North American Natural Gas Markets Under LNG Demand Growth and Infrastructure Restrictions.

• Author: Calci, Baturay

1. INTRODUCTION

   North American natural gas markets have undergone sweeping transformations over the past two decades. The 2000s witnessed the shale gas revolution, where hydraulic fracturing and horizontal drilling led to a decline in extraction costs and a surge in natural gas production in the United States (U.S.) (Huntington, 2016). Consequently, the U.S. became the largest natural gas producer in the world in 2009 (EIA, 2018b), and a net natural gas exporter for the first time in 60 years in 2017 (EIA, 2018a), This increased supply has kept the Henry Hub natural gas price below $5 per million British thermal units (MMBtu) for most of the past decade (EIA, 2019c). As U.S. natural gas production has grown, so too has its capacity to export liquefied natural gas (LNG). In 2007, the U.S. imported 771 million cubic feet (MMcf) of LNG and exported none. One decade later in 2017, U.S. LNG imports dropped to 78 MMcf while LNG exports grew to 707 MMcf, signaling a new era of LNG trade for North America.

   The first LNG export facility in the continental U.S. was the Sabine Pass terminal on the Gulf Coast, whose first phase became operational in 2016. It was followed by Cove Point on the Chesapeake Bay, which commenced commercial operations in 2018. Since then, more LNG export terminals have been approved to begin construction at Corpus Christi, Freeport, Cameron, and Elba Island on the Gulf and Atlantic Coasts. These planned additions would bring the total combined export capacity to 10.79 billion cubic feet per day (Bcf/d) (EIA, 2018a). In Canada, an LNG export facility is currently being constructed in British Columbia, adding Canada to the list of likely LNG exporters in the near future (NEB, 2019a).

   North American LNG infrastructure investment is being fueled by the rapid increase in natural gas demand around the world. Across the Pacific Ocean, Asian LNG imports are rising rapidly. China's LNG imports tripled in just six years from 2010 to 2016. By 2040, China is expected to triple its 2015 natural gas consumption to reach 57 Bcf/d, supported by roughly 11 Bcf/d of LNG imports. This projected growth would put China on par with Japan, currently the world's largest LNG importer (EIA, 2017). However, looking several decades ahead, it is very difficult to project LNG demand from importing regions. Demand will depend on uncertain future developments in their own natural gas production, economic growth energy and environmental policies, geopolitical goals, and trade relations (e.g., the current trade disputes between the U.S. and China).

   These bullish but uncertain projections raise the question of how North American natural gas markets and LNG export infrastructure will be shaped by rising global demand, especially from Asia. All currently operational North American LNG export facilities reside on the Gulf and Atlantic Coasts of the U.S., which increases the time and cost required to transport LNG to Asian markets. A natural solution to secure easier access to these markets would be to build LNG export terminals on the Pacific Coast of North America. However, proposals to construct LNG export facilities along the Pacific Coast of the U.S. and Canada have encountered fierce political and public opposition in line with general resistance to fossil fuel infrastructure development. As an example, activists in Oregon petitioned the State not to issue a proposed LNG project the necessary water quality certification (Brady, 2018), which was denied by Oregon's Department of Environmental Quality in May 2019 (DEQ, 2019). Whether LNG will be exported from the Pacific Coast of North America, and the resulting implications for North American natural gas markets overall, remain to be seen.

   In this article, we investigate how regional natural gas markets in North America will be shaped by the combined effects of LNG demand growth abroad and possible restrictions on LNG infrastructure development in the U.S. and Canada. These two drivers are believed to exert a particularly strong influence on the trajectory of North American natural gas and its relationship with the global LNG market. To conduct this analysis, we construct a nine-region equilibrium model of North American natural gas markets interacting with Atlantic and Pacific LNG demands, formulated as a mixed complementarity problem (MCP). Scenario results highlight how regional natural gas production, consumption, and prices, as well as endogenous investments in production, pipeline, and liquefaction infrastructures, vary with assumptions about future LNG demand and infrastructure restrictions. The underlying analysis presents a diversified view of the evolution of natural gas markets and infrastructure in North America that takes political opposition to fossil fuel development and rising Asian LNG demand into account, and represents our primary contribution to the related literature.

   The remainder of this paper is organized as follows. In Section 2, we briefly review the literature on natural gas market modeling. Section 3 outlines our model and explains how it is parameterized and calibrated. Section 4 delineates the scenarios investigated. We present, compare, and discuss scenario results in Section 5. Section 6 acknowledges limitations that suggest fruitful directions for future research. We conclude in Section 7, which summarizes our most important findings.

2. LITERATURE REVIEW

   This literature review is divided into two parts. In Section 2.1, we present a brief overview of methodologies for natural gas market modeling with a special emphasis on MCP models and how they have evolved over time. Then in Section 2.2, we review previous applications of natural gas market models to different contexts and questions, and then clarify the main novel contributions of our work. The literature on natural gas market modeling is quite rich; therefore, we strive to focus on the previous research that is most relevant to our model and application, and must necessarily omit some earlier work in the interest of brevity.

   2.1 Review of MCPs and Modeling Methodologies

   In this section, we outline a variety of models that use an MCP formulation to provide a realistic representation of natural gas markets and infrastructure. MCP models are widely used in the literature to represent spatially and temporally large natural gas markets with different interacting players, all maximizing their profits subject to various constraints. Karush-Kuhn-Tucker (KKT) optimality conditions of each player's optimization problem are coupled with market clearing conditions between players and solved simultaneously, giving the market equilibrium as an outcome. This is the methodology that we choose to follow, as opposed to techniques like representing the entire system as a large linear program (LP). An LP is easier to scale up and could achieve higher granularity, but it would inherently adopt the perspective of a single decision maker who controls all aspects of the market. The primary focus of our paper is to investigate the effects of various scenarios on North American natural gas markets and infrastructure investments, mediated by the strategic interactions among market players. The MCP approach is ideally suited to determine market equilibria arising from these strategic decisions. Our model also includes nonlinear cost relationships for a more realistic representation of the natural gas industry, and thus could not be modeled as an LP.

   One of the earlier models we review here is GASTALE (Boots et al., 2004), which adopts a complementarity approach and distinctly represents upstream suppliers and downstream traders in the natural gas market as an extension of the successive oligopolist model developed by Green-hut and Ohta (1979). Gabriel et al. (2005a) formulated a complementarity model with producers, storage reservoir operators, peak gas operators, pipeline operators, marketers, and consumers as strategic players, and established conditions for the existence of a solution and for the uniqueness of equilibrium prices. This work was later used by Zhuang and Gabriel (2008) as the baseline market framework for an MCP model with stochastic demand. Gabriel et al. (2005b) also developed a linear complementarity model of the North American gas market encompassing producers, storage and peak gas operators, third-party marketers, and four end-use sectors. Holz et al. (2008) developed GASMOD, a static, two-stage successive game theory model, formulated as an MCP, and investigated the effects of market power on upstream and downstream markets. A previous work in the literature that characterized the LNG players explicitly is by Egging et al. (2008), who developed a European natural gas model in which LNG players such as liquefiers, tanker operators, and regasifiers are explicitly represented as players and thus added to the usual sets of players included in previous models. LNG players such as liquefiers and tanker operators also exist in our model and play central roles in our scenario analysis.

   In earlier natural gas market models, endogenous capacity expansions were often excluded. However, some of these models were later extended to have this capability to better represent the development of natural gas infrastructure. As an example, Lise and Hobbs (2008) extended GASTALE to incorporate endogenous capacity expansion decisions for storage operators, pipeline operators, liquefiers, and regasifiers. Similarly, Holz (2009) extended GASMOD to include this feature as well. Endogenous capacity investments became more standard in relatively newer models. As an example, one work that we followed closely in the literature is the World Gas Model (WGM) (Egging et al., 2010), a multi-period complementarity model that features endogenous capacity investment decisions for pipelines, storage operators, regasifiers, and liquefiers. This model, along with some studies mentioned above, use the logarithmic production...