Evaluating an Interconnection Project: Do Strategic Interactions Matter?

• Author: Debia, Sebastien
• Position: Report - Statistical table

1. INTRODUCTION

   It is widely acknowledged that improving integration between jurisdictions is essential in the development of power markets (IEA, 2016). One of the major rationales is a comparative advantage argument: integration allows to take advantage of the complementarity between two systems. In addition, integration is considered to reduce market power (Borenstein et al., 2000) and to improve the security of supply by increasing the volume of reserve capacity (Stoft, 2002).

   The integration process involves large investments in High Voltage Direct Current (HVDC) interconnections. (1) Contrary to alternate current (AC), flows over a DC line do not cause network externalities from the Kirchoff's laws, but are controllable. Thus, third parties may explicitly (and physically) arbitrage the price differential. Most of these projects ought to be merchant-based, where the initial investors sell Physical Transmission Rights (PTRs) to third parties in charge of the arbitrage. (2) These PTRs are the maximum amount of power that each arbitrageur is able to transfer from one end (the source) of the line to the other (the sink).

   This research studies the impact of increasing integration--by adding a new HVDC line--between New York (NY) and Quebec (QC), when regulation is dichotomous with respect to local trade and exports. Such a framework applies particularly well to power markets, where local activities are thoroughly monitored, but no standard design has been developed to enforce interjurisdictional trade efficiency. This dichotomy may reinforce incentives for strategic cross-border trade, and increasing the interconnection capacity may reallocate sales volume from the (regulated) local market toward the unregulated exports.

   Despite that the New York energy market is not structurally competitive--the day-ahead market accounts for a limited number of players who bid in a multi-unit auction--the market outcome is considered as competitive (Patton et al., 2016). Since the California electricity crisis, research about market design mostly focused on the mitigation of local market power (Borenstein et al., 2002; Joskow, 2008). Mitigation measures is an important ingredient of the FERC Standard Market Design (FERC, 2008), and have been adopted in all the U.S. liberalized electricity markets. In the New York electricity market (NYISO), those mitigation rules prevent suppliers to drift away from their estimated marginal cost. (3) As a result, Patton et al. (2016) states: "The mitigation measures were generally effective in limiting conduct that would raise prices above competitive levels".

   On the other hand, inter-jurisdictional trade is insufficiently monitored, due to the lack of a common regulatory authority (Spees and Pfeifenberger, 2012). (4) The price differential constitutes the core value of HVDC interconnections, and the extent to which they are subject to imperfect arbitrage has been demonstrated empirically (Turvey, 2006; Bunn and Zachmann, 2010; Gebhardt and Hoffier, 2013). But the vast majority of interconnection evaluations assume perfect competition, and thus, perfect arbitrage of price differentials (see Billette de Villemeur and Pineau, 2016; Newberry et al., 2016 for some recent examples).

   A literature about strategic management of interconnections has grown out of Joskow and Tirole (2000), where a local monopolist maximizes a two-stage program where trade is decided before production. (5) Their main result is that an importer with market power has an incentive to withhold its PTRs in order to maintain its mark-up on the local price. Accordingly, any allocation to such a player would create inefficiencies in the absence of use-it-or-lose-it (UIOLI) rules to avoid possible PTR's withholding issues.

   However, releasing rules like UIOLI require a strong inter-jurisdictional regulatory framework to be enforced (de Hautecloque and Rious, 2011), and can be circumvented if their definition lacks commitment devices. (6) Commitment makes PTRs a very rigid instrument under uncertainty, and even more so in a context where two markets have different rules. An initial holder would run the risk of losing potentially valuable rights under this rule. The releasing provision of a PTR may be triggered just before gate closure, with a selling price provision. (7) Such a restricted operational window tends to reduce the effectiveness of the U10LI rule, while a selling price provision--as designed in the NY market--keeps intact the incentive to withhold (Joskow and Tirole, 2000). Finally, a UIOL1 rule would be void if not coupled with a must-offer provision for the line investor and initial holder of the rights. In a context where investment is lacking, such a provision may not be desirable (Brunekreeft and Newberry, 2006).

   PTRs are sold through bilateral transactions or auctions. (8) As for the US, the FERC allows investors in new merchant interconnections to allocate their entire transmission capacity at a negotiated rate through bilateral transactions, in order to induce private investment (FERC, 2013). In the case of bilateral negotiations, Joskow and Tirole (2000) shows that an importing producer buys all the rights to increase his local market power. For auctions, the outcomes depend on the chosen mechanism. In the case of pay-as-bid auctions, there is no pure strategy equilibrium, the number of rights bought by an importer is randomized. However, Gilbert et al. (2004) shows that oligopolistic generators can obtain market-power-enhancing transmission contracts in a discriminatory-price auction, whereas a contestable uniform-price auction with complete information always provides market-power-mitigating contracts. Hence, the allocation mechanism plays a critical role in the players' ability to exercise their market power.

   With bilateral negotiations and absent efficient releasing rules, the transmission rights market is not contestable, and the efficiency results of Gilbert et al. (2004) does not hold anymore. Imperfect information coupled with institutional barriers to entry may typically favor the existence of market power and explain persistent price-spread between electricity forward and spot markets, and thus allocation inefficiency (Borenstein et al., 2008; Ledgerwood and Pfeifenberger, 2013; Ito and Reguant, 2016; Birge et al., 2017). Applied to our problem, these results suggest that an integrated producer-trader is sometimes willing to trade against the price differential in order to raise its local price. The ability of a producer to store its production in a reservoir may reinforce this pattern.

   This paper investigates the short-term impact of different allocations of PTRs on the management of a merchant interconnection project under various regulatory regimes. We base our analysis on a representative HVDC project: the Champlain Hudson Power Express (CHPE). This planned underground interconnection between Quebec and southern New York state is a merchant investment project with a capacity of 1,000 MW, and an estimated cost of US$2.2 billions. This project has the interesting feature of connecting thermal and hydro-dominated systems, two excellent complements from an operational point of view. Also, the seasonality of demand is not the same in the two markets, which increases the trading potential between the two regions. Finally, 75% of the capacity can be sold bilaterally to an anchor customer like Hydro-Quebec (HQ), the state-owned monopolist in Quebec, while the remaining 25% has to be sold through an open season auction (FERC, 2010). (9)

   We model the interconnection between the power systems of the province of Quebec and the state of New York as a strategic game on a three-node network: New York North (TV), South (S) and Quebec (QC). We use an Equilibrium Problem with Equilibrium Constraints (EPEC) approach to model a bilevel trading game and to obtain subgame-perfect Cournot-Nash equilibria. (10) The lower stage corresponds to the maximization of the local welfare under production capacity constraints in each region and results in local market-clearing prices and production for a given level of interjurisdictional trade. The strategic game takes place at the upper stage, where traders of each generation company strategically trade on the interconnection. Their access is limited by their initial allocation of PTRs. (11)

   We consider three possible cases. The first case is a benchmark where cross-border trade is not subject to regulation. Integrated traders within generating companies anticipate the effect of trade on local market-clearing prices and choose their strategies as profit-maximizers of the whole company. It delivers insights about the effects of market power on cross-border trade flows. The second case accounts for an important existing regulation: the "Heritage electricity pool" (Heritage) provision in QC. The third case proposes an alternative regulation: traders and producers are functionally unbundled. (12)

   The contributions of this paper are the following. First, we propose a modeling framework to analyze the impact of PTRs allocation on the trading pattern in controllable-flow HVDC interconnections. Our approach--strategic traders anticipating regulated markets-clearing--focuses on situations where the emphasis is on local regulation while interjurisdictional trade is relatively free. Second, we show that connecting a hydro-dominated system with a thermal-dominated one--two highly complementary systems--does not create value per se. If not monitored correctly, such interconnection may even destroy wealth. This effect is largely due to an over-exporting strategy in order to increase the local price by increasing the total production, and thus the marginal cost. Third, we evaluate the current regulation, Heritage, and propose a better alternative: functional unbundling between operation and trading. To that extent, we show the critical influence of the link between traders and...