14 April 2022
Ensuring that power comes from the lowest cost sources is a clear priority for achieving a Net Zero power system in 2035 at least cost. This means ensuring that market prices reflect the true cost of producing power.
This was the motivation for changes to Contract for Difference (CfD) contract terms for Allocation Round 4 (AR4) – to remove a plant’s subsidy if the price they are assumed to receive from the market turns negative.
However, this change could have a huge impact on how power is traded, and we believe these contracts need extra reform for future allocations rounds.
In this blog, we outline the issues with the current contracts. We also argue that without reform, markets will not be able to operate efficiently, generation and supply portfolios will be harder to manage, and customers will bear the cost of increased uncertainty and inefficiency.
Different time horizon, same price
The key to the issue is that the rules apply differently to separate markets where power is traded.
Power is traded over different time horizons as suppliers and generators manage their exposure to future price changes. However, the underlying commodity for each of these products is the same unit of power at the same time. The theory goes that the prices in these markets should align on average if expectations of future prices represent the true average.
The theory turns out to be true in power markets. Taking prices from day-ahead, intra-day and balancing (cashout) markets between 2016 and 2020, the average monthly price in these markets align closely as the chart below shows. The same behaviour was seen in 2021 but has been excluded from the graph as prices were much higher.
Maintaining price alignment
So, what is the balancing market and how do generators and suppliers change their behaviour to trade power at the same price as day-ahead and intraday markets?
In the balancing market, National Grid ESO take actions to align supply and demand, and parties are required to buy or sell power at a cashout price to settle any imbalances in their traded position.
The cashout price is set based on the most expensive action required to match total supply and demand. If there is a need for more generation (the system is short) then plants will need to increase generation, and vice versa if the system is short.
Whether more or less power is required than was traded depends on unexpected changes in the supply and demand balance, for example changes in wind generation, demand, and plant outages. However, parties can also influence this by intentionally trading too much or too little.
While the cashout price is uncertain, changing the system imbalance alters the chance of the system being long or short and influences the distribution of cashout prices. This is how parties ensure that average prices in the intraday and balancing markets align and there are no arbitrage opportunities based on systematic price differences.
In theory, as the intraday market price changes, the distribution of cashout prices changes too as different generators will be required to change their generation. To account for asymmetry in cashout prices, the system imbalance will need to be adjusted. Net Imbalance Volume (NIV), the measure of system imbalance, is positive if the system is long, that it is has too much generation, and negative when short.
The chart below shows how the average NIV changed as the intraday market price changed in 2020. It also shows the average differential between the intraday price and the cashout price when the system is long and short. These differentials give an indication of the cashout price asymmetry that the market must account for.
We can see that as the intraday price increases, the distribution of cashout prices changes. The price differential when the system is long grows, and the differential when short reduces. To account for this change and align the expected cashout price with the intraday price, the market moves from a positive NIV through to a negative NIV as intraday prices increase. In doing so, the likelihood that the market will be short increases and the expected difference between cashout and intraday prices targets 0.
Over the long term, as commodity prices and the mix of technologies in the generation fleet changes, we would expect to see that this average NIV position changes. This behaviour is shown in the chart below between 2016 and 2020.
So what for new AR4 CfD rules?
The new AR4 rules state that support payments will not be paid in any period where the day-ahead market price (i.e. their reference price) is negative. If this rule continues to be applied for future CfD auctions (AR4+), it will apply to all new wind and solar generation, and they will not be willing to sell power at a negative price in the day-ahead market. This will effectively set a floor price of zero in the day-ahead markets (when any AR4+ supported plant are generating).
However, power can be traded after the day-ahead market price is set, through the intraday and balancing markets. In these markets, prices will still be able to go negative if the day-ahead price was at least 0 and their support payment is guaranteed.
The future power system is set to see many periods of oversupply of renewable generation, most of which will be supported by CfD contracts with these rules. In these periods, day-ahead prices are likely to be zero (putting aside issues around tie-breaking rules). However, the price of power in the intraday and balancing markets is all but guaranteed to be negative as these same generators will be willing to generate at a negative price to ensure they receive their support payment.
Which raises the question - why would anyone buy power at a positive price in the day-ahead market when it is very likely to be negative in the intraday and balancing markets?
An alternative approach would be to not buy any power from these generators in the day-ahead market and only buy power at a negative price in the day-ahead market. However, in this scenario, the market will require huge quantities of power from the intraday and balancing markets. This will come from the unused generators at their unsubsidised, positive cost.
So why would any generator sell their power at a negative price in the day-ahead market when it can receive a positive price in the intraday and balancing markets?
The only remaining option is that all parties will choose to forego the day-ahead market entirely and settle large imbalances in the intraday and balancing markets.
Market outcomes when AR4+ generation would be marginal in day-ahead markets
What impact would this have?
Consumers will benefit if power is traded more efficiently. This involves markets where power can be traded easily and generators and suppliers can discover and realise a more accurate price for power in each half hour period. This is at risk if the day-ahead market doesn’t function when renewables are oversupplied, and more power is traded closer to delivery in intraday and balancing markets.
If the AR4 rules are not reformed, then power will not be traded as freely or efficiently. This inefficiency will be paid for by suppliers and generators and ultimately passed to consumers.
How could this be fixed?
If these contracts should continue to disincentivise negative prices, most reform options involve adjusting the reference price that is used for contracts.
The reference price should represent the average price that the generator receives in the market for their power. It currently assumes that all renewable power is traded in the day-ahead market.
However, this could be updated to reflect the fact that some renewable power is sold in the intraday and balancing markets as well. This would involve setting the reference price at the volume-weighted average of intraday and balancing market prices. This would also disincentivise negative prices in the intraday and balancing markets as well as the day-ahead market.
This may have some technical challenges, but the current reference price accounts for volumes sold in multiple day-ahead markets with different prices and could be extended to include intraday and balancing markets.