Assessment of the potential impacts of a carbon tax in Chile using dynamic CGE model

Abstract

Carbon taxes have been proposed as a major instrument to mitigate carbon emissions and promote an energy transition to low carbon sources. However, its adoption remains politically challenging, particularly amid rising inflation and energy prices. Despite the need for more aggressive action on carbon mitigation to reach the Paris Agreement goals, few countries in Latin America have adopted carbon taxes and the tax levels are relatively low. A key concern for these countries, is to adequately assess the tradeoffs between stricter emission goals and the potential negative economy wide as well as sectoral and distributive impacts. In this context, in this paper we first propose a step by step approach to enhance an existing dynamic Computable General Equilibrium (CGE) model for Chile based on OECD's Green model. The contribution of this research is twofold. Firstly, emission factors are estimated and the development of the electricity sector is aligned with the expectations of decision makers. As a result, credible emission and energy sector development forecasts are generated by the model, that are in line with what policymakers expect to happen based on other bottom-up engineering models. Secondly, this baseline is then used in the CGE model to examine the use of a carbon tax to reach Chile's first Nationally Determined Contribution. The required tax level is determined together with CO2 emissions and the economywide, sectoral and distributive impacts. The results allow concluding about the applicability of carbon taxes and possible complementary measures.

Introduction

Growing concerns about climate change has forced the countries around the world to take decisive action plans to reduce their greenhouse gas (GHG) emissions significantly. Among the latest action plans, nationally determined contributions (NDCs) have become the most important instrument for the implementation of the Paris Agreement which contains national climate action plans of each country to reduce national emissions and adapt to the impacts of climate change. In the context of Chile, in 2020, the Chilean Minister's Council for Sustainability issued a favorable pronouncement to update the country's NDC to the Paris Climate Agreement. In this pronouncement, and to mitigate emissions, Chilean Government proposed a GHG emissions budget not to exceed 1100 Mt CO2eq, between 2020 and 2030, with GHG peak emissions by 2025, and reaching a GHG emissions level of 95 Mt CO2eq by 2030 (The Chilean Government, 2020). It also proposes that this goal is a midpoint on the path to carbon neutrality by 2050, established in the Framework Bill on Climate Change. Another important point considered in the proposal is the incorporation of a specific social pillar on fair transition and sustainable development as clear proof of the close ties between climate matters and social/environmental issues. How to adequately advance in these objectives and assessing the economic, social and environmental impacts of alternative policy instruments, in particular emission taxes, is a key concern for policymakers.

To achieve climate goals, countries face challenges that require a mix of strategies, policies and regulations specific to each country context. In the literature, policy instruments are generally classified into two categories: economic or market-based, and direct regulation or non-market-based (Hille et al., 2020; Wang et al., 2019). In particular, carbon pricing initiatives (CPIs) that are economic instruments have been pushed as a powerful tool to advance towards low emission paths. As of April 2022, there are 68 CPIs operating globally, among them 37 carbon taxes and 34 emission trading systems (ETS) (World Bank, 2022). However, carbon prices are too low to reach desired Paris agreement goals (ibid, p 20). In Latin America only Uruguay, Argentina, Colombia and Chile have CPIs operating, all of them taxes, and with the exception of Uruguay, with values of US$5 or below (ibid, p.26). This study, in line with many other relevant literature (Vogt-Schilb et al., 2019) proposes that developing and maintaining carbon pricing approaches requires a strong emphasis on ensuring carbon pricing is fair, inclusive, and well communicated. It is necessary to adequately assess the expected systemic impacts as “recent economic instability, volatile energy markets, and rising energy prices exacerbate the political challenges for policymakers” (World Bank, 2022, p.8). In Chile, over the last decade, both economic incentives and direct regulation instruments have been used. Among the former, a tax on large pollution sources of US$ 5 per ton of CO2 emitted into the atmosphere, and a tax on light vehicles based on emissions of NOx, applied as of 2018 (Mardones and Flores, 2017). In 2019, a sovereign green bond has also been established, issued by the Ministry of Finance for a total of US$ 2377 million, which aims to finance green projects in the fields of transportation, energy efficiency, renewable energy and others. In mid-2019, the Chilean government announced its plans to close 8 (out of 28) older coal-fired power plants by 2024, and the rest by 2040, a typical command and control measure (O'Ryan et al., 2020; Nasirov et al., 2020).

Although many countries have adopted policies aimed specifically at curbing greenhouse gas emission (GHG), as discussed below there is a clear evidence of the trade-offs and synergies between these policies and other economic and societal goals. For this reason, there are a growing number of modelling approaches in the literature that allow for the systematic assessment of the consequence of policy instruments. A usual distinction is between bottom-up sectoral or engineering models and economic models, in particular top-down computable general equilibrium models. The former is more popular for simulating energy sector development and generally include detailed analyses of energy technologies with both technical and economic parameters. CGE models have a stronger microeconomic foundation and allow examining economy-wide interactions among the different economic agents, explaining the long-term macroeconomic, sectoral, environmental, and social impacts associated with specific shocks and future policies. The use of CGE models has a long history worldwide, and in the last decade there has been a growing number of publications that use CGE models to examine the impact of the applying environmental policy instruments. Some CGE models are: MMRF-Green, MONASH-Green and environmental ORANI-G for Australia (Adams and Parmenter, 2013; Meng et al., 2018; Nong et al., 2017); UPGEM for South Africa (Bohlmann et al., 2015); GTAP-E, GTA-PE-Power, ENVISAGE, GTEM, G-Cubed and OECD ENV-Linkages for the Global Economy (Chateau et al., 2014; McDougall and Golub, 2007; McKibbin and Wilcoxen, 1999; Nong, 2019; Pant, 2007; Peters, 2016; van der Mensbrugghe, 2008) and ECOGEM-Chile for Chile (O'Ryan et al., 2020; Nasirov et al., 2020). Over the last decade, numerous CGE studies published on climate change policies with the main focuses on the carbon price mechanism vary depending on the participating industries and regions. Therefore, most of the studies focuses on the major polluting countries, such as the United States, China, the European Union, Australia, Canada, and some other developing countries (Nong, 2019; Yu et al., 2018; Zhu et al., 2019).

To assess the medium- and long-term impacts of carbon taxes and associated tradeoffs, it is useful to apply dynamic CGE modelling, an approach that allows including the expected evolution of a many variables over time, including population, labor productivity, different growth rates, penetration of new technologies, public polies and others (O'Ryan et al., 2020; Yu et al., 2018; Zhu et al., 2019). This approach permits handling complex interactions between different sectors of an economy and capturing the feedback effects of policy and technological changes over time (Bohlmann et al., 2015). The most common approach used is recursive dynamic modelling where agents are assumed to be “myopic”, however forward-looking dynamic stochastic CGE models have also been developed.

A carbon tax would increase the cost of emitting CO2, creating an economic incentive for economic agents to reduce their emissions (Takeda and Arimura (2021); Devarajan et al., 2011, Malerba et al., 2021). This would lead to a decrease in CO2 emissions over time relative to the baseline scenario as agents shift towards lower-carbon alternatives and become more efficient in their use of energy. However, there are tradeoffs. A carbon tax would increase the cost of some economic activities and reduce competitiveness, particularly for industries that rely heavily on fossil fuels. This could potentially lead to job losses, economic decline of specific economic sectors, or a reduction in economic growth, especially if the economic sectors that may benefit such as clean energies, do not offset the negative impacts. Additionally, there are concerns about the fairness of the tax and its impact on low-income households or low wage workers. Also, a given carbon tax may not be sufficient to meet emission reduction targets, and it would have to be increased, augmenting the expected impacts. To date, there has been a growing body of research that has analyzed the performance of carbon taxes and their potential effects in various countries using CGE models. Some of these studies examine the impacts of carbon taxes on CO2 emissions reductions using a recursive dynamic CGE model (Kim and Lim, 2021; Cao et al., 2021), double dividend (Presley and Lin, 2018), employment levels (Devarajan et al., 2011), total factor productivity (Zhu et al., 2018), and energy efficiency (Wang et al., 2022), and the effect on technological progress (Chang et al., 2023). Another set of studies focuses on the examination of different carbon tax policy designs across countries (Malebra et al., 2021) and the comparative effectiveness analysis of carbon taxes in reducing CO2 versus other mechanisms (Takeda and Arimura, 2021; Jia and Lin, 2020). Shi et al. (2019) develop a recursive dynamic CGE model to examine the potential impacts of various carbon tax conditions on the energy consumption of the construction sector and the macroeconomy of China. They find that a carbon tax of 60 RMB/t would be appropriate to achieves the emission reduction target and also minimizes the negative impact on the macroeconomy. Takeda and Arimura (2021) use a forward looking dynamic stochastic CGE model to study the effects of environmental tax reform in Japan and clarify which types of such reforms are the most desirable. Their simulations show that environmental tax reform tends to generate more desirable impacts than a pure carbon tax. The effects of carbon pricing on the Korean economy Is examined by Kim and Lim (2021). They show that it is desirable not to exempt the power sector from carbon pricing in order to avoid a severe burden to the economy in terms of real GDP and real consumption. In another study, Devarajan et al. (2011) employs a CGE model to examine the impact of a carbon tax in South Africa. They find a modest effect on South African welfare and employment levels, followed by sales taxes on energy commodities and on pollution-intensive commodities. Although Chile passed a first carbon tax bill in 2018 and plans to set more ambitious objectives starting in 2030, no studies have been found focusing on the potential economy-wide and environmental impacts of these reforms using sophisticated modelling approaches such a CGE modelling. We acknowledge that it is of vital importance to systematically assess potential effects and the trade-offs ex-ante of levying a carbon tax within the framework of socio-economic and environmental objectives of the country.

Despite the extensive use of CGE modelling for climate change policy studies, from a methodological perspective, two important issues have received little attention and are rarely subjected to thorough investigation in the literature. One key issue in CGE modeling is related to the emission factors used by these models, that are not usually fine-tuned to adequately mirror engineering model projections of total and sectoral emissions. Most CGE models in the literature lack local industry-based emission factors, adopt these factors as given from other international studies and/or use emission factors that are not updated. One of the constraints for conducting a better analysis of emission scenarios for a particular country is the lack of an emission database obtained from local sources related to air pollution and the environmental and health impacts they cause (Fæhn et al., 2020). This hampers the possibility of adequately estimating the baseline year emissions and obtaining values sufficiently similar to the available national emissions inventories, as well as credible emission trajectories for each economic sector. Consequently, there generally exist a mismatch between engineering-based projections used by policymakers and the CGE simulated trajectories. As a result, CGE modelling results are less reliable for policymakers. This is usually overcome in CGE studies by not explicitly comparing with engineering model projections, or by reporting only the changes due to policies, rather that the absolute values. Improving the alignment between local industry-based databases and CGE model emission parameters would enhance the capacity and effectiveness of CGE models, by generating emission scenarios credible for policymakers.

A second important issue is that most studies that focus on environmental policy assessments using dynamic CGE models lack robust baseline scenarios for key economic sectors, i.e. reference paths that are credible for policymakers. In fact, since CGE models are based on market variables, they do not explicitly represent physical processes about climate, energy transformation and emissions. This in contrast to dynamic sectoral-focused models or bottom-up engineering models that allow for forecasts of energy and resource sector development that correspond better with expert intuitions. Better baselines in energy and resource sectors in dynamic CGE models, i.e. with pathways over time that relate to policymakers projections, serve two purposes (Delzeit et al., 2020; Krook-Riekkola et al., 2017). First, they provide insights about plausible and internally consistent projections of future trends for key economic sectors in absence of policy changes or other shocks. Secondly, if constructed correctly, these possible pathways can be more credible for policymakers. Building sectoral-focused dynamic baseline scenarios by linking CGE to sectoral models provides an additional quality guarantee and credibility to CGE-based assessments (Raul et al., 2020; Nasirov et al., 2020). However, linking baseline scenarios has been limited in the CGE modelling community due to a lack of tradition of baseline comparison exercises, high cost of implementation and the lack of transparency of sectoral baseline modelling.

In this context, and given the research gaps discussed above, in this paper we propose a step by step approach to modify the ECOGEM-Chile, a recursive dynamic CGE model based on OECD's GREEN model (OECD, 2006), to align the expected baseline emission and energy sector development trajectories with other relevant forecasts. We then use the upgraded model to examine the impacts of a CO2 emission tax that allows reaching and maintaining Chile's Nationally Determined Contribution (NDC) at 95 Mt CO2eq up to 2050. As a result, this research makes a twofold contribution. First, it presents in detail how to improve the modelling capacity of an existing CGE model, through two specific interventions that allow it to replicate baseline results that relate to what is expected by policymakers. The first intervention involves improving the emission estimation capacity of the model based on available country specific data. For this GHG emissions factors specific to each economic sector are determined, based on a database of sectoral emissions and a current GHG emission inventory. The second intervention involves incorporating in the CGE baseline scenario the expected changes in the energy generation matrix that is currently moving strongly towards nonrenewable energy sources, specifically solar and wind power. These technological changes must be correctly incorporated in the ECOGEM-Chile model, based on forecasts from sectoral engineering models. The second contribution is determining the tax schedule required to reach and maintain Chile's NDC and quantifying the economy-wide effects of different carbon tax scenarios and the associated economic, environmental and social trade-offs vis-a-vis the expected baseline.

The rest of the paper is structured as follows: Section 2 provides a review on the description of ECOGEM-Chile model; Section 3 outlines the methodology framework which includes estimation of emission factors and construction of baseline scenarios; Section 4 discusses the analysis of the simulation results; and Section 5 highlights concluding remarks.