Renewable Energy and Market Power in the Italian Electricity Market.

• Author: Bigerna, Simona

1. INTRODUCTION

   The Italian electricity market (IPEX) is one of the largest markets in Europe and is organized according to the usual structure designed by the EU Directive on the rules for the internal market in electricity (EU, 2009) in a day-ahead market (DAM), an intra-day adjustment market and an ancillary market for resource dispatching. In the DAM the average liquidity is around two thirds of the total electricity consumption of the country. Italy is a large consumer and a net importer of electricity, because imported electricity is generally cheaper than domestic generation as confirmed in Authority and GMEs' annual reports.

   In fact, fossil fuel fired plants constitute the bulk of Italian generation capacity, especially Combined Cycle Gas Turbine plants, (1) making the marginal generation cost higher than the rest of Europe. In the recent years, besides the traditional hydro sources, there has been a dramatic development of new renewable energy sources (RES): wind and solar, spurred by generous subsidy policy. These developments have followed the localization of the availability of natural resources, wind and sun in the south and on islands Sicilia and Sardegna (Sicily and Sardinia), while the localization of consumption remained in the more industrialized north of Italy.

   There are two economic consequences of this massive RES deployment, which are interesting for our analysis. First, there has been an increasing strain on the transmission network, resulting in line congestion. (2)

   Secondly, the dispatch priority enjoyed by RES in the merit order in the DAM has determined a large fraction of supply bids recorded at zero prices. This fact has determined a structural shift of the supply function to the right and lower equilibrium prices in the DAM, especially in the daily peak hours, when there is abundance of solar RES. Indeed, in some hours the equilibrium price has been set to zero. This is obviously a wrong signal to the market, because it discourages long-term investment in new generation. In general, the traditional model of the electricity market functioning should yield equilibrium prices higher in the peak hours and lower in the off-peak hours. However, in Italy after 2012, there has been a reversal of equilibrium prices, whereby average prices in the peak hours have been structurally lower than average prices in the off-peak hours (GME, 2013). In particular, the highest prices occurred around the evening hours, with some sudden price spikes around 20:00-22:00.

   From a technical viewpoint, this fact can be explained for each generator by the need to ramp up its flexible and costly gas-fired plants at the sunset, precisely during those hours when the photovoltaic (PV) generation fades away. Nonetheless, from an economic viewpoint, there remains open the question, whether the gas-fired plants enjoy some oligopolistic market power in that period. In other words, we conjecture that generators could exercise some market power by bidding aggressively during the evening hours, in order to increase or anyway recover the profitability of their plants, which are suffering the displacement by the RES during the conventional daily peak hours.

   In summary, we deem that market power is a structural feature of electricity spot markets (see, Borestein, 2000; 2002), despite all attempts to improve competition and to introduce mitigation measures. We think that RES abundance has become a relevant issue in the electricity market for two reasons: i) determining low equilibrium prices in some hours and high prices in other hours; ii) increasing congestion in some hours.

   The aim of this paper is to investigate whether massive RES deployment has affected the exercise of market power in the Italian electricity market. Specifically, we want to understand whether or not the traditional market power of the few large generators has been weakened or concentrated in some specific periods of the day, as a strategic response to the massive competition of the myriad of small RES generators.

   In our analysis, we take into account that RES deployment and the resulting increase in the supply tends to lower prices but also that fostering congestion leads to higher market prices. Furthermore, strengthening of market power could increase prices. It is therefore necessary to construct a model to disentangle market power from congestion effects. (3) Following the methodology of Bigerna et al. (2016), we construct a measure of market power, net of the congestion effect, computing the zonal Lerner Index (ZLI).

   We investigate the period 2010 - 2013 of the IPEX. In this period the share of RES in total electricity consumption has dramatically increased from 3-4% in 2010 to 30% in 2013.

   This paper is organized as follows. Section 2 describes the main developments of the IPEX in the period analyzed. Section 3 describes the data and the method used to measure market power and congestion effects at the zonal level. Section 4 discusses the empirical analysis and results. Section 5 summarizes our primary findings. The method details and statistical information are provided in the Appendices.

2. THE ITALIAN ELECTRICITY MARKET DEVELOPMENT

   The DAM in the IPEX is organized every hour by a centralized operator, which collects the supply and demand bids and orders them according to a merit order. The operator entity in Italy is the Electric Market Operator, in Italian, Gestore del Mercato Elettrico (GME). In the IPEX there are two other operators: the Transmission system operator (TSO), which is responsible for acquiring reserves and for the network security management; and the RES operator, in Italian Gestore del sistema elettrico (GSE), which is responsible for bidding the RES supply in the DAM and for managing the RES subsidy policy of the Government.

   The price that equates aggregate supply and demand in each hour is a System Marginal Price (SMP) paid by buyers and received by generators. According to market rules, the priority dispatch of RES implies that RES supply is inserted in the merit order with zero price by the GSE. This means that the equilibrium market price can be zero, if the marginal supplier is a RES generator. Demand bids without indication of a price are recorded at the conventional price of 3000 EUR/MWh, which is the price cap imposed to generator bids by the Energy Authority. This means that the equilibrium price cannot exceed this limit.

   In case of bids that exceed the physical transmission capacity of the network across geographical zones defined by the TSO, market splitting occurs and there are different zonal SMPs. The resulting price is higher in an importing zone, where more costly plants are called to meet the demand and it is lower in an exporting zone, because some plants cannot fulfil their export potential. Generators in each zone receive the zonal SMP, while all buyers throughout Italy pay a unique SMP, which the average of the zonal SMP.

   Operationally, there are six geographical zones (see Figure A1 in Appendix) in the Italian market, according to the TSO definition of critical transmission capacity. They are defined as: North Italy (N), Center-North Italy (Cn), Center-South Italy (Cs), South Italy (S), Sicily (Si) and Sardinia (Sa).

   When line congestion is determined by the intended flows in the IPEX, market splitting can occur, segmenting Italy in two or more adjacent market zones. The first five geographical zones stretch from North to South, so that market splitting can occur along the border of any two adjacent zones. Sa is an Island connected with the mainland with two HVDC (High Voltage Direct Current) lines, one linked to Cn and the other linked to Cs. This means that for Sa there are four cases to be considered. Sa is segmented alone; it is part of a market zone that includes Cn but not Cs; it is part of a market zone that includes Cs but not Cn; it is part of a market zone that includes both Cn and Cs.

   Notice that there are different combinations of geographical zones that can form a market zone. For instance, when there is a two-zone configuration, there are 6 different ways to split Italy in two market zones (N-CnSaCsSSi; NCn-SaCsSSi; NCnSa-CsSSi; NCnSaCs-SSi; NCnSaCsS-Si; NCnCsSSi-Sa). Likewise, there are 11 different ways to split Italy in three zones, 6 ways into four zones, 4 ways into five zones and, trivially, 1 way into six zones.

   In the period 2010-2013, RES supply grew strongly (in particular PV, see Figure A2 in appendix). At the end of 2010 the number of PV plants was 156,000 (double that in 2009) for a total of 3,500 MW of installed capacity (triple than on 2009). The boom occurred in 2011 with 330,000 plants for a total of 12,700 MW of installed capacity, a growth of 268% over 2010. The growth continued with 481,000 plants and 16,700 MW of installed capacity in 2012 and 591,000 plants and 18,000 MW of installed capacity in 2013 (Table 1).

   In the period 2010-2013, there are a total of 35,064 hours (one year is leap-year). The effect of line congestion is the splitting into a total of 75,583 market zones and consequently there are 2.16 average number of market zones per hour to be considered (Table 2).

   The distribution of the hourly number of zonal markets is skewed. The situation of no congestion, i.e. one-market (ITA) occurred for 4,325 hours or 12.3% of the hours in the period. Notice that in the period 2004-2007 it was 18.2%, so the congestion problem has worsened in recent years. The most common configuration is two-market splitting, which occurred for 18,117 hours (51.7% of the total). Notice that the most frequent configurations are by far the splitting between Si and the rest of the country (NCnSaCsS-Si), which occurred for 15,898 hours (45.3% of the total). There are few other occurrences of splitting between Sa and the rest of the country (2.4%) or NCn and the rest of the country (2.0%). The three-market splitting occurred for 10,208 hours (29.1% of the total) in 11...