Market impact in short-term power trading: The Dutch case and lessons for Belgium

Main topics:

• Market impact mechanics: How collective trader behavior in short-term power markets can transform profitable strategies into systematic losses
• Dutch case study: Analysis of regulation state 2 events and TenneT’s responses to restore grid stability
• Belgian market lessons: What Belgian traders can do to avoid similar pitfalls

On some days, the Dutch power system resembles a pinball machine. High wind and solar output trigger mass curtailment decisions, followed by rapid reversals as traders react to the changing grid state. Within minutes, the system bounces between opposing imbalance states, creating events that penalize all market participants regardless of their individual logic.

These moments exemplify a fundamental reality: when hundreds of short-term power traders pursue the same profitable strategy, that strategy eventually stops paying off. Yet market impact – the collective influence of trading decisions on market outcomes – remains a blind spot for many participants.

In this article, we explore how market impact has affected the Dutch power market, analyze TSO TenneT’s technical response to restore grid stability, and draw lessons for Belgian traders looking to avoid similar pitfalls.

The hidden costs of market impact

Trading decisions don’t happen in isolation. Every order influences the market it operates within, creating ripple effects that extend far beyond immediate price movements.

Consider the mechanics at play. A sell order in the day-ahead market shifts the supply curve downward, slightly reducing the day-ahead price. A buy-back in the balancing market activates more automatic Frequency Restoration Reserve (aFRR) volume, pushing prices higher. These direct effects are well understood.

But the effects are often more nuanced. Take the common approach of selling in the day-ahead market and buying back in the balancing market, which is profitable when day-ahead prices exceed imbalance prices. Low imbalance prices often indicate that the TSO is reducing generation. In this case, your buy-back decision actually reduces the need for additional downward aFRR activations, bringing the total activation volume closer to zero. This pushes the imbalance price upward, but toward a more “neutral” level, rather than exacerbating extreme lows.

The strategy works – until everyone adopts it. When market participants rely on the same price forecasts and public signals, the predictive value of those signals diminishes, potentially turning expected profits into systematic losses. This is market impact in action, and it has played out dramatically in the Dutch power market.

Why the Dutch power market is a revealing case

The Dutch power market presents a unique environment for observing market impact due to four key characteristics:

Passive imbalance tolerance

Market participants are allowed to deviate from their scheduled positions, enabling them to go into imbalance. This flexibility is particularly useful during times of solar oversupply, offering market-based tools to support grid balancing without direct TSO intervention.

Data transparency

The Dutch TSO, TenneT, publishes minute-by-minute operational data. This openness is valuable for building forecasts but also encourages reactive trading behavior.

Marginal pricing

The imbalance price for each Imbalance Settlement Period (ISP) is set by the most extreme aFRR activation price observed across all 4-second optimization cycles within the 15-minute window. This makes Dutch imbalance prices significantly more volatile and sensitive to the effects of market impact. By contrast, countries like Germany and Belgium calculate imbalance prices using an average across all 4-second intervals.

Dual pricing or regulation state 2

When too many assets join in the passive imbalance – such as renewables curtailing during negative price events – it can lead to overcorrection. TenneT must then rebalance in the opposite direction within the same program time unit (PTU). In this situation, imbalance prices split to penalize both positive and negative positions. This is called regulation state 2, or dual pricing, and is abbreviated as RS2.

Understanding regulation state 2 events

Regulation state 2 events require two specific conditions:

• TenneT must activate both upward and downward regulating power within the same PTU.
• The system-balance delta – the cumulative minute-by-minute total of requested balancing power – must swing in both directions during the PTU. If the delta only rises or only falls, even with minor counter-activation, the period remains classified as RS +1 or –1.

Throughout 2024, we’ve seen a surge in instances of dual pricing. For traders on the Dutch market, the impact of dual pricing has been notable not only in terms of frequency but also in terms of price spreads.

A textbook example occurred on Friday, March 8, 2024, when high wind and solar output led to curtailment, quick reversals, and rapid switching between positive and negative imbalances. The result was a classic RS2 event that resembled a pinball machine, with the system bouncing rapidly between opposing states.

If market behavior leads to more RS2 events, but the system remains controllable, TenneT isn’t likely to intervene. But if fast fluctuations continue to outpace the ramping capability of reserves, threatening grid stability, that’s when action is triggered.

TenneT’s initial response: The price signal delay

To reduce aggressive passive balancing, TenneT adjusted its publication process in November 2024, delaying its minute imbalance price signal from two to five minutes. Market experts were skeptical from the start, with Jean-Paul Harreman at Montel calling it ‘a blow to transparency’.

As mentioned, passive balancers rely on up-to-date imbalance signals to make near-instant decisions: when the grid is long, they curtail; when it’s short, they decurtail. If that signal is delayed, they respond to outdated conditions, leading to overcorrection: curtailing when the grid is no longer long, or decurtailing when it’s already short.

Subsequently, the delay for balance delta publication was reduced back to two minutes, effective July 1st.

Structural reforms underway

Likely recognizing that quick fixes were insufficient, TenneT outlined comprehensive technical reforms at its May 2025 informative session. The changes target both software and hardware limitations.

At the heart of the response lies the Load-Frequency Controller (LFC), which uses a PID (Proportional-Integral-Derivative) algorithm to respond to imbalance changes by adjusting the setpoint of activated aFRR reserves. Historically, the controller emphasized the integral component – useful for slower deviations, but too sluggish for today’s fast-moving fluctuations driven by passive balancing.

TenneT is now rebalancing the controller to prioritize faster-reacting components, especially the proportional term. This allows the system to respond more aggressively to sudden swings in Area Control Error (ACE), improving frequency stability.

However, the control logic alone isn’t enough. There’s a hardware-speed mismatch: passive traders can adjust positions almost instantly, while aFRR reserves ramp up at just 20% per minute, meaning it takes up to five minutes to fully activate.

To address this, TenneT is:

• Allowing faster ramping from Balancing Service Providers (BSPs) up to 100% per minute on a voluntary basis
• Developing a new balancing product with minimum ramp rates up to 1500% per minute, enabling full activation in just 4 seconds
• Continuing PID tuning, with product changes to follow

When asked whether passive balancing could be restricted if these measures fail, TenneT’s response was: “We are not excluding any options.”

Our take: Four long-term solutions

At Dexter Energy, we support TenneT’s current efforts to stabilize the Dutch power system. We also believe additional long-term measures can address structural challenges. Here are our recommendations:

Improve data transparency and latency

Enhancing both the quality and speed of operational data would be a good step toward mitigating market impact.

First, publishing more detailed signals, such as the Area Control Error (ACE), would help passive traders adjust more precisely within each PTU. For instance, they could stop curtailing earlier if they detect that oscillating ACE and RS2 seem imminent. While this wouldn’t eliminate oscillations entirely, it would help reduce their amplitude. Open Loop (OL) ACE data is already published on the ENTSOE platform, but with a 10-minute delay, which is too slow to support real-time trading decisions.

Second, minimizing signal delay is just as critical. Instead of using a 2-minute delay, real-time or near-real-time signals would go further in preventing overcorrection.

Reform the ex-post market

Currently, Balancing Responsible Parties (BRPs) can offset much of the cost incurred during regulation state 2 by offloading their imbalance volumes in the ex-post market. This may undermine the intended design of RS2 and reduce the incentive to prevent large fluctuations. Therefore, reforming or removing this market could make RS2 more effective and discourage overaggressive passive balancing.

Introduce a volume-based penalty

The French system offers a promising model by applying a k-factor – a type of transaction cost – to each volume taken on imbalance. This factor increases with the cumulative imbalance volumes taken by all BRPs over recent months, effectively pushing less capable speculators out of the market. A similar mechanism in the Netherlands could help control excessive imbalance behavior.

Shorten the PTU

Moving from 15-minute to 5-minute PTUs would better align trading behavior with real-time system conditions. It would narrow the window for exploiting aggregate intra-PTU fluctuations and incentivize traders to respond more accurately to the grid’s actual state.

Strategic lessons for Belgian traders

The Belgian power market shares many characteristics with its Dutch counterpart: passive imbalance is allowed, and local TSO Elia publishes high-resolution near-real-time operational data. Furthermore, imbalance volumes and spreads have also increased year-on-year.

At the same time, there are notable differences:

• Belgium uses single-price imbalance settlement, which reduces volatility.
• Imbalance prices are calculated as the average of 4-second marginal prices, smoothing out price spikes that, in the Netherlands, are sometimes set by brief 2,000 MW activations.

Since both countries joined the PICASSO platform in late 2024 (the Netherlands in October, Belgium in November), imbalance prices have become more moderate on average. Still, Belgium’s relative price stability is an opportunity, not a guarantee. The Belgian grid is also exposed to systems and control theory, and oversteering can lead to the market flipping from negative to positive prices within or across PTUs.

Thus, Belgian traders can take several lessons from the Dutch experience:

• Avoid aggressive steering: Monitor for synchronized trading behavior that could lead to the market flipping.
• Diversify signals: Don’t rely solely on public TSO data. As seen in the Netherlands, widespread reliance on the same public signal leads to overcrowded trades,
• which in turn can quickly erode profitability.
• Expect regulatory shifts: If needed, Elia may follow TenneT’s lead by implementing publication delays, ramp rate requirements, or new pricing mechanisms.
• Prepare for cross-border effects: With both countries now on the PICASSO platform, instability in one market can increasingly affect the other.

Factoring market impact into trading strategies

The Dutch market’s experience with regulation state 2 events demonstrates how market impact can transform profitable trading strategies into painful losses. For short-term power traders, success increasingly depends on understanding not only market fundamentals but also how collective market behavior influences them.

A simple question driving this mindset is, “What happens when everyone else does the same thing?”

At Dexter Energy, we factor market impact into our strategies for each bidding zone, ensuring that our customers maximize the value they derive from our trading signals. This explains why we choose to work with a limited number of companies and a maximum trading mandate for our different products in each country.

Would you like to know how this applies to your portfolio? Let’s talk.