Taxing air pollutants and carbon individually or jointly: results from a CGE model enriched by an emission abatement sector

ABSTRACT

We analyse the separate and collective impacts of emissions taxation to understand the internalisation effects of externalities. The analysis is carried out using a static computable general equilibrium model, with unemployment, bottom-up abatement technologies represented by a step function, and detailed emission coefficients. Environmental and health external costs are quantified using the ExternE’s Impact Pathway Approach. Emissions, as a result of environmental taxation, fall through reduced output, production factor substitution, and increased end of pipe abatement activity. The analysis shows that a full internalisation of environmental externalities can result in modest overall economic and environmental welfare gains. There are, however, differences in terms of employment and output, depending on what combination of taxes are applied, which sectors are covered, and how fiscal revenues are redistributed. Air quality benefits range from €35–75 per ton of CO₂ abated. Total environmental benefits always exceed GDP loss and the associated welfare loss.

KEYWORDS:

• CGE modelling
• carbon taxation
• air pollution charging
• environmental benefits
• bottom-up abatement technology

Acknowledgement

We are grateful to Fusako Menkyna-Tsuchimoto for her contribution to the initial version of the paper. Finally, we are grateful to anonymous referees for many useful comments that have improved the paper. Responsibility for any errors remains with the authors.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Olga Kiuila https://orcid.org/0000-0002-4622-0928

Anil Markandya http://orcid.org/0000-0001-6304-7103

Milan Scasny https://orcid.org/0000-0001-5040-1281

Notes

1 In 2005, the share of renewable resources were only 4.3% on TPES, dominated mainly by biomass (3.9%) and large hydro power plants that share has been stable at 0.4% of TPES since 19990s till now. This is the reason why other renewable resources are not included among the factors. Share of RES has been increasing in the Czech Republic since 2009 (see Rečka and Ščasný 2016) and reached 14.2% (biomass, wastes), 0.5% (hydro) and 0.9% (geothermal, solar, wind) of TPES in 2015 (OECD/IEA 2017).

2 Available as Supplementary material.

3 The model set the 2005 unemployment rate of 8% as a baseline. The unemployment rate and labour market conditions in the Czech Republic are historically stable. Even when the crisis hit the economy (2008–2009), the unemployment rate barely reached 10%. Prior to the crisis and more recently the Czech economy experienced an unemployment low of around 4–6%. Thus we have not calibrated our model at any extreme value of unemployment.

4 An alternative technique in the literature is to fix the nominal wage (Yin 2002). We preferred to use the Rutherford and Light technique, which is also the more popular (see Partridge and Rickman, 2010; Kuester et al., 2007; Bhattarai, 2008) as it provides the possibility of unemployment if the demand for labour (which is determined according to profit maximization conditions) is less than the available supply at a gross real wage.

5 The data comes from ‘RAINS’ – the bottom-up model developed by IIASA (Amann et al., 2004).

6 Previous models have used other representations of energy technologies to track changes in emissions. One is through a soft link (Kumbaroğlu and Madlener, 2003; Vrontisi et al., 2016), another through a hard-link (Helgesen, 2013), and a third through the integration of an emission-extended bottom-up energy system model and a top-down economic model (Böhringer and Rutherford, 2008). None of them, however, included end of pipe abatement into their modelling framework. We further approach the abatement technologies by explicitly following an activity analysis approach as discussed in Kiuila and Rutherford (2013a).

7 While the equivalent variation depends directly on households demand level only, GDP measures total value of output at producer price, i.e. a value-added approach is applied, and it depends on value of output, value of intermediate consumption, and net indirect taxes.

8 The term ancillary benefits refer to those secondary or side effects of mitigation policy on problems that arise subsequent to the proposed mitigation policy (IPCC, 2001). Reductions in local air pollution associated with less use of fossil fuels due to greenhouse gas (GHG) mitigation are referred to as ancillary benefits. Conversely air quality improvement measures would generating reductions in GHG emissions, making the latter the ancillary benefit of air quality policy. In contrast, following the definition by IPCC (2001), the term “co-benefits” refers to the non-climate benefits of GHG mitigation policies explicitly incorporated into the initial creation of mitigation policies. Thus the term co-benefits reflects that most policies designed to address GHG mitigation also have other, often at least equally important, rationales involved at the inception of these policies”.

9 There has been long-standing discussion on economic standing – whose benefits and costs counts in the benefit-cost analysis, see, for instance, Gayer and Viscusi (2016). We account for all benefits regardless who is the recipient of the damage.

10 The EcoSenseWeb1.3 tool was developed within the EU-funded NEEDS project (Preiss and Klotz 2008), recently its Web2 version is being developing; see http://ecosenseweb.ier.uni-stuttgart.de. The loss due to increased mortality is estimated using the Value of Life Year (Desaigues et al. 2011) and Value of Statistical Life, while increased morbidity is valued by willingness to pay and cost-of-illness values corresponding to each health outcome (Máca et al. 2017). Crop losses are valued by the international market prices. The impacts on building materials are assessed using replacement and maintenance costs; the assessment of biodiversity impacts is based on restoration costs (see Preiss and Klotz 2008).

11 A sensitivity analysis shows that the model is sensitive to the elasticities of substitution: the higher the values of elasticities of substitution, the lower the cost of environmental policy. We also perform sensitivity analysis based on ${\sigma }_i^A=4$ for all sectors. The results are qualitatively same as for the base assumption, except the effect on GAS sector that is still positive but much smaller in magnitude (it is about five times smaller for carbon tax and by about a half for a combined policy).

12 Since emissions of air quality pollutants stemming from mobile sources are controlled by technology standards, such as EURO standards applied on vehicles in the EU, it is not realistic to assume this tax is imposed on mobile sources.

13 The rates of carbon tax correspond to a carbon price estimated by the European Commission for a 20% or a 30% emission reduction target (EC 2010), equal to €17 or €30 per tonne CO₂ respectively. These rates also cover a range of marginal abatement costs as reviewed by Carraro and Faveli (2009), corresponding to the estimates of the social cost of carbon. See, for instance, a review by Tol (2009).

14 NOx has been chosen as representative of the local pollutants. Similar results hold for other local pollutants.

15 Note the effect on unemployment is an increase on a per cent figure, so if unemployment was 8% in the benchmark, the actual increase of 3.8% change is +0.31% of labour force (resulting in the rate of 8.31%).

Additional information

Funding

This research has received funding from the Czech Science Foundation (GA CR) Republic 18-26714S (Ščasný), and from the H2020-MSCA-RISE project GEMCLIME-2020 GA No. 681228 (secondment). Previous versions of the model were built within research grant No. SPII/4i1/52/07 MODEDR funded by the Ministry of the Environment of the Czech Republic and the Czech Science Foundation P403/11/2494.