9. Carbon Taxes and Energy-Intensive Trade-Exposed Industries: Impacts and options

Ian Parry
Published Date:
March 2015
  • ShareShare
Show Summary Details
Carolyn Fischer Richard Morgenstern and Nathan Richardson

Key Messages for Policymakers

  • A unilateral carbon tax could hurt the international competitiveness of U.S. firms, and resulting production shifts could cause emissions to “leak” outside of the U.S., undermining the environmental benefit of the tax.

  • These effects are most pronounced for energy-intensive, trade-exposed (EITE) industries.

  • Economic analysis suggests that competitiveness and leakage effects from a modest carbon tax are likely to be very small. But they do exist, and may be politically important.

  • Policy options exist for addressing competitiveness and leakage under a carbon tax. None are perfect, but they have important relative advantages and disadvantages.

  • Of these policies, output-based rebates and border carbon adjustments are superior to alternatives (carbon tax exemptions or general corporate tax cuts). But between the preferred options, it is unclear which approach is better. Both have important disadvantages.

  • It is likely that domestic and international politics would be important factors in determining which of these two options for addressing competitiveness and leakage emerge.


An important consideration for any carbon-oriented tax (or regulatory policy) is its effect on the competitiveness of U.S. firms in sectors exposed to international trade. A carbon tax would increase energy costs for U.S. firms, but not their foreign competitors (until and unless those competitors also face a tax or other carbon-pricing policy). Firms in sectors that are both energy-intensive and trade-exposed (EITE) are particularly vulnerable to competitiveness effects. A consequential shift in market share to producers in other nations, particularly those with weak or nonexistent carbon pricing policies, also erodes environmental benefits from a carbon tax, to the extent that emissions are displaced rather than truly abated. This “leakage” effect, along with the loss of U.S. competitiveness itself, is cited by opponents of a carbon tax as a key argument against it.

Most research suggests that these disproportionate effects are relatively modest; however, critics are correct that they are real problems. Furthermore, addressing competitiveness-related leakage need not greatly erode the carbon tax base. To put the scale of the issue into perspective, consider that all industrial-sector CO2 emissions accounted for less than 14 percent of total U.S. emissions in 2011, up only slightly from the 2009 low but down substantially from 1990 levels of almost 17 percent.1 Of course, the relevant EITE firms form a smaller subset of that group, since energy use is highly concentrated. As shown in Figure 9.1, three-quarters of manufacturing output is from industries with energy expenditures less than 2 percent of the value of their output (roughly the average across all manufacturing), while industries with energy expenditures in excess of 5 percent of the value of their output account for only one-tenth of the value of U.S. manufacturing output and less than 2 percent of U.S. GDP.

Figure 9.1Share of GDP, emissions, abatement, and leakage by sector

Source: GTAP database, 2004 base year, and simulations from Fischer and Fox (2010).

While emissions shares represent a rough upper bound of the emissions revenues that might be in play, abatement shares (or the reductions that might be in play) are significantly smaller. Due to the abundance of lower-cost emission reduction opportunities in the electric power sector, studies estimate that the share of emissions reductions under an economy-wide CO2 pricing program coming from the industrial sector would be smaller than that sector’s share in nationwide emissions.2 At the same time, reductions in EITE sectors are predicted to result in disproportionate leakage rates – as much as 3 to 5 times the economy-wide leakage rate.3

As a result, some sort of special treatment for the affected sectors may be justified. Clearly, the most effective approach to reduce carbon leakage is to ensure that other countries take comparable carbon pricing action. But until that becomes politically viable, other policy options are available, some of which have been considered by the U.S. Congress in previous proposals for carbon pricing (in the form of cap-and-trade). This chapter examines four potential approaches for addressing competitiveness concerns for EITE sectors in the context of an economy-wide carbon tax. While not exhaustive, these options are commonly discussed in U.S. policy circles:4

  • partial or full exemption from the carbon tax;

  • rebates to firms tied to their output levels;

  • border adjustments for energy- or carbon-intensive goods; and

  • across-the-board lowering of corporate income tax rates.

The chapter begins with a short discussion of estimates of the size of both leakage and EITE sector competitiveness losses under a carbon tax. Following sections describe policy options for addressing these issues, their relative advantages, disadvantages, and practical design considerations. A final section offers broad conclusions.

The scope of the problem

In order to determine whether special policy treatment for EITE sectors is needed, and if so, to evaluate which policy options are best and how they should be structured, you first have to know the magnitude of the problems you are trying to address. Over the past few years economists have investigated this question and, while their results are sensitive to modeling assumptions, the best evidence suggests that, given a moderate carbon price, both competitiveness losses and emissions leakage are real but modest problems – though higher carbon prices undoubtedly present greater challenges.

Identifying EITE sectors

First, it is important to determine what qualifies as an EITE sector. This is critical for measuring the size of competitiveness or leakage problems, since you need to know where to look for them. But it is also important for policy. If preferential provisions are applied to sectors that are less susceptible to adverse impacts from the carbon tax, they risk weakening the tax’s ability to both raise revenue and to reduce emissions. Moreover, the legality of preferential treatment for EITEs depends on their link to environmental goals – namely, mitigating carbon leakage. The less finely such treatment is focused, the more difficult it is to justify it on environmental grounds, increasing legal risk.

Consequently, relatively narrow definitions of vulnerable sectors are preferred. Such definitions should be limited to sectors that meet the twin hurdles of (1) being highly energy-intensive, either as a result of direct combustion of fossil fuel, especially a carbon-intensive fuel such as coal, or from heavy reliance on carbon-intensive electricity as an input to production; and (2) being highly vulnerable to international competition. For example, the American Clean Energy and Security (ACES) Act (the 2009 cap-and-trade bill also known as Waxman-Markey or H.R. 2454) defines EITE sectors as manufacturing industries (excluding refining) that are at least 5 percent energy (or CO2) intensive and 15 percent trade-intensive, or 20 percent energy-intensive.5 According to an analysis conducted by the Environmental Protection Agency, the ACES definition of EITE industries included 44 presumptively eligible of almost 500 manufacturing industries defined at the six-digit level of the North American Industrial Classification System (NAICS).6 Most of the presumptively eligible industries are within the chemicals, paper, nonmetallic minerals (e.g., cement and glass), or primary metals (e.g., aluminum and steel) sectors. Refineries were not eligible for production rebates under Waxman-Markey, although they were freely granted 2 percent of total allowances without regard to current output, that is, via grandfathering. Other sectors could petition for inclusion. Taken together, these EITE industries account for 12 percent of total manufacturing output and 6 percent of manufacturing employment (half a percent of total U.S. non-farm employment). At the same time, they account for more than two-thirds of direct combustion and process emissions from the manufacturing sector, and slightly more than half when emissions from electricity are also included in the calculations. By way of comparison, the European Union used less stringent criteria for its allocation benchmarking provisions, including a broader group of eligible sectors, albeit with smaller rebates.

Further, with any special tax provision, it is useful to consider the scope for potential administrative complexity. The U.S. Census Bureau reports there were slightly less than 9,200 establishments in 2007 in the presumptively eligible EITE industries under H.R. 2454.7 In addition, Census reports 189 refiners would be covered under the provision for grandfathered allowances, bringing the total number of establishments to just under 9,400. However, that is likely a substantial overestimate of the number of actual firms that would be affected by any tax mechanism tailored to the EITE sector, as many firms own two or more individual establishments.

Carbon taxes and competitiveness

Recent modeling results have examined the impacts of domestic carbon taxes on industrial output for EITE industries (with their scope defined as in the ACES Act) under two distinct timeframes: the short run, typically less than three years, when it is assumed firms cannot alter production techniques; and the long run, when firms can alter production techniques and make a broad range of new investments in response to the carbon tax regime.

Not surprisingly, the results vary considerably by industry. For the case of a $15/ton CO2 tax imposed in the U.S. and other Kyoto Annex I (developed) countries, the estimated output decline in EITE industries ranges from less than 0.1 percent to more than 3 percent.8 Across all such firms, the average output decline is slightly less than 0.5 percent in the short run and about twice that level in the long run (Adkins et al. 2012). Some of the hardest-hit industries include cement, aluminum, lime and gypsum, petroleum refining, and chemicals, rubber, and plastics. These industries operate in highly competitive international markets and generally have difficulty passing along higher costs not also imposed on their competitors. Interestingly, in some industries the long-run output declines are smaller than in the short run, as markets and technologies respond to the new price signals to restore some of the short-run competitiveness losses. The modeling projects that the vast majority of emission reductions achieved by these industries will be from reductions in the emission-intensity of their production (e.g., increased energy efficiency, or shifts to lower-emission production methods), rather than from declines in production associated with increased imports from unregulated countries.

Carbon taxes and emissions leakage

Protecting domestic competitiveness, per se, may confer political benefits, but many attempts to do so run counter to the obligations associated with membership in the World Trade Organization (WTO). However, certain exceptions may be made for measures that are necessary to protect human health or to conserve exhaustible resources, including the global climate.9 Thus, for international legitimacy, it is important to establish a connection between competitiveness impacts and environmental impacts.

“Emissions leakage” is generally defined as the increase in foreign emissions that can be attributed to the effects of emissions reduction policies at home. The “leakage rate” is measured as the increase in foreign emissions as a percentage of (economy-wide) domestic emissions reductions – for example, if a policy led to 100 tons of domestic reductions but 10 additional tons elsewhere (the leakage), it would have a 10 percent leakage rate. Recent studies find overall leakage rates in the range of 5 to 20 percent (Böhringer et al. 2012),10 although these amounts are sensitive to the size of the international coalition taking action.

Importantly, emissions leakage can arise not only from increases in net exports from nonparticipating countries, but also from the drop in global energy prices that follows the withdrawal of demand for fossil fuels on the part of the regulating countries, inducing nonregulating countries to become more emissions-intensive with cheaper energy. Indeed, most leakage estimated in the widely used global trade models results from the latter mechanism, that is, global energy price induced leakage, which explains why in absolute terms a large amount of leakage is expected in the electricity and transportation sectors, despite the lack of trade in those goods. In the case of a carbon price in the Annex I countries only, model results suggest the countries with the largest percentage increase in their emissions are the oil-exporting nations, then India and China, in that order.11

However, leakage rates for certain individual EITE sectors – that is, the increase in emissions by foreign firms in those sectors, relative to the emissions reductions made in those sectors at home – can be much higher than the economy-wide leakage rates. Estimates of leakage for steel and nonferrous metals, for example, can run as high as 50 percent, due to strong competitive pressure and energy price effects for producers of these energy-intensive commodities (Fischer and Fox 2012a).

Put in another perspective, according to simulations by Fischer and Fox (2010), based on 2004 data), EITE sectors in aggregate contribute 7.3 percent of abatement, more similar to their 6.5 percent GDP share than their 10.3 percent share of economy-wide emissions. By contrast, increases in EITE sector emissions abroad represent 23 percent of total leakage. (Figure 9.1 displays the shares of these totals for five broad categories of industries.) Furthermore, while this model finds economy-wide leakage is roughly 8 percent of abatement, leakage rates among the EITE sectors average a much higher 28 percent.

In short, this research indicates that the emission reductions from a U.S. carbon tax are likely to be eroded somewhat by leakage. The much higher rates of leakage in EITE sectors, due in large part to the competitiveness losses discussed above, strengthen the case for targeted policies aimed at reducing the competitiveness impacts of a carbon tax on these sectors. The next section discusses some such policy options.

Description of alternative policy options

Partial or full exemption from the carbon tax

One straightforward option for addressing competitiveness concerns is to exempt certain industries from the broader GHG-reduction policy, in whole or in part. Conceptually, the mechanics of actually providing exemptions are relatively easy. If downstream entities – primarily energy users – are subject to the tax, exempt firms in EITE sectors would have reduced obligations (or perhaps none at all) to cover their emissions.12 For a tax scheme focused on upstream entities (i.e., energy suppliers), exemptions could be provided to downstream firms indirectly via a procedure of credits based on their emissions or fuel use. The credit could come in the form of a tax credit or rebate, or a tradable allowance that could be sold to firms with positive tax liabilities. However, this method would undermine some of the attractiveness of a carbon tax on fuel suppliers as it would require not only a finding of eligibility but also calculations of exempt portions of fuel payments.

The principal advantage of exemptions is that they can be used to protect vulnerable firms or industries in a convincing and targeted way, potentially making it more politically feasible to adopt a higher overall carbon tax rate on domestic emissions.

But exemptions tend to be an inefficient and costly means of addressing competitiveness and leakage issues. This is primarily because exempting certain firms or sectors would almost certainly leave untapped some relatively inexpensive options for cutting emissions in these sectors. If all emissions from the H.R. 2454 presumptively eligible industries were exempt without compensating increases on other sources of emissions, both emission reductions and revenues would decline by about 8–10 percent, depending on the treatment of emissions associated with electricity use.

This approach may also raise equity concerns: if some industries or firms are exempt from participating, a greater burden would be placed on the remaining nonexempt industries. As a result, exemptions would likely increase the total, economy-wide cost of achieving a given emissions target, unless leakage effects are very strong.13

The energy (Btu) tax proposed by the Clinton administration in 1993 is a cautionary lesson concerning the political hazards of exemptions. At that time, many firms and industries made claims of business hardship. As a result, the final House legislation included a long list of exemptions added at the request of members or recommended by the administration. Ultimately, of course, the Btu tax was defeated in the Senate and the policy was never implemented – in part because of the effectiveness and equity concerns raised by the exemptions.

Rebates tied to output levels

A less-obvious but possibly more effective policy remedy for loss of EITE sector competitiveness is a rebate tied to firms’ domestic output. Such an approach would reduce the adverse competitiveness impacts of the carbon tax, while maintaining the incentives the tax creates to reduce the carbon intensity of production (the most important source of emissions reductions for these sectors).

The ACES Act offers a starting point for illustrating how output-based rebates have been proposed in practice. That legislation would have created a national cap-and-trade system for greenhouse gases, under which covered emitters would have been required to surrender an allowance for each ton of carbon they emit. The limited number of such allowances (the cap) and their ability to be traded gives them a market value, imposing a cost on carbon emissions. This cost has similar competitiveness effects on firms in EITE industries as a carbon tax.

The ACES Act addressed this concern by allocating allowances to firms in EITE industries based on their output levels for an initial period.14 Crucially, firms did not receive block allocations simply by doing business in such an industry. Instead, these rebates would have been updated on the basis of recent output levels. If a firm in an EITE sector increases its output, it gets a bigger allowance allocation, regardless of its emissions. If the firm is able to increase its output without increasing its emissions –that is, if it is able to become less carbon-intensive – all the better, since under a cap-and-trade system it can sell its excess allowances.

The rebate thus functions like a subsidy to output of EITE firms, offsetting some or all of their costs of the remaining emissions embodied in that output. Under ACES, sector-wide average carbon intensity was chosen as the benchmark in rebate calculations.15 This allocation is relatively generous, as it implies that a large share of firms in each sector will receive net subsidies. In contrast, the European Union chose a benchmark of the performance of the top 10 percent of firms (i.e., those with lowest emissions intensities) in a sector.

Figure 9.2 shows the estimated output losses for some of the industries most affected by the ACES proposal, with and without rebates and over both short- and long-run horizons. Except for the long-run impacts on the petroleum and coal industries, a large portion of the estimated output losses are offset by the output-based rebates in ACES. At the same time, there is also a small amount of overcompensation in several industries.

Figure 9.2Output effect on EITE industries of a unilateral $15/ton CO2 tax, with and without rebates

Output-based rebates can be designed in conjunction with a carbon tax to work much like those under cap-and-trade. Several potential approaches offer similar economic incentives, but the legal treatment might differ. Importantly, the WTO Subsidies Code prohibits specific subsidies for industry and does not allow for environmental exceptions (unlike for tariffs, discussed in the next section). One approach would simply offer firms in EITE sectors direct rebates, based on the carbon tax rate multiplied by a sector benchmark emissions rate and their output. This is relatively simple, but seems likely to be viewed by the WTO as an illegal subsidy.

Another option would be to embed the rebate into the carbon tax in the determination of the tax base for EITE sectors; that is, the carbon tax liability would be for emissions above a level equal to the benchmark rate multiplied by the firm’s output. With this option, especially if the benchmark is generous, the tax would have to be refundable (allowing for negative tax payments) to preserve the incentive effects for relatively clean firms. A final (hybrid) option is to exempt EITE sectors from the carbon tax, but then to regulate them separately with a distinct market-based climate policy, like a tradable performance standard,16 along with measures to ensure the price follows the carbon tax.17 For example, if the performance standard is not very tight, allowing EITE firms to sell excess credits to firms subject to the carbon tax would preserve abatement incentives for the former and reduce the tax liability of the latter. If the performance standard is more stringent, allowing EITE firms the alternative compliance method of paying the tax would ensure that EITE carbon prices do not rise above the carbon tax rate.18

Reducing adverse competitiveness impacts of the carbon tax while preserving (most of) the incentives to reduce emissions are the key advantages of output-based rebates versus the simple carbon tax exemption for EITE industries. However, both options reduce net government revenues, by the extent of the emissions payments either rebated or forgiven; the amounts thus depend on the generosity of the rebate, but could range to 8 or 10 percent.19

An additional disadvantage of the rebate mechanism is its complexity. In addition to the challenge of accurately and equitably identifying EITE industries (a problem faced by all competitiveness policies), under output-based rebates the government must devise a dynamic rebate system and track the output of all firms in EITE industries. For example, information on current and historic output levels would be required, imposing an additional (though probably modest) reporting burden over and above that already required by the carbon tax. For some industries, it will be difficult to define appropriate output metrics across the varied products (e.g., chemicals). Presumably, the Internal Revenue Service, in cooperation with other agencies, would need to develop the needed metrics and collect the relevant data from the affected parties.

Perhaps the most important disadvantage of output-based rebating is its potential to distort firm production decisions in ways that yield unintended consequences. The rebate does discourage consumers from diverting their demand to imported alternatives, but it also discourages them from seeking less emissions-intensive alternative products/services, including conservation practices. In other words, output-based rebates may encourage production of the very energy-intensive materials that a carbon tax is designed to discourage. The costs of these distortions tend to grow as more trading partners adopt carbon pricing, and can soon outweigh the benefits from limiting any remaining carbon leakage (Böhringer et al. 2011; Fischer and Fox 2012b).

Border carbon adjustments

Border carbon adjustments (BCAs) address the leakage and competitiveness impacts of a carbon tax by ensuring that consumers face consistent carbon pricing signals, regardless of the country of origin for energy-intensive manufactured products. Import adjustments would require importers to pay an equivalent carbon tax, based on the actual or estimated emissions embodied in the product. Full border adjustment would add relief for exports in order to implement a fully “destination-based” carbon tax. However, most proposals (including the ACES Act) address only imports.20 Recent studies indicate that the additional leakage reduction benefits of export adjustments are small, relative to those from import adjustments (Böhringer et al. 2012).

BCAs entail several important design issues: To which products should they apply? How are embodied emissions calculated? Should adjustments be made to the BCA for carbon regulations in the country of origin? Which exporting countries should be exempt? As trade-related measures, BCAs are controversial and should be designed with an eye toward maintaining compliance with WTO obligations and the principle of “common but differentiated responsibilities” enshrined in the UNFCCC (see Cosbey et al. 2012). Indeed, only mitigating carbon leakage is likely to be a legally permissible justification for interfering with international trade; protecting competitiveness contradicts WTO principles.

Scope of application: Like the other policy options, a BCA should only be applied to qualified EITE sectors. This narrow scope ensures that the benefits of reduced leakage outweigh the substantial administrative costs. It also is more likely to be legal under trade law – or, specifically, to be found to justify a GATT Article XX exemption from nondiscrimination obligations.

Defining embodied emissions: The definition of embodied emissions determines not only the extent to which consumers will pay for carbon embodied in imports, but also the extent to which foreign producers have an incentive to reduce their emissions. This incentive only occurs if producers can influence their calculated carbon content. Basing the BCA on actual embodied emissions would thus create some incentive for exporters to reduce their emissions; however, such calculations may be complex, costly, and legally challenging. Ideally, all firms would accurately report their embodied emissions, credibly verified by a third party. In practice, a default measure of embodied emissions is necessary for cases when actual data are either unavailable, unreliable, or prohibitively costly to obtain. Regardless of this benchmark, foreign producers should be afforded the opportunity to demonstrate if their emissions are actually lower. Although it raises compliance costs, this opportunity may be essential for WTO compatibility; it also can preserve some incentives for firms that can become relatively clean.

Setting a high default benchmark – for example, at the level of embodied emissions of a “worst technology” producer in the foreign country – would create a powerful incentive to report actual embodied emissions. Some producers, especially those in developing countries, would probably need technical assistance to measure their emissions, or else a high benchmark would unduly penalize them. Another option is to set the benchmark at the average emissions from the sector in the country of export, although sufficient data to calculate this average may not be available. A third benchmark option is to use domestic (that is, U.S.) industry average emissions. This would avoid any discrimination by country of origin, possibly increasing the legal and political palatability of BCAs.21 But purely domestic-based benchmarks may systematically underestimate embodied emissions if domestic sectors are relatively clean. Indeed, any benchmark reduces incentives for more carbon-intensive producers, that is, those well above the benchmark, to cut their emissions, since marginal reductions have no effect on the BCA.

A reasonable compromise would be to adopt technology-based benchmarks for direct emissions related to production processes (since these are likely to be reasonably consistent across countries) and country-specific emissions factors for indirect emissions related to power generation (since this type of reporting already occurs to the UNFCCC). This option seeks to balance administrative burden and accuracy (Cosbey et al. 2012).

Adjustments to the embodied carbon tariff: The BCA should account for any output-based rebates offered to domestic industries, to ensure equal treatment with foreign producers. The BCA should also account for carbon prices paid in the exporting country; of course, this calculation may not be so simple, as corresponding tax breaks (like OBR) can undermine the price impact. A BCA need not attempt to account for the costs of non-price-based carbon regulation, as the intent is not to adjust for abatement costs that reduce emissions but rather for the cost of payments for the embodied emissions that remain.

Exemptions: One type of adjustment is simply to exempt imports from certain countries. The goals of such exemptions should be to minimize administrative costs, ensure the environmental effectiveness of the carbon tax, and foster compatibility with the UNFCCC goals of “common but differentiated responsibilities” (CBR). For example, Cosbey et al. (2012) recommend exemptions for the following types of countries:

  • those with an effective national emissions cap, and for sectors with an effective sectoral cap (as this by definition prevents leakage);

  • least developed countries (for CBR compliance); and

  • those from which imports in these sectors or goods fall below a de minimus level.

Offering any sort of exemption entails a need for trans-shipment provisions – without them, goods could be produced in countries subject to BCA, then shipped to the U.S. via exempt countries to avoid BCA. The resulting complexity necessitates a high threshold of applicability that probably excludes manufactured goods and covers only a small number of commodities.

If these design issues can be appropriately addressed, the principal advantage of BCA is the preservation of carbon price signals for consumers. The disadvantages relate to administrative complexity and an arguably more combative approach to leakage, from the perspective of international relations.

Lower corporate income taxes

U.S. corporate income taxes are higher than those in most other developed economies, and many U.S. firms, including those in EITE sectors (see Chapter 8), argue that these tax rates put them at a competitive disadvantage with foreign rivals. Accordingly, some have suggested that the U.S. should broadly reduce corporate taxes, possibly as part of a revenue-neutral policy package which would include a carbon tax.

Relative to a carbon tax alone, broad corporate tax reductions would not only make U.S. companies, including those in EITE industries, more competitive, but might also broadly reduce incentives to shift production to other jurisdictions.22 Overall corporate tax bills might be higher or lower relative to today under a policy package including both a carbon tax and corporate tax cuts, depending on the size of the corporate tax cut relative to the carbon tax, and the energy intensiveness of the firm. But corporations’ tax bills would generally be lower (and their competitiveness greater) under such a package than under an equivalent carbon tax alone. And since firms would be paying the carbon tax, they would still have an incentive to make their production less carbon-intensive.23

However, broad corporate tax reductions are a blunt instrument for addressing competitiveness concerns associated with the carbon tax, as they make no distinction between EITE and non-EITE firms. Many firms liable for U.S. corporate taxes are not in EITE sectors, and there is therefore little need for policies to shield them from the effects of a carbon tax. Even small reductions in corporate tax rates will have significant effects on the revenue the federal government receives from such taxes, but only a small part of both the change in tax receipts and increased competitiveness would occur in the EITE sectors.24 Indeed, effective U.S. corporate tax rates vary widely across sectors and firms, based on differences in tax breaks, foreign tax liabilities, and also profitability.25 Corporate taxes are based on profits, so a lower rate only has a significant effect in sectors with significant profits. If firms in EITE sectors are already struggling, the benefits of a lower tax rate are limited; in fact, it can reduce the benefits from operating losses they would carry forward.

As a result, it is difficult to justify such across-the-board corporate tax reductions purely based on their ability to reduce negative effects on EITE firms. If one thinks that U.S. federal corporate taxes are currently set about right, then cutting those rates to address EITE competitiveness and leakage concerns arising from a carbon tax is an expensive, poorly directed policy. Alternatively, if you think that federal corporate taxes should be lower for reasons unrelated to a carbon tax, you might favor a revenue-neutral package that trades lower corporate taxes for a carbon tax. However, if the rationale for reducing corporate taxes is to improve the competitiveness of U.S. firms in EITE industries, a carbon tax erodes some or all of these benefits for those sectors. This puts us back where we started – one of the policy options discussed above is still needed. Cutting broad corporate taxes even more deeply could achieve the desired competitiveness benefits for firms in EITE sectors. But those cuts would be deeper than necessary to achieve the target competiveness boost for the larger group of trade-exposed (but not energy-intensive) firms (or the desired economy-wide efficiency improvements of lower corporate taxes). In short, the cart would be driving the horse.

A strong business advocate might argue that nothing is wrong with this approach – that lower corporate taxes are always a good thing. But the mismatch between broad corporate tax reductions and the problem of EITE sector competitiveness under a carbon tax exists whatever level of general corporate taxes one considers ideal. Once corporate taxes have been set at this ideal point, further cuts to address EITE sector competitiveness overshoot the target for other firms. In the extreme case, if one believes the correct corporate tax rate is zero, further cuts to benefit EITE sectors are not possible (setting aside negative taxes – that is, direct subsidies to U.S. firms – as both politically implausible and a clear violation of trade law).

In other words, the general competitiveness of an economy relative to other countries is a different (and possibly much larger) issue than the competitiveness of EITE sectors under a carbon tax. A general reduction in corporate taxes helps with the former, but only indirectly with the latter. And any help it would provide must be weighed against other uses for corporate tax revenues – cuts in other, other distortionary taxes, deficit reduction, infrastructure or technology investment, or anything else worthwhile that the government may do. More targeted policies for addressing EITE competitiveness would almost certainly have a smaller revenue impact.

One important advantage of broad corporate tax reductions, however, is that they could not be construed as breaches of international trade law. Nothing in U.S. WTO obligations restricts the freedom to set domestic corporate tax rates. Measures that specifically target EITE sectors are much more likely to trigger legal scrutiny.

At the same time, lobbyists for EITE sectors may argue that under a carbon tax + corporate tax cut package, their industries are still disadvantaged relative to other industries, since all industries benefit roughly equally from the tax cuts, but EITE sectors are disproportionately harmed by the new carbon tax. As the 1993 Btu tax example mentioned above illustrates, heavy lobbying is likely under any new tax proposal, and this may be a superficially compelling equity argument. However, firms making this argument are complaining about their tax position relative to firms in other industries – that is, firms against which they do not compete. EITE firms’ domestic competitors would be similarly (though not necessarily identically) affected by a carbon tax. And as noted above, EITE firms might be more or less competitive with foreign firms in the same industry, but would be better off than they would be under a carbon tax alone.

Comparing the options

The goal of a carbon tax would be to raise revenue for public finances while reducing an important distortion to the economy, namely, the lack of a price reflecting the damages from greenhouse gas emissions.

Both exemptions and output-based rebates would forego part or all of the revenue that might be raised from a carbon tax on those sectors. Since EITE sectors in the U.S. account for perhaps 10–15 percent of emissions (depending on how broadly or narrowly the category is defined), the foregone revenue would be significant but not extraordinary. At the same time, with border adjustments, the revenue consequences would be minor, particularly if tariff revenues are returned to exporting countries (as suggested by some), but even if retained domestically, since imports represent only a share of production in EITE industries. A corporate income tax reduction, on the other hand, would absorb the large part (or all) of the carbon tax revenue.

In terms of cost-effectiveness, a broad literature has analyzed the design of climate policies in the context of unilateral or Annex I actions. It is well understood that the optimal policy requires uniform emissions pricing across sectors and countries, ensuring that marginal abatement costs are equal.26 When such coverage cannot be complete, so-called second-best policies have been considered as complements to emissions pricing among the covered sectors.

Although analysis has not yet been done for the effects of corporate tax reforms on EITE industries, it is clear that the least efficient of the options discussed is setting lower carbon prices (that is, lower carbon tax rates) for the EITE sectors or exempting them altogether. The reason is that it forgoes some important incentives for those sectors to reduce their emissions. Still, when leakage effects are very strong, differentiated emissions pricing can lower the costs (both domestically and globally) of meeting a given global emissions reduction target, by shifting more of the burden toward sectors with less leakage potential. Although some studies find substantial justification for emissions price differentiation to deter leakage, from a global perspective, using these differentiated prices only slightly lowers the economic cost of unilateral climate policy versus uniform pricing.27

Unlike exemptions, output-based rebating retains the incentive effects of the carbon price on encouraging firms in EITE industries to find cost-effective ways to reduce their emissions intensity. At the same time, like exemptions, rebates help keep product prices from rising, and thus discourage substituting toward uncovered alternatives as a means of reducing emissions. However, this de facto subsidy also discourages conservation as a means of reducing emissions, resulting in an efficiency trade-off. In fact, it may actually encourage production of the energy-intensive products. While this may be attractive to some firms, it may be inconsistent with the overall aims of the carbon emissions reduction policy. Thus, this treatment is only recommended for sectors in which a large share of the production loss that might be associated with a carbon tax would arise from import displacement rather than reduced domestic consumption. Furthermore, it may only be recommended while the group of countries with carbon regulation is relatively small: as the climate coalition grows larger, the costs of the lost conservation incentives ultimately outweigh the benefits of reducing leakage (Böhringer et al. 2011).

BCAs tend to be significantly more cost-effective than these other options, since they not only address competitiveness-related leakage by discouraging carbon-intensive imports from nonparticipating nations, but also encourage those same countries to adopt a carbon pricing regime. In addition, they also preserve the incentives for consumers to find less carbon-intensive products. They are, however, the most politically controversial option and face significant legal and administrative barriers.

A few economic studies compare BCAs with output-based rebates and exemptions. In combination with a carbon tax, all the options promote domestic production to some extent and reduce leakage. However, none would necessarily reduce global emissions in a given sector: while they reduce emissions abroad, they expand domestic firms’ emissions. The net effect depends on the relative responses of domestic and foreign producers to price changes and their relative emissions intensities. Fischer and Fox (2012a) find that for most U.S. sectors, full border adjustment would be most effective at reducing global emissions, but output-based rebates can be more effective than import adjustments (particularly when those are limited by WTO law to a weaker, nondiscriminatory standard) at reducing emissions leakage and encouraging domestic production (see Figure 9.3).

Figure 9.3Carbon leakage rates by sector and policy option

Source: Based on Fischer and Fox (2012a), which assumes that a $14/ton CO2 carbon tax is applied to energy-intensive industries.

Figure 9.3 also illustrates the point that while these policies can address some of the leakage related to competitiveness impacts, they do not eliminate all leakage. That is because the carbon tax also causes global energy prices to fall, encouraging more carbon-intensive fossil fuel use among foreign EITE sectors.

Of course, the effect of policies on leakage rates is distinct from the effect on economic well-being, which recognizes the benefits of emissions reductions as well as the costs of mitigation. Many trade models have compared these policies in a cost-effectiveness framework that holds global emissions (and thus the environmental benefits) constant, requiring the domestic carbon price to adjust in response to antileakage measures. From a global cost-effectiveness standpoint, border adjustment for imports achieves most of the benefits associated with full border adjustment, and results in lower costs regardless of coalition size. However, BCAs also entail the largest shifting of the burdens of climate regulation toward developing and emerging economies (Böhringer et al. 2012).28 At the same time, requiring that the tariff revenues be returned to exporting countries can allow developing countries overall to enjoy net benefits from the policy (Fischer and Fox 2012b). When the regulating coalition is sufficiently small, output-based rebating offers a middle ground of improving cost-effectiveness without shifting the burden excessively (see also Böhringer et al. 2011).


Critics point to competitiveness and leakage as major problems that would undermine the effectiveness of a carbon price. While economic modeling indicates that the magnitude of these impacts may not be as severe as these critics suggest, they are legitimate concerns that merit attention.

There are, however, policy options available for mitigating these twin problems. At least three specific policy options focused on EITE industries exist: sector-specific exemptions, output-based rebates similar to the ones advanced in the ACES Act, and border carbon adjustments. Broad corporate tax reform not aimed exclusively at EITE industries also addresses loss of competitiveness, and could be viewed as a crude pro-competitiveness/anti-leakage policy.

As shown in Table 9.1, some of these policies are likely to be more effective than others. Broad-based corporate tax reform, while attractive for many reasons, is not an effective mechanism for addressing the disproportionate impacts imposed on EITE industries by a carbon tax. Exempting firms from carbon taxes altogether is also a fairly blunt instrument that can be both inefficient and inequitable.

Table 9.1Pros and cons of alternative policies
Tax ExemptionSimpleReduces revenues
EITE firms exempt from some/all of the carbon taxPolitical resonanceUndercuts incentives to reduce emissions
Output-based RebatesMaintain incentivesComplex
EITE firms subsidized based on their outputto cut emissionsReduces revenues
Border Carbon AdjustmentsDirectly targetsLegally vulnerable Internationally
Imports taxed/exports subsidized for EITE sectorsleakage/competitivenesscontroversial Complex
General Corporate Tax Cuts Lower corporate taxes for all firmsNonePoorly targeted to EITE sectors Reduces revenues

Two other policies are relatively attractive, although each has its own advantages, disadvantages, and challenges. BCAs offer the potential to address both competitiveness and leakage concerns most directly, although their design poses many challenges and their use may be constrained by a variety of legal problems. Output-based rebates are more effective in addressing competitiveness than leakage impacts, could be implemented simultaneously with the carbon tax itself, and might be more likely to survive legal challenge. But they are relatively complex to administer and might be more prone to unintended consequences from firms’ production decisions. For either policy, the devil is largely in the details. But both appear to offer viable methods for addressing legitimate concerns about the adverse impacts of a U.S.-only carbon tax.


This paper was presented at an October 23, 2012, workshop at Resources for the Future “Roundtable Discussion on the Economics of Future US Fiscal and Carbon Policy” and at the November 12, 2012, conference “The Economics of Carbon Taxes” sponsored by the American Enterprise Institute, Brookings Institution, IMF, and Resources for the Future.

EPA, “Draft Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990–2011,” Washington, DC, February 13, 2013.

For example, the U.S. Energy Information Administration (EIA) in its 2009 analysis of H.R. 2454 (see below) estimated the industrial sector would contribute 9.7 percent of total reductions in 2020 and 5.5 percent in 2030. Reflecting recent declines in natural gas prices and the corresponding increase of gas use in the electric power sector, today’s outlook probably calls for slightly greater shares from the industrial sector, although EIA analysis is not yet available on this issue.

Other less commonly discussed options include cuts in employer payroll taxes or subsidies to energy-efficient technologies for EITE industries.

Energy intensity is defined as the share of total costs that is from energy use (combustion and electricity). Trade intensity is defined as exports plus inputs divided by the sum of the value of shipments and imports.

U.S. EPA, 2009.

Output losses in the case of unilateral taxation would be somewhat higher.

These exceptions are described in Article XX of the General Agreement on Tariffs and Trade (GATT).

For a coalition of Kyoto Annex I (developed) countries, excluding Russia, with a reduction target of 20 percent below 2004 emissions; the corresponding carbon prices range from $15 to $60 per ton CO2. Studies of similar unilateral U.S. policies find slightly larger leakage rates (Fischer and Fox 2012b).

Sweden has adopted this approach. Its industries are exempt from 80 percent of the CO2 tax.

The freely allocated allowances were designed to start phasing down in 2025, and to end completely by 2030.

For example, CO2 emissions per ton of product, although the precise definition would be left to EPA.

Under a tradable performance standard, regulators would set a benchmark carbon intensity or other measure of efficiency on a sector-by-sector basis. Firms that do better than this standard would receive credits tradable to firms that do not reach it.

A feebate is sometimes mentioned as another possible option, although it is almost identical to a carbon tax combined with output-based rebates. The principal distinction is that with a feebate you could have a lower tax rate by blending exemptions and output-based rebates, likely resulting in lower abatement incentives.

For example, the Alberta intensity-based emissions trading scheme allows the compliance option of contributions to the Climate Change and Emissions Management Fund at a price of CND 15 per metric ton of CO2e.

Rebates in excess of emissions liabilities on average would likely run afoul of WTO law on subsidies.

For one thing, the legal basis for export rebates is shakier than that for import adjustments. Rebates fall under the WTO Subsidies Code, which does not have an environmental exception akin to Article XX in the GATT, which might allow for discrimination among “like” products if necessary in order to protect health or conserve natural resources.

Opinions vary as to whether BCAs that differentiate across countries can be compatible with WTO obligations to treat imported goods no less favorably than “like” domestic products and to accord Most Favored Nation Treatment to all WTO members. Exceptions might be made for policies deemed necessary to preserve the environment, as long as they are the least trade restrictive option.

On the other hand, if lower corporate taxes also mean similarly reduced taxes on multinational corporations’ repatriated foreign income, there might be little or no additional incentive to avoid shifts of production (and therefore leakage) that a carbon tax would promote.

Carbon and corporate tax interactions could be complex. One relatively straightforward observation is that carbon tax payments (both direct and indirect) are partly corporate-tax-deductible, depending on their effect on corporate profits. Market power, the availability of substitutes, and other factors determine the extent that firms will be able to pass carbon tax costs on to consumers or be forced to absorb them with corresponding reductions in profits and, therefore, corporate tax payments.

Of course, it is possible to cut corporate taxes only for EITE sectors. Doing so is broadly similar to a carbon tax plus rebates tied to output levels (discussed above), except that it is achieved via a different part of the tax code and is only useful to firms that are making profits (and therefore paying corporate tax).

Of course, this assumes an otherwise ideal world, barring additional major distortions.

In one analysis, second-best carbon prices for EITE sectors were found to be roughly 40 percent lower than for other sectors overall (Böhringer et al. 2010).

See also the full special issue in Energy Economics (December 2012).


    AdkinsL.R.GarbaccioM.HoE.Moore and R. D.Morgenstern.2012. Carbon Pricing with Output Based Subsidies: Impact on U.S. Industries over Multiple Timeframes. Discussion Paper 12–27. Washington, DC: Resources for the Future.

    BöhringerC.E.Balistreri and T. F.Rutherford2012. “The Role of Border Carbon Adjustment in Unilateral Climate Policy: Insights from a Model-Comparison Study”.Discussion Paper 12–54The Harvard Project on Climate AgreementsCambridge, MA.

    BöhringerC.J. C.Carbone and T. F.Rutherford.2012. Unilateral Climate Policy Design: Efficiency and Equity Implications of Alternative Instruments to Reduce Carbon Leakage. Energy Economics34(Supplement 2): S208S217 (December).

    BöhringerC.C.Fischer and K. E.Rosendahl.2011. Cost-Effective Unilateral Climate Policy Design: Size Matters. Discussion Paper 11–34. Washington, DC: Resources for the Future.

    BöhringerC.A.Lange and T. F.Rutherford.2010. Optimal Emission Pricing in the Presence of International Spillovers: Decomposing Leakage and Terms-of-Trade Motives. Working Paper 15899. April. Cambridge, MA: National Bureau of Economic Research.

    CosbeyA.S.DroegeC.FischerJ.ReinaudJ.StephensonL.Weischer and P.Wooders.2012. A Guide for the Concerned: Guidance on the Elaboration and Implementation of Border Carbon Adjustment. ENTWINED Report. Stockholm.

    FischerC. and A. K.Fox.2010. On the Scope for Output-Based Rebating in Climate Policy: When Revenue Recycling Isn’t Enough (or Isn’t Possible). Discussion Paper 10–69. Washington, DC: Resources for the Future.

    FischerC. and A. K.Fox.2012a. Comparing Policies to Combat Emissions Leakage: Border Tax Adjustments versus Rebates. Journal of Environmental Economics and Management64(2): 199216 (September).

    FischerC. and A. K.Fox.2012b. Climate Policy and Fiscal Constraints: Do Tax Interactions Outweigh Carbon Leakage?Energy Economics34(Supplement 2): S218S227 (December).

    U.S. Environmental Protection Agency. 2009. The Effects of H.R. 2454 on International Competitiveness and Emission Leakage in Energy-Intensive Trade-Exposed Industries: An Interagency Report Responding to a Request from Senators Bayh Specter Stabe-now McCaskill and Brown. Washington DC: EPA.Available at:

    Other Resources Citing This Publication