Life After Kyoto: Alternative Approaches to Global Warming Policies

200704072126The figure at left shows the estimated costs and benefits of the original Kyoto Protocol estimated in the RICE-2001 model for different regions. The figure shows the costs and benefits of the Kyoto Protocol (with full Annex I trading) for the major regions. Costs are production costs (measured negatively), benefits are the environmental benefits of reduced climate change, and net benefits are the difference between costs and benefits.

Note on regions:

“OHI” is other high-income countries, including Japan and Canada.

“Europe” is primarily the European Union

“EE” is Eastern Europe and the countries of the former Soviet Union

“ROW” is the rest of the world

The captioned study by William D. Nordhaus was published December 2005 [PDF, 32 pages]. The purpose of this post is to highlight the discussion of the relative merits of alternative methods for implementing a global price signal for carbon emissions. In particular, to compare:

[1] Harmonized carbon taxes [price-type mechanisms]

[2] Cap-and-trade [quantity-type mechanisms]

The study abstract states:

This study reviews different approaches to the political and economic control of global public goods like global warming. It compares quantity- oriented control mechanisms like the Kyoto Protocol with price-type control mechanisms such as internationally harmonized carbon taxes. The pros and cons of the two approaches are compared, focusing on such issues as performance under conditions of uncertainty, volatility of the induced carbon prices, the excess burden of taxation and regulation, potential for corruption and accounting finagling, and ease of implementation. It concludes that, although virtually all discussions about economic global public goods have analyzed quantitative approaches, price-type approaches are likely to be more effective and more efficient.

Nordhaus examines the [probably fatal] structural flaws of the existing Kyoto scheme, then lays the groundwork for a discussion of the two main alternatives for a successor to Kyoto: price- or quantity-type mechanisms. Nordhaus writes:

Quantity approaches are the norm in environmental policies today, and policies toward global warming are no exception. Policymakers, environmentalists, and economists are so accustomed to quantity constraints in environmental policy that the fundamental advantages of price-type approaches have been largely overlooked. This section reviews ten differences between quantity and price approaches with an emphasis on the advantages of the price-type mechanisms for climate-change policies.

1. The fundamental defect of the Kyoto Protocol is that the policy lacks any connection to ultimate economic or environmental policy objectives. The approach of freezing emissions at a given historical level for a group of countries is not related to any identifiable goals for concentrations, temperature, costs, damages, or “dangerous interferences.” Nor does it bear any relation to an economically oriented strategy that would balance the costs and benefits of greenhouse-gas reductions. It is not inevitable that quantity-type arrangements are inefficient. In principle, they might be designed to choose an emissions path that meets some well-defined and well-designed economic and environmental objectives. However, in practice, quantity approaches in environmental policy tend to be technologically oriented.

Price-type systems such as taxes have a mixed record of efficiency. In this context, the ideal system is relatively simple, as has been described in the section on harmonized carbon taxes above, and is simply the dynamically efficient Pigovian tax. Because of its conceptual simplicity, it might prove simpler to target an efficient tax in this area than efficient quantities.

2. A related issue concerns the baseline policy against which countries set their policies. Quantity limits are particularly troublesome in a world of growing economies, differential economic growth, and uncertain technological change. These problems have become evident under the Kyoto Protocol, which set its targets thirteen years before the control period and used baseline emissions from twenty years before the control period. Base year emissions have become increasingly obsolete as the economic and political fortunes of different countries have changed. The 1990 base year penalizes efficient countries (like Sweden) or rapidly growing countries (such as Korea and the United States). It also gives a premium to countries with slow growth or with historically high carbon-energy use (such as Britain, Russia, and Ukraine).

The baselines for future budget periods and for new participants are further profound problems for the Kyoto Protocol. The natural baseline, were it feasible to calculate, is the zero-restraint level of emissions. That level is in practice impossible to calculate or predict with accuracy, particularly when

abatement policies are in place. Problems would arise in the future as to how to adjust baselines for changing conditions and to take into account the extent of past emissions reductions.

Under a price approach, the natural baseline is a zero-carbon-tax level of emissions, which is a straightforward calculation for old and new countries (more on this below). Countries’ efforts are then judged relative to that baseline. It is not necessary to construct a historical base year of emissions. Countries are not advantaged or disadvantaged by their past policies or the choice of arbitrary dates. Moreover, there is no asymmetry between early joiners and late joiners.

3. One key difference between price and quantity instruments concerns the structure of the uncertainties — and uncertainty is clearly a central feature of climate-change policy. As is well-known in the static Weitzman problem, if the curvature of the benefit function is small relative to the curvature of the cost function, then price-type regulation is more efficient; conversely, if the benefit functions are highly nonlinear while the cost functions are close to linear, then quantity-type regulation is more efficient.

While this issue has received little attention in the design of climate-change policies, the structure of the costs and damages in climate change gives a strong presumption to price-type approaches.16 The reason is that the benefits are related to the stock of greenhouse gases, while the costs are related to the flow of emissions. This implies that the marginal costs of emissions reductions are highly sensitive to the level of reductions, while the marginal benefits of emissions reductions are essentially invariant to the current level of emissions reductions. More generally, where the damages are caused by stock externalities (as is the case for climate change because damages are a complicated function of the stock of greenhouse gases), then the damage function is likely to be close to linear with respect to current emissions. Abatement costs, by contrast, are likely to be highly nonlinear as a function of emissions. This combination of nonlinearities means that emissions fees or taxes are likely to be much more efficient than quantitative standards or auctionable quotas when there is considerable uncertainty, as is clearly the case for climate change.

4. Closely related to point about uncertainty is that quantity-type regulations are likely to show extremely volatile prices for the trading prices of carbon emissions. Carbon prices are likely to be extremely volatile because of the complete inelasticity of supply of permits in the quantity case along with the presumption of quite inelastic demand for permits in the short run.

We have preliminary indications that European trading prices for CO2 are highly volatile, fluctuating in a band and + 50 percent over the last year. More extensive evidence comes from the history of the U.S. sulfur-emissions trading program. SO2 trading prices have varied from a low of $70 per ton in 1996 to $1500 per ton in late 2005. SO2 allowances have a monthly volatility of 10 percent and an annual volatility of 43 percent over the last decade. Figure 7 shows that sulfur prices are much more volatile than oil prices or stock-market prices. This is analogous to a carbon-trading program because the supply is virtually fixed and the demand is inelastic because of the low substitutability of other inputs for sulfur in the short run. Both programs build in some banking features, which can in principle moderate price volatility.

Such rapid fluctuations would be extremely undesirable, particularly for an input (carbon) whose aggregate costs might be as great as petroleum in the coming decades. An analogous situation occurred in the U.S. during the “monetarist” period of 1979-82, when the Federal Reserve targeted quantities (monetary aggregates) rather than prices (interest rates). During that period, interest rates were extremely volatile. In part due to the increased volatility, the Fed changed back to a price-type approach after a short period of experimentation. This experience suggests that a regime of strict quantity limits might become extremely unpopular with market participants and economic policymakers if carbon price variability caused significant changes in inflation rates, energy prices, and import and export values.

5. A fourth advantage of tax mechanisms is the strong fiscal-policy preference for using revenue-raising measures rather than quantitative or regulatory measures. When prices are raised and real incomes are reduced by regulations, this increases the inefficiency losses from the overall tax system. This effect is the “double burden” of taxation (misnamed as the “double dividend” from green taxes).If the carbon constraints are imposed through taxes that are then rebated in taxes with approximately the same marginal deadweight loss as the carbon taxes, then the overall efficiency loss from taxation will be unchanged. If the constraints under a quantity-based system are imposed by allocations that do not raise revenues, then the conventionally calculated abatement costs will underestimate the economic costs and the efficiency losses from the price-raising elements should be added to the abatement costs. The impacts are likely to be large.

While it is possible that emissions permits will be auctioned (thereby retaining the revenues and removing the double burden of taxation), history and current proposals suggest that most or all of the permits are likely to be allocated at zero cost to “deserving” parties, or will be distributed to reduce political frictions. In the cases of SO2 allowances and CFC production allowances, all the permits were allocated to producers. The point here is that using tax approaches rather than quantity-type approaches will help promote a more efficient collection and recycling of the revenues from the carbon constraints.



6. An additional question applies particularly to international environmental agreements and concerns the administration of programs in a world of where governments vary in terms of honesty, transparency, and effective administration. One of the subtle and overlooked problems with quantity-type systems is that they are much more susceptible to corruption than price-type regimes. An emissions-trading system creates valuable tradable assets in the form of tradable emissions permits and allocates these to different countries. Limiting emissions creates a scarcity where none previously existed – in essence printing money for those in control of the permits. Such wealth creation is potentially dangerous because the value of the permits can be used by the country’s leaders for non-environmental purposes rather than to reduce emissions. It would probably become common practice for dictators and corrupt administrators to sell part of their permits, pocket the proceeds, and enjoy wine, partners, and song along the Riviera. Some analysts even believe that the presence of rents of this kind is harmful to economic growth (the “resource curse”).



A few examples will show the perils in the quantitative approach. Simulations suggest that tens of billions of dollars of permits may be available for export from Russia under the Kyoto Protocol. A Russian scientist recently reported that people in Moscow were already considering how to profit from the “privatization” of the Russian carbon emissions permits. Alternatively, consider the case of Nigeria, which had emissions of around 90 million tons of CO2 emissions in recent years. If Nigeria could sell its allowances for $20 per ton under a “clean development mechanism,” this would raise around $2 billion each year of hard currency. This is in a country whose non-oil exports in 2000 were around $600 million.

To prevent unacceptable diversions of funds, any broad-based emissions-trading plan would undoubtedly lead to a major monitoring system and might get bogged down in concerns about the diversion of funds to arms purchases, drugs, money laundering, and terrorism. It would be tempting to make participation and receipt of permits conditional on “good behavior” with respect to terrorism, human rights, environmental concerns, child and prison labor, and other worthy causes du jour. Reducing emissions permits would be a tempting target for sanctions for countries who violate international norms. Of course, the more burdensome are the “ethical” restrictions on the sale of the permits, the less attractive participation becomes for countries, so the plan could easily founder.

A price approach gives less room for corruption because it does not create artificial scarcities, monopolies, or rents. There are no permits transferred to countries or leaders of countries, so they cannot be sold abroad for wine or guns. Any revenues would need to be raised by taxation on domestic consumption of fuels. In fact, a carbon tax would add absolutely nothing to the instruments that countries have today. The only difference would be the international approval of carbon taxes, which probably adds little to their acceptability in corrupt or weakly governed countries. The dangers of quantity as compared to price approaches have been shown frequently when quotas are compared to tariffs in international trade interventions.

7. Problems of financial finagling are not limited to poor, weak, and autocratic states. Concerns arise in the wake of the recent accounting scandals in the U.S. A cap-and-trade system relies upon accurate measurement of emissions by all relevant parties. If firm A (or country A) sells emissions permits to firm B (or country B), where both A and B are operating under emissions caps, then it is essential to monitor the emissions of A and B to make sure that their emissions are in fact within their specified limits. Indeed, if monitoring is ineffective in country A but effective in country B, a trading program could actually end up raising the level of global emissions because A’s emissions would be unchanged while B’s would rise.

It was generally supposed that monitoring would be relatively straightforward in countries with strong legal and enforcement systems such as the United States. This was probably naïve and overly optimistic. The accounting scandals of the last decade have not been limited to dollar scandals, but these have also spilled over into emissions markets.

Some recent cases were described by Ruth Greenspan Bell:

PSEG Fossil LLC, the biggest player in [New Jersey’s emissions trading system], apparently had not installed necessary pollution controls or obtained proper permits. The U.S. Justice Department discovered this and brought an enforcement action, which was resolved in the form of a consent decree. PSEG, without admitting any wrongdoing, agreed to stop selling its credits to other firms and to stay out of the trading system. When PSEG was forced to withdraw, its sheer size and status as one of the largest “suppliers” of credits in New Jersey brought that state’s system close to collapse.

[A]ccording to … Electricity Daily, [authorities] are looking into charges that a Pasadena broker cheated several firms who paid for emissions credits that were never delivered…. A similar example from the United Kingdom was reported … in an account of a government-sponsored auction in which participating companies bid by offering greenhouse gas reductions. An independent review by Environmental Data Services noted strong grounds to suspect that at least half of the claimed emissions reductions were not real, and blamed the inaccuracies on shortcomings in the Department of Environment, Food, and Rural Affairs regulatory controls and “poorly thought through rules.”

If emissions finagling takes place in countries with relatively solid legal systems like the United States and the United Kingdom, it would be foolish to overlook the likelihood of emissions cheating in Russia, Ukraine, and many developing countries.

Such cheating will probably be pandemic in an emissions-trading system that involves large sums of money. There are very poor intrinsic incentives for honesty in a cap-and-trade system. The purchasing unit gets a permit whether or not any true reductions take place by the selling unit. Emissions evasion has even worse incentives than tax evasion. Unlike the emissions-permit case, the recipient of the tax wants the payer to dispense the funds just as much as the taxpayer dislikes dispensing the funds. Tax cheating is a zero-sum game for the two parties, while emissions evasion is a positive sum game for the two parties. If tax evasion in the U.S. is in the order of 10 or 20 percent of taxes due, is there reason to believe that emissions evasion in Ukraine or Romania would be substantially less?

8. One objection to the carbon-tax approach concerns its administration. The issue has been analyzed by David Victor in his analysis of the Kyoto Protocol:

Monitoring and enforcement [of a carbon tax approach] are extremely difficult. . . . In practice, it would be extremely difficult to estimate the practical effect of the tax, which is what matters. For example, countries could offset a tax on emissions with less visible compensatory policies that offer loopholes for energy-intensive and export-oriented firms that would be most adversely affected by the new carbon tax. The resulting goulash of prior distortions, new taxes, and political patches could harm the economy and also undermine the goal of making countries internalize the full cost of their greenhouse gas emissions.



I believe such concerns are serious but can be overcome. The major obstacle to enforcement is the measurement of “net carbon taxes.” As Victor notes, we would need to measure net carbon taxes in the context of other fiscal policies (such as fuel taxes and coal subsidies). For example, suppose that Germany imposed a $50 carbon tax, which would fall primarily on coal. It might at the same time increase its coal subsidies or reduce its gasoline taxes to offset the carbon tax, thereby reducing the level of net carbon taxes. Alternatively, Canada might argue that it has met its carbon-tax obligations by raising provincial stumpage charges on timber. How would the carbon tax be calculated in such circumstances?

One approach would be to calculate that net taxation of carbon fuels including all taxes and subsidies on energy products but not going to second-level tax impacts except in exceptional cases. Such a calculation would require two steps. First, each country would provide a full set of taxes and subsidies relating to the energy sector; second, we would need an appropriate methodology for combining the different numbers into an overall carbon tax rate.

The first issue – obtaining tax rates – is relatively straightforward for market economies. One of the proponents of the tax approach, Richard Cooper, describes the monitoring issue as follows:

Monitoring the imposition of a common carbon tax would be easy. The tax’s enforcement would be more difficult to monitor, but all important countries except Cuba and North Korea hold annual consultations with the International Monetary Fund on their macroeconomic policies, including the overall level and composition of their tax revenues. The IMF could provide reports to the monitoring agent of the treaty governing greenhouse gas emissions. Such reports could be supplemented by international inspection both of the major taxpayers, such as electric utilities, and the tax agencies of participating countries.



Additionally, the levels of taxes and subsidies are generally public knowledge, particularly in market democracies, where they are part of the legislative process. On the other hand, countries with closed political systems might attempt to hide their subsidies. This problem would be particularly troublesome in non-market economies or sectors where fuels are allocated by quantitative measures rather than by the price mechanism. Direct allocation is becoming the exception rather than the rule.

The second issue, calculating the effective carbon tax from the underlying data, is essentially a technical economic issue. Calculations would require certain conventions about how to convert energy taxes into their carbon equivalent. Some of the calculations involve conversion ratios (from coal or oil to carbon equivalent) that underpin any carbon-based system, whether price-based or quantity-based. Others would require input-output coefficients, which might not be universally available on a timely basis. On the whole, calculations of effective carbon tax rates are straightforward as long as they involve first-round or impact calculations (i.e., the rate of tax per unit of carbon emitted) and do not need to involve substitution effects.

To go beyond first-round calculations would require assumptions about supply and demand elasticities and cross-elasticities, might engender disputes among countries, and should be avoided if possible. The procedures would probably require mechanisms similar to those used in WTO deliberations, where technical experts would calculate effective taxes under a set of guidelines that would evolve under quasi-legal procedures.24 Overall, measurement and calculation of effective carbon tax rates seems more tedious than insuperable.

9. An important issue involves the question of how to count initial carbon taxes. Some countries — particularly those in Europe — might claim that they already have high carbon-equivalent taxes because of high taxes on gasoline. They would argue for taking existing taxes into account before requiring them to undergo further obligations.

While this looks like a subterfuge, counting pre-existing taxes as compliance is appropriate and is easily seen as such in the carbon-tax framework. From the point of view of global efficiency, it makes no sense for countries with high existing taxes to add further penalties on top of existing ones before countries with subsidies or no penalties impose their carbon taxes. Therefore, the first step, and one absent from analysis of the Kyoto Protocol, would be a calculation of existing equivalent carbon taxes and subsidies. Our data suggest that, even without its CO2 taxes, Europe is taxing carbon at a rate of approximately $100 per ton carbon more than the United States.25 Given that disparity, it would make no economic sense to require Europe to add even higher carbon taxes on top of its existing ones before other countries had raised their carbon taxes. Moreover, the fact that Europe might be overtaxing carbon today would never come up in the quantity-type approach.

10. The fundamental intuitive concern about price-type approaches is that they fail to “solve” the climate-change problem because they do not limit emissions. This objection is wrong because there is no “correct” level of emissions or of emissions reductions. Indeed, there is today no agreed-upon upper limit on concentrations or temperature change. The price approach reflects the view that we have a better estimate of the size of the penalty on carbon emissions that should be imposed over the next one or two decades than we do of the level of allowable emissions over that period.

Putting this differently, emissions limitation is only an intermediate objective. It is preferable to steer policy toward the ultimate objectives of reducing concentration or temperature changes or limiting net environmental damages rather than at intermediate and intrinsically unimportant objectives like emissions. And this point is emphatically reinforced by the large uncertainties and evolving scientific knowledge. A control mechanism should allow iterative adjustment and movement toward evolving goals, which can be accomplished using either prices or quantities. However, while either prices or quantities can be used as a control mechanism, emissions taxes are more efficient in the face of massive uncertainty because of the relative linearity of the benefits with respect to emissions and the resulting high volatility of prices under an emissions-targeting approach. In other cases, quantities would be more appropriate, but in the case at hand – with a stock externality and vast uncertainty – using quantity controls gives a false impression that the problem is under control.

Non-economists will probably always be uncomfortable with using indirect instruments like prices, just as patients may wonder how little yellow pills can cure their disease. Nonetheless, the fact that prices are more indirect than quantity restraints should not prevent us from recognizing their superior power as a coordinator and motivator for global warming.

A shorter version of this paper was published by Foreign Policy in Focus in March, 2006 as “After Kyoto: Alternative Mechanisms to Control Global Warming”.

Technorati Tags: