Locational Investment Signals: How to Steer the Siting of New Generation Capacity in Power Systems?

• Author: Eicke, Anselm
• Position: Report

1. INTRODUCTION

   The construction of network infrastructure is costly, subject to lengthy permitting processes, and often met by public resistance. However, the integration of newly constructed power generation facilities results in a significant rise in transmission infrastructure needs in many parts of the world. There are three major reasons for this increase. A first is the expansion of wind and solar energy. These energy sources are produced at least cost where land is cheap and resource availability is high, which is often far away from energy consumption centers. (1) Second, the legacy of regional energy monopolies is fading 20 to 30 years after restructuring the electric power industry in many parts of the United States and Europe. Historically, these utilities were the primary investors in local generation capacity. Last, European power markets are increasingly closely integrated, which has resulted in a rising long-distance trade of power, and by association, in load flows.

   As a consequence of the increasing distance between load and generation, strain on transmission infrastructure is increasing and generators in many regions have experienced significant curtailment over the last years (Bird et al., 2016). One way to reduce the pressure on the infrastructure is to site generation assets and end users closer to each other. Such locational steering can be applied to generators and to consumers. The primary purpose of this paper is to providing a comparative review of locational investment signals applicable to generators. Politically, it is much easier to introduce cost (or revenue) differences for generation compared to discriminating against consumers.

   In Europe, zonal electricity markets prevail. A textbook zonal electricity market does not provide any locational incentive within a zone, which implies that the choice of where to invest is purely driven by the costs of power generation. In the case of wind and solar energy, these are cost of land and resource quality. Cost differences between coal-fired power plants mostly stem from coal transportation costs.

   However, the social cost of power supply not only comprises generation costs but also includes the costs of the accompanying network. (2) The latter includes infrastructure costs and costs associated with congestion and transmission losses. Socializing these costs implies they are not accounted for in investment decisions. This is a classical externality problem (Alayo, Rider, and Contreras, 2017). It leads to a cost-inefficient distribution of generation investment, with generation capacity being installed too far from consumers and grid investments being too large.

   The socially optimal location of power generation, in terms of (social) cost-efficiency, depends on the difference in generation costs between locations and on network costs. If generation costs were unrelated to network costs, the (trivial) cost-optimal location of generation would be at sites where generation costs are lowest. Yet, the cost of power generation is in practice often low at sites which strain the network, e.g. wind generation at windy but remote sites. In these cases, social costs are minimized when the marginal cost of relocating generation equals the marginal cost saving in the transmission system (Figure 1). In other words, it is cost-efficient to relocate generators if the resulting savings in network costs exceed the additional costs of generation at the new site. Well-designed locational signals therefore compel generators to consider transmission- and congestion-related costs in their siting decision.

   Unlike textbook zonal markets, most real-world power systems provide locational signals to generators. Such signals often stem from regulations outside the market, including grid connection charges, grid usage charges, capacity mechanisms, and renewable energy support schemes. Some power markets have also introduced spatial granularity into the market itself in the form of smaller zones or locational pricing. We use locational instruments as an umbrella term for this variety of regulation. The aim of this paper is to discuss locational instruments from a theoretical perspective and review their use empirically.

   A vast amount of published literature covers many of these instruments--Google Scholar reports 4,100 papers related to "locational marginal pricing" alone. However, this body of research often does not consider the instruments as locational investment incentives. Locational pricing, for example, is generally viewed as a dispatch incentive (Joskow, 2008), and grid charges are widely considered to be a cost recovery mechanism (Olmos and Perez-Arriaga, 2009).

   Only a few authors compare locational investment incentives across instruments. Hadush et al. (2011) examine market splitting, loss factors, grid usage charges and grid connection charges associated with European case studies. The authors assess each instrument's effect on investment decisions based on the criteria stability, predictability and strength. Brunekreeft et al. (2005) propose that additional locational instruments complement locational marginal pricing to signal the cost-efficient location of generation investments. The authors base their argument on the observation that locational marginal prices do not recover all grid costs, and therefore do not fully internalize the actual locational value differences of generation. They discuss grid usage charges and deep grid connection charges as supplementary locational instruments. Keller and Wild (2004) assess how coordination between transmission and generation investment can take place in liberalized power markets. To do so, the authors examine locational investment signals arising from transmission pricing. Nikogosian et al. (2019) analyze grid connection charges, regional quotas, and regional premiums with respect to steering the siting of renewable energy assets in Germany. Their study concludes that among these three, regional quotas are the most effective and easiest to implement in the context of the German energy market. Locational investment signals are considered as a means of reducing grid congestion by Hirth et al. (2018). In their categorization, the authors cluster instruments in a manner similar to that which was employed in this study. To the best of our knowledge, a review of the different classes of instruments has not been conducted so far.

   The objective of this paper is to close that gap in the literature by providing a comparative review of locational investment signals applicable to generators. More specifically, our contribution to the literature is three-fold. First, we propose ten dimensions to characterize locational instruments. Secondly, we review the locational instruments used in twelve power systems and finally, we introduce a simple methodology to quantify the strength of these instruments and employ it.

   We find that every power system employs at least one instrument, and most systems use multiple locational instruments in parallel. In practice, most of the analyzed locational electricity markets apply regulation on top of a granular market to steer the location of investments. We further observe that instruments differ significantly in design, and there does not appear to be a "silver bullet" instrument, which represents the best option for all systems. The effect of many of the locational instruments on investment decisions is reduced due to lack of predictability, low levels of transparency, and insufficient spatial and temporal accuracy.

   The remainder of this paper is structured along our three contributions: section 2 presents our analytical framework, section 3 identifies which instruments are used where, and section 4 quantifies their impact.

2. TEN RELEVANT CHARACTERISTICS OF INSTRUMENTS

   The effect of locational instruments on investment decision-making depends on their design. This section proposes ten distinct design characteristics that influence efficacy and nature of the locational signals that such instruments provide. In the following section, we apply these characteristics to structure the review of locational instruments in our sample and discuss their implications.

1. Price or quantity. Locational instruments can be designed as price or quantity instrument. For price-based instruments, the difference in the cost or revenue between locations is determined by the regulator, for example grid usage charges that are differentiated by location. By contrast, quantity instruments are characterized by upper or lower capacity thresholds in a region. Regional quantity limits for renewable energy deployment are an example of this. In an efficient market, such a quantity constraint translates into a "virtual" price signal (Neary and Roberts, 1980). Price-based instruments benefit investors by making it easier to ascertain value differences between locations. Integration costs that result from connections at certain locations can be transferred directly to project developers. By contrast, quantity-based instruments are valuable because they provide a simple and effective way to steer generation investment and account for quantity constraints (e.g., limited transmission capacity is easy to account for). Most instruments identified in this study are price-based.

2. Per energy or per capacity. Locational signals remunerate or charge generators based on the total amount of energy produced (MWh) or the generation capacity installed (MW). Depending on the design of these instruments, technologies are affected differently. Capacity-based signals have a stronger impact on technologies with a low capacity factor, such as peaking plants or renewable energy sources, while energy-based instruments have a more significant effect on generators with high capacity factors such as base load plants. To support this point, compare two hypothetical generators. Peak generator P with a capacity of 2 MW is operated 1000...