Backtest: how accurate were Montel’s European power price scenarios?

Scenarios are not forecasts 

Their purpose is to outline coherent, plausible development paths and define a range within which future prices may evolve. None of the scenarios is intended to predict the exact price path. In power markets, that would be unrealistic given the influence of short-term factors such as weather, commodity prices and electricity demand. 

Montel calculates its power price scenarios using its in-house fundamental power market model, Power2Sim. Different assumptions are applied to the key market drivers for each of the three scenarios: Central, Tensions (High) and Go Hydrogen (Low). More information about our scenarios can be found in our EU Energy Outlook. 

The challenge of assessing a period of three years or less 

In our scenarios and their quarterly updates, commodity prices play an important role and have a strong influence on power price development, particularly in the short term. This is consistent with observable market dynamics, where short-term price fluctuations are partly strongly driven by commodity prices.  

At the same time, our power price modelling is not based solely on commodity prices. Even in the short term, additional fundamental factors are taken into account, including the development of installed capacity of renewable energy and thermal power plants; the expansion of battery energy storage systems (BESS); electricity demand; and other drivers which are explained in detail in the article linked above. 

With a focus on how commodity prices are modelled, Montel uses current future market prices for commodities for the first three years (the period that is liquidly traded on the futures market). Thereafter, the price paths transition to the respective long-term scenario trajectories, which for coal, oil and CO₂ certificates, are based on scenarios from the International Energy Agency’s World Energy Outlook. For long-term gas price developments, a separate modelling approach is applied, initially based on LNG price developments and, from 2050 onwards, on hydrogen prices. 

These commodity futures reflect volatile expectations for future day-ahead prices and are therefore only ever a snapshot of market sentiment. Montel uses different methodological approaches for the three power price scenarios. The Central scenario is based on the average future price over the two weeks leading up to a defined cut-off date, Tensions on the high of that period, and Go Hydrogen on the low. 

As a result, the spot power prices modelled on the basis of commodity prices do not have to match the power futures traded in the market at the same time. That is particularly true when the market is pricing in different short-term assumptions for other drivers (such as the generation mix, demand or weather) than those embedded in the scenario path. At the same time, because the High and Low scenarios use market extremes, actual power futures will usually still fall within or close to the corridor opened up by the three scenarios. 

Another point matters: the scenario path for the first three years cannot fully capture spot-market volatility in extreme conditions. That is why the explanatory power of a short-term backtest is inherently limited. A longer time horizon is more meaningful. Over time, historical power prices should mostly fall within the modelled range of the three scenarios. 

The chart below compares historical future and spot prices for Germany (both annual averages) with the European power price scenarios published by Montel Energy Brainpool in each annual Q4 since 2012. All data points represent annual averages of spot or future prices. The connecting lines are purely visual and do not show intra-year movements. 

Key takeaways from the backtest 

The historical picture is fairly clear. From 2013 to 2016, both historical future prices and the respective Central scenarios were consistently above realised spot prices. Since 2016, Montel Energy Brainpool has published three coherent scenarios to reflect the range of possible market developments. Over that period, historical annual average spot prices mostly remained within the modelled corridor. Even the Covid-related dip did not fall below the Low scenarios. 

The clear exception is the extreme price spike in 2022 and 2023 following Russia’s invasion of Ukraine. This is where the limits of scenario-based models become visible in the face of rare political shocks. At the same time, comparing the dashed black line of future prices with the solid black line of average spot prices for 2022 to 2025 highlights the modelling challenge: future prices reflect expectations for future spot prices based on the information available at the time, so market effects tend to appear with a lag. 

That said, the sharp rise in future prices from 2021 to 2022 also shows that markets had already started to price in Russia-related risks by the fourth quarter of 2021, at least in part. As a result, the modelled power prices across all three scenarios also rose sharply from 2021 to 2022. 

The chart also shows that the scenarios published for 2023 and 2024 remained above the later realised average spot prices because they were based on the elevated future curve at the time. After that, market expectations gradually adjusted lower as spot prices materialised. Actual spot prices moved back close to the 2022 scenario corridor from 2023 onwards. By 2024, they were only slightly above the High and Central scenarios published between 2012 and 2021. 

Our conclusion 

Outside extreme events and short-term volatility, power price scenarios provide a robust range of possible price paths. Which scenario is most suitable for valuation and planning depends on risk appetite and market view. Even with their limitations, they remain the best available basis for long-term investment decisions in power markets.