Perspectives of the European natural gas markets until 2025.

1. INTRODUCTION: THE EUROPEAN NATURAL GAS MARKET

The European natural gas market lends itself particularly to the EMF 23 study design. It is in the middle of a deep structural change that comprises both, restructuring and vertical unbundling, as well as changing supply relations. Contrary to the reform process in the U.S., restructuring in continental Europe has only started seriously with the second European Gas Directive (2003/55/EC, so-called "Acceleration Directive") whereas the UK had started the reform of its natural gas sector in the early 1990s already. In continental Europe, a small number of players still dominate the national wholesale markets; vertical unbundling is pursued by most member states, though with varying degrees of success. The individual countries are poorly interconnected, and the limited access to pipeline capacity prevents liquid hubs from emerging.

The second aspect, supply structures, also plays an important role in the energy policy debate, and it is the focus of this paper. Europe is a relatively mature pipeline market, with a significant increase in LNG regasification capacity and imports over the last years (IEA, 2004, 2007). In the next decades, the demand for natural gas is generally expected to rise, albeit with some uncertainty on the extent given new developments that may reduce the relative benefit of natural gas in environmental or cost terms (e.g. competition with coal with CCS for power production). In institutional terms, European natural gas supplies are also undergoing the global trend from long-term contracts towards shorter-term trading and a more important role for spot markets. "Energy supply security" is a particularly sensitive issue in European natural gas, especially with a view to the dominant supplier, Russia.

These issues have been studied in the previous literature. Several models have indicated that market power is indeed an issue in the European natural gas market, amongst them Boots et al. (2004), Egging and Gabriel (2006), and Egging et al. (2008). Smeers (2008) summarized and discussed the papers that develop strategic models of European natural gas supply. Hubert and Ikonnikova (2003) and Hubert and Suleymanova (2006) have focused on the specific role of Russia as a supplier to Europe, and the strategic role of transit countries such as Ukraine or Poland. OME (2001, 2005) have provided in-depth numbers of potential prices and quantities of gas supply options for the EU. Stern (2007) provides a balanced discussion of the true "supply security" issues.

In this paper, we report simulation results for European natural gas supplies to 2025, following the EMF 23 study design (EMF, 2007). We apply a strategic model of European gas supply, called GASMOD, that was developed in the early phase of the EMF 23 study, and then slightly adopted to suit the requirements of the EMF 23 study design. The GASMOD model is described in detail in Holz et al. (2008), and therefore will not be presented in detail in this paper. Instead, we focus on the results of GASMOD with regard to the EMF 23 Reference case, and most of the EMF 23 scenarios (see EMF, 2007, p. 30). The next section provides a non-technical model description and discusses data sources and assumptions. Section 3 then summarizes the model results for the EMF Reference case, and five scenarios: higher demand growth, Russian exports constrained, Middle East exports constrained, Middle East & Russian exports constrained, and liquefaction constrained. We put particular emphasis on the future role of Russia, potential alternative supply sources, and model results for the UK market in transition.

In general, our results suggest rather modest changes in the overall supply situation of natural gas to Europe. This also indicates that current worries about energy supply security issues may be overrated:

* LNG will likely increase its share of European natural gas imports in the future, but stay relatively stable beyond 2015;

* Russia will continue to play an important role as a supplier to Europe (~ 1/3 of imports), but it will not play the dominant role that many studies (and politicians) suggest it might play;

* In the time frame of our analysis (2025), the Middle East will continue to be a rather modest supplier, and its exports are more likely to be directed to the Asian and the North American markets;

* the UK is in the process of successfully converting from being a natural gas exporter to become an importer and a transit node for LNG towards continental Europe;

* congested pipeline infrastructure, and in some cases LNG terminals, will remain a feature of the European gas markets, but less than in the current situation;

* the diversification of natural gas supplies, already observed in this decade, should continue and contribute to supply security.

2. THE GASMOD MODEL: MODEL DESCRIPTION AND DATA SPECIFICATION

The model used is a modified version of the static GASMOD model. It corresponds to the description by Holz et al. (2008), except for the regional and technology aggregation (pipeline vs. LNG), the demand function, the time frame and the market power assumptions for certain countries.

GASMOD is a model of the European natural gas trade on a yearly basis. (1) It is programmed in GAMS in the mixed complementarity format and solved using the PATH solver (Ferris and Munson, 2000). We include data for all relevant exporters to Europe, which can supply pipeline gas and/or LNG (Table 1). An exporter can use both technologies simultaneously, but each technology is modeled as a separate player, contrary to Holz et al. (2008) where both technologies were aggregated to one player per country. The importing market in Europe is represented by a disaggregated representation of continental Europe, assuming one wholesale company (marketer) per country that can import from both technologies. Figure 1 shows the structure of the model, exemplified by two exporters (Russia by pipeline and Algeria by LNG) and two European markets (Germany and France), with imports and wholesale trade between each other. European importers are detailed in Table 1 with their import technologies in 2025. In addition, we include the possibility for endogenous domestic production in all European countries. Final consumption is aggregated to total demand of all sectors in each importing country. We model the trade relations in bilateral pairs of exporters-importers, or marketers-final markets, (2) and use aggregated and calibrated capacity bounds for each pair and technology.

GASMOD is a game-theoretic partial equilibrium model of the European natural gas market. Exports to Europe and wholesale trade within Europe are represented as successive markets in a two-stage structure. Market power can be assumed in both market stages, thereby leading to double marginalization of the final customers. We assume market power to be exerted in a Cournot framework. A Cournot market model typically yields higher prices than the perfect competition model (or Bertrand models), thereby giving an incentive to more (higher cost) players to participate in the market. The results of this equilibrium model correspond to the Nash equilibrium of the Cournot game in each market stage. The model results must therefore be interpreted as long-term market equilibrium that does not reflect the short-term adaptation path to the equilibria. Hence, this model type is also not appropriate to simulate short-term market shocks.

[FIGURE 1 OMITTED]

In Holz et al. (2008) we consider three stylized cases of market power in each market stage in order to assess the most realistic scenario for the current European natural gas market: Cournot competition in both market stages, perfect competition in both market stages, and EU liberalization (Cournot competition in export market, and perfect competition in the wholesale market). In line with the market observation we identify the successive Cournot market model as the most realistic representation, but with exceptions for certain countries where the double marginalization structure leads to very high prices and low imports or consumption. Hence, in the GASMOD version used for the EMF simulations, we use a successive Cournot model with a competitive fringe in the export market, and the assumption of perfect competition for certain final markets. On the production market, next to the Cournot players, we assume the small players to be the competitive fringe (Libya, Egypt, Iran, Iraq, Nigeria, Trinidad, and Venezuela, and all domestic European producers except the UK and the Netherlands). On the wholesale market level in Europe, we assume the following markets to be competitive: in the UK, Denmark, Sweden/Finland, Romania/Bulgaria, the Baltic countries, and Turkey. (3)

In this paper, we apply the method of comparative static simulations for the time period 2003--2025. We simulate the years 2003, 2010 and continue in five-year steps up to 2025. For each year, we adapt the data input, namely the reference demand and import volumes and prices, the production and transport capacities and costs. In the absence of founded knowledge about the future market structure, we assume the same market structure prevailing in all model periods.

In particular, as agreed within the EMF group and based on EIA (2005) projections, we assume the reference demand volumes (needed to specify the demand function) to increase by 1.8% p.a. in Western Europe and by 2.2 % p.a. in Eastern Europe. The increase of the reference prices (that are also included in the demand function) is based on projections by the European Commission (European Commission, 2003) with an annual growth rate of 0.8% until 2010, of 2.06% between 2010 and 2020, and 1.25% between 2020 and 2025. (4) The production and transport cost data are based on OME (2001) for 2003 and OME (2005) for all other periods. They mainly include a cost reduction over time of LNG supplies relative to pipeline supplies to Europe.

Print Get the Mag Weekly Updates