The double dividend of a new environmental tax reform with promotion of resource substitution

Susana Silva,[*] Isabel Soares,[†] and Oscar Afonso[‡]

Abstract

We examine the existence or not of a double dividend for a new type of environmental tax reform (ETR) were tax revenues are used to finance a renewables subsidy to extraction/production. In our model, production uses non-polluting renewable and polluting non-renewable resources. Initially, resource extraction costs are constant and in a model extension, renewables producers can invest in knowledge to reduce costs and knowledge becomes complementary to the subsidy. We show that the choice of indicators for the first and second dividends is critical to the conclusions. Focusing on emissions per output as the indicator for the first dividend and utility as the indicator for the second dividend, we show that environmental policy is desirable when compared to the “laisser-faire” situation. Additionally, for the same tax levels, it is preferable to implement the ETRwhen compared to the tax used alone. Once the government chooses to implement the ETR, the higher the policy levels, the higher the dividends. These effects are achieved through a transition to a more renewables intensive production.

JEL Classification: O13, Q20, Q32, Q58.

Keywords: Renewable and non-renewable resources; Environmental Tax Reform; Pollution; Economy; Welfare

May 2016

1

1. Introduction

Environmental problems have, over the last decades, been at the center of the political agenda worldwide. To deal with these problems, there has recently been a strong emphasis on the promotion of Renewable Energy Sources (RES). However, the present economic crisis represents the risk of sending environmentalproblems to a second place in the political agenda. Additionally, there are constraints to finance RES. Since both the economy and the environment are crucial issues, the debate about the relationship between the two is a top priority. The analysis of the evolution of emissions per output shows a continuous decrease in this ratio for most countries, as shown in Figure 1. This indicates a potential compatibility between the two aspects, being critical to know what has been done to achieve this outcome and what can be done to achieve an even better outcome in the future.

Figure 1. Emissions per output ratio

Source:from

One side of this economy-environment relationship is the fact that production uses polluting non-renewable resources. The pollution problem can be assessed using environmental policy, which sometimes harms output and also technical advances (e.g., Schou, 2000, 2002).

An important route to reduce pollution without reducing output, which is the focus of this paper, is the substitution of polluting non-renewable resources for non-polluting renewable resources. Governmental intervention in this area faces some challenges. Subsidies to renewables are common and generally accepted. However, when financed through general tax funds, such as income and labor, may generate distortions (Galinato and Yoder, 2010). At the same time taxes on polluting resources may face public opposition. Some studies consider the policy instruments as independent alternatives (e.g., Silvaet al.,2013;Abolhosseini and Almas, 2014), but do not study the advantages of using the instruments simultaneously. In particular, it is possible to, following anEnvironmental Tax Reform (ETR),[§] tax polluting resources and use the revenues (maintaining revenue-neutrality) to finance renewables.

There is a vast literature on traditional ETR. This reform implies shifting the tax burden from economic “goods” such as employment and income to environmental “bads” such as carbon emissions. This literature has been developed after the seminal work of Pearce (1991), which focuses on revenue-neutral carbon taxes.

One of the main themes studied is the existence or not of a double dividend associated with the ETR(Bovenberg, 1999). The double dividend consists of an environmental improvement (the first dividend) and an economic/welfare improvement (the second dividend). Bosquet (2000) performs a survey of the empirical evidence on this theme and concludes that, in the short term, the second dividend is very small while results are less certain in the long term.

More recently, a new generation of ETR has appeared in which tax revenues may be used to, for example, promote renewable energies, energy efficiency, distributional offsets or even fiscal consolidation(Gagoet al., 2013).

The literature on this topic is scarce and consists essentially on empirical applications and analysis. For instance, Convery (2010) shows an Irish application were carbon taxes are used to fiscal consolidation, given the economic crisis in the country. In this case, there is not a pure case of revenue recycling.Shaheen (2012) presents a proposal of the Italian government to introduce a carbon tax whose revenues would be mainly used to promote low-carbon technologies and procedures. According to FOEN (2010), Switzerland introduced a carbon tax on 2008 whose revenues are partially used to fund energy efficiency improvements in buildings.

Few authors perform a modelization of such new ETR and some of them do not label their option with that name. For instance, Galinato and Yoder (2010) consider a net-revenue constrained carbon tax (for high emitting energy sources) and subsidy program (for low emitting energy sources) which is theoretically close to an ETR. The authors find welfare gains from this program relatively to the no-tax-scenario together with emissions reductions. These authors do not study, however, the substitution between non-polluting and polluting resources since emissions result from the output of each sector and not from resources use.

Kalkuhlet al. (2013) evaluate several policy options to promote renewables in terms of their welfare (which does not account for disutility from emissions) and energy prices effects. The authors show, for example, that a RES subsidy when used alone implies significant welfare losses. At the same time, a ‘carbon trust’, i.e., a certain combination of carbon prices/taxes and a subsidy to renewables allows achievinghigh emissions reductions at relatively moderated costs. However, revenue-neutrality is not always maintained, as it is in an ETR. The negative policy results on welfare may be influenced by the fact that emissions are not included in the utility function, i.e., agents do not value a clean environment.

Finally, Bohringeret al. (2013)consider several options (a lump-sum tax, a labor tax, an electricity tax or a revenue-neutral abolishment of a coal subsidy) to finance renewables and study the employment and welfare impacts of those options for Germany. In the context of an ETR only the last two options are of interest. The authors find that,for small renewable subsidy levels, there are minor benefits while for higher subsidy levels there are substantial losses.

We contribute to this literature of a new ETR, studying the economy-resources-environment relationship through astylized analytical model which analyses the fuel-switching responses when tax revenues are used to finance a renewables subsidy.[**]In particular, we study the possibility of achieving a double divided through the new ETR, We also study the transformations in the production side/structurethat occur when the ETR is implemented. To our knowledge, this analysis has never been done before.

Initially extraction/production costs are constant for both types of resources (other authors assume costless extraction,e.g., Aarrestad, 1990; Agnani et al., 2005; Andre and Cerda, 2006). Then, in an extension of the model, we include technical change (TC) to increase the realism of the model. In particular, we consider that producers in the renewables sector can invest in knowledge to decrease extraction/production costs. In both cases, only non-renewable resourcesgenerate pollution.

Our results show that the choice of indicators for the first and the second dividend are critical to the conclusions. If we choose emissions per output as the indicator for the first dividend, and welfare/utility as the indicator for the second dividend, results show that even the tax used alone is preferable to the “laisser-faire” situation, since it decreases emissions per output and increases welfare. Hence the ETR is not indispensable to achieve a double dividend. However, for the same tax levels, the ETR achieves a better performance. Once the ETR has been adopted, the higher the policy levels, the higher the dividends. If we choose absolute emissions as the indicator for the first dividend and output as the indicator for the second dividend, then, only with the ETR there is the possibility (but not certainty) to achieve the double dividend. This happens because the tax used alone decreases both emissions and output – achieves a first dividend but not the second one.

The article is organized as follows. Section 2 presents the main features of the base model and develops its extended version, exploring the equilibrium conditions and their implications. Section 3performs a sensitivity analysis to the tax, the subsidy and the renewable resources’ knowledge stock. Section 4 shows a numerical simulation, focusing on two ratios – the renewables intensity of production and the emissions per output level - and on welfare effects. Section 5 concludes and highlights some policy implications.

2. The model

We consider a model in continuous time with three sectors: homogenous final-goods, non-polluting renewable resources (R-sector) andpolluting non-renewable resources(F-sector). The final-goods sector is perfectly competitive, but we introduce monopoly power in the resources sectors. The reasoning to assume a monopolistic structure in the resources sectors has two main arguments. Firstly, empirically, firms related to energy generation or natural resources exploitation have a certain monopoly power. Secondly, in theoretical terms, the firms need to have monopoly rents (at least temporarily) to have an incentive to invest in knowledge accumulationthat appears in our extended version.

The consumers of this economy are also shareholders of the monopolistic firms, which guarantee that profits remain in the economic system. We focus on the decentralized equilibrium. This equilibrium is defined by the path of resources allocation and prices, such that: (i) consumers and firms solve their maximization problems; (ii) Research and Development (R&D) free-entry conditions are met; and (iii) markets clear.In the extended version of the model, the endogenous growth mechanism is as follows: by decreasing the production costs through knowledge accumulation, there is a saving for the firms/shareholders in the R-sector which allows having more resources for investment and consumption. We ignore unnecessary aspects in order to highlight key features.

There is a mass [0,1] of identical individuals who own assets and there is no population growth such that all aggregate variables can be interpreted as per capita quantities.We abstract from capital accumulation and other production factors in order to isolate the effects of natural resources on the economy and the environment. Consumers have the following instantaneous utility function:

/ (1)

The consumers’ utility increases with consumption, C, which, in equilibrium is a given proportion of the output (), and decreases with emissions, E, i.e., agents value a clean environment. Marginal consumption utility is positive, UC0, but decreasing,UCC0. On the other hand, emissionsreduce utility, UE0, but at anincreasing rate, UEE0. This utility function will be used to perform a welfare analysis., where is the growth rate of any variable .

The government imposes a tax () on the consumption of F and gives a direct subsidy to R extraction/production (). The government’s budget is balanced each time and, since it is following an ETR, the tax revenues are used only to finance the subsidy to renewables:

2.1 Constant costs

Initially extraction/production costs are constant. Therefore, there is no endogenous TC mechanism. In equilibrium, output or final-goodsare used for consumption and resources production/ extraction, , whereY is the final output; is the constant extraction cost faced by the R monopolist; is the cost faced by the F monopolist; is the amount of R consumed at time t and is the amount of F consumed at time t; the final-good price is normalized to one.

Final-goods sector

There areN (n = 1,..., N) final-goods producers who face perfect competition. Each firm has the following production function:

/ (3)

where is a parameter representing the general efficiency of the economy, can be interpreted as the elasticity of output in relation to R and the elasticity of output in relation to F.

The use of fossil fuels generates polluting emissionswith a given emission factor, , : .

These final-good firms maximize the profit function:

subject to the production function, where: (i) as referred, the final-good price has been normalized to one; (ii) and are, respectively, the price paid for R and F. The first order conditions (FOCs) give the demand functions ofR and F, respectively, which may by aggregated for all economy ():[††]

/ (4)
/ (5)

The demand functions show the degree of complementarity between the two types of resources since the amount used of one resource increases the use of the other. Both resources are always necessary for production and no resource will be completely driven out of the market. However, since F consumption generates pollution, to achieve higher output without harming the environment, it will be necessary to replace F for R. This is the governments challenge in this model. To carry on with the equilibrium analysis, next, we determine theR and F supply functions.

Renewable resources sector

In this sector there is a monopolistic firm which “extracts” resources and sells them to final-goods producers. For now, this firm faces a constant extraction cost,.We do not consider scarcity or regeneration for R and, hence, we are not focusing on truly extractable resources. Those costs are not necessarily physical extraction costs, they may refer to costs of building wind parks, or dams, or electricity generation costs.

In each period, the monopolist chooses the price to maximize profits:

subject to the renewables demand function. From the FOC we havetheRsupply function:

/ (6)

As expected, the cost of using R increases with extraction/production costs, but decreases with the subsidy. Additionally, the subsidy has to be lower than , otherwise the price would be negative. It is also possible to see that , i.e., the R price decreases at a rate proportional to the increase in the subsidy.

Gathering the demand and the supply functions, we obtainR consumption:

/ (7)

Given the government’s goals of maintaining or increasing output and decreasing emissions, the subsidy will be fundamental in order to promote an increase in R use.

Non-renewable resources sector

In the F sector there is also a monopolistic firm which extracts resources and sells them to final-good producers. The behavior of this firm is similar to the one in the renewables sector.However, it faces an additional aspect which is resource scarcity.The F stock/reserves evolves according to the following motion law: . Thus, . As before, extraction costs for F, , are constant.

The monopolist intertemporal profits are given by:

Where is the discount rate. The firm maximizes intertemporal profits subject to the demand function of final-good producers, and the reserves motion law. After substitution, the Current Value Hamiltonian(CVH) is:

where is the dynamic multiplier of the Fstock, i.e., the variation in profits induced by an infinitesimal change in F reserves. The FOCs give, respectively, the F supply function and the law motion of the shadow price:

/ (8)
/ (9)

The F supply function shows, as expected, that the cost of using F increases with extraction costs and with the tax on emissions. The higher the tax level, the more expensive it will be to use F because they generate pollution. Both in the R and in the F sectors, the policy instruments are directly reflected in the cost of using the resources. However, the tax is paid by final-goods firms, hence its effect is affected by . The demand and supply functions, together, give F consumption:

/ (10)

Both the tax and the subsidy change relative prices and may be used to stimulate R use and discourage F use.From F and R expressions, we obtain the renewables intensity of production:

/ (11)

Both the tax and the subsidy, when used alone are able to increase .

Given the previous relationship, and using the ETR logic where , we obtain the following relationship between the tax and the subsidy:

/ (12)

As expected, the subsidy always increases with the tax level.

For the first dividend, we consider the emissions per output ratio:

/ (13)

Both the tax and the subsidy, when used alone are able to decrease.

For the second dividend weanalyze welfare. Since, in equilibrium, consumption is a fraction, , of output, we may write:

/ (14)

2.2 Extended version

Now, we extend the model to include endogenous TC. Here, Rfirms can invest in R&D activities to reduce extraction/production costs. The consideration of decreasing costs with the accumulation of knowledge is common in the literature since, throughout time, technical advances have made the use of resources more efficient or cheaper (e.g., Chakravorty et al., 1997; Popp, 2006).In line with Fisher and Newell (2008), we simplify the analysis by assuming that there is only TC in the R-sector. Even though this is not strictly true, considering TC for both sectors, but higher in the R-sector and lower in the F-sector, would significantly complicate the analysis without changing the qualitative results. For final-good producers everything remains as before, thus, the demand functions for Rand Fare observed.The same happens with theoptimization problem in the non-renewable resources sector. In equilibrium, output is used for consumption, resources extraction/production and for R&D investment. Therefore, the general equilibrium condition is , where is the amount invested in knowledge accumulation in the R-sector.