Analysis: Core Flow-Based Market Coupling 18 months since go-live

21 December 2023 Consultancy.eu 10 min. read
Profile
More news on

In June 2022, Europe saw a major milestone in its ambition to develop an integrated electricity market with the go-live of the Core Day-Ahead flow-based capacity calculation and market coupling. 18 months on from the introduction of the new approach, Iacopo Bertelli from Magnus Energy analyses the impact on clearing prices, price spreads at bidding-zone borders, and net positions.

Launched and supported by Nominated Electricity Market Operators (NEMOs), electricity Transmission System Operators (TSOs), and regulatory authorities, the so-called Core Flow-Based Market Coupling program (Core FB MC) has as its objective to run a Flow-Based (FB) day-ahead market coupling across 15+ European member states.

From a technical perspective, coupling electricity markets increases their resiliency and that of the transmission network, while from an economic perspective, importing and exporting electricity flows lowers the overall cost of electricity generation and increases overall socio-economic welfare.

In order to compare what has changed since the introduction of the Core FB MC program, Magnus Energy compared data across two time intervals:

  • 9 June 2022 - 31 December 2022 (i.e. the part of 2022 after the go-live of flow-based day-ahead market coupling), denoted in the plots as 2022
  • 1 January 2023 – 8 December 2023, denoted in the plots as 2023

The first data set holds 4,944 market time units (MTUs), the second 8,208. Data was drawn from ENTSO-E’s Transparency Platform and the JAO Publication Tool.

Clearing prices

The clearing prices in 2023 have been significantly lower than in 2022, due to the softening of the energy crisis.

Distribution of day-ahead clearing prices across Core bidding zones

Additionally, there have been significantly more instances of negative prices in 2023 than 2022. Negative prices are generally an indication of inflexible generation, often due to a combination of high RES generation (e.g. more wind than forecasted) and grid congestion, preventing export of electricity.

Because the clearing price levels are so different between 2022 and 2023, it is more insightful to rescale the clearing prices around the mean Core price per period.

Distribution of rescaled day-ahead clearing prices across Core bidding zones (+ indicates the mean)

The rescaled plot shows how much higher or lower prices are in one bidding zone compared to the Core mean (~105 €/MWh, 0 in the plot above). Generally, scaled clearing prices are aligned across Core, with the exception of Poland (more on this later).

In 2023, clearing prices have been more homogeneous, as shown by the medians being closer to the Core mean (‘0’ in the plot above), and by the lower variance (i.e. shorter boxes), with Poland being again the only exception.

More focus on the mean clearing prices (rather than the distributions) provides further insights.

Mean clearing prices across Core bidding zones

Excluding PL, the mean prices in 2023 stayed within roughly 5% of the Core mean. On average, AT, CZ, HR, RO stayed within 2% of the Core price. In 2023 the Austrian price is the closest to the Core mean. However, the best predictor for Core, for both 2022 and 2023, seemed to be CZ, overall.

Belgium, the Netherlands, Germany, France (and the Czech Republic to a lower extent) show lower relative clearing prices than other bidding zones.

In 2022, Poland had (by far) the lowest prices on average (35% lower than the Core mean of 293 €/MWh), due to the allocation constraints limiting its (SDAC) (export) net position.

Note: SDAC is the Single Day-Ahead Coupling, the pan-European electricity market for the day-ahead timeframe. While the allocation constraint shielded the Polish market from high prices in 2022, in 2023 the allocation constraint condemned Poland to the highest prices (10% higher than 105 €/MWh), due to constrained import possibilities.

France, on the other hand, shows a reversed picture when it comes to clearing prices. While in 2022 median clearing prices were the highest in Core, in 2023 they are among the lowest (following DE and NL).

Notably, France switched from being a net importer in 2022 to a net exporter. Several nuclear power plants were not operational in 2022 due to maintenance, while in 2023 they have been running. On average, nuclear generation was ~28 GW in the second half of 2022, while it was ~36 GW in 2023.

Further readingFrench energy giant brings in consultancies for nuclear strategy.

Clearing price spreads across Core

The below plot shows duration curves of the maximum price spread across Core CCR per MTU and per period (2022 versus 2023). MTUs are normalised and shown as “frequency”.

Duration curve with maximum clearing price spread across Core

In 2023, the maximum price spreads at bidding-zone borders were significantly lower than in 2022. Excluding Poland (again because of the allocation constraint), the 2023 curve shows that the maximum spreads per MTU were above 100 €/MWh only about 10% of the time, below 50 €/MWh for 80% of the time, and below 25 €/MWh for about 60% of the time.

For around 1/3 of the time (similarly to 2022), spreads were 0 across all of Core (flat horizontal line at the bottom right corner), indicating fully coupled clearing prices (again excluding Poland).

On which bidding zone borders did price spreads materialise? An overview of the spreads per bidding zone border:

Price spreads at Core bidding zone borders

In 2023, the spreads became significantly lower. Excluding the bidding zone borders of Poland sheds more light:

Price spreads at Core bidding zone borders excluding PL

A few borders showed a reversing of the prevailing spread direction: HR-HU, HR-SI, HU-RO and HU-SI are the most evident ones, alongside the bidding zone borders of FR that went from large spreads (with FR having a higher clearing price) to spreads with mean close to zero.

We can also look at the clearing price spreads after grouping spread ranges (zero, low, medium, high). Inspiration is taken form the ACER Market Monitoring Report.

Price spreads in 2022 in Core, grouped by spread ranges

As in the duration curve with the highest spread above, we can see that (excluding PL bidding zone borders) all bidding zone borders have zero spread around 30% of the time (similarly to 2022) and spread below 1 €/MWh around 60% of the time (compared to ~40% in 2022).

HU-RO, HU-SK and HR-SI are the borders where spreads are lowest (suggesting that they are the least congested): on these borders, in 2023, spreads are below 1 €/MWh around 2/3 of the time, and are “very” high (red colour) only about 5% of the time for HU-SK, HR-SI and ~15% for HU-RO.

Price spreads in 2023 in Core, grouped by spread ranges

On the other hand, the borders where spreads are more often very high are (excluding the Polish borders) DE-FR, BE-DE, BE-NL and AT-DE

Net positions

Analysis of net positions provides an outline imports and exports (respectively negative and positive values).

Distribution of net positions across Core bidding zones

The plot shows distributions of the Core net positions. Note: These are not SDAC net positions of Core bidding zones (i.e. any exchanges on non-Core SDAC bidding zone borders are not included). Taking France as an example, the exchanges with Spain (in SDAC) or UK and Switzerland (not in SDAC) are not taken into account.

When comparing the two periods of data, SK exported more, while SI, AT and HU imported less, and France went from a net importer to net exporter (again because of the aforementioned increase in nuclear generation in 2023). On the other hand, CZ, RO, PL exported less, while BE and DE went from exporters to importers. The situation in NL and HR was virtually unchanged.

Allocation constraints

To provide some context: PL and BE use(d) so-called allocation constraints. For both bidding zones, these allocation constraints limit the SDAC net position (i.e. max imports/exports) of the respective bidding zone (note: not the Core net position is limited, and “transits” across the bidding zones are not restricted by the allocation constraints).

Poland uses allocation constraints for both import and export while Belgium uses an allocation constraint for import only.

Limiting behaviour of allocation constraints in PL

The figure shows that the PL allocation constraint is limiting a good amount of time: approximately 78% of time in 2022 and 47% in 2023. In 2022, the export constraint was limiting most of the time (~72%), while in 2023 the import is most limiting (~40%). These constraints result in the PL price being the lowest in 2022, and highest in 2023, which we observed when looking at the clearing prices earlier.

The BE import constraint has not been limiting at all. For this reason, BE (and NL, which had constraints on the Core net position) stopped using allocation constraints as of 18 November 2023 and 15 December 2023, respectively.

The allocation constraint limiting the import/export from PL from/to its neighbours effectively ”locks-in” power prices in Poland, preventing the market from reaching a global optimum.

Key take aways

Overall, in 2023 clearing prices levels have been lower and price convergence has greatly improved compared to 2022. The exception is Poland, caused by its allocation constraint that limit imports/exports. While in 2022 the allocation constraint caused low prices in Poland, in 2023 it had the opposite effect.

Negative prices have been far more common in 2023 than 2022.

France’s nuclear power plants have been generating in 2023 on average ~8 GW more than in the second half of 2022, causing shifts in the import/export balance of the FR bidding zone and of nearby countries (DE and BE above all).

Price spreads have been significantly lower in 2023 than in 2022, with HR-SI and HU-SK re-confirmed to be the least congested bidding zone borders.