Non-binding voluntary agreements

Abstract

In reality, most voluntary agreements (VAs) with polluters are not enforceable in the sense that no legal tools are available to enforce firms’ commitments. We examine whether such VAs are able to achieve an efficient level of environmental protection when they are obtained under the legislative threat of a pollution quota. We show that they can improve social welfare relative to legislative intervention when lobbying congress is very effective and when the polluter and the regulator do not discount future costs and benefits heavily. These findings suggest that VAs should be used selectively, taking into account sector characteristics and the degree of influence of lobbying on congress.

Introduction

In environmental policy, a voluntary agreement (VA hereafter), whereby polluting firms voluntarily commit to control pollution, has become a major policy innovation of the last decade. While the use of VAs was limited initially to a few countries (e.g., Germany, Japan), they are now used extensively around the world, particularly to deal with industrial greenhouse gas emissions and waste. The use of the term “voluntary” has long been disputed since many agreements are in fact obtained under the threat of an alternative legislative intervention. The present paper focuses on such agreements preempting legislation.

In most countries, VAs are not binding. As a result, no legal tools are available to enforce firms’ commitments. For instance, this is the case in Canada, France, Germany and the USA. To the best of our knowledge, the only exceptions are the UK Climate Change agreements and the Dutch “covenants” which are made enforceable through their connection with mandatory regulation. A few countries (e.g., Belgium) have tried to set up a legal framework to promote enforceable agreements with varying results essentially because companies are reluctant to enter into binding schemes.

Non-enforceability contributes to a widespread suspicion among observers about VAs ability to genuinely improve the environment. However, this property does not imply that compliance incentives are completely lacking. The legislative threat which initiates the voluntary commitment of polluters also promotes compliance ex post since the parties to the VA contract are all aware that, in case of non-compliance, the threat will be acted upon. But non-compliance cannot be observed immediately after the contract comes into force. Furthermore, once it is discovered, enacting a new legislation takes time. In the end, several years necessarily elapse before the non-complying polluters bear the cost of the legislation. This obviously creates adverse incentives. In particular, firms may enter strategically into VA without any willingness to comply, just to postpone legislative intervention.

In this paper, we develop a model of a non-enforceable agreement between a polluter and a benevolent regulator to address these issues. We examine whether this type of VA can lead to an efficient level of pollution abatement and how this level compares to both the first-best level and the level that might have been imposed legislatively. We make two crucial assumptions. First, the legislative threat is determined endogenously. More specifically, we explicitly model a legislative process in which the polluter lobbies a congress, thereby reducing the stringency of mandated abatement. Otherwise, a benevolent regulator would have absolutely no reason to use a VA since it could impose the first-best legislation directly. Second, we assume that the regulator can only punish a non-complying polluter by implementing the threat in the future. In this context, the polluter's propensity to comply is driven by the endogenous stringency of the legislative quota and by the rate at which he discounts the cost of future legislation.

We do not know of any previous contribution dealing with non-enforceable VAs.¹ Some work in this field has explored the role of legislative threats in triggering voluntary abatement [4], [9], [10], [14] .² But they all assume perfect compliance. This obviously influences the analysis and the results obtained. In particular, polluters do not enter into perfectly enforceable VAs to delay legislative intervention as in our case.

Amacher and Malik [1] or Arguedas [2] do not specifically deal with VAs but address related issues. They examine bargaining between a polluter and a regulator over the value of an emission standard. In contrast with the papers on VAs previously mentioned, they do not assume perfect compliance. But in contrast with ours, the standard is enforceable, albeit imperfectly. In fact, they deal with the negotiation taking place during the process of setting traditional mandatory emission standards. In this context, they analyze a situation in which the regulator is ready to accept a more lenient standard if it leads the polluter to adopt an abatement technology which reduces enforcement costs. In our setting, the regulator's gain is totally different. It enters into the VA in order to avoid a politically distorted legislative quota.

The paper is organized as follows. Section 2 provides an overview of real-world VAs. Section 3 introduces the model. Section 4 identifies the circumstance under which a non-enforceable VA can emerge in equilibrium. The analysis rests on the key property that entering into a VA is a dominant strategy for the polluter. Indeed, either the polluter enters into the VA to postpone legislative intervention without any intention to abate pollution, or it does so to comply with its commitments because the discounted cost of the legislation is sufficiently high. Accordingly, Section 5 focuses on the regulator's motives to rely on VAs. We show that the VA is more efficient than legislation in cases where lobbying congress is very effective and when polluters and the regulator do not discount the future heavily. In Section 6, we discuss the robustness of these results and present an extension of the model in which the polluter competes with a green lobby group to influence the congress.

In Section 7, we conclude and discuss policy implications, particularly for climate change policies where VAs are widespread. The key lesson is that non-enforceable voluntary schemes are weak instruments that are potentially useful when political constraints are severe. This is probably the case when regulators seek to cut carbon emissions of energy-intensive industries.