Analysis of Mean and Volatility Price Transmissions in the MIBEL and EPEX Electricity Spot Markets.

• Author: Ciarreta, Aitor

1. INTRODUCTION

   The European Union (EU) set the ambitious target of fully integrating national energy markets by 2014. Market integration is intended to encourage more intense competition and security of supply. The progressive harmonization of electricity market rules and the increased interconnection of national grids are at the heart of promoting an effectively competitive internal market.

   However, electricity markets in Europe are not yet actually harmonized. Some progress has been made since the first steps were taken with the passing of Directive 96/92/EC, which set common rules for the construction of the single European electricity market by which country members must favor the interconnection and interoperability of systems. However, this Directive was not very successful because the degree of openness and the speed of reforms were very uneven. Moreover, cross-border electricity trade did not increase significantly, and this jeopardized the ultimate goal of building a single market. Subsequently, Directive 2003/54/EC established "common rules for the generation, transmission and distribution of electricity", but set no further targets for interconnections. (1) In 2007 the European Commission (EC) published its third energy package, in which it identified the lack of electricity market integration as one of the factors distorting competition. This mainly results from insufficient interconnecting infrastructures between national grids, and incompatible market design between transmission system operators (TSO) and spot market operators.

   In the EU, much work still needs to be done with respect to aligning national market and network operation rules and cross-border investment. However, regional market integration has already been launched: MIBEL, the Iberian electricity market formed by Spain and Portugal was set up in July 2007, and the EPEX power exchange, which includes France, Germany, Austria and Switzerland, in 2010. These are two examples of the Price Coupling of Regions (PCR) initiative launched in 2010. (2) Our aim here is to analyze the degree of market integration between bordering countries, with reference on the one hand to the MIBEL and EPEX markets and on the other hand to the situation of the Spanish and French electricity market interconnection for the period from July 1, 2007 to February 29, 2012. Within MIBEL and EPEX, transmission capacity has sharply increased. However, the interconnection capacity of the Spanish and French markets has barely increased since 2007. Moreover, the Spanish Energy Regulatory Commission (CNE, 2012) has raised concerns as to the efficiency of the management of the French-Spanish interconnections (3). We compare the degree of integration of these markets, seeking to determine the extent to which the common rules of Directive 2003/54/EC lead to a single electricity market. It is possible that shocks in one country may be transmitted to more distant non-bordering markets, however we would expect this to be a second-order effect. Nevertheless, multivariate models for the four electricity markets of the EPEX together are also considered.

   There are several studies which confirm the lack of full European electricity market integration. Zachmann (2008) finds evidence of price convergence for several pairs of countries for 2002-2006, but the target of a common electricity market remained unfulfilled. Bunn and Gianfreda (2010) observe increasing market integration for 2001-2005 and conclude that the better interconnection capacities become, the greater integration will be. Balaguer (2011) focuses on the pricing behavior of two non-EU exporter countries--Switzerland and Norway--for 2003-2009. He finds market integration with respect to the two EU regions where they sell. Nevertheless, a single electricity market has not yet been reached. Lindstrom and Regland (2012) model prices in six European electricity markets and conclude that market integration is only partial, as they do not find strongly dependent prices. Outside Europe, De Vany and Walls (1999a) analyze US electricity market integration for the period 1994-1996 and find stronger integration for the off-peak market pairs. Worthington et al. (2005) study volatility transmissions between the regional electricity markets in the NEM Australian market using data from 1998 to 2001 and conclude that regional markets are not fully integrated due to the lack of a well developed network of interconnections. Later, in the same market but for the extended period of 1999 to 2007, Higgs (2009) finds that interaction between markets is highly dependent on geographical proximity and interconnections.

   The methodology used in this paper is based on multivariate GARCH models (MGARCH), which enable mean and volatility transmissions of prices between different markets to be studied. An analysis of the presence and degree of interdependence between markets and a study of price convergence are used to determine whether the regional markets selected are integrated. Constant and dynamic conditional correlation models in which seasonal effects are taken into account are estimated and compared, among others, to select the best models and analyze the mean and volatility transmissions. Previous studies using MGARCH models include Bystrom (2003) and Malo and Kanto (2006) for Nord Pool spot and future prices, Sadorsky (2012) for oil prices and the stock prices of clean energy companies and technology companies, Worthington et al. (2005) and Higgs (2009) for NEM Australian spot prices, and Veka et al. (2012) for Nord Pool, EEX and ICE electricity contracts. Frommel et al. (2014) forecast the daily price volatility in the EPEX market using Realized GARCH-type models.

   The rest of the paper is structured as follows. Section 2 explains the market design and the interconnection between MIBEL and EPEX countries. Section 3 presents the data and provides some descriptive statistics. Section 4 describes the methodology and Section 5 summarizes the results. The paper concludes with some policy implications of the study.

2. MARKETS AND INTERCONNECTIONS

   The MIBEL and EPEX electricity markets comprise all transactions resulting from the participation of agents in the day-ahead market and the sessions of the balancing markets, respectively. In the day-ahead market, OMIE and EPEX agents submit offers for each hour of the following day. Offers are sorted in increasing (decreasing) order if they are selling (purchasing) bids. Then both work as uniform price day-ahead markets where a system marginal price is set for each hour of the following day. In the absence of congestion a single marginal price emerges in each market. However, if there is congestion then there is market splitting and each pole has a different marginal price.

   For MIBEL, the proportion of hours with congestion is 34.5%, of which 32% result in a lower price on the Spanish market than on the Portuguese one. For EPEX, the proportion of hours with congestion is close to 99% of total hours, being in general with higher prices in Switzerland (68.2% in the French interface, 72% in the Austrian interface and 72.2% in the German interface). This highlights the fact that despite EPEX being one of the European interconnections in which the share of total consumption has increased most rapidly, congestion is still significant, as compared to MIBEL.

   The relative importance of each market differs. On average, the share of electricity traded as spot during the sample period is 80% for OMIE, and 25% for EPEX. It should be noted that market rules establish a maximum price of [euro]180.3/MWh for MIBEL and [euro]3000/MWh for EPEX. There is also a minimum price of [euro]0/MWh for MIBEL and [euro]--3000/MWh for EPEX, which means that negative prices can emerge in equilibrium for some hours and, consequently, negative daily prices can occur, as is the case in the German market.

   In order to achieve market integration it is important to have sufficient inteconnections such that price-reponsiveness is achieved. Figure 1 plots the share of total imports from total consumption and the fraction that comes from the rest of countries within the same power exchange.

   It is noteworthy that electricity imports represent a larger fraction of total consumption in the smaller countries (on average 16% for Portugal, 34% for Austria and 57% for Switzerland) than in the larger countries (on average 3% for France and Spain, and 8% for Germany). Geographical reasons explain why Switzerland and Portugal import all their electricity from other power exchange members. However, although Spain has increased the share of imports from Portugal, in EPEX this has not been the case. Austria's share of imports has fallen, in France the figure is lower in 2011 than in 2007 and in Germany it remains more or less constant. Thus, there is still potential for further increases in interconnection capacity, especially with the increase in market share of intermittent generation from renewables. (4)

   As mentioned, the explicit auction mechanism at the Spanish-French interface results in inefficient allocation of electricity on many days. However, this is not the case within MIBEL or EPEX where electricity flows follow price signals. Table 1 reports the number of days on which there is inefficient exchange between Spain and France.

   Note that on between 29% and 46% of the trading days between Spain and France there is non efficient allocation of electricity in one direction or the other, although the figure has declined significantly since 2010 as a result of the reinforcement of the market allocation mechanism in the 2009 new capacity allocation rules.

3. DATA AND DESCRIPTIVE STATISTICS

   The data used consists of daily electricity prices (measured in [euro]/MWh) from Spain, Portugal, Germany, Austria, Switzerland and France. Prices for Spain and Portugal are obtained from OMIE (2012) and those for Austria, France, Germany and Switzerland...