Electricity (cross-product) price volatility[1] has historically been closely linked to overall price levels. This trend seems to have ended in 2024 in Germany, as low marginal cost renewables are pushing the overall wholesale price level down, while peakers such as gas and batteries need to finance their investments in relatively few but increasingly expensive production hours. In this article we explain this mechanism and argue that this decoupling is the markets’ way for asking to build short term flexibility.
Figure 1 shows the average Day Ahead daily spot price levels for the German market in blue and the daily Day Ahead High-Low Price spread as measure of volatility in green. One can see that the two moved together over the last 5 years, however in 2024 they are decoupling.
[1] In this article we define volatility as the within day product price spread between the hours in Spot markets. Strictly speaking volatility in finance refers to the movement of a price curve over time. For instance, the hourly product 12-13 can be traded at various prices from week ahead-to day ahead and on Intraday until delivery. This is NOT the kind of volatility that we are referring to here. We mean the price spreads BETWEEN various products at any given point in time. The two concepts are closely linked; however we want to focus on cross-price volatility.
Figure 1: Price Levels and cross product price volatility
Why high prices are correlated with high volatility
Think of two balls bouncing in a contained space like the box in t1 (see Figure 2). We think of cross product volatility as the average measured distance between these two bouncing balls d1. Assume we increase the ceiling or the height of my box in t2. Now with all other things equal both balls have more space to move around so the average distance between my balls will increase so that in t2 d2 is greater than d1. This analogy works well for electricity prices, where an increase in the overall price level due to an increase in input costs such as coal, gas or CO2 simply lifts up the ceiling and allows prices to fluctuate more wildly in an open space.
Figure 2: Volatility Illustration
Prices are function of supply, demand and their respective elasticities
Let us take a step back. In economics, prices are the outcome of the interplay of supply and demand and their respective elasticities at various price levels. To put it very simply, prices are the result of producers wanting to supply a certain amount of electricity at a certain price level and the consumers’ willingness to purchase at these price levels. Elasticity of demand measures how much electricity consumption would change when prices change.
Demand is usually inelastic because people and businesses need electricity regardless of price and often do not see real-time market prices when making their consumption decisions as they are hedged in fixed price contracts by their suppliers.
Elasticity of supply measures how much electricity production changes when prices change. Supply depends on the merit order, where cheaper sources (like renewables) are used first, and expensive ones (like gas turbines) are used later as demand rises. So, when we are looking at differences in prices we are necessarily looking for changes in demand and supply and their respective elasticities.
So why do higher price levels tend to cause higher price volatility? We forget about renewables for a second and just look at what a supply price shock does to price volatility. The energy crisis of 2021 – 2023 serves as a textbook example.
First, we look at a very cozy merit order supply curve from 2018 when gas was trading at around 18 EUR/MWh and CO2 Costs were at around 16 EUR/ton. A drop in demand from 60 GW to 40 GW would have led to a drop in prices of 20 EUR. Or to put it in other terms: Supply was quite price elastic: i.e. a change of prices of 20 EUR leads to a high change of supply of 20 GW. Figure 3 illustrates this point. A demand of 60 GW (without renewable production) would have shifted us from gas trading at 50 EUR/MWh into coal trading then at around 30 EUR/MWh.
Figure 3: A change in demand causes moderate price shifts at low price levels (Source: FfE)
Now we jump into the energy crisis caused by the Russian-induced war in Ukraine in 2022. The war induced gas supplies to freeze up, lifting up gas prices to 123 EUR/MWh (in this example) and thereby the marginal production costs of gas and also other conventional energy sources such as coal [2]. In addition, CO2 prices had risen to 90 EUR per ton further lifting up the merit order curve.
Looking at Figure 4, first note that the price level on the Y axis has changed quite a bit and there we can observe our bouncing balls effect. We increased the ceiling of our box and now the average distance between offers on the merit order are simply farther apart all other things being equal . The merit order has also a slightly different shape as some nuclear had left the market, however this is not material to the points made in this article. The price at 60 GW demand was at 470 EUR/MWh and at 280 EUR/MWh at a demand of 40 GW. Observe how the change in demand from 60 GW to 40 GW now leads to a price change of 190 EUR compared with the 20 EUR before. A massive increase that can be seen on Figure 1. In 2018, a 20 EUR price change was associated with a 20 GW change in supply while in 2022 a 20 EUR price change would have barely moved supply. This means that the elasticity of supply decreased immensely.
[2] Coal can act as a substitute for gas (and vice versa) so an increase of cost in one can lift up demand for the other thereby also increasing its prices.
Figure 4: A change in demand causes high price shifts at higher price levels
It is interesting to note that it would be theoretically possible to have an upward shift in prices, without an increase in volatility. If we just shifted the 2018 merit order in Figure 3 up by 400 EUR/MWh at all places, then the change of demand would still only lead to a 20 EUR drop in prices, but in practice we see that the more space prices have to move, the greater is their average distance. Prices tend to bounce in the space that is available to them.
Hence looking at Figure 1, we can now understand why and how higher price levels tend to lead to higher cross product price volatility. But now to the key question: Why are these two decoupling since 2024? The simple answer is, renewables, but we will go into more depth.
Adding low-margin renewables into the mix
So how do renewables change all this? Well, first renewables are very cheap in marginal costs. For all purposes we can assume that wind and solar marginal costs are zero. This means that they drive out all other sources of supply whenever they are available. This decreases the annual production hours of all conventional producers meaning that even without any political intervention, baseload technologies such as nuclear (and lignite) are driven out of the market by renewables as they need lots of production hours to be profitable. This means that the conventional supply stock shrinks and requires more flexible peak load capacities, mostly gas (and batteries) to fill the decreasing gaps in renewable production.
In addition, the supply curve of conventional technologies gets steeper i.e. elasticity of supply decreases. You can think of this effect as a shift from long run marginal costs to short run marginal costs which tend to be higher across various markets. The logic is that suppling lots of additional output within a short term of time (less runtime hours) is more expensive. Imagine having to serve 100 guests in a kitchen built for 30. Each extra meal will have higher marginal costs as too many chefs crowd a very small kitchen. For gas peakers the same effect can be seen because cycling costs (ramping up and down for short periods of time) need to be recuperated and also their capital costs need to be financed in fewer hours.
We need some microeconomics to illustrate the situation. In a conventional power supply world, we have a quite constant merit order curve that gets activated depending on the hourly level of demand. So, changes in demand drive differences in prices.
Let us take the 20 GW shift in demand within a day from before and theorize around what happens. We see a moderately elastic conventional supply curve that reacts to prices. As demand decreases prices fall until some conventional plants adjust their production down just enough to cover demand. This is the classic merit order model and Figure 5 is merely a simplified illustration with elastic supply and very inelastic demand at two points in time.
Figure 5: Moderate price shifts in a conventional supply universe
Now we add large capacities of wind and solar into the mix as Germany has done over the last 15 years. In a renewable supply world, we must look increasingly more at shifts in supply in addition to shifts in demand in order to understand price movements. In Germany, production from wind and solar fluctuated between 0.5 GW to 68 GW last year indicating extreme short-term supply shocks in a country with average demand of 55 GW. So, imagine a situation in which most of demand is covered by very cheap wind and solar supply. Without any market distortions, as often seen in Germany, where negative prices are common, renewables curtail their production at roughly 0 EUR/MWh. So, at this point supply is extremely elastic: a switch from 1 EUR/MWh to -1 EUR/MWh could mean a supply change of tens of GW of power. However, renewables can not ramp up, so the residual demand i.e. the difference between total demand and renewable supply still needs to be covered by conventional production. In our example in Figure 6, high demand in t1 is covered to a large part by renewables and then some relatively expensive conventional supply. The supply curve is steeper than in the previous example as the power plants are only needed for a few hours and therefore have higher short run marginal costs.
As demand decreases from Q1 to Q2, renewables are fully able to cover all demand and prices drop down to zero. Because of very low prices at t2, the average price level in the renewable universe is lower than in the conventional universe above despite P1 being higher than before i.e. renewables drive down the average price level.
Figure 6: Extreme price shifts in a renewable supply environment
What we can see very clearly though is that the difference between P1 and P2 is much larger in the renewable universe than in the conventional universe. Price volatility has increased.
After some theorizing we should now understand what we have termed the “The Great Decoupling”: It is not driven by an ending of the relationship between high prices and high volatility. This relationship should generally remain intact, because the reasoning still holds in today’s world. However, we see a new force that is pushing overall price levels down while at the same time increasing volatility, as we will tend to have very expensive power for only a few hours during the day. This force has a name, and it is wind and solar power.
The need for flexibility
So, what does this “Great Decoupling” signal to the market? The answer is very simple, and it is a cry for flexible demand from a market that speaks in prices. In a renewable power system, there will be lots more capacity than average demand meaning that when renewables are available, prices will be close to zero (or negative) most of the time. This is not a market failure, or a “cannibalization” effect, it is just very simple demand and supply. But when renewables are scarce, conventional power plants will be needed for only few hours a day or year, and in order to recoup their investment and cycle costs, they will need high prices to finance themselves. How to finance production capacities in this environment is a fascinating, however different question.
In order to take advantage of this new and quickly changing environment, we need to push significant short-term demand from electric vehicles, heat pumps and flexible industrial processes into the times when renewables (especially solar) are available. Forget about your fridge and washing machine, they don’t make a dent. Prices should be allowed to send this signal across all customer segments which requires smart metering and flexible tariffs in order to give the proper incentives to customers.
The technology solution that most benefits from this increasing volatility are of course energy storage system or batteries (BESS), that can really be thought of as flexible demand: They increase load in times of high renewable supply and decrease load in times when renewables are scarce. With current battery configurations they are not (yet) suitable to get us through any prolonged “Dunkelflaute”, but they are the cheapest answer to flexible demand that we have seen at this point in time.
What we have learned
Observe that in 2024 German power price volatility is decoupling from overall power price levels. This does not mean that higher price levels due to high gas or CO2 no longer lead to high volatility; they do. We rather see an additional effect with the increase in renewable capacity. On the one hand average power prices are pushed down due to many hours with very low prices. On the other hand power prices are momentarily lifted up due to a steep short run supply curve for very few hours. So, we are depressing average price levels while increasing volatility. This “Great Decoupling” is great news for everyone working on flexibility: be it demand side solutions or battery storage. Their time has just begun.