Carbon Tax and Energy Intensity: Assessing the Channels of Impact using UK Microdata.

• Author: Adetutu, Morakinyo O.

1. INTRODUCTION

   It is well established from previous studies that climate change mitigation will require substantial abatement of greenhouse emissions from different sectors of the economy (see Pacala and Socolow, 2004). However, because sectors differ in terms of their levels of energy intensity, the abatement effort required across different sectors would vary accordingly. (1) For instance, achieving the much-needed global emissions reduction will require significant emissions abatement in the production technologies of manufacturing plants. This is because manufacturing is a major contributor to worldwide pollution, accounting for around 20% of global greenhouse gas (GHG) emissions (See IEA, 2010, IPCC, 2014). Similarly, it accounted for around 17% of UK GHG emissions in 2015, mainly dominated by carbon dioxide ([CO.sup.2]) emissions (2) (MacCarthy et al. 2016).

   However, considering that manufacturing output is largely tradable, there are valid concerns that policy instruments aimed at curbing industrial emissions could harm international competitiveness, as well as result in job losses and plant closures (Martin et al., 2014). Consequently, the implicit challenge in designing an optimal industrial climate policy centers on the regulatory dilemma arising from the trade-off between the joint policy objectives of pollution abatement and preserving international competitiveness. This issue underscores the preference of economists for market-based policy instruments (3) in the textbook approach for designing optimal climate change policies.

   The most common type of market-based policy instrument is the Pigouvian tax (4) which can be imposed on energy units or carbon content in order to signal the social marginal cost (SMC) of pollution arising from its impact on climate change. While the negative relationship between carbon taxes and energy intensity is well-established in the literature, (5) existing studies are unable to shed much light on the channels through which a moderate carbon tax leads to reductions in energy intensity. (6) As a consequence, important open questions remain about the behavioral components that drive or dominate firm energy intensity reductions: how do firms achieve reductions in their energy intensity when they are faced by a moderate carbon tax liability? How do industrial climate policies place binding constraints on firm behavior? Are the carbon tax-induced changes in actual firm behavior consistent with predicted policy outcomes?

   In practice, there exists a range of responses by firms to a moderate tax on carbon. For instance, firms may adjust the input mix within their production technologies in response to changes in the relative price of energy arising from a carbon tax liability. Secondly, they might install new capital with lower energy-using technologies. A third alternative is that firms may pursue low carbon innovation efforts or knowledge through RandD investments that deliver efficiency improvements in existing production technologies. Furthermore, it is also possible that some firms may choose to exploit scale economies in order to absorb the shocks to energy costs due to the carbon tax.

   In this paper, we evaluate the components of energy intensity reductions arising from the UK carbon tax, using a panel of 493 manufacturing firms over the period 2001-2006. In the first place, the empirical evaluation of market-based climate policies on manufacturing is scarce, due in part to the lack of suitable microdata (Martin et al., 2014). In particular, the dearth of studies on the impact of UK carbon tax on the components of energy intensity reduction is even more severe, such that the empirical literature is unable to shed any light on the crucial policy discussions above. Using a two-stage econometric approach, we provide the first comprehensive analysis of the five components of industrial energy intensity gain (EIG) due to the UK CCL. In the first stage of our research design, we propose an energy intensity decomposition based on a stochastic energy cost frontier. In the second stage, we estimate the impact of the carbon tax on the EIG components using an instrumental variables (IV) approach that addresses the endogeneity of the UK CCL rules.

   Contrary to the much-touted idea that energy conservation through efficiency improvements is the most effective energy policy approach to tackling global emissions, we find that the dominant firm responses to the UK CCL are factor substitution and technological progress. These findings signal a need for broader energy policy objectives towards improving overall resource allocation, as opposed to the narrow objective of energy efficiency improvements.

   The remainder of the paper is organized as follows. In section 2, we provide a background and description of the UK CCL scheme. Section 3 set out our two-stage econometric methodology. In section 4, we describe our estimation strategy with emphasis on how we address the econometric issues arising from the unobserved heterogeneity and endogeneity problems within our specified models. Section 5 describes the data set employed, and section 6 contains the econometric results and study findings. Section 7 concludes.

2. BACKGROUND ON UK CLIMATE CHANGE LEVY (CCL) PACKAGE

   During the last 20 years, the reduction of greenhouse emissions from manufacturing has become a top priority of the UK energy and climate policy agenda. In the same vein, it seems that reduction in industrial energy use features prominently in the policy plans towards meeting ambitious carbon reduction targets. For instance, a recent paper by the Committee on Climate Change (CCC) stated inter alia:

   The CCC recommend the implementation of a stronger policy framework for industrial energy efficiency in order to meet the fifth carbon budget.

   (HM Government, 2016) (7)

   However, the reduction in industrial sector energy remains a source of political debate, especially as it mirrors the policymaker's dilemma between low-carbon economy and industrial sector competitiveness or job creation. (8) These considerations are at the core of the CCL package as a climate policy response to the joint objectives of reducing industrial GHG emissions and enhancing industrial competitiveness. The package, which was introduced in 2001, has two components: (i) a carbon tax component namely climate change levy (CCL), which is a tax per unit (9) of industrial fuel (10) purchased and (ii) the Climate Change Agreement (CCA), which is an alternative scheme available to preserve the competitiveness of energy intensive manufacturing plants through a reduced carbon tax liability.

   While the CCL is a straightforward tax on energy and carbon content, CCAs (11) entail negotiated energy (or energy intensity) reduction targets between manufacturing firms and the UK Environment Agency, (12) in exchange for up to 80% discount on the CCL liability. (13) The negotiated CCAs are undertaken at two levels. Firstly, 'umbrella agreements' on sector-wide energy use or [CO.sup.2] emissions targets are agreed between sector/trade associations and the environmental agency. Secondly, at the micro level, plant-level 'underlying agreements' are negotiated between firms and the environmental agency for specific energy reduction targets by the plant.

   One critical feature of the CCL package, which raises vital questions in our research design, is the self-selection of firms encountered in typical voluntary emissions abatement schemes such as the CCA. Effectively, plants under the CCL scheme are liable to pay the full carbon tax rates, whereas CCA plants receive discounted tax liabilities in exchange for binding energy reduction or efficiency targets. One the one hand, these targets (if stringent) could influence firm opt-out decisions from the CCA. On the other hand, a further compounding feature of the CCA's design is the non-eligibility of some plants for CCA participation. CCA eligibility is based on polluting activities regulated under the Pollution Prevention and Control (PPC) act of 1999, such that a manufacturing firm is eligible if at least one of its installation is engaged in a PPC activity (e.g. a blast furnace). This implies that, albeit CCA participation is voluntary, not every plant is eligible. This results in a selection endogeneity into the CCA scheme, with the implication that non-eligible plants incur the full CCL tax liability by default. In short, the design of the CCL-CCA package design embodies non-random selection (14) of plants into one of the schemes. (15)

   Turning now to the CCL component, one of its important features worth mentioning is the non-uniformity of tax rates across fuels. For instance, Martin et al. (2014) demonstrated that the carbon tax rates per unit of energy varied significantly across fuel types, ranging from 6% on coal to 10% on electricity; and approximately 17% on natural gas. These tax rates indicate, for instance, that the carbon contained in gas is on average taxed at more than twice the rate as the equivalent carbon content of coal. The variation in tax rates across fuel types is a second source of endogeneity concern since the effective or overall carbon tax rates per unit of total energy consumed is dependent on firm fuel mix. In this case, coal reliant firm pay lower carbon tax rates whereas the converse is the case for firms with relatively high levels of electricity or gas consumption.

3. LITERATURE REVIEW

   There is a substantial body of literature on the impact of carbon taxes in general. (16) However, there is a dearth of studies on the impacts of carbon taxes on industrial sectors despite their widespread adoption across advanced economies. In terms of the UK literature, a small body of literature on the CCL (17) exists, although the micro-econometric evaluation of its impact on industrial sectors is sparse. Agnolucci, et al. (2004) and Ekins and Etheridge (2006) provide empirical evidence on "announcement effect" and...