The Impact of Imperfect Competition in Emission Permits Trading on Oligopolistic Electricity Markets.

• Author: Limpaitoon, Tanachai

1. INTRODUCTION

   In the recent years, growing concerns around the issue of climate change have led to numerous transformations in the electric power industry. These changes are partly driven by regulatory policies such as renewable portfolio standard and emission trading programs that rely on market-based mechanisms to mitigate emissions and/or promote renewable energy. One concern over the implications of these regulations is the possibility that some firms in the market posses market power in both electricity and permit markets. Such market power may manifest itself when a dominant firm can deliberately consume permits in order to raise other firms' production cost or it can withhold its capacity to drive up electricity prices. There is substantial empirical evidence that market power is a relevant issue in both permit and electricity markets (see e.g., Borenstein et al. 2002, Joskow and Kahn 2002, Mansur 2007, Puller 2007 on market power in electricity market; and Kolstad and Wolak 2003 on permit market).

   Since the onset of California electricity crisis, the presence of market power, even within the electricity markets alone, has drawn considerable attention. This is mainly because the complexity of market rules and regulations in the electricity markets have made the markets fundamentally different from other commodity markets. In such a scenario, where the issue arises from the existence of market power, a Nash-Cournot game is commonly applied to model an electricity market. Such models can be justified by empirical evidence suggesting that the performance of several U.S. regional electricity markets, e.g., Pennsylvania, New Jersey, and Maryland (PJM), New England, and California, is consistent with Nash-Cournot market outcomes (see Bushnell et al. 2008). Nevertheless, finding Nash-Cournot oligopoly equilibria could be complicated by the presence of a congestion-prone transmission network. For example, Neuhoff et al. (2005) showed that the inclusion of transmission constraints into the strategic models, used in the analysis of impacts of regulatory mechanisms, can result in unexpected equilibria. Cardell et al. (1997) and Borenstein et al. (2000) also demonstrated the effect of transmission congestion on firms' strategic behavior. The introduction of any emission regulation to a transmission-constrained electricity market can lead to some unintended consequences for market outcomes when transmission congestion occurs (see e.g., Downward 2010; Limpaitoon et al. 2011). These studies highlight the importance of incorporating transmission network into strategic models used in the analysis of market power in electricity markets.

   As mentioned earlier, a market for trading emission permits has also faced concerns over potential market power. Studies have shown that dominant firms are able to employ manipulative strategies in order to move prices in their favor. For instance, Kolstad and Wolak (2003) provided empirical evidence that firms used permit prices to justify higher costs of electricity produced, and for raising electricity prices. The earliest work on market power in the permit markets was introduced by Hahn (1984). He considered one dominant polluting firm and other price-taking firms in a static model. Sartzetakis (1997) studied a two-stage duopoly model, where the leader sets the permit price in the first stage and both firms' output and abatement decisions are made in the second stage. Resende and Sanin (2009) studied a three-stage model. This differs from Sartzetakis (1997) in that the follower in the permit market can anticipate the strategic response from its rival in the output market in the third stage, when deciding the optimal permit quantity in the second stage. Montero (2009) extended Hahn's model to accommodate the more contemporary market settings such as permit auctions. Tanaka and Chen (2011) considered a model for electricity market in which permit prices can be manipulated through fringe producers.

   Although existing models consider imperfect competition in permit markets, most have not taken into account neither the transmission network nor realistic network scale. The exception is the work by Chen and Hobbs (2005) who considered electricity and permit markets with a realistic, yet coarse, scale of transmission network. Our work differs from those existing models in that we examine simultaneous interactions, instead of a multi-stage model, of oligopoly competition between multiple players of transmission-constrained electricity market participating in a carbon permit market at a more realistic scale. Our approach takes into account essential market characteristics such as electrical loopflow, resource ownership, and transmission constraints based on thermal ratings. To account for the transmission network, we employ a direct current (DC) approximation, which is a common approach to analyze the impact of market power because of its tractability (Wei and Smeers 1999; Pang et al. 2001).

   In modeling market power in the permit market, we propose a conjectural variation approach, whereby the conjectural variation parameters are derived "empirically" through simulations. (1) The approach allows us to estimate conjectural variations under different permit endowments because our conjectural variation parameters are estimated endogenously, whereas the parameters derived in Chen and Hobbs (2005) are based on historical emissions. Our model is especially useful when a regulatory agency wants to examine the extent of market power, if exists, in joint electricity and permit markets. Furthermore, the proposed model is readily extended to account for other economy sectors participating in the permit market.

   The model is applied to a reduced 225-bus representation of the Western Electricity Coordinating Council (WECC) 225-bus network system for two reasons. (2) First, evidence has shown that the California market is among the least competitive restructured US electricity markets. (3) Second, California has recently implemented a cap-and-trade program (C&T) under AB 32, where market power is a concern. (4) As for the issues of permit allocation, (5) the paper is primarily focused on characterizing the equilibrium when emission permits are initially allocated for free to generation firms. This is because permits allocation in California under AB 32 is intended for various reasons other than efficiency and equity (e.g., mitigating price impact). Our main interest is to understand the potential strategies taken by firms, when they are given different permit endowments. A sensitivity analysis of permit endowments of firms may shed light on how firms strategically plan their permit positions. This is relevant to the issue of "holding limit" of emission permits that is currently under discussion by the California Air Resource Board in Resolution 11-32 (ARB 2011). We also look at various market scenario assumptions in attempt to investigate the extent to which simultaneous exercises of market power in both electricity and permit markets impact market outcomes.

   Our analysis shows that when an efficient firm (less polluting and low production cost) is "grandfathered" a substantial number of permits, it tends to strategically withhold the "unused" permits in order to place upward pressure on permit price, hence driving up the electricity price. The degree to which firms strategically withhold permits may be lessened, when a stringent cap is imposed, a situation in which permit prices tend to be relatively expensive. Also, the effect of the degree of competition in the permit market on social welfare is ambiguous and depends on the cap level. Finally, patterns of transmission congestion can be influenced by trading activities in the permit market. However, given that the scope of a C&T policy covers more than one sector, the case study may underestimate the price elasticity of emission permits, therefore inflating permit prices. Along with the adoption of other complementary measures to mitigate potential for market power, the possible market outcomes, as we argue, are likely bounded by our simulation results.

   This paper is organized as follows. Section 2 describes an equilibrium model that includes optimization problems and market equilibrium conditions. The concept of conjectural variation in the context of permit market is also introduced. Section 3 describes a case study of the California market through the WECC 225-bus network in conjunction with a C&T program. In this section we simulate various market scenarios, propose the estimation procedure for conjectural variations, and discuss the economic implications of a C&T program on the electricity market in the light of market power in the joint markets. Section 4 concludes.

2. MODEL

   We extend the model proposed by Limpaitoon et al. (2011) to account for market power in an emission permits market via conjectural variations. In particular, instead of having the permit price being determined endogenously within a perfectly-competitive equilibrium framework using a complementarity constraint, (6) imperfect competition in the permit market, herein, assumes that each firm makes a conjecture about other firms' demand for permits. Each firm's decision regarding the number of permits to hold at equilibrium is, therefore, based on its perceived residual supply of permits. A firm's conjectured price response in the permit market--which relates the price and quantity of permits that the other firms are willing and able to purchase--is represented by a downward-sloping demand curve. These curves are constructed from the simulation results of a Nash-Cournot game of an electricity market and a perfectly-competitive permit market by varying levels of emission cap. The simulation of the joint markets yields the demand for permits, resulting from electricity producers' unabated emissions as they behave a` la Cournot in the electricity market but behave as...