Effects of Liberalizing the Natural Gas Markets in Western Europe* Rolf Golombek**, Eystein Gjelsvik and Knut Einar Rosendahl***

This paper uses a numerical model to examine the long-run impact of a radical liberalization of the West-European natural gas markets. We study profit maximizing Cournot producers facing an ideal third party access regime .for gas transport. Producers sell gas either to large users in the manufacturing industry and to gas-fired thermal power plants, or to local distribution companies. We first examine the case where no traders exploit arbitrage possibilities and some producers have limited access to the markets. In this equilibrium net prices differ across markets. These differences disappear in the second case where traders are introduced. 7he third case focuses on a complete European market for natural gas in which traders exploit all arbitrage possibilities and all producers can sell gas in all markets. We also study the impact on the complete European market of changes in costs for production, transport, and distribution. Finally, welfare implications from a liberalization of the West-European natural gas markets are discussed. We argue that a radical liberalization could increase economic welfare in Western Europe by 15 % to 20% in the long run.

INTRODUCTION The natural gas industry in Western Europe is subject to a number of government regulations and controls. As of 1994, in Italy one subsidiary company of the state-owned corporation ENI is responsible for most of the purchase, transport and sale of natural gas to the distributors, whereas another subsidiary of EN1 has a monopoly on domestic gas production (Cameron, 1990). The Energy Journal, Vol. 16, No. 1. Copyright

*

** ***

lg 1995 by the IAEE. All rights reserved.

We are indebted to Jan Bdten for his research assistance and to Morten Berg and Michael Hoe1 for comments and suggestions. Earlier versions of this paper were presented at SNFOslo, Statistics Norway, the 17th annual conference of IAEE and the 16th North American conference of IAEE. We thank the participants for their comments. Research support of the Research Council of Norway under the program PETRO is gratefully acknowledged. Foundation for Research in Economics and Business Administration (SNF). Oslo, Norway. Statistics Norway, P.O. Box 8131-Dep, 0033 Oslo 1, Norway.

85

86 / T%e Energy Journal

In France, the state owned Gaz de France still has a legal import monopoly and virtually a monopoly over transport. In the Netherlands, the state owns 50% of Gasunie, the sole company which is involved in gas trading and transport. On the other hand, in past years there have been some tendencies to stricter gas competition in Western Europe. In 1989, Wintershall, a former small German gas producer, announced that it planned to build a transmission line from Emden (the landfall of Norwegian gas supplies to Continental Europe) to Ludwigshafen, the headquarters of Wintershall’s parent company BASF (Stem, 1992). One year later, it became evident that Wintershall was a major player in German gas transportation when the company announced an agreement with Gazprom (the organization which controls gas exports from the former USSR) to market Soviet gas in the eastern part of Germany. In the UK, a liberalization process started in 1986 when British Gas Corporation was transferred to the private sector intact as a 100% monopoly supplier. Two years later, the Monopolies and Mergers Commission reported that competition had been slow to emerge. The Office of Fair Trading drew the same conclusion in 1991. The lesson from the UK is clearly that it is necessary with a radical new regulation (of tariff principles, price transparency etc.) to achieve a more competitive market for natural gas (Price, 1994). Also the Commission of the European Union (EU) has made efforts to liberalize the natural gas industry. This became apparent in 1988 when the Commission published a working document on the International Energy Market, cf. Commission of the European Communities (1988). The document, which encompassed a number of new initiatives like harmonization of taxation, price transparency and interconnection of grids, made it clear that the position of the Commission from the mid 1980s to largely exclude the energy sector from the Single European Market, had been changed. The proposals from the EU met opposition from several Member States as well as from part of the gas industry. Four years later the Commission issued a Draft Directive which introduced a strategy to liberalize the gas market in three stages, cf. Commission of the European Communities (1992). The first stage took place in 1990 and 1991 when the Directive on price transparency and the Directive on transit of gas between high pressure transmission grids were passed. In stage 2, originally intended to enter into force on January 1 1993, enhanced competition should be accomplished by a new way of organizing natural gas transportation. The basic idea was that pipeline companies agree to carry gas-which is owned by another agent-in return for payment (to the extent that there is idle capacity available). Other principal elements were a transparent and non-discriminatory licensing system and separation within vertically integrated undertakings ofmanagement andaccounting (“unbundling”). Finally, on the basis of an evaluation of stage 2 a further liberalization was

The Effects of Liberalizing

/ 87

planned (stage 3). The idea was to “complete the internal market for gas and electricity. “I The importance of the passed Directives on price transparency and transit has so far been limited. In December 1992, the Council of Energy Ministers asked the Commission to revise the Draft Directive of 1992, stressing the importance of the principle of subsidiarity. According to Hopper (1994), early in 1994 a new version had been worked out by the Commission. The proposal has weaker impositions on Member States’ transmission and distribution companies than the Draft Directive of 1992. On the other hand, observers do not rule out that over the next lo-20 years the general process in the EU of removing barriers to trade could lead to significant institutional changes in the natural gas markets. Hence, it is of interest to examine how new measures to promote competition may affect the natural gas markets in Western Europe. This paper gives rough estimates of the long-run impact on prices, quantities and welfare that could follow from a radical liberalization of the natural gas markets in Western Europe. In analyzing the future markets for natural gas in Western Europe, one should be aware that it is virtually impossible to predict the final outcome of the liberalization process. On the other hand, data availability strongly limits the set of feasible numerical models of the natural gas industry. In this study we focus on a feasible case that reflects theprincipal elements of the 1992 Draft Directive of the EU Commission, namely the corner case of a non-discriminatory regulation in which all economic agents are free to rent gas transport services and companies in the gas industry are forced to compete.* One main difference between the present European gas market and the Draft Directive concerns natural gas transport. At present, there are long-term (“take-or-pay”) contracts between the producer and a pipeline company. The pipeline company buys natural gas from the producer and resells the gas to local distribution companies (LDCs) and large end users. In several countries, one pipeline company (virtually) controls the outlet to all end users in the country. In this study we focus, however, on a regime in which pipeline companies only carry gas which is owned by another agent (“a third party”, e.g. a large end user or a producer) in return for money (the tariff). Such an institutional

1. The Draft Directive also emphasized the principle of subsidiarity: “The Commission should not impose rigid mechanisms, but rather should define a framework enabling Member States to opt for me system best suited to their natural resources, the state of their industry and their energy policies. ” 2. In a companion paper we study an even more radical liberalization of the European natural gas markets where also gas sales consortia in Western Europe are assumed banned, cf. Section 5 in Golombek, Gjelsvik and Rosendahl (1994) for preliminary results and Gjelsvik, Golombek and Rosendahl (1995) for a broader discussion.

88 / The Energy Journal

arrangement is termed third party access (TPA) to gas transport. A key question is then: how are tariffs determined? The EU proposal for an internal natural gas market suggests that transport tariffs for transmission companies (high pressure pipelines) should be “reasonably related to the long-term costs incurred in the provision of the relevant service, together with a reasonable rate of return on capital employed in the provision of that service”, Article 12.6. A similar principle should be used for transportation of gas in local and regional pipeline systems (Article 19.6). This means that a pipeline company and a (large) customer will bargain over the tariff, given that the tariff should not give the pipeline (slightly) more than a normal rate of return. From a theoretical point of view, the outcome of tariff negotiations are not clear because theory does not offer many guidelines for the equilibrium of bilateral bargaining games with several buyers and sellers. Moreover, although the tariff question has been addressed in the applied literature, see e.g. Stem (1992), a unique workable principle has not been reached. In this study we assume that tariffs are set by a benevolent regulator so that (in equilibrium) a pipeline company obtains a normal rate of return per dollar invested. As we are forced (due to limited data availability) to assume that for each pipeline company total cost of transporting one unit of gas one km is constant (although unit. costs differ across pipeline companies), tariffs are simply set equal to total long-run unit costs. 3 The rest of the paper is organized as follows. In the second section we describe how we model the natural gas industry after a radical liberalization has been implemented and explain how the parameters of the model are determined. In the model, natural gas is consumed in six major Western European countries. In each of these countries there are two end-user markets, one for small consumers and one for large users of gas. We distinguish between two groups of natural gas producers according to level of production and degree of export. The major producers are modelled as Coumot players. In the model there are ideal TPA systems for transportation, storage and load balancing that allow equal opportunity to all players, allow no form of discrimination and are entirely transparent. All services are offered at fixed rates which reflect total long-run unit costs, i.e. the owners of the facilities obtain a normal rate of return. Producers of gas, which buy transport services and services necessary to meet fluctuations in demand, sell gas to large end-users and to local distribution companies (which resell the gas to small end users).

3. Our analysis shows the outcome from a radical liberalization, given that the regulator knows the true unit costs. It is beyond the scope of this numerical paper to examine the outcome when the regulator has incomplete knowledge of unit costs. For a discussion about regulation under asymmetric information, see e.g. Laffont and Tirole (1993) and Chermak and Patrick (1994).

The Effects of Liberalizing

/ 89

Note that it may take several years to liberalize the natural gas markets. This is because natural gas is currently sold under long-term contracts, and a transition period is necessary where existing contracts are phased out and new contracts are phased in. Hence, we study a hypothetical long-run equilibrium. To be more specific, the model simulates a static long-run equilibrium in which costs reflect a “typical” year between 1990 and 2010, e.g. the year 2000. Moreover, because we focus on long-run aspects, it is legitimate to neglect bottle-necks in the transport grids: Assume that demand for transport (in the short-run) is higher than capacity. Because a pipeline obtains the fixed transport tariff, we assume in line with standard theory that the pipeline company expands capacity until it equals demand. Needless to say, a pipeline company will not expand capacity beyond demand as the company will not receive any income from its idle capacity.4 The third section examines three stylized scenarios. We first examine the case where gas users in the UK are served by domestic production and Norwegian exports (the present situation), whereas consumers in other countries could be served by a number of Coumot producers. With profit maximizing Coumot producers, net prices (end-user prices less of costs of storage, load balancing and distribution) differ both across end-user markets in a country and across countries. On the other hand, in the next scenario net prices are equal because all arbitrage possibilities are assumed exploited by traders. Finally, in the third case (the complete market scenario) traders exploit all arbitrage possibilities and all major gas producers are in a position to sell gas in all markets. In the complete market scenario, production is around 20 % higher than in the base year 1990. The average price for large gas consumers in Western Europe is (almost) unchanged, whereas the average price for small consumers has decreased by approximately one third. The fourth section examines the impact on the complete market scenario of changes in costs of production, costs of transport and costs of distribution. In particular, we focus on how the equilibrium changes when transport tariffs in Western Europe are changed: what is the impact on total production, market shares and profits of a drop in the transport tariffs by e.g. 25%? Finally, the fifth section provides a welfare assessment of measures examined in this study. A radical liberalization leads to increased consumer surplus but lower total profits to producers and transport companies in Western Europe. By adding

4. If capacity is lower than demand in the short run, either capacity has to be rationed or a market clearing tariff has to be used (assuming that the difference between the market clearing price and the fixed tariff is collected by the regulator). Because a pipeline makes zero profit under our regulation (when capacity is less than or equal to demand), to give the pipeline an incentive to increase capacity the regulation scheme should actually be slightly modified; tariffs could e.g. be marginally higher than total long-run unit costs.

90 / The Energy Journal

consumer surplus and profits we find that economic welfare in Western Europe may increase by 15 % to 20% in the long run. DATA AND CALIBRATION Demand At present, natural gas is consumed in almost all Western European countries. Consumption varies, however, substantially: while there is no consumption of natural gas in Norwegian households, 96% of Dutch households use gas heating, cf. Intergas Marketing (1993). To simplify, in this study we focus on natural gas consumption in six major countries; Belgium, France, (West) Germany, Italy, the Netherlands and the United Kingdom. In 1990 (the base year of the model), natural gas consumption in these countries amounted to 75% of total gas consumption in Western Europe, cf. BP Review of World Gas. In each country we distinguish between small and large users of natural gas: a small user is an agent in the residential, commercial or public sector, whereas large users are found in the manufacturing industry and in electric power production (gas-fired thermal power plants). Henceforth, we term the first group the household market, and the second group the industry market. In Western Europe, natural gas competes mainly with light fuel oil in the household markets and heavy fuel oil in the industry markets. According to several observers, cf. e.g. Vrieling, Munksgaard and Hopper (1989), as a rule of thumb gas prices are set slightly below the prices of the competing energy fuels. If this is the case, a marginal increase in the price of gas should trigger a substantial drop in consumption, i.e. the demand curve is very elastic above the observed price. This could be modelled by a kinked demand curve. In this study we have, however, used linear demand curves. Because our equilibrium prices in the household markets are below the observed 1990 prices, and the equilibrium prices in the industry markets are mainly around the observed 1990 prices, the kink in the demand curve is of limited concern in this study. The chosen energy elasticities (at the observed consumption p&rs) build on Birkelund, Fuglestvedt, Gjelsvik and Aaserud (1991) who identititi long-run elasticities by simulating price increases in a disaggregated econom&ti& model for energy demand in Western Europe. In general, our price elasticities for the industry markets are lower and show less variation across countries than the elasticities in the household markets, cf. Table 1. The total direct price elasticity for natural gas (aggregated over all twelve markets) is -0.93. This is mainly in line with several other studies, see e.g. Al-Sahlawi (1989) for a survey on price elasticities for natural gas. Finally, to make quantities in our scenarios comparable with the 1990 outcome, demand functions are calibrated for 1990.

The Effects of Liberalizing

Table 1. Key Parameters Demand Price elasticities

country

Belgium

France

Netherlands

Italy

UK

West Germany

household

-1.16

-1.53

-0.5

-1.2

-1.18

-1.2

industrv

-0.75

-0.88

-0.66

-0.72

-0.73

-0.70

sector

costs Costs are measured

in 1990 USD

Costs of production Marginal cost (in USD per toe) is given by a + bq + c. In (l-q/Q), where q is production and Q is capacity

Q

a

b

Algeria

11

0

0

CIS

12

0

-22.4

loo 70

Parameter

C

Country

Netherlands

3

0

-12

Norway

38

0.75

-10

80

UK

60

1.38

-10

47

Costs of international transportation Western

Europe Onshore: Offshore: - to Germany - to UK

2.49 USD per 100 km per toe 3.74 USD per 100 km per toe 7.47 USD per 100 km per toe

CIS Onshore:

0.62 USD per 100 km per toe

Onshore: LNG: - to Belgium - to France - to UK

1.25 USD per 100 km per toe

Algeria

50.1 USD per toe 50 USD per toe 50.1 USD per toe

/ 91

92 / The Energy Journal

Table 1. continued. Costs of national transportation (USD per toe) Country

Belgium

France

Netherlands

Italy

UK

West Germanv

1.59

15.03

4.41

11.86

7.99

28.77

Costs of storage and load balancing 23.62 USD per toe 3.22 USD per toe

Household markets Industry markets

Costs of distribution (USD per toe) Country

Belgium

France

Netherlands

Italy

UK

We,st Germany

household

80

200

40

40

60

100

industry

6

6

6

6

6

6

Sector

Production Natural gas is supplied by indigenous production, as well as imports from Algeria, the CIS, the Netherlands and Norway. As production from these nine countries vary substantially, we distinguish between two groups of producers according to degree of export and level of production. The large exporters of natural gas, Algeria, the CIS, the Netherlands and Norway are placed in the same group. Due to a substantial production of natural gas, the United Kingdom also belongs to this group.5 In Western Europe, governments have been involved in natural gas agreements, particularly with respect to cross-border transactions. Hence, political considerations, in addition to profitability, tend to influence supply agreements. In this study we focus, however, on comer cases where crossborder agreements are determined without government interventions. Hence, for

5. One could argue that as there are several producers of natural gas in the UK, it may be more adequate to assume price taker behavior. However, because we want to study the importance of price discrimination, the UK producer cannot be modelled as a price taker; if the domestic producer is large and acts as a price taker, there will be no price discrimination in equilibrium.

The Effects of Liberalizing

/ 93

the major producers extraction is determined by profitability only: we model all major producers as profit-maximizing Coumot players that act simultaneously. Needless to say, (particularly) for countries outside Western Europe this is a pure simplification. The second group of producers are those with limited production and no export (Belgium, France, (West) Germany and Italy). Production from each of these countries is taken as an exogenous delivery to the domestic markets. This could be interpreted as production is (basically) determined according to political considerations, e.g. to ensure a minimum security of supply. Alternatively, production is always at full capacity because market prices cover average costs (and capacity is exogenous). Note that most of our results are valid also for the case where the producers in group two are price takers. Since we assume that all economic agents can rent transport services and services to manage demand fluctuations at fixed rates, in the model Coumot producers sell natural gas directly to LDCs and to end users in the industry markets. Hence, producers are faced with costs associated with production of natural gas, transportation of natural gas and services to cope with demand fluctuations in the end-user markets. Table 1 shows the cost parameters of the model. In line with all observations of the gas industry, we assume that in each country natural gas is extracted simultaneously from several fields (with predetermined production capacities) that may have different unit costs. For the Netherlands, Norway and the UK, available information on production costs clearly justifies an assumption of increasing marginal costs, cf. e.g. North Sea Service and North West Europe Service from Wood Mackenzie and annual reports from the Norwegian Petroleum Directorate. For these countries, the parameters in the marginal cost functions are determined so that cost curves fit well to available information. In particular, we have assumed that marginal cost of production for the first unit of natural gas is 3, 38 and 60 USD per toe for the Netherlands (cf. Blitzer 1986), Norway (the Troll field) and the UK, respectively. Turning to costs of production outside Western Europe, according to Kalim (1991) “it is virtually impossible to establish a reasonable figure for Soviet production costs but they themselves are clear on one point: costs have substantially increased, by perhaps as much as fourfold over the past 15 years. ” Kalim relates increased costs to “problems of developing gas fields in regions that are remote and inhospitable, and getting more so with every discovery. 11For the CIS we assume that marginal cost of production doubles as deliveries to the European market increase from zero to the actual 1990 export quantity. Finally, we assume that marginal cost of production in Algeria is constant. This reflects that Algeria has huge reserves of natural gas in the desert, which could be extracted at a low cost (11 USD per toe, which is based on Dahl and Gjelsvik (1993)).

94 / The Energy Journal

Transportation In this study we distinguish between international and national transportation of gas. By international transportation we mean transportation of gas from the well-head to the border of the importing country. By national transportation we mean transportation of gas from the border to a point where all LDCs/large gas users are assumed to be situated. Hence, the estimates for national transportation costs should be considered as average values as “in the real world” consumers in a country are not located at the same point. Dal-11and Gjelsvik (1993) made a comprehensive survey on costs of production and costs of international transport of natural gas. In general, several of the referred transport cost estimates differ significantly. For pipelines, cost per mile per Mcf starts at 0.0005 and ends (significantly) above 0.002. The estimate of 0.0005 was used by MacDougall and Linder (1992), who also assumed a multiplicative factor of 2 and 3 for new onshore pipelines and underwater pipelines, respectively. According to our knowledge, underwater pipelines are more expensive than onshore pipes, but partly because of costs of land the difference should not be exaggerated. Moreover, our industry sources claim that actual transport rates in Europe are (on average) higher than 0.0005 per mile per Mcf. In this study, the tariff for international onshore transportation in Western Europe is set equal to 0.001 1990 USD per mile per Mcf, which corresponds to 2.49 USD per 100 km per toe. Due to lack of information on the cost distribution, this rate is used for all international onshore transportat,ion in Western Europe. For most pipelines, the tariff for offshore transportation is taken to be 50% higher than the common onshore rate. This estimate is based on cost information from the Norwegian Petroleum Directorate, cf. Golombek, Gjelsvik and Rosendahl (1994). Turning to transport costs in Algeria and the CIS, these could be lower than in Western Europe because costs of e.g. land are cheaper. According to Blitzer (1986), transport costs in the CIS are approximately 1.25 USD per 100 km per toe, i.e. 50% below our estimate for Western Europe. Moreover, as the EU does not have jurisdiction outside Western Europe, the basis of transport tariffs in Algeria and the CIS could differ significantly from the practice in Western Europe: Since transport lines have already been set up, it cannot be ruled out that producers only have to cover variable costs, which are negligible. Since it is virtually impossible to predict how much producers in Algeria and the CIS will pay for (domestic) transport, it is important to examine the impact of changes in these costs. As a starting point, we set the transport tariff in Algeria and the CIS to 50 % and 25 % of the estimate for Western Europe, respectively. The cost estimates for LNG transport (from Algeria) mainly follow the estimates in Adelman and Lynch (1986), cf. Table 1.

The Effects of Liberalizing

/ 9.5

For national transportation of gas, we rely on a study published by the Ministry of Industry and Foreign Trade in France where the average cost of national gas transportation in France is estimated to 15.03 USD per toe, cf. Direction de Gaz, de I’Electricitt et du Charbon (1993). Again, we assume that this rate gives owners of the national transport line a normal rate of return (when capacity is equal to demand, cf. the discussion above). For other countries we assume that the tariff for national transportation is given by the estimate for France multiplied by the size of the country relative to France (size is measured by number of kilometers of domestic pipelines). Finally, to serve end-users producers of natural gas must also be able to meet (seasonal and daily) fluctuations in demand. This requires storage capacity and load balancing, services which we assume are offered at fixed rates. Again, these rates should give owners of the facilities a normal rate of return. In this study, costs of storage and load balancing are taken from the report from the French Ministry of Industry and Foreign Trade. Due to limited information, we use these estimates for all countries, cf. Table 1. Distribution Gas producers first extract natural gas and then rent (international and national) pipeline capacity to transport the gas either to a distribution line, which connects a large end user to the transmission grid, or to a LDC (located at the “city-gate”). For end users in industry markets, cost of distribution is set equal to 6 USD per toe. This is simply a compromise between the report from the French Ministry of Industry and Foreign Trade (1 USD) and IIASA (1987) (5.6 and 16.8 USD for commercial users in developed and new areas, respectively). Customers in the household markets are served by local distributio:n companies. As each of these companies is the sole supplier of gas to some end users, we assume that the margin between the end-user price in the household market and the price paid to the producer at the city-gate is regulated so that LDCs earn a normal rate of return (when the capacity of distribution equals demand from the end users). To identify distribution costs in the household markets, we have compiled cost information from three sources. Our main source is the statistical yearbook from Figaz, which for several European countries provides time series for investments and number of gasmeters. Assuming a real rate of interest at 7% and 30 years (economic) life time of capital, for the period 198588 total distribution costs for new customers (measured in 1990 USD per toe) are in the range of 28 USD (the Netherlands) to 263 USD (France). For the period 1976-79, the cost ranking is identical t’o the one for 1985-88 (the Netherlands, Italy, UK, Belgium, (West) Germany and France). Moreover, cost levels are quite similar in the two periods.

96 / The Energy Journal

Hopefully, the data series from Figaz provide information that are comparable across countries. As a test, we compared two of our cost estimates with estimates based on national sources. First, according to the French study from the Ministry of Industry and Foreign Trade, costs of distribution to French households are 180 USD per toe, i.e. almost one third lower than the estimate based on the Figaz data. For households in West Germany, we used information from Bundesverband der deutschen Gas- und Wasserwirtschaji e V (BGW). For the period 1985-88, the BGW estimate for distribution cost is slightly higher than the estimate based on the Figaz data. For 1980-90, the difference was, however, opposite and significant. In this study we primarily rely on the estimates based on the Figaz data. We have, however, increased the lowest cost estimate and decreased the highest cost estimate, cf. Table 1. As these estimates are quite uncertain, we examine below how the equilibrium changes when distribution costs are equal across countries.

THIRD PARTY ACCESS IN THE EUROPEAN GAS MARKET The main purpose of this section is to illustrate the role of traders and market access in a liberalized European natural gas market where transport, storage and load balancing services are offered at fixed rates. In the model, the major producers of natural gas are Coumot players. Let xii be supply of gas from Coumot producer i to market j, and let x.~ be supply of gas to market j from all producers except i. Profits of producer i is then given by xi = zDj(x~+x-ii)xii i

- Ci(ZXU) - YzC& i i

Cl)

where Dj(.) is the inverse demand function of market j and ci(-) is cost of production for producer i, c’~ > 0, c”~ 1 0. Moreover, cii is the total cost of moving one unit of gas from the well head of producer i to end users in market j, i.e. the total unit cost of transportation, distribution, storage and loadbalancing. Profit-maximizing yields Dj + xiiDj’ s ci- + cii

M,j

Hence, if it is profitable to sell gas in a market ((2) holds with an equality) marginal revenue should equal total marginal costs, i.e. the sum of marginal cost of production, transport, distribution, storage and load balancing.

The Effects of Liberalizing

/ 97

It is clear from (2) that several producers may serve the same market. However, for each market supply differs across producers because total marginal cost,s differ. Moreover, total trade differs across markets because demand (and hence marginal revenues) differs. Finally, it is straight forward to show that SX,/~X-,~ < 0, i.e. the optimal response to increased production from other producers is to reduce your own production; increased production from the competitors drives down the market price and hence lowers marginal revenue. Since marginal revenue should always equal marginal cost, it is optimal to reduce own production as this action increases marginal revenue and decreases marginal cost if c”i > 0.

Table 2. The Market Scenarios Scenario

1

No traders.

UK imports gas only from Norway.

Scenario

2

Traders.

UK imports gas only from Norway.

Scenario

3

Traders.

Free trade in gas (the complete

market).

The analysis of this section proceeds in stages, cf. Table 2. In scenario 1, large producers are profit-maximizing Coumot players and small producers have exogenous production. Hence, the equilibrium in scenario 1 is characterized by price discrimination, i.e. the price difference between two markets exceeds the cost difference of serving these two markets. Thus arbitrage profits could be collected by traders; it is possible to buy gas from a producer (or a LDC or a large user of gas), rent transport capacity and sell the gas at a profit. In this study traders are introduced in scenario 2, i.e. by comparing scenario 1 and 2 we can identify the impact of traders. To be more specific, in scenario 2 we assume that (i) traders exploit all arbitrage possibilities and (ii) Coumot producers know that the final outcome is characterized by no price discrimination. In scenario 1 and 2, consumers in Belgium, France, Italy, the Netherlands and West Germany can purchase gas from Algeria, the CIS, the Netherlands and Norway (and from the domestic producer). On the other hand, like at present only Norway exports gas to the UK and the UK does not export any natural gas. As it is profitable for the other major producers to export gas to the UK, in the long-run this option should be exploited. We term the scenario where all Coumot producers can export natural gas to the UK, the UK producer can export gas to the Continent and traders exploit all arbitrage possibilities, the complete market (scenario 3).

98 / The Energy Journal

Price Discrimination The equilibrium without any traders (scenario 1) is shown in column 2 in Table 3.6 In contrast to the present situation, Algeria, the CIS, the Netherlands and Norway sell gas to all consumption countries on the Continent. The largest quantity of gas is extracted in the Netherlands (59.3 mtoe) where the rate of capacity utilization is 85 % . This reflects that (i) cost of production is quite low in the Netherlands as long as the rate of capacity utilization is below 90-95%, and (ii) the Netherlands are located near all markets. Next, production in Algeria, the CIS and Norway is 44.5, 40.0 and 38.7 mtoe, respectively. While Algeria and the CIS have low marginal costs of production but substantial costs of transportation to several markets, Norway has high costs of production but lower transport costs. A closer look at the equilibrium reveals that almost two thirds of the Norwegian production are exported to the United Kingdom, where Norwegian gas only meets competition from the domestic producer.’ UK production is only 22.9 mtoe, which is around 50% of the present level of extraction, cf. Table 3. Turning to profits, the profit ranking differs significantly from the ranking of production. Measured in billion 1990 USD, profits for the Netherlands and Norway are 7.2 and 5.1, whereas for the other producers profits are around 3.5. Profit per unit of production (toe) therefore varies across producers, ranging from the CIS and Algerian producer (around 85 USD), via the Norwegian and Dutch producer (around 125 USD) to the UK producer (157 USD). The differences reflect that (i) marginal costs differ across producers, (ii prices of natural gas differ across markets, and (iii) the composition of gas sai’es differs across producers. For the UK producer, the high profit per unit of production mainly reflects less competition in the UK markets than on the Continent. As an illustration of the various degrees of competition, observe that the difference between price and total marginal cost for Norwegian supply to LDCs is 164 USD in the UK and 35 USD in West Germany. In equilibrium, there are a number of arbitrage possibilities that could be exploited by traders. In e.g. West Germany, the netprice (end-user price less of costs of distribution, storage and load-balancing) in the household and industry market is 167 and 144 USD, respectively. Hence, there is a potential arbitrage profit that could be collected by purchasing gas from a producer (or from a large user) and sell the gas to a LDC. Moreover, there are also international arbitrage possibilities. To take one example, as the net price in the

6. comer 7. Recall

We used GAMS to find all equilibria described in this paper. In particular, we checked for cases; in e.g. scenario 1 it is not optimal for Norway to sell gas to large end users in Italy. Increased Norwegian exports to the UK require an extension of British North Sea pipes. that the model is not restricted by existing pipe capacities.

The Effects of Liberalizing

/ 99

industry market is 127 USD in Belgium and 144 USD in West Germany, and the cost of transporting gas from end users in Belgium to end users in West Germany is around 6 USD per toe, it pays off to move gas from Belgium to West Germany (127 +6 < 144). What happens when all arbitrage possibilities are exploited by traders?

Table 3. Production (in mtoe) and Profits (in billion 1990 USD) Scenario

1990

Producer

(1) price discrimination

(2) no price discrimination

(3) complete market for natural gas

Algeria production

21.8

profits

44.5

44.4

60.1

3.8

3.3

4.7

40.0

39.9

48.3

3.4

2.9

3.4

59.3

59.2

66.7

7.2

6.7

7.5

38.7

39.1

29.3

5.1

4.4

1.6

22.9

22.9

16.2

3.6

3.4

0.7

205.4

205.5

220.6

23.1

20.7

17.9

CIS production

41.5

profits Netherlands production

53.4

profits Norway production

22.9

profits U. K. production

42.5

profits Sum Coumot producers tot. prod. tot. profits

Sources

182.1

for 1990 data: BP Statistical Review of World Energy.

100 / The Energy Journal

Table 4. Natural Gas Price (1990 USD per toe) and Consumption (mtoe) Scenario

(1) 1990

Country

house -hold

(2)

price discrimination

indus -try

house -hold

indus

(3)

no price discrimination

-try

house -hold

indus -try

complete market for natural gas house -hold

indus -tly

Belgium price consumption

395 3.8

154 5.1

286 5.0

136 5.5

250 5.4

155 5.0

250 5.4

156 5.0

France price consumption

457 13.8

172 13.2

392 16.8

145 15.0

377 17.5

162 13.9

373 17.7

158 14.1

Italy price consumption

484 16.7

152 25.0

258 26.1

121 28.7

195 28.7

140 26.3

197 28.7

142 26.1

Netherlands price consumption

311 15.3

132 18.2

295 15.7

131 18.3

209 17.8

154 16.2

208 17.8

153 16.3

U. K. price consumption

366 32.4

172 17.0

355 33.4

217 13.8

329 36.1

255 11.0

242 45.3

167 17.3

West Germany price consumption

366 20.6

174 29.7

291 25.7

153 32.2

274 26.8

160 31.5

276 26.7

161 31.2

Sum av. price tot. consum.

390 102.6

160 108.2

320 122.7

150 113.5

282 132.3

169 103.9

254 141.6

156 110.0

Sources for 1990 data: IEA Energy Prices and Taxes (prices),

IEA Oil and Gas information 1992

(consumption).

No Price Discrimination Assume now that traders are able to exploit all arbitrage possibilities (scenario 2). Hence, there is no price discrimination in equilibrium, i.e. net prices in a country are equal, and price differences between countries are less than the costs of transporting gas between the end users.

The Effects of Liberalizing

/ 101

We assume that all Cournot producers maximize profits, subject to the constraint that the final outcome is characterized by no price discrimination. Due to the constraint of no price discrimination, in equilibrium total marginal cost of a producer is now not equal to marginal revenue. Table 3 shows that compared with scenario 1, production of each Cournot producer is almost unchanged when all arbitrage possibilities become exploited. Hence, total production and total consumption are only slightly altered. On the other hand, in each country consumption in the household market increases, whereas there is a drop in the industry market, cf. Table 4. These changes reflect that arbitrage possibilities are now exploited: in each country, the net price in the household market was (in scenario 1) higher than the net price in the industry market. In scenario 2 where net prices (in each country) are equal, relative to scenario 1 prices in the household markets have decreased, whereas prices in the industry markets have increased. Hence, consumption in the household markets increase (by 8 %), whereas consumption in the industry markets decrease. Moreover, with no price discrimination profits decrease. Total profits to Cournot producers decrease (relative to scenario 1) by 10%) cf. Table 3. The drop (measured both in USD and in percent) is lowest: for the UK producer, i.e. the producer which faces least competition. Thus Algeria, the CIS and the Netherlands have now even stronger incentives to export natural gas to the UK. The Complete Market Assume now that an offshore pipeline is laid between Zeebrtlgge in Belgium and the UK (scenario 3).8 Again, producers can rent capacity at a rate which gives the owners a normal rate of return. As this pipeline is short (only 120 km), it could be relatively costly to construct (per km) due to fixed costs. On the other hand, the pipeline should be based on a more cost efficient technology than previous pipelines in the North Sea. To balance these two1 effects, we simply assume that the imposed tariff is equal to the rate for offshore transportation to the UK (7.47 USD per 100 km per toe). This pipeline is used by the CIS, the Dutch and the UK producer, whereas we assume that the Algerian producer relies on LNG transport, cf. Table 1. Compared with scenario 2, there is a substantial increase in production from Algeria, the CIS and the Netherlands, i.e. from the new exporters to the markets in the UK. On the other hand, Norwegian and UK production decrease, cf. Table 3. A closer look reveals that as Algeria, the CIS and the Netherlands initiate exports to the UK, Norwegian exports to the UK decrease. A reduction

8. This assumption is adequate as the so-called Interconnector transport gas (both ways) between Zeebriigge and the UK.

project

will be designed

to

102 / T?ze Energy Journal

in Norwegian exports simply reflects that the response curve of a Coumot producer is downward sloping, cf. the discussion above. However, part of the drop in Norwegian exports is compensated by increased sales in other markets. While the export opportunities increase (cet. par.) UK production, increased competition in the UK markets decreases (cet. par.) UK production. The net effect is a fall in UK production, which partly reflects that the UK producer has high marginal costs. In total, production from the five Coumot players increases by 7%. Table 4 shows that compared with scenario 2, except for the UK markets, consumption is only slightly altered. Total gas consumption in the UK increases by around one third, making the UK the largest gas consumer in Europe. Turning to profits, as a new market is opened up for Algeria, the CIS and the Netherlands, profits of these producers increase (relative to scenario 2), whereas the opposite is the case for the Norwegian and the UK producer. In particular, profits of the UK producer decrease from 3.4 to 0.7 billion USD, i.e. by 80%. Because the overall degree of competition has increased, total profits of all Coumot producers drop (by 14%). Compared with the observed outcome of 1990, total production is almost 20% higher. While production from the CIS, the Netherlands and Norway has increased by 15-30 % , Algerian production has more than doubled whereas the UK production has decreased by more than 50%.g Moreover, producers from Algeria, the CIS, the Netherlands and Norway now sell gas in all consumption countries, cf. Table 5. As total consumption has increased, the average West-European price of natural gas has decreased. However, as there was substantial price discrimination in 1990, but there is no price discrimination in scenario 3, some prices of natural gas could be above the 1990 level. It turns out that all prices of gas in the household markets have decreased; the weighted average household price has decreased by one third. In the industry markets, the weighted average price has only decreased marginally. However, in the Netherlands (which had the lowest price initially), there is a significant price increase (19%). There are a number of reasons why the equilibrium of scenario 3 differs from the observed outcome of 1990. First, the objective and strategy of the

9. As Algeria and the CIS are (at present) frequently regarded as high-risk sources, supply security may be important to Western Europe also after a radical liberalization (which moves the market structure closer to a spot market). This is ignored in our study. Primarily, supply security should be examined within a framework of uncertainty, cf. Hoe1 and Strom (1986). However, one ad hoc way to model supply security is to risk adjust costs of production. As one example of such an exercise, we increased the cost of production for Algeria and the CIS by an amount corresponding to 5% of the average gas price in our six consumption countries (in 1990). Relative to scenario 3, production from Algeria and the CIS decreases by 9%. whereas total production from the Netherlands, Norway and the UK increases by 4%.

The Effects of Liberalizing

/ 103

major gas producing countries differ: in our analysis we have assumed that (the major) producers are Coumot players which simply maximize profits. Second, gas is now sold directly to LDCs and large end-users, not to transmission companies. Moreover, in scenario 3 there are no bottle-necks in the pipeline network, whereas at present bottle-necks may restrict supply. Fourth, in the complete market scenario all major producers have the option to sell gas in all consumption countries. Fifth, in contrast to the present situation there is no price discrimination. Finally, costs of production, transportation and distribution (may) differ. Taking the first five factors for granted, in the next section we examine how the equilibrium outcome changes when the cost assumptions are altered.

Table 5. Natural Gas Production, Consumption and Trade (in mtoe) in a Complete European Market (1990 values in parentheses) Market

Belgium

France

Italy

Nether -lands

8.7

U.K.

West Germany

Total deliveries

15.7

10.9

60.1 (21.8)

Producer Algeria

as

2.8

8.3

13.7

(3.5)

(8.3

(9.9)

7.1

9.3 (12.3)

7.2

10.5

11.9 (20.3)

48.3 (41.5)

10.9 (22.8)

16.3

15.6 (17.7)

66.7 (53.4)

10.5

5.6

(Z)

(6.7)

(7.1)

29.3 (22.9)

9.6 (42.5)

0.6

16.2 (42.5)

13.1 (13.1)

30.9 (30.9)

57.7 (58.2)

251.5 (213.0)

2.3

(8.9) Nether -lands

Norway

3.3

9.3

11.3

(3.5)

(4.1)

(5.3)

1.4

(2.0) United Kingdom

0.6

Domestic

Consumption

Sources 1992.

10.4 (9.0)

(0.1)

3.5 (K) 1.2

1.8

2.7 (2.7)

15.1 (15.1)

31.9 (29.1)

54.7 (42.6)

2.4

34.2 (24.8)

62.6 (49.3)

for 1990 data: BP Statistical Review of World Energy, and IEA Oil and Gas Information

IO4 / The Energy Journal

COSTS OF PRODUCTION, DISTRIBUTION

TRANSPORTATION

AND

The complete market scenario in the third section differs from the present European natural gas market with respect to market structure. This is why we did not study the liberalization of the European gas market by first constructing a model for the present situation, and then examine effects of a liberalized market by simply shifting some parameters of the model. Instead, we constructed a model that captures the main features of an ideal nondiscriminatory regulation in which all economic agents can rent gas transport services and companies in the gas industry are forced to act competitively. In calibrating the model, we relied on a number of sources for cost information. Because these are uncertain, it is important to test the robustness of the complete market equilibrium. The main purpose of this section is thus to identify, through comparative static, the impact on equilibrium prices and quantities of shifts in costs of production, transportation and distribution, cf. Table 6.

Table 6. Shifts in the Complete Market Scenario Scenario Scenario Scenario

3 4 5 6

Complete Complete Complete Complete

market. market. market. market.

Positive resource rents. Lower transport tariffs iu Western Equal costs of distribution.

Europe.

Costs of Production In the second section above, production cost estimates were based on (the implicit) assumption that costs reflect only exploration, development and extraction activities. However, there is-both for society and for a private producer-an additional cost which is called the Hotelling rent or the (marginal) resource rent. This cost reflects that increased present extraction affects the possibility and profitability of future extraction. First, this could be a cost due to the limited quantity of gas reserves. Second, if natural gas is extracted from fields with increasing unit costs, increased extraction today accelerates extraction from more expensive fields in the future. In general, the resource rent at time c is the value of getting one more unit of the resource at time t (with the same cost as the unit to be extracted), evaluated at time t. Golombek and Hoe1 (1990) estimate the resource rent for Norwegian natural gas by using a dynamic optimization model with a planning period running from 1990 to 2070. In their reference scenario, the resource rent is 24 USD in the year 2000 (at 1990 prices). In this study we use Golombek and

ILKhe Effects of Liberalizing

/ 105

Hoel’s estimate for the Norwegian resource rent. Moreover, we also use their model to estimate the resource rent for the other major producers. By changing the assumptions about costs, demand and initial stock of gas, the resource rent in 2000 for the CIS, the Dutch and the UK producer was estimated to 1, 64 and 61 USD per toe, respectively. These resource rents should only be taken as very rough estimates. Finally, as Algeria has huge reserves of natural gas, and its marginal cost of production is assumed constant in this study, the resource rent for Algeria is zero. When resource rents are included in costs of production (scenario 4), compared with scenario 3 production in Algeria and the CIS increases whereas the opposite is the case for the other Coumot producers, cf. Table 7. The drop is particularly strong for the UK producer, which now has very high marginal costs of production. In total, production from the Coumot producers decreases by only 5%.

Table 7. Robustness. Production (in mtoe) and Profits (in billion 1990 USD) Scenario

(6)

(3)

(4)

(5)

complete market for natural gas

resource rent

lower transport tariff

equal costs of distribution

Algeria production profits

60.1 4.7

72.0 6.7

58.7 4.5

60.2 4.7

as production profits

48.3 3.4

56.5 4.7

52.9 4.1

48.4 3.4

Netherlands production prOfitS

66.7 7.5

46.9 6.2

66.5 7.4

66.8 7.5

Norway production profits

29.3 1.6

26.0 1.9

31.2 1.8

29.3 1.6

U. K. production profits

16.2 0.7

7.5 0.5

16.9 0.7

16.3 0.7

220.6 17.9

208.9 20.0

226.2 18.5

221.0 17.9

Producer

Total Coumot producers production profits

106 1 The Energy Journal

Table 8. Robustness. Price of Natural Gas (in 1990 USD per toe) and Consumption (in mtoe) Scenario

Countty

(3) complete market for natural gas house -hold

indus -try

(4) resource rent

house -hold

indus -try

(5) lower transport tariff

house -hold

indus

(6) equal costs of distribution

-try

house -hold

-try

indus

Belgium price consumption

250 5.4

156 5.0

264 5.2

170 4.7

245 5.5

151 5.1

261 5.3

1.59 4.9

France price consumption

373 17.7

158 14.1

391 16.9

177 12.9

366 18.0

152 14.6

268 22.6

167 13.6

Italy price consumption

197 28.7

142 26.1

210 28.1

156 24.4

194 28.8

140 26.4

246 26.6

145 25.8

Netherlands price consumption

208 17.8

153 16.3

222 17.5

168 15.0

203 18.0

148 16.8

258 16.6

1.56 16.6

U. K. price consumption

242 45.3

167 17.3

259 43.5

184 16.2

233 46.2

159 18.0

264 43.0

162 17.7

West Germany price consumption

276 26.7

161 31.2

290 25.8

176 29.6

269 27.2

154 32.1

265 27.4

164 31.0

Sum average price tot. consump.

254 141.6

156 110.0

270 137.0

171 102.9

247 143.8

149 113.6

259 141.7

156 110.2

Decreased production gives higher end-user prices in all markets, cf. Table 8. As the (positive) price effect is stronger than the (negative) quantity effect (i.e. production “moves towards the monopoly point”), total profits (i.e. revenues minus costs of exploration, development and extraction) of the Coumot producers increase. Turning to each producer, profits of the Algerian and the CIS producer increase because both prices and production increase. On the other hand, for the Dutch and the UK producer (i.e. producers with high resource rents), profits decrease because lower quantities dominate the price effect.

i%.e Effects of Liberalizing

/ 107

Costs of Transport To identify the importance of transport tariffs, in scenario 5 we examine the case where all onshore and offshore tariffs in Western Europe are reduced by 25 % . Moreover, all national transport tariffs in Western Europe are reduced by 25% as well. Lower transport tariffs in Western Europe yield increased extraction from some producers, but decreased extraction from others. Extraction from producers which are located far from the markets and use pipelines increases because the relative cost position of these producers (CIS and Norway) has improved. On the other hand, production in Algeria decreases as this producer primarily uses LNG. Finally, for producers located near the markets (the Dutch and the UK producer) extraction is almost unchanged as the small partial effect of lower transport tariffs is counteracted by a net increase in production from the competitors. The total effects of a significant drop in all transport tariffs in Western Europe are moderate. Total production and profits increase by around 3 % , and all end-user prices fall by 2-6%. Note that the main part of these effects is due to lower international (not national) tariffs. The impact of lower (international and national) transport tariffs is limited because total transport costs as a share of gas prices are quite low. To take one example, both for imports from the CIS and Norway total transport costs as a share of the French household price amount to around 15 % . Turning to tariffs for onshore transportation outside Western Europe, in the complete market scenario international onshore transportation in Western Europe is four times as costly as transportation in the CIS. If, however, the CIS producer is charged the same tariff in the CIS as in Western Europe, relative to the complete market scenario total production decreases by only 4 % . On the other hand, CIS production decreases by 54% (and profits by 82%), whereas e.g. Norwegian production increases by 17%. Turning to the other comer case, if the CIS producer is not charged for transportation in the CIS, his production increases by 18 % . In this case other producers respond by cutbacks; Norwegian production decreases by e.g. 5 % . Costs of Distribution So far we have assumed that distribution costs of large end users are quite small and equal across countries, whereas distribution costs of LDCs are much larger and not equal across countries. As explained above, our distribution costs for LDCs are derived from present value calculations based on data from Figaz. These estimates-ranging from 40 USD per toe (the Netherlands) to 200 USD per toe (France)-are uncertain. To take some examples, according to an IIASA study by Strubegger and Messner (1986), M&rig (1984) estimated the

108 / The Energy Journal

costs of storage and distribution for Germany to 192 USD per toe and Safoschnik (1985) obtained 170 USD per toe for Austria. Strubegger and Messner used the average of these two estimates when they examined the cost of supplying a Central European consumer with natural gas. However, in a 1987 study also from the IIASA, for the year 2000 town distribution costs (which included gas storage costs) were for Western Europe estimated to 77 USD per toe, i.e. less than half of the estimate used by Strubegger and Messner. Gas distribution costs are difficult to estimate because consumption patterns, population densities and the topography substantially influenc.e the costs. Moreover, marginal cost of distribution is surely dependent on whether new customers are linked up to the old network, or the network is expanded into a new area. As an alternative to the estimates of the second section above, we now examine the case where distribution costs of LDCs are equal in all countries (scenario 6). The common value (87 USD per toe) is simply estimated as the average of the costs of distribution in the complete market scenario.” Table 7 shows that a common cost of distribution in the household markets has little impact on the equilibrium. Both production and profits of each Coumot producer are almost unchanged. There are, however, some changes in prices and consumption: in France, where cost of distribution has been reduced significantly, consumption in the household market increases by almost 301%)cf. Table 8. The corresponding changes in the other countries are much lower. To sum up, differences in national distribution costs tend (for a given average value) to be of little importance. CONCLUSIONS This paper uses an empirical model to examine the impact of a significant liberalization of European natural gas markets. As transmission and distribution of gas are natural monopolies, a liberalization requires changes in the regulatory regime: Hence, even if the gas markets in Western Europe are liberalized, the industry will still be under strict regulation. In this paper we study the West-European natural gas markets under different assumptions for degree of liberalization. First, we focused on the case of profit maximizing Coumot producers and an ideal TPA system for transport of gas (scenario l), which yields an equilibrium characterized by significant price discrimination (net prices differ across markets). Because end-user prices

10. Since we use data from company accounts, and distribution companies may differ with respect to efficiency and accounting practice, for several companies the tme costs in a competitive environment may be significantly lower than the estimates used in this study. This is also the case for other secton in the natural gas industry.

The Effects of Liberalizing

/ 109

are (on average) lower than in the base year 1990, total consumer surplus in the six consuming countries increases by 32% relative to 1990, cf. Table 9. Turning to profits, in comparing total profits (to producers, transporters and distributors) in 1990 with scenario 1 we need a number of (simplifying) assumptions. The most important (from a numerical point of view) are the following.” First, we assume that costs of production, transportation and distribution are identical in these two equilibria. Moreover, we also assume that for small producers unit cost of production is 25 USD per toe (old onshore fields). Finally, in scenario 1 profits in the transport and distribution sectors are zero because owners of such facilities are assumed to make a normal rate of return on their investments. Under these assumptions, total profits in WestemEurope decrease by 6% (relative to 1990), whereas welfare in Western Europe, i.e. the sum of consumer surplus and profits in Western Europe, increases by 17%. In scenario 2, traders enter the markets and exploit all arbitrage possibilities that are stemming from the net price differences in scenario 1. Relative to scenario 1, production from each producer is almost unchanged. The big winners are end users in the household markets (lower prices). Compared with 1990 and scenario 1, consumer surplus and welfare have increased, whereas profits have decreased. ‘*

Table 9. Welfare Effects of Market Liberalizations. Changes in Consumer Surplus, Profits and Welfare Relative to 1990 (in per cent) Scenario

Consumer

(1) price discrimination

(2) no price discrimination

(3) complete market

surplus

32

42

57

Europe

-6 - 11

- 14 - 20

- 36 - 29

17 11

20 12

20 15

Profits Western total Welfare Western total

Europe

11. For a further discussion about these assumptions and an examination of the impact on the profits of tbe Norwegian producer, confer Gjelsvik, Golombek and Rosendahl (1995). 12. Note that profits to traders are zero: in line with standard economic theory we assume that traders keep entering the market until arbitrage profit is driven down to zero.

110 / The Energy Journal

In scenario 3, all Coumot producers (not just Norway) export gas to the UK, and the UK producer is free to sell gas on the Continent. Relative to scenario 2, total consumer surplus is higher, whereas profits to producers in Western Europe decrease due to more competition. On the other hand, profits to producers outside Western Europe increase because these have now access to a new market. Both welfare in Western Europe and total welfare, i.e. the sum of welfare in Western Europe and profits of non-European producers and pipeline/LNG companies, are higher than in the three other cases (1990 and scenario 1 and 2).13 Relative to 1990, welfare in Western Europe has increased by as much as 20%.

REFERENCES Adelman, M. A. and M. C. Lynch (1986). “Natural gas trade in Western Europe: The permanent surplus.” In Wesfern Europe Nafurul Gus Trade. Final report. Center for Energy Policy Re:;earch. Energy Laboratory. MIT. AI-Sahlawi, M. A. (1989). “The demand for natural gas: A survey of price and income elasticities.” The Energy Journal 10(l): 77-90. Ballard, C. L. and D. Fullerton (1992). “Distortionary taxes and the provision of public goods.” The Journal of Economic Perspectives 6(2): 117-13 1. Birkelund, H., J. Fuglestvedt, E. Gjelsvik and M. Aaserud (1991). Energiettersp0rselsmodell for Vest-Europa (energy demand model for Western Europe). Unpublished report. Statistics Norway. Blitzer, C. R. (1986). “Western European natural gas trade model.” In Western Europe Natural Gas Trade. Final report. Center for Energy Policy Research. Energy Laboratory. MIT. British Petroleum. Review of World Gas (annually). London. British Petroleum. Statistical Review of World Energy (annually). London. Bundesverband der deutschen Gas- und Wasserwirtschaft e V. Gasstatistik (annually). Bonn. Cameron, P. (1990). Gas Regulation in Western Europe. A country-by-country guide. Financial Times Business Information. London. Chermak, J. M. and R. H. Patrick (1994). “Incentives in pipeline pricing and capacity.” In M.A. Crew, ed., Incentive Regulation for Utilities. Boston: Kluwer Academic Publishers. pp. 12’7-184. Commission of the European Communities (1988). The Internnl Energy Market. COM (88) 238 Final, Brussels, 2 May 1988. Commission of the European Communities (1992). Completion of the Internal Market in Electricity and Gas. Brussels, 21 January 1992. Dahl, C. and E. Gjelsvik (1993). “European naturalgas survey.” Resources Policy. September 1993: 185-204. Direction de Gaz, de I’Electricitk et du Charbon (1993). Rapport d’activitt!. Ministirre de I’Industrie et du Commerce Extkrieur. France.

13. Because (part of the) profits in the gas industry are kept by governments, one could argue that in calculating welfare the weight of profits should be higher than the weight of consumer surplus (government revenues could be used to reduce distortionary taxes). At present, it is customary to let the weight of profits be between 20% and 50% higher than the weight of consumer surplus, cf. e.g. Ballard and Fullerton (1992). When the weight of profits is as high as 1.5. the welfare ranking of Western Europe is as follows: scenario 2 (top), scenario 1, scenario 3 and 1990.

The Effects of Liberalizing

/ 111

Figaz. Stutisticul Yearbook (annually). Bruxelles. Gjelsvik, E., R. Golombek and K. E. Rosendahl(l995). “Liberalization of the natural gas markets in Western Europe: The impact on Norwegian profits.” Unpublished paper. SNF. Oslo. Golombek, R., Gjelsvik, E. and K. E. Rosendahl (1994): “Effects of liberalizing the natural gas markets in Western Europe.” Memorandum from Department of Economics, University of Oslo, 15194. Golombek. R. and M. Hoe1 (1990). “The resource rent for Norwegian natural gas.” In Bjerkholt, 0.. 0. Olsen and J. Vislie, eds., Recent Modelling Approaches in Applied Energy Economics. Chapman and Hall. Hoel, M. and S. Strem (1986). “Supply security and import diversification of natural gas.” In Golombek, R., M. Hoe1 and J. Vislie, eds., Natural gas Markets and Contracts. North-Holland. pp. 151-172. Hopper, R. (1994). “EU TPA battle draws to an end.” Petroleum Economist. March 1994. IEA. Energy Prices and Taxes (annually). Paris. IEA. Oil and Gas Information (annually). Paris. Intergas Marketing (1993). The Domestic Gas Market. Gaz de France. International Institute for Applied Systems Analysis IIASA (1987): “International natural gas market.” Working Paper. WP-87-102. Kalim, 2. (1991). Natural Gus in the European Community. Financial Times Business Information. London. Laffont, J.-J. and J. Tirole (1993). A Theory of Incentives in Procurement and Regulation. Cambridge, MA: MIT Press. MacDougall, M. W. and P. T. Linder (1992): Long-term Outlookfor World Gas Trade 1990-2015. Canadian Energy Research Institute. Study no. 46. Miinig, W. (1984). “Internal report on energy conversion and delivery costs.” (Kemforschungsanlage Jtilich GmbH). International Institute for Applied Systems Analysis, Laxenburg, Austria. Norway (various years). The Norwegian Petroleum Directorate. Annual Reports. Price, C. (1994). “Access charges in the U.K. gas industry.” Paper presented at the 17th annual conference of IAEE, Stavanger, Norway. Safoschnik, R. (1985): “Gegenwartiger stand des erdgasfemtranspottes aus technischer und wirtschaftlicher sicht.” (Present status of the natural gas long-distance transportation from a technical and economical view). Osterreichische Mineral61 Verwaltung (OmV), Vienna, Austria. Stem, J. P. (1992). Third Party Access in European Gas Industries. Regulation-driven or Marketled? The Royal Institute of International Affairs. London. Strubegger. M. and S. Messner (1986). “The influence of technological changes on the cost of gas supply.” Working paper from the International Institute for Applied Systems Analysis IIASA. WP86-38. Vrieling, E.B., J. Munksgaard and R. J. Hopper (1989). Transparency in Natural Gas Prices in Western Europe. Komgas and Vegin. Wood Mackenzie. North Sea Service and North West Europe Service.