Price Formation in Auctions for Financial Transmission Rights.

• Author: Opgrand, Jeff

1. INTRODUCTION

   Since the passage of FERC Order 888 in 1996, competitive electricity markets have expanded in the United States to serve roughly two-thirds of electricity consumers in the country. The Order encouraged open access to transmission facilities, the divestiture of vertically integrated utilities, and the creation of Independent System Operators to administer competitive markets. A key feature of competitive electricity markets is a location-based pricing system. For competitive market participants, location-based pricing implies location-specific price risk due to potential network congestion that can cause price differences across nodes. The presence of uncertain network congestion inspired the creation of a financial product to hedge locational price differences (Hogan, 1992). In U.S. electricity markets, this financial product is called a Financial Transmission Right (FTR). These financial products are used by market participants to manage exposure to the risk of price differences between two locations on a transmission network.

   FTRs are sold in auctions administered by an Independent System Operator. The revenue raised in these auctions is allocated to load-serving entities (LSEs) to reimburse their electricity customers for expected congestion payments they will incur in the energy market. However, recent analysis shows that FTR auctions are persistently profitable for speculators, and that, on average, electricity customers are not fully reimbursed for their congestion payments. (1) One common explanation for the auction revenue shortfall is that the FTR auction process is inefficient (Deng et al., 2010; Olmstead, 2018). (2) In this paper, we propose an alternative explanation for persistent congestion reimbursement shortfalls, which is the role of trading premiums demanded by auction participants. Essentially, the trading premium of an FTR adjusts the FTR's bid price to account for the market participant's risk aversion and/or transaction costs.

   Our main contribution is a conceptual and empirical analysis of the market mechanism used to reimburse electricity customers for their expected congestion payments. This mechanism, called the Auction Revenue Right (ARR) process, gives an LSE a choice between acquiring an FTR at no cost or selling the same FTR in the annual FTR auction and receiving the associated auction revenue. Given that the choices made by LSEs in the ARR process determine fundamental supply conditions in the FTR auction market, we develop a conceptual framework that describes how different auction equilibria emerge under different ARR decision-making regimes. A key insight is that even if the FTR auction market is fully competitive, an LSE selling an FTR through the ARR process may result in a financial transfer from electricity customers to FTR buyers through a buyers' trading premium. One component of the trading premium is a risk premium adjustment due to the extreme difficulty of forecasting the future payout of an FTR.

   We test the predictions from our conceptual model using data from the PJM market. PJM is a wholesale electricity market in the eastern United States serving 65 million customers. We study ARR management strategies and outcomes in PJM using publicly available data on auction results, realized network congestion, auction participant classifications, and various other components. We find robust empirical evidence that variation in ARR management strategies helps explain differences between an FTR's auction price and its realized ex post value.

   Previous studies have examined the efficiency of FTR auctions (Adamson et al., 2010; Deng et al., 2010; Olmstead, 2018) and analyzed the presence of abnormal returns in FTR markets (Baltadounis et al., 2017). While our empirical finding that FTR auction prices diverge from their ex post value is consistent with the literature, we differentiate ourselves from these previous studies by focusing on the role of FTR supply (or lack thereof) in determining an FTR's equilibrium auction price.

   To explain the role of the ARR process in price formation in FTR auctions, we organize the rest of the paper as follows. Sections 2 and 3 provide an overview of competitive electricity markets as well as a review of the existing literature that examines FTR auction markets. Section 4 provides a conceptual representation of how decisions made in the ARR process influence equilibrium FTR auction outcomes. Sections 5 and 6 describe the data, empirical approach, and results regarding ARR management strategies in the PJM market. Section 7 concludes.

2. INSTITUTIONAL SETTING

   Competitive wholesale electricity markets are based on a system of locational marginal prices (LMPs). An independent system operator (ISO) collects offers from generators to produce power and bids from LSEs to consume power and then solves an economic dispatch optimization problem to settle the market. The essence of economic dispatch is that it selects the leastcost, or welfare-maximizing, mix of generation resources to meet electricity demand. Coordination of power flows by an ISO to achieve least-cost dispatch guarantees the transmission network is used most efficiently. Efficient use of the transmission network in a competitive setting cannot be achieved without the coordination of an ISO (or similar entity) because electricity travels according to KirchofFs Laws, which makes the enforcement of physical property rights to transmission capacity impractical on an interconnected grid.

   In an LMP system, generation resources are dispatched in merit order in terms of marginal delivery cost, starting with the cheapest units. When a transmission element reaches its rated carrying capacity, the ISO may have to dispatch a generation resource out of merit order to avoid damaging the transmission element. In the economic dispatch optimization problem, this limiting transmission element is called a binding constraint. In the absence of binding transmission constraints, all LMPs (ignoring losses) will be equal to the same price throughout the network, namely the marginal cost of generation. Whenever there is a binding transmission constraint in the economic dispatch problem, LMPs at each node reflect the opportunity cost of scarce transmission capacity in addition to the marginal cost of generation. In general, prices at load nodes increase and prices at generator nodes that contribute to congestion decrease with a binding transmission constraint.

   The nodal price fluctuations faced by market participants due to congestion represent price risk that is ubiquitous in electricity markets. Generators and power utilities often engage in bilateral contracts or purchase futures contracts to mitigate this price risk. However, these contracts are typically settled at a node that is different from the node at which the generator or load settles physical power transactions with the ISO. Market participants must forward contract at nodes different from their own because there are thousands of nodes and forward contracts at each individual node would be too thinly traded. So, after forward contracting for energy, generators and load face locational basis risk that cannot be hedged with bilateral contracts or exchange-traded products. To fill this gap, most ISOs act as counterparty to a hedging product called a Financial Transmission Right (FTR) that can be used as a hedge against locational basis risk.

   2.1 Financial Transmission Rights

   An ISO sells FTRs in periodic auctions up to three years before the FTR begins generating cash flows. Market participants submit offer 'schedules' into the auction to buy (or sell previously acquired) FTRs. A schedule is a series of bids where each bid includes a source node, a sink node, a MW quantity, a reservation price, and potentially other characteristics (e.g., on-peak hours or offpeak hours, a particular month or season, etc.). Which characteristics are available in a given auction vary by ISO and auction type (e.g., long-term, annual, or seasonal auction). There are no restrictions as to which nodes can be source or sink nodes, nor do source or sink nodes need to correspond to where generators or load physically reside on the network.' Most ISOs sell both FTR obligations and options. FTR options are unique because they can never have a negative value. The following description and focus of this paper is on FTR obligations because FTR obligations make up the vast majority of the FTR market and are the relevant product type when discussing ARRs.

   A mathematical programming model whose objective function is to maximize the FTR auction revenue determines auction-clearing prices. To see a mathematical formulation of the FTR auction problem, see Appendix A. The mathematical program that determines cleared transactions in the FTR auction calculates a price for every source/sink combination simultaneously. The auction-clearing price for an FTR is the nodal price difference between the source and the sink determined in the auction:

   [Please download the PDF to view the mathematical expression]. (1)

   where [Please download the PDF to view the mathematical expression] is the nodal price at the sink node in the auction, and [Please download the PDF to view the mathematical expression] is the nodal price at the source node in the auction.

   The payoff to an FTR is determined in the day-ahead energy market over the time period that the FTR covers. The payoff, called the Target Allocation, is defined as the difference between the congestion components of LMP in the day-ahead energy market for every hour the FTR is a valid obligation (as defined by the contract):

   [Please download the PDF to view the mathematical expression] (2)

   where t is the index of hours during which the FTR is a valid obligation as defined by set [Please download the PDF to view the mathematical expression] is the congestion component of the LMP at the sink node in hour t, and [Please download...