Price Adjustments and Transaction Costs in the European Natural Gas Market.

• Author: Garaffa, Rafael
• Position: Report - Statistical table

1. INTRODUCTION

   A unified and liberalized natural gas market in continental Europe is still a distant prospect that requires the development of regional markets. Notwithstanding the notable progress, increasing the number of physical transactions between markets is still required to reduce bottlenecks in transport networks and interconnection points. Under the regulatory reforms promoted by the European Commission, hubs with high liquidity play an important role on market integration, reducing price uncertainty and the transaction costs (1) of natural gas trade.

   Moreover, hold-up problems (2) may arise in the absence of long-term, price-indexed contracts. Hold-up and limited liquidity issues may introduce price uncertainty in natural gas markets. Thus, the European transition from long-term contracts to hub-based spot markets raises some concerns regarding the presence of transaction costs.

   The transaction costs encompass many more aspects than just the (natural gas) transport expenditure. They can be any costs with respect to temporal and financial expenses connected to the search for information, financing the trading process and legal duties (Shepherd, 1997). A possible way to evaluate the role played by transaction costs in market integration is through the degree of (non-)price convergence between hub markets.

   This paper tests the hypothesis that asymmetric price responses in the continental European hubs are derived from transaction costs. Asymmetric price responses might reflect different transaction costs depending on what is the trade direction and are particularly interesting to observe because of the increasing role of non-operational agents (traders etc.) in the European hubs. To do so, we assess the degree of integration between the German, the Belgium and the Dutch natural gas spot markets and estimate nonlinear error correction models that allow for short-term regime specific price dynamics. Through this methodology we can identify price asymmetries and transaction costs, possibly offering arbitrage opportunities. To our knowledge, this methodology was mostly applied to agricultural commodities' markets, with few applications to energy markets. Therefore, by applying this methodology to the European gas market we argue that this can be used in energy markets to evaluate transaction costs and asymmetric price responses, providing case studies for other markets that plan to go through liberalization processes, like the Brazilian market (Mathias and Szklo, 2007) or the Chinese market (Ishwaran et al., 2017).

   Besides this introduction, section 2 summarizes the evolution of natural gas markets in Europe and presents different views of market integration found in the literature. Section 3 describes the data selection and section 4 the methodological procedure adopted. The results of the tests and the models employed in the analysis are discussed in section 5, with special attention to the cointegration relations, the price regimes and the asymmetric price responses between the German, the Belgian and the Dutch markets. Section 6 brings final remarks.

2. INTEGRATION OF THE NATURAL GAS MARKETS IN EUROPE

   During the 1960's and 1970's natural gas became an important source in the European energy mix. Throughout those decades, the European natural gas industry grew vertically integrated based on natural monopolies of national companies. Despite the pressures for introducing more competition in the continental European natural gas industry during the second half of the 1980's and the 1990's (3), the market liberalization process started only in 1998, with Directive 98/30 of the European Commission (EC). Bourbonnais and Geoffron (2007) showed that, until the mid-2000's, the integration of the European natural gas markets was a sparse phenomenon. Throughout the 2000's, under new regulatory reforms, spot markets based on hubs emerged, challenging the long-term contracts structure based on take-or-pay clauses.

   High investments in transport infrastructure are required for expanding supply in the natural gas industry, so that the economic feasibility of these investments strongly depends on the pricing structure and predictability (Komlev, 2013). Over the last years, some companies have introduced flexibility clauses and started trading in the spot market, but the substantial part of their contracts is still long-term based (Statoil ASA, 2015). Meanwhile, the EC kept pushing on regulatory reforms to promote a single integrated market for natural gas. In this sense, the benefits associated with reduced market power of the national (and vertically integrated) oil and gas companies may be offset by price uncertainties associated with the liberalization process, thus reducing incentives for investments in infrastructure expansion.

   Several empirical studies rely on the Law of One Price (LOP) (4) to evaluate the effectiveness of natural gas market integration (Asche et al., 2002; Siliverstovs et al., 2005; Bourbonnais and Geoffron; 2007; Robinson, 2007; Brown e Yucel, 2009; Renou-Maissant, 2012; Asche et al., 2013; Growitsch et al., 2013). Firstly, most of these studies do not consider a complete interpretation of what market integration stands for. They usually focus on price series and do not account for physical factors related to market integration in their analysis--flow constraints, for example. Secondly, in practice, market failures and economic inefficiency are pervasive and perfectly competitive markets are hard to find. So, how to assess price transmission asymmetries within an integrated market framework where the LOP prevails?

   Regarding the first aspect, different definitions for market integration can be found in the economic literature. Market integration can be a measure of the expected rate of price transmission (Fackler and Goodwin, 2001) or related to price transmission efficiency and equilibrium in the presence of trade flows (Barrett, 2001). Another view comes from Gonzalez-Rivera and Helfand (2001) that consider both geographic elements--such as space--and economic elements, defining integrated markets as the set of places that share the same commerce and the same long-term information. Moreover, in Ihle (2012) a fourth different interpretation emerges, defining the integration of markets as a relationship characterized by long-term dynamics. In this study, we share Ihle's (2012) view, thus referring to market integration as a dichotomous variable for which there is (or there is not) a physical and informational long-term trade flow. This definition is particularly interesting especially because for Ihle (2012) the informational flow (i.e., the price transmission) incorporates the dynamics of the error correction models (as further described in section 4).

   As for the second aspect, economic inefficiencies may hinder price convergence, leading to a departure from the LOP. Several studies in the literature are based on the idea that arbitrage is the force which eliminates deviations from the LOP, occurring when the price differential is sufficiently greater than the transport cost of the goods (Lo and Zivot, 2001). This could be the case for the natural gas trade in continental Europe hubs: if transaction costs play a significant role, arbitrage opportunities might occur (Nick 2016), leading to a milder form of the LOP. The asset-specific investments on pipelines, valves, compressor stations, underground storage facilities and city gates leads to uncertainty, opportunism strategies and hold-up problems, increasing these transaction costs (Joskow, 1988; Von Hirschhausen and Neumann, 2008). Hubbard and Weiner (1991), for instance, identified that price negotiations in the United States (US) natural gas industry were often driven by market power and/or transaction costs. In the natural gas industry, transaction costs are commonly associated with its transport infrastructure. Long-term contracts, take-or-pay provisions and ex-ante prices or price adjustments provide the mechanisms to reduce contractual conflicts (Crocker and Masten, 1991) and, consequently, to avoid informational, administrative, bureaucratic, bargaining, among other transaction costs incurred by the parties in the natural gas trade.

   Suffice it to say that the concept of transaction costs is at the heart of such inefficiencies. Nick (2016) remarks that short-run price deviations from equilibrium are persistent in the European market and assesses intertemporal arbitrage through a non-linear approach--a feature that should be taken more heavily into account in empirical studies. However, although mentioned (5), the role of transaction costs in the equilibrium is left behind and not fully explored.

   In this sense, this study contributes to the literature in different aspects. First, methodologically, when it explicitly models transaction costs in a market integration perspective. Second, it explores price equilibrium through a nonlinear modeling approach, exploring asymmetries which are not perceived by linear models. Finally, the methodological framework relies on a conceptualization of market integration that is mostly applied in agricultural economics and could be explored by energy economic studies.

3. DATA SELECTION

   To evaluate the liberalization policy pursued by the EC we focus on the hub based continental Europe spot market, therefore not including the British NBP hub, yet a fully liberalized market. Following the interpretation of market integration, we assessed the existence of physical trade in both directions (i.e., entry and exit) in continental Europe. We chose the period between 2013 and 2014 because of data availability and due to the new regulatory guidelines for trans-European energy infrastructure (6). The hubs that best meet these criteria in this period are the Dutch TTF, the Belgian ZTP and the German NCG (ACER, 2014; ENTSOG, 2014). Figure 1 shows (in dark circles) their location and the available natural...