The political economy of emissions trading
L'économie politique des marchés de permis d'émission négociables
Résumé
Since three decades, the protection of the environment has become a major subject of concern in industrial societies, as shown by public polls . Thanks to the rise in their income, to the higher awareness towards health effects of pollutions and to major ecological disasters , citizens of industrialized countries have asked for more preservation of the environment and of natural resources. In their answer, governments first preferred the use of command and control instruments (norms and standards). Today, they tend to prefer the use of market instruments . This is what environmental economists have been advising for a long. They have shown, theoretically and empirically, that economic instruments are less costly for the control of pollutions such as greenhouse gas (GES) emissions (Baumol and Oates [1988]). The two principal economic instruments are taxes and tradable permits. In theory, they are equivalent in term of efficiency. In practice, they raise different problems of implementation. Tradable emissions permits systems have been used in the United States since the 80s, with significant results in term of economic efficiency and environmental effectiveness. (Tietenberg [1985]). Following American experiences, Europeans are now organizing also systems for emissions trading. The 25 member States are supposed to launch a wide market of GES emissions allowances in January 2005. It will be organized for firms pertaining to six industrial sectors and it is a key part of the European climate policy. The US market for sulfure dioxide emissions allowances remains, until today, the most sophisticated experience with emissions trading. It has served as “the reference” to promote the use of this instrument at the domestic and international levels. We shall use it in this way in this thesis concerned with markets for emissions trading. Two variables appear as central in the implementation of this instrument. The first is the environmental constraint level: the global quantity of permits created. The second is the choice of an initial allocation rule for the permits. We shall distinguish between two main rules: the sale (auction) and the free grant of allowances. Environmental economics have a lot to say about decisions on these two variables. (Tietenberg [1985]). It suggests that society reaches an optimal pollution level defined by the comparison between the marginal costs and benefits of depollution. However, this criterion is difficult to apply and as we shall see, one prefers an objective that is “socially acceptable” for the concerned parties. The initial permits allocation determines the distribution of an “ecological rent” that is equal the to market value of allowances. Following the Coase theorem [1960], the theory has shown that under restrictive conditions, the choice of an allocation rule is neutral (Montgomery [1972]). In a partial equilibrium analysis, with exogenous prices, this choice only has distributive effects, without incidence on economic efficiency. In reality, conditions of pure and perfect competition are not verified and this neutrality disappears easily. This is the case when we allow, in the analysis, the influence of distributive effects on allocation phenomena. It is also the case when we take into account equity considerations or the existence of markets distortions. The governments involved in the implementation of the European market of GES emissions trading should be concerned with the consequences of the high levels of taxation. Indeed, economists show that in presence of fiscal distortions on factor markets, the use of the ecological rent matters for the economic efficiency of the environmental policy. They suggest recommend to sell the permits and to “recycle” the receipt in the cut of other pre-existing taxes. The “ecological tax reform” is a condition ensuring that the costs of emissions abatement are minimized. The argumentation, borrowed from the optimal taxation theory (Sandmo [1975]), is based on a property of the ecological rent: it is equivalent to an implicit tax that reduces the whole efficiency of the tax system. It appears that the practice of economic instruments of environmental regulation differs from economists’ prescriptions: the constraint level is weak (Hahn [1989]) and most of tradable permits are granted for free. This is true for the European market, as it has been design until 2012. The reality of environmental policies contradicts the economic analysis’ recommendations. How shall we explain such “policy failures” as quoted by Dixit [1996]? This is the concern of this thesis. We investigate the reasons of the weak level of the environmental constraint and of the free grant of permits. Interviewed during discussions on the European system’s design, the environment commissioner, Margot Wallström declared: “Any amount of auctioning, however small, will make it even harder for businesses to accept emissions trading, […] the method of (allocating allowances) will not change the environmental benefits of trading” . This declaration underlines European policy markers’ concern to treat tactfully particular interests (industries and ecologists). It characterizes suitably this thesis subject. According to the economic analysis, markets for emissions trading are efficient. Nonetheless, their implementation faces numerous difficulties and political impediments are among the most important. The legal emissions abatement imposes costs to polluting industries, raising their opposition. This hostility is certainly amplified by the use of financial instruments (taxes and auctioned permits). Indeed, they induce payments, by polluters, that can be important in comparison with abatement expenditures. This opposition is weaker with tradable permits granted for free. Indeed, they drive transfers at the benefit of polluter and this reduces their resistance in comparison with a policy based on auctioned allowances or taxes. In an imperfect world, the ecological rent allocation matters for the economic efficiency of the environmental policy. It matters as well in term of political acceptability. It is necessary to consider the design of an emissions trading system as a set of political decisions, in a “public choice” perspective. Our thesis must start with the normative analysis of emissions trading. It will provide us with the referential necessary in order to conduct the positive analysis of policies. So as to explain the gap between what economists say and what politicians do, we could assume an incomprehension of the firsts by the seconds. We could also doubt on the results of the economic analysis if they were relying on unrealistic hypothesis. We shall assume, in the frame of this study and except from the strict point of view of the firm, that the choice between auction and free grant of permits is non-neutral. It does matters on the environmental regulation’s efficiency. We need a common denominator in order to be able to value and compare policy choices. We shall use welfare. We shall define the economic efficiency of a policy as the cost minimization, in term of forsaken welfare, to reach a given environmental objective. The normative analysis relies on a particular representation of the public decision making process. It makes the unrealistic assumption of a “benevolent dictator” (Mueller [1989 : 4]) who chooses the society’s optimal decisions. It is necessary to use a more realistic characterization of this process if one pretends to understand public choices. It is the objective of the political economy. Then we must follow the most relevant approach of the environmental regulation positive analysis. According to Portney and Oates [2003]), it is based on the study of interest groups behaviours and their interaction in a given political framework. We shall consider the modalities for permits markets’ implementation as decisions influenced by special interest groups. Two questions must then retain our attention: 1. What are interest groups’ preferences concerning emissions trading systems and the modalities of their implementation? 2. Given these preferences and economists recommendations, how can we explain the weak level of the environmental constraint and that permits are granted for free? Economists have focused most of their attention to the first question. Numerous studies are concerned with the attitude towards tradable emissions permits of polluting industries shareholders and wage earners, of ecologists and of bureaucrats. Some of them focus more specifically on the ecological rent allocation (Brandt and Svendsen [2004], Markussen, Svendsen and Vesterdal [2002] and Svendsen [2002]). The existence of interest groups opposed to the sale of permits is a necessary condition in order to explain the free grant. However, it is not sufficient. One should consider the mechanism by which certain agents’ preferences materialize in political compromises. This is the second question’s subject. There is an abundant literature on the positive analysis of environmental regulation and of instrument choices. But as far as we now, there are no such formal studies of emissions trading systems’ design and of the ecological rent allocation. The empirical paper by Joskow and Schmalensee [1998] is an exception but the authors don’t formalise the policy making process and don’t take into account the existence of interest groups. This thesis constitutes therefore an original contribution to the literature as it studies formally the incidence of political factors on the choice of an allocation rule for tradable emissions permits. We will explain why permits are granted for free even though economic analysis recommends selling them in their majority. We first show empirically that free permits, because they are a windfall gain, motivate rent-seeking behaviours. Bhagwati [1982] call them “directly unproductive profit seeking” (DUP) activities as they are wasting society’s resources. This result does not allow to conclude that auction is a superior permits allocation method as other factors must be considered as well. However, it confirms the assumption (Cramton and Kerr [1998]) that the free grant raises a political contentious on the distribution of permits. Our second contribution is theoretical. We formalize the “political” interaction that exists between the choices of an environmental constraint level and of an allocation rule for the permits. These two variables motivate shareholders’ interest (for the rent) in polluting industries. We show that a decision on one of them modifies the “desirability” of a political favour on the other. We conclude that the choice between auction and free grant of permits is not neutral as it affects the effective level of pollution through a biased policy making process. Our third contribution, also theoretical, proposes an answer to our initial question. In the political process that we have formalized, the “corrupted” government makes an arbitrage between the two variables: decisions on the global quantity of permits and the way to allocate them. He chooses in priority the free grant in order to get a higher freedom in his choice of a legal emissions level. In other words, he gives (all) the permits so as to limit the deviation of the environmental constraint in comparison with its optimal level. Our last contribution to the positive analysis of emissions trading underlines the political interaction between the behaviours of environmentalist organizations and industries interests groups. Ecologists may buy emissions allowances so as so to retire them and obtain more emissions abatements. Such a strategy may nonetheless be counter-productive in term of the environmental constraint level. In addition, it is certainly less efficient than ecologists’ lobbying for the same aim. Structure of the thesis We devote the first chapter of this thesis to the literature on the normative analysis of emissions trading systems. A first section proposes a brief reminder of the efficiency conditions of an environmental regulation. We present markets for pollution allowances and the main methods for allocating the permits. In a second section, we study the neutrality property of the initial repartition. It explains this instrument’s interest in term of economic efficiency, but its validity relies on restrictive conditions. Indeed, it has been shown that this neutrality disappears in presence of markets imperfections. We examine the empirical relevance of these situations and in particular in the perspective of the implementation of the European GHG emissions trading system. We focus our attention on the effect of pre-existing fiscal distortions as several member states are particularly concerned with this question. We present the principle and mechanisms of a “green fiscal reform” and we shall explain the concept of a “weak double dividend”. It justifies the sale of permits and the appropriate recycling of the rent. This is a necessary for the cost minimization of any emissions reduction policy. We end this chapter with an argument that justifies the free grant of part of the permits, even I presence of fiscal distortions. The objective of this measure would be to compensate the polluting industries losses of profit in order to raise the political acceptability of the environmental constraint. Economists (Bovenberg and Goulder [2001]) have shown that it would only have a weak incidence on the economic efficiency of the environmental policy. Despite this argument, one can observe that a vast majority of permits are granted for free. One must therefore looks for an explanation other than this acceptability criterion. This is the aim of this thesis that places rent seeking and political influence at the centre of the analysis. The second chapter deals with the political economy of emissions trading. In a first section, we study the reality of environmental policies and the implementation of economic instruments and we observe the existence of political failures. We define this concept using a representation of the political market. For a long time, the term political economy has been referring to the discipline economy. Today, it refers to a more restricted scoop that studies collective and political public decisions making processes. In a second section, we present the various approaches of the positive analysis of environmental regulation and of emissions trading, focusing on the ones that relies on interest groups. The models of rent competitions appear to be the most relevant in the frame of our analysis. One of the most comprehensive, according to the political economy paradigm (Rodrick [1995]), is the common agency model of politics (Grossman and Helpman [1994]). Indeed, this frame specifies public decision making processes and microeconomic foundations in order to explain interest groups behaviours. Because we place the rent at the centre of our analysis, we must first demonstrate that tradable emissions permits motivate lobbying for their distribution. In the third chapter, we present an empirical study showing that the free grant of tradable emissions permits motivates rent seeking. It is based on the US experience with sulfure emissions allowances trading that was launched in 1993, in the frame of the acid rain policy. This market was organised between regional power producers and it remains the most sophisticated experience in this area. The initial allocation rule was set as a result of discussions in the US Congress. We formalize it as the endogenous sharing rule of a rent. In order to test this relation, we measure the lobbying effort of power sector shareholders using their political (money) contributions. The results confirm the incidence of this variable on US congressmen’ decisions. During discussions of the law, the shareholders could expect that they would be able to retain part of the ecological rent despite prices and profit regulation in the power sector, thanks to a climate of regulatory uncertainty. In the following three chapters, we conduct a theoretical analysis of an emissions trading system design. A regulator under political influence must still set the constraint level falling on the collective emissions of a group of polluting industries and choose a method to allocate the allowances. This frame, in which the instrument is chosen before the constraint level, corresponds to the European case. We formalize this decisions making process using the common agency model of politics. We can thus characterize a public choice as the arbitration between the social welfare effects of a policy favour and the private benefits that are expected as a reward. In the fourth chapter, we first show that the choice of an initial allocation method for the permits is not neutral in presence of a political market distortion (lobbying). The decision of a “corrupted” government to sell the permits or to grant them for free modifies the shareholders’ incentive, in polluting industries, to pressure for or against legal emissions abatement. The equilibrium values of the permits market variables are therefore modified by the bias of the imperfect political process. We assume the existence of a unique lobby that represents polluters’ interests and that captures completely the regulator’s decisions. The fifth chapter consider this relation of political influence in presence of several industry lobbies. They participate to a partially non-cooperative game for the permits distributions. Pressure groups have common interests: they want that the government create a maximum of emissions allowance (doesn’t tighten the environmental constraint) and that he grants them for free (rather than auction). However, their interests are diametrically opposed concerning the allocation of the free permits. For that reason, they play a zero sum game of political influence. We apply the “general theory of common agency and coordination” (Digit, Grossman and Helpman [1997]) to this frame and we analyse the distributive and efficiency characteristics of the political compromise. We show that in order to satisfy pressure groups, the regulator chooses in priority to grant the permits for free, before manipulating the environmental constraint. The first solution determinates a lump sum transfer, less distorting than the second solution that modifies the marginal cost of emissions abatement. We find the standard result of the optimal transfers theory (Diamond and Mirrlees [1971]) but in the frame of an imperfect political market. This result offers an explanation to the observation that permits are granted for free. Despite efficiency appearances, this situation is the worst collective outcome for the lobbies that face a “prisoner dilemma”. By giving the emissions allowances for free, the government divides shareholders and can therefore collect all the ecological rent. This result is based on a condition of the common agency model. The competition between pressure groups is formalized as a price war à la Bertrand. Industry does not have the monopoly of lobbying for environmental regulation and it must take into account the power of ecologist organizations. Chapter 6 focuses on the interaction between industrial and ecologists lobbies. We assume that environmentalists can purchase permits in order to retire them from the market. This phenomenon happens on the US market, but with a modest width. We show that this behaviour may be counter-productive. It exacerbates the polluting industry lobbying that obtains an even less constraining level for the environmental regulation. In a second part of the analysis, we assume that environmentalists lobby the government in order to get a lower legal ceiling on global emissions. This strategy always appears less onerous than a purchase of permits resulting in the same effective level of abatement. This result is based on recent empirical studies (Riddel [2003]) showing that in the US, monetary political contributions offered by ecologists are more “productive” than those offered by other interest groups. We conclude on the necessity to take into account the political factor during the implementation of tradable emissions permits systems and on the implications for the choice of an allocation method. The importance of our results is underlined by the political process for the design of the European market for GES emissions trading. It reveals the existence of complex political interactions between industries and governments for the distribution of permits. New direction for research will be proposed.
L'Union européenne a décidé d'organiser un marché de permis pour contrôler les émissions industrielles de gaz à effet de serre. Ce choix témoigne de la popularité croissante de cet instrument de réglementation environnementale. Sa mise en œuvre nécessite de prendre des décisions sur un certain nombre de variables et deux d'entre elles sont centrales : le niveau de la contrainte collective, c'est-à-dire la quantité globale de permis créés ; la règle d'allocation initiale des droits. On distingue ici deux méthodes principales de répartition : la vente (aux enchères) et la distribution gratuite. L'analyse économique normative suggère que les choix sur ces deux variables importent pour l'efficience de la politique environnementale. Ils apparaissent également au cœur d'interactions politiques complexes révélées notamment lors du processus de transposition de la directive pour la création du marché européen. Suivant Coase [1960], les économistes suggèrent que sous des conditions très restrictives, la distribution initiale est neutre (Montgomery [1972]). Dans la réalité, cette neutralité disparaît aisément. C'est le cas en présence de distorsions fiscales sur les marchés de facteurs. Ce problème concerne particulièrement des États membres européens et les économistes préconisent de vendre la majorité des permis pour financer la baisse d'autres impôts (Bovenberg et de Mooij [1994]). Cette « réforme fiscale écologique » est la condition pour minimiser les coûts d'une politique de réduction des émissions (Goulder et al. [1997]). On constate pourtant que la mise en œuvre des politiques environnementales diffère des recommandations des économistes. Le niveau de la contrainte est faible (Hahn [1989]) et les permis négociables sont distribués gratuitement pour l'essentiel. Le marché européen, tel qu'il a été défini jusqu'en 2012, n'échappe pas à la règle. Cette thèse est une analyse d'économie politique des systèmes de permis négociables. Elle étudie les raisons des « défaillances politiques » lors de leur mise en œuvre : le faible niveau de la contrainte environnementale et l'offre gratuite des permis. En raison des transferts de richesse qu'elle engendre, l'affectation de la rente écologique (la valeur marchande des permis) importe pour l'acceptabilité politique de la réglementation environnementale. Il est donc nécessaire d'envisager l'élaboration d'un système de droits négociables comme un ensemble de décisions politiques, dans une perspective de « public choice ». Dans l'analyse normative, le processus des choix publics n'est pas véritablement spécifié. C'est une « boîte noire » qui garantie la prise des décisions optimales pour la société. Il est nécessaire d'avoir une représentation plus réaliste de ce processus pour comprendre la réalité des décisions politiques. L'approche la plus pertinente pour l'analyse positive de la réglementation environnementale, consiste à envisager le comportement des groupes d'intérêt et leurs interactions dans un cadre politique donné (Oates et Portney [2003]). C'est pourquoi nous étudions formellement les choix du niveau de la contrainte environnementale et de l'allocation des permis comme des décisions influencées par des lobbies. Cette thèse constitue de ce fait un apport original à la littérature. Nous expliquons pourquoi les permis sont offerts gratuitement quand l'analyse économique suggère d'en vendre la majorité. Le premier chapitre présente les résultats de l'analyse normative des systèmes de permis négociables et des conditions de neutralité et de non-neutralité du mode d'allocation initiale des droits. Nous consacrons le second chapitre à l'analyse positive de la politique environnementale et des marchés de permis. Nous montrons empiriquement (chapitre 3) que les permis gratuits, parce qu'ils sont un gain « tombé du ciel », motivent un comportement de recherche de la rente, source d'un gaspillage de ressources pour la société (Bhagwati, [1982]). L'étude porte sur le marché américain pour l'échange, entre producteurs d'électricité, de droits de rejet de SO2. Il a été lancé en 1993 dans le cadre du programme de lutte contre les pluies acides. La répartition des permis, donnés gratuitement, a fait l'objet de discussions au Congrès américain et ces tractations avaient les caractéristiques d'un jeu à somme nulle, profitant aux producteurs locaux d'électricité. Nous formalisons la règle d'allocation finalement adoptée comme une règle endogène de partage d'une rente. Pour tester cette relation, nous mesurons l'effort de lobbying des actionnaires de l'industrie électrique à l'aide des contributions politiques (monétaires). Les résultats confirment l'influence de cette variable sur le choix des parlementaires. Au moment des discussions de la loi, les actionnaires pouvaient espérer capturer une part de la rente écologique malgré la réglementation des prix et des profits et en raison d'un climat d'incertitude réglementaire. Ce résultat ne permet pas de conclure sur la supériorité de la vente aux enchères comme mode d'allocation initiale car d'autres facteurs doivent être pris en compte. Cependant, il confirme l'hypothèse (Cramton et Kerr [1998]) selon laquelle l'offre gratuite induit un contentieux politique au sujet de la répartition des droits. Notre contribution théorique repose sur la formalisation de l'interaction « politique » entre les choix du niveau de la contrainte environnementale et du mode d'allocation des permis. Ces deux variables suscitent chacune l'intérêt (pour la rente) des actionnaires des industries polluantes. Nous envisageons le cadre d'un régulateur ayant choisi d'appliquer cet instrument et devant encore décider sur ces deux variables. Nous formalisons ce processus de décisions dans le cadre d'un modèle d'agence commune de politique (Grossman et Helpman [1994]). Nous pouvons ainsi caractériser une décision publique comme un arbitrage entre l'effet sur le bien-être social de faveurs politiques accordées à des intérêts particuliers et les bénéfices privés attendus en échange. Nous montrons tout d'abord (chapitre 4) que le choix d'une méthode d'allocation initiale de permis d'émission négociables n'est pas neutre en présence d'une imperfection du marché politique (le lobbying). La décision d'un gouvernement « corrompu » de vendre les droits ou de les offrir gratuitement influence l'incitation des actionnaires d'une industrie polluante à faire pression pour ou contre la réduction des émissions. Les valeurs d'équilibre des variables sur le marché des permis sont alors modifiées par le biais du processus politique « imparfait ». Nous supposons l'existence d'un unique lobby représentant les intérêts des pollueurs et il parvient à capturer complètement les décisions du régulateur. Le chapitre 5 envisage la relation d'influence politique en présence de plusieurs lobbies industriels. Ils participent à un jeu partiellement coopératif pour la distribution des permis. Les groupes de pression ont en commun de souhaiter que le gouvernement crée un maximum de quotas et qu'il les offre gratuitement (plutôt que les vendre). Leurs intérêts sont en revanche diamétralement opposés au sujet de la répartition des permis gratuits et celle-ci a les caractéristiques d'un jeu à somme nulle. Nous appliquons la « théorie générale de l'agence commune » (Dixit, Grossman et Helpman [1997]) à ce cadre et nous analysons en particulier les caractéristiques d'efficience et de distribution du compromis politique. Nous montrons que pour satisfaire les groupes de pression, le régulateur choisit en priorité d'offrir gratuitement les permis, avant de manipuler la contrainte environnementale. La première solution détermine un transfert forfaitaire, moins distordant que la seconde modifiant le coût marginal de la dépollution. Nous retrouvons un résultat standard de l'analyse normative de la théorie des transferts optimaux (Diamond et Mirrlees [1971]) mais appliqué au cadre d'un marché politique imparfait. Ce résultat apporte une explication au constat de l'offre gratuite des permis. Le régulateur choisit en priorité d'offrir gratuitement les droits pour avoir une plus grande liberté pour fixer le niveau légal des émissions. En d'autres termes, il donne (tous) les permis pour ne pas devoir trop relâcher la contrainte environnementale par rapport à l'objectif socialement optimal. Notre dernière contribution (chapitre 6) rend compte de l'interaction politique possible entre les comportements de mouvements écologistes et ceux de groupes de pression industriels. Les groupes écologistes peuvent acheter des droits d'émission pour les retirer définitivement et obtenir ainsi davantage de dépollution. Ce phénomène se produit sur le marché américain. Nous montrons que ce comportement peut être contre-productif. Il exacerbe le lobbying de l'industrie polluante qui obtient que la réglementation environnementale soit fixée à un niveau encore moins contraignant. Le lobbying auprès du régulateur pour qu'il abaisse le plafond légal sur les émissions apparaît comme une stratégie plus efficace pour les groupes écologistes. Ce résultat s'appuie sur ceux d'études empiriques récentes (Riddel [2003]) montrant qu'aux États-Unis, les contributions politiques monétaires offertes par les mouvements écologistes sont plus « productives » que celles données par d'autres groupes d'intérêt.