Application of a Structural Model to the Spanish Electricity Wholesale Market

Abstract

1. Framework The aim of this study is to analyze the behavior of agents in the Spanish electricity market during the period from January 1999 to June 2007 before MIBEL (Iberian Electricity Market) started. The main questions that we tried to give an answer to are: * Did market power exercise occur during a long period in this market? * What kind of long term strategies have been observed during the analyzed period? The analysis is carried out in the theoretical framework of structural models. This framework is based on the existence of causal relationships between related variables, explained by economic theory. These relationships are, in general terms, expressed by the resolution of a system of equations, thus implying economic equilibrium. Structural models provide a heuristic and causal approach about the market's economic relationships. Then also allowing for the answer to other questions, namely for the price elasticity of demand. Section 2 presents the organization of the Spanish wholesale market. The methodological approaches are presented in section 3, namely the models developed. This section includes the survey of the various methodological approaches, and presents the particularities of the methodology applied: the research is carried out through the structural model methodology, and the results are tested with the direct estimations of the main variables. The structural model has two equations, one with the demand function and the other with the profit maximization function. The price elasticity of demand is estimated in section 4 through the first equation. The behavioral factor is estimated in section 5, through the profit maximization function 2. Framework The former Spanish wholesale market can be characterized by being an Uniform Price Auction (UPA) Market. In these markets, the generator which sells the marginal quantity defines the system marginal price. This price is paid to all producers with accepted bids. In that case the price is defined for each hour. This market presents a strong regulatory framework. The main regulatory drivers were the stranded costs compensations (CTC), (from 1998 (Ley 54/1997), until 2006 (Real Decreto-ley 7/2006), with a decreasing influence in producers income over this period). It also has to be referred the price cap imposed to the transaction prices between producers and buyers belonging to the same companies (Real Decreto-ley 3/2006) and the limit imposed at the biggest companies in order to not increase their market share (Real Decreto-Ley 6/2000). This market was highly concentrated. There were 4 main producers. However the two biggest, Endesa and Iberdrola represented about 3/4 of the production sold in the wholesale market. But their importance tended to decrease (75% in 2002 and 60% in 2006). 3. Methodological Aspects 3.1 The Games The market for power generation is very much like a market with Cournot strategies, with capacity constraints (Kreps and Scheinkman, 1983). Within this framework, the quantities correspond to the decision variable. Even when the price is assumed as a strategic variable, the results of the strategies are similar to the Cournot game, as there is capacity constraint in that kind of market (see Wolak and Patrick, 1997). Starting with the Nash-Cournot solution for the Spanish oligopolistic market and reprising the Cowling-Waterson formula (1976), the Lerner index and the strategies developed by companies can be correlated using the [theta]4 index (5): (P - [bar.Cmg])/P = [bar.[theta]]HHI/[parallel][epsilon][parallel] = [lambda] (1) in which [bar.Cmg] is the weighted marginal cost for the industry, HHI the Herfindahl Hirschman index and [lambda] the factor measuring the level of market power, i.e. which corresponds to the Lerner index. In this case, the Lerner index is directly related to the level of concentration and also related to the firms' conjectural variations. …