Cap and Trade is Better Climate Policy than a Carbon Tax

Transcription

1 Cap and Trade is Better Climate Policy than a Carbon Tax Robert Repetto Senior Fellow United Nations Foundation May 2013 This paper explains why a cap-and-trade system, not a carbon tax, should be the foundation for a comprehensive U.S. policy to reduce greenhouse gas emissions and the growing threat of global climate change. While both approaches would create a comprehensive price on carbon and provide the economic incentive needed to motivate economy-wide efforts to reduce emissions, a cap-andtrade policy has political advantages and is more consistent with policies already adopted in the U.S. and abroad. A cap-and-trade system also provides more assurance that greenhouse gas emissions would decline sufficiently to avoid catastrophic damages from climate change the most compelling risk that such policies must address. Although the cap-and-trade option has been demonized by political opponents in the recent past, it would be a much better approach than the partial and less effective efforts now in place and would be a better policy choice than the carbon tax that is once more under discussion. The Need for Rapid and Sustained Reductions in Greenhouse Gas Emissions National climate policies should be adopted to ensure a rapid and sustained reduction in greenhouse gas emissions. The need is increasingly obvious and growing. The latest National Climate Assessment, 1 recently released in draft form by the U.S. Global Change Research Program, states unequivocally that temperatures are increasing, precipitation patterns are changing, and the frequency and intensity of storms have been altered as the result of increases in greenhouse gas concentrations. The Assessment states that the impacts and damages are already being felt throughout the country, especially from extreme weather events made more frequent and, in many respects, more intense because of global warming. Heat waves and droughts, such as those enveloping Texas and the Southwest in 2012, 1

2 are becoming more likely and will be even worse in the future if global warming continues. The most intense Category 4 and 5 hurricanes in the Atlantic have become more frequent, threatening the kinds of catastrophic damages caused by Sandy and Katrina. Intense rains have become more common, resulting in more damaging floods in the Midwest and Northeast. Human impacts on the climate are already evident and highly damaging to America s people and economy. The Assessment also warns that in the future it will only get worse. Despite the dangers, total greenhouse gas emissions are increasing at a pace more rapid than in the most pessimistic scenarios previously accepted by climate modelers, which means that increases in atmospheric concentrations are actually accelerating, not stabilizing. 2 If this continues, without sustained and significant cuts in emissions, the National Climate Assessment warns that in the decades beyond 2030 the country will suffer potentially catastrophic damages. Increases in average temperatures would reach 5-10 degrees F., making heat waves normal throughout the country and conditions in the Midwest, South, and Southwest insufferable. Water shortages would become acute; fires would rage as long as forests survived; pests and diseases would proliferate; crop yields would fall. Along the seaboards, the rise in sea level could reach six feet over the century as melting of ice accelerates at high latitudes. As more intense hurricanes and winter storms would then ride in on the higher tides and storm surges, life and property would be unsafe for millions in coastal cities and settlements. These ominous threats call for a vigorous, comprehensive, and sustained greenhouse gas abatement policy to be implemented as soon as possible. The goal should be to stabilize emissions within the next five years and then reduce them by 3 percent per year over the next 50 years. If this is accomplished, emissions levels would be about two-thirds lower than their current level by 2050 and 80 percent lower by According to climate researchers, this is the trajectory required of all major emitters if global warming is to be kept to 2 degrees Centigrade (3.6 o F.), the upper limit of what would avoid major damages. Unilateral actions by the United States would not be sufficient, of course, but most other major emitting countries have already begun to adopt national climate policies or have such policies in place and would be encouraged to go further if the United States takes action. 2

3 This transition is feasible. Starting gradually, it can be accomplished by faster improvements in energy efficiency and a gradual but sustained shift toward low-carbon and non-carbon energy sources. The United States, because of its historically low energy prices, would begin this transition from a startlingly low level of energy efficiency, compared to its peers. There is room for much improvement. The United States is also favored with an abundant resource base of wind and solar potential and natural gas. Though a transition of this scope might seem unrealistic to some, the United States has had half-century energy transitions of similar magnitude in its past: from water power to steam power, from steam power to electricity, from animal traction to motorized traction. In each case, the transitions led not to economic stagnation but to surges in investment, productivity, and prosperity. The latest transition is already under way. In 2012, for example, for the first time in history, more wind generating capacity than any other energy option was added to meet U.S. electricity demand. Thirteen gigawatts of additional wind capacity came on line, outpacing all other sources of electricity generation. During 2012 solar PV installations increased 84% over the previous year, and 50 coal-fired plants were retired, while only one new coal-fired power plant went online. 3 The Need for a Price on Carbon In order to bring about a rapid and sustained reduction in emissions without burdening the economy unnecessarily, all opportunities to reduce emissions at reasonable cost must be implemented, addressing not only carbon dioxide emissions from power plants and vehicle tailpipes but also emissions from the heating and cooling of buildings, from farming, and from all industrial and commercial activities. Low-cost opportunities to reduce emissions exist throughout the economy. Ignoring them by concentrating efforts only on power plants and vehicles, which together contribute about 60 percent of the total, would mean significantly higher overall costs. An effective economy-wide program requires that an economic penalty be attached to all or nearly all greenhouse gas emissions, whatever their source. That penalty is the so-called price on carbon. It would raise the cost of releasing 3

4 carbon dioxide from fossil fuel combustion; it could also be used to raise the cost of releasing methane and other greenhouse gases from activities where such releases can be measured and monitored. There is no equally effective alternative to a price on carbon: It would be administratively and politically impossible to regulate all emission sources, large and small. Establishing an economic penalty, or price, on greenhouse gas emissions would have four advantages. First, it would engage the thoughtful attention of every household and enterprise to the effort to reduce emissions. Each emitter, direct or indirect, would face higher prices for fossil-based energy and for energyintensive goods and services. Each one would then be motivated to find ways to minimize those cost increases in the least onerous way. All the resourcefulness and creativity that the economy and population can muster would be engaged in the effort, not because of environmental commitment or citizenship but because of economic self-interest. Although a large majority of Americans believes that greenhouse gases should be reduced, that alone has not been enough to motivate their actions on nearly the scale required. Values must be supported by selfinterest. Second, it would promote the most cost-effective solutions to the problem. Facing a price on carbon, each party would be motivated to explore whether to change behavior, change equipment, develop novel solutions, or simply pay the higher cost and continue as before, either temporarily or permanently. Changes that would be cheaper than paying the price on carbon, without sacrificing other advantages, would be adopted; those that would not be economically advantageous would be rejected. The decisions would be made by those who know best the advantages and disadvantages of each option in each circumstance, not by regulatory authorities who would have to promulgate rules of general applicability. As a result, with all enterprises and households facing the same price on carbon, only those solutions would be adopted that are economically superior to paying the price. This would ensure that all reductions take place at costs that are at or below the price on carbon and none at higher cost. Third, establishing a price on carbon would use the strength of the marketplace to correct an economically costly and potentially disastrous failure in the marketplace. The failure is that fossil energy and goods using fossil energy are 4

5 sold at prices that don t reflect their true economic costs. The omitted costs are most notably those of increasing damages from climate change. Each ton of carbon emitted raises those damages and risks, yet the price to users fails to reflect that cost. Therefore, the costs to society and the economy from additional fossil energy use exceed the price paid by the purchaser. A price on carbon can correct that discrepancy, bringing costs and benefits into alignment and allowing private markets to operate more efficiently. Fourth, a price on carbon would stimulate technological advances. The flexibility for businesses to decide how, where, and when to mitigate their emissions permits innovation and provides an economic incentive to do so. Finding cheaper and better ways to reduce emissions would lower a firm s carbon tax liability or leave it with less need for carbon permits in a trading system. A price on carbon also generates business opportunities for firms that provide equipment, materials, and services to other businesses and that are frequently the main source of innovation. The advantages of a price on carbon over more limited regulatory approaches are easy to understand. Vehicle fuel efficiency standards, for example, reduce fuel use and tailpipe emissions for every mile the vehicle is driven, but they do nothing to discourage vehicle use, as higher fuel prices would. In fact, better fuel efficiency encourages drivers to drive more miles because costs per mile are lower. This is the so-called rebound effect and applies to energy efficiency standards for appliances and other equipment as well. Also, the standards apply only to new vehicles, which make up less than 10 percent of the vehicles on the road in any year. It takes a decade or more for higher standards to work their way through the entire fleet. Worse, if the higher fuel efficiency standards result in a higher initial price for vehicles, some drivers keep their older vehicles on the road longer or purchase used cars. A price on carbon would escape all such pitfalls. It would encourage people to buy, and automakers to sell, more fuel-efficient vehicles. It would encourage reductions in vehicle miles traveled as well. Finally, it would encourage people to trade in their older, less efficient vehicles. A price on carbon provides a comprehensive set of incentives to reduce emissions. 4 Establishing a price on carbon is not just the preferred policy of economists and most environmentalists, it is also endorsed by major U.S. and international 5

6 companies. At the UN climate talks last year in Doha, Qatar, a coalition of more than 100 large international corporations submitted a statement recognizing that there are multiple ways to set a price on emissions, but stating a strong preference for market-based approaches such as emission permit trading, which offer both environmental integrity and flexibility for business. 5 They are joined in this support by investor groups managing tens of trillions of dollars in assets. 6 Investors increasingly recognize that their assets are now jeopardized by global warming and see their financial risks and exposures growing sharply unless effective policies are put in place. The Essential Choice: Cap and Trade or Carbon Tax There are two ways in which an economy-wide price on carbon can be created. One way is to put in place a national cap-and-trade program; the other is a carbon tax. Both would impose the necessary economic penalty on releases of carbon dioxide from fossil fuel combustion and accomplish the objectives described above. Both could be structured in ways that could help achieve additional goals as well, such as penalizing other major sources of greenhouse gases and encouraging the capture of carbon dioxide from the atmosphere as well as encouraging the wider use of non-fossil energy sources. Both could be designed to shield vulnerable households and industries from rapid cost increases. Both could be designed to generate public revenues that could be used for deficit reduction, tax relief, or investments in research and development or public infrastructure. These are important similarities, but there is a crucial difference. The national cap-and-trade system envisaged here would require all producers and importers of fossil fuels to have annual permits issued by government matching the year s sales. Permits would reflect the tons of carbon released when the fuel is burned i.e., more emissions per ton for coal than petroleum, more emissions per ton for either than natural gas. The total number of permits issued in any year would equal the total tonnage of carbon that could be released as carbon dioxide in that year. The total tonnage permitted would decline by 3 percent each year. 6

7 The federal government would oversee the creation of a market, similar to other commodity markets, in which such permits could be bought and sold. These trading options would enable firms whose operations are marginal or that have low abatement costs to sell unneeded permits to firms whose abatement options are more expensive and that could continue to sell fossil fuels profitably. In this way fossil fuel use would decline in the most cost-effective way. A market price would be established and regularly quoted, just as market prices for crude oil or gold are. The cost of permits would add to the cost of fossil fuels because sellers would have to recover the cost of their permits or compensate themselves for lost revenue in not selling their permits. Buyers would adjust their purchases to the higher prices. The market price for permits that would emerge would be the price on carbon. As the number of permits falls by 3 percent per year, the availability of fossil fuels would shrink relative to market demand, and the price of permits would tend to rise over time. Permit prices would also change in response to market conditions, such as business cycles and underlying market demands for fossil fuels. Under this cap-and-trade regime the higher cost of fossil fuels due to the price on carbon would work its way through the economy to all activities using fossil energy directly or indirectly. Businesses and households facing higher costs would have unfettered options about how, when, where, and if to respond to those higher costs. Costs would rise in electricity generating plants burning coal or natural gas, encouraging fuel switching and greater investments in higherefficiency plants and in renewable energy. Electricity customers would see somewhat higher rates, motivating them to consider cost-effective efficiency upgrades. Vehicle owners would face somewhat higher gasoline and diesel prices, encouraging changes in driving and equipment. The price on carbon created by a cap-and-trade system would prompt flexible economy-wide responses. It is not the purpose of this paper to explain in detail the many design choices that would be made in installing a national cap-and-trade system. That has already been done very ably elsewhere. 7 Most of those choices have already been faced and resolved in the several cap-and-trade systems already operating in this country and in the systems in operation or being designed in other countries. Rather, this paper examines the critical differences between such a cap-and-trade system and the alternative, a carbon tax. 7

8 A carbon tax also would be applied on the basis of the amount of carbon emitted when a ton of fossil fuel is burned, and so would be higher on coal than petroleum and higher on petroleum than natural gas. It would be levied on the same producers and importers of fossil fuels that would be subject to a cap-andtrade system. The cost of the tax would be paid by the initial seller and then added to the selling price of the fuel, establishing a price on carbon. This price would subsequently affect the costs and prices of all goods and services using fossil fuels directly or indirectly. Like a cap-and-trade system, it would provoke and allow flexible and cost-effective responses by energy users. This is the basic similarity of the two approaches. The two systems would be equally easy to administer because the companies subject to permit requirements and those subject to the tax would be the same 2,000 or so companies engaged in the upstream supply of fossil fuels to the U.S. economy. Both systems would also be based on the same information regarding annual sale volumes. A carbon tax, however, sets the price on carbon directly through the tax rate, while a cap-and-trade system does so indirectly, based on the cost of avoiding or mitigating emissions from fossil fuels. Therefore, if the policy objective was to increase the price of carbon gradually over time, in a carbon-tax system that would require raising the tax rate; under a cap-and-trade system, price increases would be effected by reducing permit availability. The crucial difference between the two approaches is that a carbon tax directly affects the price of carbon and relies on the response of market forces to bring about a reduction in fossil fuel use and carbon emissions; a cap-and-trade system directly reduces fossil fuel use and carbon emissions and relies on market forces to set the price on carbon. Thus, with a carbon tax, one would expect the impact on emissions to vary from year to year. The impact would be smaller when the economy is booming and the demand for energy is strong, and also smaller when the price of natural gas (or renewable energy) is high relative to that of coal, making fuel switching by electric utilities less attractive. The impact would be greater when the opposite conditions exist. Under a cap-and-trade system, the declining supply of permits would ensure that emissions fall steadily, little affected by market conditions, although banking and borrowing of permits would allow some response to market conditions because permits banked in slack times could be utilized in boom times. Market conditions would primarily affect the price of 8

9 permits and, thus, the price on carbon. Prices for permits would rise when the demand for permits and the underlying fuels is strong; they would fall when, for one reason or another, the demand is weak. A carbon tax would allow the government greater control over the price on carbon and thereby over the costs that businesses and households would incur to reduce their use of fossil fuels. This, however, comes with a disadvantage: less control over the extent and pace of emission reductions. A cap-and-trade approach gives the government considerably more direct control over the extent and pace of emission reductions but less control over the costs that will be incurred in the economy to bring them about. If permit prices are high, higher cost abatement options will seem attractive; if permit prices are low, even quite cost-effective investments might have little appeal. The results of government miscalculation would therefore be quite different between the two options: If, for one reason or another, the government issued an oversupply of permits as occurred under the European Union s Emissions Trading System permit prices would be low; if the government set the carbon tax too low, the reduction in carbon emissions would be meager. A second important difference is that a carbon tax, like other taxes, would automatically generate federal government revenues. These potential revenues would be significant, of the order of magnitude of tax increases adopted under the recent fiscal cliff negotiations. For this reason, some economists and others have found a carbon tax to be an attractive way to reduce the deficit or to reduce reliance on distortionary taxes within the federal tax structure while also helping to reduce climate change. Using revenues to reduce reliance on distortionary taxes would significantly reduce the economic impacts of climate policy. 8 A cap-and-trade policy could also generate federal revenues, but the extent would depend on government decisions whether to auction permits or to give some or all of them out freely. The government would have to decide what portion of them should be auctioned, what portion should be given away, and to whom they would be given. Permits conferring the right to sell increasingly scarce fossil fuels would be valuable marketable assets. The recipients might be fossil fuel sellers or their customers such as electric utilities or energy-intensive industries or the general public as compensation for higher costs of living. 9

10 Which approach is better in combating global warming? The choice is important because climate change is a long-term problem demanding a long-term solution. If one option is chosen, it is very likely that it will persist for decades and very unlikely that it will be abandoned in favor of the alternative. 9 When any important policy is adopted, it prompts the creation of institutions and systems that are specific to that approach and not transferable to another. Businesses and households learn how to manage under that policy, not the alternative. Interest groups that benefit under the policy grow up to defend it against change. Whichever approach is chosen will be the foundation for climate policy for decades. The Choice: Political and Policy Considerations The strongest political argument against the carbon-tax option is that its chances of being passed by Congress in the foreseeable future are negligible. 10 In the House of Representatives, the Republican majority has displayed its strong opposition to tax increases of any kind. A small increase in the tax rate on only the very wealthiest people, while preserving most of their lucrative tax breaks, required a Presidential election and then a major political battle. Most members of the Republican caucus have signed a pledge never to vote for any tax increases. With the election of many Tea Party adherents in 2010 and the increasing gerrymandering of districts into safe seats, the degree of polarization between Republican and Democratic Representatives is higher than ever. Most Representatives represent safe districts where any electoral challenge would come not from a moderate but from a challenger even more extreme. This polarization affects not only issues related to taxes and government spending but also environmental policy, especially climate policy. Denial of climate change is nearly a litmus test of Republican right-wing conservatism. The Republican leadership has already announced that it would not even consider a bill to enact a carbon tax. Moreover, it is doubtful that a meaningful carbon tax could command a majority in the Senate, despite the slim Democratic majority, let alone the 60 votes necessary to overcome an inevitable filibuster. Several Democratic senators from fossil energy-producing states have consistently opposed efforts to address climate change at the expense of coal and oil interests and would align with Republicans. 10

11 The Obama administration recently announced that it has no intention of proposing a carbon tax, recognizing lack of Congressional or popular support for tax increases. In surveys, most people who say that climate change is a serious problem and who support government action to stop it still do not favor a carbon tax or higher energy taxes as a policy option. 11 Even if a carbon tax could be enacted, the fact that Congress could revisit it every year or two in the context of tax legislation does not inspire confidence that it would be allowed to ratchet steadily upward in a predictable way to reduce emissions. Prospects for enactment of a cap-and-trade system admittedly are not much better. When last attempted, a bill narrowly passed the House despite a Democratic majority and considerable concessions to gain reluctant votes. The Senate version did not have sufficient votes for passage over a probable filibuster and was not even brought up for consideration. Nonetheless, the cap-and-trade option has several political advantages. Opponents have recognized this by attempting to rebrand cap-and-trade as cap-and-tax during the recent presidential campaign. Capping and reducing emissions is a direct and obvious way to address the climate problem, and a direct approach is easier for people to understand. In a capand-trade system the public policy discussion can be focused on the environmental and scientific issue of how fast emissions should decline, not on the price on carbon that would result although economic impact estimates will inevitably compete for center stage. This is an advantageous frame: A majority of voters has consistently favored environmental regulations that reduce pollution-causing emissions, and a growing majority recognizes that greenhouse gas emissions are responsible for damaging climate changes. Extreme weather events, such as hurricanes, floods, heat waves, droughts, and fires, have increasingly brought home to people that climate change is causing damages even now, not just at some time in the distant future. The media is reinforcing this recognition by a shift in its coverage of climate-related events. In contrast, the public policy discussion of a carbon tax will inevitably focus on its economic impacts, immediately raising the specter of job losses and industrial flight. Other environmental regulations also raise production costs, but the public approves of them nonetheless, largely because any effects on consumer prices are mediated through complex market processes and are usually small and difficult to 11

12 attribute specifically to regulation. These partially hidden cost repercussions are politically more acceptable to the public than those that are explicit and readily tied to political decisions. For example, the most prominent cap-and-trade system now in operation in America, created to limit emissions that cause acid raid, has encountered minimal public opposition, partly because its costs are only one of many factors affecting electricity prices and partly because the costs have been considerably smaller than projected under a conventional, site-specific emissions cap. Both a cap-and-trade system and a carbon tax would raise costs and prices throughout the economy. These increases would be mildly regressive because lower-income households spend a larger proportion of their incomes on energyusing goods and services than wealthier households do and because low-income households may have fewer ways to adjust. These cost-of-living effects could be neutralized if permits were auctioned and some of the revenues were recycled to low-income households through changes in the tax code or in income support programs. Economists who have studied the issue find that the most effective way to offset impacts on low-income households would be through the Earned Income Tax Credit. 12 This approach, by accurately targeting the most vulnerable households, would be economically efficient but would have less political appeal than a broader distribution of revenues, as proposed in the cap-and-dividend scheme, under which auction revenues would be distributed uniformly to all households annually, maximizing public political support. 13 However, in the absence of a groundswell of public support, the decisive factor in Washington is the lobbying of organized industry groups that can muster squadrons of former Congressmen and Congressional aides as lobbyists and spend scores of millions of dollars in campaign contributions and lobbying campaigns. Until now, these groups have been able to soundly defeat the efforts of environmental groups and their allies. Yet a cap-and-trade option could capture more business support than a carbon tax. In the previous effort, a group of about two dozen large and influential corporations, including some electric utilities, allied with major environmental groups under the banner of the U.S. Climate Action Partnership to endorse the proposed cap-and-trade legislation. 12

13 Impending regulations may persuade these same corporations and more to support a cap-and-trade system again. A Supreme Court decision empowered the Environmental Protection Agency to develop regulations to reduce greenhouse gas emissions, and the agency is now finalizing rules applicable to new power plants. 14 After that, it is likely to begin drafting rules applicable to existing power plants and other large industrial emission sources. Facing this prospect, large companies potentially affected by regulation will recognize the cost advantages of a cap-andtrade regime, which would allow them much more flexibility over how, where, when, and if to reduce emissions. Their previous experience will tell them that this added flexibility would significantly diminish their compliance costs. 15 If the EPA does move forward, it will be important that the program provide maximum flexibility, said a spokesperson for American Electric Power, one of the country s largest electric utilities. 16 Recognizing this advantage, such corporations would be likely to drop their outright opposition to a cap-and-trade regime and turn their lobbying efforts to gaining the maximum advantage from the specific design of the system. This would probably involve, as before, efforts to persuade Congress to grant them as many free permits as they could get. Adding big business to environmental and other interest groups arguing for a cap-and-trade system would improve its chances of adoption, relative to those of a carbon tax. Studies have estimated that if companies subject to the cap are allocated even a minor fraction of their permits, rather than having to buy them at auction, their profits and stock prices would be insulated from any cost increases. The reason is that in the early years, the required reductions of 3 percent per year would increase the price of fossil fuels, offsetting the loss in sales volume and leaving the remaining permits on the balance sheet as valuable tradable assets. 17 The government s flexibility in distributing permits provides a tool for muting potential political opposition and easing the short-term impacts on particularly vulnerable sectors. Of course, companies could be granted rebates or exemptions from a carbon tax as well, but again, there is an important difference. Exemptions from a carbon tax would weaken its effectiveness in reducing emissions, but the distribution of permits in a cap-and-trade system does not affect the emission cap; in most circumstances it only affects the distribution of cost impacts. This is an additional advantage of the cap-and-trade approach. 13

14 This advantage also seals off the program from the effects of many significant subsidies to fossil fuel-producing companies that are embedded in the tax code and have proven politically impossible to dislodge. Under a cap-and-trade system, such subsidies would not affect total emissions, which would be capped, but would only affect the distribution of income and level of profits among producers and consumers. Even better, if permits were auctioned, the government would regain some of the revenue lost through the subsidies because energy companies bids would tend to be higher, the more heavily subsidized they are. Since the companies receiving the most lucrative subsidies would find fossil fuel sales more profitable, they would also tend to be among the highest bidders for permits, paying relatively high prices to the government. Therefore, the government would be able to negate some of the harm of politically entrenched fossil fuel subsidies and recapture some lost revenues, while avoiding a frontal assault on energy company interests. Better yet, any subsidies provided to renewable energy producers would survive intact, because those companies would not require permits in order to sell their output. Another advantage of cap-and-trade systems is the experience gained from their past success in the United States at both national and regional levels. When introduced to manage the phase-out of leaded gasoline, emissions that cause acid rain, smog-forming emissions in California, and even excessive water pollution, compliance was high, emissions fell on schedule or even faster, trading markets worked, innovative approaches to pollution control were encouraged, and compliance costs were lower than they would have been under conventional regulatory approaches. 18 There are now cap-and-trade programs for carbon emissions in place on both coasts, in California and in the Northeast. These regional programs have already encountered, considered, and decided many of the design questions that a national program would face. A national program would build on this history and would also eliminate leakages from regional programs that occur when emissions are displaced from one region to another. 19 Opponents of national climate policy often claim that any action to reduce emissions would be futile, despite the fact that the United States accounts for a fifth of the global total, because emissions in the rest of the world will continue growing and would swamp our efforts. In fact, the United States now lags most of the rest of the world in enacting effective climate policies. The European Union 14

15 has by now operated an international cap-and-trade system for five years. The lessons learned there would be valuable here. 20 So would the institutional mechanism that has been created, the Clean Development Mechanism, to encourage emission reductions in developing countries as offsets. China, which is often portrayed as an environmental renegade, has already enacted stronger climate policies than Washington and will put in place cap-andtrade systems in seven of its major cities and regions within the next three years, covering most of its greenhouse gas emissions. 21 Those systems are already being designed, and preliminary identification of emission sources is under way. The experience from this pilot project will guide the creation of a national Chinese capand-trade system. Other national and sub-national governments have also adopted cap-and-trade systems, including South Korea, Australia, New Zealand, Taiwan, the Canadian province of Quebec, and the Tokyo metropolitan area in Japan. Additional countries, even including Kazakhstan, have announced their intentions to install cap-and-trade regimes. As the richest and one of the largest emitters, action by the United States would encourage still more countries to take stronger action in response. The United States could also benefit from broader market opportunities if a domestic system is linked to systems in other countries. 22 Doing so would not only promote cost-effective emission reduction globally but also provide the United States with access to cost-saving reductions in other countries. 23 Linking cap-and trade systems internationally would be much easier than linking a U.S. carbon tax with foreign permit-trading regimes. The Choice from an Economic Perspective If the world were as simple as portrayed in introductory economics textbooks, the choice wouldn t matter. The government could fine-tune the rate of a carbon tax to reduce emissions until the extra pollution control costs of doing more would just offset the extra environmental damages that would be avoided. That would maximize the policy s net benefits. Or it could just set a cap on emissions to achieve precisely the same degree of pollution control and then auction the permits. Both ways would achieve the same results and work equally well. 15

16 The world is not so simple, however. The climate problem will persist for many decades, far beyond our abilities to predict the future. Even now, there is considerable uncertainty about the pace and extent that the climate will change in response to emissions and about the damages that will result, especially from extreme weather events. If atmospheric concentrations rise above the levels previously experienced during human history, the risks are unfathomable, but the outcome is unlikely to be a happy one. On the other side of the equation, there is also considerable uncertainty about the extent and speed with which the economy would respond to higher fossil energy costs, about the extent, speed and cost with which renewable energy systems could be expanded, and about the pace at which new technologies will be developed and introduced into the economy. The choice of policy options is a momentous problem of dealing with uncertainty. Placing uncertainty at the center of their analyses, many economists have treated the problem as one of dealing with the strong probability of a policy error. Recognizing that policymakers probably would not be able to identify either the perfect tax rate or cap, economists have asked which error would be worse: a deviation from the ideal that allowed damages to increase too quickly or one that allowed pollution control costs to rise too quickly. Framing the problem this way highlights the essential difference between a cap-and-trade system, which exercises more control over the level of emissions, and a carbon tax, which exercises more control over the pollution control costs that would be incurred. To reach the economic ideal, emissions should be reduced until the resulting rise in control costs just offsets the additional reduction in damages. Going further would raise control costs more than the savings in averted damages. Doing less would allow greater damages than the costs of more emission reductions. So, if damages would rise sharply beyond that ideal emissions level but control costs were rather flat, then a cap on emissions would be preferable to a carbon tax, which would allow emissions to deviate widely if the tax rate were set wrongly. If the situation were reversed and control costs changed drastically beyond the ideal emissions level but damage costs were rather flat, then a carbon tax, which allows more control over emission control costs, would be a better option than a direct emissions cap, which would be very costly if set wrongly. 16

17 This reasoning implies that the key economic question is how climate change damages and greenhouse gas control costs vary when the cap or the tax deviates from the ideal. If damages vary sharply but control costs don t, then a cap on emissions would probably result in the smaller excess loss, counting both pollution control costs and climate change damages. However, if control costs vary sharply but damages don t, then a carbon tax would probably produce a smaller loss. Following this line of reasoning, economists have compared the likely behavior of control costs and climate damages. The question is not just whether they rise or fall rapidly as emissions vary near the ideal level. What matters more is whether they rise or fall non-linearly, at an accelerating rate: the sharper the acceleration, the greater the gains and losses when the policy deviates from the ideal. If damage costs change more non-linearly when emissions change, then a cap is the better way to deal with uncertainty; if control costs vary more nonlinearly, then a tax is the better option. 24 Unfortunately, this analysis has proven to be of little practical help. The problem is that there is great uncertainty both about the extent of climate damages as emissions and concentrations increase and about how sharply those damages would rise or fall. There is also great uncertainty about the economic costs of reaching any particular emission target and perhaps even more uncertainty about how those costs would change for different emission targets. We face these uncertainties today, in our present imperfect state of knowledge, and are even more uncertain about how they will change in the future. Despite these difficulties, economists have combined conjecture, assumptions, and what few relevant facts they can muster to generate more insight for the policy debate. They note that carbon dioxide remains in the atmosphere for many decades, even centuries, and short-term variations in emissions may not matter much. What matters, they say, is that overall atmospheric concentrations stay within some safe or tolerable limit or budget over the next half-century and beyond; excessive emissions in one period can be balanced out by lower emissions in another. They also note that change in the climate takes place rather slowly, with considerable lags in the system, so impacts are likely to be gradual, and considerable adaptation to changing conditions is possible. On these grounds they 17

18 tentatively suggest that any acceleration or deceleration of damages in response to increasing or decreasing emissions would be rather mild. 25 Looking at the energy system, economists note that fossil-based energy is ubiquitous and essential to the economy. They note that wind, solar, and other renewable energy sources now make only a marginal contribution to the energy mix and are generally costlier. They point out that expanding the supply of renewables rapidly would encounter obstacles and possible limits. For such reasons, they conclude that costs would escalate rapidly as attempts to reduce carbon dioxide emissions intensify. These conclusions a more rapid escalation of control costs than of climate damages have been embedded in several climate economic models. For example, the influential DICE model, created by Yale professor William Nordhaus, assumes that climate damages rise as the second (squared) power of emissions, while control costs increase as the third (cubed) power of emission reductions. 26 Naturally, since control costs are assumed to change more non-linearly, when the model is used to analyze which policy choice is superior, it yields the answer that a carbon tax would be the better choice However, these assumptions seem increasingly unrealistic in light of more recent evidence. It is now recognized that most of the damages from climate change will not come from gradual changes in temperature and precipitation but from the increasing frequency and intensity of extreme weather events: heat waves, droughts, floods, violent storms, and hurricanes. 27 Damages from these events rise very non-linearly as their intensity increases. 28 The reason is that both human and natural systems have certain ranges of tolerance for climatic variations. Fluctuations within those ranges do little damage, but weather events that go beyond the tolerable ranges cause catastrophic damage. For example, Nordhaus estimated econometrically from past hurricanes that damages increase as the 8 th power of a storm s maximum wind speed. 29 That means, for example, if the maximum wind speed of a hurricane rises 10 percent, the resulting damages will more than double. Another study by his colleague Robert Mendelsohn concluded that damages rise at the 5 th power of maximum wind speed. 30 Scientists predict that the most intense Category 4 and 5 hurricanes, which rarely strike land but already do 84 percent of the damages from all hurricanes, will become more frequent. 31 Similarly, flood damages rise catastrophically when waters rise higher than infrastructure was designed to withstand, as both Katrina and Sandy demonstrated. 18

19 In heat waves, mortality rises sharply with the duration and persistence of excessive temperatures. In agriculture, crops can withstand heat and water stress for a while with little loss in yield, but if conditions get too bad, they fail completely. So, more extreme weather provoked by climate change implies sharply escalating damages. Scientists also recognize that there are strong non-linearities in the climate system itself. Important positive feedbacks within the system imply that warming driven by rising greenhouse gas concentrations can feed on itself, provoking further warming in an escalating sequence. In the Arctic, for example, warming is reducing the extent of sea ice cover, so more of the Arctic Ocean is exposed to the sun and air for more of the year. As a result, less of the sunlight shining on the Arctic is reflected back into space by snow and ice, and more warmth is being absorbed by the darker ocean waters. Warming generates further warming, further reducing the sea ice, in a process that feeds on itself. Recent evidence indicates that this process is well under way. Peer-reviewed studies published in the last five years show that Arctic sea ice is melting 40 percent faster than the IPCC estimated in Satellites have also found that sea levels have risen globally by an average of 3.4 millimeters per year from 1993 to 2008, 80 percent more than the 1.9 mm that the IPCC estimated for that period. 32 On land, in sub-arctic latitudes, global warming is melting the permafrost, releasing carbon dioxide and methane, a more powerful greenhouse gas. Lakes formed on the melting landscape can be seen releasing these gases in bubbles rising through the water. As these gases are released, they produce more warming, which in turn causes more melting. Vast amounts of carbon and methane are locked in the permafrost or in the Arctic seabed. Their release over time will be another process that feeds upon itself. 33 Similar positive feedbacks can also be seen on land at lower latitudes. Warming, drought, and pest infestations are increasing the forest area burned each year, an area already as large as Texas. Forest fires not only release carbon from the trees and underlying forest soils, they also reduce the amount of carbon dioxide removed from the atmosphere by vegetation. All of these feedbacks and nonlinearities together imply that keeping within an overall long-term emission budget may not be enough to prevent a sharp increase in damages. Any level of emissions 19

20 that sets off or strengthens these self-reinforcing processes may accelerate global warming and its impacts dramatically. Since these feedbacks are already clearly at work at current concentration levels, today s emissions may carry exactly those risks. The evidence suggests that these non-linearities in the climate system, together with the non-linearities in damages caused by more frequent extreme weather events, mean that incremental damages from rising emissions may rise sharply. What about the other side of the equation? Would the costs of reducing emissions rise sharply if policy is enacted to achieve the reduction needed? Recent experience suggests otherwise. It was previously claimed that moving away from coal-fired generation would be difficult and costly, but precisely that process is already well under way. Plans for dozens of new coal-fired plants have been cancelled, and few will be built in the years ahead. Dozens more existing plants have been retired. It turns out that coal-fired power isn t so cheap after all. Most of the plants now being retired are old, inefficient, dirty facilities that were kept in service mainly by a quirk in environmental regulations that allowed them to avoid installing expensive pollution controls. The environmental damages caused by their releases of mercury and other heavy metals, particulates, and smog-forming exhaust gases were found to be greater than the value of the electricity they produced, even ignoring the cost of their greenhouse gas emissions. 34 When their owners were faced with the necessity of installing pollution controls and the possibility of future climate regulations, they chose to shut down the plants instead. High-efficiency power plants burning natural gas are now cheaper both to build and operate than coal-fired plants and will have a much greater cost advantage when there is a price on carbon emissions, as they produce only half as much greenhouse gas emissions. If the natural gas can be extracted and transported without significant methane leakage, the switch toward gas would lower emissions on balance. Estimates of the reserves of natural gas available in the U.S. and other countries others have risen dramatically, as have production levels, and natural gas prices have fallen by more than half in the last five years. These large reserves imply that the replacement of coal with gas in electricity generation can continue for many years without any constraint on resource availability. 20

21 There is an even larger potential for increased use of wind and solar energy. The resource base for expansion is huge: Just one county in Arizona has enough solar potential to power the entire country. This expansion is already under way: Installed solar and wind capacity has been increasing by about 20 percent per year for many years now. Costs have fallen sufficiently that the economics of wind and solar are attractive in many applications, and a price on carbon would make them more so. Skeptics point out that because wind and solar power are intermittent and not dispatchable when needed to meet demand, their role is inherently limited, but this constraint is overstated: Renewable energy plants can be complemented by natural gas plants that are dispatchable or by energy storage technologies or demand-side management options. Experts conclude that at least a ten-fold increase in their share of electricity markets is feasible. The impediments to rapid expansion are institutional, not technical or economic. These include, for example, the complexities of siting and financing new interstate transmission lines to bring solar and wind power from their rural sources to urban load centers; the overlapping jurisdictions of federal, state, and local agencies that must review and issue permits for offshore wind farms; and the need for wind power developers to go back to Congress every two years or so to plead for renewal of the tax credits that put them on a par with fossil fuels in the absence of a price on carbon. These institutional barriers can be corrected, allowing clean energy options to expand without facing substantially higher costs. Indeed, the premise that costs will escalate rapidly as emissions are reduced is questionable. Renewable energy costs have fallen dramatically over the past three decades as installed capacity has expanded. Much of this cost reduction can be attributed to the expansion itself economies of scale in production, learning by doing efficiencies in manufacturing and installation, and investments in research and development that were stimulated by growing market opportunities. Thousands of patents have been granted in recent years for improvements in clean energy technologies. Any cost curve for emission reductions that could be estimated today would not be applicable to emission reductions taking place years in the future. Technological change will probably mitigate any tendencies for costs to rise substantially, especially if climate policy is adopted that allows full flexibility regarding the methods used to control emissions. 21

22 For all of these reasons, one is led to the conclusion that damage costs increase more sharply as emissions rise than control costs increase when emissions are reduced. This implies that from an economic perspective a cap-and-trade policy would be superior to a carbon tax as a means of avoiding costly errors in setting targets in a world of great uncertainty. The economic advantages reinforce the political and other policy considerations discussed previously. Further Uncertainties and Risks The advantage of the cap-and-trade option is multiplied if an additional condition holds true: the greater the uncertainty regarding the appropriate carbon tax rate, the greater the advantage of cap and trade. 35 If an error in setting the tax rate is likely to be more costly than an error in setting the cap, then it follows that the advantage of a cap is greater, the greater the range of uncertainty about the appropriate tax rate. The wider that uncertainty, the more likely a serious error would be made if a carbon tax were attempted. In fact, there is great uncertainty regarding the appropriate carbon tax, which should not be an unchanging rate but rather one that rises gradually over time. The Energy Modeling Forum at Stanford University has for years invited the country s leading economic modelers to compare their models forecasts when simulations are based on identical assumptions. All models are run using the same assumptions about such conditions as the rate of future economic growth, the level of world energy prices and other relevant variables. The models are then asked what tax rate schedule would be needed to achieve a specified emission reduction trajectory. Unfortunately, even when these important underlying assumptions are identical and the models are asked to predict the tax rates needed to achieve the identical emission reductions, the predictions vary by a factor of five from the highest to the lowest. 36 The range might be from $15 to $75 per ton in 2030, for example, and the range widens as predictions go out further into the future. This high degree of uncertainty regarding the tax rate needed to hit any emission control target strengthens the economic argument for a cap-and-trade system. Moreover, the effect of a carbon tax will vary substantially, depending on a) the current rate of economic growth or decline, b) the pace at which the economy 22

23 is undergoing structural change, c) the effect of ancillary measures already in place, such as energy efficiency standards, d) fossil fuel prices, and e) the current availability and costs of alternative energy, including nuclear. All these are variable and difficult to forecast. Another reason why the government would find it difficult to set the appropriate tax rate, even if everyone agreed on the emission target, is that the effect of a carbon tax works through the economy very gradually. Consider a hike in the gasoline tax, for example. The short-run response might be some change in people s driving patterns. Over time, as people buy new cars and perhaps choose more efficient models, the effect might be greater. Still later, when changing demand induces the auto companies to alter the models they make and market, the effect would be greater still. Studies have found that the long-run response to changes in energy prices is three to five times greater than the short-run response. 37 Policy makers would be uncertain even about the effects of the tax rates they adopt. It is difficult to imagine any tax-writing committee of the Congress, working amid all the political pressures, coping successfully with these wideranging uncertainties. In contrast, concerns about potential price spikes or unintended price levels can be handled better within a cap-and-trade system. If emitters are able to bank and borrow permits from one time period to another, then market operations can help to smooth out price spikes. In addition, the government can constrain price variations by adding to permit availability when prices reach upper limits borrowing the additional permits from future supplies or by withdrawing permits when prices are low and adding them to future supplies. Such flexibility can provide greater predictability about both the long-term emission trajectory and the short-term cost of emission reduction. 38 Finally, there is an inherent flaw in the economic analysis just described, which is based on a methodology for calculating net economic benefit in the face of an uncertain range of outcomes. This approach treats a 1% chance of losing 90% of the economy as being just as acceptable as a 90% chance of losing 1% of the economy. In other words, it assumes that the government should act with no special aversion toward risk, even catastrophic risk, in setting climate policy. This is not how government policy is or should be designed. 23

24 An even-handed weighting of possible outcomes by their likelihoods makes sense when the government is making many uncertain decisions, each of which has limited possible impact. Then, it would be reasonable to assume that as the uncertainties are resolved, some decisions will turn out well and others not so well. The favorable and unfavorable outcomes will even out: You win some and lose some. This rationale is completely inappropriate, however, when the government faces a singular momentous decision, with possibly irreversible consequences, of a magnitude that could threaten the welfare of future generations on a global scale. As a report from the National Research Council noted, in the past the Earth s climate has changed locally by 10 o C. (18 o F.) or more in the short space of only a decade or two, probably triggered by rapidly changing forces. 39 It is possible, though unlikely, that the rapid accumulation of greenhouse gases could become such a trigger. It would be inconceivably reckless for the government to be indifferent to these risks. The Clean Air Act, under which the Environmental Protection Agency is beginning to regulate greenhouse gas emissions, clearly incorporates risk aversion. Congress instructed EPA to develop primary air quality standards that, with an adequate margin of safety, protect the health of the entire population from air pollution, including the health of sensitive people, such as children, the elderly, and people suffering from respiratory ailments. Expressly refusing to require comparisons of benefits and costs in setting standards, Congress required that pollution standards be set to prevent any and all health impacts and required an adequate margin of safety as well. That expresses significant aversion to risk in the nation s most relevant environmental law. It isn t difficult to find other such examples. America s national security policies are extremely risk-averse. If they weren t, one wouldn t see 80-year-old grandmothers and 7-year-old children passing through airport body scanners to make sure they are not carrying weapons. If they weren t, our military and national security expenditures would probably not be as high as the expenditures of all other countries combined. The national security community has identified climate change as one of this country s important security issues, mainly because of the upheavals and crises it is likely to create around the world. If climate change is a national security issue, it should be treated with comparable risk aversion. 24

25 Once risk aversion is brought into the analysis, the economic advantage shifts even more decisively toward a cap-and-trade regime rather than a carbon tax. Risk aversion implies that potentially catastrophic outcomes, even if unlikely and occurring well into the future, should be given more weight in decisions than in the conventional expected net benefit approach. 40 Moreover, because the 100-year storm is now a 20-year storm, the far more devastating 500-year storm, which may never yet have been recorded, is now more likely to have become a 100-year event. 41 As the frequency of extreme weather events increases, it is reasonable and prudent to think that the likelihood of even more extreme events is also increasing and should be guarded against. Control costs are then like investments in police protection or homeland security aimed at preventing personal or national catastrophe 42 and climate policy should be designed to ensure that the worst possible outcomes do not happen. The more likely it is that climate change has potentially catastrophic consequences and the more likely it is that there are thresholds beyond which climate change could become self-reinforcing, the greater the advantage of a capand-trade that controls emissions more effectively. For all these economic, political, and policy reasons, a cap-and-trade system should be the foundation for America s climate policy. Its advantages are decisive, and its prospects for enactment and successful operation are superior. It is the better way to create a badly needed price on carbon, which will mobilize the entire economy to ward off the threats created by global warming. 1 US Global Change Research Program, National Climate Assessment, draft for public comment, 2012; 2 M. Raupach, Global Emissions Rising Faster than Expected, CSIRO, May 2007; PNAS.aspx; Global Emissions Rising Faster than Worst-Case Scenario;

26 4 Joseph A. Aldy & R. Stavins, The Promise and Problems of Pricing Carbon: Theory and Experience.Journal of Environment and Development 21(2) 2012: Business Leaders Demand a Global Carbon Price at Doha;, Business Green, Nov. 19, 2012; Robert N. Stavins, A U.S. Cap-and-Trade System to Address Global Climate Change, The Hamilton Project, Brookings Institution, October, 2007; Robert Repetto, America s Climate Problem: The Way Forward, Earthscan, 2011; Ch. 3, Choosing the Right Policy Architecture. 8 Parry, Ian W. H., and Roberton C. Williams III What Are the Costs of Meeting Distributional Objectives for Climate Policy? B. E. Journal of Economic Analysis & Policy 10(2). 9 Australia is trying a hybrid approach: In 2012 the government imposed a tax-like fixed price of $23AUD per tonne for unlimited emissions permits on selected fossil fuels consumed by major industrial emitters. The fixed price will transition to a trading system in , when the number of permits will be limited by a pollution cap. 10 Theda Skocpol, What It Will Take to Counter Extremism and Engage Americans in the Fight against Global Warming, paper presented at the Symposium on The Politics of America s Fight Against Global Warming; Harvard University, February, 14, University of Michigan, Ford School of Public Policy, National Surveys on Energy and Environment: Public Opinion on Climate Policy Options, December 2012; 12 Terry Dinan, Offsetting a Carbon Tax s Costs on Low-Income Households, Conference on the Economics of a Carbon Tax, Brookings Institution, Washington DC, December, James K. Boyce & Matthew Riddle, Cap and Dividend: How to Curb Global Warming While Protecting the Incomes of American Families, U. of Massachusetts, Political Economy Research Institute, Amherst, MA, NO USEPA, Carbon Pollution Standard for New Power Plants, last updated 5/25/ Lawrence H. Goulder, Markets for Pollution Allowances: What Are the (New) Lessons? Journal of Economic Perspectives, Volume 27, Number 1, Winter 2013, pp Goulder, L. and Bovenberg, A Efficiency Costs of Meeting Industry Distributional Constraints Under Environmental Permits and Taxes, Rand Journal of Economics, 36:4. 18 Robert N. Stavins, Experience with Market-Based Environmental Policy Instruments. In Handbook of Environmental Economics, vol. I, ed. Karl-Göran Mäler and Jeffrey Vincent, Amsterdam: Elsevier Science Justin Caron, Sebastian Rausch, Niven Winchester, Leakage from sub-national climate initiatives: The case of California, May 2012; 20 Lucas Brown et al., The EU Emissions Trading System: Results and Lessons Learned, Environmental Defense Fund, Washington DC, June 5, Jaffe, Judson, Matthew Ranson, and Robert N. Stavins Linking Tradable Permit Systems: A Key Element of Emerging International Climate Policy Architecture. Ecology Law Quarterly 36(4): Robert Stavins and J. Jaffe. A Wave of the Future: International Linkage of National Climate Change Policies. March 31, William A. Pizer, The optimal choice of policy in the presence of uncertainty. Resource and Energy Economics 21, , Richard G. Newell & Pizer, W.A., Regulating Stock Pollutants under Uncertainty, Journal of Environmental Economics and Management, 45: , William Nordhaus & J. Boyer, Warming the World: Economic Modeling of Global Warming, MIT Press, Cambridge, Mass., Summer

27 27 Pew Center on Global Climate Change, A Climate of Extreme Weather Events: U.S. Impacts and Vulnerability, June Stern Review on the Economics of Climate Change, 29 W. A. Nordhaus, The Economics of Hurricanes in the United States, Yale University; 30 Robert Mendelsohn, K. Emanuel & Shun Chonabayashi, The Impact of Climate Change on Global Tropical Storm Damages; 31 K. Emanuel & T. Jagger, 2010, On Estimating Hurricane Return Periods, J of Applied Meteorology and Climatology, 49: ; Kerry Emanuel, 2011, Global Warming Effects on US Hurricane Damages, Weather, Climate & Society, 3: Copenhagen Diagnosis Offers a Grim Update to the IPCC Climate Science, Grist; No. 24, Edward A.G. Schur et al., Vulnerability of Permafrost Carbon to Climate Change: Implications for the Global Carbon Cycle; BioScience 58(8): ; doi: 34 Nicholas Z. Muller et al, Environmental Accounting for Pollution in the United States, American Economic Review 101:5; August, Martin L. Weitzman, Prices vs. Quantities, The Review of Economic Studies, Vol. 41, No. 4 (Oct. 1974), pp Allen A. Fawcett et al., Overview of EMF 22 U.S. transition scenarios, Energy Economics 31 (2009) S198-S Kenneth A. Small & Kurt Van Dender, Long run trends in transport demand, fuel price elasticities and implications of the oil outlook for transport policy, OECD Transport Research Center, Oil Dependence: Is Transport Running Out of Affordable Fuel?, Paris, 2008, pp "Gilbert A. Metcalf & Jongsang Park, A Comment on the Role of Prices for Excludable Public Goods," International Tax and Public Finance, 14(2007): Sheila M. Olmstead and R. N. Stavins, Three key elements of a post-2012 international climate policy architecture, Rev Environ Econ Policy (Winter 2012) 6 (1): National Research Council, Abrupt Climate Change: Inevitable Surprises, 40 Martin L. Weitzman, Rare Disasters, Tail-Hedged Investments, and Risk-Adjusted Discount Rates, Harvard University, October 19, 2012; forthcoming; 41 Martin Weitzman, A Precautionary Tale of Uncertain Tail Fattening; March 10, 2012; forthcoming 42 Martin L. Weitzman, On Modeling and Interpreting the Economics of Catastrophic Climate Change, Review of Economics & Statistics, 91(1); 1-19; February 2009; Martin Weitzman, Climate change: Insurance for a warming planet, Nature, 467: , October 14,