine Kwon