In a recent statistical study of the German year-ahead forward prices, KYOS found that the speed of cointegration between power and fuel prices increased. This is contrary to the common belief that power prices are less and less dependent on fuel prices due to the renewables penetration. The figure shows that since early 2015 the daily speed of mean-reversion has gone up
from about 5% to 20%. This means that power prices stay closer to a mix of coal, gas and emission prices.
To understand how the price of energy is made up, and why power and fuel prices are cointegrated, it is important to realize that most power is (still) produced from fossil fuels. This mechanism is most commonly visualized with a merit order, where demand is fulfilled first of all by non-dispatchable sources, primarily solar, wind and most of nuclear. The production of these sources is often not enough to meet requirements so more electricity has to be produced – at gradually higher marginal production costs. In many countries this second part is built up of efficient coal plants, less efficient coal plants or more efficient gas plants, and finally less efficient gas and oil plants (peaking stations). The price of energy is therefore heavily influenced by fuel-fired producers.
KYOS uses cointegration in all of its models for generating Monte Carlo price simulations. It ensures that commodity prices and commodity spread scenarios are realistic. In particular, it avoids that spreads become extremely wide or narrow, if that is not in line with market fundamentals. This is especially relevant for conventional power plants (KyPlant), whose margins are derived from the spread between power, fuel and emission prices. The use of cointegration in statistical models is very relevant for valuation and risk management, especially with a horizon of up to about two years. It captures how market prices actually move, in terms of return volatilities, return correlations and price cointegration. With the observed levels of cointegration, spark and dark spreads are considerably less volatile. If this behavior in power prices is ignored, power plants and spark/dark spread options will be overvalued.
Over longer horizons, the statistical parameter estimates must be treated more carefully; a fundamental analysis may be more appropriate for long-term investment decisions. Fundamental analysis, however, is quite dependent on the assumptions about the future market structure and fuel prices. Because we believe in the relative strengths of both approaches, statistical and fundamental, KYOS provides tools for both: a primarily statistical model with cointegration (KySim), especially for the medium term trading oriented applications, and a primarily fundamental model (KyPowerFundamentals, KyPF), especially for somewhat longer-term applications. Both models can also be used simultaneously: the statistical model generates Monte Carlo simulations of fuel prices, which the fundamental model uses to generate power price simulations.
In the future, will the growth of renewable energy affect the price even further? Power prices have already declined sharply and the profitability of conventional generation is very poor. Will this trend of lower power prices
and lower margins continue? And if so, what mechanisms will drive the power prices, and how important will be the role of coal and gas to set the prices?
This situation as outlined for the German power market is applicable to all European countries. We believe both statistical and fundamental analysis help to provide answers. So far, even though absolute power prices have decreased, our statistical analysis shows that power prices are closely connected to fuel prices, even more so than several years ago.
Please click on the link below to read the full article (pdf):
Fuel price is driver European power price
Nieuwe Gracht 49
2011 ND Haarlem
Tel: +31 (0)23 551 02 21
Contact us via the form below. We will reply within 48 hours.