Risky Business: Energy Price Volatility Concentration and its Implications for Revenue Risk for Battery Storage Systems in Different Power Markets

Energy Price Volatility Concentration:  Key Takeaways

As energy price volatility grows with renewable buildout and ancillary prices decline with battery storage buildout, battery revenues will increasingly rely on energy arbitrage as a revenue source.  At the same time, ISOs can vary widely in their distributions of volatility throughout the year, leading to significant differences in revenue risk and variability between markets for battery storage system owners even when exhibiting similar annual price volatility metrics (and battery storage revenues).  Quantifying how market volatility is distributed across the year is critical for understanding which markets are riskier and which are more stable.

• In markets with highly concentrated energy price volatility, revenues are highly dependent on infrequent grid or weather conditions that may or may not occur within a given year, or that may occur when a battery storage system is depleted or experiencing an outage.
• The Ascend Analytics RTB Distortion metric (introduced for the first time in this article) quantifies the degree of volatility concentration and can be used to understand differences in revenue risk and uncertainty between markets.
• SPP displays the most persistent volatility throughout the year, while ERCOT exhibits the highest dependence on extreme conditions.  CAISO exhibits volatility concentration between ERCOT and SPP, with some years closer to ERCOT and recent years more similar to SPP.

Understanding the Dynamics of Energy Market Volatility

Ascend Analytics identifies four main drivers of price volatility, each of which is amplified by renewable generation.  

1. ‍Congestion occurs when there is a regional imbalance between electricity supply and demand, is location-specific with frequent occurrence.  Renewable intermittency can cause frequent over/undersupply conditions in a given location as renewable resources often concentrate in areas with a strong renewable energy resource and open land, both of which typically correlate with low load.‍
2. Scarcity occurs when power supply is tight relative to load and is infrequent but large.  Renewable growth can drive thermal retirements, but renewable production during peak load can vary and be difficult to predict, both of which lead to challenges in resource adequacy planning.
3. Forecast error is the difference between day-ahead and real-time net load and occurs continuously.  Renewable power output is more difficult to predict than thermal output, leading to larger forecast error during periods of high renewable generation.‍
4. Net load ramp is the movement through the electricity supply stack as net load changes, which occurs daily with large impact.  Renewable output can vary greatly over short time periods (e.g., sunset), leading to large and concentrated swings in net load.

While energy price volatility can be measured in a variety of ways, Ascend Analytics’ preferred metric is ‘Real-Time Top-Bottom 120 minutes’ (RTB120), which measures the cumulative daily two-hour non-contiguous top-bottom spread in the real-time electricity market. This approach sums the highest 24 five-minute prices of each day and subtracts the lowest 24 five-minute prices of the day, which can be interpreted as the revenue of a two-hour battery storage cycling once every day with 100% round-trip efficiency and perfect daily foresight.

Historical Power Market Volatility in US ISOs

Figure 1 shows historical power price volatility for representative hubs in each of the seven U.S. ISOs: CAISO SP15, ERCOT South, SPP South, MISO Indiana, PJM East, ISO-NE Mass, and NYISO Zone J. CAISO, ERCOT, and SPP show significantly higher volatility than the other markets over the past six  years, and price volatility was amplified in all markets in 2022 due to elevated gas prices.

The annual volatility metrics shown in Figure 1 reflect average price volatility across the year, but ISOs also differ in their distribution in daily price spreads because each power market has a different distribution of the volatility drivers listed above.  Moreover, annual volatility can be heavily affected by extreme weather events such as heat waves and winter cold snaps that can lead to transient but concentrated periods of very high price volatility.  Markets can therefore have similar annual volatility metrics (and battery storage revenues) while exhibiting very different distributions in volatility throughout the year.

‍Figure 2 shows the probability distribution of daily RTB120 values in 2022 for each ISO.  The mode (or  peak) and median of the distribution are less affected by extremes and thus better reflect ‘typical’ daily price spreads. During 2022, ERCOT had one of the lowest modes among the ISOs despite having one of the highest averages, while CAISO’s and SPP’s modes were the two highest.  Similarly, ERCOT had  the lowest median value while SPP had the highest.  SPP displays consistently high volatility because its high wind penetration drives persistent electricity price volatility as wind generation ramps up and down.  In comparison, ERCOT experiences lower consistent volatility than SPP, but exhibits days of extreme volatility that accumulate large energy arbitrage potential over short time periods.  When a heat wave occurs in ERCOT and the system experiences an extreme net load ramp as the sun sets with low wind generation, RTB120 values can climb to levels nearly 40 times higher than the median day.  CAISO can experience similar periods of extreme volatility during heat waves.  These differences in price spread distributions drive a need to better characterize differences in the prevalence of extreme conditions between markets and the implications for revenue risk and variability.

The effects of extreme volatility can be easily seen in Figure 3, which shows a timeline of how volatility accrued in each ISO during 2022. '‘Jumps’ in the timelines indicate brief periods of extreme weather driving very high volatility.  ERCOT exhibits a large jump in mid-July, where just three hot days contributed 14% to its annual total RTB120.  CAISO saw a large jump during a heat wave from August 30 to September 8, ten days that accounted for 42% of CAISO’s total RTB120 volatility in 2022.  SPP, the most stable of the markets, saw no major ‘jumps’ while still ending the year with the highest total RTB120.

Energy markets with large portions of revenue concentrated in a small number of days will be prone to boom-bust cycles in which annual variation in weather, load, renewable generation, and outage conditions can yield significant year-to-year variation in revenue.  Figure 4 is a sorted version of Figure 3, displaying  the cumulative volatility throughout the year in increasing order from the days with lowest to highest RTB120 spread.  The degree of upward turn at the end of the year indicates how concentrated the annual total is in the highest volatility days.  Markets with steeper slopes toward the tail end, such as ERCOT, indicate high volatility concentrated over shorter periods  of time, while markets with straighter curves, such as SPP,  have a steadier path to reaching their annual total volatility.   For example, the 30 highest volatility days in SPP, ISO-NE, and MISO account for less than 30% of the year’s total volatility. However, these two top 30 days constitute over 40% of CAISO’s total and nearly 60% of ERCOT’s.  Looking only at annual volatility obscures the risk of these major volatility events  not occurring, which would yield significantly lower revenue potential in any given year and more unpredictable cash flows for battery storage operators.

‍RTB Concentration™ and RTB Distortion™:  New Metrics for Characterizing Power Market Volatility

Because of the impact of volatility distribution on battery storage revenue risk, Ascend Analytics has developed metrics for the concentration of annual volatility to better quantify how the distribution of volatility and arbitrage revenues varies between markets.

The first metric is the ‘RTB Concentration™’, which measures the number of highest-value days needed to add up to 30% of the total annual RTB120 for each market.  Figure 5 depicts this metric for 2022, where a lower number indicates a concentration of value in a smaller number of days.  ERCOT and SPP again land in opposite ends of the spectrum, with ERCOT getting 30% of its total RTB120 in only six days and SPP needing 45 days.  The higher number of days for SPP indicates a more stable market where volatility and revenues are more consistent throughout the year instead of concentrating in brief intervals.

‍

Ascend Analytics is also introducing the ‘RTB Distortion™’ metric, which measures the mean daily RTB120 divided by the mode.  Dividing the mean by the mode indicates how much a market’s most volatile days are driving the average relative to the most typical days, capturing both the frequency and magnitude of the extremes.  A high value indicates that extreme values are driving the mean whereas lower values indicate that the mean is driven more by consistent daily values.

Figure 6 shows the RTB Distortion across all markets, with SPP and ERCOT yet again appearing on opposite ends of the spectrum.  ERCOT exhibits both high distortion overall and significant year-to-year variation, indicating that ERCOT is always dependent on extreme events, and some years are especially concentrated.  SPP exhibits the lowest RTB Distortion, while PJM and MISO exhibit the highest outside of ERCOT. CAISO historically has had high RTB Distortion, but  recent years have been lower as the system has been run more conservatively (which has been discussed in previous Ascend Analytics Market Intelligence Webinars).

‍

The Growing Importance of Battery Storage Revenue Optimization in Concentrated Markets

Understanding RTB Distortion and RTB Concentration is critical for understanding - and better managing - revenue risk and uncertainty for battery storage systems across different markets.  When battery storage revenues are concentrated in a small number of days during extreme conditions, cash flows could exhibit significant year-to-year variability depending on whether those conditions occur within a given year.  This variability in revenues could have a significant impact on the timing of returns to investors and the ability to repay debt. Moreover, battery storage revenues in concentrated energy markets can be severely affected by the project’s outage conditions, which may be more likely to occur during those periods of extreme heat or cold that drive high price volatility. State-of-charge management will also be critical, with batteries that dispatch prematurely severely underperforming the potential arbitrage opportunity and scarcity conditions increasingly being associated with empty batteries as energy storage buildout grows within a given market. These factors underscore the importance of using tools like Ascend Analytics’ Smartbidder™ to optimize energy market bid strategy and participation for operating resources, especially in markets with high volatility concentration and distortion.

‍