A dynamic simulation of market power in the liberalised European natural gas market.

Thus, the perfect competition and Cournot cases bracket the base market power case. Because it is possible that future contracting practices, exercise of oligopsonistic market power, or government action will result in price mark-ups that differ from our base case, the competition and Cournot cases can be viewed as bounds upon future prices. Figures 2 and 3 illustrate this relationship among the cases, showing prices (weighted by quantity demanded in each demand period and each EU country) and total quantity demanded in each year. Prices in all three cases increase with time by similar amounts, and demand growth is strong. The base case results, as predicted, lie between the two extremes of pure price-taking and pure Cournot oligopoly. The changes in quantities imply arc elasticities on the order of -0.4 to -0.7, depending on the year and cases compared.

Figure 4 shows disaggregated results, breaking down EU sales among the twelve producing regions. The base case shows Russia maintaining a one-quarter market share, approximately. A comparison of the competitive and base cases shows that the following producers significantly restrict output to raise prices in the base case: Russia (despite the 75% forward contracting assumption); Algeria; Iran (just in later years, when pipeline infrastructure has expanded to accommodate potential Iranian sales); and the large EU producers (Norway and UKIE), especially in the early years. Note that the largest producers are not the ones who necessarily restrict output; market power outcomes are a complex interplay of production costs, network and storage constraints, and contracting assumptions. For instance, the BENELUX countries maintain their output in all cases, even though their production is similar in magnitude to Norway and UKIE, who do exercise market power. On the other hand, some exporting regions with relatively small market shares take advantage of the higher base case prices to increase output and, of course, profits, notably Egypt and Nigeria.

Turning to a comparison of competition with unrestrained Cournot competition, a Russia that is free of long-term contracts reduces its output by as much as 40% in order to increase prices by over 30%. UKIE also significantly reduces production, particularly in the early years. All other producing regions together produce about the same amount in the perfect competition and base cases, confirming that Russia is the most important strategic player. In contrast, Iran, Nigeria, and Egypt take advantage of the higher Cournot prices to expand sales to the EU.

We now consider the distribution of price increases due to market power which, due to network constraints and relative proximity of different producers, is by no means uniform. Table 2 shows prices in the low demand seasons for each region in 2010 and 2030 for the competitive case along with the percent price increases for the other two scenarios. Interestingly, prices under low demand are most affected by market power in percent terms, but have the lowest prices. Comparing first the base and competitive cases, we see that mainly eastern European countries suffer from the exercise of market power by Russia, with the Balkans, Central, and Turkish regions (Table 1) suffering price increases between 27% to 36% in 2010, while western and southern European countries experience price increases between 15% to 19% in 2030. UKIE is an exception because of the market power that its producers exercise--its price increases are similar to eastern Europe's. If we now consider the relaxation of contract constraints that limit market power (especially of Russia) in the base case, Cournot competition causes prices to go up by an additional 15% or more in Eastern Europe, but only about 10% elsewhere. The net result is that, compared to perfect competition, unrestrained Cournot oligopoly increases prices by over 50% in eastern Europe, but by only about 20% to 30% elsewhere because of the access other countries have to alternative suppliers. This shows that market accessibility is an important determinant of prices.

[FIGURE 4 OMITTED]

Table 3 shows the storage and transport investments made by GASTALE in each year within each of the three competition cases, while Table 4 shows aggregate capacities and utilization. In general, total investment in storage and LNG facilities over the time horizon does not change between the competitive and base cases, although there are shifts in geographical distribution and specific timing of investments. For instance, under the base case, there is more storage in the Balkan-Turkey and Baltic-Central regions. This storage serves to arbitrage peak-offpeak price differences in eastern Europe that might otherwise be exacerbated by Russian market power. Meanwhile, the base case shifts LNG investment to UKIE in order to moderate the price increases that would otherwise occur due to the restriction of supply from UKIE producers. UKIE is particular sensitive to seasonal swings, mainly due to the recent drop in production from their gas fields in the North Sea. UKIE has too few existing pipeline connections to compensate for the seasonal swing in demand, and the quickest and economically most attractive way to fill this gap is to invest in storage.

On the other hand, the base case sees approximately 50% more pipeline investment on average than competition, because the exercise of market power in eastern Europe (and to a lesser extent in the UK) results in sharpened price differences between western/southern Europe and the rest of Europe. Pipelines are then built in an attempt to arbitrage those price differences, with pipeline investment within the EU increasing by over 250% over the time horizon. Meanwhile, investments in pipelines that import gas from Norway, Russia, Asia, and Africa decrease somewhat compared to the competitive case, corresponding to lower imports. Aggregate pipeline use (measured as the sum of the bcm/yr flows across all pipelines in the model) is roughly 50% higher in both market power cases than under competition, because of heightened geographical price differences (Table 4).

Contrasting the extreme (Cournot) market power case with the competitive case shows similar patterns for pipeline and LNG investment as in the base case-competition comparison. However, storage investment increases by about 30% under Cournot competition, with additional investments occurring in all four aggregate regions shown in Table 3. Intuitively, one might have expected less investment, as overall gas volumes are smaller in the Cournot case. But because oligopolistic competition can yield greater price mark-ups under supply constrained (peak) conditions, as fewer suppliers have excess capacity, the attractiveness of storage is increased. This is reflected in the higher amount of storage in the Cournot case in Table 4. As a result, prices increase less than they otherwise would during the high demand season, and prices during the low demand season are higher.

The final set of results (last rows of Table 4) concern 'flexibility' or what might be called 'swing supply'--the differences between the amount of gas provided to consumers in the medium/high and low demand seasons. Either increased production and imports or increased use of storage can provide that flexibility. The table shows that the swing supply of storage is much higher in 2030 than in 2010, while pipeline swing supplies from other sources actually fall, indicating that the pipelines are being used closer to capacity in all seasons by 2030. Comparing the competitive, base, and Cournot cases shows that the main effect of large producer market power is to decrease swing supply from Russia, which is consistent with the greater use of storage in eastern Europe that we just noted above. Under market power, LNG swing supplies also fall in importance relative to the competitive case; this is due to increased use of storage (in the Cournot case) and more intensive use of LNG import facilities throughout the year. The effect on imports via the Norwegian and other pipelines is ambiguous, with a large increase in the base case but lesser under Cournot Competition.

5. CONCLUSIONS

Equilibrium models can show what degree of market power might possibly be exercised under alternative assumptions concerning market player behavior and market structure. We have provided illustrative scenarios for the EU market which indicate that Russia has the highest incentives to exercise market power, and that such exercise will tend to increase price differences over space within the EU, which will encourage within-EU pipeline construction, and over time, will encourage storage construction. Differences in prices are due to the interaction of the (1) inherent ability of producers to exercise market power (determined by production capacity and costs) with the (2) accessibility of the market (determined by gas transport infrastructure).

Although the dynamic GASTALE model is an improvement over the previous static version because of its endogenous increase in transport and storage capacity, a number of modelling improvements are desirable in it and other gas market models. These include endogenous investment in production and representation of intertemporal production tradeoffs (Zwart and Mulder, 2006); more realistic representation of transport and storage technology (Midthun, 2007); stochastic models that represent how uncertainty and risk aversion affects market equilibria (Gurkan et al., 1999; Zhuang and Gabriel, 2008); and market power on the part of other players, including storage, traders, and pipelines. Careful validation against historical data is also desirable. However, because of structural and parameter uncertainty that are inherent to equilibrium models, their use must necessarily be for exploration of scenarios and implications of assumptions rather than point projections.