|
Article Excerpt
Abstract In this paper we analyze a time series of daily average prices in the Italian electricity market, which started to operate as a Pool in April 2004. Our objective is to model the high degree of autocorrelation and the multiple seasonalities in electricity prices. We use periodic time series models with GARCH disturbances and leptokurtic distributions and compare their performance with more classical ARMA-GARCH processes. The within-year seasonal variation is modelled using the low-frequency components of physical quantities, which are very regular throughout the sample. Our results reveal that much of the variability in the price series is explained by the interactions between deterministic multiple seasonalities. Periodic AR-GARCH models seem to perform quite well in mimicking the features of the stochastic part of the price process.
Keywords Electricity auctions * Periodic time series * Conditional heteroskedasticity * Multiple seasonalities
JEL Classification C22 * D44 * L94 * Q40
Introduction
Electricity prices as they are now determined in regulated (generally, Pool) markets, where private operators have replaced previously well established public enterprises, present specific behavioural characteristics. Prices in these Pool markets differ from the previous prices which were fixed by governments or public agencies. Despite the very limited storability and transportability of electricity, government-determined prices incorporated little uncertainty in their dynamics as they were generally capped by the imposition of price ceilings, resulting from the implementation of welfare-improving tariff policies. Market determined electricity prices, on the other hand, are strongly affected by the impossibility to arbitrage between time and space, and consequently are very volatile.
Time series of current electricity prices differ quite substantially from prices determined in the markets for financial assets or other type of commodities, since electricity cannot be treated like a stock. Hence, electricity prices have specific and somewhat unique characteristics (e.g. strong seasonalities and mean reversion), which in the recent past was the motivation for using time series modelling to study the specific features of their time patterns and to evaluate how prices are affected by temporal demand-supply imbalances, seasonality, transmission congestion and, to a lesser extent, by the features of the mechanism that generates the data (type of auction employed, price rule, degree of market concentration, etc.).
In Italy the privatization of the former public (quasi-) monopolist eventually led to the creation of an electricity Pool with some specific legal characteristics such as the presence on the demand side of a single buyer and which began operations in April 2004. In this paper we try to describe the price dynamics of the Italian Pool and compare our findings with those obtained from analysis of other European markets. We also suggest a new methodology for dealing with multiple seasonalities that may prove useful for future econometric analysis of electricity prices. The stochastic part of the price evolution is modelled through periodic AR-GARCH processes, which are able to capture the different memory patterns of each day of the week.
Our results can be exploited directly by companies and traders for forecasting and for improving their bidding strategies, by regulators for identifying possible anomalous pricing behaviours, and, indirectly, by hedgers for the pricing of energy derivatives.
The paper is organized as follows. We first discuss the main characteristics of some European electricity markets, including the Italian Pool, for which data availability permits time series analysis. We emphasize differences in the organization and regulation of the markets as well as in their production structure (specifically, electricity generation), which might be important for explaining differences in the econometric results. We then review the existing econometric literature. Our empirical analysis starts with a description of the general characteristics of the Italian data and concentrates on modelling the deterministic component of the data. We go on to describe and justify the choice of stochastic models and methods employed to account for the dynamics in the Italian prices. The paper concludes by presenting the results accompanied by some detailed comments.
The Electricity Markets in Italy and Other European Countries
The Italian electricity market (IPEX) is organized as a Pool system, managed by a market operator (GME), that collects the bids, determines the merit order for the dispatch of electricity and is responsible for all auxiliary services. The Pool, initially planned to come into force in January 2001, actually began operations at the end of March 2004, as a one-side market. A single buyer, constituted in 1999, was responsible for guaranteeing the supply of electricity to a set of captive customers. Demand-side bidding was introduced in January 2005.
IPEX was the most recent Pool market to be created in Europe. Following the establishment in 1991 of the England and Wales (E & W) Electricity Pool, as a result of the liberalization of the British electricity market, competitive electricity markets have been organized in many European markets: Norway, Finland, Denmark, Austria, France, Germany, Netherlands, Spain. The key features of most of these changes, including the Italian Pool market, were the privatization and restructuring of the existing vertically integrated monopolistic supplier. The second step was to organize the exchange of physical electricity in competitive wholesale spot markets through auctions. Competition was introduced at the retail level while transmission/distribution, which were still considered natural monopolies, remained under government regulation. This reorganization of the industry led to a separation of potentially competitive elements from natural monopolies.
The wholesale exchange of electricity poses some problems for regulators in relation to market architecture and design. First a decision must be made about whether to opt for a centralized Pool or for a decentralized market. In the former arrangement, all the electricity is allocated through the Pool, which is then mandatory; thus bilateral contracts are not allowed. All operators, both on the demand and the supply sides, submit hourly or half-hourly bids which are matched by a procedure that minimizes the cost of despatch. A decentralized electricity market such as NETA (England), on the other hand, is organized as a series of voluntary forward and spot markets and bilateral contracting is allowed. The advantages of a Pool over a decentralized market are that demand and supply are continuously matched so that all coordination problems disappear. Advocates of the decentralized market structure, however, emphasize that the Pool may be affected by strategic bidding on the part of those operators with market power and, as a consequence, Pool prices do not...
|
