A Framework for Evaluating Mitigation Options
To devise a coherent strategy for mitigation of greenhouse warming, it is necessary to have an analytical framework that compares the alternatives available. This chapter develops such a framework and discusses a number of issues that arise in the development of a plan to respond to the potential effects of greenhouse warming.
Economics and engineering are central to the comparison of alternatives. The connections between human activities and their environmental consequences are technological in character; engineering is consequently required to imagine, design, and implement alternatives. Economic concepts are central to choosing among the technically feasible alternatives. A variety of social and cultural factors are also important in the interactions of humans with their physical environment, but these are not the primary focus of this inquiry.
The chain of causation from human activities, to the release of greenhouse gases, to changes in the composition of the atmosphere, and to climate change is long, often indirect, and complex. For this reason, estimating the relationship between human activities affected by policies or shifts in markets and far-removed changes in climate is a difficult technical task. Indeed, the very human difficulty of perceiving this indirect and long-term relationship is an important component of the problem of greenhouse warming.
It is easier to see the direct costs of decreasing CO2 emissions than to estimate the benefits of doing so. There is, accordingly, an emphasis in this report on the direct costs of change rather than on the potential benefits and secondary costs of changing. Readers should bear in mind that the picture presented by the panel is skewed in this respect.
Even if the relation between human activity and climate change were
readily quantifiable, there would still be the matter of selecting the most effective, least costly response strategies. Here economic concepts are central. Proposed responses to greenhouse warming include ideas that would affect national economies, international trade, and the life-styles of people in both developing and industrialized societies. Moreover, selecting some of these alternatives would mean that other highly valued objectivessuch as improving economic status or national securitywould have to be altered. Making choices in the face of scarcitythe problem at the heart of economic scienceis inescapable.
Much of this chapter is devoted to explaining the difficulties of carrying out a conceptually straightforward approach. There are three critical problems: (1) markets are imperfectthat is, neither the prices observed nor the responses of markets are the simple result of demand and supply operating unimpeded; (2) uncertainties abound in the technical realm, in social responses to policy instruments, in environmental changes due to changing climate, and in markets; and (3) consideration of most alternatives requires comparing costs and benefits at different times, paid for or enjoyed by different people.
Although it has been possible to assemble an overview of the options for mitigating greenhouse warming, the panel urges readers to bear in mind the formidable problems of theory and practice limiting the precision of the estimates that can be provided at this time and even the qualitative accuracy of the picture that can be presented.
Greenhouse warming is a phenomenon of the atmosphere, taking place in a global "commons." Similar emissions of greenhouse gases have similar potential to affect global climate, regardless of their country of origin. Thus mitigation strategies must be global in scope, at least implicitly involving both developed and developing countries. Indeed, many of the lowest-cost mitigation options may be found at first in some of the poorest developing countries. For example, the efficiency of wood-burning cookstoves can potentially be raised at very low cost (Reid, 1989). Because these countries may be unwilling or unable to afford such policies, the developed countries may choose to underwirte such efforts. This targeted redistribution of economic resources could be efficient and less costly to the developed countries than mitigation strategies directed solely toward their domestic economies.
Because of the limited availability of information on a global basis, however, and the scope of the panel's responsibilities, the analysis of mitigation options in the chapters that follow is devoted largely to the United States. With a few exceptions, information on mitigation costs and estimates of
mitigation potential are derived entirely from U.S. experience and data. Similar analyses should be effected for other countries. Indeed, the analytical framework used in this report is generally applicable to the analysis of mitigation options in a global context.
The Role of Cost-Effectiveness
Principal among the objectives of policymakers in designing a mitigation strategy should be to minimize the adverse effects of mitigation on the domestic or world economy. This requires designing a strategy that is "cost-effective"one in which the incremental costs of reducing radiative forcing are minimized. Because the cost per unit of mitigation for most options is not likely to be constant over the entire range of measures, estimates of incremental cost per unit of mitigation will depend on the degree of mitigation obtained, and may rise rapidly as measures are used more intensively. In this report, all cost estimates are based on changes from current levels of emissions, although in many cases these cost estimates are for substantial increments of potential mitigation.
The cost of mitigation may include a number of components, some of which are difficult to measure. First, there are direct expenditures, such as the increased cost of chemical substitutes for CFCs; these costs reduce CFC concentrations below what they would otherwise be and do so promptly. Direct expenditures can be measured readily when market transactions are available to provide data on the prices consumers pay for the benefits of these expenditures. Second, there are investments whose benefits are delayed. For example, higher energy efficiency in an industrial facility will return benefits in the form of reduced emissions of greenhouse gases and energy costs as the facility reduces its energy consumption over the life of the plant. In estimating the value of a stream of benefits and costs over time, a discount rate (interest rate) is used to compute the present value equivalent in order to compare alternative investments. Third, there are implicit costs and benefits in substitutions among final goods or services that imply different levels of greenhouse gas emission. For example, inducing urban commuters to switch from automobiles to mass transit would reduce an important source of greenhouse gas emissions. Yet experience in the United States suggests that such a switch would not occur at the energy prices observed in recent time. If changes in government policy are necessary to change behavior, however, the social cost would include the net loss in value to consumers of changes in their behavior that would not have occurred without changes in policy. That cost is difficult to measure because there is no market transaction that directly reflects such changes in value to customers.
Because most of the mitigation options discussed in this report involve a
reduction in energy consumption, there may also be reductions in other undesirable externalities of energy production. The social and economic costs of these reductions in energy consumption may even be outweighed by their benefits. For example, the decision to limit highway speed to 55 miles per hour (mph) in the 1970s and 1980s reduced traffic fatalities considerably. In principle, these reductions in externalities should be deducted from the direct and indirect costs of mitigation. Where possible, these favorable offsetting effects are identified in the chapters that follow; however, they are not generally quantified and applied as offsets to the estimates of cost per unit of mitigation. Of course, many of the issues bear on societal and individual preferences having components that extend beyond quantifiable costs.
The scope of the task of cost-effective choice can be seen through a review of the work done to date by economists who have estimated the costs and, less often, the benefits of mitigating greenhouse warming. There have been relatively few attempts to estimate these costs by energy modeling. In energy modeling the energy sector of the economy is represented in terms of technological activities such as space heating or transportation services. By using a mathematical programming or other algorithm, the models then solve for the "optimal" trajectory of prices, output, fuel mix, and technologies. It can be shown that, under certain conditions, the optimal trajectory would correspond to the outcome of perfectly competitive markets. (See Appendix R for more details.) Recently, a few economists have begun to work on estimating the costs and benefits of various CO2 reduction scenarios in this way. Most of these modeling exercises are still in rather preliminary form.
A major problem in measuring the costs of reducing emissions of greenhouse gases lies in establishing the baseline from which these reductions are to be measured. If the object is to measure the costs of restricting CO2 emissions by some year, such as 2030, to some percentage of current emissions, it is necessary to begin by predicting unconstrained emissions for 2030. The costs of limiting CO2 will then be dependent on the assumptions made about economic and population growth until that date; the prices of oil, natural gas, and coal; technological changes in energy-using industries; and numerous other parameters that drive emissions in the unconstrained baseline scenario.
A simpler approach is to estimate the cost of reducing CO2 from current emission levels. This procedure eliminates the necessity for predicting unconstrained CO2 emissions in some future year, but it does not provide estimates of the cost of restricting future emissions to some fixed level.
In either case, it is necessary for energy modelers to make assumptions about the costs of certain activities at scales outside recent experience or to predict technical change over some horizon. The estimated costs of reforestation or ocean-modification options are highly speculative. So are the estimates of additional costs of technologies to replace current fossil-fuel-based electric power generation. As a result, costs are usually estimated for various scenarios of technical change. These scenarios include a range of estimates of fuel prices, growth in gross national product (GNP), and other parameters.
There are four relatively recent attempts to model the prospective costs of CO2 reductions: Nordhaus (1989), Manne and Richels (1990), Jorgenson and Wilcoxen (1989), and Edmonds and Reilly (1983). These provide a perspective on the current state of the art in modeling the cost of CO2 abatement. (Another review that reaches similar conclusions in Darmstadter (1991).)
The Nordhaus Study
Nordhaus's work on global warming began in 1977. In recent papers, he has presented the beginnings of a major modeling effort designed to estimate the cost of controlling CO2 and other greenhouse gases, assuming efficient markets and taking into account the costs and benefits of various rates of abatement. As part of this exercise, he has attempted to synthesize the results from eight studies of CO2 abatement, which draw upon data on current practice and extrapolations into the future. A log-linear ordinary least-squares regression is fitted to these estimates and shown as a relation between CO2 reductions and a "tax rate" per ton of carbon at 1989 prices. This tax rate purports to measure the minimum cost of reductionan estimate of the marginal cost for the whole economy of the most efficient approach to CO2 reduction. His results are shown in Figure 20.1.
Nordhaus also estimates the marginal costs of achieving efficient reductions in greenhouse warming through reductions in CFCs and through reforestation. He then combines the three cost curves for CFCs, reforestation, and CO2 abatement into a single efficient marginal cost curve for greenhouse gas reductions. These results are shown in Figure 20.2 (The chart estimates, for example, that a 30 percent reduction of greenhouse gases would not be equivalent of $150/t CO2 equivalent.)
Nordhaus's results on the costs of mitigation should be thought of as his estimate of "the best we can do" to reduce carbon usage at minimum cost, by using currently known technology or expert estimates of the technology that can reasonably be expected to be available. Because his model assumes the consumption of resources at current levels and constant exponential growth of the economy with this resource constraint, the resulting estimates
of mitigation costs must be viewed as tentative. Nordhaus stresses that the actual costs of regulatory approaches are likely to be higher, because government-mandated reductions in emissions are likely to be less efficient than a carbon tax.
Analyses like Nordhaus's, however, implicitly ignore institutional barriers that impede efficient economic adjustments to change. Such barriers exist because information is imperfectly distributed, there are regulatory restrictions on transactions, and buyers or sellers can possess monopolistic control over markets. For example, the adoption of such energy efficiency measures as the installation of better-insulated windows suffers from several impediments: homeowners are generally uninformed about many possible efficiency measures, and building codes may not permit the installation of windows that would be suitable. If existing barriers to adaptation can be lowered, both buyers and sellers can gain from the resulting transactions, which leads to estimates of negative costs for some mitigation steps. That does not mean that mitigation requires no monetary outlay: it is still
generally an investment, requiring the commitment of capital. It does mean that, relative to the present situation, everyone can gain from lower barriers to adaptation. There is a strong case, prima facie, for adopting suchpolicies, although experience over the past decade suggests that resistance to doing so often exists.
The Manne and Richels Study
Manne and Richels have built simulation models of CO2 reductions for the United States and for the entire world. These models estimate the cost of various CO2 abatement scenarios from the present through the twenty-first century and thus include forecasts of economic growth, fuel prices, and new technology.
Manne and Richels examine the economic cost of holding carbon emissions constant from 1990 to 2000 and then gradually reducing these emissions to 80 percent of 1990 levels by 2020. Using a discount rate of 5 percent, they estimate the present value of aggregate loss to U.S. economic consumption due to this constraint under a variety of scenarios concerning U.S. energy technology and policy. The most restrictive and pessimistic
scenarioone that assumes no autonomous improvements in energy efficiency, no development of low-cost nuclear power, and no cost-effective means of removing CO2 from utility waste gasespredicts very small reductions in U.S. consumption until 2010, but sharply rising losses thereafter. The discounted present value of U.S. consumption losses through 2100 is $3.6 trillion under this scenario, or about 1 year's current consumption (i.e., less than 1 percent of total consumption during the century). Using the most optimistic combination of assumptions, Manne and Richels estimate that the present value of the cost of the carbon emission reduction through 2100 would fall to $0.8 trillion. In that case, technical progress in and wider use of nuclear power, improved CO2 removal technology, and overall energy efficiency in the economy could reduce the cost of mitigation by 78 percent in their model.
It should be stressed that Manne and Richels's results depend heavily on their assumptions concerning U.S. economic growth, world fuel prices, and the prospects for technical progress in the energy sector. They assume a substantial slowing of U.S. economic growth from 3 percent annually in the period from 1990 to 2000 to only 1 percent annually in the last half of the twenty-first century, in the absence of a carbon constraint. A higher rate of economic growth would increase the estimated costs of a carbon constraint substantially, but probably not proportionately with future GNP.
Manne and Richels's estimate of the carbon tax required to produce a 20 percent reduction in carbon (C) emissions, compared to 1990 rates, is shown in Figure 20.3. It shows the carbon tax rising from nearly zero in 2000 to nearly $400/t C in 2010 and peaking at about $600/t C in 2020.1 The tax falls thereafter, presumably due to a slowing of economic growth and the expansion of more efficient (lower emissions) energy supply technology.
The Jorgenson and Wilcoxen Study
Jorgenson and Wilcoxen have built a long-term simulation model of the U.S. economy to measure the effects of energy and environmental policies on U.S. economic growth. Although this model was not constructed with the goal of estimating the effects of CO2 reductions, it can be used for this purpose. Jorgenson and Wilcoxen's model is by far the most disaggregated and complete model discussed here. It also has the most sophisticated treatment of capital formation, an important determinant of long-term economic growth.
Carbon dioxide emissions from fossil fuel consumption plus cement manufacture were virtually the same in 1972 and 1987, according to Jorgenson and Wilcoxen, in large part because of a doubling of the relative price of oil between these two years. This observation can be used to simulate the cost of a freeze on CO2 emissions, given the central role of energy prices in the
Jorgenson-Wilcoxen growth model. An analysis prepared by the authors concludes that long-term growth of the gross domestic product in the United States was reduced by about 1.3 percent per year by the 1972–1987 doubling of oil prices (Jorgenson and Wilcoxen, 1989).
None of the three energy models described above is likely to provide precise estimates of the effects of a reduction in CO2 emissions in the next 30 years or beyond. Nevertheless, they are quite helpful in judging the first approximations of those effects. Of the models described, Nordhaus's is the best for projecting the immediate costs of reducing CO2 emission from any given current level. Manne-Richels and Jorgenson-Wilcoxen provide longer-term simulation models of the effect of energy-environment policies, including the limitation of CO2 emissions.
The Edmonds and Reilly Study
The Edmonds and Reilly (1983) model analyzes long-term, global emissions of CO2 by adopting a simplified picture of an economy that generates CO2 from fossil fuel burning. Because it was developed early in the current cycle of attention to greenhouse warming, the model has been widely used (e.g., Lashof and Tirpak, 1991).
The model divides the global economy into nine regions, each of which is assumed to be a single market for energy. Six primary energy categories, three of which emit CO2 in varying amounts per unit of energy consumed, are analyzed. Demand for energy is driven by a simple model of population, regional economic activity (GNPs), energy productivity (a measure of the pace of technological change), and taxes. Supply of energy is governed by regional resource availability and economic descriptions of ''backstop" technologies available within each region. (A backstop technology provides the price at which unlimited quantities of energy are assumed to be available from an inexhaustible resource; in general, the backstop technology is more expensive than other resources currently available from domestic production or international trade.)
Edmonds et al. (1986) discuss the behavior of this model when a significant subset of those assumptions is systematically varied. Projected CO2 emissions range from 5 to 20 Gt C/yr in the 400 scenarios examined. (These values cover the span between the twenty-fifth and seventy-fifth percentiles of the 400 scenarios.) Thus, by employing assumptions that are not inconsistent with current estimates, this widely used model projects a sizable uncertainty in CO2 emissions 60 years in the future, ranging from values close to those emitted today to values 4 or more times larger.
Problems in Comparing Options
The energy models described above yield results that appear to be strikingly different from those presented in subsequent chapters. For example, Figure 20.4 shows a curve for energy efficiency (discussed in Chapter 21) that indicates that significant amounts of carbon mitigation are available at negative net costs ("net savings") to society. (Net costs in Figure 20.4 are described in two ways: dollars per ton of CO2 saved; and costs per kilowatt-hour of electricity needed to achieve those savings.) As shown on the right-hand vertical scale, energy efficiency in the buildings sector saves money because, although energy-efficient appliances cost more than those currently in use, the additional cost (at a 6 percent real rate of interest) is less than the cost of the energy saved. Compare Figure 20.2, which estimates the carbon tax required to induce carbon emission reductions; that such a tax must be imposed to reduce emissions of greenhouse gases means that there is a positive net cost to society. Which perspective, if either, is correct?
The answer lies in understanding the inherent limitations of each approach with respect to the task of evaluating specific mitigation options. Energy modeling, in its current state of development, is limited in its ability to evaluate the direct reduction or offset of greenhouse gas emissions achievable by different options. The approach used in this study, which the panel
calls "technological costing," is better suited to this task but limited in its ability to assess overall consequences for the economy. Several related problems are discussed below: (1) deviations of real markets from the idealized bargaining assumed in economic theory, (2) uncertainty, and (3) comparisons of current costs with future benefits. These problems lead to the conclusion that there is no single formula or method for choosing the best alternatives, although comparisons among alternatives are necessary to make informed and prudent decisions.
A common limitation on evaluations of mitigation options is the assumption of unconstrained markets, in which buyers and sellers can arrive at mutually agreeable prices for the exchange of goods without external influence. Roughly half the world's energy is not sold through unconstrained markets, howevera reflection of the widespread perception that energy is too important to be left outside the sphere of government control. Evalutions that do not allow for such factors can produce misleading results.
Market Imperfections and Regulatory Distortions
Markets can be constrained directly by price setting, as in the policies of the Organization of Petroleum Exporting Countries (OPEC) cartel, or indirectly, as in the regulations that limit the production of CFCs in the United States. These are imperfect marketsone caused by monopolistic practices, the other a result of an international agreement to protect the stratospheric ozone layer. Interventions in an unconstrained market may or may not be justified as a matter of social policy, but all such interventions move prices away from the levels that would prevail otherwise. This creates important complications in estimating the costs of abating greenhouse gases.
Energy prices are not what they would be without regulationand, consequently, their levels after mitigation steps are taken become more difficult to estimate. Other imperfectionsin capital markets, building codes, subsidies, and taxesall bend the behavior of producers and consumers away from the equilibrium that would obtain in their absence. Moreover, consumer and commercial discount rates are generally higher than the discount rates typically used in the societal cost-benefit analyses discussed below in the section on rates of return (Meier and Whittier, 1982; Train, 1985; Ruderman et al., 1987; Electric Power Research Institute, 1988; Peters, 1988; Ross, 1989; Gladwell, 1990; Koomey, 1990). Indeed, the market imperfections already in place provide significant room for mitigation.
The analyses of potential mitigation options contained in the chapters that follow are often addressed to such market failures. Some utilities and governments have accumulated significant experience with such policies in the past 15 years (Vine, 1985; Hirst et al., 1986; Vine and Harris, 1988; Koomey and Levine, 1989; Krause et al., 1989; Rosenfeld et al., 1989; Wilson et al., 1989; American Council for an Energy-Efficient Economy, 1990; Nadel, 1990). Experience shows that these policies can offer high returns in the short rungreater than the cost of capital in many cases (Geller et al., 1987; Train and Ignelzi, 1987; Geller and Miller, 1988; U.S. Department of Energy, 1988; Krause et al., 1989; Nadel, 1990). This is the origin of the net negative cost estimates in Figure 20.4. Note, however, that
evaluations of energy efficiency programs are usually made in terms of public utility outlays, rather than total social costs. The latter is the appropriate standard of measure of energy efficiency programs because both the benefits and the costs of energy efficiency are widely distributed, affecting firms and households that are neither shareholders in utilities nor purchasers of their energy services.
Technological Costing Versus Energy Modeling
The limitations of energy modeling indicate the need for an additional method to supplement the information provided by the energy modeling method in the effort to estimate the costs of mitigation. This approach, "technological costing," bases its estimates on a variety of assumptions about the technical aspects of the mitigation option, together with estimatesoften no more than guessesof the costs of implementing the required technology. This approach can be useful whenever there are insufficient data on the actual costs realized in markets or when it is difficult to use statistical methods to estimate the costs of future policies from the historical behavior of markets. Technological costing does not escape the limitations of economics, however, because this approach also relies implicitly on economic assumptions. The most important of these is the assumption that direct costs are a good measure of total cost.
The technological costing approach should be seen as complementary to the modeling described above, which uses simulation techniques to make inferences from economic data and assumptions about economic structure. Energy modeling studies are likely to be useful whenever observations of market behavior are available. As shown in Chapter 29, the two approaches are in rough agreement, given the large uncertainties in the best available knowledge. This lends some comfort with respect to the general validity of the results, although the degree of uncertainty limits the utility of these estimates for the purposes of making specific policy recommendations.
Neither the energy modeling nor the technological costing approach is perfect. Technological cost studies can often be criticized for scanting informational, adjustment, or managerial costs. Their main weakness is that they fail to allow for impacts on quantities and prices in other markets and therefore neglect "general equilibrium" effects of any major actions undertakensomething that the full-scale economic models allow for. Energy modeling analyses are subject to criticisms concerning model specification, measurement errors, and the relevance of historical data and behavior for future untested policy actions. In the chapters that follow, most of the cost-effectiveness measures derive from technological costing rather than energy modeling. The estimates of cost-effectiveness and emission reduction potential from the analyses in this report are combined in Chapter 29, which
traces out a "supply curve" of mitigation possibilities. In some instances, these analyses find a negative cost of mitigation, implying that mitigation actions sometimes yield a positive economic return. Negative mitigation costs assumes certain current public or private market imperfections will be overcome. If such imperfections could be corrected at low enough cost, society could achieve substantial benefits from these mitigation measures even if greenhouse warming were not a problem.
The energy models set a baseline for more detailed analysesa baseline, discussed below, that is still fuzzy. Yet energy models do attempt to capture responses to changes in technology and policy from markets reaching throughout society. Analyses that use prices only as static indicators of current value can miss dynamic adjustments of great significance on time scales of decades.
Uncertainties in Energy Modeling
Although the timing of greenhouse warming is not known with precision, it is highly likely that any effects will unfold over times much longer than the normal horizons of economic forecasting. Accordingly, mitigation efforts may have to be spread out over decades as well. Analyzing such possibilities requires extrapolation of economic conditions well into the coming century, an enterprise of doubtful accuracy given our current understanding of economic dynamics. The existence of so many uncertainties constitutes the second major problem in policy design.
There are no facts about the future. Estimates of important parameters, such as the level of carbon tax needed to induce large-scale (20 to 50 percent) reductions in CO2 emissions, vary by large amounts. Nordhaus (1990) estimates $40/t C ($11/t CO2 equivalent) of coal equivalent; Manne and Richels (1990), $250/t C ($68/t CO2 equivalent). These variations illustrate the current state of the art. Not only do all models indicate large quantitative uncertainty, the narrow base of validation raises questions about the qualitative uncertainties and systematic errors that may be embodied in the models. For instance, several models incorporate a rate at which energy efficiency improvements enter the economyan ad hoc but influential assumption. Edmonds et al. (1986) find that the exogenous energy efficiency improvement rate explains much of the variation in the behavior of the widely used Edmonds-Reilly forecasting model. Yet there is no theoretical reason to think that energy efficiency improvements should grow exponentially or that the rate of improvement is independent of government policy. The assumption appears to be made simply for convenience in modeling.
Thus, for the purpose of policy design in the near future, energy models are essential to framing a conceptually sound approach, but neither individually nor collectively do they provide comprehensive guidance on the choices to be made.
Uncertainties in Technological Costing
Although the panel believes that the technological costing approach provides the most information for evaluating the mitigation potential of current options, this approach does have certain limitations or uncertainties. Perhaps most important, such option-driven assessments do not do a good job of incorporating social responses. For example, despite the proven benefits of a variety of energy efficiency measures, consumers have been slow to adopt them. Lack of information, financial constraints on households, and a variety of other barriers impede their use. Such factors are difficult to incorporate in the technological costing approach, which leads some to criticize this approach as having too often led to optimistic estimates of options.*
A second uncertainty in the technological costing approach involves interactions among different options. Implementing energy efficiency options, for example, would likely reduce demand for energy. This, in turn, could alter the price of energy, which would affect the panel's calculations of cost-effectiveness for energy options, which assume a constant price for the cost of electricity, natural gas, gasoline, and so on.
This constant cost assumption is an additional uncertainty that affects not only the interaction among the prices of different options, but also the cost of each individual option. The technological costing approach assumes that the implementation of a particular option does not affect the cost of that option. For example, a substantial increase in the number of natural gas-fired electricity generating plants could increase demand for natural gas, and therefore the price of natural gas would likely rise. Such price increases could affect the cost-effectiveness calculation for natural gas, perhaps enough to alter its ranking.
The energy modeling approach takes such issues of demand, supply, and consumer behavior into account, whereas the technology costing approach does not. Although the panel recognizes the possibility of such interactions, it believes the technological costing approach is better suited to responding to the panel's charge of evaluating the comparative advantages and disadvantages of specific mitigation options so that it can identify and rank the available mitigation strategies. Since the panel realized that such interactions are likely, however, it developed reasonable targets (e.g., replacement
*Panel members Douglas Foy and Robert Crandall have further views on the use of technological costing: Foy wishes to emphasize that the approach can overestimate the net social cost (i.e., social costs less the social benefits) of mitigation, because it fails to include many possible quantifiable and unquantifiable social benefitssuch as reduced air pollution and enhanced energy security. In contrast, Crandall believes that the approach is likely to provide estimates of the social cost of mitigation that are overly optimistic. He believes that such excessive optimism is particularly noticeable in, but not necessarily limited to, the transportation sector, given the economic studies that reach very different conclusions from those advanced in this report.
of 3.5 incandescent light bulbs per household with compact fluorescent lighting) and cost assumptions in its calculations, leaving a full assessment of such interactions to future studies. In addition, the panel presents the implications of less than full implementation of those reasonable estimates.
Comparisons over Time
One important conceptual contribution of economic theory addresses the third problem in policy design: the matter of comparing benefits and costs occurring at different times. Because the time scales of greenhouse warming are long in comparison with most of the rhythms of human events, it is particularly critical to think through clearly the question of intertemporal comparisons of value.
Steps to mitigate future changes in climate are investments: we incur the costs now, and we and future generations enjoy the benefits. If our principal concern is the effects of our current actions on our grandchildren and great-grandchildren, we should compare greenhouse mitigation actions with alternative legacies that can be left to future generations. The conventional approach to making this comparison is to apply a discount rate, comparing the value of investments against the value of an equivalent sum put into an interest-bearing instrument. The discounting procedure reduces the value of future benefits, because the alternativeearning intereststeadily raises the level that must be achieved to be competitive.
Some argue that discounting may not be appropriate. For example, if taking option A today results in destruction of a particular ecosystem in 2020, whereas taking option B results in destruction of the same ecosystem in 2030, should a lesser cost be assigned to option B? The discounting procedure does so. If options A and B differ in no other respect, it would be a mistake to apply the discounting approach blindly. But as discussed below, it remains sensible in practice to require returns on investment of at least 10 percent per year, except under special circumstances. That is, mitigation actions should be compared with other investments we can make now. The availability of alternative investments establishes a minimum rate of return, which mitigation actions should yield if they are to compete favorably with these alternative choices to improve life in the future.
Rates of Return
What are the rates of return on alternative investments? Macroeconomic calculations for the United States suggest a return on capital investments of about 12 percent. The U.S. government operates under a guideline, both for government investments and for regulatory requirements on the private sector,
suggesting that a project or a regulation with a prospective rate of return below 10 percent in real terms (i.e., corrected for inflation) should be rejected (Office of Management and Budget, 1974; Reischauer, 1990). The World Bank lends extensively to developing countries around the world, mainly but not exclusively for large-infrastructure projects. The informal guideline in project evaluation is that the estimated rate of return must exceed 10 percent in real terms, although in some circumstances, where nonquantifiable benefits are expected to be important, lower estimated rates of return are accepted. A survey of over 1000 projects undertaken during the 1970s and 1980s yielded an average expected return on completion of 16 percent (Pohl and Mihaljek, 1989).
Other rates of interest have been suggested for evaluating public investments. An extensive and complex literature exists that attempts to sort out which of several possible discount rates should be used in the United States, depending on whether the investment in question is deemed to displace private investment, personal consumption (where consumers are borrowing at rates higher than 10 percent to consume now), or government investment (e.g., see Hausman, 1979; Lind, 1982). Recently, the Congressional Budget Office has suggested that the appropriate rate of return for evaluating U.S. government projects is about 3 or 4 percent in real terms, the rate at which the United States is able to borrow from the rest of the world, and hence the real cost to Americans (Reischauer, 1990); however, this standard may be irrelevant to a global issue such as greenhouse warming (see the discussion of discount rates in Chapter 4, Part One).
The appropriate criterion in establishing a discount rate should be the ability to provide for future generations. Determining that rate requires deciding the scale of investment and, at any given scale, comparing each proposed investment in the future with alternative investments. What is relevant, again, is the return on feasible investments that the mitigating policies displace anywhere in the world. The World Bank experience and current criteria suggest that any proposed mitigation investment should be expected to yield at least 10 percent per year, except under special circumstances.
The time frame for global climate change, at 50 to 100 years or more, is a long one by the standards of most investments and most public policy. At a discount rate of 10 percent, the present value of dollar a 50 years from now is less than a cent, and that of a dollar 100 years from now is less than one-hundredth of a cent. In other words, far distant payoffs from current investments are worth practically nothing today.
Suppose, however, that the cost of greenhouse mitigation rises sharply from current levels as greenhouse gas concentrations increase. Does that not warrant doing something now? The answer is negative, as long as alternative investments continue to have a higher yield. Indeed, that criterion
will be easier to meet if, as postulated, the cost of mitigation increases in the future and the yield from mitigation declines.
This conclusion would fail to hold only if the yield on alternative investments in the future could be expected to fall, and to fall by so much that over the total 50- to 100-year period the yield on alternative investments would be lower than the yield on mitigation actions taken in the near future. The "at least 10 percent" criterion is applicable to the late twentieth century; it might not be applicable to the mid-twenty-first century if there were a decline in the yield on alternative investments. Although this is a theoretical possibility, there is little historical evidence to support it. The world's capital stock has been growing relative to its labor force for two centuries, and although the return to capital has fallen over one or two decades in countries of exceptionally high investment such as Japan and Korea, it has not fallen significantly over the longer period of world industrialization.
Growth and Uncertainty
Two qualifications may, however, have to be introduced to the "at least 10 percent" criterion. The first derives from the fact that gross world product can be expected to grow over the next century. Because the driving forces of climate change appear to be related to economic activity in general, many components of the cost of global climate change may also be proportional to gross world product. To the extent that is true, the cost of climate change will increase with economic output. Although discounting reduces the present value of more distant costs, the possibility that they may grow over time cuts in the other direction.
It is ordinarily assumed in economic analyses that gross world product per capita will grow at a rate between 1 and 3 percent per year over the next 50 years, a magnitude that depends partly on advancing technology and partly on the rate at which existing technology is absorbed by economies that are not operating at the frontier of existing technology. If the costs of climate change are proportional to per capita gross world product, and if that quantity grows, for example, at 3 percent per year, then the yield criterion for mitigation actions drops to "at least 7 percent."
It should be noted that the costs of global climate change may not be proportional to gross world product. As change begins to be evident, societies will take adaptive action, and it is the cost of future adaptive action that is in part avoided by mitigation action now. However, the costs of some adaptive actions will not be proportional to gross world product. For example, the costs of building sea walls to protect against rising sea levels, if that should occur, may not increase with gross world product, even though flood damage in the absence of sea walls might increase with gross world product.
The second qualification concerns the uncertainty that attends predictions about future climate change and its associated costs. Uncertainty is an important topic, one that is conceptually different from comparing present costs with known future benefits. However, one aspect of uncertainty deserves mention because it could influence the choice of a discount rate, especially when the uncertainty cannot be entered directly into the cost-benefit calculation, as it should be.
The uncertainties associated with mitigating global climate change and its attendant costs are, in the current state of knowledge, at least as great asand probably greater thanthe uncertainties associated with other forms of investment that could be undertaken today. Accordingly, the investor averse to risk might conclude that costly mitigation actions should not be undertaken. However, the payoff from mitigation actions now will be greatest if the magnitude of global climate change and the associated costs turn out to be high, even if that is judged to be a contingency of low probability. Of course, if the costs associated with global climate change are low, any investment in mitigation actions will have a low return. Yet such investment may still be worthwhile as insurance against an uncertain but possibly costly contingency.
How do these considerations influence the discount rate? The precise answer is not at all straightforward, unless the uncertainty itself is related in a particular way to the passage of time. Roughly speaking, however, one can say that where an uncertain outcome (the future payoff from mitigation actions) is negatively correlated with overall economic prospects (as measured by future gross world product per capita), and where the uncertainty grows exponentially with time, some deduction from the discount rate used to evaluate mitigation actions is warranted. How much? That depends in detail on the nature of the uncertainty, an issue that remains to be clarified, and on the degree of our aversion to risk. Presumably, it was this sort of consideration that led U.S. policymakers in 1980 to stipulate a discount rate of only 7 percent for energy-related projects, 3 percent lower than the general standard for government investments.
In this report, most mitigation policies are analyzed at three different discount rates3, 6, and 10 percentso that decision makers may choose as they wish. In addition, the efficiency options are also analyzed at 30 percentto represent the rate of return that studies show consumers currently need before they will invest in energy efficiency.
Generic Alternatives in a Least-Cost Strategy
Investment involves choosing among alternatives uses of resources. As discussed in the preceding section, these choices are significantly influenced by policy choices, including the discount rate. With policy guidance
on this and other key assumptions, it is possible to develop a cost-effective portfolio of investments in mitigation (for a similar application, see Northwest Power Planning Council, 1986).
Finding the least-cost mix of responses to greenhouse warming entails comparing all the different mitigation responses. Figure 20.5 illustrates that the least-cost plan will probably involve a mix of responses. For simplicity, only two hypothetical options are plotted. They are shown as curves giving the cost for achieving various reductions in greenhouse gas emissions (or the equivalent: removal of greenhouse gases from the atmosphere, blocking of incident radiation, or changing of the earth's reflectivity). For comparability, all responses are translated into CO2-equivalent emissions.
Both options exhibit increasing cost for increasing reductions in emission (the curves gradually bend upward). If the only alternative were to achieve the desired level of reduction by choosing one option, the clear preference would be the hypothetical option B. Option B produces each level of reduction at lower cost than option A.
Several analysts (Edmonds and Reilly, 1986; Nordhaus, 1990) have pointed out the technical complications of making sensible comparisons among different greenhouse gases. The cost of reductions has been plotted along the vertical axis in terms of a ''levelized" cost (i.e., total cost over the period of analysis, divided by the number of years). Responses to greenhouse warming should be evaluated as investments, because the benefit that is sought will generally take a long time to appear. Consequently, it is important to compare costs over time, rather than simply in the particular years in which expenditures are made. Discounting the costs and benefits allows such a comparison. As discussed above, the choice of discount rate influences the comparisons made.
Figure 20.6 extends the comparison to additional options with different characteristics. Option C shows the "negative cost" or net positive benefits, associated with achieving the initial reductions in CO2 emissions. An example is energy efficiency, such as variable speed motors or compact fluorescent lighting. The cost of these measures would be less than the cost
of adding electricity generating capacity if the conservation measures were not implemented.
Option D illustrates a "backstop technology." A backstop technology provides an unlimited amount of reduction at a fixed cost. An example would be an abundant energy source that provides electricity with no CO2 emissions at all. Where a backstop technology exists, its cost sets a ceiling on the investment in reducing emissions. Only options costing less than D should be considered, no matter how much emission reduction is desired.
The heavy line labeled S in Figure 20.6 shows the cost-effective combination of options. Option C is selected up to the point at which option B becomes more cost-effective. Option A is added when it becomes cost-effective. S becomes horizontal when the cost reaches that of the backstop technology.
As the comparison of curves A and B indicates, the cost-effective portfolio contains a mix of alternatives. The level of expenditures is established by governments, who are guided by estimates of the benefits to be derived from mitigation, as well as budgetary considerations, international commitments, and other factors. The level of expenditure translates into a number of tons of greenhouse gas reductions; the objective is to get the largest reduction for that expenditure. This is shown by curve S in Figure 20.6, which outlines the mix of investments that produces any specified reduction at the least cost. At the point labeled b, for instance, all of the pairings from options B and C below the dashed line have been obtained, and acquisition of the alternatives at the bottom of curve A is beginning to be added. Additional savings from options B and C would also be pursued as the level of spending moves upward. (This discussion assumes that curves A through D describe independent activities, so as to avoid double-counting of savings. See Chapter 29 for an additional discussion of the problem involved in double-counting.)
Curves A through C all reflect a conventional assumption: that the cost of obtaining reductions in greenhouse gas generally increases as the size of the reduction is increased. Note, however, that curve C begins below zero. As discussed in Chapters 21, 22, and 23, there may be options available that are of net benefit to society even without accounting for the benefits of reduced greenhouse warming. These include some energy efficiency measures, such as variable speed motors or compact fluorescent lighting. As mentioned above, these actions may be worth more to electric utilities than the costs of producing and installing them because the improved efficiency allows the electric utility to defer expensive additions to generating capacity. In principle, they can therefore be provided at no cost to the homeowner because they reduce the total cost of serving that customer, provided the utility can reap a reward on its investment in energy efficiency. A substantial portion of the energy savings would reduce emissions of CO2 while
simultaneously producing economic benefits. There may be other options available whose costs are lower than the cost of the energy saved, even without placing a value on the reductions in greenhouse gas emissions.
What prevents these measures from being taken now are lack of information, inadequate economic incentives for utilities, high discount rates for personal consumption, and resistance to changing established methods. However, even technically feasible measures that benefit the national economy as a whole may not benefit every individual.
Research and development should both lower and flatten the supply curves in Figure 20.6, reducing the cost of alternatives and raising the scale at which they can be economically introduced. Research and development here includes social experimentation in areas such as mass transit, marketing of energy efficiency, and planting of trees on residential property, where consumer behavior has a substantial effect on the reductions achieved. More generally, research affects uncertainty. Although the supply curves in Figure 20.6 are drawn as lines, there is actually considerable uncertainty about how much reduction is available and at what price. The lines should be bands. The distinction between two technologies may not be as clear in practice as shown for curves A and B in Figure 20.6.
Timing of Mitigation Policy and Transient Effects
As further described in Appendix B, another important consideration in designing a mitigation policy is the timing and targeting of mitigation activities so that they have the desired impact on greenhouse warming. Therefore an important distinction to make is that human activities affect both the stocks and the flows of greenhouse gas emissions.
Greenhouse gas emissions occur in one time period, with some portion of the emissions sequestered immediately by the natural system (e.g., oceans) but with the remainder augmenting the much larger stock in the atmosphere that has developed over geological time due to the natural occurrence and long lifetime of many of these gases.
The response of climate may depend in complicated ways on both stock (atmospheric concentrations) and flow (emissions and absorptions into oceans, plants, and other reservoirs). Changing bothas is done in most mitigation approachesmay therefore produce nonlinear effects. Lowering emissions by 10 Mt/yr for 10 years may not have the same effect on greenhouse warming as lowering the stock of greenhouse gases by 100 Mt in a single year. This implies that different CO2 reduction patterns will have different effects on greenhouse warming with time and thus different benefits.
Therefore, in evaluating a prospective mitigation measure, one must examine the relationship between both the timing and duration of its reduction in greenhouse gases and the policy outcome desired.
Not only are there nonlinear effects due to the response of the natural environment, there also are important nonlinearities in social dynamics. An important body of knowledge has been accumulated on the reaction of various national economies to the energy price shocks of the 1970s. This analysis suggests that gradual change is likely to be significantly less costly than sudden imposition of a carbon tax or any other policy instrument designed to bring about a rapid change in CO2 emissions (Jorgenson and Wilcoxen, 1991). More generally, the transient effects of policy can be a large fraction of the total impact of attempts to mitigate greenhouse warming, particularly if the economic changes occur on a time scale of a year or shorter.
Thus timing is an important policy consideration. Climate change is a slow process in comparison with the rates of price fluctuations or changes in the business cycle. To the extent that institutions permit slow phasing in of policies such as carbon taxes, gradual changes are likely to be less disruptive economically.
Uncertainty and Choice of Parameters
Uncertainty cannot be ignored in responding to greenhouse warming. Errors of doing too much can be as consequential as errors of doing too little; the error of trying to solve the wrong problem is as likely as the error of failing to act. Above all, errors are inevitable, whether one acts or not, but inevitable errors are also occasions to learn. Therefore policy design that incorporates these lessons of the past helps to increase the resilience of the decision-making system and to foster future learning (Holling, 1978).
An initial step is to choose the range of parameters to be used in the analysis. The case of discount rate has been discussed here at some length, illustrating the social judgments at stake in making these quantitative assumptions. Note that what is needed is a range, rather than a single "best" value. If uncertainty cannot be avoided, one needs to know what would happen under different circumstances, so that serious errors can be forestalled and affordable ones identified.
Therefore, as illustrated in Chapter 29, after using the best information that the Mitigation Panel had available to evaluate the cost-effectiveness and emission potential of the various mitigation options at discount rates ranging from to 3 to 30 percent, the panel used its judgment as shown in Figures 29.1 to 29.3 to provide a range of values for the cost and potential of mitigation. This process culminates in Figure 29.5, which shows two curves: one with the highest cost and lowest emission reduction, the other with the lowest cost and highest emission reduction. This technological costing curve range is compared with the range developed using energy modeling as an accuracy check.
It is important to note that the mitigation options evaluated are merely technical choices. It is the policy judgments that are of instrumental importance, first, because a judgment of what to study shapes the kinds of conclusions that can be reached (Selznick, 1947; Kingdon, 1984) and, second, because governments are likely to be held accountable for their actions, including actions taken in analyzing large-scale changes. Both require policy-level involvement, as well as competent technical execution.
Because of this, the Mitigation Panel believes it is important to also evaluate various policy instruments that can be used in implementing the mitigation options. A list of some of the alternatives that have been proposed appears in Table 20.1. The list includes command-control instruments, economic incentives, revenue-neutral incentives, information programs, and redefinition of the mission and profits of utilities. The potential of these policy options for reducing the barriers to implementing the mitigation option is discussed in the evaluation of each option.
The charge to the Mitigation Panel was to "examine the range of policy interventions that might be employed to mitigate changes in the earth's radiation balance, assessing these options in terms of their expected impacts, costs, and, at least in qualitative terms, their relative cost-effectiveness." In this chapter, the panel has examined the two primary methods that can be used to evaluate greenhouse gas mitigation options: technological costing and energy modeling. While the energy modeling approach uses models that predict society's responses based on past societal behavior, technological costing attempts to determine the cost-effectiveness and emission reduction potential of future behavior and assumes that current public or private market imperfections can be overcome. The panel believes that the technological costing approach is better suited to evaluating the comparative advantages and disadvantages of specific mitigation options because current energy models do not have the spedicificity needed for such an analysis. For example, they look at the impact of a given carbon tax across the economy, but not the cost of specific methods for responding to that tax. However, there are reasons to be skeptical of the degree to which option-driven assessments can incorporate social responses (including market responses) to alternative courses of action. For example, although it is technically feasible at some cost to replace all coal-fired plants with nuclear power plants, social opposition to the installation of nuclear plants could prevent the option from being implemented. Yet, because energy modeling draws inferences from past behavior, the total cost of a shift to nuclear power may be overestimated, if there were to be widespread public reevaluation of the relative risks of climate change and energy technology, and if
TABLE 20.1 Potential Greenhouse Gas Mitigation Instruments
(Table 20.1 continued on page 197)
(Table 20.1 continued from page 196)
investments in nuclear engineering were to produce technical alternatives that were widely regarded as acceptable.
In conducting the analyses in subsequent chapters, the panel used the best and most reliable information available. But because more and better information is needed to determine the full social costs of mitigating greenhouse warming, the analysis presented in this report should be seen as a starting point on which future assessments can build. Despite the uncertainties described in this chapter, the components of a reasonable policy approach can be inferred from the discussion above:
• Although U.S. national policy is important, it is not by itself the determining factor in global greenhouse gas emissions.
• There are likely to be substantial economic impacts from controlling greenhouse gas emissions. Transient effects and transaction costs are important and potentially large, but they are highly uncertain, and methods for making usable predictions of these dynamic effects do not exist.
• Mixed strategies, aimed at cost-effective reductions of greenhouse gas emissions, are likely to be the best approach to mitigation. The timing and precise design of such a mix of policies are both significant and uncertain at present. It makes sense, accordingly, to emphasize that set of policies that is cost-effective.
• The ranking of options in terms of cost-effectiveness is strongly dependent on the choice of discount rate and a variety of uncertainties concerning technology, energy prices, and economic growth.
• The substantial uncertainties in both science and social science make errors inevitable. It is important, accordingly, to shape policies that can be resilient and that foster learning.
1. Throughout this report, tons (t) are metric; 1 Mt = 1 megaton = 1 million tons; and 1 Gt = 1 gigaton = 1 billion tons.
American Council for an Energy-Efficient Economy (ACEEE). 1990. Proceedings of the 1990 ACEEE Summer Study on Energy Efficiency in Buildings. Washington, D.C.: American Council for an Energy-Efficient Economy.
Atkinson, S. E., and R. Halvorsen. 1984. A new hedonic technique for estimating attribute demand: An application to the demand for automobile fuel efficiency. Review of Economics and Statistics 66(3):416–426.
Barker, B. S., S. H. Galginaitis, E. Rosenthal, and Gikas International, Inc. 1986. Summary Report Commercial Energy Management and Decision-making on the District of Columbia. Washington, D.C.: Potomac Electric Power Company.
Darmstadter, J. 1991. The economic cost of CO2 mitigation: A review of estimates for selected world regions. Discussion paper ENR91-06. Washington, D.C.: Resources for the Future, Inc.
Edmonds, J. A., and J. M. Reilly. 1983. A long-term global energy-economic model of carbon dioxide release from fossil fuel use. Energy Economics 5:74–88.
Edmonds, J. A., and J. M. Reilly. 1986. The IEA/ORAU Long-Term Global Energy CO2 Model: Personal Computer Version A84PC. Report ORNL/CDIC-16 CMP-002/PC. Institute for Energy Analysis, Oak Ridge Associated Universities. (Available from National Technical Information Service, Springfield, Va.)
Edmonds, J. A., J. M. Reilly, R. H. Gardner, and A. Brenkert. 1986. Uncertainty in Future Global Energy Use and Fossil Fuel CO2 Emissions for 1975 to 2075. Report DOE/NBB-0081. Carbon Dioxide Research Division, Office of Basic Energy Sciences, Office of Energy Research, U.S. Department of Energy. (Available from National Technical Information Service, Springfield, Va.)
Electric Power Research Institute (EPRI). 1988. DSM Commercial Customer Acceptance, Volume 1, Program Planning Insights. Report EM-5633, Project 2548-1. Palo Alto, Calif.: Electric Power Research Institute.
Geller, H. S., and P. M. Miller. 1988. 1988 Lighting Ballast Efficiency Standards: Analysis of Electricity and Economic Savings. Washington, D.C.: American Council for an Energy-Efficient Economy.
Geller, H. S., J. P. Harris, M. D. Levine, and A. H. Rosenfeld. 1987. The role of federal research and development in advancing energy efficiency: A $50 billion contribution to the U.S. economy. In Annual Review of Energy 1987, J. M. Hollander, ed. Palo Alto, Calif.: Annual Reviews, Inc.
Gladwell, M. 1990. Consumer's choices about money consistently defy common sense. Washington Post. February 12, 1990. A3.
Hausman, J. A. 1979. Individual discount rates and the purchase and utilization of energy-using durables. Bell Journal of Economics 10(1):33–54.
Hirst, E., J. Clinton, H. Geller, and W. Kroner. 1986. Energy Efficiency in Buildings: Progress and Promise. Washington, D.C.: American Council for an Energy-Efficient Economy.
Holling, C. S., ed. 1978. Adaptive Environmental Assessment and Management. New York: John Wiley & Sons.
Jorgenson, D. W., and P. J. Wilcoxen. 1989. Environmental Regulation and U.S. Economic Growth. Cambridge, Mass.: Harvard Institute of Economic Research.
Jorgenson, D. W., and P. J. Wilcoxen. 1991. U.S. Environment Policy and Economic Growth: How Do We Fare? Paper presented to the American Council for Capital Formation Center for Policy Research, September 12, 1991.
Kingdon, J. W. 1984. Agendas, Alternatives, and Public Policies. Boston: Little, Brown.
Koomey, J. 1990. Energy Efficiency Choices in New Office Buildings: An Investigation of Market Failures and Corrective Policies. Ph.D. dissertation. University of California, Berkeley.
Koomey, J., and M. D. Levine. 1989. Policies to Increase Energy Efficiency in Buildings and Appliances. Report LBL-27270. Berkeley, Calif.: Lawrence Berkeley Laboratory.
Krause, F., E. Vine, and S. Gandhi. 1989. Program Experience and Its Regulatory Implications: A Case Study of Utility Lighting Efficiency Programs. Report LBL-28268. Berkeley, Calif.: Lawrence Berkeley Laboratory.
Lashof, D. A., and D. A. Tirpak, eds. 1991. Policy Options for Stabilizing Global Climate. Washington, D.C.: U.S. Environmental Protection Agency.
Lind, R. C., ed. 1982. Discounting for Time and Risk in Energy Policy. Washington, D.C.: Resources for the Future and Johns Hopkins University Press.
Manne, A. S., and R. G. Richels. 1990. Global CO2 emission reductionsThe impacts of rising energy costs. Preliminary draft. February 1990.
Manne, A. S., R. G. Richels, and W. W. Hogan. 1990. CO2 emission limits: An economic cost analysis for the USA. Energy Journal 11(2):51–85.
Meier, A., and J. Whittier. 1982. Purchasing patterns of energy efficient refrigerators and implied consumer discount rates. In Proceedings of the 1982 ACEEE Conference. Washington, D.C.: American Council for an Energy-Efficient Economy.
Nadel, S. 1990. Lessons Learned: A Review of Utility Experience with Conservation and Load Management Programs for Commercial and Industrial Customers. Albany: New York State Energy Research and Development Authority.
Nordhaus, W. D. 1989. The economics of the greenhouse effect. Preliminary draft, June 4, 1989. Available from W. D. Nordhaus, Department of Economics, Yale University, New Haven, Conn.
Nordhaus, W. D. 1990. To slow or not to slow: The economics of the greenhouse effect. Paper presented at the 1990 Annual Meeting of the American Association for the Advancement of Science, New Orleans, La., February 15–20, 1990.
Northwest Power Planning Council. 1986. Northwest Conservation and Electric Power Plan. Portland, Oreg.: Northwest Power Planning Council.
Office of Management and Budget (OMB). 1974. Discount Rates to Be Used in Evaluating Time-Discounted Costs and Benefits. OMB Circular A-94. Washington, D.C.: Office of Management and Budget.
Peters, J. S. 1988. Lessons in Industrial Conservation Program Design. In Proceedings of the 1988 ACEEE Summer Study on Energy Efficiency in Buildings. Washington, D.C.: American Council for an Energy-Efficient Economy.
Pohl, G., and D. Mihaljek. 1989. Project Evaluation in Practice. Washington, D.C.: World Bank. December.
Reid, W. V. C. 1989. Sustainable development: Lessons from success. Environment 31(4):29–35.
Reischauer, R. D. 1990. Statement of the Director, Congressional Budget Office, before the Committee on Energy and Natural Resources, U.S. Senate. March 1990.
Rosenfeld, A. H., R. J. Mowris, and J. G. Koomey. 1989. Policies to improve energy efficiency and reduce global warming. Strategic Planning and Energy Management 9(2):7.
Ross, M. 1989. Improving the efficiency of electricity use in manufacturing. Science 244:311–317.
Ruderman, H., M. D. Levine, and J. E. McMahon. 1987. The behavior of the market for energy efficiency in residential appliances including heating and cooling equipment. Energy Journal 8:101–124.
Selznick, P. 1947. TVA and the Grass Roots. Berkeley: University of California Press.
Stone, D. A. 1988. Policy Paradox and Political Reason. Glenview, Ill.: Scott, Foresman.
Train, K. 1985. Discount rates in consumers' energy-related decisions: A review of the literature. Energy 10(12):1243–1253.
Train, K., and P. C. Ignelzi. 1987. The economic value of energy-saving investments by commercial and industrial firms. Energy 12(7):543–553.
U.S. Department of Energy (DOE). 1988. Technical Support Document: Energy Conservation Standards for Consumer Products: Refrigerators, Furnaces, and Television Sets. Report DOE/CE-0239. Washington, D.C.: Building Equipment Division, Conservation and Renewable Energy, U.S. Department of Energy.
Vine, E. 1985. State Survey of Innovative Energy Programs and Projects. Report LBL-19126. Berkeley, Calif.: Lawrence Berkeley Laboratory.
Vine, E., and J. Harris. 1988. Planning for an Energy-Efficient Future: The Experience with Implementing Energy Conservation Programs for New Residential and Commercial Buildings, Volumes 1 and 2. Report LBL-25525. Berkeley, Calif.: Lawrence Berkeley Laboratory.
Wilson, D., L. Schipper, S. Tyler, and S. Bartlett. 1989. Policies and Programs for Promoting Energy Conservation in the Residential Sector: Lessons from the Five OECD Countries. Report LBL-27289. Berkeley, Calif.: Lawrence Berkeley Laboratory.