آلودگی هوا محلی و تغییرات آب و هوایی جهانی: تجزیه و تحلیل ترکیبی هزینه فایده

This article presents the findings of a combined cost-benefit analysis of local air pollution and global climate change, two subjects that are usually studied separately. Yet these distinct environmental problems are closely related, since they are both driven by the nature of present energy production and consumption patterns. Our study demonstrates the mutual relevance of, and interaction between, policies designed to address these two environmental challenges individually. Given the many dimensions air pollution control and climate change management have in common, it is surprising that they have only little been analyzed in combination so far. We attempt to cover at least part of the existing gap in the literature by assessing how costs and benefits of technologies and strategies that jointly tackle these two environmental problems can best be balanced. By using specific technological options that cut down local air pollution, e.g. related to particulate emissions, one may concurrently reduce CO2 emissions and thus contribute to diminishing global climate change. Inversely, some of the long-term climate change strategies simultaneously improve the quality of air in the short run. We have extended the well-established MERGE model by including emissions of particulate matter, and show that integrated environmental policies generate net global welfare benefits. We also demonstrate that the discounted benefits of local air pollution reduction significantly outweigh those of global climate change mitigation, at least by a factor of 2, but in most cases of our sensitivity analysis much more. Still, we do not argue to only restrict energy policy today to what should be our first priority, local air pollution control, and wait with the reduction of greenhouse gas emissions. Instead, we propose to design policies that simultaneously address these issues, as their combination creates an additional climate change bonus. As such, climate change mitigation proves an ancillary benefit of air pollution reduction, rather than the other way around.

Two interrelated environmental policy problems are global climate change (GCC) and local air pollution (LAP). Both are discussed in the political arena: the first notably in the United Nations Framework Convention on Climate Change (UNFCCC) and the second in, e.g. the United Nations Economic Commission for Europe's task-force on Long-Range Transboundary Air Pollution (UNECE-LRTAP). Emissions from the combustion of fossil fuels contribute to both GCC and LAP. Options to mitigate these problems are typically chosen to address each exclusively. For example, to reduce the emissions of SO2, NOx, or particulates, one often uses end-of-pipe abatement techniques specifically dedicated to these respective effluents, but not to CO2. Their application thus only contributes to diminishing LAP, not GCC. Alternatively, one of the ways to cut down emissions of CO2 is to equip fossil-fired power plants with CO2 Capture and Storage (CCS) technology, which in principle only addresses this greenhouse gas, and not the emissions of air pollutants. CCS equipment installed in isolation therefore alleviates GCC, not LAP. Still, options exist capable of simultaneously addressing both environmental problems, such as the substitution of fossil fuels by various types of renewables or nuclear energy. This paper investigates, through an integrated cost-benefit analysis of GCC and LAP, to what extent synergies between solutions for these environmental challenges can be created by using technologies that are beneficial to both at once. Nordhaus became one of the early protagonists in the cost-benefit analysis of GCC by deriving an analytical solution to a simple climate change maximization problem (Nordhaus, 1977 and Nordhaus, 1982). The answer to the problem involved an optimal time-profile for the concentration of CO2 in the atmosphere. Nordhaus later developed a numerical model (DICE) that simulated a rudimentary world climate–economy system (Nordhaus, 1993). Estimates for climate change damage costs, however, fundamentally determined his modeling results, like those of others who meanwhile had undertaken similar research (see, for example, Fankhauser, 1995, Manne and Richels, 1995, Tol, 1999 and Rabl et al., 2005). The reason was a very incomplete scientific understanding of potential climate change impacts, resulting in large cost uncertainties. Another shortcoming of this type of work was, and still is, that none of the GCC cost-benefit analyses cover the LAP problem, even while these two issues are closely linked. Indeed, they are both much driven by current energy production and consumption patterns. This paper attempts to correct for this, by presenting a single model that includes detailed descriptions of the costs and benefits of both GCC and LAP control strategies. In 1999, the EU adopted the Gothenburg Protocol to Abate Acidification, Eutrophication and Ground-level Ozone. This protocol set emission ceilings for the year 2010 for SO2, NOx, NH3, and VOC (volatile organic components). A few years later, the EU developed the National Emission Ceiling Directive that stipulated more stringent targets for these pollutants. The multi-national negotiations, leading to the agreement of these targets, used insights from scientific assessments and estimates for the economic costs of pollutant abatement options obtained with the LAP model RAINS ( Amann et al., 2004a and Amann et al., 2004b). Recently, results from RAINS have been used for restricted cost-benefit analyses of LAP, notably to serve the Clean Air For Europe program (CAFE, see Holland et al., 2005). Other studies of costs and benefits of air pollution policy packages have been performed that focused on isolated environmental problems or single pollutants (such as RIVM, 2000). All these analyses conclude that the monetary benefits of air pollution policies can be much larger than their costs. They all imply that the benefits are dominated by the avoided number of premature deaths from the chronic exposure of the population to concentrations of particulate matter (PM). A few studies merely signal potential LAP benefits resulting from GCC policies ( Criqui et al., 2003 and van Vuuren et al., 2006). They typically fix the carbon price, however, and restrict their analysis to Europe and the year 2010. These analyses therefore disregard the potential benefits of other and more costly options that simultaneously avoid GCC and LAP. Burtraw et al. (2003), in a similar study, also fix the carbon price and restrict themselves to the electricity sector in the United States for the year 2010. They find ancillary benefits from a decline in SO2 and NOx emissions, as well as avoided compliance costs under existing or anticipated emission caps. The authors also conclude that the initial carbon prices are significantly lowered because of these ancillary benefits. However, their analysis does not consider longer term or non-electric energy options. Thus, they give little guidance on how to design optimal strategies for addressing global warming and local air pollution. To our knowledge no multi-region model exists, that (1) covers the world and has a long time horizon, (2) jointly analyzes optimal greenhouse gas and PM emission reductions, and (3) allows balancing the costs of abatement with the benefits of avoided damages for both GCC and LAP. Our study aims to fill this gap. To be able to analyze the dual GCC-LAP problem, we judged it best to employ a global top-down model, but with a sufficiently large number of bottom-up technology features. For this purpose, we adapted the climate change model MERGE (Model for Evaluating the Regional and Global Effects of greenhouse gas reduction policies) as developed by Manne and Richels (1995). We employed MERGE in its cost-benefit mode, rather than in its cost-effectiveness format, which allows for an investigation of the balancing between the costs of abatement technologies and the benefits reaped from avoiding environmental damages. Hence, we did not impose a climate constraint under which total costs are minimized, as done in some of the other energy-environment models (such as DEMETER, see van der Zwaan et al., 2002 and Gerlagh and van der Zwaan, 2006). We expanded MERGE with a module dedicated to LAP including mathematical expressions for: • emissions of primary PM from energy use in electricity and non-electricity sectors; • chronic exposure of the population to increased PM concentrations; • number of people prematurely dying from chronic PM exposure; • monetary estimates for the damages resulting from premature PM deaths. We calibrated the LAP module to estimates from studies by the World Health Organization (WHO, 2002 and WHO, 2004) and the RAINS consortium (Amann et al., 2004a), as well as several other sources (Pope et al., 2002 and Holland et al., 2004). Since GCC and LAP damage cost estimates, as well as most of our other modeling assumptions, are subject to uncertainties, we performed an extensive sensitivity analysis with respect to all these elements. To mention just a few: our discounting assumptions, the climate sensitivity parameter, the costs of implementing CO2 and PM abatement options, the willingness-to-pay (WTP) for avoiding GCC damages, the number of premature LAP-related deaths, and the monetary valuation of these deaths. Reaping the welfare benefits from avoiding LAP-related damages constitutes the main mechanism at work in our new version of MERGE. LAP damages result from emissions of PM. Abatement of these emissions implies costs incurred by the implementation of end-of-pipe measures or switches from fossil fuels to the use of cleaner forms of energy. When benefits exceed costs for certain regions, an incentive is created for lowering the emissions of PM. A similar and synchronous balancing between costs and benefits occurs for CO2 emission reductions. At the same time the new model allows for balancing the incentives to act on LAP and GCC, while interactions and spill-overs between these two add to the overall optimization process. The analysis in this paper involves a stylized version of LAP, since it is restricted to one pollutant only (primary PM) and disregards other pollutants (e.g. secondary formed aerosols). We thus employ several abstractions: • We focus on emissions from fossil-fuel combustion, in both the electricity and non-electricity sectors, as these have an impact on (mainly) urban exposure to PM, but are also the main source of greenhouse gas emissions, and constitute as such the principal driver of both GCC and LAP. • While recognizing that LAP also includes pollution such as acidification, we restrict ourselves to PM only, as the monetary health benefits from PM emission reductions are much larger than those for other pollutants. • Mostly fine PM is responsible for deaths resulting from particulates in the ambient air, that is, PM with a diameter smaller than 2.5 μm (henceforth labeled as PM2.5), so that in principle we will only focus on this category of PM. • We disregard the contribution to PM concentrations from secondary aerosols, as the corresponding related health impacts are more uncertain than for primary PM. The impact of secondary aerosols on mortality may even turn out to be very limited (see WHO, 2006, p. 242). • Another reason for disregarding secondary aerosols is that their inclusion would necessitate addressing their interregional diffusion, and thus require an in-depth version of an air-transport model, which is beyond the scope of this paper. • Whereas theoretically PM can travel thousands of kilometers before being deposited, the major contribution to local PM concentrations comes from emissions that remain close to their source. Indeed, the high concentrations of primary PM in cities and densely populated urban areas mostly result from local transportation systems and power plants in their direct vicinity. We therefore make the assumption that regional PM emission reductions contribute to a decrease in PM concentrations within the region under consideration only. Among our other approximations are: • We have purposefully modeled LAP at a highly aggregated level, since this enables us to integrate LAP and GCC into a single modeling framework. The drawback hereof is that the PM emissions problem is modeled in a more rudimentary fashion than in, e.g. RAINS, as we simplify its detailed bottom-up abatement cost information for EU countries to only a few sectors and regions. The advantage, however, is that with this approach we are able to also introduce more economic realism than available in RAINS, as our simplification allows for an enrichment in terms of the simulation of time-dependent abatement technology costs. • The RAINS PM emissions information we use only covers Europe. Since few reliable data are available on PM emissions and activities for countries outside Europe, we have ourselves derived emission coefficients for all other world regions, based on those for Europe. • The concentration of PM2.5 has a larger mortality impact in comparison to PM10, but data on the former are generally scarce, while amply available for the latter. We therefore use PM10 data as proxy for PM2.5 data, as done in (WHO, 2006). • At an intermediate level of PM emissions a linear relationship exists between emissions and concentrations. PM concentrations, however, depend not only on regionally produced air pollution, but also on local factors such as meteorology. As a result, it proves that at low PM emission levels an increase hardly alters the PM concentration, the latter mainly being determined by regional PM background values. For our calculations we nevertheless restrict ourselves to a linear dose–response relationship. • The valuation of premature deaths from chronic exposure to PM concentrations is a contentious issue, since there are basically two ways to value health impacts, either through a ‘Value of a Statistical Life’ (VSL) or a ‘Value Of a Life Year lost’ (VOLY) method. In the first, one values a premature death against the VSL, while in the second, one estimates the number of ‘Years Of Life Lost’ (YOLL) and multiplies these with the VOLY. The European Commission decided for the CAFE program to adopt the precautionary principle, and thus employed the higher damage estimates from the VSL approach, also because they argue it to be more statistically reliable than the VOLY method. In this paper we also choose to follow the VSL approach, and test the robustness of our major conclusions through a detailed uncertainty analysis. Modeling a stylized version of LAP and restricting our analysis to one but dominant LAP-related substance – primary PM from fossil energy, with a very local character – allows us to explore and test the potential significance of the synergy aspects between policies mitigating LAP and GCC in an integrated cost-benefit framework. This framework enables us to derive optimal pathways for CO2 and PM emissions, under varying parameter values and modeling assumptions, on the basis of a trade-off between costs associated with mitigation efforts and benefits obtained from avoiding mid-term air pollution and long-term climate change damages. Section 2 of this article gives an overview of our adapted version of MERGE, and explains in detail how we extended the original MERGE model with a module covering air pollution. We highlight our most important results in Section 3, in terms of simulated CO2 emission levels and calculated costs and benefits of GCC and LAP policy. In Section 4 we present our uncertainty analysis, while reserving Section 5 for our main conclusions and recommendations.

To our knowledge, this article is the first to present a cost-benefit analysis that combines the damages resulting from global climate change and local air pollution. We demonstrate that MERGE, originally a global welfare optimization model of the energy–economy–environment system capable of investigating climate change policies only, can be extended with pollutants other than greenhouse gases. With our adapted version of MERGE we perform an integrated assessment of the long-term conundrum of climate change mitigation and the short-term challenge of reducing local air pollution, including for each the associated costs and benefits. Since these environmental problems are both driven by present energy production and consumption patterns, they constitute an inseparable pair that should, as we have pointed out, ideally be studied together. Our first main result is that the benefits of policies mitigating the emissions of CO2 and PM10 largely outweigh the costs of these policies, even while they induce important reallocations of resources to new (e.g. renewable) energy technologies and end-of-pipe abatement techniques (rendering fossil-fuel usage clean). Our second finding is that, as expected, GCC policy significantly reduces CO2 emissions and to some extent also PM emissions, while LAP policy induces radical PM emission reductions with negligible effect on the level of CO2 emissions. Third, combining GCC and LAP policies generates little further PM emission reductions, but clearly achieves extra CO2 emission reductions, that is, more than the sum of the reduction levels generated by either policy alone. Thus, a beneficial synergy between GCC and LAP policy can be created, with, as it proves, an additional energy-related CO2 emission reduction of 15% in Western Europe and 20% in China. Fourth, we find that GCC policy also delivers a welfare co-benefit in terms of lower LAP, while LAP-directed policy only generates welfare gains in terms of LAP benefits. The explanation is that under GCC policy modest PM emission reductions are achieved as a result of the installation of new technologies like renewables that simultaneously reduce CO2 and PM emissions. Fifth, we find that LAP policy leads to global environmental benefits that largely outweigh the benefits from GCC policy, by half an order of magnitude. Sixth, also in terms of costs and benefits we observe that a bonus can be created through a synergy of GCC and LAP policy, as the net welfare gain of combined GCC and LAP policy is higher than the sum of the gains of GCC and LAP policy alone. This welfare gain proves to be mostly employed to further mitigating climate change. Our overall finding is that it is more urgent to address the problem of local air pollution than that of global climate change. The main reason is that the short-term benefits that may be obtained from air pollution control are much larger than the long-term benefits obtainable through strategic climate change measures, while the associated costs are in both of these policy cases much lower than the achievable benefits (even with very low discount rates, see also our sensitivity analysis). So, most environmental and human health policy today should be dedicated to local air pollution. We do certainly not suggest, however, that climate change policies should be neglected or postponed. Rather we advise to combine already today our first priority (LAP control) with our second (GCC mitigation), because there is a clear bonus to be gained in terms of climate change control by jointly implementing both policies. In this article we suggest that climate change mitigation is an ancillary benefit of air pollution policy, rather than the other way around: LAP control combined with GCC policy creates an extra early kick-off for the transition towards climate-friendly energy supply. The benefits of climate change policy will be experienced much further in the future than those of air pollution policy, and thus are subject to more substantial discounting. This of course much contributes to our finding that the difference between the monetized benefits of avoided air pollution and precluded climate change is large. Given the importance of discounting assumptions for this principal result, we have modified our descriptive approach to one of a prescriptive nature in our sensitivity analysis (that is, replacing high discount rate values with low ones). But still we find essentially the same outcome. Since there are many other uncertainties involved in cost-benefit analysis, we have changed our assumptions regarding all main modeling parameters, which allows for an assessment of the robustness of our conclusions. We have reported in detail the specific variations applied to our assumptions concerning the principal driving forces behind our results. All of these confirm our conclusion that the benefits obtainable through LAP policy largely outweigh those of GCC policy, at least by a factor 2, and in most cases of our sensitivity study much more. Our investigation has revealed the mutual relevance of policies designed to address the associated challenges of GCC and LAP. Strategies restricting to long-term climate change are likely to improve air quality, as emissions of both CO2 and PM are often reduced at once. Alternatively, however, by only controlling local air pollution one only little helps to reducing emissions of CO2 and hence to mitigating climate change. PM emission reductions are typically achieved through end-of-pipe applications that do not simultaneously affect emissions of CO2. Yet even while the latter may be true, we have shown that a combined GCC plus LAP policy generates extra benefits in terms of climate change mitigation. Given this effect, we thus advise (1) policy makers to design and implement combined GCC and LAP strategies, and (2) analysts and scientists to correspondingly study these environmental challenges jointly. With this article we hope to have made an insightful first step. An interesting corollary is a comparison of our results with those of Rabl et al. (2005). They report, like we do here, that uncertainties in damage costs distinctly affect cost-benefit analyses of environmental pollution. Still, they point out that, for a range of different pollutants, the social cost penalty is remarkably insensitive to errors in the assumed damage costs. Their main finding, namely that it is optimal to achieve significant emission reductions for all effluents analyzed, continues to hold under large variations of external environmental costs. The results presented here have also been subjected to an extensive robustness analysis regarding a range of possible uncertainties that relate to air pollution and climate change damage costs. Also our main finding, the predominance of LAP concerns above those for GCC, remains unaffected under a wide span of parameter values related to CO2 and PM induced damages. In this article we employed conservative estimates for the impact of ambient concentrations of PM2.5 on mortality by neglecting some important contributing sources. Among these are the use of traditional fuel-wood in non-Annex I countries, the second-order formation of fine particulates through emissions of SOx, NOx, and NH3, as well as (and in particular) process-related emissions. Still, even with these conservative estimates, our results point at LAP being the primary concern, and GCC the secondary. While we definitely do not want to discard the problem of GCC, LAP policy should be given clear priority. Furthermore, a GCC plus LAP policy can ‘lock’ the world deeper into climate change mitigation than GCC policy alone. The two large developing countries, China and India, deserve a last remark, as they are likely to soon become dominant players in the global economy and will almost certainly increasingly become dependent on fossil fuels. They will without doubt continue their use of coal throughout much of the 21st century, also given their large domestic coal resources (see e.g. van der Zwaan, 2005). The sense of urgency to deal with local and regional pollution will be felt especially in these countries: already now their large cities are plagued by a severe deterioration of the ambient air. Several commercial end-of-pipe technologies exist that constitute clean complements to the traditional use of coal, which allows these nations in the short term to switch away from dirty coal combustion and benefit from the corresponding avoided air pollution damages. Still, they will not solely want to focus on LAP, but also need to start considering GCC, and thus contemplate the use of renewable energy resources like solar energy and wind power, or options like hydropower and nuclear energy, or the continued use of fossil fuels but complemented with CCS technology. This study shows that such climate mitigation options, however desirable and necessary, should first and foremost be carefully considered against the simultaneous benefits they engender in terms of their potential contribution to reducing local air pollution.