بسته بندی و ادغام در بازارهای انرژی

Does bundling trigger mergers in energy industries? We observe mergers between firms belonging to various energy markets, for instance between gas and electricity providers. These mergers enable firms to bundle. We consider two horizontally differentiated markets. In this framework, we show that bundling strategies in energy markets create incentives to form multi-market firms in order to supply bi-energy packages. Moreover, we find that this type of merger is detrimental to social welfare.

A trend towards deregulation of utilities industries, such as energy or telecommunications, is observed worldwide. This has an impact on market structures and pricing strategies. In particular, market structures shift from monopolies1 to oligopolies. Moreover, in energy industries, we observe multi-market mergers between firms belonging to various oligopolies. Such mergers enable firms to bundle several energies.2 For instance, they provide packages of two different energies like gas and electricity. A significant example is the merger case between E.ON and Ruhrgas on the German market. Although the merger proposal is rejected in 2002 by the competition authority, the Federal Minister of Economics and Technology even so approves the merger in 2003 (Marsden et al., 2007). Before this acquisition, Ruhrgas was the first gas producer in Germany while E.ON, the first electricity one. E.ON only owned holdings in a few local subsidiaries of gas supply. Thus, the new leader of the German energy market now supplies bi-energy bundles. We also remark that other mergers become effective on this market straight after. Indeed, electricity supplier EWE merges with gas suppliers Cuxhaven and SWB in 2003. This substantiates the merger wave phenomenon. The following question therefore becomes important: do bundling strategies trigger mergers in energy markets? This type of incentive could better explain the convergence phenomenon in energy industries. In this paper, we study the emergence of these mergers. In order to carry out our analysis, we use a horizontally differentiated model derived from Reisinger (2006). It allows to study bundle competition. This analysis can be interpreted as a modelling of a competition between two electricity firms and two gas firms. 3 We build a merger game allowing to underline a merger wave phenomenon. This phenomenon is due to the ability to supply bi-energy bundles 4 once a merger is achieved. Bundling entails two effects. The first is a price discrimination one. The second is a competition one. The trade-off between these effects and merger choices causes an increase in profits. The results which we have just evoked allow to better assess a relevant phenomenon in the energy markets: the convergence phenomenon. Usually, convergence refers to a process that reduces differences between activities. It corresponds to a gradual integration of formerly separate industries. To describe convergence in the energy industry, we analyze a specific trend: the convergence between gas and electricity.5 This trend is widely observed during the 1990s in the US and is now described in Europe too (Verde, 2008). Multi-market mergers in energy industry participate in this convergence phenomenon. Indeed, downstream mergers allow the diversification of energy supplies and clearly participate in the convergence phenomenon. For instance, the inter-industry merger Dong/Elsam/EnergiE2 (European Commission, 2006) refers to the integration between the Danish gas incumbent and Danish companies active in the electricity sector. The firm could now exploit their complementarities and supply bi-energy bundles. Another example is Gaz de France/Suez merger proposition submitted to the European Commission in 2005. This corresponds to a national merger but concerning both midstream and downstream markets. Colette Lewiner (a senior vice president at Capgemini in Paris) says that this merger could have been "a plus for competition if Suez and GDF bundle their offerings to give customers like industry better offerings, perhaps in the form of a single bill for electricity, gas and water" (Kanter, 2006). As the proposition was declined by the authorities, Gaz de France has to purchase electricity to Electricité de France in order to supply bi-energy bundles. In 2008, the Gaz de France/Suez merger becomes effective because European Commission approves it. This example suggests that bundling strategies may incite to merge. Empirical studies show that a lot of consumers6 use several types of energies (Bernard et al., 1996 and Nesbakken, 2001). So bundling strategies may be a fundamental reason for merger decisions.7 Despite the prevalence of this particular type of merger, to our knowledge they are not analyzed by the theoretical literature. The aim of this paper is to fill this gap. Before modelling the competition with bundles, we give more details about the bundling literature. Bundling refers to the practice of selling two or more goods at a unique price.8 The economic literature on bundling isolates several effects. One of the main effects is price discrimination. Bundling allows to sort consumers according to their willingness to pay. This characteristic is analyzed by Adams and Yellen (1976) for a monopoly producing two goods. In analysis dealing with specific cases, they show that mixed bundling is generally the optimal strategy9 since the correlation between the goods is negative. Nalebuff, 2004, Peitz, 2008 and Whinston, 1990 underscore the fact that a two-market monopolistic firm can deter the entry of competitors by bundling10 if the potential entrant can enter only one market. In this framework, Nalebuff shows that pure bundling is optimal. A second effect of bundling to consider is, in competitive environments, a competition effect. Anderson and Leruth (1993) analyze bundling in a complementary-goods duopoly. In their view, independent pricing is a dominant strategy in the commitment case. Economides (1993), in the same framework, shows that firms follow mixed bundling strategies in the Nash equilibrium. Firms, however, make lower profits than they do when adopting an independent pricing strategy. Armstrong and Vickers (2008) examine principally a unit-demand model where consumers may buy one product from one firm and another product from another firm under nonlinear pricing. They show that bundling generally acts to reduce profit and welfare and to boost consumer surplus.11 However, they consider an intrinsic extra shopping cost when consumers purchase each good at different locations. Thanassoulis (2007) finds that if buyers incur firm specific costs or have shop specific tastes then competitive mixed bundling lowers consumer surplus overall and raises profits. Reisinger (2006) also studies a duopoly that produces two types of horizontally differentiated goods. He analyzes a framework for which consumers buy one unit of each good with neither substitutability, nor complementarity effects created by variants choices for each type of goods. The correlation of the reservation prices is expressed by the correlation of consumers' location on each market. He shows that there are two effects created by bundling: the well-known "sorting effect" and the "business-stealing effect," which results from bundle competition. Reisinger shows that firms have an incentive to adopt a mixed bundling strategy. Nonetheless, the effect on profits is ambiguous. If the correlation of reservation prices is negative, then the competition effect dominates and the bundling strategy lowers profits. Such firms are in a prisoner's dilemma situation. On the other hand, if the correlation of reservation prices is positive, then the sorting effect allows firms to make higher profits. We use the model of Reisinger (2006) in order to analyze the impact of bundling on merger incentives. We therefore consider two horizontally differentiated markets, that are electricity and natural gas markets. As Reisinger (2006), the link between these two markets is the correlation of consumers' locations. Nevertheless, four firms are present. Two firms produce electricity, and the two others supply natural gas. In their respective markets, firms compete in prices. We build an endogenous merger game and assume that monopolization was illegal. First, we exclude the post-merger bundling strategy. Second, we remove this assumption in order to analyze the effect of bundling strategy on merger incentives. In a basic model in which bundling is not considered, we find that there is no incentive to merge. Once a merger is achieved, however, as we show, there is an incentive to adopt a mixed bundling strategy. Otherwise, the bundling strategy triggers a merger wave. Moreover, we show that relative to the correlation of reservation prices, two types of mergers are achieved. Furthermore, while Reisinger (2006) shows that there is a prisoner's dilemma, we show that the different types of mergers allow this dilemma to be removed. Finally, from a welfare point of view, we show that bundling is less harmful than Reisinger (2006) suggests. In order not to neglect merger interactions in our model, we endogenize merger decisions. In this sense, our study is closely linked to the endogenous merger literature, some of which seeks to explain mechanisms preventing mergers as the "insider's dilemma12" previously evoked in the exogenous merger model of Stigler (1950). For instance, Kamien and Zang, 1990, Kamien and Zang, 1993 and Fridolfsson and Stennek, 2005b also consider the "insider's dilemma". Moreover, Kamien and Zang, 1990 and Kamien and Zang, 1993 add auction mechanisms to take into account firms' acquisitions processes. We care about the "insider's dilemma" but without any auction mechanism. Indeed, we are not interested in surplus sharing rule. On the other hand, we did deal with other characteristics found in the endogenous merger literature, such as taking all firms' combinations into consideration. For instance, some endogenous merger models allow merger interactions to be revealed (Nilssen and Sorgard, 1998). More particularly, some models attempt to emphasize the phenomenon of preemptive mergers (Fridolfsson and Stennek, 2005a, Brito, 2003 and Matsushima, 2001). Finally, other models, such as those of Fauli-Oller, 2000 and Nilssen and Sorgard, 1998, focus on merger waves phenomena. As the same type of merger interactions are possible in our framework, we build a merger game based upon Nilssen and Sorgard (1998). Contrary to Nilssen and Sorgard (1998), we do not restrict merger possibilities in an ad hoc fashion. Indeed, the only restriction concerning merger choices is due to the prohibition of the monopolization. Some merger choices are then mutually exclusive but this is not determined in an ad hoc way. This is due to the fact that a homogeneous merger is de facto incompatible with a heterogeneous merger. The following section presents the basic model. The section 3 introduces the bundling strategy on energy markets. Section 4 gives the equilibrium of the game and the social welfare analysis. The final section presents some concluding remarks.

We observe an increasing number of mergers in the energy markets. A lot of them concern firms from different markets such as gas and electricity. Whatever the merger, it allows firms to supply bi-energy bundles. Our paper studies the bundling effects on merger incentives in energy markets. We show that bundling strategies always create merger incentives for specialized firms (electricity or gas firms). The intuitive explanation is the following. Although competition effects of a merger involving firms from two independent markets are non-existent with an independent pricing, competition effects appear if the merger allows product bundling. First, we show that there is always an incentive to follow a bundling strategy once the merger is achieved. When bundling is not possible, we find that there is no incentive to merge. From these two results, we deduce that bundling strategy generates not only a merger incentive but also a merger wave. This incentive comes from the sorting effect of the bundling strategy. However, a competition effect, which is negative for firms' profits, is also generated by the bundling strategy. We find that, in order to take better advantage of the sorting effect and to avoid the competition effect, energy firms choose between two merger types. This choice is function of the correlation of consumers' reservation prices for the two energies. An electric firm can merge with a gas firm at the same location on the other circle. An electric firm can also merge with a gas firm at opposite location on the other circle. We call homogeneous and heterogeneous mergers, respectively. These merger opportunities remove the prisoner's dilemma created by the dominance of the competition effect which is emphasized by Reisinger (2006). Our model has important implications concerning competition policy. We show that bundling strategies have a negative effect on social welfare, but in our model this effect is weaker than in Reisinger (2006). Competition authorities should pay more attention to mergers in domestic markets, such mergers may be authorized by the governments in order to promote national champions. However, we must note that our analysis don't take into consideration potential efficiency gains following a merger. A direction for future research could be the introduction of these effects. The introduction of other effects like efficiency gains could affect the results.