Italy-Usa: comparing electricity market designs to improve efficiency

The Italian transmission system operator (Terna) and researchers from the Program on Energy and Sustainable Development (PESD) at the renowned Stanford University have been successfully cooperating on no fewer than three projects since 2019. The main goal has been to detect and assess trends in electricity markets, by seeking solutions capable of improving market efficiency relevant for businesses and consumers alike.

The presentation of the results of these projects was the focus of a Tech Talk organised by Terna end of May 2022. During these periodic digital internal communication events, Terna hosts high-profile personalities from institions and the academia to discuss topics of greatest interest. This edition specifically focused on the outcome of the last project carried out in 2021: Enhancement of the Ancillary Service Procurement Process. Terna together with researchers from PESD analysed offer strategies in the Italian electricity market adopting an innovative regressions-based approach capable to identify key variables that drive the offer decisions made by market participants. Taking part in the talk were Christoph Graf, a researcher at PESD, and Federico Quaglia, Manager of the Operational Analysis area of the Dispatching Department, who coordinated and contributed to the research from Terna side.

«The project focused on assessing the offer strategies adopted by market participants in the electricity market with the goal of understanding how they adapt their offers according to expected operating conditions (e.g., Demand and renewable infeed) but also based on how they expect Terna to act in the Dispatching Services Market (MSD)», explains Federico Quaglia.

Two additional topics have also been addressed in their recent research project: (1) if and how the launch of the capacity market and the MSD incentive scheme at the start of 2022 impacted operators' offer strategies, and (2) the development of a flexible algorithm based on machine-learning techniques, which allows Terna to predict expected ancillary service market costs. A scenario risk analysis using this tool serve as the basis for Terna's actions.

The researchers showed that market participants' offers in the electricity market are strongly influenced by both the expectations regarding ancillary services market needs as well as expected system conditions. «We have developed a model that is able to predict offer curves at the generation unit level», announced Christoph Graf. «After reviewing the literature, a model based on predictive machine-learning techniques was developed, which takes past data as input to successfully identify the base parameters for predicting future offer curves». A useful approach, particularly in the medium term, with monthly and annual time frames, to evaluate possible reactions from market participants to changes in system conditions.

Finally, with regard to the impact of the two major game-changers of 2022 (i.e., the launch of the capacity market and the MSD incentive scheme), the researchers have found that the capacity market has led to a general reduction in the average price offered in the electricity market (given equal variable costs). Moreover, the change was found to be most pronounced in the south of the country, where Terna implemented various steps to improve efficiency over the course of 2021 and 2022. «This means we’ve shown that the efficiency-improving steps implemented by Terna in the ancillary services market automatically pushed market participants to behave more competitive in the market», explains Quaglia. In summary, the three-year project in collaboration with Stanford University has yielded 7 significant results:

* a better understanding of the US electricity market design, which is based on the main alternative paradigm to the one currently adopted in Europe;

* an overview of the main mechanism introduced in the United States to mitigate the exercise of market power (a phenomenon which can cause a huge money transfer from consumers to producers);

* a benchmark approach for assessing the impact of the exercise of market power in the Italian electricity market over recent years;

* a quantitative analysis of the link between day-ahead market offer strategies and ancillary service market conditions;

* a preliminary proposal for a mechanism to monitor market power which can also be applied to the Italian context;

* a tool to predict day-ahead market offer strategies, potentially useful for simulations of medium-term scenarios;

* an evaluation of the main drivers affecting offer-related behaviours, confirming the positive impact of the steps implemented by Terna in 2022.

Terna's collaboration with Stanford University will not end here: the two organisations have already announced the launch of a fourth project aimed at further investigating and improving the mechanisms for analysing and simulating the electricity markets.