Large volumes of renewable power capacity with rock-bottom marginal costs can push modern power grids into negative pricing. On windy days, the design of the UK’s CfD system could send prices spiralling deep into negative territory.
I noted in our recent blog on Contracts for Difference (CfDs) that generators in the UK are not paid a top-up to their strike price in a negative price period (NPP). This is defined as six consecutive hours in the day-ahead auctions for contracts from Round AR1 to AR3 when the auction clears below £0/MWh, or just one negative hourly price for AR4 contracts.
So far this hasn’t materialised as a major issue. But, unless the overall provisions of the CfD are amended in future rounds, dysfunctional behaviour and outcomes seem inevitable as new CfD generation comes online, and if legacy wind and solar is transferred to CfDs with the same provision.
Why is this? In practice, the marginal cost of production of wind and solar photovoltaic generators is slightly below zero, because there is some cost, effort and risk in shutting down; generators would rather pay a small amount for their power to be disposed of than to shut down. If Renewable Energy Guarantees of Origin (REGOs) are taken into account, any disposal costs for surplus power can be netted against the revenue from sold REGOs. So, in theory, CfD generators should be prepared to offer volume into the day auction at slightly below £0/MWh.
However, they’re very unlikely to actually offer below this level, as there is a risk of creating an NPP for their own CfD. This would mean they would receive no payment under the CfD. They will therefore offer at exactly £0/MWh out of commercial self-interest, and a recognition that their offers may contribute towards the auction clearing price. If they offer at £0/MWh or above, they know they can’t be responsible for creating a negative price; by definition, all the bids in the auction would have to be matched by offers below their price, at negative prices. Conversely, if they do offer below £0/MWh, they may well influence the auction to clear at a negative price – either at their own offer price, or just above. This means that we can expect CfD generators to offer into the auction at precisely £0/MWh, regardless of conditions.
Incentives to go to zero
The likely consequence of this is that there will be a strong gravitation of auction clearing prices to exactly £0/MWh.
When the market clears at a particular price – £0/MWh in this case – the auction operator will scale back either bids or offers, depending on whether there is more bid or offer volume submitted at that price or better.
As an example, suppose in an auction there are 2,000MW of bids and 4,000MW of offers placed at exactly £0/MWh for hour 1, as well as 25,000MW of bids above £0/MWh, and 24,000MW of offers below £0/MWh. Then all bids priced at £0/MWh or above will be satisfied in full, and all offers below will be satisfied in full too. But any offers at exactly £0/MWh will have to be scaled back to make the total volume of accepted bids and offers equal. The auction clearing volume will be 27,000MW, and all bids at £0/MWh or above will be satisfied at £0/MWh. All offers below £0/MWh will also be satisfied at £0/MWh; but only 3,000MW of the zero-priced offers can be matched with bids – therefore all offers at £0/MWh will be scaled back to 75% of the volume offered.
This means that, if it is windy, CfD generators will offer their forecast volume at £0/MWh in the auction, and will often succeed in selling some, whilst keeping the clearing price at £0/MWh, and preserving their CfD payment.
But they will now have some unsold potential generation output, which they will be keen to produce and sell (because it will be eligible for a CfD payment). Once the auction is over, the generator’s indifference price suddenly switches from just below £0/MWh to just below minus their CfD strike price.
Why is this? If, for example, their CfD strike price is £75/MWh and the auction clears at £0/Mwh, they will be eligible for a £75/MWh CfD payment on all output. They’ll also still get a REGO and face unwanted cost and risks if they do shut down. So, if they have to, they should be prepared to sell all remaining unsold output at just below -£75/MWh, knowing that the CfD payment and REGO will offset the cost of disposal.
The emergence of the Minus Strike?
What will that mean?
In these circumstances, the post-auction price will tend to crash after the auction towards exactly negative one times the most active CfD strike prices – let’s call this the ‘The Minus Strike’ level.
Speculators will be tempted also to offer volume into the auction at £0/MWh, hoping to sell some successfully (scaled back) and carry a short position into the day. They may even be tempted to offer into the auction a little at below £0/MWh, to avoid scaling back – although not too much, or too negative, as they won’t want to accidentally create a negative clearing price and make wind farms willingly turn off. Any volume they successfully sell at £0/MWh they will then look to buy back just above the Minus Strike level.
The more volume that gets offered at £0/MWh by wind generators in windy conditions, the greater the scaling back. Generators will observe the effect and deliberately offer more volume than they can generate. Speculators will offer more and more, knowing scaling back will be severe. There is no clear way to stop this upward spiral and, where there is a regular outcome, tens of gigawatts could be offered at £0/MWh in the auction, with more and more severe scaling back. Perversely, less and less volume may actually clear in the auction, as buyers may prefer to wait to buy below £0/MWh within the day.
Generators may be tempted to sell volume bilaterally in advance of the auction, but they will be wary that the buyers of that volume don’t offer the power back into the auction at negative prices. It’s therefore most likely that generators with CfDs will hold back until the auction, and offer at exactly £0/MWh.
It shouldn’t happen all the time. There are situations where a £0/MWh clearing price won’t occur. For example, there has to be a clear overhang of CfD generation compared to demand for it to materialise – but with plans to install over 100GW of CfD generation in the coming years, that will inevitably occur more and more frequently.
A nuance of the auction mechanism is that someone has to bid £0/MWh for at least one megawatt of capacity, or the auction will probably not clear at exactly that level. (If no one bids to buy at exactly £0/MWh, the most likely result is that the auction actually clears above £0/MWh, with offers at £0/MWh partly “paradoxically rejected” and scaled back just as they would have been at a clearing price of £0/MWh). The full range of auction tools, such as linked bids and offers, may also occasionally cause different results.
Prices can and probably will occasionally fix below zero which means the headline rate of the CFD will actually be lower; there are some forecasters that believe this could be as much as 5%, so a headline CfD rate of £80/MWh would actually deliver at £76/MWh.
Things may improve eventually when large-scale energy storage arrives in the form of hydrogen electrolysis, rebalancing supply and demand somewhat. But even an economically optimised grid is likely to have policy planned curtailment levels of 15% or more to ensure sufficient capacity is available, so times of surplus wind power will still occur.
Heading off dysfunction
Overall, it seems very likely that a dysfunctional pattern will occur during windy weather where prompt over-the-counter power prices are slightly negative, the auction clears at exactly £0/MWh, and post-auction spot prices trade at deeply negative numbers. This will lead to increased balancing costs for generators further reducing their net revenues and undermining the energy transition.
What’s the solution? It must lie in a change to the terms set out in forthcoming rounds of CfDs. If there is too much concentration of wind farms doing the same thing in response to the same incentives, there will be no moderating influence.
So, in response, the government could amend the terms of future auction rounds to change the definition of NPPs so that, below a low number above £0/MWh, no top-up payment is received. Or, that at exactly £0/MWh, only a scaled back proportion of output will receive a top-up payment, in line with auction oversubscription. That said any changes to the rules may create uncertainties over the revenues of projects that get a CfD.
There are many other alternatives, but policy makers need to start working on a remedy before NPP gaming materialises as a problem. They can’t be caught napping with a sub-optimal support instrument, as they have been with biomass CfDs.