Boom and Bust? Understanding the Power Plant Construction Cycle
cyclical pattern, producing periods of excess and lean capacity.
The staff also argued that this cyclical pattern would occur primarily because profitability of new generating units would remain heavily dependent on the prices during the summer peak demand season. They warned that summer peak prices would not necessarily reach a level to support new entry until reserve margins drop below the levels normally associated with reliable service. A.F.
The results in Figure 1 are typical of the short-term results from more detailed models.1 These models predict that the PX price will remain exactly at the levelized cost of a new CCCT, even as the horizon extends beyond the short term, but that result is questionable. This prediction is plausible only if we believe that new units will come on line in exactly the amount to counter the growth in demand. But in a competitive market, investors are not striving to bring exactly the right amount of capacity on line. Their goal is to invest in a profitable manner. We need to extend the simulation to learn if their actions lead to a stable price trajectory.
Long-Term Reaction. Figure 2 extends the simulation over 20 years to reveal that in the longer term, investor activity would likely appear in waves, causing cyclical variations in the PX price.
The first wave of investor activity appears around the years 2004_2009. Capacity in the permitting process reaches a peak of around 10,000 MW in 2008. Capacity in construction reaches a peak one year later. Installed CCCT capacity begins to grow around 2007. By this time, however, the PX price has climbed to $32 per megawatt-hour, well above the cost of a CCCT.
By the year 2008, installed capacity is growing rapidly. These new units allow the PX price to fall from the peak. By the year 2009, the PX price has returned to the cost of a new CCCT and is on a downward path. The PX price reaches a low of $22 per megawatt-hour before returning to a gradual upward trend. This renewed upward trend is similar to the upward trend at the start of the simulation; it is caused by the gradual growth in demand.
When the first wave of construction ends, about 14,000 MW of CCCTs are in operation, but there are no new CCCTs in the permitting process. Investors are reluctant to get involved in new projects until the PX price approaches profitable values. This reluctance does not reverse again until around the year 2013, when a second wave of activity appears. This wave appears near the end of the simulation, when the PX price peaks at $37 per megawatt-hour. This second wave of CCCTs comes on line around the year 2017, and drives the PX price down to $23 per megawatt-hour by the end of the simulation.
This base case simulation reveals an inherently unstable systemnot only do we see oscillations in the PX price, but the oscillations grow larger over time. The oscillations appear even though we have created an artificial environment with no surprises. There are no unexpected changes