## 1. Summary of Instruments to Address Environmental Risks of Shale Gas

## Source details

**Canonical URL:** [1. Summary of Instruments to Address Environmental Risks of Shale Gas](https://www.imf.org/-/media/files/publications/wp/2017/wp17254.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2017/wp17254.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2017/wp17254.pdf.json)

---

### Introduction: questions addressed and high-level conclusions
- Central questions:
  - Do environmental issues associated with shale gas extraction warrant corrective taxes in fiscal regimes for upstream production activity, or are they better addressed through other instruments?
  - Do commercial and technical features of shale gas development suggest that principles of fiscal regime design for conventional oil and gas development need modification?
- High-level conclusions:
  - Many upstream environmental risks (e.g., water contamination) are generally best dealt with through regulation, though ex post liability for individual firms and industry-level compensation funds collected through production fees can play an important role.
  - Novel tax schemes for methane leakages are feasible despite limited metering capabilities.
  - Levies on downstream CO2 emissions from gas combustion on fuel supply are administratively appealing (at extraction, processing, or distribution).
  - Research on the nature and magnitude of environmental risks from shale gas is needed to guide policy design.
  - Simulations suggest special fiscal regimes and incentives for shale gas have, at best, only slight effects on investment while producing significant foregone government revenues.
  - For the most part, there is little evident need to consider specific incentives for shale gas versus conventional gas, provided the overall fiscal regime responds substantially to realized profit or rent.

### Brief overview of the shale gas industry (key facts and drivers)
- Technology and technical features:
  - Shale exploitation requires hydraulic fracturing, seismic imaging, and horizontal drilling.
  - Shale development typically uses these technologies more intensively than conventional exploitation, with more rapid well productivity decline and a greater number of wells per unit of gas.
  - Capital intensity, operating costs, water consumption, and chemical use are generally higher in shale exploitation than in conventional land-based drilling.
- US production and reserves:
  - Production ramped up from 300 billion of cubic feet (Bcf) to 15,213 Bcf by 2015.
  - Shale accounted for about 40 percent of US natural gas production by 2015.
  - Proven shale reserves stand at 175,000 Bcf, or 31 percent of total US gas reserves.
- Global resources and other countries:
  - Technically recoverable reserves totaling roughly 7.3 million Bcf have been identified across the globe, equal to about one third of total world gas resources.
  - Canada produced 765 Bcf in 2013; other countries exploring shale include Argentina, Australia, China, Mexico, Poland, and the UK.
- Factors that favored US development:
  - private ownership of subsurface mineral rights (in other countries these rights are state-owned or may be ill defined, creating legal challenges to rapid development);
  - an abundance of independent gas companies with the requisite technical knowledge or the incentive to develop it;
  - pre-existing gas pipeline infrastructure and capacity;
  - fairly abundant water resources;
  - low population density (and hence public opposition);
  - deep capital markets; and
  - (not least) high natural gas prices in the early 2000s.
- Policy responses in some jurisdictions:
  - Bans or restrictions on fracking apply at national level in Bulgaria, France, Germany, and South Africa and at subnational level in Australia, Ireland, Italy, Spain, Switzerland, and the United States (including New York, Vermont, and cities and counties in California, Colorado, Texas, and Ohio).

### Environmental risks: differences with conventional gas and specific issues
- Differences with conventional gas (environmental perspective):
  - Different pad footprint, water use, chemical use, and pressure below ground.
  - Unconventional development permits multiple wells on a single well pad — could lead to a lower footprint per unit of gas extracted depending on geological features.
  - Fracking uses a mixture of water, sand and a variety of chemicals at high pressure; conventional drilling uses far less water and additives.
  - Greater concentration of wells and high pressures might increase risks of groundwater contamination from casing and cementing failures.
  - Return of fracking fluid up the well raises concerns for liquid wastes handling, treatment, and disposal.
  - Pumping liquid wastes into deep disposal wells might affect seismic activity risks.
  - Use of powerful diesel engines for fracturing contributes to local air pollution.
  - Higher water demands imply more infrastructure (pipelines, roads, water storage) and greater potential for habitat disruption, community impacts, roadway wear and tear, and damages to land and waterways.
  - There appear to be no environmental risks posed by conventional gas-only wells that are not also posed by shale gas wells.
- Water quantity:
  - Shale gas development requires large quantities of water depending on maturity of the shale and formation thickness, and technologies used.
  - Water use per well: "Around 3-5 million gallons for development of a well (Mielke and others 2010)."
  - Per unit of energy, shale gas uses a lot more water than conventional gas, similar amounts to coal, and far less than enhanced oil recovery (Kuwayama and others 2013).
  - Water consumption for shale gas is highly concentrated at a point in time (occurring over several days during fracking).
  - Opportunity costs of water use depend on usage relative to local water availability; shale activity can overlap areas of high water stress in the United States.
  - Water demands for energy extraction are generally trivial compared with agriculture in many regions.
- Water quality:
  - Liquid and solid wastes from fracking or naturally occurring chemicals released by fracking can affect surface and groundwater quality (Burton and others 2013).
  - Liquid wastes can contain fracking fluids and highly salted water, radioactive materials, heavy metals and volatile organics from the formation itself, called “produced water.”
  - Surface water can be affected by runoff, spills, and discharge from treatment plants (e.g., Olmstead and others 2013); impacts are highly site-specific.
  - Groundwater can be affected through leakages due to faulty casing and cementing around the well bore (a standard oil and gas drilling problem).
  - Evidence on whether fracking itself can pollute groundwater by methane or substances in produced and flowback water is mixed (e.g., Osborn and others 2011, Schon 2011).
  - Current data gaps and uncertainties prevent a broad assessment of frequency and severity of fracking harms to water quality (US EPA 2016a).
- Climate change impacts:
  - Downstream combustion of natural gas (conventional and unconventional) is responsible for 27 percent of US fossil fuel CO2 emissions.
  - Combustion of natural gas produces about 40 percent less CO2 emissions than coal per unit of energy.
  - Net effect of shale gas on CO2 is modestly negative for the United States where gas displaces coal (Newell and Raimi (2104) and Brown and others (2009)).
  - Leakage or venting of natural gas (mostly methane) can occur across the lifecycle (drilling, production, processing, distribution, storage).
  - Methane is about 25 times more potent in trapping heat than CO2.
  - Leakage rates are uncertain; a typical estimate is "around 1.5 to" (text ends mid-phrase in supplied content).
- Local air pollution:
  - Sources and health links:
    - Combustion of diesel in engines used to pump fracking fluids and leaks of volatile organic compounds from storage and gas processing facilities contribute to local air pollution.
    - Ambient air pollution is associated with mortality risks (e.g., from heart and lung diseases).
  - Illustrative damage estimate:
    - If average population exposure to emissions released in shale extraction is one-tenth of that for emissions from the average road vehicle (and emission rates are comparable), pollution damages for the United States might be around 3 cents per liter (12 cents per gallon).
  - Policy implication:
    - Diesel emissions could be targeted via fuel taxes, emission-rate standards, or a combination to incentivize emissions control technologies and fuel-efficiency improvements.
- Ecological and seismic risks:
  - Land conversions for shale gas infrastructure potentially disrupt ecosystems, particularly in areas of high well density, though documentation is limited.
  - Fracturing per se produces quake magnitudes that are small to trivial, but wastewater deep-injection disposal has contributed to earthquake risk and has caused shutdowns of several disposal wells.
  - Areas with endangered or highly valued habitat, significant populations exposed to pre-existing seismic risks, or lacking viable means to dispose of liquid wastes, are probably not good candidates for shale gas development.
- Community and broader impacts:
  - Local communities benefit economically from shale gas development but also bear environmental, health, and broader impacts like noise, visual degradation, traffic congestion and accidents, property value changes, and overloaded schools.
  - Quantified local impacts (examples cited):
    - Extra truck traffic estimated to result in 9 extra road fatalities and 12 extra (serious but non-fatal) injuries in Pennsylvania counties with shale gas development.
    - Ground-water dependent property values fall by 16 percent as they become closer to shale wells, while values for properties with piped water access rise by 10 percent.
  - Indirect vs direct impacts:
    - Distinction for policy: direct environmental impacts (require corrective policies; e.g., health effects) vs indirect/pecuniary effects (largely transfers among regions/firms/individuals; may warrant equity-based compensation but not efficiency correction).

### Policy instruments and fiscal considerations
- Preferred instruments for upstream environmental risks:
  - Regulation as primary tool for many upstream risks, e.g., water contamination prevention and waste management.
  - Ex post liability for individual firms; for large risks, compensation via industry-level funds collected through production fees.
- Carbon and methane pricing options:
  - Novel tax schemes for methane emissions from production are feasible despite limited metering capabilities.
  - Levies on downstream CO2 emissions from gas combustion should be applied on fuel supply (administratively appealing at extraction, processing, or distribution).
  - If continuous monitoring of methane leakage were possible, an emissions tax would be ideal; interim option: tax fuel suppliers based on a default leakage rate with rebates for firms demonstrating lower leakage via mitigation and continuous monitoring systems.
  - For both CO2 and methane, ideally tax levels would be set in line with countries' emissions mitigation pledges for the 2015 Paris Agreement.
- Fiscal regime design and simulation findings:
  - Simulations suggest special fiscal regimes and incentives for shale gas have only slight effects on investment but cause significant foregone government revenues.
  - Little evident need to design specific incentives for shale gas relative to conventional gas if the fiscal regime responds to realized profit or rent.
  - Principal exception: jurisdictions that rely heavily on gross royalties or on initial or continuing flat fees or bonuses may need to consider special treatment to avoid discouraging shale development.

### Role of fiscal policy: instruments and design trade-offs
- Controlling development:
  - A unit production tax t_X can induce production at the efficient level X* and generate an economic welfare gain (savings in supply/environmental costs less forgone consumer benefits).
  - Where uncertainty over appropriate tax rate exists, a ban in fragile areas preventing production beyond X* may be more efficient.
  - Combining a ban with a production tax can capture rents left in producers' hands; rent taxation can be integrated into the broader fiscal regime.
- Pricing water use:
  - A usage tax reflects opportunity cost of water and can promote efficient input mixes and site development extent.
  - Efficient price-setting depends on regional water supply/demand characteristics (examples illustrated with prices p1 and p2 and consumption differences X0 − X1 and X0 − X2).
  - Practical relevance may be limited given typically small share of water resources used by shale gas developers.
- Pricing air pollution:
  - Prefer taxing emissions from diesel engines directly; taxing diesel fuel may be more practical but does not incentivize lower emission rates per unit of fuel.
  - A combined approach (fuel tax + emission-rate standard) can mimic an emissions tax.
  - Benefits may be limited because air pollution damages from shale activities are relatively small, especially with emissions controls and remote operations.
- Taxing GHGs:
  - Administrative preference is to levy CO2 charges on fuel supply proportional to carbon content to cover all sources and minimize administrative burdens.
  - Choice of collection point (extractors, processors, distributors) depends on national circumstances and existing tax collection points.

### Alternatives to taxes: regulation and liability
- Regulation:
  - Attractive when environmental taxes are impractical and government targets specific, observable behaviors.
  - Instruments include technology requirements, case-by-case permitting for sites, location and spacing of well sites, drilling methods, casing, fracking, plugging wells, disposal of most oil and gas wastes, site restoration.
  - In the United States, extractive industries are traditionally regulated at the state level; federal authority covers air and surface water quality, endangered species, and federally owned lands; municipalities have additional powers.
- Liability:
  - Litigation can penalize non-compliant firms and deter accidental leakage.
  - Firm-specific ex post penalties can be efficient when producers have better information about risks and mitigation costs.
  - Obstacles: collective action problems, legal/institutional hurdles, difficulty establishing causality, and “judgment-proof firms.”
  - Measures to improve liability effectiveness:
    - Relax legal restrictions on class-action suits.
    - Require information disclosure by companies on environmental risks.
    - Condition operating licenses on minimum asset requirements or liability insurance.
    - Apply vicarious liability to hold principals accountable for agents when agents lack resources.
    - Create industry-level supplementary insurance funds collected through production levies, analogous to the International Oil Pollution Compensation Funds.

### Box 3 — Private information and implications for instrument choice
- Efficiency under private information:
  - If regulator misperceives marginal costs or benefits, standards or taxes can lead to too much (R_H) or too little (R_L) mitigation, with welfare losses (blue and red triangles respectively).
  - Under strict liability, if producers correctly perceive mitigation costs and environmental risks, private incentives can lead to efficient mitigation R*.
- Obstacles and remedies:
  - Difficulty quantifying damages ex post; judicial capacity and civil society constraints in developing countries.
  - Remedies include legal reforms, information disclosure, asset/liability requirements, vicarious liability, and industry insurance funds.
- Integration:
  - Most efficient response is a combination of ex-ante (taxes, insurance and/or regulation) and ex-post (liability) controls, given diverse site-specific risks and instrument strengths/weaknesses.

### Box 6 — A closer look at shale incentives in Oklahoma: regime and modeling findings
- Oklahoma regime mechanics:
  - The state allows a reduced royalty rate of 1 percent for production stemming from horizontal drilling up until all well development costs are recovered or the well’s production reaches 4 years.
  - Footnote: Since 2015 the reduced royalty rate has been increased to 2 percent and applies for a maximum of 36 months.
  - Rationale: shale has a much lower EUR and thus lower production value per well; the incentive applies to a greater percentage of a shale gas well’s total production.
  - Constraint: well cost recovery limit means low-cost wells will not receive a full exemption.
- Comparative modeling results — key findings:
  - North America (shale): Oklahoma breakeven prices are $2.42 and $2.51 per MMBtu, and METRs are 49 and 52 percent, under the shale and conventional tax regimes respectively.
  - General North American finding: almost no difference between fiscal regimes for shale and conventional gas in Texas and North Dakota.
  - Canada example: Alberta’s favorable regime for shale reduces breakeven price from $1.99 to $1.91 compared with the conventional regime.
  - AETR observations:
    - AETRs are upwards of 80 percent in half of the jurisdictions modeled.
    - Projects are deemed marginal if AETRs exceed 90 percent.
    - The breakeven prices in Figure 9 are not that different from the assumed gas price $2.86/MMBtu in the North American baseline.
  - Europe:
    - United Kingdom: METR reduced from 28 to 8 percent (relative to the conventional fiscal regime) for the shale project.
    - Algeria: METR reduced from 59 to 42 percent for shale.
    - China: METR reduced from 59 to 36 percent for shale.
    - Despite reductions, in Algeria and China the breakeven prices are close to the assumed European gas price $4.79/MMBtu, suggesting incentives may not overcome unfavorable economics (AETR only slightly below 100 percent in Figure 8).
    - United Kingdom example: AETRs are 30 and 62 percent for shale and conventional regimes respectively.
  - Net fiscal cost (tax expenditure) of shale regimes vs conventional:
    - United Kingdom: revenue forgone $145 million or 26 percent of pre-tax cash flow (discounted) for the analyzed project.
    - Poland: fiscal subsidy $33 million or 6 percent of pre-tax cash flow.
    - Alberta: tax expenditure $55 million or 13 percent of pre-tax cash flow; Saskatchewan’s tax expenditure is about half of Alberta’s.
- Quantitative project parameters (selected):
  - Project life: Shale 44 Years / Conventional 40 Years.
  - Production Years: Shale 37 / Conventional 32.
  - Gas production: 1.0 Tcf (per project) for all four modeled project cases.
  - Oil production: 50 MMBbl (per project) for all four modeled project cases.
  - Total production (MMBOE): 22 for each project case.
  - Wells drilled (Wells): Shale North America 286; Conventional North America 40.
  - Success Rate: 95% (Shale North America).
  - Gas price used in North America scenarios: $2.86 $/MMBtu; Europe: $4.79 $/MMBtu.
  - Exploration $bn: Shale 0.1 / Conventional 0.2 (North America).
  - Capex $bn: Shale 1.5 / Conventional 0.2 (North America).
  - Opex $bn: Shale 1.0 / Conventional 0.6 (North America).
  - Total cost $bn: Shale 2.7 / Conventional 1.0 (North America).
  - Pre-tax cash flow $bn: Shale 2.8 / Conventional 4.4 (North America).
  - Pre-tax IRR%: Shale 25.2% / Conventional 27.3% (North America).
- Well-level assumptions (selected):
  - Well Life Years: Shale 15 / Conventional 40.
  - Gas production per well Bcf: Shale 3.5 / Conventional 119.1 (North America).
  - Total production per well MMBOE: Shale 0.82 / Conventional 26.5.
  - Drilling/Completion costs $mm per well: Shale 5.0 / Conventional 6.0 (North America).
  - Total cost $mm per well: Shale 8.8 / Conventional 107.9 (North America).
  - Pre-tax cash flow $mm per well: Shale 10.2 / Conventional 95.3 (North America).
- Box 7 comparative highlights:
  - Exploration cost and success rates differ by region: Europe success factors for shale are lower (70 percent compared with 95 percent for conventional) and exploration costs are 50 percent higher; in the United States exploration costs for shale are 50 percent lower than for conventional gas and success rates are the same.
  - Well life: shale 15 years vs conventional 40 years.
  - Production per well: conventional 26.5 MMBOE vs shale less than 1 MMBOE.
  - Operating expenditure: about 60 percent higher per unit of gas recovered for shale.
  - US versus Europe: geology yields lower gas flow rates in Europe (2.5 Bcf per well) vs United States (3.5 Bcf per well); natural gas price United States around $3/MMBtu; Europe $5-10/MMBtu.

### Appendix B — representative project assumptions and fiscal analysis details
- Production, reservoirs, and product definitions:
  - Production aims to maintain a plateau (though this may not necessarily be true for shale gas in the US).
  - For Canadian regimes, reservoirs are taken to be at 3000 meters.
  - Associated liquids (“wet gas”) are included in shale production.
  - Fixed ratio of 1 barrel of oil for every 20 Mcf of natural gas is assumed (no substitutability between oil and gas flows is modeled).
- Project finance, tax treatment, and cash flows:
  - Tax regimes apply throughout project life with taxes paid by one entity, the “investor”.
  - The investor receives all cash flows before paying taxes and interest.
  - Exception when there is state participation (NOC equity stake); in Algerian and Chinese regimes the NOC has exploration costs “carried” by the investor and thereafter covers its share of costs, including repayment of carried costs, with cash.
  - 70 percent of development expenses are financed through debt, while the rest are paid from investor equity.
- Economic environment, prices, and rates:
  - Inflation and interest rates assumed constant throughout the project.
  - Petroleum prices are in constant 2016 real dollars and annually adjust for inflation based on long-term forecasts published in the April 2016 World Economic Outlook.
  - Price levels used:
    - North American gas: US $2.86/MMBtu
    - North American oil: $48.00/bbl
    - Non-US gas: $4.79/MMBtu
    - Non-US oil: $51.01/bbl
- Sensitivity and scenario notes:
  - Saskatchewan simulation: well depth of 2,000 meters (vs 1,500 meters) causes pre-tax IRR to decrease by 0.4 percentage points due to higher costs; under the unconventional regime a large portion of oil becomes eligible for a lower royalty rate at greater depth, increasing investor’s post-tax return while effective royalty rate sharply falls under 2,000 meters scenario.
  - Some substitutability between oil and gas flows is ignored (fixed oil/gas ratio used).

### Key numeric facts and figures (preserved)
- US shale production growth: from 300 Bcf to 15,213 Bcf by 2015.
- Shale as share of US production: about 40 percent.
- Proven shale reserves: 175,000 Bcf (31 percent of total US gas reserves).
- Global technically recoverable shale gas: roughly 7.3 million Bcf (about one third of total world gas resources).
- Canada production (2013): 765 Bcf.
- Water per well: Around 3-5 million gallons for development of a well (Mielke and others 2010).
- Gas price trends:
  - Peak over $12 per thousand cubic feet (Mcf) in 2005.
  - Current levels around $4 per Mcf (as reported in Box 1).
  - Pre-2000 prices around $2.50 per Mcf.
  - Supply curve elasticities cited in Hausman and Kellogg 2015: between about 0.4 and 1.6.
- US fossil fuel CO2 emissions from natural gas combustion: 27 percent.
- CO2 emissions: natural gas combustion produces about 40 percent less CO2 per unit of energy than coal.
- Methane global warming potency: about 25 times CO2.
- Leakage typical estimate fragment provided: "around 1.5 to"
- Oklahoma breakeven prices: $2.42 and $2.51 per MMBtu (shale and conventional regimes respectively).
- North American baseline gas price used in modeling: $2.86/MMBtu.
- European baseline gas price used in modeling: $4.79/MMBtu.
- Project-level examples:
  - Gas production per modeled project: 1.0 Tcf.
  - Oil production per modeled project: 50 MMBbl.
  - Total production (MMBOE) per project: 22.
  - Wells drilled (North America shale): 286.
  - Pre-tax IRR% (North America sample): Shale 25.2% / Conventional 27.3%.
  - AETRs are upwards of 80 percent in half of jurisdictions modeled; projects marginal if AETRs exceed 90 percent.
- Local impact examples:
  - Extra truck traffic estimated to result in 9 extra road fatalities and 12 extra (serious but non-fatal) injuries in Pennsylvania counties with shale gas development.
  - Ground-water dependent property values fall by 16 percent as they become closer to shale wells; properties with piped water access rise by 10 percent.
- Illustrative local pollution damage estimate: around 3 cents per liter (12 cents per gallon) under stated exposure assumption.

*Source: IMF Working Paper — "1. Summary of Instruments to Address Environmental Risks of Shale Gas" (wp17254 excerpts).*

### 1. Summary of Instruments to Address Environmental Risks of Shale Gas ...........................21

### 1. Summary of Instruments to Address Environmental Risks of Shale Gas

### Introduction: questions addressed
- Two central questions the paper attempts to answer:
  - Do the environmental issues associated with shale gas extraction warrant corrective taxes in fiscal regimes for the upstream production activity, or are they better addressed through other instruments?
  - Do the commercial and technical features of shale gas development suggest that the principles of fiscal regime design for conventional oil and gas development need to be modified?
- High-level conclusions (from the supplied text):
  - Many upstream environmental risks (e.g., water contamination) are generally best dealt with through regulation, though ex post liability for individual firms and industry-level compensation funds collected through production fees can play an important role.
  - Novel tax schemes for methane leakages are feasible despite limited metering capabilities.
  - Levies on downstream CO2 emissions from gas combustion on fuel supply are administratively appealing (at extraction, processing, or distribution).
  - Research on the nature and magnitude of environmental risks from shale gas is needed to guide policy design.
  - Simulations suggest special fiscal regimes and incentives for shale gas have, at best, only slight effects on investment while producing significant foregone government revenues.
  - For the most part, there is little evident need to consider specific incentives for shale gas versus conventional gas, provided the overall fiscal regime responds substantially to realized profit or rent.

### Brief overview of the shale gas industry (key facts and drivers)
- Technology and technical features:
  - Shale exploitation requires hydraulic fracturing, seismic imaging, and horizontal drilling.
  - Shale development typically uses these technologies more intensively than conventional exploitation, with more rapid well productivity decline and a greater number of wells per unit of gas.
  - Capital intensity, operating costs, water consumption, and chemical use are generally higher in shale exploitation than in conventional land-based drilling.
- US production and reserves:
  - Production ramped up from 300 billion of cubic feet (Bcf) to 15,213 Bcf by 2015.
  - Shale accounted for about 40 percent of US natural gas production by 2015.
  - Proven shale reserves stand at 175,000 Bcf, or 31 percent of total US gas reserves.
- Global resources and other countries:
  - Technically recoverable reserves totaling roughly 7.3 million Bcf have been identified across the globe, equal to about one third of total world gas resources.
  - Canada produced 765 Bcf in 2013; other countries exploring shale include Argentina, Australia, China, Mexico, Poland, and the UK.
- Factors that favored US development (listed exactly):
  - private ownership of subsurface mineral rights (in other countries these rights are state-owned or may be ill defined, creating legal challenges to rapid development);
  - an abundance of independent gas companies with the requisite technical knowledge or the incentive to develop it;
  - pre-existing gas pipeline infrastructure and capacity;
  - fairly abundant water resources;
  - low population density (and hence public opposition);
  - deep capital markets; and
  - (not least) high natural gas prices in the early 2000s.
- Policy responses in some jurisdictions:
  - Bans or restrictions on fracking apply at national level in Bulgaria, France, Germany, and South Africa and at subnational level in Australia, Ireland, Italy, Spain, Switzerland, and the United States (including New York, Vermont, and cities and counties in California, Colorado, Texas, and Ohio).

### Environmental risks: differences with conventional gas and specific issues
- Differences with conventional gas (environmental perspective):
  - Different pad footprint, water use, chemical use, and pressure below ground.
  - Unconventional development permits multiple wells on a single well pad — could lead to a lower footprint per unit of gas extracted depending on geological features.
  - Fracking uses a mixture of water, sand and a variety of chemicals at high pressure; conventional drilling uses far less water and additives.
  - Greater concentration of wells and high pressures might increase risks of groundwater contamination from casing and cementing failures.
  - Return of fracking fluid up the well raises concerns for liquid wastes handling, treatment, and disposal.
  - Pumping liquid wastes into deep disposal wells might affect seismic activity risks.
  - Use of powerful diesel engines for fracturing contributes to local air pollution.
  - Higher water demands imply more infrastructure (pipelines, roads, water storage) and greater potential for habitat disruption, community impacts, roadway wear and tear, and damages to land and waterways.
  - There appear to be no environmental risks posed by conventional gas-only wells that are not also posed by shale gas wells.
- Water quantity:
  - Shale gas development requires large quantities of water depending on maturity of the shale and formation thickness, and technologies used.
  - Water use per well: "Around 3-5 million gallons for development of a well (Mielke and others 2010)."
  - Per unit of energy, shale gas uses a lot more water than conventional gas, similar amounts to coal, and far less than enhanced oil recovery (Kuwayama and others 2013).
  - Water consumption for shale gas is highly concentrated at a point in time (occurring over several days during fracking).
  - Opportunity costs of water use depend on usage relative to local water availability; shale activity can overlap areas of high water stress in the United States.
  - Water demands for energy extraction are generally trivial compared with agriculture in many regions.
- Water quality:
  - Liquid and solid wastes from fracking or naturally occurring chemicals released by fracking can affect surface and groundwater quality (Burton and others 2013).
  - Liquid wastes can contain fracking fluids and highly salted water, radioactive materials, heavy metals and volatile organics from the formation itself, called “produced water.”
  - Surface water can be affected by runoff, spills, and discharge from treatment plants (e.g., Olmstead and others 2013); impacts are highly site-specific.
  - Groundwater can be affected through leakages due to faulty casing and cementing around the well bore (a standard oil and gas drilling problem).
  - Evidence on whether fracking itself can pollute groundwater by methane or substances in produced and flowback water is mixed (e.g., Osborn and others 2011, Schon 2011).
  - Current data gaps and uncertainties prevent a broad assessment of frequency and severity of fracking harms to water quality (US EPA 2016a).
- Climate change impacts:
  - Downstream combustion of natural gas (conventional and unconventional) is responsible for 27 percent of US fossil fuel CO2 emissions.
  - Combustion of natural gas produces about 40 percent less CO2 emissions than coal per unit of energy.
  - Net effect of shale gas on CO2 is modestly negative for the United States where gas displaces coal (Newell and Raimi (2104) and Brown and others (2009)).
  - Leakage or venting of natural gas (mostly methane) can occur across the lifecycle (drilling, production, processing, distribution, storage).
  - Methane is about 25 times more potent in trapping heat than CO2.
  - Leakage rates are uncertain; a typical estimate is "around 1.5 to" (text ends mid-phrase in supplied content).

### Policy instruments and fiscal considerations (summary from supplied text)
- Preferred policy instruments for upstream environmental risks:
  - Regulation (primary tool) for many upstream risks, e.g., water contamination prevention and waste management.
  - Ex post liability for individual firms; for large risks, compensation via industry-level funds collected through production fees.
- Carbon and methane pricing options:
  - Novel tax schemes for methane emissions from production are feasible despite limited metering capabilities.
  - Levies on downstream CO2 emissions from gas combustion should be applied on fuel supply (administratively appealing at extraction, processing, or distribution).
- Fiscal regime design:
  - Simulation results (across US and non-US jurisdictions) suggest special fiscal regimes and incentives for shale gas have only slight effects on investment but cause significant foregone government revenues.
  - There is little evident need to design specific incentives for shale gas relative to conventional gas if the fiscal regime responds to realized profit or rent.

### Key numeric facts and figures (preserve original values)
- US shale production growth: from 300 Bcf to 15,213 Bcf by 2015.
- Shale as share of US production: about 40 percent.
- Proven shale reserves: 175,000 Bcf (31 percent of total US gas reserves).
- Global technically recoverable shale gas: roughly 7.3 million Bcf (about one third of total world gas resources).
- Canada production (2013): 765 Bcf.
- Water per well: Around 3-5 million gallons for development of a well (Mielke and others 2010).
- Gas price trends:
  - Peak over $12 per thousand cubic feet (Mcf) in 2005.
  - Current levels around $4 per Mcf (as reported in Box 1).
  - Pre-2000 prices around $2.50 per Mcf.
  - Supply curve elasticities cited in Hausman and Kellogg 2015: between about 0.4 and 1.6.
- US fossil fuel CO2 emissions from natural gas combustion: 27 percent.
- CO2 emissions: natural gas combustion produces about 40 percent less CO2 per unit of energy than coal.
- Methane global warming potency: about 25 times CO2.
- Leakage typical estimate fragment as provided: "around 1.5 to"

*Source: IMF Working Paper — "1. Summary of Instruments to Address Environmental Risks of Shale Gas" (excerpts provided).*

### 3.5 percent) and vary across firms (e.g., due to technologies used) and sites.

### wp17254 - 3.5 percent) and vary across firms (e.g., due to technologies used) and sites.

### Local Air Pollution
- Sources and health links:
  - Combustion of diesel in engines used to pump fracking fluids and leaks of volatile organic compounds from storage and gas processing facilities contribute to local air pollution.
  - A voluminous literature documents an association between ambient air pollution and mortality risks (e.g., from heart and lung diseases).
- Illustrative damage estimate:
  - If average population exposure to emissions released in shale extraction is one-tenth of that for emissions from the average road vehicle (and emission rates are comparable), pollution damages for the United States might be around 3 cents per liter (12 cents per gallon).
  - Downstream combustion of natural gas causes local air pollution but damages are typically modest relative to global warming damages because natural gas combustion does not produce direct fine particulate or sulfur dioxide emissions, and nitrogen oxide emission rates are smaller than those for coal.
- Policy implication:
  - Diesel emissions could be targeted via fuel taxes, emission-rate standards, or a combination to incentivize emissions control technologies and fuel-efficiency improvements.

### Ecological and Seismic Risks
- Land and ecosystems:
  - Land conversions for shale gas infrastructure (wells, roads, pipelines, rights of way) potentially disrupt ecosystems, particularly in areas of high well density, though documentation is limited.
- Seismicity and wastewater:
  - Fracturing per se produces quake magnitudes that are small to trivial, but wastewater deep-injection disposal has contributed to earthquake risk and has caused shutdowns of several disposal wells.
  - More evidence is needed on probabilities of added wastes causing seismic events of different severities.
- Site suitability takeaway:
  - Areas with endangered or highly valued habitat, significant populations exposed to pre-existing seismic risks, or lacking viable means to dispose of liquid wastes, are probably not good candidates for shale gas development.

### Community and Broader Impacts
- Local costs and benefits:
  - Local communities benefit economically from shale gas development but also bear environmental, health, and broader impacts like noise, visual degradation, traffic congestion and accidents, property value changes, and overloaded schools.
- Quantified local impacts (examples from studies cited):
  - Extra truck traffic estimated to result in 9 extra road fatalities and 12 extra (serious but non-fatal) injuries in Pennsylvania counties with shale gas development.
  - Ground-water dependent property values fall by 16 percent as they become closer to shale wells, while values for properties with piped water access rise by 10 percent.
- Indirect vs direct impacts:
  - Distinction for policy: direct environmental impacts (require corrective policies; e.g., health effects) vs indirect/pecuniary effects (largely transfers among regions/firms/individuals; may warrant equity-based compensation but not efficiency correction).

### B. The Role of Fiscal Policy
- Objective:
  - Use fiscal policies to reduce overall level of development, water use, air pollution, and GHGs.

- (i) Controlling Development
  - Conceptual framework:
    - Market price fixed at p_c; unit supply costs and environmental costs rise as reserves in progressively fragile areas are exploited.
    - Competitive production at X_c; efficient production at X*, where supply and environmental costs equal p_c.
  - Policy instruments:
    - Unit production tax t_X can induce production at X* and generate an economic welfare gain (savings in supply/environmental costs less forgone consumer benefits).
    - Where uncertainty over appropriate tax rate exists, a ban in fragile areas preventing production beyond X* may be more efficient.
    - Combining a ban with a production tax can capture rents left in producers' hands; rent taxation can be integrated into the broader fiscal regime.

- (ii) Pricing Water Use
  - Rationale:
    - A usage tax reflects the opportunity cost of water in firms’ input costs and promotes efficient input mixes and site development extent.
  - Price-setting depends on water stress:
    - If regional demand intersects the horizontal portion of supply curve → efficient price is unit supply cost p1; welfare cost from failing to charge producers is the red triangle (extra consumption X0 − X1).
    - If regional demand intersects the upward sloping/vertical part → efficient price is higher at p2; failing to charge producers yields larger welfare costs (brown + red areas) associated with excess consumption X0 − X2.
  - Practical note:
    - Argument’s practical relevance is questionable given typically small share of water resources used by shale gas developers.

- (iii) Pricing Air Pollution
  - Instrument design:
    - Ideally tax emissions from diesel engines directly to promote emissions controls and cleaner fuels; taxing diesel fuel may be more practical but does not incentivize lower emission rates per unit of fuel.
    - A combined approach (fuel tax + emission-rate standard) can mimic an emissions tax.
    - Where taxing combustion is impractical, taxes could be imposed at extraction/processing/distribution with rebates for downstream mitigation.
  - Expected net benefits:
    - May be limited because air pollution damages from shale activities are relatively small, especially when emissions controls exist and operations are located away from urban centers.

- (iv) Taxing GHGs
  - Administrative preference:
    - Charges for CO2 emissions are generally best levied on fuel supply (proportional to carbon content) to cover all sources and minimize administrative burdens.
    - Levies downstream at point of combustion miss many emissions sources and increase administrative complexity.
  - Choice of collection point:
    - Whether to tax extractors, processors, or distributors depends on national circumstances and existing tax collection points.
  - Methane leakage:
    - If continuous monitoring of methane leakage were possible, an emissions tax would be ideal; current monitoring technologies provide only discrete measurements at limited sites.
    - Interim option: tax fuel suppliers based on a default leakage rate with rebates for firms demonstrating lower leakage via mitigation and continuous monitoring systems.
  - Alignment with international commitments:
    - For both CO2 and methane, ideally tax levels would be set in line with countries' emissions mitigation pledges for the 2015 Paris Agreement.

### C. Alternatives to Taxes
- (i) Regulation
  - When attractive:
    - Regulatory approaches are attractive when environmental taxes are impractical to administer and government is targeting specific, observable behaviors.
  - Instruments:
    - Technology requirements, case-by-case permitting for sites, location and spacing of well sites, drilling methods, casing, fracking, plugging wells, disposal of most oil and gas wastes, site restoration.
  - U.S. context (summary):
    - Extractive industries traditionally regulated primarily at the state level; federal authority significant for air and surface water quality, endangered species, and on federally owned lands; municipalities can restrict equipment weights, repair roads, tax operations, and constrain well pad locations and methods.
    - Rapid expansion has produced heterogeneity among state regulations of similar elements.

- (ii) Liability
  - Role:
    - Litigation can penalize firms out of compliance and deter one-off events like accidental leakage.
  - Efficiency conditions:
    - Firm-specific ex post penalties can be more efficient than upfront regulations or taxes when producers have better information about risks and mitigation costs.
  - Obstacles to efficient liability:
    - Collective action problems and legal/institutional hurdles in lawsuits affecting many parties.
    - Difficulty establishing clear causality between pollution source and delayed negative effects; producer may no longer exist when long-term effects manifest.
    - Ideally, penalty would equal ex post environmental damages divided by probability of lawsuit, but punitive penalties in excess of damages are rare.

*Source: Excerpt from IMF working paper wp17254.*

### Box 3. Private Information and the Relative Efficiency of Liability vs. Regulations/Taxes

### Box 3. Private Information and the Relative Efficiency of Liability vs. Regulations/Taxes

### Efficiency of mitigation policy under private information (Figure 7 scenario)
- True marginal cost: MC_T.
- True marginal environmental benefit: MB_T.
- Efficient level of risk mitigation: R*.
- If the regulator perceives marginal costs as MC_L (but knows environmental benefits) or perceives marginal benefits as MB_H (but knows mitigation costs), and sets a standard (or equivalent tax) to maximize expected welfare:
  - Risk mitigation will be too high at R_H.
  - Resulting welfare loss indicated by the blue triangle.
- If the regulator perceives marginal costs as MC_H, or marginal benefits as MB_L, and sets the standard (or tax) to maximize expected welfare:
  - Risk mitigation will be too low at R_L.
  - Resulting welfare loss indicated by the red triangle.
- Under strict liability, if producers correctly perceive mitigation costs and environmental risks:
  - The private marginal benefit from risk mitigation coincides with the true marginal benefit.
  - Mitigation will be efficient.

### Obstacles to relying on legal liability for internalizing environmental risks
- Environmental damages may be difficult to quantify ex post:
  - Baseline level of environmental quality at different sites is frequently unavailable.
- Some damages (e.g., to ecosystems) may extend beyond those suffered by victims of the lawsuit.
- Existence of “judgment-proof firms”:
  - Firms may declare bankruptcy if liabilities exceed their net worth.
  - Hazardous-industry businesses may have incentive to organize to contract out risky activities to small, undercapitalized firms.

### Measures to address obstacles to liability-based approaches
- Enhance prospects for lawsuits:
  - Relax legal restrictions on class-action suits.
  - Require information disclosure by companies on environmental risks to enable third-party monitoring and reduce the cost of discovery for plaintiffs (Olmstead and Richardson 2014).
- Conditioning operating licenses:
  - Require minimum asset requirements.
  - Require liability insurance (already a regular feature of many environmental protection laws).
- Vicarious liability:
  - Hold principals accountable for harms committed by agents when agents’ resources are insufficient to cover liabilities.
  - Can discourage contracting out risky activities to undercapitalized firms.
- Industry-level supplementary insurance fund:
  - Collected through production levies on individual shale gas developers to pay compensation for large environmental damages impractical to recover from individual firms.
  - Industry incentive to self-regulate by monitoring member firms’ efforts to limit environmental risks, thereby containing production fees.
  - Arrangement analogous to the International Oil Pollution Compensation Funds, financed by contributions paid on the amount of oil received in the relevant calendar year.

### Considerations for developing and some emerging countries
- Additional constraints on internalizing environmental risks through legal liability:
  - Judicial capacity may be constrained or subject to capture.
  - Civil society organizations that might monitor petroleum companies are likely underdeveloped.
  - Government capacity constraints may prevent public lawsuits and effective regulatory oversight.
- Policy implications:
  - These countries may need to rely more heavily on ex-ante measures, particularly tax and/or insurance instruments that generate a revenue flow to help finance regulation, oversight, prosecution and/or remediation.

### Integration with overall assessment (Section III summary excerpt)
- Numerous and diverse environmental risks are associated with shale gas development, but:
  - Their severity is highly site specific.
  - There is scant quantitative evidence on the risks.
- Most efficient response: some combination of ex-ante (taxes, insurance and/or regulation) and ex-post (liability) controls, given:
  - The diverse nature of the environmental risks.
  - The varying strengths and weaknesses of different instruments in addressing each of them.

*Source: wp17254 - Box 3. Private Information and the Relative Efficiency of Liability vs. Regulations/Taxes*

### Box 6. A Closer Look at Shale Incentives in Oklahoma

### Box 6. A Closer Look at Shale Incentives in Oklahoma

### Oklahoma regime design and mechanics
- The state allows a reduced royalty rate of 1 percent for production stemming from horizontal drilling up until all well development costs are recovered or the well’s production reaches 4 years.
- Footnote: Since 2015 the reduced royalty rate has been increased to 2 percent and applies for a maximum of 36 months.
- Rationale: because shale has a much lower EUR and thus lower production value per well than onshore conventional gas, the incentive applies to a greater percentage of a shale gas well’s total production.
- Constraint effect: the well cost recovery limit means low-cost wells will not receive a full exemption, so low-cost (less likely marginal) wells are not fully subsidized.

### Figure 8 — effective royalty dynamics (qualitative summary)
- The effective royalty rate under the unconventional (shale-targeted) regime increases with gas price when applied to a shale gas project.
- For a conventional gas project, the unconventional regime performs nearly identically to the conventional-regime rate.
- The chart labels include gas prices in $/MMBtu (values shown on the x-axis): 1.71, 2.06, 2.40, 2.74, 3.08, 3.43, 3.77, 4.11, 4.46, 4.80 (displayed as contiguous ticks in the figure).

### Comparative modeling results — key findings (Figures 9–10 and text)
- North America (shale): Oklahoma breakeven prices are $2.42 and $2.51 per MMBtu, and METRs are 49 and 52 percent, under the shale and conventional tax regimes respectively.
- General North American finding: almost no difference between fiscal regimes for shale and conventional gas in Texas and North Dakota.
- Canada (example): Alberta’s favorable regime for shale reduces breakeven price from $1.99 to $1.91 compared with the conventional regime.
- AETR observations:
  - AETRs are upwards of 80 percent in half of the jurisdictions modeled.
  - Projects are deemed marginal if AETRs exceed 90 percent.
  - The breakeven prices in Figure 9 are not that different from the assumed gas price $2.86/MMBtu in the North American baseline.
- Europe:
  - Fiscal regimes in Europe would provide stronger incentives for shale gas in several cases. Examples:
    - United Kingdom: METR reduced from 28 to 8 percent (relative to the conventional fiscal regime) for the shale project.
    - Algeria: METR reduced from 59 to 42 percent for shale.
    - China: METR reduced from 59 to 36 percent for shale.
  - Despite reductions, in Algeria and China the breakeven prices are close to the assumed European gas price $4.79/MMBtu, suggesting incentives may not overcome unfavorable economics (AETR only slightly below 100 percent in Figure 8).
  - United Kingdom example: AETRs are 30 and 62 percent for shale and conventional regimes respectively; marginal projects might not be developed without a favorable regime.
- Conventional European projects: significant differences between shale and conventional fiscal regimes exist in some countries (e.g., China: METRs 49 and 21 percent respectively), while in Poland METR and AETR are almost identical under both regimes.
- Net fiscal cost (tax expenditure) of shale regimes vs conventional:
  - United Kingdom: revenue forgone $145 million or 26 percent of pre-tax cash flow (discounted) for the analyzed project.
  - Poland: fiscal subsidy $33 million or 6 percent of pre-tax cash flow (smallest outside North America in the sample).
  - Alberta: tax expenditure $55 million or 13 percent of pre-tax cash flow; Saskatchewan’s tax expenditure is about half of Alberta’s.

### Quantitative parameters (selected highlights from Table 2 and project assumptions)
- Project-level assumptions (selected entries):
  - Project life: Shale 44 Years / Conventional 40 Years (North America sample columns).
  - Production Years: Shale 37 / Conventional 32.
  - Gas production: 1.0 Tcf (per project) for all four modeled project cases.
  - Oil production: 50 MMBbl (per project) for all four modeled project cases.
  - Total production (MMBOE): 22 for each project case.
  - Wells drilled (Wells): Shale North America 286; Conventional North America 40; Shale Europe 403 (implied elsewhere); Conventional Europe 38.
  - Success Rate: 95% (Shale North America) vs 70% (Conventional Europe noted in Box 7 context).
  - Gas price used in North America shale/conventional scenarios: $2.86/$2.86 $/MMBtu; Europe: $4.79/$4.79 $/MMBtu.
  - Exploration $bn: Shale 0.1 / Conventional 0.2 (per project in North America columns).
  - Capex $bn: Shale 1.5 / Conventional 0.2 (North America sample).
  - Opex $bn: Shale 1.0 / Conventional 0.6 (North America sample).
  - Total cost $bn: Shale 2.7 / Conventional 1.0 (North America sample).
  - Pre-tax cash flow $bn: Shale 2.8 / Conventional 4.4 (North America sample).
  - Pre-tax IRR%: Shale 25.2% / Conventional 27.3% (North America sample).
- Well-level assumptions (selected entries):
  - Well Life Years: Shale 15 / Conventional 40.
  - Gas production per well Bcf: Shale 3.5 / Conventional 119.1 (North America sample).
  - Total production per well MMBOE: Shale 0.82 / Conventional 26.5.
  - Drilling/Completion costs $mm per well: Shale 5.0 / Conventional 6.0 (North America sample); other entries show higher conventional well costs in Europe.
  - Total cost $mm per well: Shale 8.8 / Conventional 107.9 (North America sample).
  - Pre-tax cash flow $mm per well: Shale 10.2 / Conventional 95.3 (North America sample).

### Box 7 — major differences (summary of comparative characteristics)
- Shale vs conventional (selected points):
  - Exploration cost and success rates differ by region: Europe success factors for shale are lower (70 percent compared with 95 percent for conventional) and exploration costs are 50 percent higher; in the United States exploration costs for shale are 50 percent lower than for conventional gas and success rates are the same.
  - Well schedule and production life: shale wells drilled throughout project life; shale shorter production life 15 years vs conventional 40 years.
  - Production per well: conventional much higher—26.5 MMBOE vs less than 1 MMBOE for shale.
  - Operating expenditure: about 60 percent higher per unit of gas recovered for shale.
- US versus Europe (selected points):
  - Well drilling cost higher in Europe.
  - Geology: lower gas flow rates in Europe (2.5 Bcf per well) vs United States (3.5 Bcf per well).
  - Natural gas price: United States around $3/MMBtu; Europe $5-10/MMBtu (contextual ranges used in modeling).

### Conclusions and policy implications
- Modeling suggests special fiscal regimes and incentives for shale gas development have, at most, only a slight impact on likely investment decisions, while potentially imposing substantial tax expenditures on government budgets.
- Other factors likely more important in promoting rapid shale development in the USA: private ownership of resource rights and availability of extensive gathering and distribution infrastructure.
- Policy prescription from results:
  - For the most part, there is little evident need for shale-specific fiscal incentives if the overall fiscal regime responds substantially to realized profit or rent.
  - Principal exception: jurisdictions that rely heavily on gross royalties or on initial or continuing flat fees or bonuses may need to consider special treatment to avoid discouraging shale development.
- Overall assessment: the case for providing relative fiscal incentives to unconventional over conventional natural gas is not especially compelling.

*Source: IMF Working Paper — Box 6. A Closer Look at Shale Incentives in Oklahoma (contained in wp17254).*

### Appendix B. Further Details on Representative Projects for Fiscal Regime Analysis

### Appendix B. Further Details on Representative Projects for Fiscal Regime Analysis

### Production, reservoirs, and product definitions
- Production aims to maintain a plateau (though this may not necessarily be true for shale gas in the US).
- For Canadian regimes, where fiscal parameters depend on depth, reservoirs are taken to be at 3000 meters.
- Associated liquids, so called “wet gas” (e.g., ethane, butane) are included in shale production.
- A fixed ratio of 1 barrel of oil for every 20 Mcf of natural gas is assumed (no substitutability between oil and gas flows is modeled).

### Project finance, tax treatment, and cash flows
- Tax regimes are taken to apply throughout the project life with taxes paid by one entity, the “investor”.
- The investor receives all cash flows before paying taxes and interest.
- Exception: when there is state participation, meaning that the national oil company (NOC) has an equity stake in the project. In the Algerian and Chinese regimes, the NOC has its share of exploration costs “carried”, or paid for, by the investor and thereafter covers its share of costs, including the repayment of carried costs, with cash.
- 70 percent of development expenses are financed through debt, while the rest of project costs are paid from investor equity, reflecting an assumption that investors are not able to borrow to fund exploration and prefer to avoid excessive leverage (capitalization in line with prevailing debt-to-equity tax ratios).

### Economic environment, prices, and rates
- Inflation and interest rates are assumed to be constant throughout the project.
- Petroleum prices are in constant 2016 real dollars and annually adjust for inflation based on long-term forecasts published in the April 2016 World Economic Outlook.
- Price levels used:
  - North American gas: US $2.86/MMBtu
  - North American oil: $48.00/bbl
  - Non-US gas: $4.79/MMBtu
  - Non-US oil: $51.01/bbl

### Sensitivity and scenario notes (from footnotes)
- Additional simulations for Saskatchewan: a well depth of 2,000 meters, rather than 1,500 meters, causes the pre-tax IRR to decrease by 0.4 percentage points due to the rise in costs. However, under the unconventional regime a relatively large portion of oil becomes eligible for a lower royalty rate at the greater depth, so the investor’s post-tax return increases while the effective royalty rate sharply falls under the 2,000 meters scenario. A less significant impact is seen for the conventional regime due to the lower portion of petroleum eligible for the lower royalty rate.
- There is some substitutability between oil and gas flows which becomes important when gas prices decline sharply relative to those for oil; this feature is ignored in the analysis (fixed oil/gas ratio is used as noted above).

*Source: wp17254 - Appendix B. Further Details on Representative Projects for Fiscal Regime Analysis*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2017/wp17254.pdf_