## wpiea2021239-print-pdf

## Source details

**Canonical URL:** [wpiea2021239-print-pdf](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021239-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021239-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021239-print-pdf.pdf.json)

---

### Introduction and framing
- Definition: A ‘BCA’ is a charge on the carbon content of imported products intended to ensure treatment equivalent to domestic carbon pricing, potentially combined with rebates for the carbon content of exports.
  - The term ‘charge’ reflects implementation either as an explicit tax or a requirement for importers to purchase allowances from a domestic emissions trading system (ETS) or separate allowance pool.
  - Remission of tax on exports is optional; without rebates a BCA differs from a VAT-style border adjustment and raises origin/destination tension under the Paris Agreement.
- Policymaker objectives for BCAs:
  - Preserve competitiveness of domestic industries facing domestic carbon pricing (especially EITE industries).
  - Reduce emissions leakage (offsetting of domestic mitigation by foreign emissions increases).
  - Strengthen incentives for carbon pricing and mitigation action in other countries.
- Key design choices to address:
  - Sectoral coverage; measurement of embodied carbon; treatment of exports; accounting for foreign mitigation; preservation of domestic mitigation incentives; impact on revenue; administrative/compliance costs; WTO/legal risk.

### Rationales: Competitiveness
- EITE characteristics and context:
  - Primary examples: iron, steel, aluminum, refined petroleum products, pharmaceuticals, plastics, glass, ceramics, cement, textiles, wood products.
  - EITE industries are typically 80 percent or more of manufacturing emissions.
  - Manufacturing is usually around 10-30 percent of nationwide emissions.
  - EITE industries typically account for around 10-20 percent of GDP.
- EU ETS EITE classification rule examples:
  - EITE if ETS raises production costs at least 5 percent and trade share with non-EU countries is above 10 percent; also deemed EITE if production cost increases or trade share exceed 30 percent.
- Mechanics of carbon pricing on firm costs:
  - Carbon price P raises cost from C(E) to C(E) + P·E; unit production cost increase decomposed into efficiency cost C and transfer payment T = P·E1 (with direct and indirect components P·Edir and P·Eind).
- IMF staff illustration for a carbon price of $50 per ton in 2030 (estimated unit cost increases):
  - Basic metals: around 25-30 percent in India.
  - Basic metals: 12-15 percent in China.
  - Basic metals: less than 10 percent in the EU and US.
  - Textiles, machinery, fabricated metals: less than 10 percent in each case.
- Empirical evidence notes:
  - Studies generally have not identified large production effects to date, often due to low carbon prices and compensating measures (e.g., free allowances).
  - Pass-through rates for EITE industries are typically between about zero and 50 percent; power sector pass-through in EU has been large.
- Role of BCAs for competitiveness:
  - A BCA charging embodied carbon at the difference between domestic and foreign carbon prices, with symmetric export rebates, would adjust differences in transfer payments.
  - Competitiveness impacts depend critically on how embodied carbon is measured.
- Non-price mitigation policy contrast:
  - Non-price policies impose efficiency cost triangle C but avoid transfer payment T; charging imports the domestic shadow price would generally not be appropriate and could conflict with WTO rules.

### Rationales: Leakage
- Leakage fundamentals:
  - Leakage arises when production migration or expansion abroad offsets domestic emission reductions; most relevant for EITE sectors.
  - Domestic emission reductions decompose into intensity improvements and production reductions; only migration-caused production reduction causes leakage.
- Illustrative leakage examples:
  - If domestic intensity falls 25 percent and 5 percent of production shifts abroad, leakage rate will be 35 and 70 percent respectively if foreign emissions intensity is 200, and 400 percent of domestic production respectively. (Annex 2 approximation: gives 33 and 67 percent for e*/e equal to 2 and 4.)
- Empirical evidence on leakage rates:
  - Mixed; leakage can be significant and tends to be larger for small open economies.
  - Misch and Wingender (2021) reported:
    - On average, carbon leakage amounts to 25 percent.
    - Leakage rates vary from 20 to almost 50 percent in individual European countries.
    - Leakage rates are less than 15 percent in China, the EU14+UK aggregate, India, and Japan.
    - Leakage rate is 7 percent in the US.
- International fuel-price channel:
  - Leakage can arise from international fuel price responses; zero for a small price-taking country, but potentially significant for large-country groups and essentially unaffected by BCAs.
- Paris Agreement link:
  - Nationally determined pledges may address leakage if binding, but pledges are voluntary and may not equate to nationwide caps.

### Promoting carbon pricing abroad and signaling/coordination roles
- Fiscal incentive channel:
  - BCAs create a fiscal incentive for exporters to adopt domestic carbon pricing to avoid BCA liabilities, but this incentive appears modest given small shares of emissions embodied in trade.
- Illustrative statistics (from Figures cited):
  - Carbon embodied in EITE exports from China and India to the EU and US: about 3 percent of China and India’s domestic carbon emissions.
  - Formal incidence on China and India of a $50 BCA imposed by the EU and US: only 0.1-0.15 percent of GDP.
  - Embodied carbon in EITE exports to all trading partners: 10 and 8 percent of domestic carbon emissions for China and India respectively; formal incidence approximately 0.45 and 0.3 percent of GDP for China and India respectively.
  - Embodied carbon in EU-27 and US EITE exports to the world: 5 and 2 percent of domestic emissions respectively; formal incidence of a rest-of-world BCA is small (figure truncated in source).
- Signaling and coordination:
  - BCAs may signal carbon pricing priority and encourage coordination (e.g., border-free trading zones, International Carbon Price Floor (ICPF)).
  - ICPF could scale global carbon pricing more effectively than BCAs because BCAs price only emissions embodied in trade flows.

### Design choices and empirical considerations
- Key design elements:
  - Sectoral coverage (initial EITE focus vs. broader coverage).
  - Measurement approach (domestic benchmark vs. country-specific vs. firm/plant-level).
  - Export treatment (rebates).
  - Accounting for foreign mitigation and free allocations.
  - Exemptions (e.g., least developed countries).
- Sectoral coverage recommendations:
  - Limit initially to EITE industries to target leakage, reduce administrative burden, and lower legal risks; consider broader coverage later.
- Measuring embodied carbon:
  - Country-specific benchmarks best align incentives but are administratively complex.
  - Domestic industry-average benchmarks are simpler and pragmatic initially but give little incentive for foreign emission reductions and may benefit emerging market exporters.
  - Firm- or plant-level measurement is most precise but adds significant administrative complexity; a rebuttability provision (third-party verification) can improve WTO defensibility but risks gaming.
- Export rebates and revenue implications:
  - Rebates justified for competitiveness and potentially environmental reasons; should be based on firm-wide production or industry benchmarks to avoid perverse export incentives.
  - Empirical points:
    - Embodied carbon in EITE exports: 10 percent of domestic emissions in China, 8 percent in India, 2 percent in the US (2015).
    - A $50 per ton BCA on imports in 2015 would have raised import-charge revenues of around 0.1-0.2 percent of GDP in China, India, EU-27, and US.
    - Export rebates would offset 25 and 60 percent of import-charge revenues on EITE products in the US and EU-27 respectively; in China and India rebate-related revenue losses would substantially outweigh import-charge revenues.
  - Export rebates reduce BCA revenue but may enable higher domestic pricing and net revenue through broader domestic pricing.
- Adjusting for foreign carbon pricing and mitigation:
  - Reducing the BCA rate by the amount of foreign carbon pricing is justified when exporting countries do not rebate exporters.
  - Differentiating BCA rates across countries raises legal challenges; an alternative is separate BCAs with export rebating by each trading partner.
- Exemptions for LDCs:
  - Lower BCA rates for LDC exporters could be WTO-compatible and better align with equity principles; rules of origin needed to prevent trans-shipment.

### Legal considerations and WTO interaction
- Legal uncertainties and possibilities:
  - WTO rules permit harmonizing measures for indirect taxes; uncertainty exists whether carbon pricing (especially ETS) qualifies as an indirect tax and whether country-differentiated charges violate MFN.
  - If not WTO-consistent, Article XX exceptions (notably sub-paragraph (g)) could be invoked but require strict tests and avoiding arbitrary/unjustifiable discrimination.
- Specific WTO implications:
  - BCAs with carbon taxes: may be characterized as customs duties or charges under GATT Article II:2(a); taxing embodied carbon raises interpretive uncertainty (is embodied carbon an input?).
  - Export rebates: could be allowable under SCM footnote 1 for indirect taxes but must be offered on same terms to all domestic firms covered by the tax.
  - BCAs with ETSs: import allowance purchase requirements may be domestic regulation under GATT Article III:4 and risk treating imports less favorably; export rebates could be prohibited export subsidies under SCM if not equally available for domestic sales.
- MFN and Article I concerns:
  - Differentiating imports by country-of-origin embodied carbon may violate MFN absent objective criteria; rebuttal mechanisms and objective tests (e.g., development indicators for LDC exemptions) can reduce risk.
- Article XX considerations:
  - Article XX(a) and (g) could justify BCAs but measures must effectively address climate change and satisfy the chapeau (no arbitrary or unjustifiable discrimination).
  - Few historical measures survive chapeau scrutiny, emphasizing careful design.

### BCAs versus alternative instruments (comparative assessment)
- Instruments described:
  - Exemptions or rebates (e.g., South Africa carbon tax exemption for first 70 percent of emissions).
  - Tradable emissions rate standard (e.g., Canada).
  - Free allowance allocations under ETS (e.g., California, EU, Korea, New Zealand).
  - Feebates and output-based rebates.
- Comparative matrix (summary of effects):
  - Protecting competitiveness of EITE industries: BCAs—Yes; Exemptions—Partial/less effective if indirect emissions not covered; Tradable standard—Partial; Free allowances—Partial.
  - Limiting leakage: BCAs—Yes (more so if country-varying and with export rebates); Others—Partial or less effective.
  - Promoting mitigation abroad: BCAs—Yes; Exemptions—No; Tradable standards/free allowances—Mixed.
  - Domestic mitigation incentives for EITE: BCAs—Modest incentive; Exemptions/free allowances—No incentive; Tradable standard—Reduces emissions per unit.
  - Revenue: BCAs—Preserves revenue (but export rebates can offset); Others—forgo revenue.
  - Administrative burden: BCAs—Significant if beyond EITE; Others—Modest.
  - Legal risk: BCAs—Higher WTO challenge risk; Others—have faced no legal challenges to date.
- Effectiveness conclusions:
  - BCAs are potentially more effective for competitiveness and leakage if they vary by trading partner embodied carbon and include export rebates.
  - Other instruments often reduce domestic mitigation incentives and forgo revenue; their effectiveness declines as deeper decarbonization raises efficiency costs relative to transfers.
- Administrative and transitional recommendations:
  - Start with domestic industry benchmarks and EITE coverage for simplicity; transition to country-specific measures over time.
  - Design legislation emphasizing environmental objectives and include rebuttal/verification options for foreign firms.

### Analytical results, leakage scenarios, and Annex summaries
- Algebraic import charge approximations:
  - τ = (P − P*)·E* + {C(E) − C(E*) − P·(E* − E)}  (A1.1)
  - If P = 0: τ = −P*·E* + {C(E) − C(E*)}  (A1.2)
  - First-order approx with −C'(E) = P yields τ ≈ (P − P*)·E* (traditional BCA).
- Leakage formula highlights (Annex 2):
  - Leakage depends on relative emissions intensity e*/e, changes in production and emissions intensity; can exceed 100 percent if foreign intensity is much higher.
  - Numerical scenarios show wide potential leakage outcomes, including over 100 percent in extreme intensity differentials.
- Annex 3 empirical literature headline:
  - Large literature estimates leakage rates mostly around 10-30 percent (including production location and fuel-price channels); ex post EU evidence sometimes shows limited leakage but may underestimate future risks as prices rise.
  - EU ETS price rose from $6 per ton in 2017 to over $70 per ton in 2021; EU 2030 pledge tightened from 40 to 55 percent reduction vs 1990.
- Annex 4 (ICPF rationale/design):
  - Two obstacles: many parties and non-comparable pledges; competitiveness/economic scaling constraints.
  - ICPF design: focus on a small group of key emitters and on a minimum carbon price floor; include equity/flexibility arrangements and possible exemptions for participants from BCAs applied to non-participants.
- Annex 5 (WTO compatibility key points):
  - WTO allows rebates for indirect taxes on exports and charges on imports not in excess of indirect taxes on like domestic products, but embodied-carbon measures raise interpretive issues.
  - Article I (MFN), Article III (national treatment), Article XX (exceptions), and SCM Agreement interactions create multiple legal constraints and possible failure modes for BCAs if designed poorly.

### Conclusions and policy implications
- BCAs offer advantages over alternatives for addressing competitiveness and leakage, with appeal rising as decarbonization deepens, but design details are critical.
- Recommended transitional and design choices:
  - Initially benchmark BCAs to domestic industry embodied carbon (industry average), limit coverage to EITE industries, design legislation with environmental objectives, allow rebuttal via third-party verification, and plan transition to country-specific measures.
- Limitations and trade-offs:
  - BCAs alone provide limited incentives for rapid, global mitigation; they attenuate but do not remove free-rider problems.
  - Unilateral BCAs covering limited trade shares may produce modest fiscal incentives for exporters and risk international price dispersion and legal challenges.
  - An International Carbon Price Floor (ICPF) could more effectively scale up global carbon pricing through coordinated minimum prices among key emitters.
- Policy judgment:
  - Early-stage carbon pricing may not justify the administrative, political, and legal complexities of BCAs; pressure for BCAs will rise with more ambitious domestic pricing, potentially increasing interest in formal price coordination mechanisms.

*Source: wpiea2021239-print-pdf - https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021239-print-pdf.pdf*

### 1.    Introduction

### 1.    Introduction

### Overview
- As countries consider more aggressive climate mitigation policies, the question of whether some form of ‘border carbon adjustment’ (BCA) is appropriate has become central to the wider climate debate.
- The EU’s recent announcement of a BCA proposal in July 2021, as well as BCA proposals in the United States, have heightened interest in this instrument, not least as countries are revising their climate strategies in the run up to COP26 in November 2021.
- Underlying this interest is a concern that more ambitious unilateral actions—higher domestic carbon pricing, in particular—will be discouraged by cross-border effects. BCAs are considered as a possible way to limit harm from lack of uniformity in and coordination of national policies.
- Definition used in this chapter: A ‘BCA’ is taken to be a charge on the carbon content of imported products intended to ensure treatment equivalent to domestic carbon pricing, potentially combined with rebates for the carbon content of exports.
  - The term ‘charge’ reflects that the BCA could be implemented either as an explicit tax, or as a requirement for importers to purchase allowances from a domestic emissions trading system (ETS) or separate allowance pool.
  - The remission of tax on exports is treated as an optional feature—and many proposals do not allow for such rebating. Without such an adjustment, a BCA differs from a ‘border adjustment’ in the VAT sense (destination basis) and points to a tension with the Paris Agreement’s assignment of responsibility on an ‘origin’ basis.

### Policymakers’ stated objectives for BCAs
- To help preserve the competitiveness of domestic industries in the presence of domestic carbon pricing, particularly for energy-intensive, trade-exposed (EITE) industries—this improves economic efficiency in the sense of preventing distortions in the relative prices of domestic and foreign goods (i.e., clean and polluting industries at home and abroad are treated alike) and can aid the political acceptability of carbon pricing.
- To reduce the risk of emissions leakage, that is, partially offsetting emissions increases in foreign countries induced by domestic mitigation policy—this objective signals a concern not only with national welfare but with global welfare more generally.
- To strengthen incentives for carbon pricing and mitigation action in other countries—there is a direct fiscal incentive to the extent that non-BCA countries effectively forgo revenue on their exports collected by the importing BCA country, and indirectly BCAs might help to strengthen the international credibility of carbon pricing schemes.
- Note: These objectives are related but distinct; for example, leakage may be significant even if competitiveness effects (decline in domestic production) are small, and vice versa.

### Key design and policy choices to be addressed
- Sectoral coverage (which sectors and products to include).
- Measurement of embodied carbon in traded goods.
- Treatment of exports (whether to rebate tax on exports).
- Accounting for mitigation actions in foreign countries (how to credit foreign policies or prices).
- Considerations also include:
  - Preservation of domestic mitigation incentives.
  - Impact on revenue.
  - Moderating administrative and compliance costs.
  - Limiting risks of challenges under World Trade Organization (WTO) rules or of trade retaliation.

### International reception and trade considerations
- Trading partners may be concerned that BCAs could camouflage protectionist measures.
- The impacts of carbon pricing and the BCA should be considered jointly: a BCA may be better thought of as mitigating a competitive disadvantage that domestic carbon pricing would otherwise create for the implementing country by raising costs on domestic producers.
- Where countries already use measures such as free emissions permit allocations, a BCA would in effect replace one mechanism with another; appropriate design that avoids over-compensation may alleviate trading partners’ concerns.

### Analytical focus and empirical issues
- The paper focuses on analysis of BCAs from a national rather than a collective perspective.
- Key empirical issues highlighted include:
  - Leakage rates.
  - Burdens of BCAs on trading partners.
  - Emissions shares of traded products.
  - Embodied carbon in imports and exports for different countries.
  - Impacts of BCAs on industrial costs.

### Organization of the paper
- Sections 2, 3, and 4 focus on potential rationales for BCAs, design issues, and instrument choice issues respectively.
- Section 5 provides brief concluding remarks.
- Although the focus is on the tax policy aspects of BCAs, the legal context is recognized and a brief account is given in Annex 5.

*Source: wpiea2021239-print-pdf - 1.    Introduction*

### 2. Rationales for BCAs

### 2. Rationales for BCAs

### Competitiveness
- Around 30 carbon pricing schemes had been implemented by 2021 at the national and EU levels, with prices and coverage varying widely (and many not applying to the industrial sector).
- Implicit carbon prices in mitigation pledges for 2030 also vary widely.
- EITE (emissions-intensive, trade-exposed) industries:
  - Primary examples include iron, steel, aluminum, refined petroleum products, pharmaceuticals, plastics, glass, ceramics, cement, textiles, and wood products.
  - EITE industries are typically 80 percent or more of manufacturing emissions.
  - Manufacturing is usually around 10-30 percent of nationwide emissions.
  - EITE industries typically account for around 10-20 percent of GDP.
- EU ETS EITE classification example:
  - Industries are classified as EITE if the ETS increases their production costs at least 5 percent and their trade share with non-EU countries (imports plus exports relative to production) is above 10 percent.
  - Sectors are also deemed EITE if production cost increases or trade share exceed 30 percent.
- Mechanics of carbon pricing on firm costs (schematic):
  - Carbon price P raises cost curve C(E) to C(E) + P∙E, reducing emissions per unit from E0 to E1 and raising unit costs from UC0 to UC1.
  - Increase in unit production cost comprises:
    - Efficiency cost, C (cost of switching to cleaner technologies/fuels).
    - Transfer payment to government or allowance sellers, T = P∙E1, which can be divided into P∙Edir (direct emissions) and P∙Eind (indirect emissions, chiefly electricity).
  - At moderate abatement levels, the efficiency cost is small relative to the transfer payment; efficiency cost rises with extent of abatement.
- Estimated distributional impacts for a carbon price of $50 per ton in 2030 (IMF staff illustration):
  - Basic metals unit costs increase by around 25-30 percent in India.
  - Basic metals unit costs increase by 12-15 percent in China.
  - Basic metals unit costs increase by less than 10 percent in the EU and US.
  - Cost increases for textiles, machinery, and fabricated metals are less than 10 percent in each case.
- Empirical evidence:
  - Empirical studies have generally failed to identify large production effects of carbon pricing to date, albeit at generally low levels of carbon pricing and often with compensating instruments such as free allowances.
  - Most studies suggest pass through rates for EITE industries of between about zero and 50 percent, in contrast to the power sector where carbon pricing in the EU has been largely passed forward in higher consumer prices.
- Role of BCAs:
  - A BCA charging the carbon content of imports (direct and indirect) at a rate equal to the difference between domestic and any foreign carbon prices, and symmetrically for exports, would fully adjust for differences in carbon prices.
  - For imports from jurisdictions without carbon pricing, such a charge would make foreign producers face the same transfer payment component (T) as domestic producers with the same emissions intensity.
  - Export rebates in the BCA can put domestic producers on level terms with foreign producers in external markets.
  - Competitiveness impacts of BCAs depend critically on measurement of embodied carbon.
- Non-price mitigation policies:
  - Non-price policies increase production costs by the efficiency cost triangle C (a shadow price) but generally impose lower private costs than carbon pricing because they do not impose the transfer payment rectangle T.
  - Charging imports the domestic shadow price on embodied emissions would generally not be appropriate because domestic firms are not paying that shadow price; the objective of restoring competitiveness would be better met by charging imports some estimate of the efficiency cost faced by domestic firms.
  - Practical constraints: efficiency costs are unobserved, and charging them would not fit within WTO rules except potentially under the “environmental exception.” Compensation for efficiency costs on exports would likely conflict with WTO rules on subsidies.
  - A country using carbon pricing that adopts a BCA could choose to apply it to imports from a country achieving equivalent emissions reductions through regulations.

### Leakage
- Unilateral carbon pricing creates the risk that domestic emissions reductions are partly offset by increased emissions abroad (leakage) via migration of production or expansion of foreign production; BCAs can reduce such leakage.
- Leakage is most relevant for EITE industries; CO2 emissions from domestic transportation and buildings are largely immobile.
- Decomposition of domestic industry emission reductions:
  - Reduction in emissions intensity of domestic production (firms adopt cleaner technologies/fuels).
  - Reduction in domestic production due to lower domestic demand.
  - Reduction in domestic production due to migration of production to foreign firms — only this channel causes leakage.
- Leakage rates need not scale with competitiveness impacts; example (holding domestic demand constant):
  - If carbon pricing incentivizes a 25 percent reduction in domestic industry emissions intensity, and 5 percent of production shifts abroad, then the leakage rate will be 35 and 70 percent respectively if foreign emissions intensity is 200, and 400 percent of domestic production respectively.
  - (Annex 2 approximation: with D̂ set at zero, e*/e · Ŷ/(Ŷ + ê) gives 33 and 67 percent for e*/e equal to 2 and 4 respectively.)
- Empirical evidence on national-level leakage rates:
  - Mixed results; leakage can be significant and tend to be larger for small open economies.
  - Most empirical literature finds modest or no evidence of leakage, possibly reflecting limited scope of policies and methodological limitations.
  - One recent study (Misch and Wingender (2021)) reports:
    - On average, carbon leakage amounts to 25 percent.
    - Leakage rates vary from 20 to almost 50 percent in individual European countries.
    - Leakage rates are less than 15 percent in China, the EU14+UK aggregate, India, and Japan.
    - Leakage rate is 7 percent in the US.
- Leakage via international fuel price effects:
  - Leakage might also result from increased foreign fossil fuel demand in response to downward pressure on international fuel prices from mitigating countries.
  - This form of leakage would be zero for unilateral mitigation by a small country that is a price taker in international fuel markets but could be significant for a group of larger countries.
  - This form of leakage depends on aggregate reductions in fossil fuel consumption and is essentially unaffected by the form of mitigation instrument (carbon pricing or other) or by accompanying measures such as BCAs.
- Paris Agreement and leakage:
  - National limits on emissions under the Paris Agreement may address leakage to the extent that they bind.
  - Under Paris, countries are responsible for production emissions within their borders; potential foreign leakage might be neutralized by stronger mitigation if foreign countries honor binding nationwide targets.
  - In practice, pledges are voluntary, may not be fully achieved, and do not always take the form of nationwide emissions caps.

### Promoting Carbon Pricing in Other Countries
- BCAs create a fiscal incentive for trading partners to impose carbon pricing by allowing them to eliminate BCA liabilities and capture tax revenue instead; however, the incentive appears modest given small shares of emissions in trade flows.
- Illustrative statistics (from Figure 9a and b):
  - Carbon embodied in EITE exports from China and India to the EU and US are about 3 percent of China and India’s domestic carbon emissions.
  - The formal incidence on China and India of a $50 BCA imposed by the EU and US would be only 0.1-0.15 percent of the former’s GDP.
  - The effective incidence on Chinese and Indian producers is likely much lower because much of the import charge is likely passed forward to domestic consumers in the EU and US in the form of higher product prices.
  - If a broader range of countries impose BCAs, embodied carbon in EITE exports to all trading partners is 10 and 8 percent of domestic carbon emissions for China and India respectively, with formal incidence approximately 0.45 and 0.3 percent of GDP for China and India respectively.
  - By contrast, embodied carbon in the EU-27 and US EITE industry exports to the world is only 5 and 2 percent of domestic emissions, and the formal incidence of a BCA imposed by the rest of the world on them is less than (figure truncated in source).
- Overall implication: the fiscal incentive from BCAs for major exporters like China and India to scale up economy-wide carbon pricing appears modest unless BCAs are widely adopted by many trading partners.

*Source: IMF staff (chapter 2, "Rationales for BCAs", as provided in the supplied content).*

### 0.05 percent of their GDP.

### wpiea2021239-print-pdf - 0.05 percent of their GDP.

### BCAs as signals and coordination mechanisms
- BCAs may promote pricing in other countries by signaling that carbon pricing is the centerpiece of mitigation strategy, thereby influencing other countries’ mitigation choices.
- Unilateral BCAs can lead to coordination among countries to create border-free trading zones with a common external charge, potentially evolving into more formal arrangements for coordinating over carbon pricing.
- A BCA combined with other incentives could promote participation in an international carbon price floor (ICPF) among large emitting countries; an ICPF would scale up global carbon pricing more effectively than BCAs because BCAs price only carbon embodied in trade flows rather than all emissions.
- BCAs might be applied by ICPF participants to non-participants, which could complicate ICPF design due to the need to agree on terms for both BCAs and the ICPF.

### Key design challenges and legal considerations for BCAs
- Multiple objectives must be balanced: preserving domestic mitigation incentives, raising revenue, limiting administrative/compliance burdens, and minimizing WTO legal risks.
- WTO rules permit harmonizing measures for indirect taxes; key uncertainties include whether carbon pricing counts as an indirect tax (likely more difficult for an ETS than a carbon tax) and whether charges varying by exporting country’s carbon intensity would violate the Most Favored Nation (MFN) principle.
- If a BCA does not meet WTO rules, it might still qualify as an exception under Article XX if viewed as addressing environmental issues, though demanding legal tests must be met.
- Design forms (import tax vs. allowance purchase requirement) can be operationally equivalent if allowance pools are set up appropriately; allowance purchase requirements may increase price uncertainty and administrative complexity due to collaboration needs between customs and environment ministries.

### Sectoral coverage: EITE initially vs. broader coverage
- Limiting BCA to EITE industries initially is justified on competitiveness, targeted leakage, administrative, and legal grounds.
  - Determining EITE vs. non-EITE classification is straightforward with clearly specified criteria.
  - Measuring embodied carbon for raw materials in many EITE industries is relatively straightforward using input-output tables and emissions factor data.
  - Narrow focus reduces administrative burdens and legal risks since leakage motivation is more transparent for EITE products.
- Broader BCA coverage would more comprehensively address competitiveness and leakage, provide stronger incentives for carbon pricing elsewhere, and increase charges collected from trading partner imports, but:
  - Administrative practicality is questionable for non-EITE manufacturing, services, mining, and electricity.
  - Challenges include measuring embodied carbon for services and high-value manufacturing products and increased compliance burdens.
- Preliminary recommendation: EITE (at least initially); domestic initially to ease transition; later aim for country-specific.

### Measuring embodied carbon: domestic vs. country-specific benchmarks
- Country-specific benchmarks most directly address BCA objectives:
  - Preserve relative costs of equivalent domestic and foreign products despite carbon pricing.
  - Impose higher charges on trading partners with higher embodied carbon, incentivizing them to implement carbon pricing.
- Domestic emissions-intensity benchmarks are simpler administratively and may be appropriate during transition, but:
  - Provide little or no incentive for foreign exporters to reduce emissions.
  - If updated, domestic abatement could lead to lower charges on competing imports even if foreign emissions unchanged.
  - Emerging market economies would face much lower formal burdens if BCAs use domestic rather than country-specific benchmarks.
- Pragmatic approach: use domestic embodied carbon initially (industry average rather than cleanest firms) with view to transition to country-specific BCAs over time.
- Firm- or plant-level measures would be more precise given production heterogeneity, but greatly increase administrative complexity; for now, industry-level data may be more practical.
- A ‘rebuttability’ provision allowing firms to claim rebates if their embodied carbon is lower than the industry average (subject to third-party verification or audit risk) should improve WTO compatibility but risks gaming via production shifts between plants.

### Rebates for domestic exporters and revenue implications
- Rebates for domestic carbon pricing on embodied carbon in domestic exports are warranted on competitiveness grounds and potentially on environmental grounds; they:
  - Offset increased export costs from domestic carbon pricing, preserving competitiveness of the average exporter and limiting leakage.
  - Should be based on firms’ overall production or industry-wide benchmarks to avoid incentives to use more emissions-intensive production for export.
- Empirical points:
  - Embodied carbon in EITE exports is 10 percent of domestic emissions in China, 8 percent in India, and 2 percent in the US (2015).
  - A $50 per ton BCA on imports would have raised revenues from import charges of around 0.1-0.2 percent of GDP in China, India, EU-27, and US in 2015.
  - Export rebates would offset 25 and 60 percent of the revenues from import charges on EITE products in the US and EU-27 respectively; in China and India revenue losses from export rebates would substantially outweigh revenues from import charges.
- Export rebates reduce BCA revenues directly but may enable higher domestic carbon pricing and net revenue through broader domestic pricing.

### Use of BCA revenues and legal risk mitigation
- Using revenue for green investment, just transitions, or international climate finance may reduce legal risks by framing the measure as environmental rather than protectionist, though earmarking challenges and additionality concerns remain.

### Adjusting import charges for foreign carbon pricing and other mitigation efforts
- If exporting country does not rebate its carbon tax to exporters, reducing the BCA rate by the amount of foreign carbon pricing is justified to preserve competitiveness.
- If exporting country rebates (e.g., via its own BCA arrangement), charging the full domestic tax upon import is appropriate.
- Politically or environmentally, exemptions might be considered for countries that have “done enough” under the Paris Agreement even if their carbon pricing is lower or uses non-price measures; no single best approach exists.
- Practical and legal considerations:
  - Up-to-date details on carbon pricing for power and industry sectors are available and most relevant for EITE embodied emissions; residential and transport fuel pricing have little relevance for EITE production costs.
  - Adjustments needed when foreign firms receive free allocations; conventions needed to account for exchange rate volatility and overseas emissions price volatility.
  - Legal justification for differentiating BCA rates across countries may be difficult.
- Alternative: trading partners each maintain separate BCAs with export rebating (VAT-style border adjustment):
  - Economically similar to adjusting the charge for foreign carbon pricing, but more straightforward legally and administratively.
  - Accommodates cases where foreign carbon price exceeds domestic price.
  - Potential issue: inconsistent EITE lists across countries could cause double taxation or lack of rebate, suggesting need for a common list.

### Exemptions for least developed countries (LDCs)
- Applying lower BCA rates for exporters in LDCs would improve competitiveness for LDC exporters with little at stake for BCA-implementing countries and might be WTO compatible.
- Excluding LDCs would align with equity and “common but differentiated responsibilities” principles and might be consistent with the WTO’s Enabling Clause if exemption criteria use objective development indicators.
- Country-based exemptions need design features (rules of origin) to prevent trans-shipment of goods through exempted countries; these rules may be burdensome but could be warranted.

*Italic: Source: IMF staff and referenced materials as contained in the supplied content.*

### 4.    BCAs versus Alternative Instruments

### 4.    BCAs versus Alternative Instruments

### Alternative instruments currently used and their mechanisms
- Exemptions or rebates:
  - Exempting all, or some, of the emissions from EITE industries from carbon pricing (in a downstream pricing program), as in South Africa (exemptions for the first 70 percent of emissions from the carbon tax).
  - Rebates for carbon prices implicit in fuel and electricity inputs (in an upstream pricing program).
- Tradable emissions rate standard:
  - Allowing EITE industries to participate in a tradable emissions rate standard (firms can fall short of the standard if they buy credits from firms exceeding the standard), as in Canada.
- Free allowance allocations under ETS:
  - Granting free allowance allocations related to industry benchmarks and past emissions for relatively clean producers for EITE industries (under an ETS) which are cancelled if firms shut down or move abroad, as in California, the EU, Korea, and New Zealand.
- Other equivalents:
  - Feebates (sliding scale of fees/rebates on products with above/below average emission rates).
  - Returning revenues from carbon pricing collected from EITE industries in output-based rebates to those industries.

- Examples from Table 2 (assistance measures for EITE industries with carbon pricing):
  - Canada: A tradable emission rate scheme.
  - California: Free allowances under the ETS. BCA applies to imported electricity.
  - EU: Free allowances under the ETS but planning transition to BCA.
  - Korea: Free allowances under the ETS to qualifying EITE industries.
  - Netherlands: Levy on emissions above benchmark level (based on relatively clean firms).
  - New Zealand: EITE facilities receive free allowances of 60-90 percent of the industry benchmark.
  - South Africa: Exemptions for the first 70 percent of emissions from the carbon tax.

### How instruments compare on key policy metrics (summary of Table 3)
- Protecting competitiveness of EITE industries:
  - BCAs: Yes
  - Exemptions for EITE industry emissions from pricing: Full exemption is less effective (if it does not apply to indirect emissions)
  - Tradable emission rate standard: Partially
  - Free allowances under ETS: Partially
- Limiting leakage:
  - BCAs: Yes (noted that if BCA varies across trading partners according to embodied carbon and includes export rebates, effectiveness increases)
  - Exemptions: Full exemption is less effective (if it does not apply to indirect emissions)
  - Tradable emission rate standard: Partially
  - Free allowances under ETS: Partially
- Promoting mitigation and carbon pricing in other countries:
  - BCAs: Yes
  - Exemptions: Full exemption removes incentives for direct (but not indirect) emissions
  - Tradable emission rate standard: Reduces emissions per unit of output
  - Free allowances under ETS: Maintains all incentives
- Mitigation incentives for domestic EITE industries:
  - BCAs: Modest incentive
  - Exemptions: No incentive (full exemption removes incentives for direct emissions)
  - Tradable emission rate standard: Reduces emissions per unit of output
  - Free allowances under ETS: No incentive
- Revenue implications:
  - BCAs: Preserves revenue from pricing domestic emissionsa
  - Exemptions, tradable emission rate standard, free allowances: Forgoes revenue
  - Note a: The BCA itself raises additional revenue (unless forgone revenue from export rebates exceeds collections from import charges).
- Administrative burden:
  - BCAs: Significant if coverage beyond EITE products; depends on design features
  - Other instruments: Modest (they largely build off existing capacity)
- Legal risk:
  - BCAs: Could be challenged as subsidy but has not been; Risk of legal challenge under WTO
  - Other instruments: Have faced no legal challenges to date (free allowance allocations might be interpreted as a subsidy under WTO law)

### Effectiveness of BCAs relative to alternatives
- BCAs are potentially more effective than other instruments in addressing competitiveness and leakage, especially if:
  - The BCA varies across trading partners according to embodied carbon, and
  - The BCA includes export rebates.
- Exemptions for EITE industries from carbon pricing are less effective unless they also include compensation for charges on indirect emissions and unless import prices vary across countries depending on emissions intensity.
- Tradable emissions standards and free allowance allocations under ETSs are partially effective:
  - Firms are not charged for a large portion of their direct emissions that remain after compliance, but they are charged for indirect emissions and incur abatement costs.
  - Their relative effectiveness declines with deeper decarbonization, as efficiency costs become more significant relative to transfers.
- Most other instruments reduce mitigation incentives for domestic industries and forgo revenue:
  - Full exemptions and free allowance independent of current emissions remove mitigation incentives for direct emissions.
  - Tradable performance standards promote reductions in emissions intensity but do little to reduce output levels of emissions-intensive products.
  - Other instruments forgo revenues that could be collected from pricing domestic industry emissions.

### Administrative, legal, and transitional considerations
- Administrative and legal burdens for other instruments are relatively modest as they build on existing capacity; they have faced no legal challenges to date.
- BCAs raise administrative and legal complexities:
  - Benchmarking against domestic industry embodied carbon may be advisable initially for administrative simplicity and to ease the transition for emissions-intensive trading partners.
  - Transitioning later to country-specific measures based on embodied carbon could improve efficiency but increase complexity.
  - From a WTO perspective, BCA motivation and design in legislation should be based on environmental, rather than protectionist or revenue-raising, considerations.
  - Allowing foreign firms to “rebut” industry-level assessments with third-party certifications on their individual emissions intensity can help credibility and legal defensibility.

### Key analytical results and scenarios (Annex highlights)
- Graphical and algebraic analysis comparing carbon pricing to regulation:
  - Case 1 (domestic firm subject to regulations): Regulations achieving the same emissions reduction per unit as a carbon price leave the firm’s cost curve unchanged; private cost increase is much less than under actual carbon pricing (efficiency cost C much smaller than transfer payment T = P∙E at moderate reductions).
  - Case 2 (foreign firm subject to regulations while domestic firm faces carbon pricing): Foreign firm operating under equivalent regulation may still enjoy a cost advantage at moderate abatement levels (difference between UC1 and UCr), meaning exempting it from a BCA would not be warranted on competitiveness grounds; environmental motivation for a BCA is diminished if foreign regulations keep emissions intensity aligned.
- Algebraic expression for an import charge τ that levels deemed unit costs:
  - τ = (P − P*)·E* + {C(E) − C(E*) − P·(E* − E)}  (A1.1)
  - If home country uses only regulation (P = 0): τ = −P*·E* + {C(E) − C(E*)}  (A1.2)
  - If home country deploys a carbon tax with −C'(E) = P, first-order approximation yields τ ≈ (P − P*)·E* (traditional BCA).
- Leakage and competitiveness interactions (Annex 2):
  - Leakage defined as L ≡ −ΔE*/ΔE.
  - Change in domestic emissions: ΔE ≈ e·ΔY + Y·Δe, with Δe assumed negative.
  - Change in foreign emissions: ΔE* = −e*·ΔNX = −e*(ΔY − ΔD).
  - Approximation for leakage (A2.4) shows leakage depends on:
    - Relative emissions intensity (e*/e)
    - Percent changes in production and emissions intensity (signs noted: (+) for e*/e; (−) for other terms)
  - Key observations:
    - Leakage is symmetric in exports and imports.
    - Leakage is proportional to original relative emissions intensity of foreign production.
    - Leakage is positive if home’s net exports fall; can be negative only if domestic falls more than domestic production (unlikely).
    - If emissions intensity abroad greatly exceeds home, leakage can be over 100 percent (total emissions increase).
  - Numerical/illustrative findings:
    - Figure 2.1 (scenario with a 5 percent decline in domestic demand, half of local demand met by domestic production and half by net imports) illustrates a wide range of leakage outcomes including over 100 percent when relative intensity of foreign emissions is very high.
    - Figure 2.2 shows leakage can be high for low domestic emissions reductions even with small production shifts; higher domestic emissions cuts increase the denominator and can lower leakage, but larger cuts may cause larger production shifts, making final effect ambiguous.
    - Figure 2.3 shows that if foreign emissions intensity remains lower than domestic after pricing (e.g., hydro or nuclear abroad), shifts of production can reduce total emissions even if counted as leakage; conversely, very dirty foreign production raises total emissions.

### Conclusions and policy implications
- BCAs have appeal over other instruments for addressing competitiveness and leakage, and this appeal will likely rise with greater decarbonization, but design details are critical.
- Recommended design and transitional considerations:
  - Initially benchmark BCAs against domestic industry embodied carbon for administrative simplicity and to ease transition for trading partners; consider transitioning later to country-specific measures.
  - Limit BCAs to EITE industries to moderate compliance costs and increase credibility as a leakage-targeting measure.
  - Design legislation with environmental rather than protectionist or revenue-raising motivations to align with WTO perspectives.
  - Allow foreign firms to rebut industry-level assessments with third-party certifications of individual emissions intensity.
- Limitations of BCAs:
  - BCAs by themselves provide limited incentives for rapid, global mitigation; they attenuate the free-rider problem but do not remove it.
  - BCAs covering only a minor fraction of trading partners’ emissions and imposed unilaterally by multiple countries could create significant international price dispersion.
  - An International Carbon Price Floor (ICPF) could potentially offer more comprehensive coverage, with coordinated and progressively ramped prices to encourage the ambition needed for the global challenge.
- Trade-offs:
  - The scale of competitiveness and leakage effects may not justify the administrative, political, and legal complexities of a BCA in early stages of carbon pricing.
  - Pressure for BCAs will rise as regions and countries adopt more ambitious emissions pricing; unilateral BCAs emerging may increase interest in formal price coordination mechanisms as a route to effective and efficient global mitigation.

*Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021239-print-pdf.pdf*

### Annex 3. Empirical Literature on Emissions Leakage: A Quick Summary

### Annex 3. Empirical Literature on Emissions Leakage: A Quick Summary

### Empirical findings on leakage rates
- A large empirical literature has estimated leakage rates, mostly for large countries or broad groups of advanced countries implementing carbon pricing, at around 10-30 percent—but reflecting leakage from both changes in the international location of production and in international fuel prices.
- Some ex post studies suggest little evidence of leakage for EU climate policy; instead other factors (e.g., proximity to market, transport costs, quality of the local labor force, availability of raw materials) appear more important determinants of production location decisions.
- Ex post econometric work by Misch and Wingender (2021) estimates leakage from production migration using data on how changes in sectoral energy prices in different countries and over time affect the carbon embodied in trade flows.

### Methods and limitations highlighted
- Much of the literature relies on ex ante analyses using computable general equilibrium models that:
  - combine estimates of the impacts of carbon pricing on industrial production costs, and
  - make assumptions about the degree of substitution between goods produced in different countries.
- Ex post studies that find limited leakage often examine previous periods when the EU ETS price was relatively low and EITE industries were receiving free allowance allocations (conditional on remaining in the EU), limiting applicability to periods with higher carbon prices.
- As EU ETS conditions change (price and allowance allocation), the empirical evidence from earlier low-price periods may understate future leakage risks.

### Evidence on EU ETS dynamics and implications
- The EU ETS price jumped from $6 per ton in 2017 to over $70 per ton in 2021.
- The EU recently tightened its 2030 emission pledge from a 40 to a 55 percent reduction relative to 1990 levels.
- Implication: As EU ETS prices increase and allowance allocations become less effective at preserving EITE profitability, potential emissions leakage (in the absence of a BCA) would likely increase.

### Annex 4 — International Carbon Price Floor (ICPF): rationale and design elements
- Two main practical obstacles to scaling up global mitigation under the Paris Agreement:
  - Many parties (195), negotiating over many pledges (one per party), with pledges for 2030 difficult to compare.
  - Difficulty for countries to scale up mitigation unilaterally due to competitiveness concerns and fears of trading partners free-riding or reneging.
- An ICPF could complement and reinforce the Paris Agreement by addressing both obstacles via two key elements:
  - Focus on a small number of key emitting countries (candidates: China, India, and the United States; other participants might include the EU, UK, and some other G20 countries).
  - Focus on a minimum carbon price (price floor) as an efficient, easily understood parameter; simultaneous coordinated action on carbon pricing tackles competitiveness and free rider concerns.
- Design considerations for equity and flexibility:
  - Stricter requirements for higher income countries and/or simple and transparent (financial or technical) mechanisms to assist lower income countries.
  - Flexibility to accommodate differing national approaches (different combinations of pricing and sector-based fiscal and regulatory incentives) so long as they achieve the equivalent emissions outcome as meeting the price floor (as verified by third parties).
  - Exempting participants from a common BCA applied to all those outside the arrangement (except low income countries) could promote participation.

### Legal interaction and potential constraint
- Differentiation of a BCA based on country-of-origin of imports may violate GATT’s Most Favored Nation principle; reliance on an Article XX defense may then be necessary.

### Annex 5 — Compatibility of BCAs with Trade Law: key points
- WTO rules generally allow:
  - rebates for indirect taxes on products that are exported (not to exceed the domestic tax paid on like products consumed domestically), and
  - a charge on imported products not in excess of the indirect tax on like domestic products.
- Possible channels for compatibility:
  - BCAs with carbon taxes:
    - Charges on imports may be characterized under GATT Article II:2(a) as a ‘customs duty’ or ‘charge imposed on or in connection with importation’ allowing equivalence to domestic taxes, but:
      - The BCA must be imposed on a specific product or input to that product; taxing embodied carbon raises interpretive uncertainty (embodied carbon might be viewed as a by-product rather than an input).
      - Article III:2 prohibits exceeding the tax rate on ‘like’ domestic products, creating uncertainty about applying higher charges to imports with higher embodied carbon unless those imports are ‘unlike’.
    - Export rebates for carbon taxes might be allowable under the SCM Agreement footnote 1 (rebates of domestic indirect taxes should not be deemed export subsidies), but the rebate would have to be offered on the same terms to all domestic firms covered by the carbon tax; if ‘like’ products are interpreted by characteristics other than embodied carbon, the rebate could not exceed the lowest tax rate levied on domestic producers (the rate assessed on the cleanest producer).
  - BCAs with ETSs:
    - If a BCA requires importers to purchase allowances from a domestic ETS or separate allowance pool, it would likely be considered domestic regulation under GATT Article III:4, which requires imported products receive no less favorable regulatory treatment than like domestic products; viewing imports as ‘like’ domestic products but taxing based on imports’ carbon content could breach WTO rules.
    - A BCA on exports taking the form of a rebate for ETS costs could be a prohibited export subsidy if rebates were not available for like products sold domestically (WTO law contains no provision for border rebates of regulatory costs).

### Article I (MFN), Article XX exceptions, and design risks
- Article I prohibits discrimination among imports based on country of origin; differentiating imports using country-specific estimates of embodied carbon (vs. applying the same embodied carbon to all countries) could violate MFN if measures are viewed as arbitrary.
- Allowing relatively clean individual exporters to request lower BCAs might lower the risk that measurement procedures are viewed as arbitrary.
- Special treatment for some countries (e.g., those meeting ambitious Paris pledges) might also violate MFN absent an objective test applicable to all.
- Exemptions for least developed countries might be allowable under the WTO’s Enabling Clause if criteria are based on development indicators and treat countries in similar conditions the same way.
- Article XX (General Exceptions) could justify measures that breach other GATT obligations if they:
  - fall under sub-paragraph (a) to protect human, animal, or plant life or health, or
  - fall under sub-paragraph (g) relating to conservation of exhaustible natural resources (most analysts view (g) as easier to comply with than (a) due to the necessity test in (a)).
- The BCA must demonstrate it effectively addresses climate change (for example, by containing leakage) and must satisfy the chapeau of Article XX (not applied as arbitrary or unjustifiable discrimination or as a disguised restriction on international trade).
- Historically, very few measures have survived scrutiny under the chapeau, emphasizing the importance of WTO-compliant BCA design and treating Article XX as a fallback.

### Specific ways a BCA may fail Article XX
- A BCA may fail if it:
  - Requires specific policy changes as a basis for exemption, which could constitute arbitrary discrimination because measures tied to country-level policies punish all producers from targeted countries regardless of individual performance; instead the BCA should offset the differential between foreign and domestic carbon pricing.
  - Assesses adjustments based on country of origin rather than objective criteria applicable to all countries, which may include emissions-related policies or the environmental performance of individual producers.
  - Is implemented without having tried to negotiate in good faith toward a multilateral solution to carbon leakage (negotiations under the Paris process could be considered steps in this direction).
  - Allows exemptions from coverage of the BCA (e.g., for parties with ambitious Paris goals) or for specific domestic producers that are not based on the objective of mitigating climate change by preventing leakage.

### Interaction with other WTO agreements and potential constraints
- GATT Article XX does not cover breaches of other WTO Agreements like the SCM Agreement; a BCA that constitutes a prohibited export subsidy under SCM would have no recourse to Article XX.
- Under the SCM Agreement, a carbon tax would likely be an indirect tax; export adjustments would therefore likely be legal provided the adjustment amount is not more than the domestic carbon tax incurred.

*Annex 3, Annex 4, and Annex 5 content as provided in wpiea2021239-print-pdf - Annex 3. Empirical Literature on Emissions Leakage: A Quick Summary*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021239-print-pdf.pdf_