## wp18203 — EXECUTIVE SUMMARY

## Source details

**Canonical URL:** [wp18203 — EXECUTIVE SUMMARY](https://www.imf.org/-/media/files/publications/wp/2018/wp18203.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2018/wp18203.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2018/wp18203.pdf.json)

---

### Executive summary and purpose
- Advocates consideration of market-based mechanisms (MBMs), specifically an international maritime carbon tax or fuel levy (a tax/levy on shipping fuel in proportion to carbon content, referred to here as a carbon tax).
- Notes international maritime fuels are underpriced from an environmental perspective because there is no charge for greenhouse gas (GHG), particularly carbon dioxide (CO2), emissions; maritime CO2 emissions are significant and expected to expand steadily without policy action.
- References the International Maritime Organization (IMO) April 2018 pledge to cut emissions by 50 percent by 2050 relative to the 2008 level and frames the paper as promoting dialogue on carbon taxes as part of meeting that pledge.
- Paper structure: case for the tax, comparison with alternative instruments, design options, and quantitative estimates of impacts.

### Key policy design considerations (options and trade-offs)
- Responsibility for implementation:
  - Domestic collection (extending administrative capacity for domestic fuel taxes).
  - International collection from ship operators (based on required reporting of their fuel consumption) through establishment of an IMO-administered fund.
- Tax rates:
  - Economic models available to assess future emissions impacts of carbon taxes.
  - Practical implementation may be challenging for prices considerably higher than typical current pricing schemes (typically around $5-$30 per tonne of CO2 at present).
- Compensation for vulnerable countries:
  - Compensation mechanisms could reconcile the principle of common but differentiated responsibilities and respective capabilities (CBDRRC) with global application of the maritime carbon tax.
  - Burden of maritime carbon taxation is generally small relative to countries’ GDP, suggesting practical compensation options.
- Revenue use:
  - Allocation of potentially sizable revenues is contentious: options include national budgets, climate finance, or industry retention for clean-technology R&D/deployment.
  - Revenue-limiting designs: charge ship operators for the difference between their emissions and a benchmark level to preserve mitigation incentives while limiting total revenues.

### Environmental and economic rationale
- A carbon tax is favored because it:
  - Promotes the full range of mitigation opportunities across new and existing ships (technical and operational improvements, fleet composition shifts).
  - Strikes a cost-effective balance among mitigation responses by providing the same reward per tonne of CO2 reduced.
  - Provides price certainty and is simpler to administer and comply with relative to some pricing instruments (e.g., emissions trading systems, offset schemes).
- Behavioral responses induced by higher prices for carbon-based fuels include:
  1. Improvements in technical design efficiency of new vessels (lower empty weight, increased engine/propulsion efficiency, accommodation of lower carbon technologies like batteries, biofuels, liquefied natural gas (LNG), hydrogen).
  2. Improvements in operational efficiency for existing ships (optimizing vessel speeds, route lengths, port dwell time; maintenance and retrofitting engines, propellers, hulls).
  3. Shifts to larger (more fuel-efficient) ships and increased load factors.
  4. Shifting consumer demand away from heavy/long-distance products toward light/short-distance products and non-shipped goods and services.
- Carbon taxes can raise significant revenues; efficient clean-technology spending is likely a small fraction of potential carbon tax revenues.

### Quantitative illustrative results (selected modeling outcomes)
- Monetary figures in constant, 2016 US$.
- Illustrative pure carbon tax trajectory (starts 2021, rising $7.5 per tonne of CO2 each year):
  - Reaches US$75 per tonne of CO2 in 2030 (equivalent to $240 per tonne of bunker fuel), and $150 per tonne in 2040 ($480 per tonne of bunker fuel).
  - Reduces maritime CO2 emissions below business-as-usual (BAU) levels by nearly 15 percent in 2030 and 25 percent in 2040 (model tables: 14.07 percent in 2030; 23 percent in 2040 referenced elsewhere).
  - Raises revenues of about $75 billion in 2030 and $150 billion in 2040 (model results: $76 billion in 2030; $155 billion in 2040).
  - Increases shipping costs by 0.075 percent of global GDP in 2030.
- Revenue-neutral carbon tax with same emissions price:
  - Slightly less effective at reducing CO2.
  - Increases average shipping costs by 0.005 percent of global GDP in 2030.
- Performance standard for new ships (IMO EEDI) has only one-third of the effectiveness of carbon taxes for the same implicit CO2 price.

### Institutional and equity considerations
- Global application aligns with IMO principle of non-discriminatory treatment of all ships regardless of flag state.
- Member states emphasize CBDRRC; reconciliation options include compensation schemes, possibly limited to cases with significant burdens.
- Developing countries account for 63 percent of unloaded tonnage in 2015; full compensation for middle- and high-income developing countries is questioned.

### Comparative policy context and complementary instruments
- ICAO’s CORSIA for aviation relies on international emission offsets to stabilize aviation emissions at 2020 levels but faces uncertainty over credit prices, verification procedures, double counting, and phased implementation with exemptions.
- IMO implemented the Energy Efficiency Design Index (EEDI) tightened in phases: 10 percent reduction for ships built 2015–2020; 20 percent for 2020–2025; 30 percent after 2025.
- Carbon taxes differ from standards and offset schemes by promoting a broader set of mitigation responses and offering price certainty; complementary measures are necessary to develop and deploy alternative fuel technologies for deep mid-century reductions.

### Conclusions and recommendations
- Maritime carbon taxes are economically and administratively promising and merit serious attention at upcoming IMO deliberations as part of a comprehensive mitigation strategy.
- Consider different candidate tax designs, including revenue-limiting taxes that preserve mitigation incentives while constraining total revenues.
- Taxes should be accompanied by policies to develop and deploy alternative fuel technologies to achieve deep emissions reductions by mid-century.

---

### Box 2 — Revenue-Recycling and the Costs of Carbon Taxation (key points)
- Harberger triangle:
  - Fuel tax reduces fuel use and causes an economic welfare cost—the “Harberger triangle” (loss of consumer benefits less reductions in fuel production costs).
  - The Harberger triangle increases approximately in proportion to the square of the tax rate.
- Revenue recycling:
  - Tax raises revenue; a portion recycled can yield an economic efficiency benefit.
  - Numerical illustration: if the efficiency gain is "30 cents per dollar of revenue recycled", then by simple geometry the revenue recycling benefit is "5.4 times" and "2.4 times" the Harberger triangle for fuel reductions of "10" and "20 percent" respectively.
- Tax-interaction effect:
  - Higher product costs dampen overall economic activity and compound distortions on labor and capital.
  - Efficient revenue recycling is needed because the tax-interaction effect can be large relative to the Harberger triangle.
- Industry retention and revenue-limiting design:
  - Little basis for full retention of carbon tax revenues within the industry if most tax passes forward into higher fuel prices.
  - Design alternative: tax the difference between emissions and a benchmark level to avoid raising large revenues while preserving mitigation incentives.
- Instrument comparisons (summary):
  - ETS: auctioning can generate same revenue as a carbon tax; drawbacks include price volatility, depressed prices historically, market manipulation, transaction costs; Fell and others (2012) estimate short-term price volatility in a typical ETS raises costs by "around 15 percent" relative to a carbon tax for the same cumulative emissions reduction.
  - Carbon intensity standards: less effective unless applied broadly, limited to newbuilds unless expanded, do not automatically equalize marginal abatement costs; can be improved with flexibility (fees, rebates, trading).
  - Offsets: in theory cost-effective, but supply and credibility uncertainties are large.
  - Overall: a carbon tax is preferable if a robust and predictable price is established and (large) revenues are either used efficiently or the tax is applied on the difference between emissions and a benchmark to limit revenues.

### Design issues for carbon taxes — administration and formulas
- Collection options:
  - At refinery gate as extension of established fuel tax administration (small number of large, identifiable taxpayers).
  - International collection from ship operators through an IMO-administered fund to avoid national coordination difficulties.
- Measurement and enforcement:
  - Capacity for measuring shipping fuel use by trip is being developed.
  - Payment frequency: annual or individual route basis.
  - Enforcement: denial of port access or ship arrest for non-compliant operators.
- Tax liability formula (notation preserved as in source):
  - A ship operator’s tax liability would be given by:
    - 휏휏
      퐶퐶퐶퐶 2
      ∙퐹퐹
      푆푆푆푆푆푆푆푆
      ∙훽훽
      퐶퐶퐶퐶 2
  - Where:
    - 휏휏
      퐶퐶퐶퐶 2 is the tax rate on CO2 emissions;
    - 퐹퐹
      푆푆푆푆푆푆푆푆 is the ship’s fuel use; and
    - 훽훽
      퐶퐶퐶퐶 2 is the emissions factor for the fuel being used.
- Revenue-neutral (benchmark-based) variant:
  - Tax liability:
    - 휏휏
      퐶퐶퐶퐶 2
      ∙
      (
      퐹퐹
      푆푆푆푆푆푆푆푆
      ∙훽훽
      퐶퐶퐶퐶 2
      −퐵퐵퐵퐵퐵퐵퐵퐵퐵퐵
      푆푆푆푆푆푆푆푆
      )
  - 퐵퐵퐵퐵퐵퐵퐵퐵퐵퐵
    푆푆푆푆푆푆푆푆 is an exogenous benchmark level; setting benchmark equal to average ship emissions on an operator’s routes yields revenue neutrality.
  - Benchmark calculation:
    - 퐵퐵퐵퐵퐵퐵퐵퐵퐵퐵
      푆푆푆푆푆푆푆푆
      =푇푇푇푇
      푆푆푆푆푆푆푆푆
      ∙�
      퐶퐶퐶퐶 2
      푇푇푇푇
      �
      퐴퐴퐴퐴
- Incentive effects of revenue-neutral tax:
  - Preserves same incentives as pure carbon tax for many responses but:
    - Essentially fails to promote consumption-shift channel (response (4)) because there is no large pass-through into higher shipping costs.
    - Promotes shifts within vessel classifications (response (3)) as operators gain from reduced tax payments or rebates.

### CO2 price-setting considerations and illustrative price ranges
- Possible bases for tax rate trajectories:
  - Social cost of carbon (SCC) — difficult to agree across IMO members.
  - Global emissions price trajectories consistent with the 2oC target: a recent review suggests global CO2 prices of "$40-80 per tonne" in 2020.
  - Current global average price around "about $1 per tonne of CO2".
  - Modeling assessments: extremely high prices (perhaps "over $300 per tonne") might be needed without additional technology deployment policies.
- Pragmatic approach: aligning with prices in other carbon pricing schemes may be most practical.

### Addressing differentiated responsibilities (CBDRRC) and compensation options
- Options include remitting revenues to the GCF for allocation to adaptation and mitigation projects in targeted developing countries.
- Other approaches: reimbursing target countries for taxes on their maritime fuel sales or basing compensation on shares of global import values—each has distributional pitfalls.
- Practical note: workable compensation schemes should be feasible given the generally modest to tiny incidence of carbon taxation as measured in the text.

---

### Model assumptions, scenarios, and key baseline statistics
- Model features:
  - Distinguishes main shipping types (wet/dry bulk and container shipping) and behavioral responses.
  - Discrete time-period model used out to 2040; static fuel adjustment to price changes (long-run elasticities).
- Baseline (model starts) values:
  - 2016 maritime fuel use: "334 million tonnes".
  - Equivalent 2016 emissions: "1,051 million tonnes of CO2".
  - Allocation: "55 percent" of fuel use to bulk shipping and the rest to container shipping.
- Key parameters and assumptions:
  - Emissions factor for bunker fuels: 3.15 tonnes of CO2 per ton of bunker fuel.
  - Crude oil price assumed constant in real terms at $70 per barrel ($513 per tonne).
  - A (permanent) one-off price increase of $13.4 per barrel ($100 per tonne) from 2020 onwards reflecting low sulfur requirements.
  - Elasticities for technical design, operational, and other improvements: each -0.15 (combined elasticity, -0.45).
  - Autonomous decline in carbon intensity: 0.5 percent a year.
  - Income elasticities for shipped products: 0.5 for bulk and 0.8 for container products.
  - Global GDP assumed to expand "20 percent between 2017 and 2023" and grow at "2.9 percent a year thereafter".
  - World fleet: "about 50,000 ships in 2016".
  - About "700 oil refineries worldwide".
  - Initial fuel cost share in price per tonne-mile: about 5 percent on average (bulk 6 percent; container 2 percent).

### Policy scenarios modeled
- Pure (revenue-raising) carbon tax:
  - Starts 2021, rises $7.5 per tonne of CO2 each year to reach $75 per tonne of CO2 ($240 per tonne of fuel) by 2030 and $150 per tonne by 2040 ($480 per tonne fuel).
- Revenue-neutral carbon tax:
  - Causes same fuel price increase as pure tax but no first-order pass through of revenues into higher shipping costs.
- Carbon intensity standards (CIS) — three variants:
  - CIS—DES: technical design efficiency of new ships only.
  - CIS—DES/OP: technical design and operational efficiency for new and used ships.
  - CIS—DES/BULK: design efficiency for bulk ships only.
- Offsets:
  - Modeled with two supply scenarios:
    - ‘Low cost’: marginal cost of offsets approximately the same as that for reducing CO2 intensity of shipping.
    - ‘High-cost’: marginal cost of offsets three times as high.

### Model results — BAU (no new mitigation beyond recent observed fuel use)
- World GDP relative to 2020:
  - 2030: 33 percent higher.
  - 2040: 77 percent higher.
- Bunker fuel use or CO2 emissions relative to 2020 under BAU:
  - 2030: 14 percent higher.
  - 2040: 31 percent higher.
- Absolute CO2 emissions:
  - 2030: 1,172 million tonnes.
  - 2040: 1,343 million tonnes.
- CO2 to GDP falls by:
  - 2030: 14 percent below 2020 level.
  - 2040: 26 percent below 2020 level.

### Model results — Carbon taxes (pure and revenue-neutral)
- Pure carbon tax (rising at $7.5 per tonne from 2021):
  - CO2 emissions change relative to BAU:
    - 2030: reduces CO2 emissions by 14 percent below BAU levels.
    - 2040: reduces CO2 emissions by 23 below BAU in 2040.
  - Emissions roughly stabilize at just over 1,000 million tonnes in these years (approximately the BAU level in 2020).
  - Bunker fuel price increase above BAU:
    - 2030: about 40 percent above BAU.
    - 2040: about 75 percent above BAU.
  - Composition of CO2 reductions:
    - 96 percent from reductions in carbon intensity per tonne-mile (behavioral responses (1)-(3)).
    - 4 percent from reductions in tonne-miles (response (4)).
    - In 2030, half of CO2 reductions come from bulk shipping and half from container/other shipping.
- Revenue outcomes:
  - Pure carbon tax revenue (global average level):
    - 2030: raises 0.07 percent of world GDP, or $76 billion.
    - 2040: raises 0.11 percent of world GDP, or $155 billion.
  - Revenue-neutral carbon tax: no revenue implications by design (given no pre-existing fuel taxes).
- Economic welfare cost (triangle in Box 2; excludes climate benefits):
  - Both carbon tax policies:
    - 2030: roughly 0.006 percent of GDP, or $6.2 billion.
    - 2040: 0.016 percent of GDP, or $23.4 billion.
- Interpretation:
  - Revenue plus welfare cost indicates overall burden/incidence of the carbon tax at the global average level is relatively small compared with global GDP under the pure tax.

### Policy comparisons and sensitivity
- CIS—DES outcomes:
  - Reduces CO2 emissions by about 5 percent below BAU in 2030 and about 8 percent in 2040 (about one-third effectiveness of pure carbon tax).
- CIS—DES/BULK:
  - About 15 percent of the effectiveness of the pure carbon tax.
- CIS—DES/OP:
  - Twice as effective as CIS—DES.
- Offsets:
  - Similar within-sector impact as revenue-neutral carbon tax when priced equivalently, but extend reductions outside maritime sector:
    - Total emissions reductions are 28 percent higher (high-cost offset case) and 85 percent higher (low-cost offset case) than under carbon taxation.
  - Caveat: supply and additionality uncertainties.
- Sensitivity (2030 carbon tax impacts):
  - BAU emissions under crude oil prices between $35 and $105 per barrel are between 14 percent lower and 7 percent higher than when oil prices are $70 per barrel.
  - Percent reduction induced by the carbon tax:
    - 25 percent smaller under the higher oil price ($105 per barrel).
    - 50 percent greater under the lower oil price ($35 per barrel).
    - Increasing carbon intensity elasticities by 50 percent increases percent reduction by around 40-45 percent; decreasing elasticities by 50 percent decreases percent reduction by around 40-45 percent.
  - Revenues vary between $68 and $84 billion across different cases.
  - Welfare losses vary between $3.5 and $8.4 billion under different elasticity assumptions.
  - Policy cost examples: 0.075 percent of GDP in 2030 for non-revenue-neutral; under revenue-neutral variant, 0.005 percent of GDP.

### Implementation and practical recommendations
- Technical feasibility: reporting capacity for fuel use and emissions by ship trip is being developed.
- Administrative options: refinery-gate collection (few taxpayers) or IMO-administered fund (international collection).
- Enforcement: denial of port access or ship arrest for non-compliance.
- To increase likelihood of implementation:
  - Develop comprehensive strategy with clear objectives and use of revenues.
  - Address sensitivities of small island developing states and LICs via direct or indirect compensation mechanisms.
  - Phase in tax gradually to allow fleet mix adjustments.
  - Pair tax with policies to develop and deploy alternative fuel technologies for mid-century deep reductions.

---

*Source: wp18203 - EXECUTIVE SUMMARY and Box 2 (IMF working paper), content as provided.*

### EXECUTIVE SUMMARY .....................................................................................................3

### wp18203 - EXECUTIVE SUMMARY

### Executive summary and purpose
- Argues for considering market-based mechanisms (MBMs), specifically an international maritime carbon tax or fuel levy (a tax/levy on shipping fuel in proportion to carbon content, referred to here as a carbon tax).
- Notes that international maritime fuels are underpriced from an environmental perspective because there is no charge for greenhouse gas (GHG), particularly carbon dioxide (CO2), emissions; maritime CO2 emissions are significant and expected to expand steadily without policy action.
- References the International Maritime Organization (IMO) April 2018 pledge to cut emissions by 50 percent by 2050 relative to the 2008 level and frames the paper as promoting dialogue on carbon taxes as part of meeting that pledge.
- Describes the paper’s structure: case for the tax, comparison with alternative instruments, design options, and quantitative estimates of impacts.

### Key policy design considerations (options and trade-offs)
- Responsibility for implementation:
  - Carbon taxes could be collected domestically (extending administrative capacity for domestic fuel taxes).
  - A more immediately relevant option is international collection from ship operators (based on required reporting of their fuel consumption) through establishment of an IMO-administered fund.
- Tax rates:
  - Economic models are available to assess future emissions impacts of carbon taxes.
  - Practical implementation may be challenging for prices considerably higher than typical current pricing schemes (typically around $5-$30 per tonne of CO2 at present).
- Compensation for vulnerable countries:
  - Compensation mechanisms could help reconcile the principle of common but differentiated responsibilities and respective capabilities (CBDRRC) with global application of the maritime carbon tax.
  - Burden of maritime carbon taxation is generally small relative to countries’ GDP, suggesting practical compensation options.
- Revenue use:
  - Allocation of potentially sizable revenues is contentious.
  - Some view maritime taxes as a source of climate finance; others see them as funding for maritime technology and programs.
  - An option to permit more aggressive pricing is to limit revenues raised (while preserving mitigation incentives) by charging ship operators for the difference between their emissions and a benchmark level.

### Environmental and economic rationale
- A carbon tax is favored because it:
  - Promotes the full range of mitigation opportunities across new and existing ships (technical and operational improvements, fleet composition shifts).
  - Strikes a cost-effective balance among mitigation responses by providing the same reward per tonne of CO2 reduced.
  - Provides price certainty and is simpler to administer and comply with relative to some pricing instruments (e.g., emissions trading systems, offset schemes).
- Behavioral responses induced by higher prices for carbon-based fuels include:
  1. Improvements in technical design efficiency of new vessels (lower empty weight, increased engine/propulsion efficiency, accommodation of lower carbon technologies like batteries, biofuels, liquefied natural gas (LNG), hydrogen).
  2. Improvements in operational efficiency for existing ships (optimizing vessel speeds, route lengths, port dwell time; maintenance and retrofitting engines, propellers, hulls).
  3. Shifts to larger (more fuel-efficient) ships and increased load factors.
  4. Shifting consumer demand away from heavy/long-distance products toward light/short-distance products and non-shipped goods and services.
- Carbon taxes can raise significant revenues; revenue allocation options include national budgets (if collected domestically), climate finance (if collected internationally), or industry retention for clean technology R&D/deployment. The paper notes efficient clean-technology spending is likely a small fraction of potential carbon tax revenues.

### Quantitative illustrative results (selected modeling outcomes)
- All monetary figures are expressed in constant, 2016 US$.
- An illustrated carbon tax rising to US$75 per tonne of CO2 in 2030 ($240 per tonne of bunker fuel), and $150 per tonne in 2040:
  - Reduces maritime CO2 emissions below business-as-usual (BAU) levels by nearly 15 percent in 2030 and 25 percent in 2040.
  - Raises revenues of about $75 billion in 2030 and $150 billion in 2040.
  - Increases shipping costs by 0.075 percent of global GDP in 2030.
- A revenue-neutral carbon tax with the same emissions price (taxing operators with relatively high emissions intensity and subsidizing operators with relatively low emissions intensity):
  - Is only slightly less effective at reducing CO2.
  - Increases average shipping costs by 0.005 percent of global GDP in 2030.
- A performance standard for new ships (currently implemented by IMO) has only one-third of the effectiveness of carbon taxes for the same implicit CO2 price.

### Institutional and equity considerations
- Global application of a maritime carbon tax aligns with the IMO principle of non-discriminatory treatment of all ships regardless of flag state.
- Member states emphasize CBDRRC; reconciliation options include compensation schemes, possibly limited to cases where higher shipping costs impose significant burdens.
- Developing countries account for a large role in shipping (63 percent of unloaded tonnage in 2015), but full compensation for middle- and high-income developing countries is questioned.

### Comparative policy context and complementary instruments
- ICAO’s CORSIA for aviation relies on international emission offsets to stabilize aviation emissions at 2020 levels, but faces uncertainty over credit prices, verification procedures, double counting, and phased implementation with exemptions.
- IMO has implemented the Energy Efficiency Design Index (EEDI) for ship-specific carbon intensity standards, tightened in phases: 10 percent reduction for ships built 2015–2020; 20 percent for 2020–2025; 30 percent after 2025.
- Carbon taxes differ from standards and offset schemes by promoting a broader set of mitigation responses and offering price certainty; however, complementary measures to develop and deploy alternative fuel technologies are necessary for deep mid-century emissions reductions.

### Conclusions and recommendations
- Maritime carbon taxes are economically and administratively promising and merit serious attention at upcoming IMO deliberations as part of a comprehensive mitigation strategy.
- Different candidate tax designs should be considered, including revenue-limiting taxes that preserve mitigation incentives while constraining total revenues.
- Taxes should be accompanied by policies to develop and deploy alternative fuel technologies to achieve deep emissions reductions by mid-century.

*Source: wp18203 - EXECUTIVE SUMMARY (IMF working paper), content as provided.*

### Box 2. Revenue-Recycling and the Costs of Carbon Taxation

### Box 2. Revenue-Recycling and the Costs of Carbon Taxation

### Revenue recycling, Harberger triangle, and illustrative geometry
- Imposition of a fuel tax reduces fuel use and causes an economic welfare cost—the “Harberger triangle”—defined prior to netting out benefits from reduced future climate change.
- The Harberger triangle equals the loss of consumer benefits (the trapezoid area under the demand curve integrated over the fuel reduction) less reductions in fuel production costs (the area under the supply curve).
- The Harberger triangle increases approximately in proportion to the square of the tax rate.
- The tax also raises revenue (grey rectangles in the figure). A portion of that revenue can yield an economic efficiency benefit when recycled (darker grey rectangle), reflecting efficiency gains per dollar recycled (e.g., reductions in distortions from cutting taxes on labor income).
- Numerical illustration from the text:
  - If the efficiency gain is "30 cents per dollar of revenue recycled", then by simple geometry the revenue recycling benefit is "5.4 times" and "2.4 times" the Harberger triangle for fuel reductions of "10" and "20 percent" respectively.

### Tax-interaction effect and implications
- The ‘tax-interaction effect’ refers to the efficiency loss from higher product costs reducing overall economic activity, which compounds the dampening effect of taxes on labor and capital.
- Under plausible assumptions, up to a point the revenue-recycling benefit exceeds the tax-interaction effect, but the tax-interaction effect can be large relative to the Harberger triangle—hence the need for efficient revenue recycling.
- If the carbon tax is designed not to raise revenues (revenue-neutral variant discussed below), the tax-interaction effect is substantially reduced, as there is no pass through of (large) tax payments into higher product costs.

### Industry retention and revenue-limiting design
- There is little basis for full retention of carbon tax revenues within the industry on compensation grounds if most of the tax passes forward into higher fuel prices rather than squeezing shipping margins.
- A suggested way forward if diverting significant revenues from the industry is not initially viable:
  - Design the carbon tax to avoid raising large revenues in the first place by taxing the difference between emissions and a benchmark level.
  - This limits the broader burden of the tax on economic activity because of the weaker impact on shipping costs (less pass through of tax revenue into shipping costs).

### Other mitigation instruments (summary and comparative points)
- ETS (Emissions Trading System)
  - Requires operators to acquire allowances; total allowances (and emissions) are capped and trading establishes an allowance price.
  - Auctioning of allowances generates the same revenue as a carbon tax in principle.
  - An ETS can be made revenue-limiting by setting allowance requirements relative to baseline emissions.
  - Potential drawbacks:
    - Short-term price volatility can raise dynamic costs (empirical estimates suggest moderate importance).
    - Prices in ETSs to date have often been depressed.
    - Market manipulation and transaction costs might be issues given sector concentration (e.g., "eight companies account for about 60 percent of shipping capacity—UNCTAD 2016, Table 2.4") and many small ship owners with limited trading capacity.
  - Example estimate: Fell and others (2012) estimate short-term price volatility in a typical ETS raises costs by "around 15 percent" relative to a carbon tax for the same cumulative emissions reduction.
- Carbon intensity standards
  - As currently implemented by IMO for new ships, they are less effective than carbon taxes for the same implicit CO2 price.
  - Limitations:
    - Unless applied to existing ships and operational efficiency, standards are limited to newbuild ships and do not promote all response channels.
    - Do not provide automatic mechanism to equate incremental costs of CO2 reductions across operators, undermining cost effectiveness.
    - Non-pricing instruments are generally less effective at promoting clean technology investment than pricing instruments.
  - Potential improvements:
    - Apply standards to existing ships and account for both technical and operational efficiency.
    - Include flexibility provisions: out-of-compliance fees and rebates; allow trading of credits between operators.
  - Practical challenges:
    - Disaggregation by vessel type increases administrative complexity and may limit incentives like shifting to larger ships.
    - Difficulty obtaining a carbon intensity metric that is environmentally effective, compatible with available data, and not distortive.
- Offsets
  - Operators purchase credits for emission reduction projects outside the maritime sector for any emissions above a benchmark.
  - In theory can be cost-effective and promote outside-industry reductions.
  - In practice the supply and future price of offsets is highly uncertain (see Box 1).
- Summary conclusion on instruments:
  - Design details matter.
  - A carbon tax is preferable if a robust and predictable price is established and (large) revenues are either used efficiently or the tax is applied on the difference between emissions and a benchmark to limit revenues.
  - An ETS could be reasonable but needs price-stability mechanisms.
  - Carbon intensity standards can reinforce pricing instruments if designed flexibly and applied broadly.
  - Offsets are an alternative but face price and credibility uncertainties.

### Design issues for carbon taxes — administration and formulas
- Collection options:
  - Collect on shipping fuels at the refinery gate as extension of established fuel tax administration (small number of large, identifiable taxpayers).
  - Alternatively, collect at international level from ship operators through an IMO-administered fund (more relevant given IMO delegation of maritime mitigation strategy and to avoid national coordination difficulties).
- Measurement and enforcement:
  - Capacity for measuring shipping fuel use by trip is being developed.
  - Operators could pay tax on an annual or individual route basis.
  - Enforcement mechanisms might include denial of port access or ship arrest for non-compliant operators.
- Tax liability formula (as given in the source):
  - A ship operator’s tax liability would be given by:
    - 휏휏
      퐶퐶퐶퐶 2
      ∙퐹퐹
      푆푆푆푆푆푆푆푆
      ∙훽훽
      퐶퐶퐶퐶 2
  - Where:
    - 휏휏
      퐶퐶퐶퐶 2 is the tax rate on CO2 emissions;
    - 퐹퐹
      푆푆푆푆푆푆푆푆 is the ship’s fuel use; and
    - 훽훽
      퐶퐶퐶퐶 2 is the emissions factor for the fuel being used (well known; e.g., lower for LNG per unit of energy than for conventional heavy fuel oil sold as bunker fuel).
- Revenue-neutral variant (benchmark-based)
  - Tax liability would instead be:
    - 휏휏
      퐶퐶퐶퐶 2
      ∙
      (
      퐹퐹
      푆푆푆푆푆푆푆푆
      ∙훽훽
      퐶퐶퐶퐶 2
      −퐵퐵퐵퐵퐵퐵퐵퐵퐵퐵
      푆푆푆푆푆푆푆푆
      )
  - 퐵퐵퐵퐵퐵퐵퐵퐵퐵퐵
    푆푆푆푆푆푆푆푆 is an (exogenous) benchmark level of emissions assigned to the operator, so operators pay taxes or receive subsidies depending on whether their emissions are above or below their benchmark.
  - If the benchmark is set at the emissions that would have been generated on the operator’s routes by the average ship within a cargo classification then overall the scheme will be revenue neutral (payments from above-average emitters offset rebates to below-average emitters).
  - Benchmark calculation provided:
    - 퐵퐵퐵퐵퐵퐵퐵퐵퐵퐵
      푆푆푆푆푆푆푆푆
      =푇푇푇푇
      푆푆푆푆푆푆푆푆
      ∙�
      퐶퐶퐶퐶 2
      푇푇푇푇
      �
      퐴퐴퐴퐴
    - Where 푇푇푇푇
      푆푆푆푆푆푆푆푆 is tonne-miles for the individual operator and (퐶퐶퐶퐶 2/푇푇푇푇)퐴퐴퐴퐴 is emissions per tonne-mile for the average ship within the relevant classification.
  - Scaling back the benchmark (for all ships) yields positive net revenue; reducing the benchmark to zero converges the scheme to a pure revenue-raising carbon tax.
- Incentive effects of the revenue-neutral tax:
  - Provides the same incentives as a pure revenue-raising carbon tax (for a given CO2 price) for certain responses, but:
    - Essentially fails to promote the response channel where pass-through of tax into higher shipping costs shifts consumption patterns (response (4) in the text) because there is no large pass-through.
    - Promotes response (3) (e.g., shifting to larger, more efficient ships within a vessel classification) as operators benefit from reduced tax payments or increased rebates.

### CO2 price-setting considerations and illustrative price ranges
- Possible bases for CO2 tax rate trajectories:
  - The ‘social cost of carbon’ (SCC) — agreement across IMO member states would be challenging given widely differing SCC estimates in the literature.
  - Global emissions price trajectories consistent with the 2oC target in the Paris Agreement:
    - A recent review suggests global CO2 prices of "$40-80 per tonne" (in addition to any pre-existing fuel taxes) would be needed in 2020.
    - This is highly ambitious given the current global average price of "about $1 per tonne of CO2".
  - Estimated emissions prices countries will need to phase in by around 2030 to implement their Paris pledges — estimates are uncertain and vary considerably across countries.
  - Modelling assessments specific to the maritime sector:
    - Extremely high prices (perhaps "over $300 per tonne") might be needed in the absence of other technology deployment policies.
  - Prices in other carbon pricing schemes:
    - Political practicality may favor tax rates not considerably higher than prices elsewhere.
- Pragmatic approach:
  - On pragmatic grounds, aligning with prices in other carbon pricing schemes might be most practical; this approach is used to infer an illustrative price path in the modeling mentioned.

### Addressing differentiated responsibilities (CBDRRC) and compensation options
- One indirect solution: remit carbon tax revenues to the GCF, which would allocate funds for climate adaptation and mitigation projects in targeted developing countries; allocations might be skewed toward countries vulnerable to higher shipping costs for finer-tuned compensation.
- Alternative approaches to direct compensation:
  - Reimbursing target countries for taxes attributed to their maritime fuel sales—this may overcompensate hub refueling countries and undercompensate small island developing states.
  - Basing compensation on countries’ shares of global import values—import value is not necessarily a reliable predictor of CO2 incidence.
- Practical note: workable compensation schemes should be feasible given the generally modest to tiny incidence (as measured here) of carbon taxation.
  - Incidence measured in the text = loss of consumer surplus = first-order revenue payment plus the second-order economic welfare cost (combined gray rectangles and red triangle in Box 2 figure), assuming full pass through of fuel taxes into higher prices (no producer surplus losses).

### Quantitative policy analysis — model setup and key baseline statistics
- Model features:
  - Distinguishes main shipping types (wet/dry bulk and container shipping).
  - Distinguishes main behavioral responses for reducing emissions.
  - Discrete time-period model used out to 2040.
- Baseline (model starts) values:
  - 2016 maritime fuel use: "334 million tonnes"
  - Equivalent 2016 emissions: "1,051 million tonnes of CO2"
  - Allocation: "55 percent" of fuel use to bulk shipping and the rest to container shipping.
- Projection assumptions (BAU):
  - Fuel use projected forward in a BAU scenario using global GDP assumptions:
    - IMF forecasts global GDP to expand "20 percent between 2017 and 2023" (with further growth assumptions beyond that noted in the model).

### Administrative scope and compliance scale
- World fleet size and refineries (as provided):
  - World fleet comprised "about 50,000 ships in 2016" (UNCTAD 2016, Table 2.3) — majority are small ships that might be excluded initially from the tax.
  - About "700 oil refineries worldwide" (source indicated).
- Compliance costs:
  - International collection compliance costs might be somewhat larger than national collection due to greater number of taxpayers, but this is considered a fairly minor consideration relative to administrability.

*Source: wp18203 - Box 2. Revenue-Recycling and the Costs of Carbon Taxation*

### 2.9 percent a year thereafter) along with income elasticities

### wp18203 - 2.9 percent a year thereafter) along with income elasticities

### Model assumptions and parameters
- Emissions factor for bunker fuels: 3.15 tonnes of CO2 per ton of bunker fuel.
- Crude oil price assumed constant in real terms at $70 per barrel ($513 per tonne).
- A (permanent) one-off price increase of $13.4 per barrel ($100 per tonne) from 2020 onwards reflecting low sulfur requirements.
- Elasticities for technical design, operational, and other improvements due to higher fuel prices: each elasticity taken to be -0.15 (combined elasticity, -0.45).
- Autonomous decline in carbon intensity: 0.5 percent a year.
- Income elasticities for shipped products: 0.5 for bulk and 0.8 for container products.
- Other fuels are expressed in bunker fuel equivalents; CO2 emissions factor in the model is fixed.
- Model is static: fuel use adjusts instantly to fuel price changes; elasticities represent long-run responses allowing for significant fleet turnover.

### Policy scenarios considered
- Pure (revenue-raising) carbon tax:
  - Trajectory illustrated: starting in 2021 and rising at $7.5 per tonne of CO2 each year (equivalent to $24 per ton of bunker fuel) to reach $75 per tonne of CO2 ($240 per ton of fuel) by 2030 and $150 per tonne of CO2 ($480 per ton of fuel) by 2040.
  - Comparison context: prices in 2017 around $5-$20 per tonne of CO2 in ETSs and $5-$30 per tonne in carbon tax regimes; examples noted (no additional data invented).
- Revenue-neutral carbon tax:
  - Causes same fuel price increase as the pure carbon tax but no first-order pass through of tax revenues in higher shipping costs (no revenue raised in first order).
- Carbon intensity standards (CIS) — three variants implemented via shadow prices aligned with CO2 prices under the pure carbon tax:
  - CIS—DES: applies to technical design efficiency of new ships only.
  - CIS—DES/OP: applies to technical design efficiency and operational efficiency of new and used ships.
  - CIS—DES/BULK: promotes technical design efficiency improvements for bulk shipping only.
- Offsets:
  - Modeled to have equivalent effect on promoting carbon intensity reductions as the pure carbon tax, but with no pass through into shipping costs of charges for infra-marginal emissions under revenue-neutral treatment.
  - Two illustrative offset supply scenarios:
    - ‘Low cost’: marginal cost of offset reductions approximately the same as that for reducing CO2 intensity of shipping.
    - ‘High-cost’: marginal cost of offsets three times as high.
  - Note: lack of data for parameterizing future offset supply curve; additionality of offset projects may be difficult to establish.

### Caveats and modeling limitations
- Static model: instantaneous adjustment of fuel use to fuel price changes rather than gradual turnover; justified by focus on longer-term impacts and long-run elasticities.
- Most price-responsiveness of fuel use reflects reductions in carbon intensity rather than reductions in shipping volumes.
- Consequently:
  - Differences in environmental effectiveness between revenue-neutral carbon taxes/offset prices and a pure carbon tax are not very significant because infra-marginal emissions are not heavily charged and carbon intensity reductions dominate.
  - Omitting capital and labor costs of efficiency improvements when computing fuel use changes from mitigation policies is reasonable for this analysis.
  - Fiscal or market power distortions in the shipping market have limited effect on computed economic welfare effects because emissions reductions mainly come from carbon intensity changes.
- Model does not capture potential non-linear responses from sudden switching to a clean fuel alternative; such switching is considered a distant prospect likely requiring carbon prices far above those considered.

### Results — BAU (no mitigation beyond recent observed fuel use)
- World GDP relative to 2020:
  - 2030: 33 percent higher.
  - 2040: 77 percent higher.
- Bunker fuel use or CO2 emissions relative to 2020 under BAU:
  - 2030: 14 percent higher.
  - 2040: 31 percent higher.
- Absolute CO2 emissions:
  - 2030: 1,172 million tonnes.
  - 2040: 1,343 million tonnes.
- CO2 to GDP falls by:
  - 2030: 14 percent below 2020 level.
  - 2040: 26 percent below 2020 level.
- Drivers: below-unity income elasticities and improving energy efficiency.

### Results — Carbon taxes (pure and revenue-neutral)
- Pure carbon tax rising at $7.5 per tonne from 2021 onwards:
  - CO2 emissions change relative to BAU:
    - 2030: reduces CO2 emissions by 14 percent below BAU levels.
    - 2040: reduces CO2 emissions by 23 below BAU in 2040. 
  - Emissions roughly stabilize at just over 1,000 million tonnes in these years (approximately the BAU level in 2020).
  - Bunker fuel price increase above BAU levels:
    - 2030: about 40 percent above BAU.
    - 2040: about 75 percent above BAU.
  - Composition of CO2 reductions:
    - 96 percent of the CO2 reductions reflect reductions in carbon intensity per tonne-mile (behavioral responses (1)-(3)).
    - 4 percent reflects reductions in tonne-miles (response (4)).
    - In 2030, half of CO2 reductions come from bulk shipping and half from container/other shipping (bulk emissions share gradually declining under BAU).
- Revenue outcomes:
  - Pure carbon tax revenue (global average level):
    - 2030: raises 0.07 percent of world GDP, or $76 billion.
    - 2040: raises 0.11 percent of world GDP, or $155 billion.
  - Revenue-neutral carbon tax: no revenue implications by design (given no pre-existing fuel taxes).
- Economic welfare cost (triangle in Box 2; does not account for climate benefits):
  - Both carbon tax policies:
    - 2030: roughly 0.006 percent of GDP, or $6.2 billion.
    - 2040: 0.016 percent of GDP, or $23.4 billion.
- Interpretation:
  - Revenue plus welfare cost indicates overall burden/incidence of the carbon tax at the global average level is relatively small compared with global GDP under the pure tax.

*Source: wp18203 - 2.9 percent a year thereafter) along with income elasticities (IMF working paper content provided).*

### 0.075 percent of GDP in 2030, and under the revenue-neutral variant, a pretty tiny 0.005

### wp18203 - 0.075 percent of GDP in 2030, and under the revenue-neutral variant, a pretty tiny 0.005 percent of GDP.

### Policy comparisons
- CIS—DES policy outcomes:
  - Reduces CO2 emissions by about 5 percent below BAU levels in 2030.
  - Reduces CO2 emissions by about 8 percent below BAU levels in 2040.
  - Has about a third of the effectiveness of the (pure) carbon tax (as it only promotes one of the four behavioral responses).
- CIS—DES/BULK (limited to bulk ships):
  - Further reduces environmental effectiveness by about half relative to CIS—DES.
  - Implies this policy has about 15 percent of the effectiveness of the pure carbon tax.
- CIS—DES/OP (carbon intensity standard promoting technical and operational improvements across new and existing ships):
  - Is twice as effective as the CIS—DES policy.
- Offset policies:
  - Have about the same impact on reducing within-sector maritime emissions as the revenue-neutral carbon tax, given they are taken to establish the same emissions price.
  - Do not (to a first approximation) pass through a first-order tax payment into tonne-mileage prices.
  - Reduce CO2 emissions outside the maritime sector as well, implying total emissions reductions that are 28 percent higher (high-cost offset case) and 85 percent higher (low-cost offset case) than those under carbon taxation.
  - Caveat: "Whether offset schemes could establish the level of prices assumed here is highly questionable" and "most likely, not all offsets would be fully additional."

### Sensitivity analysis for carbon taxes (2030)
- BAU emissions sensitivity:
  - Under crude oil prices between $35 and $105 per barrel, BAU emissions are between 14 percent lower and 7 percent higher than when oil prices are $70 per barrel.
- Percent reduction in emissions below BAU induced by the carbon tax is sensitive to:
  - Oil prices:
    - 25 percent smaller under the higher oil price ($105 per barrel).
    - 50 percent greater under the lower oil price ($35 per barrel).
  - Carbon intensity elasticities:
    - Increasing elasticities by 50 percent increases the percent reduction in emissions by around 40-45 percent.
    - Decreasing elasticities by 50 percent decreases the percent reduction in emissions by around 40-45 percent.
- Revenues raised by the carbon tax:
  - Vary between $68 and $84 billion across the different cases in Table 3.
- Welfare losses:
  - Vary between $3.5 and $8.4 billion under the different elasticity assumptions.
- Numerical context from opening lines:
  - Policy cost examples cited: 0.075 percent of GDP in 2030 (non-revenue-neutral), and under the revenue-neutral variant, 0.005 percent of GDP.

### Conclusion and policy implications
- Rationale for serious scrutiny of a carbon tax for international maritime sector:
  - Can cost effectively exploit the full range of behavioral responses to reduce emissions within the sector, given available technologies.
  - Can be designed to raise significant revenues (if there is agreement on productive use of these revenues), or limit revenues (if dispute over revenue use would otherwise hold up introduction of an environmentally effective tax).
  - Is straightforward to implement from a technical perspective (given that capacity for reporting of fuel use and emissions by ship trip is being developed), through establishment of an IMO-supervised fund.
- Ingredients to increase likelihood of successful implementation:
  - Develop, in consultation with stakeholders, a comprehensive strategy with clear objectives (e.g., for future tax rates) and use of revenues.
  - Address sensitivities and concerns of small island developing states and LICs, potentially via direct or indirect compensation mechanisms.
  - Phase in the tax gradually to give shipping companies time to adjust (e.g., altering fleet mix) to minimize disruptions.
  - Recognize that the tax alone will not be sufficient: alternative fuel technologies will ultimately be needed to meet deep emissions reductions envisioned by mid-century.

### Analytical model — structure and key equations (overview)
- Aggregate fuel use:
  - F_t = sum_i F_ti, with F_ti = TTM_ti * f_ti = (TTM_ti / TTM_0i) * (f_ti / f_0i) * F_0i.
- Tonne-miles (TTM) drivers:
  - TTM_ti / TTM_0i = (GGDP_t / GGDP_0)^{υ_i} * (p_t_TTMi / p_0_TTMi)^{η_TTMi} * (p_t_TTM/ p_0_TTM)^{η_TTM_iT_j} ...
  - υ_i is income elasticity of demand (percent change in tonne-mileage per one percent change in GDP).
  - η_TTMi is own-price elasticity (< 0); cross-price elasticities η_TTM_C_B > 0 and η_TTM_B_C > 0.
- Effective bunker fuel price p_t_F consists of:
  - Average crude oil price p_t_Crude,
  - Additional low-sulfur cost δ_SC2 (fixed from 2020 onwards, zero earlier),
  - Carbon charge τ_t_CO2 · β_CO2 (emissions factor),
  - Offset price τ_t_OFFSET · β_CO2.
- Price per tonne-mile p_t_TTMi evolves with unit fuel costs and initial fuel cost share θ0_Fi (with revenue-neutral tax and offset policies producing downward adjustments to price per tonne-mile via benchmark consumption rates).
- Fuel consumption rate f_ti / f_0i varies with:
  - Autonomous technical design efficiency (DES), operational efficiency (OP), other factors (OTHER),
  - Autonomous annual improvement α_i,
  - Price elasticities η_DES_i, η_OP_i, η_OTHER_i (< 0),
  - Carbon intensity standard represented by shadow price λ_t_DES_i (reward per tonne CO2 reduced from technical design improvements), and λ_t_DES/OP_i for design+operational standard.
- Offsets supply:
  - τ_t_OFFSET = β_t_OFFSET * sum_i OFFSET_ti, with β_t_OFFSET slope constant; supply expands in proportion to global GDP.
  - Actual CO2 reductions from offsets may be less than purchased offsets due to non-additionality.
- Maritime emissions and revenues:
  - BUNKER_CO2_t_TA = β_CO2 · F_t.
  - BUNKER_CO2_t_NET = BUNKER_CO2_t_TA − sum_i OFFSET_ti.
  - Revenues R_BTR_t = τ_t_CO2 · β_CO2 · F_t.

### Data, parameterization, and baseline assumptions
- Base-year and shares:
  - International maritime fuel use for 2016: 334 million tonnes.
  - Bulk and container shipping account for 55 and 45 percent of fuel use respectively.
- GDP and elasticities:
  - (Real) global GDP assumed to expand 20 percent between 2017 and 2023 and grow at 2.9 percent a year thereafter.
  - Income elasticities: 0.8 for container; 0.5 for bulk; average about 0.65.
  - Tonne-mile price elasticity for both shipping types: -0.7.
- Oil and fuel prices:
  - Global crude oil price taken as $513 per tonne ($70 per barrel) from 2018, assumed to stay at this level to 2040.
  - Assumed one-off permanent price increase of $100 per tonne ($13.6 per barrel) from 2020 onwards to reflect the 0.5 percent sulfur requirement.
- Fuel cost shares in product price:
  - Initial share of maritime fuel costs in the price per tonne-mile: about 5 percent on average.
  - Bulk shipping fuel cost share: 6 percent.
  - Container shipping fuel cost share: 2 percent.
- Emissions and factors:
  - Average emissions factor for bunker fuel: 3.15 tonnes CO2 per tonne of bunker.
  - CO2 emissions taken as 938 million tonnes in 2012, implying fuel use of 297 million tonnes in 2012.
  - 2016 fuel use and emissions assumed to be 12 percent higher than 2012 levels: 334 million tonnes and 1,051 million tonnes respectively.
- Autonomous and price-response parameters:
  - Autonomous efficiency improvements reduce fuel consumption rates by 0.5 percent a year.
  - Elasticity of bunker fuel consumption rate w.r.t. fuel price: -0.45 (aggregate).
  - Individual elasticities for technical design efficiency, operational efficiency, and other factors: each -0.15.
- Offsets scenarios:
  - Low-cost scenario: marginal cost of offset reductions approximately the same as that for reducing the CO2 intensity of shipping.
  - High-cost scenario: marginal cost of offsets three times as high.

### Mitigation policy modeling notes
- Carbon tax rate schedule:
  - Applies to both carbon tax variants, the shadow price for CES policies, and the emissions offset price.
  - For revenue-neutral carbon tax, benchmark consumption rate f̅_t_CO2i = f_ti (no first-order impact on shipping prices per tonne-mile).
  - For offset policy, benchmark f̅_t_OFFSETi = f_ti − OFFSET_ti · β_CO2^{-1} (offset payments expressed relative to fuel costs).
- Impacts calculation:
  - Policy impacts on CO2 emissions and revenue calculated from equations (7) and (8).
  - Economic welfare costs calculated using standard formulas; broader fiscal linkages are not considered.

*Source: IMF Working Paper wp18203 (excerpts provided).*

### REFERENCES

### REFERENCES

### Key literature cited
- Acemoglu, Daron; Philippe Aghion; Leonardo Bursztyn; David Hemous, 2012. “The Environment and Directed Technical Change.” American Economics Review 102: 131–166.
- AGF (2010). Report of the Secretary-General’s High-level Advisory Group on Climate Change Financing. New York, United Nations.
- Aldy, Joseph, et al., 2016. “Economic Tools to Promote Transparency and Comparability in the Paris Agreement.” Nature Climate Change 6: 1,000–1,004.
- Burtraw, Dallas; Art Fraas; Nathan Richardson, 2012. “Tradable Standards for Clean Air Act Carbon Policy.” Discussion paper 12-05, Resources for the Future, Washington, DC.
- Clements, Benedict; David Coady; Stefania Fabrizio; Sanjeev Gupta; Trevor Serge Coleridge Alleyne; Carlo A Sdralevich, eds. 2013. Energy Subsidy Reform: Lessons and Implications. Washington DC, International Monetary Fund.
- Dechezleprête, Antoine and David Popp, 2017. “Fiscal and Regulatory Instruments for Clean Technology Development.” In I. Parry, K. Pittel and H. Vollebergh (eds.), Energy Tax and Regulatory Policy in Europe: Reform Priorities. MIT Press, Cambridge, MA.
- Farid, Mai; Michael Keen; Michael Papaioannou; Catherine Pattillo; Anna Ter-Martirosyan, 2016. After Paris: Fiscal, Macroeconomic, and Financial Implications of Climate Change. Staff Discussion Note 16/01, IMF, Washington, DC.
- Goulder, Lawrence H.; Ian W.H. Parry, 2008. “Instrument Choice in Environmental Policy.” Review of Environmental Economics and Policy 2: 152–74.
- Harberger, Arnold C., 1964. “The Measurement of Waste.” American Economic Review 54: 58-76.
- Nordhaus, William D., 2014. “Estimates of the Social Cost of Carbon: Concepts and Results from the DICE-2013R Model and Alternative Approaches.” Journal of the Association of Environmental and Resource Economists 1: 273-312.
- Parry, Ian W.H., 2003. “Fiscal Interactions and the Case for Carbon Taxes Over Grandfathered Carbon Permits.” Oxford Review of Economic Policy 19: 385-399.
- Pindyck, Robert S., 2016. “The Social Cost of Carbon Revisited.” National Bureau of Economic Research Working Paper No. 22807, Cambridge, MA.
- Stern, Nicholas, 2013a. “The Structure of Economic Modeling of the Potential Impacts of Climate Change: Grafting Gross Underestimation of Risk onto already narrow Science Models.” Journal of Economic Literature 51: 838-859.
- Weitzman, Martin L., 2011. “Fat-Tailed Uncertainty and the Economics of Climate Change.” Review of Environmental Economics and Policy 5: 275–292.
- WBG, 2017. State and Trends of Carbon Pricing 2017. World Bank Group, Washington, DC.

### Empirical and modelling studies on international transport, shipping, and freight emissions
- Calleya, J.; Pawling, R.; Greig, A., 2015. “Ship Impact Model for Technical Assessment and Selection of Carbon Dioxide Reducing Technologies (CRTs).” Journal of Ocean Engineering 97, 82-89.
- Cristea, A.; D. Hummels; L. Puzzello; M. Avetisyan, 2013. “Trade and the Greenhouse Gas Emissions from International Freight Transport.” Journal of Environmental Economics and Management 65: 153-173.
- Eide, M.; Dalsøren S.; Endresen Ø.; Samset B.; Myhre G.; Fuglestvedt J.; Berntsen T., 2013. “Reducing CO2 from shipping—do non-CO2 effects matter?” Atmospheric Chemistry and Physics 13: 4,183–4,201.
- IMO, 2014. Third IMO Greenhouse Gas Study. International Maritime Organisation, London.
- Smith, T.; Raucci, C.; Haji Hosseinloo S.; Rojon I.; Calleya J.; Suárez de la Fuente S.; Wu P.; Palmer K., 2016. CO2 emissions from international shipping. Possible reduction targets and their associated pathways. Prepared by UMAS, London.
- ICS, 2017. ICS Commits Shipping to Ambitious CO2 Reduction Objectives. International Chamber of Shipping, London.

### Policy design, carbon pricing, and instruments
- Fell, Harrison; Ian A. MacKenzie; William A. Pizer, 2012. “Prices versus Quantities versus Bankable Quantities.” Resource and Energy Economics 34: 607–23.
- Fischer, Carolyn; Ian W.H. Parry; William A. Pizer, 2003. “Instrument Choice for Environmental Protection when Technological Innovation is Endogenous.” Journal of Environmental Economics and Management 45: 523-545.
- Gately, D.; Huntington, H.G., 2001. The Asymmetric Effects of Changes in Price and Income on Energy and Oil Demand. C.V. Starr Center for Applied Economics, New York University.
- Green, Jessica F., 2014. “Don’t Link Carbon Markets.” Nature 543, March, 484-486.
- Jaffe, Adam B.; Richard G. Newell; Robert N. Stavins, 2005. “A Tale of Two Market Failures: Technology and Environmental Policy.” Ecological Economics 54: 164-174.
- Keen, Michael; Ian W.H. Parry; Jon Strand, 2013. “Ships, Planes, and Taxes: Charging for International Aviation and Maritime Emissions.” Economic Policy 28: 701–749.
- Lamotte, Henri, 2011. “Carbon Pricing in International Air and Maritime Transportation.” Ministry of Economy, and Finance and Industry, Government of France.
- Newell, Richard G., 2015. “The Role of Energy Technology Policy Alongside Carbon Pricing.” In I. Parry, A. Morris and R. Williams (eds.), Implementing a US Carbon Tax: Challenges and Debates, Routledge, London, 2015.
- Stochniol, Andre, 2011. “Securing additional climate financing, beyond budgetary contributions, by Ensuring No Net Incidence on Developing Countries from Carbon Pricing of International Transport (Rebate Mechanism), Outline and RM keys.”

### Databases, institutional reports, and technical notes
- IEA, 2017a. World Energy Outlook 2017. International Energy Agency, Paris, France.
- IEA, 2017b. World Energy Balances. International Energy Agency, Paris, France.
- IMF-WBG, 2011a. Market-based Instruments for International Aviation and Shipping as a Source of Climate Finance. International Monetary Fund and World Bank Group.
- IMF-WBG, 2017b. World Economic Outlook Database. International Monetary Fund, Washington, DC.
- ITF, 2017. ITF Transport Outlook 2017. International Transport Forum, OECD, Paris.
- OECD, 2016. 2020 projections of Climate Finance towards the $100 billion goal: Technical Note. Organisation for Economic Cooperation and Development, Paris, France.
- UNCTAD, 2016. Review of Maritime Transport 2016. United Nations Conference on Trade and Development.
- US IAWG, 2016. Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866. Interagency Working Group on Social Cost of Carbon, United States Government, Washington, DC.

### Key figures and numeric results reproduced in the unit
- Figure 1. BAU GDP, Energy Efficiency, and Fuel Trends, 2020=100
  - Timeline shown: 2020, 2024, 2028, 2032, 2036, 2040.
  - Series plotted: GDP, Energy efficiency, Fuel use or CO2, CO2 to GDP.
  - Fuel/CO2 tonnage markers: 1,024 mn tons; 1,172 mn tons; 1,343 mn tons.
- Figure 2. Impacts of Carbon Taxes
  - Panel (a) CO2 Reductions, percent below BAU: plotted across 2020–2040; two instruments compared: Carbon tax and Revenue neutral carbon tax.
  - Panel (b) Revenue, percent GDP: range plotted 0.00 to 0.12 percent GDP; monetary markers $76bn and $155bn.
  - Panel (c) Welfare Cost, percent GDP: range plotted 0.000 to 0.018 percent GDP; monetary markers $6.2 bn and $23.4 bn.
  - Note: Welfare calculations abstract from linkages with the broader fiscal system (see Box 2).
  - CO2 quantities noted: 1,031 mn tons and 1,008 mn tons.
- Figure 3. CO2 Reductions Under Alternative Mitigation Instruments
  - Percent reduction axis spans 0 to 45 percent below BAU levels.
  - Instruments compared include: Carbon tax; Revenue neutral carbon tax; CIS--DES; CIS--DES/BU LK; CIS--DES/OP; offset--low cost; offset--high cost.
  - Series reported for 2030 industry reductions, 2030 offsets, 2040 industry reductions, 2040 offsets.
- Table 1. Mitigation Pledges for the 2015 Paris Agreement, G20 Countries (selected entries with exact figures)
  - Argentina: Mitigation pledge: Reduce GHGs 30% below BAU in 2030; 2014 share of global CO2: 0.60; CO2/$1,000 GDP: 0.39; tons CO2 per capita: 4.7.
  - Australia: Reduce GHGs 26-28% below 2005 by 2030; 2014 share of global CO2: 1.00; CO2/$1,000 GDP: 0.25; tons CO2 per capita: 15.4.
  - Brazil: Reduce GHGs 37% below 2005 by 2025; 2014 share of global CO2: 1.50; CO2/$1,000 GDP: 0.22; tons CO2 per capita: 2.6.
  - Canada: Reduce GHGs 30% below 2005 by 2030; 2014 share of global CO2: 1.50; CO2/$1,000 GDP: 0.30; tons CO2 per capita: 15.1.
  - China: Reduce CO2/GDP 60-65% below 2005 by 2030; 2014 share of global CO2: 28.50; CO2/$1,000 GDP: 0.98; tons CO2 per capita: 7.5.
  - France: Reduce GHGs 40% below 1990 by 2030; 2014 share of global CO2: 0.80; CO2/$1,000 GDP: 0.11; tons CO2 per capita: 4.6.
  - Germany: Reduce GHGs 40% below 1990 by 2030; 2014 share of global CO2: 2.00; CO2/$1,000 GDP: 0.19; tons CO2 per capita: 8.9.
  - India: Reduce GHG/GDP 33-35% below 2005 by 2030; 2014 share of global CO2: 6.2; CO2/$1,000 GDP: 1.10; tons CO2 per capita: 1.7.
  - Indonesia: Reduce GHGs 29% below BAU in 2030; 2014 share of global CO2: 1.30; CO2/$1,000 GDP: 0.52; tons CO2 per capita: 1.8.
  - Japan: Reduce GHGs 25% below 2005 by 2030; 2014 share of global CO2: 3.40; CO2/$1,000 GDP: 0.25; tons CO2 per capita: 9.5.
  - Russia: Reduce GHGs 25-30% below 1990 by 2030; 2014 share of global CO2: 4.70; CO2/$1,000 GDP: 0.83; tons CO2 per capita: 11.9.
  - Saudi Arabia: Reduce GHGs 130 million tons below BAU by 2030; 2014 share of global CO2: 1.70; CO2/$1,000 GDP: 0.79; tons CO2 per capita: 19.5.
  - South Africa: Reduce GHGs 398-614 million tons in 2025 and 2030; 2014 share of global CO2: 1.4; CO2/$1,000 GDP: 1.40; tons CO2 per capita: 9.0.
  - Turkey: Reduce GHGs up to 21% below BAU by 2030; 2014 share of global CO2: 1.00; CO2/$1,000 GDP: 0.37; tons CO2 per capita: 4.5.
  - UK: Reduce GHGs 40% below 1990 by 2030; 2014 share of global CO2: 1.20; CO2/$1,000 GDP: 0.14; tons CO2 per capita: 6.5.
  - US: Reduce GHGs 26-28% below 2005 by 2025; 2014 share of global CO2: 14.50; CO2/$1,000 GDP: 0.30; tons CO2 per capita: 16.5.
  - Note: BAU denotes business as usual with no new mitigation measures. Some developing countries specify both conditional and unconditional pledges—in these cases the conditional pledges are included above.
- Table 2. Carbon Prices, Selected Countries and Regions, 2017 (selected entries and exact labels)
  - Sources: WBG (2017) and previous editions of this publication, and authors calculations (for Colombia and Canada).
  - Notes include: a Slated price for 2022 (in 2017$).
  - Examples of entries (as labeled in table): Sweden 1991 140 42; Colombia 2017 54 0; Switzerland 2008 87 33; Chile 2014 5 5; Mexico 2014 1-34 6; California 2012 15 85; EU 2005 64 5; Japan 2012 36 6; Korea 2015 18 68; Canada 2016 40 a 80.
  - Categories shown: PRICE FLOORS; CARBON TAXES; TRADING SYSTEMS.
- Table 3. Sensitivity of (Pure) Carbon Tax Impacts to Alternative Parameters, 2030 (selected exact figures)
  - Central case:
    - BAU CO2 emissions, mn tonnes: 1,172
    - CO2 emissions under carbon tax, mn tonnes: 1,008
    - % CO2 reduction below BAU: 14.07
    - revenue, $billion: 66.2
    - welfare cost, $billion: (blank in table extract)
  - Annual GDP growth rate beyond 2023:
    - 3.5 percent: BAU CO2 emissions 1,208; CO2 emissions 1,039; % reduction 14.07; revenue $86.3; welfare cost $ (not shown)
    - 2.3 percent: BAU CO2 emissions 1,141; CO2 emissions 981; % reduction 14.07; revenue $46.0
  - Income elasticities:
    - increased 33 percent: BAU 1,277; CO2 1,099; % reduction 14.08; revenue $26.7
    - decreased 33 percent: BAU 1,075; CO2 924; % reduction 14.06; revenue $95.6
  - Annual rate of autonomous carbon intensity reduction:
    - increased to 0.75 percent: BAU 1,132; CO2 974; % reduction 14.07; revenue $35.9
    - decreased to 0.25 percent: BAU 1,212; CO2 1,042; % reduction 14.07; revenue $86.4
  - 2030 crude oil price (exc. low sulfur costs):
    - $105 per barrel: BAU 1,202; CO2 1,076; % reduction 10.58; revenue $14.7
    - $35 per barrel: BAU 1,093; CO2 863; % reduction 21.06; revenue $58.6
  - Carbon intensity elasticities:
    - increased 50 percent: BAU 1,127; CO2 902; % reduction 19.96; revenue $88.4
    - decreased 50 percent: BAU 1,218; CO2 1,125; % reduction 7.78; revenue $43.5

*Source: REFERENCES (wp18203 - REFERENCES).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18203.pdf_