Optimal bidding on hourly and quarter-hourly day-ahead electricity price auctions: trading large volumes of power with market impact and transaction costs

Electricity exchanges offer several trading possibilities for market participants: starting with futures products through the spot market consisting of the auction and continuous part, and ending with the balancing market. This variety of choice creates a new question for traders - when to trade to maximize the gain. This problem is not trivial especially for trading larger volumes as the market participants should also consider their own price impact. The following paper raises this issue considering two markets: the hourly EPEX Day-Ahead Auction and the quarter-hourly EPEX Intraday Auction. We consider a realistic setting which includes a forecasting study and a suitable evaluation. For a meaningful optimization many price scenarios are considered that we obtain using bootstrap with models that are well-known and researched in the electricity price forecasting literature. The own market impact is predicted by mimicking the demand or supply shift in the respectful auction curves. A number of trading strategies is considered, e.g. minimization of the trading costs, risk neutral or risk averse agents. Additionally, we provide theoretical results for risk neutral agents. Especially we show when the optimal trading path coincides with the solution that minimizes transaction costs. The application study is conducted using the German market data, but the presented methods can be easily utilized with other two auction-based markets. They could be also generalized to other market types, what is discussed in the paper as well. The empirical results show that market participants could increase their gains significantly compared to simple benchmark strategies.