Title: Cost-at-Risk estimation on the Italian ancillary services market
Authors: Luigi Grossi - University of Padova (Italy) [presenting]
Francesco Lisi - University of Padova (Italy)
Federico Quaglia - Terna SpA (Italy)
Abstract: The aim is to explore how to evaluate the cost risk related to a market-based ancillary services procurement by an electricity Transmission System Operator (TSO). We consider the case of Terna, the Italian TSO, which operates the Italian electricity ancillary service market. We propose a two-step procedure moving from the time series of incurred costs for the years 2017-2020 and some possible costs drivers such as actual consumptions, levels of reserve requirements, network nodal and zonal constraints, as well as prices of the primary commodities affecting the Italian electricity market (natural gas and carbon dioxide). Calendar variables are also included to account for periodic and other deterministic effects. In the first step of the procedure, we model the dynamics of the cost conditional mean, which can be viewed as the predictable part of costs. A non-parametric model describing the relations among variables impacting costs, possibly non-linearly, is identified and estimated. This model relies on the basis and periodic splines. Models residuals represent the unpredictable costs component and are used to quantify the Cost-at-Risk (CaR), the equivalent of Value-at-Risk in the case of costs. For this purpose, we compute and test different approaches for 1-day and 30-days CaR. Models for conditional quantiles dynamics were identified and estimated on the residuals of the conditional mean model and were based on lagged values.