New price formation mechanisms in Sweden: Insights from the first months of FBMC

Since the introduction of Flow-Based Market Coupling (FBMC) in the Nordic region in October 2024, price formation in Sweden has undergone a fundamental transformation. Since the transition towards a more granular, short-term mechanism, the day-ahead market remains a cornerstone for power pricing.

It is still used as a core reference for mid-term hedging products in financial power markets and as a major source of operational profit for physical market participants. Transitioning towards more complex market mechanisms also brings new transparency challenges and modelling opportunities, allowing us to move from assumptions to calculations and from expectations to forecasts.

What is Flow-Based Market Coupling?

Before FBMC, cross-border capacities were explicitly calculated by TSOs using Net Transfer Capacities (NTC). These were based on predefined safety margins and predicted system conditions.

FBMC uses a more dynamic model. It allocates cross-border capacity based on physical grid functioning laws and real-time market conditions to maximise socio-economic welfare and grid throughput.

Even though the change was prepared and communicated well in advance of go-live, it had quite a rough start. It took market participants a while to recognise the new reality. But, after 6 months of operation, we seem to be past the “denial” stage and are on the way to full “acceptance”.

The FBMC transition in the nordics

Navigating FBMC requires mastering new terms such as:

• CNEC (Critical Network Element and Contingency)

• PTDF (Power Transfer Distribution Factor)

• Shadow Price, Active Constraint, and Area Net Position

In contrast to the Net Transfer Capacity (NTC) based system, FBMC connects all bidding zones via a zonal PTDF matrix. We also need to adjust to the new idea of the “market coupling”, where we do not usually see the areas coupled by price with the neighbours, but we observe all areas being connected through the active constraints.

Flow-based market coupling introduces a new way of defining transmission constraints that influence price formation. However, many fundamental drivers still apply when analysing market trends.

Hydrology, seasonal weather changes, production outages and many other factors are still playing important role in traditional, or “intuitive” power market price formation. At the same time, some phenomena are entirely new under FBMC. The zonal PTDF matrix connects all zones, unlocking interactions that were not feasible under NTC. This has permanently changed our understanding of price formation.

Sweden’s price formation: Before and after FBMC

By analysing area price spreads from November 2024 to March 2025, we observe significant changes, especially within Swedish bidding zones.

The impact of this is evident immediately after the launch - mostly affecting Sweden’s power prices.

Key observations:

• New price patterns emerged, particularly in SE2, SE3, and SE4

• Spreads are influenced heavily by a few recurring CNECs

• These CNECs were active during ~76% of all hours in the analysis period 

Here are two things to keep in mind, before diving into the following research. One, is that we are still operating the same grid as before, with NTCs, and the overall target of the flow-based market coupling is to improve the utilisation of this grid.

And two, is the security measures in Sweden, that prevent transparency on some data (#WeNeedCnecNames) which are crucial to assess the longer-term grid behaviour patterns. These are important for further forecasting, and therefore slow down the analysis and modelling development.

Therefore, all the conclusions provided in this analysis are subjects to “authors’ opinion”, or to be more precise, "subject to opinion of the mathematical models, developed by Authors".

Area price spreads in Sweden

Let’s have a deep dive into Swedish price formation after the launch of FBMC. What we are witnessing in the development of area price spreads is entirely new.

These dynamics appeared in the flow-based domain revealing something that wasn’t possible under the previous NTC setup.

The impact of Flow-Based Market Coupling in Sweden on price formation is visible through the daily analysis of the active constraints – CNECs with non-zero shadow prices. Although CNEC names in Sweden remain encrypted, we were running our analysis for five months of operation and focused on six elements, that were frequently appearing as active constraints.

Over the course of 160 days, there were only 24% of hours when none of this CNECs was active. And, they are likely to be responsible for a major part of the volatility in the area price differences in Sweden after the start of FBMC.

Our data from November 2024 to March 2025 clearly shows that when these CNECs are inactive, area prices remain largely aligned. But once they are active, significant price spreads emerge. One notable example is the negative SE2–SE1 spread, a phenomenon not observed under the old NTC regime.

Allegedly, the spreads are usually linked to structural cuts in the Swedish power system. One at the border of SE2-SE3 and another one is an internal cut in SE3. Other important constraints include the 400kv grid lines on the SE2-SE3 transmission corridor.

However, it’s important to note that this is not what FBMC does or what started solely in November 2024. This was also included in cross-zonal NTCs before. The fact is that only a small part of these activations is caused by maintenances or emergencies.

In reality, this is how Swedish power system has been designed, and how it has been functioning for decades. The only difference is that before the change, these constraints were interpreted at the bidding zone’s borders by allocated NTCs. While now we observe these constraints geographically (on particular grid elements) where the actual bottlenecks are monitored by TSOs.

(Note: CNEC locations mentioned here are approximate and based on public data from JAO, ENTSO-E, and SVK.)

Interpreting spread patterns with the PTDF matrix

With the launch of FBMC, we have learned quite fast to read the cause and the share of the area price differences from the overwhelming FB domain data. By evaluating the PTDF matrix, we can now identify recurring spread patterns.

To analyse such patterns, we will concentrate on the interaction of Sweden’s areas with direct neighbours in the following examples, using SE3 price as a reference. Let’s explore some examples:

• If we analyse the time when only SE3 East-West cut is active out of all six major CNECs (that is about 20% of time), we can see that the price in SE3 roughly equals the average of SE2 and SE4.

• Another case is when only CNEC number 2 (ROTBOR) is active. We can still nearly match SE3 to the weighted average of SE2 and SE4, but coefficients will be different.

• And, price in SE3 can be also interconnected with Finland. This takes place on hours when both Swedish cuts are active. We once again obtain SE3 price as a weighted average of SE4 and FI.

New price formation logic

These price relationships are not coincidental. They aren’t the result of statistical fitting or regression analysis. Instead, they directly follow from the mechanics of FBMC.

Let’s consider the theoretical foundation:

Zonal price is expressed through shadow price and PTDF:

This set of linear equations can be solved, excluding  from the price differences.

Based on the FB domain on 23rd of April (hour 10) we will calculate price differences between SE2, SE3 and SE4 caused by conveniently named CNEC ‘c124715e1164446d9fe815287cd3bb3b’ as an example.

Based in (1) price differences between Swedish bidding zones are

Then 

And

Finally,

Such expressions show the interconnection between neighbouring prices without dependency from CNEC’s shadow price.

Conclusions

So, if we apply the theoretical background, let’s look at what we can deduce.

After five months of analysis, reviewing domain data, grid patterns, and time series for open and encrypted CNECs, we identified several recurring cases. These are now relatively easy to process and evaluate.

The table below concludes how the interaction between SE areas could be expressed mathematically, depending on the system conditions from the start of Nordic FBMC. 

These formulas highlight the value of FBMC data for modelling. They also demonstrate how price formation under FBMC in Sweden can now be tracked, analysed, and predicted more reliably. However, full transparency, especially regarding encrypted CNEC names, is crucial to unlocking the true potential of these models.