The thesis analyzes the effectiveness of the Segmented Pay-as-Clear (SPaC) mechanism as an alternative to traditional Pay-as-Bid and Pay-as-Clear regimes in the Italian electricity market. During the 2022-2023 energy crisis, the Pay-as-Clear mechanism generated excessive inframarginal rents for renewable sources, passing on gas prices in full to the entire electricity production sector. SPaC proposes a segmentation between sources with negligible marginal costs (renewables) and non-negligible marginal costs (fossil fuels), determining separate prices to limit these rents. Through simulations based on reinforcement learning (Q-learning), two scenarios are analyzed: an idealized one with two symmetrical operators according to the PNIEC 2030, and a realistic one with ten Italian market operators reconstructed through clustering of 2024 public offers. The results show that SPaC significantly reduces inframarginal rents compared to Pay-as-Clear, keeps costs low even with aggressive strategies (higher than Pay-as-Bid), and provides positive incentives for all operators, regardless of the technological mix of the portfolio, suggesting greater political sustainability of the reform.
La tesi analizza l'efficacia del meccanismo Segmented Pay-as-Clear (SPaC) come alternativa ai tradizionali regimi Pay-as-Bid e Pay-as-Clear nel mercato elettrico italiano. Durante la crisi energetica 2022-2023, il meccanismo Pay-as-Clear ha generato rendite inframarginali eccessive per le fonti rinnovabili, trasmettendo integralmente i prezzi del gas all'intera produzione elettrica. Lo SPaC propone una segmentazione tra fonti a costo marginale trascurabile (rinnovabili) e non trascurabile (fossili), determinando prezzi distinti per limitare tali rendite. Attraverso simulazioni basate su reinforcement learning (Q-learning) vengono analizzati due scenari: uno idealizzato con due operatori simmetrici secondo il PNIEC 2030, e uno realistico con dieci operatori del mercato italiano ricostruiti tramite clustering delle offerte pubbliche 2024. I risultati dimostrano che lo SPaC riduce significativamente le rendite inframarginali rispetto al Pay-as-Clear, mantiene costi contenuti anche con strategie aggressive (superiore al Pay-as-Bid) e garantisce incentivi positivi per tutti gli operatori, indipendentemente dal mix tecnologico del portafoglio, suggerendo maggiore sostenibilità politica della riforma.
PREZZI DELL'ENERGIA E DISACCOPPIAMENTO ELETTRICITÀ-GAS: NUOVI MODELLI DI MERCATO DEL GIORNO PRIMA
ALTAMURA, ANDREA
2024/2025
Abstract
The thesis analyzes the effectiveness of the Segmented Pay-as-Clear (SPaC) mechanism as an alternative to traditional Pay-as-Bid and Pay-as-Clear regimes in the Italian electricity market. During the 2022-2023 energy crisis, the Pay-as-Clear mechanism generated excessive inframarginal rents for renewable sources, passing on gas prices in full to the entire electricity production sector. SPaC proposes a segmentation between sources with negligible marginal costs (renewables) and non-negligible marginal costs (fossil fuels), determining separate prices to limit these rents. Through simulations based on reinforcement learning (Q-learning), two scenarios are analyzed: an idealized one with two symmetrical operators according to the PNIEC 2030, and a realistic one with ten Italian market operators reconstructed through clustering of 2024 public offers. The results show that SPaC significantly reduces inframarginal rents compared to Pay-as-Clear, keeps costs low even with aggressive strategies (higher than Pay-as-Bid), and provides positive incentives for all operators, regardless of the technological mix of the portfolio, suggesting greater political sustainability of the reform.I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14253/88047