انتشار کربن محدود شده و حمایت R & D هدایت شده؛ تجزیه و تحلیل تعادل عمومی اعمال شده

We analyse welfare effects of supporting general versus emission-saving technological development when carbon emissions are regulated by a carbon tax. We use a computable general equilibrium model with induced technological change (ITC). ITC is driven by two separate, economically motivated research and development (R&D) activities, one general and one emission-saving specified as carbon capture and storage (CCS). We study public revenue neutral policy alternatives targeted towards general R&D and CCS R&D. Support to general R&D is the welfare superior. However, the welfare gap between the two R&D policy alternatives is reduced with higher carbon tax levels. For sufficiently high levels of the carbon tax equal subsidy rates are preferred. Research Highlights ► Support to general versus emission-saving technological development influence welfare. ► Induced technological change is driven by general and emission saving R&D. ► Support for general R&D is welfare superior. ► The welfare gap between the R&D policy alternatives falls with higher carbon tax. ► For sufficiently high levels of the carbon tax, equal subsidy rates are preferred.

In several European countries there is substantial governmental support towards the development and implementation of new environmentally sound energy technologies such as emission-saving carbon capture and storage (CCS) and new renewable energy sources.1 Technology policy can be a costly approach, however, if it is used as a substitute for, rather than complementary to, an emission-reducing policy (Jaffe et al., 2005). In our policy analysis, we address governmental support for the development of CCS technologies by investigating innovation policy reforms in the presence of a first-best emission-reducing policy. We ask two questions: 1) What are the economic welfare effects of distributing a given amount of innovation support to the development of general technologies compared to the development of CCS technologies? 2) How will the economic welfare effects depend on the carbon tax level? First-best carbon policy is characterised by uniform carbon pricing that equals the marginal environmental damages of carbon emissions which are independent of source, while first-best arguments for subsidising innovation activities are imperfections in the research markets. Examples of such imperfections are external spillovers from previous research and development (R&D) activities, learning externalities and other market imperfections that make the level of R&D effort too low,2Romer, 1990 and Jones and Williams, 2000. The literature on efficient carbon policies when emission-saving technological change is present has until recently mainly disregarded the innovation policy issue and concentrated on second-best optimal carbon policy design and the carbon policies’ influence on the timing and direction of technological change ( Goulder and Schneider, 1999, Popp, 2004, Popp, 2006b, Nordhaus, 2002, Rosendahl, 2004, Otto et al., 2007, Hart, 2008 and Gerlagh, 2008). In general, the second-best optimal carbon tax is higher than the first-best rate if positive learning or other innovation externalities are present. The second-best optimal carbon tax may differ between different end uses and should be largest for the technologies with largest innovation externalities ( Rosendahl, 2004). Gerlagh et al. (2007) find in a Romer (1990) based R&D model that in the absence of explicit innovation policies the second-best optimal carbon tax should be higher than the first-best level when an abatement industry is developing. More recent analyses, such as Kverndokk and Rosendahl (2007), find that a first-best subsidy for the adoption of different competing technologies should be larger for technologies that are newly adopted and where the learning effects are large. A falling time path of first-best innovation subsidies is supported by Gerlagh et al. (2007) in their R&D based model with a finite lifetime of patents. Otto et al., 2008 and Otto and Reilly, 2008 analyse second-best and third-best combinations of carbon and innovations policies to reach a domestic carbon emission target. In a Romer (1990) inspired CGE model, Otto et al. (2008) find that it is second-best optimal to subsidise non-emission intensive R&D, combined with differentiated carbon pricing. The different policy alternatives in Otto et al. (2008) are not revenue neutral, making welfare comparisons difficult. Offering large subsidies to R&D without paying attention to the public revenue effect of the subsidies (incl. lump sum taxation) will overvalue the positive welfare effects of R&D subsidies. Further analyses of public revenue neutral innovation policy alternatives combined with optimal carbon policy are necessary in order to draw more general conclusions about optimal directions of innovation policies. Several of the earlier analyses of carbon policies and induced technological change were based on ad hoc modelling of the innovation processes without specifying profit maximising innovative producers (e.g. Nordhaus, 2002, Popp, 2004, Popp, 2006a, Popp, 2006b and Hart, 2008). In line with Gerlagh et al., 2007, Otto et al., 2007, Otto et al., 2008 and Otto and Reilly, 2008, we model the innovation processes as R&D based growth of the Romer (1990) type with imperfect competition in the markets for new technologies embodied in variety-capital, Jones and Williams (2000). Our contribution to the literature is three-fold. Firstly, our carbon policy is first-best and we study first-best innovation policies even though the subsidy rates are not first-best levels. In order to identify first-best subsidy rates we have to rely on uncertain estimates of the externalities in the research markets. In addition, optimising subsidy rates without limiting the total governmental support for R&D activities is not a realistic policy alternative.3 Secondly, we perform analyses of carbon and innovation policies that are revenue neutral and perform consistent welfare comparisons. Thirdly, we specify a computable general equilibrium (CGE) model with endogenous technological change based on the Romer (1990) approach for a small open economy. The small open economy characteristic adds interactions with the rest of the world into the analyses. Export possibilities of new technologies and commitment to international climate treaties introduce new elements into the policy analyses not elaborated on in this literature. In the presence of environmental and innovation market externalities and imperfections, the relative performance of different innovation and carbon policy alternatives is not obvious. Further, other market imperfections and existing public interventions will affect outcomes. Our CGE model incorporates these relevant market imperfections and public interventions. We specify R&D producing processes in both general and CCS technologies.4 The CGE model is calibrated to the small, open economy Norway with all its special characteristics. In spite of the included endogeneities of growth, the dominant growth impulses in the model are driven by external factors, though, in accordance with the findings for small, open countries (Coe and Helpman, 1995 and Keller, 2004).5 The small, open economy approach allows us to model exports of the domestically developed CCS and general technologies as an important channel for product diffusion. The new technologies embodied in capital varieties are exported at given world market prices. A global price on carbon secures the export demand for CCS technologies. The export channel plays a crucial role in expanding domestic technology production since the domestic market is limited. Our carbon and innovation policy analyses show that reallocating R&D support from CCS R&D to general R&D improves welfare, while reallocating support from general R&D to CCS R&D reduces welfare. With a higher carbon tax, the welfare gap between the two policy alternatives is reduced and flattens out for large values of the carbon tax. Our results do not, however, contradict the first-best result that the carbon tax should target the environmental externality and the R&D subsidy should target the imperfections in the research markets. Rather, it states that the carbon tax influences the productivity of both general and CCS R&D, and that the welfare effects of R&D support for the development of CCS technologies depend on the carbon emission restriction represented by the price of carbon emissions. The effect that the productivity of CCS R&D increases with the carbon tax level is confirmed by other studies (Greaker and Rosendahl, 2006 and Heggedal and Jacobsen, 2008). The paper is organised as follows: Section 2 presents the main structure of the CGE model, while Section 3 describes the calibration and baseline growth path. The policy reforms are presented and discussed in Section 4, while Section 5 concludes.

In this study, we ask how the R&D policy should be directed in an economy where carbon emissions are regulated by a carbon tax and R&D activities take place in the development of both general and CCS technologies. We also ask how the economic welfare effects will depend on the carbon tax level. In our study, the carbon tax is first-best, so that the argument for subsidising innovation activities is imperfections in the research markets that make the level of R&D effort too low. The main conclusions from our study can be summarised as follows: firstly, given a low carbon tax, reallocating R&D support to general R&D improves welfare, while reallocating R&D support to CCS R&D reduces welfare. Different channels of transmitting productivity to the rest of the economy and lower export of capital varieties, combined with decreasing returns to scale and decreasing returns to knowledge all contribute to dampening the positive welfare effects of the CCS R&D support scenario compared to the general R&D support scenario. Secondly, with a higher carbon tax, the welfare gap between the two policy alternatives is reduced. Stricter international carbon policies (higher carbon tax levels) are inciting production of both kinds of technological development, and especially of CCS technologies. The productivity of CCS R&D becomes higher relative to the productivity of general R&D. Our results do not, however, contradict the first-best result that the carbon tax should target the environmental externality and the R&D subsidy should target, the imperfections in the research markets. Rather, it states that the carbon tax influences the productivity of both general and CCS R&D. Our analyses show that even for very high carbon tax levels there is no welfare gain of redirecting R&D support to CCS R&D. At sufficiently high levels of the carbon tax, equal subsidy rates are best. In contrast to Otto et al., 2008 and Otto and Reilly, 2008, we focus on first-best global carbon policies, and our R&D policy reforms are public revenue neutral, implying that we can perform consistent welfare comparisons. We find that it is welfare preferable to subsidise general R&D relatively more than CCS R&D for carbon tax levels up to 175 euro. Otto and Reilly (2008) also find that introducing R&D subsidies throughout the entire economy is preferable compared to only offering subsidies to CCS R&D. Major explanation factors for the positive effects in Otto and Reilly (2008) are the rebound effect on the carbon price and the high level of R&D subsidising. A considerably higher carbon price on CO2-intensive sectors stimulates early introduction of CCS. This rebound effect is absent in our analyses. Rather, in our analyses, equal subsidy rates are preferable for sufficiently high carbon tax levels. The high R&D subsidising generates a public revenue loss that is not properly corrected for in the analyses in Otto and Reilly (2008), leading to in-consistent welfare comparisons. The use of CCS technologies is limited to CCS in production of gas power in our analyses. If CCS technologies could be used in other areas, such as in the processing industries, both the domestic and world market demand could expand. In our analyses there are no limitations on world market demand for CCS technologies embodied in capital varieties since the producers can export at a given world market price. We have disregarded any interactions between the world market use of CCS capital and the international carbon price. It is left for future research to elaborate more on these issues. It is essential to levy a price on carbon emissions in order to stimulate the demand for CCS R&D. Even with identical innovation and market imperfection parameters for the two kinds of R&D, it is welfare superior to reallocate R&D support from CCS R&D to general R&D for quite high carbon tax levels. This is partly explained by decreasing returns to scale and decreasing returns to knowledge. The repeated pressure from different politicians and environmental interest groups to support CCS R&D relatively more than other R&D may only be justified if the innovation and market imperfections are significantly larger for CCS R&D. It is left for future research to investigate whether there are sufficiently large differences in these externalities to favour CCS R&D compared to other R&D.