Thoughts on the contract-for-differences (CfD)
On the importance of captured prices for CfDs and some associated reflections
In this post, we will dig into the two-way contract-for-difference (CfD). This support scheme has been promoted by the EU Commission as a means to bring online new investments in power generation (wind, solar, geothermal, hydropower without reservoir, and nuclear) as explicated in Article 19b of the regulation to improve the Union’s electricity market design.
What is a CfD?
The EU Commission is defining a CfD as follows:
A two-way contract for difference is a contract signed between an electricity generator and a public entity, typically the State, which sets a strike price, usually by a competitive tender. The generator sells the electricity in the market but then settles with the public entity the difference between the market price and the strike price. It thus allows the generator to receive a stable revenue for the electricity it produces, while at the same time it provides a revenue limitation for generators when market prices are high. In a two-way CfD, if the market price is below the strike price, the generator receives the difference; if the market price is above the strike price, the generator pays back the difference.
The payments are therefore a function of the strike price and the reference price as can be seen in the following graphs.
The following schematic presents the relationships between the different parties. The generator is receiving two financial streams: one from the power markets for the sale of electricity and another one, from the counterparty, a general public entity. The latter is based on the difference between a strike price (resulting in general from a competitive auction and constant over time) and the market price. For intermittent renewables such as wind and solar, the market price will generally be the hourly prices of the day-ahead exchanges.
The CfD is removing entirely the market price risks from the generator. Therefore, in order to maximize its revenues, the generator would mainly focus on generating as much as possible regardless of the actual value of electricity on the market. Consequently, the generator does not seek to maximize the total value of solar generation1. For the generator, this transfer of pricing risk provides certainty and increases the project's bankability.
Interestingly, the EU Commission is envisaging the introduction of CfD as a means to avoid the windfall profits that happened during the energy crisis. As written in the Q&A:
Also, the reform introduces that obligation for Member States to provide public support for new investments in low-carbon, non-fossil fuel electricity generation in the form of two-way contracts for difference. The ultimate objective is to provide secure, stable investment conditions for renewable and low carbon energy developers by bringing down risk and capital costs while avoiding windfall profits in periods of high prices.
And it is then emphasized that consumers will benefit from lower prices from the introduction of CfD:
At same the time, the pay-out that is generated by CfDs when market prices become high will have to be used by Member States to directly lower the electricity bills of all electricity customers (including companies and industry).
This reduction in electricity bills is only possible if the captured price is higher than the strike price. This was the case during the energy crisis but will it continue to be the case in the coming years?
A week in May 2023
Hereunder we present the case of Hungary, a country not particularly at the forefront of the solar revolution2. On the week from Monday 22 May to Sunday 28 May, we observed the following prices on HUPX, the day-ahead market in Hungary (blue line). We can clearly see that prices tend to be lower when solar generation is at its highest, especially during the weekend (the last two days). On Sunday, prices were negative for 8 straight hours.
By knowing the total hourly solar generation and the hourly price, we can compute the value of this solar output as if it was sold on the market (similarly to a CfD). The purple line on the graph below represents this value. We can see that the value of solar is dropping, especially on the weekend, as the demand is generally lower.
The table hereunder presents the average daily prices and the solar-weighted average price. Over the week, the solar-weighted price was 54.17 EUR/MWh, while the average market price was 80.14 EUR/MWh. Interestingly, the solar-weighted price was negative on Sunday.
Example from the UK
The UK has been at the forefront concerning the introduction of CfD. They are currently in the fifth round of allocations. The results of the fourth round are available here with the associated strike price per technology. Interestingly, in order to calculate the budget for the public counterpart3, the administration is forecasting the captured prices for the different technologies. Two elements are important:
Prices for all technologies are going down with time4, probably reflecting the increased integration of renewable energy into the grid.
Captured price of intermittent renewables is expected to be lower than baseload power. For solar, the captured price would be 78% of the baseload price (27.01 GBP/MWh compared to 34.47 GBP/MWh).
Let’s calculate the financial flows based on the strike price of round 4 for solar (45.99 GBP/MWh) and the projections in the table above. For the year 2025/26, it is expected that for each MWh sold, the generator would have to pay on average the public counterpart 3 GBP (48.99-45.99). The year after, the strike price is expected to be higher than the captured price, therefore, the generator would receive an extra payment of 4.67 GBP for every MWh sold (45.99-41.32) on average.
Some thoughts and open questions
The CfDs are signed for a relatively long period. In the UK, the standard is a 15-year fixed-price contract. Therefore, a new auction launched today would result in a plant generating in a few years until after 2040. It is therefore critical for governments to estimate correctly the market prices for almost the next two decades, a task that is utterly impossible. Nevertheless, I will suggest some thoughts and further reflections on that:
If projections of renewables, especially solar which is set to quadruple in the EU by 2030 according to the sector proponents5, are correct, periods of extremely low and negative prices would be common6. A solar output of less than 100 GW across the EU at the end of May 2023 was enough to send prices to negative territory. Such a solar output would look very low by 2030 when the solar capacity would be multiplied by four7, leading to extended periods of low and negative prices.
Would the surge of offshore wind in the North Sea lead to extended periods of low prices as well? The countries bordering the North Sea have committed to installing 120 GW of offshore wind by 2030 and 300 GW by 2050.
Flexibility, such as storage, demand-side management, and electric vehicles would counteract low prices by boosting the demand during low-price hours. Nevertheless, the speed of renewable development is simply orders of magnitude higher than flexibility options. As an example, we can just consider the time it requires for the largest European economy to roll out smart meters, the fundamental building block to enable flexibility.
Could constantly cheap energy when the renewables are producing reverse the objective of energy savings? I discussed this eventuality here.
Should we incentivize generators to install storage on-site? Or to incentivize them to produce more when prices are higher8? If yes, how can we introduce such an incentive?
In general, there is a provision of no payment when negative prices occur (or at least for a number of consecutive hours). Could it represent a material risk for project developers?
Could an extensive use of CfD by the public authorities reduce the appetite for commercial PPA? Indeed, there will be less incentive for large consumers to sign a commercial PPA with a renewable producer if the electricity on the market is cheap when renewables produce.
The generator would also have to take into account the balancing costs, which might influence its generation profile.
Hungary has 3.9 GW of solar capacity installed, or 400 W per capita, still far from the European champion, the Netherlands with 1044 W per capita.
In the UK, it is Low Carbon Contracts Company.
Inflation has not been taken into account. The prices shown in the table are with a 2012 reference.
SolarPower Europe estimates that total market capacity would be 920 GW in their Medium Scenario 2030 overshooting the EU Commission’s REPowerEU strategy’s 750 GW solar target. The current installed capacity is 209 GW at the end of 2022.
This is already common in part of Australia, see my post here. Indeed, the average price at 1 PM was negative for Q1 2023 in South Australia.
100 GW would result in a utilization rate of only 11% across the EU in 2030 if the projection from SolarPower Europe is correct (920 GW). Such utilization rate would be achieved most of the hours during the daytime.
This could be achieved by different means: enforce maintenance only in periods of low prices, choose different technologies such as East-West orientation for solar, etc.
On your point 6, I think this is the key feature of the future power market. The prices will be bimodal, near zero when there is excess negligible marginal cost power available and a high price point when there is deficit.
The main cost/revenue risk is then how much of the time/energy produced is in each of these modes. And this answer depends on things outside each generators control (system balance issues).
Someone has to eat this risk.
Dispatchable offtake users, like H2 will be important, but H2 can't really sustain more than ~$10-20/MWh power cost, so power produced in the regime where there is enough to cover all normal users, and part of H2 capacity should be a low positive value. A poorly designed CfD could cost the public entity a LOT of $ effectively subsidizing H2 production under those conditions, even if there is a don't pay if proces are negative backstop.
Very interesting article.
Have you ever looked into the simultanious effects of the energy power market price being above the 2012-2013 CfD for offshore wind while below that for Hinkley Point C, should the nuclear power plant ever be started up?
I assume EDF and the French government will gladly fill the UK-France interconnectors with the prower produced by the plant.