The evolution of real-time locational marginal prices (LMP) in PJM Interconnection’s wholesale electricity markets demonstrates the characteristics of a well functioning and competitive market; prices are driven by supply and demand fundamentals. Although critics of wholesale electricity markets frequently have pointed to increases in LMP as proof that the wholesale markets aren’t working, in reality, the markets are working as they should. Increases in LMP are driven by higher demand for electricity or increasing fuel prices. Critics of wholesale markets, however, often fail to acknowledge that decreases in demand and fuel prices also immediately lead to declining LMP.
The connection between decreasing demand and fuel prices and lower LMPs is strong evidence of how well PJM’s wholesale electricity markets are functioning. Often lacking in the debate about the performance of wholesale electricity markets, such as PJM’s, is an objective examination of LMP data side-by-side with fuel price and demand data. Such an analysis of publicly available data shows the unmistakable effects of fuel prices and electricity demand in driving LMP in PJM’s wholesale electricity market. The dynamic nature of electricity prices in PJM since May 2005 clearly demonstrates this connection—with distinct periods of fuel- and demand-related increases and decreases over the last four years. The latest example is the significant decline in fuel prices and demand concurrent with the recession that has driven the marked decrease in LMP in PJM since July 2008.
Assuming demand is held constant, as fuel prices rise, LMPs also rise. Conversely, as fuel prices fall LMPs fall in response (see Figure 1).1 From May 2005 to January 2006, the trend in fuel and power prices was upward. During 2006 the trend in fuel and LMP prices was downward relative to 2005 prices. In 2007, fuel prices remained steady through September before trending upward, while LMPs trended upward through most of 2007. The increase in LMP while fuel prices remained steady through most of 2007 can be attributed to a 4.6-percent increase in average load over 2006. From the autumn of 2007 through the summer of 2008, LMPs and fuel prices increased rapidly, before the fast decline of LMP and fuel prices concurrent with the acceleration of the economic downturn beginning in early autumn 2008.
Averages on an annual basis also demonstrate the relationship between LMP and fuel prices (see Figure 2).2
The Independent Market Monitor for PJM consistently has recognized the influence of fuel prices on LMP in its annual State of the Market Report, and each year it adjusts the load-weighted average LMP using fuel costs from the previous year.3 Using fuel-cost adjusted LMP provides a consistent comparison of price performance between two years by holding fuel prices constant. Moreover, an examination of fuel-cost adjusted LMP provides insights about the influence of demand on LMP (see Figure 3).4 For example, if natural gas and other fuel prices had remained at their 2004 levels, the load-weighted average LMP in 2005 would have been comparable to 2004. However, in 2005, gas prices increased by $2.41/MMBtu on average, which led the actual load-weighted average LMP to increase more than $19/MWh. In 2006, the market witnessed a retreat in gas prices and load-weighted average LMP. Likewise, if the higher 2005 fuel prices prevailed, the load-weighted average LMP would have been higher—as represented by the fuel-cost adjusted LMP. In 2007, despite a relatively small change in fuel prices—as evidenced by the small uptick in gas prices—on a fuel-cost adjusted basis, LMPs were much higher than in the previous year. This relatively large increase in LMP in spite of steady fuel prices is explained by the load growth between 2006 and 2007.
In 2008, gas prices increased—leading to an increase in load-weighted average LMP—but on a fuel-cost adjusted basis, LMP was down from the previous year. This result was driven by a drop-off in hourly average demand from 2007 to 2008. Through 2009, the steep fall in gas prices led to a steep fall in load-weighted average LMP. On a fuel-cost adjusted basis, through June of 2009, fuel-cost adjusted LMP also is down from the previous year’s load-weighted average—again driven by a continuing decrease in electricity demand from the previous year.
While fuel prices explain a great deal of price movements, changes in demand also factor into LMP movements. As demand increases, LMPs also increase if fuel prices remain constant. In contrast, as demand decreases—as was the case during 2009 and the last half of 2008—LMPs also should decline if fuel prices stay constant. In addition to long-term trends, demand in the electricity market also shows seasonal variability when demand is higher during the hot summer months (e.g., June, July, and August) and during the height of winter (e.g., December, January and February) while the spring and autumn months experience lower demand as heating or cooling needs are reduced.
Figure 4 shows the hourly average load by month in PJM from May 2005 through May 2009 mapped against the monthly load-weighted average LMP over the same period.5 When average hourly load (shown by the red line) is at its highest during the summer and winter months, LMP (blue line) also is at its highest. Similarly, during the fall and spring months, when loads are at their lowest, LMP also is down relative to the summer and winter months.
Much like with fuel prices, it’s helpful to examine annual hourly average load alongside LMP on an average and fuel-cost adjusted basis (see Figure 5).6 As has been discussed in the context of fuel prices, changes in hourly average load also explain movements in LMP.
For example in 2007, the fuel-cost adjusted LMP was almost $10/MWh higher than the 2006 load-weighted average LMP. The explanation for the jump in LMP, holding fuel costs constant, is the more than 3,000-MW (e.g., 4.6 percent) increase in hourly average load from 2006 to 2007. The increase in average hourly load from 2006 to 2007, holding fuel prices at their 2006 levels, means that more expensive units had to be dispatched on average to satisfy the higher demand leading to the increase in LMP.
In 2008 and in 2009, fuel-cost adjusted LMP is below the load-weighted average LMP at the same prices from the previous year. In this case, declining load in 2008 and in 2009 explains this drop in fuel-cost adjusted LMP. For example, in 2008 the average hourly load was lower than the average hourly load in 2007. Holding prices at their 2007 levels, this means that less expensive units were required to satisfy demand on average. However, in 2008, fuel prices increased, leading to an increase in overall load-weighted average LMP in spite of the decrease in demand. Similarly, if hourly average demand in 2008 had stayed at the higher 2007 level, the fuel price increases observed in 2008 would have led to an even larger increase in load-weighted average LMP than the increase from $61.66/MWh in 2007 to $71.73/MWh in 2008.
Critics of wholesale markets only lament increases in prices and usually don’t recognize the periods when prices decline as they have at times during the past four years. In reality, LMPs in PJM are driven by supply-side fundamentals of fuel prices and demand-side fundamentals—as evidenced by LMP changes that correlate to the sharp increase in average demand in 2007 followed by declines in demand in 2008 and 2009 related to mild weather and the recession.
The Independent Market Monitor for PJM also recognizes both of these fundamentals in its annual State of the Market Report. Moreover, the market monitor always has found the results of PJM’s wholesale electricity markets to be competitive, meaning that the price outcomes are driven by competitive behavior and supply-and-demand fundamentals.
Discussions about how wholesale electricity markets are working should begin with a discussion of the underlying fundamentals driving changes in LMP. When that occurs, it’s clear that LMP responds to those fundamentals and that wholesale markets produce competitive results.