مالیات انرژی و اثر سود دو برابر در سیاست حفاظت از انرژی تایوان؛مطالعه تجربی با استفاده از یک مدل تعادل عمومی قابل محاسبه

Faced with pressure from greenhouse gas reductions and energy price hikes, the Taiwan government is in the process of developing an energy tax regime to reflect environmental external costs and effectively curb energy consumption, as well as mitigate CO2 emissions through an adequate pricing system. This study utilizes a CGE model to simulate and analyze the economic impacts of the draft Energy Tax Bill and its complementary fiscal measures. Under the assumption of tax revenue neutrality, the use of energy tax revenue generated for the purpose of reducing income tax is the best choice with double dividend effects since it will effectively stimulate domestic consumption and investment, and, consequently, mitigate the negative impacts of the distortionary tax regime. The double dividend effect is less significant, however, when the supplementary measures being used are for government expenditure. Nevertheless, all supplementary measures have effectively reduced energy consumption, which means they have delivered at least the first dividend—in the sense of CO2 emissions control. It has been verified in this study that having adequate public-finance policy measures is the key to realizing the double dividend effect.

Sustainable development as an ideal has been highly valued by many countries, as noted in the Environmental Basic Law in Taiwan: “Sustainable development refers to development that meets the needs of the current generation without damaging those of the future generations”. Consequently, developing energy policies requires vigorous assessment and considerations and must strive for a sound balance between the three dimensions of economic development, energy saving, and environmental protection. Pigou (1932) was the first to propose levying taxes (i.e., a Pigouvian tax; see below for an illustration) to redress market inefficiency created by negative externalities by suggesting that polluters should be accordingly taxed to offset their underestimated input prices. However, Pigou's study did not elaborate on uses of such environmental taxes, but assumed that such taxes would be fed back into the economy in a lump-sum fashion. Tullock (1967) argued that a Pigouvian tax could create a double dividend by internalizing external costs. In other words, by means of levying Pigouvian taxes, the issue of external costs would be redressed and tax revenue thus collected could be used to reduce other distortionary tax revenues to eventually improve economic efficiency. The notion of a double dividend originally came from Tullock (1967) but was first proposed by Pearce (1991), who maintained that the government should adopt a revenue neutrality approach to levying carbon taxes and use such revenues to reduce other distortionary taxes, which would curtail environmental pollution and reduce the distortionary costs of taxation. Furthermore, the costs derived from levying a new environmental tax would be offset by the benefit from reducing the costs of other distortionary taxes, thus creating the double dividend effect. According to environmental economics, green tax reform delivers a double dividend mainly through two types of effect: 1. Pigouvian effect (first dividend): this effect uses economic incentives as a tool to reduce pollutants produced by the polluter until the marginal external cost is equal to the pollution tax rate. This is the primary objective when levying pollution/environmental taxes. 2. Revenue effect (second dividend): the revenue collected from levying environmental taxes could reduce the inefficiency of distortionary taxes (such as an income tax or social welfare tax) on the market so as to increase household income. According to Goulder (1995a), a double dividend can be verified as long as the revenue effect exists. It does not matter whether the costs of levying a new environmental tax are positive (they must be less than the transaction cost of redirecting the lump-sum tax revenue back into the economic system) or negative (they must effectively reduce the total social cost). After the mid-1990s, however, the second effect of the double dividend was questioned by some economists, such as Bovenberg, Goulder, and Parry. Bovenberg and de Mooij (1994a) claimed that an environmental tax often worsened the problem of tax distortion even when such revenues were used to reduce other distortionary tax revenues. Parry (1995) identified the “interdependency effect” out of the second dividend effect to emphasize that the benefit derived from replacing the environmental tax with the labor income tax was no match for the deteriorating effect of the environmental tax on the current distortion. Oates (1995) maintained that most measures for increasing the polluters’ production costs might have certain unevaluated major negative impacts, and, therefore, the double dividend hypothesis was considered to be quite unreliable. Parry and Oates (1998) even argued that the double dividend hypothesis should be rejected since their model indicated that the distortionary effect of environmental taxes exceeded the tax reduction effect of labor income taxes. Parry et al. (1999) further summarized the above arguments as referring to the so-called ‘tax interaction effect’, which dictates that levying environmental taxes results in increases in the production costs of the industries concerned and subsequently leads to higher product prices, lower real income, and reduced labor supply. This means that the double dividend effect disappears when the welfare decreases due to the tax interaction effect exceeding the benefit created through the revenue cycling effect. Lomborg (2001) pointed out that the double dividend hypothesis was incorrect and that carbon taxes should be charged at a level lower than the Pigouvian tax. He went further to suggest that an adequate pollution tax be even lower unless it was used to reduce taxes with a great distortionary effect. Bovenberg and de Mooij (1994b) came up with another contentious analysis on the reasons why the interaction effect had influenced the second dividend, and maintained that levying the environmental tax lowered the real labor income, thus affecting the employment level. However, since the environmental tax has a smaller tax base than the labor income tax, the tax reduction benefit for labor income cannot fully offset the negative impact of the environmental tax on employment. Consequently, using the environmental tax to reduce other distortionary tax revenues has effectively aggravated the distortion in the tax regime rather than reducing the excessive burden. Goulder (1995b) pointed out that although substitution between taxes might be able to reduce the social cost of the environmental tax, it would not be effective in improving the overall economic efficiency, the reason for this being that the environmental tax as an indirect tax for intermediate goods has a stronger distortionary effect on the market. Hence the revenue effect of the environmental tax is above zero. In further analyzing the second dividend, Goulder et al. (1999) maintained that the second dividend was decomposed into three effects: the revenue recycling effect, the tax interaction effect, and the tax shifting effect. The second dividend might exist if the benefit to employment generated by the tax shifting effect is higher than the negative effect of the revenue recycling effect and tax interaction effect combined. The above discussions in the literature were heated but unsubstantiated since they included only those theoretical assumptions that were over simplified and did not offer any empirical evidence (e.g., the positive external effect of the mitigation of pollution on society and the bio-system, which was indeed a fundamental issue related to the environmental problem) to support their arguments. On the other hand, many empirical studies in the literature have adopted a more positive view, and they differ in the ways in which they design complementary measures for the green tax regime. Terkla (1984) maintained that the pollution tax was an efficient tax regime and that using tax revenue to reduce other distortionary tax revenues could improve welfare. Based on the US-EPA data in 1970s, the simulation analyses performed in his study indicated that the value created by replacing an individual income tax with a pollution tax ranged from US$630 million to US$3.05 billion at the 1992 level, and the amount would be increased by between US$1 billion and US$4.87 billion when a pollution tax was used to replace the business income tax. Repetto et al. (1992) believed that the environmental tax, when used prudently as a substitute for all kinds of distortionary taxes, could offset most losses of national income due to the levying of an environmental tax. Shah and Larsen (1992) used a static partial equilibrium model for their stimulation study. By assuming that a carbon tax of US$10 per ton was imposed in both developed countries (the US and Japan) and in developing countries (India, Pakistan, and Indonesia) with a concurrent reduction in the capital income tax, the results showed that the welfare level improved in most countries, with more significant improvements in developing countries where profit tax rates were already high. Barker et al. (1993) used a multi-sector model to study the effects of levying a carbon tax (with the target of a 15% reduction in CO2 emissions by 2005) in the UK and Europe. The research findings showed that gross domestic product (GDP) decreased by 0.4% against the base-year when the revenue was used to reduce the deficit; it increased by 0.1% when used to offset the income tax; and it also increased by 0.2% when used to offset the value-added tax. Jorgenson and Wilcoxen (1993) used an MF-12 (Energy Modeling Forum 12) dynamic general equilibrium model to examine the effects of a carbon tax on energy and an environmental economy, and their findings indicated that higher carbon taxes would be required to reduce CO2 emissions by large margins. Furthermore, higher carbon taxes, while significantly boosting tax revenue, also resulted in GDP losses, but such a negative impact could be mitigated if the carbon tax revenue was used to offset other distortionary taxes. Mckitrick (1997) adopted a static CGE model to empirically verify if the carbon tax regime in Canada gave rise to a double dividend effect. The research findings showed that, assuming that CO2 emissions in 2000 were to be turned back to their 1990 levels, using a carbon tax to reduce other distortionary taxes would lead to a significant drop in CO2 emissions as well as offset a drop in output and a welfare loss in all sectors. The conclusion was that it was likely that a double dividend did exist, and that a carbon tax would reduce the welfare level by 0.3% and GNP by 0.8% when its lump-sum revenue was channeled back into the economic system, but that the welfare level would remain unchanged and GNP would increase by 0.6% when the carbon tax was used to offset the payroll tax. Bye and Bruvoll (2008) pointed out that, in general, the effect of an instrument depends heavily upon the elasticities of the demand and the supply sides of the markets. The sequence of the introduction of the instruments changes the effect of the instrument itself. This increasing complexity calls for theoretical and empirical research on efficiency over several simultaneous instruments of energy taxes/subsidies. In spite of the double dividend of green tax reform remaining a contentious issue, as indicated by the literature review in this study, European countries, including Denmark, Finland, the Netherlands, Norway, and Sweden, as early as the 1990s launched the green tax reform that reduced the usage of goods that pollute or damage the environment through the levying of a tax and, at the same time, alleviated the burden of social welfare taxes or reduced other distortionary revenues in the hope of benefiting from the double dividend effect of both environmental protection and economic development. Years after such green tax policies were implemented, it is undoubtedly evident to the world that both their economic development and public welfare have improved as a result. Consequently, the mode of green tax reform, especially in terms of the various combinations of complementary fiscal measures, will have various effects on the industrial structure and on different economic factors, such as economic growth, the price level, employment, tax revenue, and CO2 emissions control. In this study the ORANI-recursive dynamic (ORANI-RD) type general equilibrium model developed by Monash University has been adopted as the basis for incorporating relevant modules such as the public finance and public R&D investments to develop a computable general equilibrium model for investigating the double dividend effect of energy tax and its complementary measures in Taiwan. The ORANI model was first developed in the 1970s by Dixon et al. (1977) under the IMPACT project sponsored by the Australian government. ORANI-G, a simplified version of the ORANI model, is a suitable choice for developing a general equilibrium model for other economies. It comes with a dynamic version, or the ORANI-RD model that employs a recursive dynamic method to find solutions. As for the energy dynamic general equilibrium model: EnFore (Energy Forecast)–CGE in this study, the sectors were broken down into 21 sectors and the products were broken down into 48 commodities in Taiwan (see Table 4 for 21 sectors). This paper is divided into five sections and begins with the introduction in Section 1. Section 2 introduces the structure of the EnFore–CGE model. Section 3 elaborates on structure of the data used in the EnFore–CGE model. Section 4 describes the policy simulations and the findings regarding the economic impacts of the energy tax and its complementary measures. Section 5 provides the concluding remarks for this study.

This study has adopted a dynamic EnFore–CGE general equilibrium model, along with I–O Tables and the fiscal tax database complied by DGBAS and DOF in Taiwan, to simulate the effects of a proposed energy tax regime and its complementary fiscal measures on the economy, energy consumption, and CO2 emissions. The empirical findings under the basic energy tax scenario in the model have indicated that the mere use of an energy tax policy will have negative impacts on economic variables such as real GDP, the real wage rate, and employment, as well as lead to increases in inflation year-on-year. The outcomes will be different, however, when the energy tax regime is implemented along with other supplementary fiscal measures. The measure of reducing both business and individual income taxes, due to its effect of encouraging investment and consumption, has not only offset the negative impact on real GDP but has even led to positive GDP growth. It is also noteworthy that all types of energy consumption have consistently decreased under all scenarios, which indicates the positive effects of the policy in terms of energy saving and environmental protection. In other words, the first dividend of energy saving and environmental protection as well as the second dividend of positive economic growth have been found in Scenarios 2 and 3, which implies that a double dividend effect is achievable under the tax revenue neutrality principle. When the subsidies for public transportation and public R&D investment are used as the supplementary fiscal policies, however, only the first dividend of effective suppression of energy consumption can be achieved, while the negative impact on economic growth in the cost dimension exceeds the positive effect of income tax reduction. Hence the second dividend effect is absent or not complete. From the industry perspective, due to the substitutive relationship or output effect between industries, energy taxes will have negative impacts on energy-intensive industries such as the petrochemical, oil and coal, non-metallic, metallic, water, electricity, and gas industries, as well as transportation services. When considering the net business tax burden, however, energy-intensive industries have still been able to enjoy the benefits of a lower tax burden under all scenarios. This indicates that an energy tax regime devised under the tax neutrality principle does not impose extra net tax burdens on the energy-intensive industries. This will ameliorate the objection from those industries. Among all the scenarios included in this study, reducing the individual income tax as a complementary measure has had the greatest double dividend effect, followed by reducing only the business income tax and reducing both the individual and business income taxes. It has been found in this study that levying an energy tax may have negative impacts on the output of energy-intensive industries. The situation is inevitable amidst the trend of ever-increasing energy prices and their costs. An energy tax regime coupled with a business income tax refund has, however, led to a decrease in the net tax burden for energy-intensive industries. It is therefore suggested that the authorities should design appropriate complementary measures that can generate the double dividend effect, so as to achieve the objective of green tax reform, improve public wellbeing, and strive for sustainable development by striking a balance between economic growth, energy saving, and environmental protection. To sum up, the empirical evidence in this study shows that levying an energy tax does not necessarily lead to a negative economic impact. The actual effects depend on whether or not adequate complementary measures are in place. This study has offered new empirical evidence for the various possible impacts of a green tax policy and its complementary measures, and the findings indicate that the argument of a green tax reform is determined not by its compliance with the theory of internalizing externalities but by the design of the complementary fiscal measures.