06 Apr 2022
Alex Done

Dynamic Containment - price caps and demand elasticity

On Monday 4 April, we saw an all-time high price of £60/MW/h in Dynamic Containment (DC). The following day, that figure rose again, this time to £70/MW/h. At the time of writing, DC prices have hit £103/MW/h for a single EFA block. Figure 1 (below) shows the DC high and low clearing prices since the start of April: the increase is evident.

Figure 1 - Dynamic Containment clearing prices from 28 March 2022 - 6 April 2022.

In the last week, we have seen the total market value in DC go from £830k (24 - 30 March) to £2.4m (31 March - 6 April). This three-fold increase is due to a change in the way that National Grid Electricity System Operator (ESO) is running the DC auction. So, why are we now seeing these high prices?

In this article, we explore:

  • DC price caps and how they can distort the market.
  • What ‘elastic demand’ means, and the changes to the DC auction structure.
  • The reasons for these changes, and why this has impacted prices so much.
  • What this means for battery energy storage systems.

In the video below, you can watch us discuss why we’re seeing these prices:

Alex and Robyn discuss recent DC prices.

Price caps

By 'price cap', we mean the maximum price the ESO is willing to pay to procure DC. The price cap reflects the cost of alternative actions - the cost to the ESO of securing the system by other means, in the event it could not procure DC. Figure 2 (below) shows how the price cap in DC has evolved since the service launched in October 2021.

Figure 2 - Historical values for NG ESO price caps from October 2020 - March 2022.
  • At launch, the market established early on that the ESO’s price cap sat at £17/MW/h - though we didn’t have any official confirmation of this.
  • With the move to EFA block procurement, the ESO began to publish its buy order curve in September 2021, detailing the precise volume it sought and the maximum price it was willing to pay, with the price cap holding steady at £17/MW/h.
  • Following downwards revisions to requirements in November 2021, the ESO also revised its price cap from a flat £17/MW/h across the day, to vary per EFA block, as high as £48/MW/h across EFA blocks 4 and 5.
  • At the same time, the ESO launched the high-frequency DC service (as opposed to low frequency). Dynamic Containment High (DCH) had a different price cap, again which varied across the day.

What is a ‘buy order’?

A 'buy order' simply refers to the amount of DC the ESO requires, and the price it is willing to pay. Essentially this data can be used to draw the DC demand curve for a given service (high or low) for a given EFA block, as shown in Figure 3 (below).

Figure 3 - Example Dynamic Containment Low buy curve, 29 March EFA block 5.
  • In this instance, we have a simple buy order. The ESO will pay up to £48/MW/h and will buy up to 392 MW. So, they will spend up to £75,264 across a 4-hour EFA block.
  • The demand curve pictured above is inelastic - the price the ESO is willing to pay does not change as a function of how much it procures. In this case, the first MW of DC is worth as much to the ESO as the 392nd MW.

Exploiting the price cap

There are two components of buy curves like that in Figure 3 (above) that make them particularly prone to exploitation:

  • Static - Since the DC service started, price caps have remained largely unchanged day-to-day (as can be seen in Figure 1). This has given market participants a lot of information on how to avoid rejection due to exceeding the cost of alternative actions.
  • Inelastic - With no variation in the price cap as a function of procurement volume, market participants can use bidding techniques that minimise the risk of rejection and maximise the likelihood of setting a high clearing price. This is known as ‘hockey-stick’ bidding.

DC is not the first ancillary service auction where we’ve seen participants gaming the buy order. In the weekly FFR auctions, a similar thing happened - we’d recommend checking out the ESP Consulting report on it. So, what have the ESO done to address some of the issues with the static, inelastic demand curves?

Elastic demand curves

On April 1st 2022, the ESO changed the shape of its buy order curve, an example of which is pictured in Figure 4 (below).

Figure 4 - New, elastic buy curve (04 April 2022, EFA 5) compared to an indicative, simple buy curve.

There are a couple of differences with this buy curve as compared to the one shown in Figure 3:

  • The price the ESO is willing to pay for DC changes as a function of the volume it procures - in the language of microeconomics, we now have an elastic demand curve.
  • In the example above, this means the ESO are willing to buy up to 259 MW at a maximum price of £105/MW/h; between 259-396MW at a price up to £96/MW/h; and so on.

Comparing the simple, pre-April demand curve to the new, elastic demand curves we see a crucial implication of this new style of buy order from the ESO:

DC providers can now be paid a lot more for delivering DC - in this example, up to £105/MW/h! This is more than double the previous cap at £48/MW/h. However, such massive prices are only seen at times when the DC market is short on capacity.

Why have prices hit new highs?

To understand how these new buy orders have led to DC prices reaching £103/MW/h, let’s take a look at the auction for delivery over EFA block 5 on 06 April 2022.

Figure 5 - Supply and demand curves from DCL auction for delivery on 06 April 2022, EFA block 5.
  • According to the ESO’s buy order, it was willing to buy up to 606 MW at a maximum price of £110/MW/h.
  • 517 MW of bids were submitted for the above auction with only 1 MW rejected on grounds of exceeding the ESO’s buy curve price.
  • The highest bid accepted was submitted at a price of £103/MW/h, setting the clearing price at a new all-time high for the DCL service.

DC Market implications

What do these new buy orders mean for the DC market as a whole? One way of exploring this is to examine procurement volume against potential ESO spend in two scenarios: with a simple price and volume cap (yellow); and using the new elastic demand curves (blue). In Figure 6 (below), we plot DC procured volume against the maximum total spend for a sample EFA block comparing the two scenarios described above.

Figure 6 - Incremental cost of procurement in DCL using elastic and (indicative) inelastic demand curves.

With a ‘simple’ buy curve with just a single price and volume cap, the amount the ESO spends on DC scales linearly with the volume it procures. With an elastic demand curve, the total spend is much more variable - and it is hugely dependent on how much volume is procured.

Let’s take the maximum value from both of these curves, representing the maximum possible ESO spend (or the maximum provider revenue) that could be extracted from the DCL market over a given EFA block. Repeating this idea for all EFA blocks in both DCL and DCH, we can examine how the potential DC value has changed over the transitionary period from simple buy order curves to elastic demand curves. Figure 7 (below) shows the maximum amount the ESO were willing to spend on DC (in both the high and low services) from 26 March to 06 April in addition to their realised spend.

Figure 7 - Maximum and realised daily DC revenues from 26 March 2022 - 06 April 2022.
  • Since the move to elastic demand curves, the maximum possible daily DC spend has increased from £460k (avg. max daily spend of the week preceding 01 April) to £0.9m (avg. max daily spend after April 1st).
  • This represents a ~2x increase in the ‘money on the table’ for DC providers, but will only be realised when the market is undersaturated.
  • 06 April shows how this potential value can be realised in periods of undersaturation with realised service costs for the day exceeding £1m!

Key takeaways for battery energy storage systems

Batteries operating in Dynamic Containment could see significant changes in their bidding patterns and revenues as a result of these changes.

The inputs to the algorithm that determine the auction’s clearing prices have changed - so who secures what contracts, and at what price, will also change. There is a new, increased, complexity in the DC market. Participants and optimisers will have to understand this and potentially adjust their business practices to maximise revenues.

Some of the bidding strategies seen historically (for example, hockey stick bids) will prove less successful as market sophistication increases and strategies will need to adjust. With a demand curve that changes daily, participants may have trouble predicting where the market will clear and as such, we’d expect market competition and efficiency to increase. In normal circumstances, this would represent a loss of value for participants, however, this also comes at a time when the new demand curves show a lot more value in the market, and for the time being, the latter driver is the dominant force.

Cross-market optimisation between DC and wholesale markets (in addition to the new DR and DM markets) becomes more complex too. Should all sites exit DC in favour of exploiting arbitrage opportunities in the wholesale market, saturation in DC could fall dramatically, thereby increasing the scope for higher prices in DC. We may see some assets sticking around in DC to capitalise on this, but if the whole market has the same idea it could have the opposite effect. We could get something akin to game theory in cross-market battery optimisation.

Whichever way you cut it, strategies will begin to adapt over the coming weeks as the market adjusts to these changes. As ever, Modo will be following this closely as things develop - one thing is for sure, April’s Leaderboard is going to be interesting!