ELEXON insights: How tagging changes the System Price
Emma Tribe’s SPAR insight article takes a detailed look at how tagging affects the System Price calculation. By removing volumes from the top or bottom of the pricing stack the System Price decreases or increases.
Published: May 2018
Tagging between May 2016 and May 2018
95.7% of Balancing Volume was removed from the System Price Calculation by Tagging between May 2016 and May 2018. There are four tagging operations to remove balancing volume from the System Price calculation:
- De-Minimis Acceptance Threshold (DMAT) tagging
- Arbitrage Tagging
- Net Imbalance Volume (NIV) Tagging
- Price Average Reference (PAR) Tagging
After all of these tagging operations were taken in May 2018, 4.3% of Energy Balancing Volume remained. In each Settlement Period a volume weighted average of the remaining actions are taken to calculate the Imbalance Price.
While the tagged 95.7% of Balancing Volumes was not used to directly set the System Price, the removed volume did influence the System Price. By removing these volumes from the top or bottom of the pricing stack the final weighted average is made higher or lower.
Graph 1: Energy Balancing Volumes by Tagging operation
Graph 1 shows that in every month NIV tagging removes the majority of balancing volume. Over the two years NIV tagging removed 72.7% of Balancing Volume.
82.8% of balancing volume was removed by NIV tagging in October 2017. This is the highest percentage in a month.
There is a monthly standard deviation of 6.3% in the percentage removed by NIV tagging; a greater percentage is removed when a greater volume of actions are taken.
During the two years, PAR tagging removed 22.3% of the total Balancing Volume. PAR tagging is the final tagging operation, carried out after the pricing stack has already been NIV, Arbitrage and DMAT tagged.
PAR tagging removed 83.8% of the remaining volume after all other tagging operations had already removed volume.
Arbitrage removed 0.6% of Balancing Volume over the two years. The monthly percentage removed by Arbitrage tagging has increased from 0.3% in May 2016 to 3.6% in May 2018.
DMAT tagging removed 0.1% of Balancing Volume. The volume of balancing actions removed by DMAT tagging is much smaller than by NIV and PAR. As DMAT tagging only removes actions with a volume less than 1MWh this is expected.
Tagging in May 2018
Graph 2: Percentage of Settlement Periods where each tagging occurs
Graph 2 shows the percentage of Settlement Periods in May 2018 that each of the tagging operations occur in.
No tagging operation occurs in every Settlement Period. Volumes are NIV tagged in 94.2% of Settlement Periods; NIV tagging is not used in Settlement Periods where balancing actions are only taken in one direction.
Arbitrage tagging occurred in 27.1% of Settlement Periods. In the majority of Settlement Periods the price of buy actions have been greater than the price of sell actions.
The purpose of these tagging operations is to make the weighted average price:
- more marginal (PAR Tagging)
- representative of cost to resolve the net energy imbalance (NIV Tagging and Arbitrage Tagging)
- remove price distortions caused by errors (DMAT tagging)
Graph 3: Average Price of buy actions against short System Price
Graph 3 compares the average weighted price of buy or sell actions against long or short System Prices in May.
In 63% of Settlement Periods the weighted average price of buy actions is greater than short System Prices.
When the System Price is less than £80/MWh, the majority (79.9%) of average buy action prices are greater than the System Price. However, when the System Price is greater than £80/MWh, the majority (87.1%) of average buy action prices are less than the System Price.
The weighted average price of sell actions is less than the System Price in 79.9% of long Settlement Periods. This means that the System Price is less expensive than the weighted average price in these Settlement Periods.
Graph 4: Average Price of tagged buy actions against short System Price
Graph 4 shows the weighted average price of tagged buy actions compared to the System Price, when the market is short, for NIV, PAR and Arbitrage tagging.
DMAT tagging was not included as DMAT tagged actions can come from anywhere in the stack.
NIV tagging removes the most expensive volumes from the stack; hence the weighted average price of NIV tagged actions is greater than the System Price in 99.1% of Settlement Periods.
The System Price can be greater than the weighted average price of NIV tagged actions when a Buy Price Price Adjuster (BPA) has been applied.
The weighted average price of NIV tagged actions is greater than the System Price by an average of £12.62/MWh.
PAR tagging and Arbitrage tagging remove the least expensive actions. In all Settlement Periods the weighted average price of Arbitrage tagged actions are less or equal to the System Price.
The average Arbitrage tagged action price is £36.77/MWh.
The weighted average price of PAR tagged actions are less than or equal to the System Price in 99.6% of Settlement Periods. The weighted average price of PAR tagged actions can be greater than the System Price when second stage flagged actions have been repriced by the Replacement Price.
Graph 5: Average Price of tagged sell actions against long System Price
Graph 5 shows the weighted average price of tagged sell actions compared to the System Price, when the market is long, for NIV, PAR and Arbitrage tagging. A clear difference between this graph and graph 4 is that the distribution of long prices is much narrower than for short prices.
NIV tagged weighted averages are less than or equal to the System Price in all Settlement Periods, as the Sell Price Price Adjuster (SPA) has not been applied.
PAR tagged weighted averages are less than the System Price in 24.7% of Settlement Periods, and Arbitrage tagged weighted averages are less than the System Price in 6.3% of Settlement Periods. This occurs when the Replacement Price has been used to reprice Second Stage Flagged actions.
Tagging operations increase or decrease the final weighted average price by changing what volumes remain in the calculation. In contrast, flagging can change the final weighted average by repricing actions.
The Replacement Price was used in 5% of Short Settlement Periods and 37% of Long Settlement Periods. Hence, there are more Settlement Periods where the weighted averages of PAR and Arbitrage tagged actions are more expensive than the System Price.