Open access in the European Union

In the European Union, open access was established in the Directive 98/30/EC of the European Parliament and of the Council of 22 June 1998 (the First Gas Directive).

The first directive only defined that pipeline access shall be provided in manner determined by Member States, with the objective of fair and open access, achieving a competitive market in natural gas and avoiding abuse of dominant positions. In this first moment, the meaning of open access was vague and allowed varied interpretations. The second directive, (EC, 2003), defined the implementation of regulated Third Party Access. The interpretation of this directive led the EU Member States to define an ‘ex ante’ set of rules for infrastructure use. The definition of these rules depends on a third party (frequently a regulatory body but also may include other governmental bodies such as the Ministry). The third directive, (EC, 2009), did not change the main concept of the regulated Third Party Access, but it established the need to establish common (or harmonized) rules in the EU countriesFootnote ⁵. It meant that every shipper could access the network following a regulated set of procedures and paying a regulated tariff. According to the article 35 of EC (2009), the refusal to grant access would be possible whether there was lack of capacity and/or it could bring safety risk. However, duly substantiated reasons shall be given for any such a refusal. In the long run, Member States may take the measures necessary to ensure that the necessary enhancements are promoted in order to avoid refusals. That is, there is a strong effort to increase the difficulty to exclude players from accessing network services, as the network access is seen as a key element for promoting the wholesale gas market.

Open access in the United States

In the United States, the open access of network services has been claimed also as one of the main elements of the gas market liberalization and developmentFootnote ⁶. However, even if the meaning of open access in USA is different from the EU, we also observe an increased difficulty to exclude players from using the infrastructure.

In the US, the liberalization started almost two decades before Europe. In 1985, the Federal Energy Regulatory Commission (FERC) announced Order No. 436, which introduced voluntary open access for the interstate gas pipelines (Marston, Reference Marston1991). Pipeline companies could choose to be only a transporter of the gas (open access) or carry their own gas (remaining private carriers but keeping their gas price regulated). In 1992, FERC introduced Order No. 636, which required pipeline companies to fully unbundle gas trading from pipeline operation activities, and to set up separate transportation and trading affiliates. In other words, pipelines were forbidden to carry their own gas and forced to carry gas of third parties (Makholm, Reference Makholm2012). The key consequence of the Order N. 636 was the complete separation between the ownership of the infrastructure and the services provided.

The Federal Energy Regulatory Commission (FERC) regulates many aspects of interstate gas transmission pipeline operations, including approval, permitting and site for new pipeline facilities, as well as transmission rates that pipelines are permitted to charge for interstate shipments. The rules of service access, however, are negotiated among pipeline owners and shippers. Hence, the rules of network use are defined in the contract between the shipper and the pipeline and are supervised by FERCFootnote ⁷. So, differently from the European scheme, the rules of infrastructure use are defined by negotiation between the implicated players, and the third party (FERC) plays a role in case there is no agreement. The exclusion of a player from the access to the pipeline depends on the pipeline available capacity and in the rules defined between the pipeline and the users (in case that the exclusion is not justified by the pipeline, the shipper may require the intervention of FERC). Therefore, the ownership of the infrastructure does not define the rights of use of the infrastructure; those rights are defined by the service rules. Consequently, the definition of these rights becomes more costly, because the costs of measuring and enforcing those rights are greater than the costs to define rights of infrastructure ownership.

Subtractability

As explained by Ostrom et al. (Reference Ostrom, Schroeder and Wynne1993b), it is important to highlight that, in general, an infrastructure that can be used by many players can be considered a subtractable service,Footnote ⁹ and it is especially important in the case in which the flow of services is scarce with respect to demand. Natural gas networks are a combination of pipelines and compressors. The gas flows from the higher pressure point to the lower pressure point, limited by the technical characteristics of the pipeline system. If the pressure differential between two points is lower than the pipeline material resistance, the gas will remain stored inside the pipeline. The limit of gas volume that can be transported/stored is largely limited by the pipeline characteristics and the compressor potential.Footnote ¹⁰

The flow of services that a gas network may offer is limited, that is, the use of the flow of services by a player decreases the amount of the available flow of services in the same period of time. Hence, the overuse of the gas network services may generate congestion or even safety risks, which is negative externality for all the network users and may be a problem even for players geographically close but not economically related with the infrastructure.Footnote ¹¹

Excludability

Gas networks are ‘naturally’ excludable goods, in the sense that it is not particularly difficult to exclude individuals from the use of the gas network. However, restructuring gas industries implied opening the access to gas pipelines (for instance, United States and European countries) in order to promote competition. As explained by Ostrom and Hess (Reference Ostrom and Hess2010), open access can be a consequence of conscious public policy to guarantee the access of all players for the use of a resource. That is the case of gas transmission pipelines. The precise meaning of that ‘open-access’ is quite diverse, as was shown in Section 2. However, all share the characteristic that the infrastructure owner has no right to discriminate users, and should allow a large number of players to access the infrastructures. Therefore, despite some differences in the implementation of open access, all cases are characterized by the choice of creating difficulties to the exclusion of players from accessing the network services. The other side of this process of decreasing excludability is the creation of access's rules. The logic for that can be thought to be that, in the presence of common-pool resources, it is necessary for the definition of some rules.Footnote ¹² In the absence of suitable rules, incentives may lead to network overuse.

In that view, we may classify exclusion in the context of gas pipelines under two headers: exclusion to access the resource system (physical connection and permission to carry gas), and exclusion to access resources units (right to use the system). Open access policies required a series of characteristics and rules for the player to fulfill in order to obtain permission to carry gas. Thus, it is possible to argue that excludability decreases but it exists. Nonetheless, once the player becomes a shipper (obtains permission to carry gas), the ability of excluding the shipper to use a specific flow of service becomes much lower.Footnote ¹³ Consequently, after the player becomes a member of the group of users (a shipper), she has incentives to overuse the service. The entry procedure does not guarantee that players use the resources efficiently. It will depend on the incentives of shippers (benefits and costs of using the services in the different ways). In gas industries, the decision of what kind of service shippers may use depends on the rules of use of the network.

Potential prisoner's dilemmas when networks offer different products

The first step of our reasoning is to put forward a simple example of a commons’ dilemma in gas networks. Consider a simplified model of a gas network that two players can use. Both players obtain the same value for the flat use of network v_flat = 6 and a value for the flexible use of the network v_flex = 10. Players decide on two possible options: using 2 MW of flat network capacity, or using 1 MW of flat capacity plus 1 MW of flexible capacity, that is, we are first supposing that flat and flex extract the same amount of services units of the available amount.

Game 1

Let us assume first that the network capacity is 4 MW regardless of the combination of flat and flexible use of the network. For the same transmission system, the amount of available flex capacity is equivalent to the amount of flat capacity.

The grey box of Table 1 represents the equilibrium of the game. In this case, there is no commons’ dilemma regarding the rules of resource allocation. But the result is based on considering that the services offered by the transmission network are equivalent for the system (the use of flexible capacity subtracts the same amount of the available capacity than flat capacity). It means that whether the use is flexible or flat is irrelevant, as they affect the others players the same way.

Table 1.

First game illustration

Game 2

Nevertheless, in practice, flexible capacity subtracts more from the available capacity than flat capacity. Hence, the impact of flexible demand for network capacity is different from the impact of flat demand. Flexible services demand more capacity from the total capacity of the system than flat services. To illustrate this, consider the following modification of the initial game: the capacity of the network, if both players choose to use it flat, is 4 MW. If one player chooses flexible use, the flat capacity is 2 MW and the flexible capacity is 1 MW. If both players use the network flexibly, the flat capacity is zero and the flexible capacity is 2 MW. It means that the use of flex capacity (1 MW) leaves less available flat capacity for further allocation than the use of flat capacity (1 MW).Footnote ¹⁸

Capacity allocation 1

If the demand for network is higher than its capacity, the rules for allocating network capacity are the following: (1) the capacity is allocated equally among users and (2) the flex capacity is allocated first (there is priority for flex capacity before the allocation of any flat capacity).

In this second game, the equilibrium is represented by the grey box in Table 2. We observe a prisoner's dilemma because players have to choose between more capacity or more valued capacity. If the other player chooses valuable capacity, the amount of your own capacity is restricted. The use of the rules allocating capacity to both shippers equivalently and giving priority to flexible capacity results in prisoners’ dilemma outputs.

Table 2.

Second game illustration

But it is possible to consider a different allocation rule.

Capacity allocation 2

If the demand for network is higher than its capacity, the rules for allocating network capacity are the following: (1) the capacity is allocated equally among users and (2) the flat capacity allocated first (there is priority for flat capacity allocation before the allocation of any flex capacity).

In this game, the equilibrium is represented by the grey box of Table 3. We observe that the previous dilemma is now solved by giving priority to the allocation of flat capacity.

Table 3.

Third game illustration

Including players’ heterogeneity in the game

The simple examples developed earlier show that the definition of the rules allocating network services may be central to avoid dilemmas in the choice of network use, when the services offered have different costs for the system. We may assume that some fixed mechanism (priority for flat capacity, in the example earlier) is able to allocate common resources among players, taking into account their preferences and the system costs. However, in the presence of players’ heterogeneity, this might not be straightforward. Consider a simplified model of a gas network that two players can use. The two players have different preferences (thus give different values) for flexible and flat capacity. The first player obtains a value for the flat use of network v _flat = 6 and for the flexible capacity v_flex = 10. The second player obtains a value for flat capacity v _flat = 6 and for flex capacity v _flex = 30.

Let us consider the two options for capacity allocation of the previous section.

First of all, the use of the network depends on the rules of network use. But in this case, contrary to the first examples, we are not facing only a prisoner's dilemma. Both equilibria are maximizing the welfare of the corresponding game. What these examples show is that the pipeline operator, in the process of calculating the capacity, is inducing different games. Moreover, the distribution of the social welfare is not the same in both allocation schemes.

The conclusion that can be drawn for the games with heterogeneous users is that there is one solution for capacity that makes the flat player end up better off than in the other solution. Therefore, the initial mechanism considered to allocate capacity, characterized by a pipeline operator using the capacity calculation to coordinate the use of the pipeline, cannot be done efficiently in the case of heterogeneity. The capacity calculation process defines the game that network users will be playing, and thus the rules for network use affect the use of the network.

Adding negotiation to reveal preferences

Consider a case where the final capacity calculation (of flat and flexible services) takes into account players’ preferences. To represent it, we allow players to sign contracts without any ‘a priori’ constraint on services. The pipeline operator does not know a priori about the use of the pipeline and the definition of services would depend on the negotiation of services users. For that it is necessary to create the mechanisms for players to reveal their preferences. Hence, players should negotiate (and commit) before the capacity calculation takes place. It is necessary to have mechanisms allowing communication in the contracting process. Consider that a pipeline operator is trying to allocate pipeline capacity efficiently. To do so, some bidding mechanism may be established in order to allow players to reveal the kind of service they need and how much they are willing to pay for them. This negotiation may be done through bilateral negotiations or through auction-like processes. In this game, we introduce the bilateral negotiation.

The illustration here uses the examples stated earlier (Tables 4 and 5), where the choice of the priority rules actually defines winners and losers. Our intention here is not arguing that negotiation necessarily improves welfare,Footnote ¹⁹ but to point out the changes in the game where decision makers change (from the system operator to users). The main point to highlight in this example is that sharing information ex ante may help define the optimal set of rules. The situation we describe builds on the possibility of two possible contracting options. Both contracts involve implicit commitment of the pipeline operator.

Table 4.

Heterogeneous players: Game 1

Table 5.

Heterogeneous players: Game 2

These contracts are:

1. (1) The first one gives priority to flat services and player 1 commits to 2 units of flat capacity. Hence, the equilibrium is that player 2 takes 1 flat capacity 0,5 of flexible capacity and the outcome is (12,21), as in Table 5 earlier.

2. (2) The second one gives priority to flex services and player 2 commits to one flex capacity. Hence, the equilibrium implies that player 1 also chooses 1Flex and the payoff is (10,30), as in Table 4 earlier.

Player 1 would prefer to commit to 2 units of flat services but player 2 would prefer 1 flex unit. As there is no exogenous rule that defines who wins, they need to negotiate. As player 1 knows that player 2 will try to induce the second equilibrium, she commits to an additional payment. Such payment is so that player 1 still prefers the first equilibrium, so if both players sign this contract player 1 pays to player 2, as illustrated in table 6. The payment need to be:

1. (1) High enough so that player 1 have a payoff at least as good as it was in the case of priority to flat services, and

2. (2) Low enough so that the payoff of player 2 is still higher than in the case of priority to flat services.

Table 6.

The role of negotiation

In order to illustrate the previous process, we will define a type of negotiation that may occur between player 1 and 2 in order to increase both payoffs. The first contract says that player 1 commits to choose 2 Flat and that the capacity will be that of the first equilibrium. As she knows that player 2 will try to induce the second equilibrium, she commits to an additional payment. Such payment is so that player 1 still prefers the first equilibrium, so if both players sign this contract player 1 pays one euro to player 2.The second contract says that player 2 commits to choose 1 Flat–1 Flex and that the capacity will be that of the second equilibrium. As she knows that player 1 prefers the first equilibrium, she specifies an additional payment of 3 in the case that both players sign the contract.

Finally, if players do not agree to a contract, the pipeline operator chooses the most valued of the previous equilibria. In this instance, the pipeline operator chooses the second equilibria.

This last example shows that, if players were able to negotiate their capacity,Footnote ²⁰ they would be able to find a better solution, which allows them to maximize the welfare through the redistribution gains.

In order to extract lessons from this last game, note that the first assumption of our analysis is that different players may use the transmission system, but they need to follow some rules. These rules can be defined ex ante (through rules on which kind of service has priority). But we showed in this last game that the rules might be defined and implemented by a system of contracts. This implies a definition of the rules for network use by users’ negotiation, so the pipeline operator is not involved in the definition of the game. Assuming that users have more information about the value they give to flat and flex services, they may be able to arrive in a better solution if they are able to negotiate.

But it is important to note that our aim is not representing a complete process of negotiation and renegotiation. As we highlight in the beginning of this section, we are describing a static situation that maybe not a contracting situation but, for instance, a one-shot auction. Actually, what one finds in practice (as described for the case of the US system) is a system of contracts between shippers and pipeline owners. The main characteristic we are underlining is that pipeline owners receive all shippers’ information before the decision on capacity offer is made, and with that information, pipeline capacity is decided on. Therefore, even if the pipeline owner is counter-party of all contracts, it is not deciding on the offer of capacity in the pipeline system. In that view, the situation is similar to the one described by our last game.

Lessons from the simple games earlier

The sequence of simple games stated earlier shows how the inclusion of complexities in the relationship between players may strongly affect the outputs. We begin with the simplest case, in which the services are homogenous (both cost the same as they withdraw the same amount of resource flow units) and players have the same preferences. It is easy to see that there is no conflict: both will demand the same kind of service.

In a second step, we considered that services can have heterogeneous cost. That is, they subtract different amount of units from the available resources.Footnote ²¹ In this case, as one of the services decreases a greater amount of available resources, when the demand prefers the most expensive service, the pipeline will not have enough services. Thus, one faces the need to decide a rule that decides the way in which available resources are allocated. The objective is to ensure open access, and one needs to define rules to use services. To that end, we study two basic examples of usage rules: rules giving priority to flexible use and rules giving priority to flat use. In the first case (which is priority for the more costly service), it leads to a prisoner's dilemma situation; in the second one, we end up with the design of an efficient solution. Consequently, the design of rules matters when we have heterogeneous services, but it is possible to find the efficient rule which leads all players to efficient solution.

In the third step, we considered heterogeneous players (they valued differently the available network services). In this case, each of the rules considered benefits one player instead of the other. That means that the rules actually change the distribution of profits among players, so it is not possible to define which the most efficient rule is. In that third step, we could not define the ‘best’ ex ante rule, but we analyzed the possibility that players reveal information. Using the possibility of negotiation as a simple description of information revelation, we allowed players to define the rule that should be used and the profit transfer among them. In this case, we illustrate how players could find the ‘best’ set of rules. Nonetheless, we do not include the costs associated with the ‘negotiated’ solution (e.g. transaction costs), so that it is not possible to assert in general that the negotiated solution is better than the regulated one. Our intention is just underlining that revealing players’ information may play a role in finding the best rules of network use.

The previous sequence of games is aimed at showing step-by-step the consequences of the different ways to implement open access. The set of games using ex ante rules is aimed at representing the EU approach. Actually, such definition of ex ante rules has become the model adopted by the EU countries and recommended by the EU institutions. The ‘regulated’ approach has led to the homogenization of the services offered. As shown in the games, if users’ profiles are similar, it is feasible to find a common set of rules that, applied to the group of users, allow an efficient management of resources. The game implementing negotiation is aimed representing (schematically) the US approach. It highlights the differences associated with using a negotiated process to define the rules. In practice, under the US approach, it is possible to observe a relatively large amount of contract types. This, in turn, allows different flexibility degrees and flexibility types. Network services consider line-pack utilization, nomination scheduling and geographical localization. Furthermore, it is possible to observe unbundled flexibility services, such as loan and parking services, which had not been offered just by network operators but also by traders managing infrastructure portfolios, storage or LNG regasification owners.Footnote ²² In that view, when players are allowed to negotiate the rules, it is easier for them designing solutions to deal with heterogeneity.