Chapter
1
Efficiency Implications of Different Policy Designs
Incentive-based policies can reduce emissions of carbon dioxide (CO2) and other greenhouse gases, thereby reducing the risks associated with global climate change, at a lower cost than less flexible alternatives. Policymakers have many options, however, for giving businesses and households an economic incentive to reduce emissions. One option is to regulate the price of emissions—for example, by imposing a tax on them. A tax would limit the cost of cutting emissions but would leave the amount of CO2 emitted in a given year uncertain. As an alternative, the government could adopt a market-based system to regulate the quantity of emissions—for instance, by combining a cap on total annual emissions with a system of tradable emission permits, or allowances. If monitoring and enforcement were effective, a cap-and-trade program would limit the amount of CO2 emitted in a given year but would leave the cost of reducing emissions uncertain. The design of a cap could be modified in various ways to make it more flexible and to adopt some of the characteristics of a tax while maintaining the structure of a cap-and-trade program.
Any of those incentive-based approaches could achieve a given cut in emissions at a lower cost than command-and-control approaches, in which the government mandated how much individual factories could emit or what technologies they should use. However, incentive-based approaches would differ in their economic efficiency (the subject of this chapter) and in the ease with which they could be implemented in the United States and coordinated with other countries’ emission-reduction policies (discussed in Chapters 2 and 3). The most economically efficient policy is the one that can best keep the marginal cost of reducing emissions—that is, the cost of cutting emissions by another ton—in balance with the marginal benefit (in terms of avoided damage from climate change). A related concept is cost-effectiveness. A cost-effective policy would minimize the cost of meeting a given target for emissions, regardless of whether or not that target was chosen to balance benefits and costs. The efficiency criterion addresses how well policies might function to minimize the cost of reducing emissions over a period of several decades; however, policymakers may choose to place more emphasis on providing certainty about the amount of emissions at specific points in time.
Neither the costs nor the benefits of reducing CO2 emissions can be known when a reduction policy is put in place. Thus, policymakers must rely on estimates of both of them. The costs of reducing emissions would occur when the reductions were made and could vary substantially depending on such factors as the amount of economic activity, market conditions, weather, and available technologies. The benefits of reducing emissions, in contrast, would be realized decades or even centuries after the reductions were made. The reason is that each ton of CO2 generates a rise in the average global temperature that peaks about 40 years after the CO2 is emitted and then dissipates slowly, with a half-life of about 60 years.1
Estimating the benefits of cutting emissions is complicated by that long-term effect. In addition, analysts who try to estimate the benefits of cutting emissions face many other challenges, including addressing numerous scientific and economic uncertainties; measuring costs, such as mass species extinction, that are difficult to quantify in economic terms; and deciding how much weight to give to changes in the welfare of future generations.2
Some experts think that the effects of climate change could be modest, especially if society is ingenious in adapting to the change. However, other experts are concerned that rising concentrations of greenhouse gases could produce far more severe consequences for the global and U.S. economies than have generally been projected. Curbing greenhouse-gas emissions would help limit not only the expected costs of future global climate change but also the chances of irreversible or potentially catastrophic damage.
In general, the possibility of significant damage provides an economic motivation for taking additional action to moderate the growth of emissions in the near future—and, potentially, to cut emissions to very low levels in the longer run. Individuals take actions (such as reducing risky behavior or buying insurance) to lessen their harm from extreme events; similarly, societies or governments should and do take actions to avoid catastrophic collective harm. The difficulty for policymakers is determining the appropriate cost to be paid today to lessen what may be a small risk of a potentially catastrophic event in the future.3
Although estimating the benefits of emission reductions is difficult, policymakers cannot avoid making a judgment about them: Policy choices about climate change will necessarily imply a value for those benefits. That value would be explicit under a tax, because the tax rate provides an indication of what the government thinks an incremental reduction in emissions is worth. By contrast, that value would be implicit under a cap. A higher (less stringent) cap would imply a lower estimate of the marginal benefit of cutting emissions—as reflected in lower prices for emission allowances—than a lower (more stringent) cap would.
When comparing emission-reduction policies, the Congressional Budget Office (CBO) generally assumes that lawmakers would design them in the most efficient way—that is, to achieve the highest possible net benefits, given the limitations of each particular policy tool. Thus, for example, this analysis compares the most efficient tax on CO2with the most efficient cap. In other words, the tax or cap is assumed to be set at a level that encourages the affected parties to reduce emissions as long as the expected cost of doing so is less than or equal to the expected benefit. Those costs and benefits will inevitably be different than anticipated. Policy designs will yield different net benefits depending on their ability to balance the costs and benefits of emission reductions when those turn out to be higher or lower than policymakers had anticipated. Designs that are relatively more efficient would also be relatively cost-effective: The characteristics of a policy design that enable it to equate the cost of additional emission reductions with their anticipated benefits also enable it to minimize the cost of achieving any given emission-reduction target.
To be most efficient, a tax would need to rise and a cap would need to decline gradually over time. The future benefits of avoiding climate-change damage by reducing CO2 emissions by a ton would have an increasingly greater present value (that is, the value today after taking into account the time value of money) as the potential for large damage drew closer in time. An increasingly stringent tax or cap would reflect that increase in present value over time. Further, a gradually rising tax or tightening cap would allow for a smoother transition to a less carbon-intensive economy. Businesses and households would have more time to replace their equipment and energy-use practices with more efficient alternatives.
A Carbon Dioxide Tax Versus an Inflexible Carbon Dioxide Cap
According to many analysts, a tax would be a more economically efficient policy for reducing CO2 emissions than an inflexible cap (with "inflexible" meaning a cap whose level was not affected by the price of emission allowances). That conclusion stems from the cumulative, long-term nature of climate change: The benefit of emitting one less ton of CO2 in a given year is roughly constant, whereas the cost of emitting one less ton of CO2 each year rises with each ton reduced. The reason for rising marginal costs is that companies that have to comply with an emission-reduction policy will make the cheapest cuts first and progressively more expensive cuts thereafter.
The contrast between constant marginal benefits and rising marginal costs means that the gap between uncertain costs and benefits is particularly sensitive to the amount of annual emission reductions. A cap that is too tight will disproportionately increase costs over benefits, and a cap that is not tight enough will disproportionately lower costs relative to benefits. A tax, by contrast, will tend to hold the costs of emission reductions in line with the constant (although uncertain) expected benefits, encouraging greater emission reductions when costs are low and allowing more emissions when costs are high.
An Illustrative Example of How a Tax Would Be More Efficient Than a Cap
To understand how a tax could offer efficiency advantages over a cap, assume that the future benefits of limiting emissions have a present value of $15 per metric ton of CO2 (or $55 per metric ton of carbon), that those benefits would be constant over the range of potential emission reductions during the initial years of the policy, and that the tax or cap would take effect in the United States in 2017.4 If the costs of cutting emissions turned out to be as expected, the tax and the cap would be equivalent. But if those costs differed from the government’s expectations, a tax would be the more efficient policy.
For example, given the assumptions above, if lawmakers imposed a tax of $15 per metric ton on U.S. emissions of CO2, and if the costs of limiting emissions equaled expectations, the $15 tax would reduce U.S. emissions in 2017 by 437 million metric tons (see the top panel of Figure 1-1). That amount represents a cut of roughly 6.5 percent from the 6.7 billion metric tons that would otherwise be emitted that year, CBO estimates.5 Alternatively, lawmakers could set a cap that was 437 million metric tons below the baseline level of U.S. emissions, and if the costs of reducing emissions were what they had expected, the incremental cost of meeting the cap would be $15 per metric ton. Under the illustrative assumption that each ton of emission reductions would produce $15 worth of avoided damage and using information about the cost of emission reductions derived from various models, CBO estimates that either policy would yield net benefits of $3.5 billion in its first year (see the lower panel of Figure 1-1).6
Illustrative Comparison of Various Policies to Reduce CO2 Emissions Under Different Cost Conditions
Source: Congressional Budget Office.
Notes: For illustrative purposes only, this example assumes that the benefit of reducing carbon dioxide (CO2) emissions is $15 per metric ton. It examines the net benefits that would result in the first year of each policy, assuming that the policy covered only the United States and took effect in 2017 after having been announced 10 years earlier. The cost of firms' emission reductions (and the response to various taxes) is derived from Mark Lasky, The Economic Costs of Reducing Emissions of Greenhouse Gases: A Survey of Economic Models, Office Technical Paper No. 2003-03 (May 2003).
A safety valve is a ceiling on the price of emission allowances.
a. Assumes that the actual marginal cost of reducing emissions by 437 million metric tons is $15 per metric ton, the cost that policymakers anticipated when they set the cap.
b. Assumes that the actual marginal cost of reducing emissions by 437 million tons is $7.50 per metric ton but that the tax induces more reductions (up to 824 million tons) at a marginal cost of $15 per metric ton.
c. Assumes that the actual marginal cost of reducing emissions by 437 million tons is $30 per metric ton but that the tax induces fewer reductions (234 million tons instead of 437 million), up to a marginal cost of $15 per metric ton.
If the costs of cutting emissions were different than expected, however—for example, if new technologies turned out to be less expensive than anticipated—the two policies would produce different outcomes.
If the costs of cutting emissions were half the anticipated level—for example, because of unforeseen technological breakthroughs—both policies would produce higher net benefits than expected.7 The increase in net benefits, though, would be greater under a tax than under a cap: The tax would give firms an incentive to keep cutting emissions as long as doing so cost less than paying the tax. CBO estimates that in this scenario a tax would cause emissions to be cut by 824 million metric tons (roughly 12 percent below the baseline level), rather than by the 437 million metric tons required by the cap. Each of those additional cuts would boost net benefits because they would cost less than, or as much as, their $15 per ton expected benefit.
Alternatively, if the cost of reducing emissions turned out to be twice as high as expected, the net benefits would be lower under each policy—but would fall much more under the cap than under the tax. In particular, under the inflexible cap, firms would be required to reduce emissions by 437 million metric tons, even though reaching that target would entail making reductions that cost up to $30 per metric ton but provided benefits of only $15 per metric ton. As a result of the higher costs, the total net benefits of the cap would fall to $0.7 billion—just one-fifth of the expected amount. A tax would also have lower net benefits if the costs of cutting emissions proved greater than expected. But net benefits would decline by less for a tax than for a cap. Because companies would have the flexibility to reduce emissions by less than 437 million metric tons, the net benefits of a tax would be more than twice those of a cap.
Like costs, benefits could also be higher or lower than anticipated; however, neither policy would adjust to that change. If actual marginal benefits turned out to be much higher than expected, either a tax or a cap would produce too few cuts in emissions, and both policies would fall short of the most efficient level of emission reductions by the same amount.8
Empirical Estimates of the Efficiency Advantage of a Tax
If the government wanted to maximize expected net benefits, it would need to set the level of a cap or a tax in a given year on the basis of its best estimate of both the costs and benefits of reducing emissions in that year. However, actual costs in any year are likely to differ from the best estimate, sometimes exceeding it and sometimes falling below it. Because a tax would motivate only emission reductions that cost less than the tax rate, it would automatically adjust the quantity of emission reductions to keep their costs in line with their anticipated benefits, whereas a cap would not.
When analysts take into account the degree to which costs are likely to vary around a single best estimate, they conclude that a tax could offer much higher net benefits than a cap. One study suggests that the net benefits of a worldwide tax on CO2 emissions in 2010 would be more than eight times larger than those of an equivalent inflexible cap. If the policies are assumed to be set in place for 100 years, the efficiency advantage of a tax declines to a factor of five.9 Another study concluded that a tax could offer up to 16 times greater expected net benefits than a cap under some assumptions.10 A third study examined outcomes when cost shocks were assumed to be correlated across time—that is, an unusually high cost of meeting the cap in any given year increases the likelihood of a higher than average cost in the following year. Using their base-case parameter estimates for factors that might affect costs (such as baseline emissions and changes in technology) and assuming a 10-year policy, those researchers estimated that the net benefits of a tax would be roughly five times higher than those of a cap.11 Taken together, those studies suggest that the net benefits of a tax could be roughly five times those of an inflexible cap (see Figure 1-2)—assuming that both policies were designed to balance expected costs and benefits.
Relative Economic Efficiency of Various Policies to Reduce CO2 Emissions, When Cost Uncertainty Is Taken Into Account
Source: Congressional Budget Office based on estimates of the relative magnitude of the net benefits of various policies found in William A. Pizer, "Combining Price and Quantity Controls to Mitigate Global Climate Change," Journal of Public Economics, vol. 85 (2002), pp. 409–434, and in Richard G. Newell and William A. Pizer, "Regulating Stock Externalities Under Uncertainty," Journal of Environmental Economics and Management, vol. 45 (2002), pp. 416–432.
Notes: The net benefits of each policy are shown in relationship to each other with the net benefits of an inflexible cap set equal to one. The inflexible cap and the tax are assumed to be set at the most efficient level—that is, at the point at which the expected marginal cost of complying with the policy would be equal to the anticipated marginal benefit of reducing emissions.
The net benefits of a cap with a safety valve (a ceiling on the price of emission allowances) are based on the assumption that the cap would be set at the level of the most efficient inflexible cap and the safety-valve price would be set at the level of the most efficient tax. Banking would enable firms to save unused allowances from one period to use in a future period.
The net benefits of a cap-and-trade program with a circuit breaker (not shown in the figure) would be greater than those of an inflexible cap and less than those of a cap with a safety valve; however, CBO lacked sufficient information to determine how much greater or less they would be.
A cap-and-trade program that included a safety valve and either a price floor or banking provisions could be significantly more efficient than an inflexible cap, although somewhat less efficient than a tax.
Viewed another way, any long-term emission-reduction target could be met by a tax at a fraction of the cost of an inflexible cap-and-trade program. That cost savings stems from the fact that a tax could better accommodate cost fluctuations while simultaneously achieving a long-term emission target. It would provide firms with an incentive to undertake more emission reductions when the cost of doing so was relatively low and allow them to reduce emissions less when the cost of doing so was particularly high.
The Impact of Price Volatility
The flexibility in reducing emissions that a tax affords is important because the cost of cutting emissions by a given amount could vary from year to year depending on such factors as the weather, the level of economic activity, and the availability of low-carbon technologies. A tax would provide a steady, predictable price for emissions. An inflexible cap, however, could result in volatile allowance prices, making a cap-and-trade program more disruptive to the economy than a tax would be.
Experience with cap-and-trade programs has shown that price volatility can be a major concern when a program’s design does not include provisions to adjust for unexpectedly high costs and to prevent price spikes. For example, one researcher found that the price of sulfur dioxide allowances under the U.S. Acid Rain Program was significantly more volatile than stock prices between 1995 and 2006 (see Figure 1-3).12
Volatility of SO2 Allowance Prices and Selected Other Prices, 1995 to 2006
(Average annual percentage rate of volatility)
Source: Congressional Budget Office based on William D. Nordhaus, “To Tax or Not to Tax: Alternative Approaches to Slowing Global Warming,” Review of Environmental Economics and Policy, vol. 1, no. 1 (Winter 2007), pp. 26–44.
Note: Volatility is calculated as the annualized absolute logarithmic month-to-month change in the consumer price index (CPI), the stock price index for the Standard & Poor’s 500 (S&P 500), and the price of sulphur dioxide (SO2) allowances under the U.S. Acid Rain Program.
Price volatility was most apparent in the summer of 2000 in Southern California’s Regional Clean Air Incentives Market (RECLAIM), a program that capped emissions of nitrous oxide (NOx) from the power sector. A heat wave caused demand for electricity to soar that summer, while the availability of imported power from other states declined. The increase in demand had to be met by running many of California’s old gas-fired generating facilities, which had not yet installed NOx emission controls. As a result, the demand for NOx RECLAIM Trading Credits for 2000 rose significantly, boosting their average annual price tenfold (from $4,284 per ton in 1999 to almost $45,000 per ton in 2000) and contributing to high wholesale electricity prices in California during that period.13 In addition to the California experience, allowance prices in the European Union’s (EU’s) Emission Trading Scheme (ETS)—a trading program that covers CO2 emissions from roughly 12,000 sources across 27 countries—fell drastically when it became evident that policymakers had overallocated emission allowances.
Price volatility could be particularly problematic with CO2allowances because fossil fuels play such an important role in the U.S. economy. They accounted for 85 percent of the energy consumed in the United States in 2006. CO2 allowance prices could affect energy prices, inflation rates, and the value of imports and exports. Volatile allowance prices could have disruptive effects on markets for energy and energy-intensive goods and services and make investment planning difficult.14 The smoother price path offered by a CO2 tax would better enable firms to plan for investments in capital equipment that would reduce CO2 emissions (for example, by increasing efficiency or using low-carbon fuels) and could provide a more certain price signal for firms considering investing in the development of new emission-reduction technologies.
Conditions Under Which a Cap Could Be More Efficient Than a Tax
To compare the net benefits of a tax and a cap, researchers must estimate the marginal benefit of reducing a ton of CO2 emissions. The efficiency advantage of a tax over a cap, however, does not depend on any particular measure of that benefit or even on the ability to place a monetary value on it. Rather, the advantage of a tax stems from the cumulative nature of climate change and from the fact that a tax is better able to reduce emissions over time without imposing potentially disruptive and unnecessarily expensive annual limits on emissions.
The relative advantages of a tax and a cap could change over time, however. One area of growing concern is that the buildup of greenhouse gases in the atmosphere could cause the global temperature to reach a critical level after which further growth in emissions could trigger a rapid increase in damage.15 The existence of such a threshold could alter the assumption that the marginal benefit of reducing emissions would be relatively constant and could make a cap more efficient than a tax.
Although concerns about thresholds exist, analysts who have tried to define more precisely the conditions that would cause a cap to be more efficient than a tax have concluded that those conditions are quite narrow and unlikely to apply in the near term. Specifically, scientists would need to have fairly precise knowledge about the location of an emissions threshold, and the threshold would have to be sufficiently close that the government would want to make very large cuts in emissions each year to avoid crossing it.16 If, instead, policymakers wanted to stabilize the concentration of greenhouse gases in the atmosphere after a period of several decades (at a level that would be expected to prevent the global temperature from rising to a trigger level), there could be considerable leeway about when the reductions took place. A tax would provide flexibility in the timing of emission reductions by encouraging companies to cut emissions more in years when the cost of doing so was low and cutting less when the cost was high. A rigid cap would not provide that flexibility over time.
A fundamental change in the cost of reducing emissions could also reverse the efficiency rankings of a tax and a cap. A cap could become more efficient than a tax if a new technology provided the opportunity to make extremely large cuts in emissions at a low and fairly constant cost, rather than at a rising marginal cost.
Other Efficiency Implications of a Tax or a Cap
Besides the efficiency advantages described earlier, a tax on CO2 emissions could offer another advantage. By generating a significant amount of revenue, it would give the government a chance to use the revenue in a way that would lower the cost to the economy of curbing emissions. For example, studies have found that the economywide cost of reducing emissions could be more than twice as high if the reduction was achieved through a cap-and-trade program (with allowances allocated for free) than if it was achieved through a CO2 tax (with the revenue used to reduce existing taxes that discourage economic activity, such as taxes on capital, labor, or income).17 A cap-and-trade program could offer a similar opportunity, but only if the government chose to sell the allowances rather than give them away.
If the government elected to tax CO2 emissions or sell allowances for them, it could opt to use some of the revenue to achieve other aims as well. One goal could be to offset the adverse financial impact of a CO2tax or cap on low-income households, who would bear a disproportionate burden (relative to their income) from the higher energy prices that the policy would trigger. In addition, lawmakers could compensate workers in carbon-intensive sectors (such as the coal industry) who might lose their jobs because of the policy.18
A cap on CO2 emissions could achieve some of the efficiency advantages of a tax while maintaining the basic structure of a cap-and-trade program by incorporating various design features to make the cap more flexible. Such policies would allow the cap to be exceeded or altered depending on economic circumstances that affect the cost of reducing emissions.
A Ceiling or Floor on Allowance Prices
Combining an emissions cap with a ceiling on the price of allowances—or safety valve—could offer some of the advantages of a tax.19 Under that approach, if the cost of cutting emissions (as indicated by the price of allowances) rose to the safety-valve level, the government would issue an unlimited number of allowances at that price, thus allowing emissions to exceed the cap. However, unlike a tax, a cap with a safety valve would not give firms and households an incentive to make additional emission cuts if the cost of doing so was lower than anticipated.
In the illustrative example described above, if a cap limiting CO2 emissions to 6.3 billion metric tons in 2017 (437 million tons below the baseline level for that year) included a safety-valve price of $15 per metric ton of carbon, it would produce the same outcome as a tax of $15 per ton if the cost of meeting the cap was higher than expected (see Figure 1-1). In that case, both the tax and the cap/safety valve policy would allow higher emissions than an inflexible cap and would limit the cost of reductions to $15 per ton. Conversely, if the cost of meeting the cap was lower than expected, the cap/safety valve would produce the same outcome as an inflexible cap. The lower-than-expected costs would cause net benefits to be higher than anticipated, but not as high as they would be with a tax.
Under some circumstances, a cap with a safety valve could offer roughly half of the efficiency gains of a tax over a rigid cap. That situation would be most likely to occur if the safety-valve price was set at the amount of the most efficient tax (assumed to be $15 per ton of CO2in this example) and the cap was set at the level of the most efficient inflexible cap (estimated to be 6.3 billion metric tons, on the basis of an assumed marginal benefit of $15 per ton of CO2 and the quantity of emission reductions that would result from that price).20 In that case, the net benefits of the cap/safety valve policy would fall roughly halfway between those of a cap and a tax (see Figure 1-2).
If the safety-valve price was kept at the level of the most efficient tax but the cap was tightened, then the cap/safety valve policy would function more like a tax and would become even more efficient (see Figure 1-4). Specifically, the amount of emission reductions would increasingly depend on the cost limit specified by the safety-valve price rather than on the quantity limit specified by the cap. At the extreme, a cap of zero emissions with a safety-valve price of $15 per ton of CO2 would provide the same incentives as a tax of $15 per ton. The cap of zero emissions would not prohibit emissions, but companies would have to purchase an allowance from the government at the safety-valve price for each ton of CO2 they emitted. (Adding banking or a price floor to a cap-and-trade program with a safety valve offers another way to capture more of the efficiency advantages that could result from an appropriately designed tax. That option is discussed later in this chapter.)
Illustrative Range of Net Benefits for a Cap With a Safety Valve Compared With a Tax or an Inflexible Cap on CO2 Emissions
Source: Congressional Budget Office based on information from Richard G. Newell and William A. Pizer, Indexed Regulation, Discussion Paper 06-32 (Washington, D.C.: Resources for the Future, June 2006).
Note: CO2 =carbon dioxide; bmt = billion metric tons.
In the other direction, if the cap in the cap/safety valve approach remained at the level of the most efficient inflexible cap but the safety-valve price rose above the level of the most efficient tax, then the cap/safety valve policy would function more like an inflexible cap and would become less efficient. In that case, the amount of emission reductions would be more likely to be determined by the cap than by the safety-valve price. At the extreme, if the safety-valve price was raised high enough that the safety valve would not be triggered, the policy would be equivalent to not having a safety valve, and the net benefits would be the same as those of an inflexible cap.
A recent criticism of a safety valve is that it could unintentionally reduce firms’ incentives to replace carbon-intensive capital equipment and to develop new technologies for lowering CO2 emissions.21 Either taxing or capping emissions would set a price on them. Researchers generally conclude that the most efficient price for CO2 emissions would be relatively low in the near term but would rise substantially over time. Expectations of higher future prices would give companies an incentive to gradually replace their stock of physical capital associated with carbon-intensive energy use (such as coal-fired generators or inefficient heating systems) and to invest in researching and developing new technologies that would reduce emissions (such as improvements in solar power, wind power, or energy efficiency).22 The higher that future allowance prices were expected to rise, the greater that incentive would be. Including a safety valve in a cap-and-trade program, however, would lower expectations about future prices by ensuring that the price of allowances would not rise above the safety-valve level, although it could fall below. In other words, the fact that the range of potential future prices would be truncated at the high end by the safety valve but not at the low end would reduce the expected price.23 As a result, the safety valve could have the unintended effect of inducing less capital-stock turnover and less investment in research and development (R&D) than would occur under an inflexible cap or a tax. 24
That problem could be addressed by adding a floor on allowance prices.25 Enforcing a minimum price for allowances could be fairly straightforward if the government chose to sell a significant share of the allowances rather than give them to affected businesses for free. If allowances were auctioned, policymakers could specify a reserve auction price and restrict the supply of allowances to maintain that price. In combination, a reserve price and a safety valve could define a band of acceptable clearing prices for the allowance market in a cap-and-trade system and could stabilize price expectations. Thus, that combined policy could capture much of the efficiency advantage offered by a tax on emissions (see Figure 1-2).26
Enforcing a minimum price would be considerably more difficult if nearly all of the allowances were given away for free. In that case, the government could attempt to enforce a minimum price only by reducing the supply of allowances—for example, it could buy back allowances from firms or decrease the value of allowances so that each allowance would permit less than one ton of emissions. Determining when such actions should be undertaken would require the government to make judgments about current and future allowance prices (for example, distinguishing short-term dips from long-term trends). To the extent that those judgments were incorrect, the adjustments to the supply of allowances might under-correct or overcorrect the allowance price. Further, some analysts are concerned that identifying a trigger price at which policymakers would alter the cap could actually promote price volatility. For example, firms might resist buying allowances once the price began to approach or exceed the trigger point, waiting for policymakers to loosen the cap. But once the demand for allowances dropped, the price would begin to fall and the possibility of intervention would diminish. As a result, purchases (and prices) would once again begin to increase.27
Alternatively, increasing the stringency of the cap, while holding the safety valve constant, would reduce the potential problem of underinvestment in R&D and insufficient capital-stock turnover. As noted above, the safety valve would become increasingly likely to determine the quantity of emission reductions and the price of allowances. It would also keep the amount of reductions from falling below the efficient level when the cost of cutting emissions was low. Another option that could help address the underinvestment problem would be to allow emitters to bank allowances for future use.
Banking and Borrowing Allowances
Banking and borrowing would give firms the opportunity to move allowances—and the emissions that correspond to them—between time periods. Each emission allowance would be valid for a specific year or alternative compliance period. (A 2017 allowance, for example, would allow the company that held it to emit one ton of emissions in that year.) With banking, a company could reduce its emissions below the amount it would be permitted to emit on the basis of its allowance holdings for a given year, thereby using fewer allowances in that year, and could bank the extra allowances to use in a future year.28 With borrowing, by contrast, a firm could exceed its permitted level of emissions in one year by borrowing from its allocation of allowances for a future year.
Emitters would want to bank allowances in years when they thought the price of allowances was low relative to that of future years (for example, because of a mild winter or a period of slow economic activity, or because they believed that tighter caps in the future would lead to higher allowance prices). Conversely, companies would want to borrow allowances in years when they thought the price of allowances was high relative to that of future years (for example, because they expected a new, low-cost technology for reducing emissions to become available later).
Banking Allowances. Banking provisions could improve the efficiency of a cap-and-trade program, regardless of whether the program included a safety valve. While a safety valve could prevent the price of allowances from climbing too high, banking could help prevent the price from falling lower than policymakers would like. Firms would have an incentive to bank allowances in a given year if the cost of making additional emission reductions in the current year—that is, reductions in excess of the aggregate amount that firms need for compliance in that year—was less than the expected present value of the cost of reducing emissions or buying allowances in the future. By providing firms with an incentive to save their own allowances—or purchase additional allowances for saving—banking would boost the demand for, and the price of, allowances in years in which that price was relatively low.29
The combination of banking and a safety valve could help keep the marginal cost of emission reductions in line with their anticipated benefits under some conditions. For example, such a policy could be effective in preventing relatively short-term lows in allowance prices, but it would be less effective in boosting the price of allowances if the cost of reducing emissions turned out to be significantly lower than anticipated in both the near term and the long term—because of the introduction of a new technology, for instance.30 In that case, the market price for allowances could stay well below the safety-valve price (that is, below the expected marginal benefits), and the policy would motivate too few emission reductions. As discussed above, policymakers could help ensure that the safety valve would be triggered by setting the cap relatively tightly in comparison with the safety-valve price (see Figure 1-4).
Provided that the safety valve was expected to be eliminated at some point, combining a safety valve with banking provisions could create an incentive for firms to purchase very large amounts of allowances through the safety-valve mechanism and bank them for use once the safety valve was removed.31 That strategy could prevent a sharp increase in the price of allowances once the safety valve was removed, but it could also mean that the cap would not be met for several years after the removal. The potential for such an outcome would be greatest if the safety valve was holding the price of allowances well below the actual cost of meeting the cap. For example, suppose that firms were allowed to buy allowances through the safety valve in 2020 for $20 but that the safety valve was expected to be removed in 2021 and that, in its absence, the price of allowances required to actually meet the 2021 cap was anticipated to be $40. In that case, firms would have an incentive to purchase very large quantities of allowances through the safety valve in 2020 and use those allowances once the safety valve was removed.
The large excess supply of allowances purchased through the safety valve would prevent the steep jump in allowance prices that would have occurred if firms had not been allowed to bank allowances, but it would also mean that the annual cap in 2021—and for a period of time thereafter—would not be met, even though the safety valve was no longer in place. If policymakers wished to reduce the potential for a multiyear delay in attaining the cap after the safety valve was removed, they could require firms to use allowances purchased at the safety-valve price in the year in which they were purchased.32
In addition, policymakers could choose to sell safety-valve allowances through an auction—rather than at a given price—and specify a reserve price for the auction that would increase as greater quantities of allowances were sold in any given year. For example, policymakers could choose to auction blocks of allowances, with increasing reserve prices, just prior to each year’s compliance deadline. The reserve price could be $22 for the first block, for instance, $24 for the second block, and so on. Such a strategy could prevent the price of allowances from jumping up once the safety valve was removed while limiting firms’ incentives to bank a large supply of allowances for use in future years.33
Borrowing Allowances. Including either borrowing provisions or a safety valve in a cap-and-trade program could help prevent spikes in the price of allowances; however, a safety valve could offer greater efficiency advantages. Borrowing would help bring down the price of allowances in a given year only if the price in that year was high relative to prices anticipated in future years. For example, if the price of allowances was $30 in 2010 and was expected to be $15 in 2015, then a firm would have an incentive to borrow 2015 allowances for use in 2010. If, however, the price was expected to be $45 in 2015, no such incentive would exist. Thus, borrowing could help avoid a price spike but would not necessarily keep the cost of emission reductions from exceeding their expected benefits. A safety valve, in contrast, could prevent the cost of emission reductions from exceeding estimates of the benefit of those reductions.
Allowing firms to make one-for-one trades between current and future allowances (and, correspondingly, between current and future emissions) would provide them with too much incentive to defer emission reductions to the future. Because firms discount future costs relative to current costs, they would have an incentive to engage in borrowing (and, thus, defer the cost of reducing emissions) simply to delay the cost of reducing emissions. The potential for excessive borrowing could be avoided if the government discounted borrowed allowances at the rate that companies use to discount future costs.34 That rate will generally vary from firm to firm; however, policymakers would need to choose a single discount rate. Some researchers suggest that the government could use a discount rate equal to the industry average interest rate used to finance medium-term capital expenditures.35 In addition, policymakers could choose to limit the amount of borrowed allowances that companies might use for compliance in any given period or the length of time over which borrowing might occur.
Policymakers could attempt to enforce a ceiling on the price of allowances (for example, keeping it roughly in line with the expected benefits of reducing emissions) by altering the terms under which firms could borrow allowances.36 Reducing restrictions on borrowing or lowering the rate at which borrowed allowances were discounted could increase the supply of borrowed allowances and thus reduce allowance prices in the near term. As described above, such a strategy could only be effective if firms anticipated that the price of emission reductions in the future would be low (in present-value terms) relative to the current price of allowances. (If that was not the case, firms would not have an incentive to borrow, even under the revised terms.) As a result, altering the terms under which firms might borrow allowances would be more effective in dealing with relatively short-term price spikes than with a situation in which policymakers had underestimated the cost of compliance—in both the near term and in the future—when they set the level of the cap.
Using such a strategy to enforce a limit on the price of allowances would require policymakers to have relatively accurate information about both the current and future prices of allowances. To the extent that those estimates were wrong, the changes that policymakers made to borrowing terms could over- or undercorrect the price. For example, if policymakers reduced restrictions on borrowing in order to lower the current price of allowances, but market conditions changed, the increased supply of allowances could cause their price to drop more than policymakers had intended. Alternatively, the increased availability of allowances might fail to reduce the current price as much as policymakers had anticipated.
Some analysts have suggested that an emissions cap that declined at a preset rate and that included a "circuit breaker" would offer economic advantages relative to an inflexible cap and perhaps relative to a cap with a safety valve as well. The circuit breaker would freeze the cap if the price of an allowance exceeded a specified level.37
Provided that the circuit breaker price was set at an efficient level (that is, the level that reflected the best available information on costs and benefits), a cap-and-trade program with a circuit breaker could be more efficient than a rigid cap. Specifically, it would offer some economic relief if the cost of meeting the declining cap was higher than the anticipated marginal benefits. Unlike a safety valve, however, a circuit breaker would not set an upper limit on the cost of reducing emissions. Once the circuit breaker was triggered and the cap stopped declining, the allowance price could continue to increase (albeit by not as much as if the circuit breaker was absent). In fact, continued price increases would be likely because meeting a constant cap would become increasingly costly as the economy grew. Thus, assuming that the circuit breaker price was set equal to the expected marginal benefits of reducing emissions, the allowance price (and the cost of achieving additional emission reductions) would be likely to rise above those expected benefits.
See William A. Pizer, "Combining Price and Quantity Controls to Mitigate Global Climate Change," Journal of Public Economics, vol. 85 (2002), p. 416.
For a more detailed discussion, see Congressional Budget Office, Uncertainty in Analyzing Climate Change: Policy Implications (January 2005).
For more discussion of policy choices in the face of catastrophic costs, see Cass R. Sunstein, Worst-Case Scenarios (Cambridge, Mass.: Harvard University Press, 2007).
The stringency of emission-reduction policies is sometimes discussed in terms of carbon and sometimes in terms of CO2. Estimated costs or benefits that appear in dollars per ton of CO2 can easily be translated into dollars per ton of carbon by multiplying by the ratio of the molecular weight of CO2 to the molecular weight of carbon (44/12, or 3.67). Thus, a tax of $15 per ton of CO2 translates into a tax of $55 per ton of carbon. Conversely, costs and benefits that are stated in terms of dollars per ton of carbon can be converted into dollars per ton of CO2 by dividing by 3.67.
For a description of how CBO calculated the emission reductions that would result from a given tax, or the price of allowances that would result from a given cap, see Mark Lasky, The Economic Cost of Reducing Emissions of Greenhouse Gases: A Survey of Economic Models, CBO Technical Paper 2003-03 (May 2003).
The cost of reducing emissions in any given year is incurred in that year, while the benefits accrue over a period of decades or centuries. Thus, comparing the costs and benefits of emission reductions involves discounting the value of future benefits to the current year. This illustrative example assumes that the benefits of reducing a ton of emissions have a present value of $15. As a result, reducing emissions by 437 million metric tons would produce benefits of $6.55 billion. The cost of achieving those reductions would be $3.07 billion, according to Lasky, The Economic Cost of Reducing Emissions of Greenhouse Gases.
The cost changes considered in this example correspond to two separate doublings of the price sensitivity parameter. Thus, the cost of cutting emissions by 437 million metric tons doubles from $7.50 to $15 per metric ton and then from $15 to $30 per metric ton.
For a more detailed discussion of the uncertainty about the costs and benefits of emission reductions, see Congressional Budget Office, Uncertainty in Analyzing Climate Change: Policy Implications (January 2005), pp. 30–31.
See Pizer, "Combining Price and Quantity Controls to Mitigate Global Climate Change." That paper considered a worldwide tax or cap on carbon emissions. In analyzing the sensitivity of his results to how long the policies are assumed to remain in place, the author assumed that the damage from climate change would rise rapidly once a certain temperature increase had occurred (in other words, that the damage function was sharply kinked). In that case, a cap would yield larger net benefits than a tax. However, the difference ($600 billion) would be small compared with the net benefits offered by either policy (roughly $34 trillion). Thus, under a sharply kinked damage function, the paramount concern would be to make drastic cuts in emissions, and the choice of policy tool would be relatively unimportant.
Michael Hoel and Larry Karp, "Taxes and Quotas for a Stock Pollutant with Multiplicative Uncertainty," Journal of Public Economics, vol. 82 (2001), pp. 91–114. Only under the assumptions of very great damage from climate change and a large initial stock of allowances do those authors conclude that a cap would be more efficient.
See Richard G. Newell and William A. Pizer, "Regulating Stock Externalities Under Uncertainty," Journal of Environmental Economics and Management, vol. 45 (2002), pp. 416–432.
William D. Nordhaus, "To Tax or Not to Tax: Alternative Approaches to Slowing Global Warming," Review of Environmental Economics and Policy, vol. 1, no. 1 (Winter 2007), pp. 26–44.
See A. Denny Ellerman, Paul L. Jaskow, and David Harrison Jr., Emissions Trading in the U.S.: Experience, Lessons, and Considerations for Greenhouse Gases (Arlington, Va.: Pew Center on Global Climate Change, May 2003), pp. 24–25, available at www.pewclimate.org/global-warming-in-depth/all_reports/emissions_trading. Some observers argue that the lack of banking provisions contributed to the price spikes. Such spikes could have been prevented by the inclusion of a safety valve as well. (Those design features are discussed later in this chapter.)
Nordhaus, "To Tax or Not to Tax," pp. 37–39.
See National Research Council, Abrupt Climate Change: Inevitable Surprises (Washington, D.C.: National Academy Press, 2002), pp. 13–14; R.B. Alley and others, "Abrupt Climate Change," Science, vol. 229 (March 28, 2003), pp. 2005–2010; and Congressional Budget Office, Uncertainty in Analyzing Climate Change, Box 2-1, pp. 10–11.
See William A. Pizer, Climate Change Catastrophes, Discussion Paper 03-31 (Washington, D.C.: Resources for the Future, May 2003).
See Congressional Budget Office, Trade-Offs in Allocating Allowiances for CO2 Emissions (April 25, 2007).
That feature is included in a cap-and-trade proposal (S. 1766) introduced by Senator Bingaman on July 11, 2007.
As determined in Lasky, The Economic Cost of Reducing Emissions of Greenhouse Gases.
See Dallas Burtraw and Karen Palmer, "Dynamic Adjustment to Incentive Based Policy to Improve Efficiency and Performance" (draft, Resources for the Future, Washington, D.C., November 30, 2006).
The amount of investment in research and development under either a tax or a cap-and-trade program could be less than the amount that would be best for society because such investment may generate "spillover benefits" to society that do not translate into profits for the firm doing the investing. For a discussion of that issue, see Congressional Budget Office, Evaluating the Role of Prices and R&D in Reducing Carbon Dioxide Emissions (September 2006).
For example, suppose policymakers set a cap on emissions in 2020, and observers generally agreed that there was a 25 percent chance that the allowance price necessary to meet the cap would be $25, a 50 percent chance that it would be $50, and a 25 percent chance that it would be $75. With no safety valve, the expected allowance price would be $50 [that is, (0.25 x $25) + (0.50 x $50) + (0.25 x $75)]. If, however, policymakers set a safety valve at $50, the expected allowance price would fall to $43.75 [(0.25 x $25) + (0.75 x $50)].
That effect is not reflected in Figure 1-4.
Burtraw and Palmer, "Dynamic Adjustment to Incentive Based Policy to Improve Efficiency and Performance."
If both the price floor and the safety valve were set at the expected marginal benefit of emission reductions, the combined policy would be analogous to a tax.
See Ian W.H. Parry and William A. Pizer, "Emissions Trading Versus CO2 Taxes Versus Standards," in Raymond J. Kopp and William A. Pizer, eds., Assessing U.S. Climate Policy Options: A Report Summarizing the Work at RFF as Part of the Inter-Industry U.S. Climate Policy Forum (Washington, D.C.: Resources for the Future, November 2007), pp. 83–84.
Uncertainty about the existence of a cap-and-trade program in the future would undermine incentives for banking.
See Henry D. Jacoby and A. Denny Ellerman, "The Safety Valve and Climate Policy," Energy Policy, vol. 32, no. 4 (March 2004), pp. 481–491.
For a discussion of this point, see Burtraw and Palmer, "Dynamic Adjustment to Incentive Based Policy to Improve Efficiency and Performance."
This observation was made by William A. Pizer of Resources for the Future in a personal communication to the Congressional Budget Office.
That requirement would reduce, but not eliminate, the delay. Firms would be able to comply in 2020 by using safety-valve allowances and then banking 2020 allowances that they had obtained by other means (such as receiving for free, making reductions, or purchasing from other firms).
This suggestion was offered by William A. Pizer of Resources for the Future.
If each allowance let firms emit one ton of CO2, a borrowed allowance could permit a firm to emit less than one ton, with the amount of the reduction depending on the discount rate that policymakers chose and the number of years in the future from which the reduction was borrowed. Alternatively, policymakers could allow firms to emit one ton of emissions for each borrowed allowance but could require that they reduce emissions by more than one ton when they pay back the allowance loan.
See Catherine Kling and Jonathan Rubin, "Bankable Permits for the Control of Environmental Pollution," Journal of Public Economics, vol. 64, no. 1 (April 1997), p. 112.
For example, that feature is included in the cap-and-trade proposal (S. 2191) introduced by Senators Lieberman and Warner on October 18, 2007.
See the statement of Joel Bluestein before the Subcommittee on Clean Air, Climate Change, and Nuclear Safety of the Senate Committee on Environment and Public Works, May 8, 2003.
