Intraday markets for electricity allow for trading of energy until shortly before the period of delivery. This offers market participants a possibility to reduce their expected imbalances and to offer own unused flexibility. Because this form of distributed balancing before the period of delivery can be profitable for market participants as well as beneficial for system operations, intraday trading is expected to gain more importance in future, especially with increasing shares of variable renewable energy sources in the generation mix. So far, intraday markets are still a research field with many open questions. This post considers some recent research done on intraday trades based upon ELBAS, one of the European intraday markets. The findings include that trading activity differs significantly between price zones, that most trades occur in the last hours before gate closure and that market participants have to handle substantial price variations during the trading period.
In most European electricity markets, dispatch of power plants is not any longer decided and optimised centrally; instead, all market participants schedule their own power plants in a profit maximising manner. As each electric power system requires a continuous balance between generation and load, instruments are needed to guarantee that the schedules of all market participant converge in a both technically feasible and economically efficient solution for the whole system. Electricity markets are used as such an instrument. Usually, they are designed as a sequence of different trading opportunities; one being intraday trading which allows for trading of energy between closure of day-ahead markets and period of delivery.
There are different reasons why intraday trading can be considered profitable by market participants: first of all, it is a possibility to reduce imbalance costs to which market participants in several European electricity markets are exposed when supplying more or less energy than they planned. These imbalance costs can comprise an important incentive for all market participants to compute appropriately accurate production and consumption forecasts as well as to schedule and trade based on these forecasts. Reducing imbalance volumes is also a way to hedge against (uncertain) imbalance prices that might deviate significantly from day-ahead prices. In a survey among Swedish balance responsible parties, reduction of imbalance costs can currently be seen as the main motivation for intraday trading.
A second motivation is the possibility to optimise own production/consumption schedules, e.g. by buying energy to reduce generation in an own power plant that would be more costly to run.
Finally, intraday trading can also be used to offer flexibility in own production/consumption to other market participants who are willing to pay more for it than the respective costs of running and rescheduling the corresponding power plants and dispatchable loads. Without intraday trading, this available flexibility might not be utilised because flexibility on intraday and balancing markets can have different characteristics: in terms of intraday trading, flexibility consists of the possibility of increasing or decreasing own energy generation/consumption measured as MWh during the period of delivery. In contrast to that, bids to the balancing markets offer flexibility in terms of increasing/decreasing own generation/consumption levels, measured in MW. Because of higher requirements on balancing bids, e.g. minimum capacity, activation times and purely physical fulfilment, not all flexibility identified by market participants during the intraday trading period can be offered on the balancing market. Even if transmission system operators (TSOs) procure balancing services either before the intraday market opens, e.g., in Germany, or after closure of the intraday market, e.g. in the Nordic countries (Norway, Sweden, Finland, and Denmark), intraday trading is important to access this flexibility and should be regarded as a complement to balancing markets rather than a substitute.
The value of intraday markets from a system’s perspective is that they can – at least on average – reduce the volume of activated balancing services; for example, if parts of the forecast errors related to variable renewable energy sources (vRES) can be handled shortly before real-time. Here, intraday trading can be advantageous because updated wind power forecasts are on average more accurate than day-ahead forecasts, which is due to reduced forecast horizons and the use of both weather prediction and recent measurements of wind power generation
In Europe, there are two predominant forms of intraday trading: discrete auctions and continuous trading
Continuous trading implies that trades can be settled whenever a market participant accepts an offer of another market participant. Therefore, prices vary from trade to trade. That is a substantial difference to auction-based intraday markets that are cleared at discrete times. Advantages and disadvantages of intraday auctions and platforms for continuous intraday trading are manifold.
The main argument in favour of continuous trading is that it allows market participants to trade whenever they can expect benefits from trading. This can, for example, be advantageous for risk-averse market participants who want to minimise price risks related to expected imbalances as early as possible. In addition, it allows market participants who face increasing cost the later they reschedule (e.g. efficiency loss and wear-and-tear) to offer intra-day flexibility at lower costs at an earlier time. For example, possibilities for rescheduling of thermal power plants are often more limited (ramping constraints, start/stop times, etc.) than for hydro units. This implies that flexibility in hydro units might be offered at comparatively low prices right before gate closure while flexibility in thermal power plants would become more expensive close to the period of delivery. The main disadvantage of continuous trading is a lower allocative efficiency due to its inherent first-come-first-serve principle. This implies that some trades with positive welfare contribution (intra-marginal trades in discrete auctions) might not to be realised while some trades with negative welfare contribution (extra-marginal trades in discrete auctions) might be settled. Using a simulation model, Weber and Schröder (2011) conclude that the larger the variations in a market participant’s willingness-to-sell and willingness-to-buy during the trading period the larger the efficiency loss of continuous trading platforms compared to discrete auctions. Other commonly discussed advantages and disadvantages refer to price transparency and ease of trade.
Intraday markets remain a research field with many open questions. For example, the majority of published models for production planning neglects intraday trading and focuses on optimising bids to day-ahead and balancing . This has two reasons: first, computational complexity which increases the more decision steps that are included in a stochastic optimisation problem. Second, modelling of trading behaviour on continuous intraday markets is not straightforward; for example, because it might always be more profitable for a market participant to trade energy for the same period of delivery at a slightly later point in time. In addition, low liquidity – which is a common peculiarity of intraday markets – complicates market modelling.
Trading behaviours on ELBAS
A recent paper by Scharff and Amelin looked at trading behaviour on ELBAS, which is the joint intraday market of Norway, Sweden, Finland, Denmark, Estonia, Lithuania, Latvia, the Netherlands and Belgium. So far, only few papers have analysed intraday trading on ELBAS. Mauritzen (2015) investigates the question whether hourly traded volumes on ELBAS are positively related to day-ahead wind power forecast errors. The econometric analysis was based on data for the Danish price zones and found that trading activity increases during hours in which wind power generation is overestimated (plausible), but also that trading tends to decrease when wind power generation is underestimated (unexpected). Holttinen and Koreneff (2012) investigated different ways a power generating company can handle its imbalances that stem from wind power forecast errors. In a case study based on wind power forecasts as well as imbalance and ELBAS prices from 2004, they compute the costs of a power generating company that balances the deviation between its day-ahead wind power forecast and an updated three-hours- ahead forecast by trading on ELBAS. They find that in an ex-post perspective intraday trading was on average less profitable for the power generating company than the corresponding imbalance costs. However, they base their assessment on a two-price system for imbalances.
Looking at data covering March 2012 to February 2013 (8736 hours of data) Scharff and Amelin found that trading activity varies strongly between price zones which in some cases would most likely be explained by the following contributing factors:
- Large share of wind power (Denmark),
- Limited transmission capacity that is available for intraday trading (Norway towards Continental Europe)
- Comparatively high balancing prices (Finland).
They also includes a more detailed study of the interrelations between intraday and balancing markets. Even though ELBAS is a continuous platform for intraday trading, most trades are done shortly before gate closure; the analysis showed that already 50 % of all trades are settled 3h before the period of delivery starts. This is an important characteristic that should be represented in market models of continuous intraday trading platforms
Trading close to gate closure is likely to be preferred by market participants who use ELBAS to achieve balanced positions according to updated wind power forecasts with short forecast horizons. For some thermal power plants, possibilities for rescheduling shortly before gate closure can be limited and more expensive. Here, more research should be done to investigate to which extent market participants with thermal power plants actually trade as early as possible, to which extent they currently offer intraday flexibility as well as to which extent and to which price they could offer intraday flexibility closer to gate closure. This could help answering the questions by which power plants an increasing demand for intraday flexibility could be covered and whether changes in the design of ELBAS and other continuous intraday trading platforms could contribute to a more efficient use of the available flexibility
Besides the areas outlined before, further research should tackle the question how to model continuous intraday markets. While the presented analysis provided several insights into continuous trading, it seems far from straightforward to model trading behaviour on ELBAS. However, this would be important for in- depth simulations of electricity market design as well as for ex- tended production planning models and bidding support tools.If you’ve found this blog helpful and would like other topics covered, please feel free to drop me an email with suggestions. You’re welcome to subscribe using ‘Subscribe to Blog via Email’ section and this will get you the latest posts straight to your inbox before they’re available anywhere else
- Trading behaviour on the continuous intraday market Elbas – Energy Policy (2016) –
- Pogosjan, D., Winberg, J., 2013. Förändringar av marknadsdesign och deras på- verkan på balanshållningen i det svenska kraftsystemet – En kartläggning och analys av de balansansvarigas arbetsgång. Master thesis, Uppsala University. URL 〈http://www.diva-portal.org/smash/record.jsf? searchId¼1&pid¼diva2:633126&rvn=3〉
- Dobschinski, J., Lange, B., Rohrig, K., Saint-Drenan, Y., Wessel, A., von Bremen, L., 2010. The potential of advanced shortest-term forecasts and dynamic predic- tion intervals for reducing the wind power induced reserve requirements. In: European Wind Energy Conference and Exhibition (EWEC)
- Henriot, A., 2012. Market Design with Wind: Managing Low-Predictability in In- traday Markets. Report RSCAS 2012/63. European University Institute and Ro- bert Schuman Centre for Advanced Studies. URL 〈http://cadmus.eui.eu/bitstream/handle/1814/24556/RSCAS_2012_63.pdf?sequence=1〉
- Scharff, R., Amelin, M., 2013. Distributed balancing of wind power forecast devia- tions by intraday trading and internal ex-ante self-balancing – A modelling approach. In: 24th International Workshop on Database and Expert Systems Applications (DEXA), pp. 176–183.
- Scharff, R., Egerer, J., Söder, L., 2014. A description of the operative decision-making process of a power generating company on the Nordic electricity market. En- ergy Syst. 5 (2), 349–369