## 1mdgea2022004

## Source details

**Canonical URL:** [1mdgea2022004](https://www.imf.org/-/media/files/publications/cr/2022/english/1mdgea2022004.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2022/english/1mdgea2022004.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2022/english/1mdgea2022004.pdf.json)

---

### Program overview and mission
- Climate Macroeconomic Assessment Program (CMAP) successor to CCPA, focused on macroeconomic implications of climate policies; Madagascar is the second pilot country.
- IMF mission: remote discussions Feb 28–Mar 21, 2022; hybrid mission Apr 1–12, 2022. Team led by Valerie Cerra and included IMF FAD, RES, AFR staff and World Resources Institute contributors.
- Extensive engagement with Malagasy authorities and institutions, including Minister of Economy and Finance and Minister of the Environment and Sustainable Development.

### Climate exposure, socioeconomic vulnerability, and sectoral impacts
- Hazard profile and trends:
  - Exposed to tropical cyclones, droughts, and floods; increase in extreme heat events and more unpredictable rainfalls.
  - Average temperature increase between 1991-2020 relative to 1960-1990: 0.53°C.
  - Mid-century mean annual temperature increase under SSP1-4.5 likely between 0.87°C and 1.67°C.
  - Warm spell duration index base period: 1979-2009; warm spell duration projected to increase by 70 days a year (median) in RCP 4.5.
  - Recent observation: 2021/2022 drought with the south experiencing 70 percent less than normal rainfall.
  - Major tropical cyclones: average about once every three years since 1998; Jan–Feb 2022 saw four cyclones (two category 3-4).
- Socioeconomic vulnerability:
  - More than 80 percent of households live in poverty.
  - Services contribute 55 percent of GDP; primary industries (including agriculture, forestry, and fishery) contribute 25 percent of GDP.
  - Agriculture employs about 64 percent of the active population; fishing contributes about 7 percent to the economy and exports and supports about 5 percent of the population.
- Sectoral impacts (selected):
  - Agriculture: projected decreases from loss of cultivable land, drought, pests, and labor productivity loss from extreme heat.
  - Water and sanitation: cyclones 2017–2019 caused drinking water pollution and disease outbreaks; 2021/2022 drought generated massive food crisis and acute malnutrition in the Grand South.
  - Ecosystems: tree loss, biodiversity loss, mangrove and marine vegetation decline, fisheries impacts.
  - Infrastructure: expected annual damages to roads and railway assets around 0.2% of GDP under current climate.
  - Energy: hydropower variability may be affected; hydropower represented 29% of Madagascar’s energy mix (contextual).

### National strategy, mitigation ambitions, and gaps
- National strategy status:
  - National development plan IEM 2019-2023 includes environmental pillar; 2016 NDC covers mitigation and adaptation and provides a global cost estimate but lacks detailed, costed individual measures.
  - National Adaptation Plan (2019) contains defined and costed projects/activities.
  - Process to update key elements (National Policy for Fighting Climate Change and NDC) started in 2021 and expected to conclude in first half of 2022 (drafts not shared with mission).
- Emissions baseline and mitigation targets:
  - Madagascar net GHG emissions in 2020: 49 mt CO2e (0.1 percent of global GHG emissions) or 0.13 metric tons per capita.
  - Comparative benchmark: Sub-Saharan African countries (excluding high income) at 0.76 metric tons per capita.
  - 2016 NDC mitigation pledge: reduce GHG emissions sources by 14 percent compared to BAU by 2030; increase GHG absorption of sinks by 32 percent by 2030, remaining a net absorber.
  - Reported mitigation cost: USD 6.37 billion between 2015-30 with government to contribute about 4 percent of indicated costs.
- Forestry and land-use:
  - 2010 tree cover: 16.4Mha (28 percent of land area).
  - 2010–2020 tree cover loss: 241,000 ha by 2020, equivalent to 125 mt CO₂ emissions (Global Forest Watch).
  - Natural forest coverage fell from 9.4Mha in 2005 to 9.2Mha in 2010 (≈40,000 ha per year).
  - Reforestation potential: 270,000 ha reforestation can absorb up to 4 mt CO2e per year (~8 percent of Madagascar's total emissions in 2020).
- Implementation gaps:
  - Climate change not mainstreamed into sectoral strategies; sectoral documents often standalone and not fully costed or forward-looking.
  - Lack of understanding of future climate impacts at sectoral level; cross-cutting standards exist but climate projections and hazard risk assessments still being developed.

### Recommended mitigation policies and fiscal measures
- Fiscal and pricing measures (timing: Short to Medium term):
  - Normalize excise taxes and other levies on diesel relative to gasoline.
  - Adopt an automatic fuel pricing mechanism (roadmap to implement by 2024Q1).
  - Eliminate the VAT exemption on the first 100 kWh of electricity consumption (Short term).
  - Gradually increase electricity tariffs to cost recovery levels (Medium term).
  - Consider recycling tax revenue from mitigation to strengthen the social safety net (Medium term).
- Market and regulatory instruments:
  - Use feebates to encourage forest carbon storage, fuel-efficient vehicles, and energy efficient appliances (Medium term).
  - Feebate design examples include vehicle feebates using emissions-rate based carbon price; appliance feebates to promote energy-efficient cookstoves and appliances.
  - Feebates for forest carbon storage: formula Y_t = τ_t^CARBON * (CAR_t^i − CAR_t,BASE^i) with design issues including defined baselines, annual payments, property rights (only 25 percent of land surface registered), and phased CO2 price aligned with NDC.
- Energy sector decarbonization objectives:
  - Government target: increase electricity access from 25 percent (2018) to 70 percent by 2030 and increase renewable energy share from around 38 percent (2020) to about 80 percent by 2030.
  - CPAT baseline: energy emissions about 5 mt CO2e for the energy sector in 2030; 2015 NDC projected BAU energy emissions of 13 mt CO2e by 2030 (noted as significantly higher than CPAT baseline).
  - Investment needs to reach 80 percent RE by 2030: around 0.4 percent of GDP on average annually over next decade; ~900 MW new hydropower capacity required (current pipeline ~240 MW; gap 660 MW).

### Fuel pricing, taxes, and recommended tax package
- Current petroleum tax rates collected by customs:
  - Gasoline: MGA 503 per liter.
  - Diesel: MGA 228 per liter.
  - Lamp oil: MGA 10 per liter.
- VAT and levies:
  - Gasoline and diesel: VAT at 20 percent; lamp oil: VAT exemption.
  - Road Maintenance Fund levy: MGA 288 per liter.
  - Environmental Fund levy: MGA 134 per liter.
- Findings on efficient pricing (as of 2021):
  - Gasoline: excise and levy taxation raised prices to around its efficient level (US 15 cents/liter reflecting GHG cost at USD 64/tCO2e).
  - Diesel, kerosene, and LPG: taxation below efficient level.
  - With recent global fuel price surge and incomplete pass-through, the gap to efficient prices likely higher now.
- Sample recommended tax package and impacts:
  - Raise diesel excise by 275 MGA/liter.
  - Raise road maintenance levy by 154 MGA/liter.
  - Eliminate VAT exemption for first 100 kWh monthly by 2023.
  - Adopt automatic fuel pricing mechanism by 2024.
  - Increase average electricity tariffs by 15 percent annually per JIRAMA recovery plan (example: raise average tariff from $0.15 per kWh to $0.22 per kWh by 2025, retaining social tariff for vulnerable households).
  - Fiscal and mitigation impacts: up to 1.7 percent of GDP in revenue by 2025; energy CO2 reductions up to 5 percent relative to baseline in 2030 (modeled under BAU scenario).

### Distributional impacts and revenue recycling
- Projected price increases from package:
  - Excise normalization and automatic fuel pricing and electricity tariff increases would raise prices by 10, 44 and 36 percent, respectively.
- Household impacts:
  - Direct burden from higher energy prices = 1.6 percent of household consumption, borne mostly by richest households.
  - Indirect costs (transport-induced food price increases) burden bottom four quintiles relatively more.
  - Overall consumption changes: poorest quintile loss 1.1 percent; richest quintile loss 2.4 percent.
- Compensation:
  - Only 10 percent of the 1.7 percent of GDP in revenues would be needed to fully compensate bottom 4 quintiles for welfare loss in 2025.
  - Remaining 1.5 percent of GDP could support electricity access, strengthen social spending, invest in resilience.
  - Example options using about 1.2 percent of GDP in revenues: universal lump sum transfer; expand Human Development Cash Transfer (HDCT) to universal grant per child below age 5 or below age 15; non-contributory old-age pension.
  - Grant per child below age 5: reduces Gini from .405 to .381 and reduces poverty rate by 1 percent; raises consumption by 14 percent for the poorest quintile and by 8 percent for the second poorest.

### Disaster risk management, financing instruments, and fiscal buffers
- Disaster exposure and fiscal costs:
  - Historical frequencies: tropical cyclone every 0.9 years on average; droughts every 2.7 years; floods every 3 years (EM-DAT, 2022).
  - Typical cyclone produces 1.8% of GDP in total losses (57% in physical asset damages); average flood 2.4% of GDP in total losses.
  - Annual average fiscal costs of natural disasters estimated at about 1 percent of GDP (SWIO RAFI: AALs over $100mn or ~1% of GDP).
  - SWIO RAFI: 10 percent probability damages exceed $230mn in any given year.
  - Recent event estimates vary widely: ARC estimate for cyclone Batsirai up to US$2.1bn (14% of GDP); World Bank estimate ~4.5% of GDP.
- Existing fiscal mechanisms:
  - Organic Budget Law provides annual contingency line but not activated in practice.
  - Advance decrees allow additional funds up to 1 percent of approved budget; intra/inter-ministerial reallocations up to 10 percent of appropriations.
  - Imprest accounts: renewable imprests up to 15mn MGA; exceptional single imprests up to open credit lines.
  - Supplementary budgets used but can take months.
  - National Disaster Fund (NDF): annual transfers of 2bn MGA (~0.003 percent of GDP); managed extra-budgetary by BNGRC.
  - FX reserves end-2021: $2.2bn (~6 months of imports).
- Risk transfer and contingent financing:
  - ARC drought insurance (2019): coverage up to $2.5mn; premia ~$0.5mn/year; Government pays full premium in 2023; $2.1mn payout July 2020.
  - ARC tropical cyclone parametric insurance: annual premium $2mn (ADRiFi financed); $10.7mn payout March 2022 (TC Batsirai).
  - WB Cat-DDO: country limit for IDA clients $250mn; IRM can mobilize up to 5 percent of undisbursed IDA projects; CERCs available in 13 of 21 ongoing IPFs with undisbursed ~ $1bn.
  - IMF ex-post emergency financing examples: PRGF $32mn (2000), ECF $42.4mn (2017), RCFs $338mn (2020), CCRT debt relief $30mn (2020-22).
  - WB CRW examples: $20mn (2016 El Niño), $50mn (2021); discussions for $250mn CRW for 2022 cyclones.
- Insurance market and public assets:
  - Public assets generally uninsured; government acts as self-insurer.
  - Insurance market: five licensed companies; two SOEs >70% market share; total assets ~2 percent of GDP; premiums ~0.7 percent of GDP; insurance access ~3 percent of adult population.
  - Recommendation: finalize National Disaster Risk Financing and Insurance (DRFI) Strategy; integrate DRFI in risk-layering framework; develop asset registries and consolidated policies.

### Public investment management (PIM), PFM, and governance gaps
- PIM/PFM context and gaps:
  - 2017 PIM strategy in place; opportunity to integrate climate-specific elements into project appraisal and selection.
  - No standardized climate methodology in project appraisal and selection; environmental impact studies (ONE) occur late and lack specific climate analysis.
  - Capital budget per 2022 budget law: 5,257 billion Ariary (MGA), 36.4 percent of total expenditure and 9.2 percent of GDP.
  - Externally financed public investment projects: 94 projects over last ten years with total commitments USD 2,756.2 million and disbursements USD 991 million.
    - Rural development: 48 projects; commitments USD 900.3 million; disbursements USD 322.2 million; 33% of total disbursement.
    - Environment: 23 projects; commitments USD 159.9 million; disbursements USD 88 million; 9%.
    - Energy: 11 projects; commitments USD 480.2 million; disbursements USD 215 million; 22%.
    - Road infrastructure: 6 projects; commitments USD 838.7 million; disbursements USD 169.5 million; 17%.
    - Fishing: 1 project; commitments USD 74 million; disbursements USD 17.4 million; 2%.
    - Others: 5 projects; commitments USD 303.1 million; disbursements USD 178.9 million; 18%.
- Identification, tracking and audit:
  - No clear budget identification or tracking of climate-related public investment; FMIS modules (budget preparation and execution) not fully integrated.
  - PIP code not always assigned or followed through execution; recommendation to implement tagging system (climate budget tagging).
  - No systematic ex post reviews of climate impacts except for some externally financed projects; no legal framework for ex-post evaluation including climate impacts.
- Subnational governments and public corporations:
  - 23 regions, 119 districts, 1,695 communes; communes' investment = 2 percent of total public investment in 2017 and largely unfunded mandates.
  - Decrees 2019-1931 and 2019-1932 require SNGs to integrate climate resiliency into investment planning, but low own resources and FDL allocation criteria do not factor climate vulnerability.
  - Combined balance sheet of majority-owned public corporations = 11 percent of GDP in 2018; no State shareholder strategy incorporating climate aspects.
- Recommendations for PIM/PFM:
  - Define methodology to integrate climate in ex ante project assessments; identify vulnerable infrastructure and maintenance methodologies; introduce climate annex to annual budget; use budget classification reform to enable climate tagging; strengthen coordination between Ministry of Environment, Ministry of Economy and Finance, and line ministries; develop centralized guidance for climate-related investments and asset registries.

### Macroeconomic modeling of climate impacts (DIGNAD) and scenario outcomes
- Model: DIGNAD (Debt, Investment, Growth, and Natural Disasters) calibrated to Madagascar; two public capital types (standard vs resilient).
- Key calibrated parameters / initial values (percent and amounts preserved exactly):
  - Return to standard infrastructure: 25
  - Return to adaptation infrastructure: 30
  - Public investment to GDP ratio: 5.2
  - Grants to GDP ratio: 1.4
  - Consumption tax (VAT) rate: 20
  - Public domestic debt to GDP ratio: 13.2
  - Public external concessional debt to GDP ratio: 22.3
  - Public external commercial debt to GDP ratio: 3.5
  - Remittances to GDP ratio: 2.6
  - Imports to GDP ratio: 17.6
  - Real interest rate on domestic debt: 0
  - Real interest rate on external commercial debt: 0.98
  - Depreciation rate of standard public infrastructure: 7.5
  - Depreciation rate of resilient public infrastructure: 3
  - Public investment efficiency: 31%
- Climate-adjusted baseline results (if government fully covers reconstruction/humanitarian needs):
  - Unconstrained financing scenario:
    - Debt-to-GDP increases by 28 percentage points relative to no-climate baseline by 2040.
    - Debt exceed 70% of GDP in 2034 and reach 84% of GDP in 2040.
    - Overall deficits average -5.4% of GDP after 2024.
    - Total cumulative GDP losses 4.4% of GDP by 2040.
    - Trend real GDP growth lower by 0.18 percentage points relative to no-climate scenario.
  - Financing-constrained scenario (indicative debt ceiling ≈ 65% of GDP):
    - Debt stabilizes around 65% by 2040 via sharp public investment cuts and revenue mobilization.
    - Durable cumulative GDP losses 5.2% by 2040; average yearly GDP growth loss 0.22 percentage points.
- Contribution to total real GDP loss by 2040:
  - Climate shocks (floods, cyclones, droughts): about 63 percent.
  - Climate change shocks (increased intensity/frequency and productivity loss): about 37 percent.
- Policy scenarios modeled:
  - Scenario 1: adaptation investments including building back better and resilient investment scale-up = 1.5% of GDP annually → debt reaches 98% of GDP; adaptation capital limited by 2040.
  - Scenario 2: Scenario 1 + public investment efficiency improvement from 31% to 63% over 10 years → total debt-to-GDP reduced by 4 percentage points relative to Scenario 1 but remains unsustainable.
  - Scenario 3: SDG-oriented investments (human capital and resilient infrastructure) amounting to 4.8% of GDP per year for 10 years, financed partly by mitigation revenues and additional grants (assumes ~1% of GDP/year in additional grants) → debt on downward trajectory in medium term; human capital investments raise productivity and potential growth (durable productivity boost of 17% by 2040 in modeling); requires substantially higher donor support and credible high-quality climate measures.
- Financing mix assumptions for Scenarios 1 and 2:
  - 67.1% concessional debt, 24.6% grants, 8.3% external commercial debt.

### Adaptation priorities, costs, and returns
- Agriculture:
  - Observed productivity fall: USD 95 per worker over past 20 years (World Bank, 2022).
  - NAP and sectoral plans: agricultural research and disaster-risk reduction costing up to USD 157 million (1.1 percent of nominal GDP in 2019) annually.
  - Gap: focus on emergency response rather than pre-disaster resilience; strategies for slow-moving effects missing.
  - Evidence: additional US$1.97 billion per year in agricultural R&D for climate resilience across emerging and developing economies could improve food supply by 6–7 percentage points between 2030 and 2050 and reduce hunger by 21–24 percentage points (IFPRI 2019).
- Water and sanitation:
  - NAP recommends expanding and climate-proofing water infrastructure; NAP total adaptation measures expected to cost $25 million over ten years.
  - Poor, outdated infrastructure vulnerable to floods and contamination.
- Ecosystems:
  - NAP ecosystem programs cost USD 90 million over 10 years.
  - Benefits of protected areas: USD35 to USD53 per hectare per year; management costs ~20% of benefits; creation costs USD27–USD101 per hectare recoverable in ~3 years.
- Infrastructure and buildings:
  - Current adaptation programmatic costs range USD110 million – USD170 million annually (~1% of nominal GDP in 2019).
  - Resilient transport infrastructure average cost increase 5% in LMICs and can reduce damage by 50%.
  - Example 2022 storms: infrastructure damage ~USD170 million; resilience could have reduced damage by USD85 million; adaptation investment estimated between USD25 million and USD115 million (depending on cost assumptions).
- Non-fiscal measures:
  - Risk zoning, early warning, weather forecasts, school emergency training exist; rollout of resilience standards constrained by local capacity.

### Public financial preparedness, social protection, and insurance-linked reforms
- Social protection and shock-response:
  - Social spending ~0.3 percent of GDP.
  - Health insurance covers ~15 percent of population (mostly private).
  - Pension assets <2 percent of GDP; public system covers 3 percent of workforce; private covers 10 percent.
  - No unemployment insurance.
  - Main cash transfer programs: HDCT and PSNP; current coverage ~2.5 million beneficiaries with target 4 million (20 percent of population under poverty line); payments ~ $100/household/year (~20 percent of poor household consumption basket).
  - Integrated beneficiary registry operational in 2020; recommendation to expand to wider social registry and leverage central digital ID to improve targeting and scalability.
- Risk transfer and DRFI progress:
  - Madagascar has used ARC, WB IRM/CERCs, IMF emergency facilities and WB CRW; recent payouts include ARC $2.1mn (2020 drought) and $10.7mn (2022 cyclone Batsirai).
  - Recommendation: finalize DRFI strategy under AfDB-ADRiFi assistance and coordinate instruments within a risk-layering framework.
- Financial sector and digital services:
  - Financial stability risks from disaster exposure concentrated in sectors (commerce, transportation, manufacturing, agriculture); prior stress tests suggested banks may incur losses up to 40 percent of 2014-15 net income.
  - Digital/mobile money widely used for transfers and inclusion; CSBF simplified customer due diligence for mobile accounts during Covid.

### Energy subsidy reform and social protection (Annex VI)
- Main recommendation:
  - Use budgetary savings from energy pricing reforms to expand social transfers to protect the poor.
- Empirical evidence:
  - Study of 25 MENA fuel subsidy reform episodes: all reforms that used cash transfers successfully removed fuel subsidies; only 17 percent of reforms without cash transfers succeeded.
- Country lessons and practical implications:
  - Iran (2010): large price increases (230–840 percent) with bi-monthly cash transfers to 70–80 percent of citizens; ~80 percent of revenue used for transfers.
  - Kenya (2001–2008): phased removal of electricity subsidies and redirected subsidies to subsidize connection costs, increasing rural connections from 650,000 (2003) to 2 million (2014).
  - Ghana (2005): 50 percent average petroleum price increase accompanied by expanding free education, transport increases, health funding.
  - For Madagascar: prioritize expanding social transfers from energy-pricing savings; consider broad-based/universal transfers when targeting capacity limited; reallocate savings to access expansion (e.g., subsidized connection costs) and combine with sector reforms (e.g., JIRAMA efficiency improvements).

### Main policy implications and recommended sequencing
- Integrate natural disasters and climate change into baseline macro-fiscal projections and annual/medium-term budgets (short-term).
- Prepare a well-costed and prioritized package of climate measures: adaptation, mitigation, public investment efficiency, and SDG-oriented human capital investments (short-term).
- Implement an automatic fuel pricing mechanism by 2024Q1 to eliminate fiscal risks from price controls.
- Normalize fuel taxation and eliminate VAT exemption for first 100 kWh; recycle revenues to social protection and resilience investments (Short–Medium term).
- Improve public investment efficiency (increase PIE from 31% toward 61% over 10 years), integrate climate in PIM/PFM (project appraisal, budgeting, tagging, asset maintenance), and finalize DRFI strategy to reduce fiscal volatility.
- Mobilize additional grants and concessional financing contingent on credible climate policies and improved governance and PIM capacity to achieve sustainable growth trajectory (Policy scenario 3).

*Source: 1mdgea2022004*

### PREFACE  _________________________________________________________________________________________ 7

### PREFACE

### Program overview
- The Climate Macroeconomic Assessment Program (CMAP) aims to assist countries, especially small and low-income countries, to build resilience and develop policy responses to cope with the economic impact of climate change.
- CMAP is a successor to the Climate Change Policy Assessment (CCPA) with more focus on macroeconomic implications of climate change policies.
- Madagascar is the second pilot country where a CMAP has been conducted.

### Mission and team
- IMF staff held remote discussions for the CMAP during February 28-March 21, 2022, followed by a hybrid mission during April 1-12, 2022.
- The mission team was led by Valerie Cerra and comprised Vybhavi Balasundharam, Simon Black, Chen Chen, Majdeline El Rayess, Joey Kilpatrick, Katja Funke, Nabil Hamliri, Amr Hosny, Claude Wendling (all FAD), Zamid Aligishiev (RES), and Dominique Fayad (AFR), and included Carter Brandon, Juan Carlos Altamirano, and Erin Gray (World Resources Institute).
- The team engaged extensively with Malagasy authorities and institutions, including Ms. Rindra Hasimbelo Rabarinirinarison, Minister of Economy and Finance; Ms. Marie-Orlea Vina, Minister of the Environment and Sustainable Development; multiple line ministries, the Central Bank of Madagascar, CSBF, BNGRC, CPGU, ONE, JIRAMA, and ARO Insurance Company.

### Climate exposure and macroeconomic framing
- Madagascar is exposed to multiple climate hazards: tropical cyclones, droughts, and floods; climate change is bringing more extreme heat events and more unpredictable rainfalls.
- Extreme events have caused sizable losses in the agriculture, ecosystems, water, and infrastructure sectors; slow-moving effects such as rising temperatures create pressure on labor productivity across sectors.
- Madagascar’s high level of extreme poverty and dependence on agriculture exacerbate vulnerability.

### National strategy and planning gaps
- Madagascar has a set of key climate related policies and strategies in place, many aligned with the Nationally Determined Contribution (NDC); several documents, including the NDC, are being updated.
- Climate change has not been mainstreamed into sectoral strategies; several sectors developed standalone climate documents.
- Lack of understanding of future climate impacts at the sectoral level hinders proactive measures and implementation capacity; Table 1 in the source summarizes strategic gaps.

### Mitigation priorities and fiscal instruments
- Madagascar is a relatively small emitter in terms of energy-based CO2 but has higher emissions per capita compared with countries at comparable income levels due to deforestation; mitigation commitments rely heavily on LULUCF, and on energy, agriculture, and waste sectors.
- Policy measures noted:
  - Adopt a feebate system for forest carbon storage.
  - Encourage uptake of energy-efficient appliances to reduce reliance on firewood and charcoal.
  - Increase the excise on diesel.
  - Gradually increase electricity tariffs to cost recovery levels.
  - Implement an automatic fuel pricing mechanism.
- Revenue from tax measures should be used to meet SDG goals such as strengthening social safety nets to mitigate adverse impacts on vulnerable households.

### Disaster risk management and financing
- Progress exists on elements of national disaster risk management, but gaps remain:
  - Integrate disaster risk assessments into the budget process to facilitate adequate provisioning for fiscal risks.
  - Expand the beneficiary registry to a wider social registry including vulnerable non-beneficiaries to enable rapid scale-up of social protection.
  - Further develop domestic insurance to increase population financial resilience.
  - Finalize the Disaster Risk Financing and Insurance (DRFI) strategy to provide centralized guidance in a risk layering framework.
- Domestic resources available for disaster response remain small despite recent improvements in external contingent financing and sovereign insurance instruments.

### Adaptation investments and prioritization
- Under the National Adaptation Plan, some key sectors have sector-specific climate documents, but these are not mainstreamed into sectoral strategies.
- Proposed investments and non-fiscal measures are insufficient to meet adaptation needs; many proposed investments could yield disaster-impact reduction, economic, and environmental benefits.
- A prioritization framework is lacking; cost estimation and benefit quantification are needed to determine financing needs and select measures aligned with government priorities.

### Public investment and public financial management implications
- Ongoing PIM and PFM reforms present an opportunity to mainstream climate considerations:
  - The PIM reform agenda has progressed since the adoption of a PIM strategy in 2017 and could integrate climate-specific elements in project appraisal and selection.
  - Strengthen relations between the Ministry of Environment, Ministry of Economy and Finance, and line ministries at the planning stage.
  - Mainstream climate concerns into comprehensive sectoral strategies.
  - Develop infrastructure maintenance strategies addressing climate-related impacts and resilience, building on road sector efforts.
  - Increase transparency on macroeconomic and fiscal information related to climate for planning and resource management.
  - Use budget classification reform to enable identification and tracking of spending on adaptation and mitigation, and to create an annex dedicated to climate change in the annual budget.

### Macroeconomic and fiscal modeling of climate impacts
- Current macroeconomic and fiscal projections do not factor in climate and climate change risks, despite frequent natural disasters.
- Under a climate-adjusted macroeconomic scenario that includes all humanitarian and reconstruction needs after a disaster, and assuming the government fully covers these needs, debt would quickly become unsustainable and exceed 85 percent of GDP by 2040.
- Natural disasters should be integrated into baseline projections rather than treated as downside risks; quantifying impacts based on historical experience and research-based assumptions supports timely policy responses.

### Policy package and sustainability
- A comprehensive policy package beyond adaptation investments is required to target SDGs and preserve fiscal sustainability:
  - Sole focus on investments risks inflating debt given low public investment efficiency.
  - Combining resilient investments (infrastructure and human capital) aimed at SDGs with revenue measures from mitigation and additional grants would increase potential growth and preserve fiscal sustainability.
  - Mobilizing grants and concessional financing should be accelerated, with effectiveness contingent on credible climate policies and improved governance and PIM capacity.

*Source: PREFACE, 1mdgea2022004*

### Box 1.  Summary of Key Recommendations 1/

### Box 1. Summary of Key Recommendations 1/

### National Climate Strategy Time Frame
- Time horizons: Short term: within 1-2 years; Medium term: within 2-5 years.
- Recommendations:
  - Complete developing and updating key elements of the strategic framework for climate change, including NDC — Short to medium term.
  - Mainstream climate change and disaster risk management into sectoral plans based on climate change projections and forward-looking hazard risk assessments — Medium term.

### Mitigation
- Fiscal and pricing measures:
  - Bring the fuel price to efficient levels by:
    - Normalizing excise taxes and other levies on diesel relative to gasoline — Short term.
    - Adopting an automatic fuel pricing mechanism that closes the gap between the true market price and the pump price — Short term.
  - Eliminating the VAT exemption on the first 100 kWh of electricity consumption — Short term.
  - Gradually increase electricity tariffs to cost recovery levels — Medium term.
  - Consider recycling tax revenue from mitigation measures to compensate for the adverse impact of price increases on the vulnerable by strengthening the social safety net — Medium term.
- Market and regulatory instruments:
  - Use feebates to encourage uptake of forest carbon storage, fuel-efficient vehicles, and energy efficient appliances — Medium term.

### Disaster Risk Management
- Recommendations:
  - Include an analysis of natural disaster risks in annual fiscal risk statements — Short term.
  - Finalize and publish a Disaster Risk Financing and Insurance (DRFI) strategy — Short term.
  - Expand the beneficiary registry to a wider social registry that includes vulnerable non-beneficiaries, to facilitate scaling-up social protection systems when needed — Medium term.

### Adaptation
- Planning and data:
  - Clearly articulate the relationship between adaptation and development throughout the adaptation planning process — Short term.
  - Enhance the National Adaptation Plan (NAP) by providing more data on climate change impacts and exploring financial resources — Short to medium term.
  - Continue with a sectoral approach to identify, assess, and prioritize adaptation measures, including by fully reflecting adaptation aspects in cost-benefit analysis — Short to medium term.
  - Include strategies that address the slow-moving effects of climate change in the NAP and the sectoral development plans — Medium term.

### Public Investment and Financial Management
- Integrating climate into PIM/PFM:
  - Define a methodology to integrate climate change into ex ante project assessments and introduce (e.g., through the ongoing work on the public investment manual) climate-related selection criteria for project prioritization — Short term.
  - Identify those infrastructure assets most vulnerable to climate change and define maintenance methodologies by main sectors — Short to medium term.
  - Strengthen climate-sensitive budgeting by introducing and gradually enriching a climate document appended to the annual budget, and by using the reform of the budget classification to pave the way towards climate budget tagging — Short to Medium term.

### Macro-Fiscal
- Policy and capacity building:
  - Prepare a well-costed and prioritized package of climate measures to anchor future reform efforts, based on existing or pending policy documents (e.g., NDC, NAP, PFM reform strategy) — Short term.
  - Develop the technical capacity to assess the impact of natural disasters on macroeconomic variables and include climate- and climate change-related shocks in all macroeconomic projections — Short to medium term.

### Strategic Assessment Questions (excerpt)
- Mitigation:
  - Does the NDC have well-specified mitigation targets?
  - Are current policies consistent with NDC?
  - Are current policies and financing consistent with moving towards net-zero emission by mid-century? (for a low income country)
- Adaptation:
  - Do adaptation plans in NDC and other climate plans adequately reflect climate shocks?
  - Are current policies consistent with NDC and other climate plans?
- Disaster Risk Management:
  - Are risk management policies adequately reflected in NDC and other climate plans?
  - Are current risk management policies adequate?
- National Process and PFM:
  - Are planning processes adequate to support NDC and other climate plans?
  - Is PIM adequate to support NDC and other climate plans?
  - Is Public Financial Management (PFM) adequate to support NDC and other climate plans?
- Macro-fiscal Implication:
  - To what extent were macroeconomic vulnerabilities eased by existing climate policies?
  - To what extent are climate projects in NDCs and other climate plans costed?
  - Are the financing plans adequate?
  - Are policies in the NDC and other climate plans consistent with debt sustainability?

### Climate Change Risks and Impact — Key Findings
- Hazard exposure and trends:
  - Madagascar is exposed to tropical cyclones, droughts, and floods, with more extreme heat events and more unpredictable rainfalls.
  - More frequent and intense extreme events have caused sizable losses in agriculture, ecosystems, water, and infrastructure.
  - Slow-moving effects of climate change add pressure to labor productivity across sectors.
- Socioeconomic vulnerability:
  - More than 80 percent of households live in poverty.
  - Services contribute 55 percent of GDP; primary industries (including agriculture, forestry, and fishery) contribute 25 percent of GDP.
  - The agriculture sector employs about 64 percent of the active population; subsistence farming is widespread.
  - Fishing contributes about 7 percent to the economy and exports, supporting about 5 percent of the population mostly in coastal areas.
- Sectoral impacts:
  - Agriculture: projected decrease in production due to loss of cultivable land, decline in land productivity from drought and pests, and loss of labor productivity from extreme heat.
  - Water and Sanitation: major cyclones between 2017 and 2019 caused serious pollution of drinking water and water-borne disease outbreaks; the 2021/2022 drought generated a massive food crisis and acute malnutrition in the Grand South.
  - Ecosystem: tree loss, biodiversity loss, declines in mangroves, marine vegetation, and fisheries; impacts on agriculture and tourism.
  - Infrastructure and Building: extreme events, landslides, drought and heat increase asset depreciation; expected annual damages to roads and railway assets are estimated around 0.2% of GDP under the current climate.
  - Energy: variability in precipitation may affect hydropower, which represents 29% of Madagascar’s energy mix.

### Madagascar: Climate Hazard Profile (selected statistics from Table 2)
- Temperature and heatwave:
  - Compared to the 30-year average (1960-1990), the average temperature between 1991-2020 increased by 0.53°C.
  - By mid-century, mean annual temperature increase likely to be between 0.87°C and 1.67°C in a moderate climate change scenario (SSP1-4.5).
  - The duration of warm spells is projected to increase by 70 days a year (median value) in a moderate climate change scenario (RCP 4.5).
  - Hot days: defined as temperature higher than 35 Celsius Degree.
  - Tropical nights: defined as minimum temperature higher than 20 Celsius Degree.
  - Warm spell duration index base period: 1979-2009.
- Precipitation and drought:
  - Annual rainfall has been steadily declining in the central and east coastal regions between 1961 and 2005; winter and spring rainfall reductions observed in most parts of the country.
  - By mid-century, annual precipitation is not expected to decline much but the maximum number of consecutive dry days is expected to rise significantly in some drought-prone areas under RCP 4.5.
  - Recent observation: The 2021/2022 drought has been happening across much of the country; the south has been experiencing 70 percent less than normal rainfall.
- Flooding:
  - By mid-century under RCP 4.5 Madagascar is expected to have higher intensity of single-day rainfall compared to the 1961-2015 baseline, increasing hydrological flooding risk.
  - Recent observation: The flood in Jan 2022 caused landslides, destruction of infrastructure and loss of life, affecting Antananarivo and other areas of Analamanga Region.
- Cyclone:
  - Projected frequency subject to uncertainty, but intensity is likely to increase.
  - Recent observation: Major tropical cyclones have hit Madagascar an average of about once every three years since 1998. In January and February 2022, four cyclones hit Madagascar, two of which are category 3-4 events.

### National Strategy for Climate Change — Status and Gaps
- Existing framework:
  - National development plan (Initiative Emergence Madagascar (IEM) 2019-2023) includes a pillar on environmental promotion and sustainable development and touches on mitigation areas.
  - The 2016 NDC covered mitigation and adaptation; provides a global cost estimate but individual measures are not detailed or costed.
  - The National Adaptation Plan (2019) contains a list of defined and costed projects and activities.
- Implementation gaps:
  - Climate change has not been mainstreamed into sectoral strategies; sectoral documents exist in some sectors (energy, waste, water, industry, forestry/REDD+) but are mostly not detailed to measures or costed.
  - Assessment of future climate implications at sectoral level is largely limited to current situations and recent events and lacks a forward-looking stance.
  - Cross-cutting norms (e.g., climate resilient construction of buildings and roads) address prevailing effects but climate projections and hazard risk assessments are still being developed.
  - COVID-19 pandemic and recent droughts and cyclones have distracted from systematic implementation and investment in climate change policies.
- Ongoing updates:
  - The process for updating key elements of the national framework, including the National Policy for Fighting Climate Change and the NDC, started in 2021 and is expected to be concluded in the first half of 2022.
  - Drafts were not shared with the mission; the policy aims include strengthening links between national development and adaptation, fostering mitigation implementation, integrating climate change at all levels of government, promoting research and technology, and supporting sustainable financing.

*Source: 1mdgea2022004 - Box 1.  Summary of Key Recommendations 1/*

### 7. Key climate change related strategies, including the National Policy for Fighting

### 7. Key climate change related strategies, including the National Policy for Fighting Climate Change

### Strategic orientation and institutional framework
- Climate change strategies, including sectoral strategies and the National Policy for Fighting Climate Change and NDC, explicitly aim to take the country’s development needs into account; links between climate measures and development objectives are mainly developed at the general level.
- Disaster risk and response framework:
  - National Risk and Disaster Management Strategy (2016) covers risk reduction, communication, planning, and skills training at provincial, local and community level.
  - Institutional framework: emergency prevention and management unit (CPGU) at the strategic level and national disaster and risk management office (BNGRC) at the operational level.
  - A National Disaster Risk Financing and Insurance Strategy is currently under preparation.

### Mitigation ambitions, emissions baseline, and costs
- Madagascar’s net GHG emissions in 2020: 49 mt CO2e (0.1 percent of global GHG emissions) or 0.13 metric tons per capita.
- Comparative benchmark: other Sub-Saharan African Countries (excluding high income) at 0.76 metric tons per capita.
- 2016 NDC mitigation pledge (summary):
  - Reduce GHG emissions sources by 14 percent compared to BAU by 2030.
  - Increase GHG absorption of sinks by 32 percent by 2030, remaining a net absorber of emissions.
- Reported mitigation cost and government contribution:
  - Madagascar estimates mitigation costs at USD 6.37 billion between 2015-30, with the government aiming to contribute about 4 percent of the indicated costs.
- Forestry and land-use context:
  - 2010 tree cover: 16.4Mha, covering 28 percent of land area.
  - 2010–2020: lost 241,000 ha of tree cover by 2020, equivalent to 125 mt CO₂ emissions (Global Forest Watch).
  - Natural forest coverage fell from 9.4Mha in 2005 to 9.2Mha in 2010, a reduction of almost 40,000 ha per year (Madagascar Energy Policy Letter 2015 – 2030).
  - Reforestation potential: reforestation of 270,000 ha can absorb up to 4 mt of CO2e emissions per year, or about 8 percent of Madagascar's total emissions in 2020.
- Accounting uncertainty:
  - Madagascar considers itself a carbon sink, but external datasets that include international estimates of LULUCF emissions suggest Madagascar is a net emissions source country; treatment of LULUCF varies substantially by methodology.

### Sectoral mitigation objectives and key actions (Table summary)
- Energy:
  - Mitigation objective: Reduce emissions by 13 percent in 2030 compared to BAU.
  - Proposed actions:
    - Increase renewable energy share from 35 percent to 79 percent.
    - Dissemination of improved stoves, with 50 percent of households adopting improved stoves in 2030.
    - Improving energy efficiency.
- Waste:
  - Mitigation objective: Reduce emissions by 85 percent in 2030 compared to BAU.
  - Proposed action: Composting for 50 percent of household organic waste in large cities.
- Agriculture:
  - Mitigation objective: Reduce emissions by 10 percent in 2030 compared to BAU.
  - Proposed actions:
    - Dissemination of improved rice cultivation systems.
    - Promotion of arboriculture (5,000 ha annually).
- LULUCF:
  - Mitigation objective: Increase absorption by 32 percent in 2030 compared to BAU.
  - Proposed actions:
    - Reforestation in indigenous species by 270,000 ha.
    - Integration of the REDD+ mechanism.
    - Reduction of wood removal from forests.

### Key risks to mitigation goals
- Rapid deforestation and land degradation from:
  - Small-scale agriculture (tavy/slash-and-burn), energy production (firewood and charcoal), illicit logging, mining, and livestock practices.
  - Weak governance, population growth, widespread poverty, uncertain land tenure, and economic instability exacerbate unsustainable harvesting.
- Governance shortcomings in forestry (Box 2) and recommended institutional responses include international mechanisms and transparency tools (examples presented in source: FLEGT/VPAs, Freedom of Information laws, Feedback and grievance redress mechanisms), noting mixed evidence on effectiveness.

### Energy sector decarbonization plans and constraints
- Petroleum and imports:
  - Madagascar imports all refined petroleum products.
  - In 2018, petroleum products met 12 percent of energy needs (IRENA 2020).
  - Total petroleum-product imports account for more than 20 percent of the value of total imports.
  - JIRAMA consumes around 75 percent of imported heavy fuel oil and some diesel for power generation.
  - JIRAMA’s oil bill: US$150 million in 2014 (an increase of more than 100 percent from 2009).
- Biomass dependence and health co-benefits:
  - Over 80 percent of total primary energy supply is based on biomass; 85 percent of that is firewood and the rest charcoal and other biomass.
  - Government target: 50 percent of households adopt cleaner cookstoves by 2030.
  - Health benefit estimate: saving about 12,000 lives annually by 2030 if cleaner cookstoves adoption target met.
  - World Bank estimation: fully replacing firewood and charcoal in cooking with ethanol would avoid degradation of roughly 1.4Mha of unmanaged forests (~10 percent of Madagascar’s forest area) and reduce emissions by around 663 million tons of CO2e over a 30-year period.
- Electricity access and renewable expansion:
  - Electricity access as of 2018: 25 percent of the population.
  - Government target: increase electricity access to 70 percent by 2030, primarily by expanding hydro and solar.
  - Renewable energy share:
    - Around 38 percent in 2020.
    - Government aim: about 80 percent by 2030.
    - Achieving 80 percent RE by 2030 estimated to reduce Madagascar’s energy emissions by around 18 percent relative to baseline in 2030.
  - Investment needs:
    - Reaching 80 percent renewables by 2030 estimated to require investments of around 0.4 percent of GDP on average annually over the next decade.
    - More than 240 MW of hydropower installations are in the pipeline over the next 4 years.
    - To meet the NDC target of 80 percent RE by 2030, around 900 MW of new hydropower capacity will be required — gap of 660 MW based on current pipeline.
- Investment incentives and tax policy:
  - Tax concessions to encourage private RE investment include exemption from customs duties and VAT on imported equipment for RE and tax reduction of 50 percent of the investment made by private companies investing in RE.
  - The 50 percent investment tax credit (ITC) for RE is currently bundled with similar credits for other industries (Article 01.01.14C in the tax code).
  - Concerns: current ITC allows full depreciation on top of the tax credit, potentially yielding a negative effective marginal tax rate and costly foregone revenue.
  - Policy options noted: reduce the base for depreciation by amount of ITC, lower ITC rate well below 50 percent, or replace the ITC with an investment allowance while ensuring depreciation base is net of the allowance.
  - Structural reforms to complement tax concessions: improve infrastructure quality, human capital, and governance to attract green energy investments.
  - Decentralized Rural Electrification Program underway; off-grid solar kits increasingly prevalent in rural areas; government studying VAT and customs duty exemptions on solar kits.
- Energy emissions baseline and comparators:
  - CPAT baseline: energy emissions of about 5 mt CO2e for the energy sector in 2030.
  - 2015 NDC projected BAU energy emissions of 13 mt CO2e by 2030 (noted as significantly higher than CPAT baseline and making the 2015 NDC energy component non-binding under CPAT assumptions).

### Waste, agriculture, and other sectoral measures
- Waste sector: National Climate Change Strategy for waste focuses on strengthening institutional capacity for waste management.
- Madagascar participates in World Bank’s Forest Carbon Partnership Facility (FCPF) to reduce carbon emissions from deforestation and forest degradation between 2020 and 2024.
  - Under a US$50 million Emission Reductions Payment Agreement (ERPA) aligned with the National REDD+ strategy, Madagascar is expected to reduce 10 million tons CO2 emissions through improved and sustainable agriculture, as well as energy supply (mainly wood charcoal).
- Replacing biomass cooking with efficient stoves and cleaner fuels yields welfare and environmental benefits; excise and VAT exemptions exist for combustible ethanol and reduced VAT of 5 percent on butane gas for households from 2020, though VAT exemptions create inefficiencies and revenue losses and targeted direct subsidies are suggested as more cost-effective for low-income households. Developing domestic production capacity for stoves and fuels is recommended to improve accessibility and affordability.

### Carbon pricing, fuel taxation, and subsidies
- Petroleum tax (TPP) rates collected by customs administration:
  - Gasoline: MGA 503 per liter.
  - Diesel: MGA 228 per liter.
  - Lamp oil: MGA 10 per liter.
- VAT on fuels:
  - Gasoline and diesel: subject to VAT at 20 percent.
  - Lamp oil: VAT exemption.
- Additional charges applied to retail consumption of gasoline and diesel:
  - Road Maintenance Fund: MGA 288 per liter.
  - Environmental Fund: MGA 134 per liter.
- Assessment: Madagascar’s tax system delivers appropriate carbon pricing for road fuels, but VAT exemptions and other fuel-specific tax treatments can create distributional and efficiency concerns.

*Source: IMF staff (content extracted from "1mdgea2022004 - 7. Key climate change related strategies, including the National Policy for Fighting Climate Change").*

### 20. Fuel prices are below their socially optimal levels in Madagascar for diesel,

### 20. Fuel prices are below their socially optimal levels in Madagascar for diesel, kerosene, and LPG, but not for gasoline.

### Findings on fuel pricing and efficient (socially optimal) prices
- Socially optimal prices = retail prices that reflect full costs to supply fuel plus negative externalities (supply costs + ‘Pigouvian’ taxes = efficient price).
- As of 2021:
  - Gasoline: level of excise and levy taxation in Madagascar raised prices to around its efficient level; US 15 cents/liter reflects the cost of GHG emissions priced at USD 64/tCO2e.
  - Diesel, kerosene, and LPG: taxation was below the efficient level.
  - Note: kerosene and LPG estimates exclude positive externalities from shifting people away from fuelwood, so the implied gap may be smaller.
- With the recent surge in global fuel prices and incomplete pass-through to consumers, the gap between pump prices and efficient prices is likely higher now for all fuels.

### Problem with current pricing mechanism and fiscal risks
- Pump prices are not adjusted to keep pace with rising international prices, producing:
  - Distortionary price signals.
  - A growing government liability due to delays in fulfilling obligations to fuel distributors.
- Government roadmap:
  - Prepare to implement an automatic fuel pricing mechanism by 2024Q1 to eliminate fiscal risks from fuel price controls.
  - Implementing this mechanism would help ensure pump prices are fully cost-reflective and efficient, complementing mitigation ambitions.

### Vehicle taxation and emissions-related reforms
- Current taxes on passenger motor vehicles:
  - New vehicles: 5 percent excise tax based on CIF value and 20 percent VAT declared at customs.
  - Used vehicles: 10 percent excise tax based on CIF value and an additional 20 percent VAT.
- Reforms suggested:
  - Rebalance excise structure to reflect emissions intensity and vehicle age (e.g., higher excise for diesel vehicles relative to petrol vehicles).
  - Penalize older and larger vehicles with higher social costs (local air pollution, road fatality/injury risks).
  - Introduce a rebate component toward more efficient vehicles (feebate).

### Feebates and other carbon-pricing strategies
- Revenue-neutral feebates (‘bonus-malus’):
  - Aim: encourage take-up of fuel-efficient cars by taxing less efficient vehicles and subsidizing more efficient ones relative to a pivot (historical average or standard).
  - Example: With an emissions-rate based carbon price of US$3 per g/CO2e and a pivot at 95g-CO2e/km:
    - Tax of US$2,625 on a vehicle emitting 200g-CO2e/km.
    - Subsidy of around US$1,125 for a vehicle emitting 50g-CO2e/km.
    - No tax or subsidy on a car at 95g-CO2e/km.
  - Administrative costs: expected to be modest because feebates can build off the existing excise tax; subsidy can be deducted from excise duty; data on fuel economy of imported used and new vehicles is readily available.
  - Caveat: other obstacles to EV uptake exist (lack of reliable electricity access; high cost of EVs).
- Feebates for appliances and production:
  - Can encourage energy-efficient cookstoves, light bulbs, electrical appliances, high-efficiency carbonizer machines for charcoal makers, and improved rice-cultivation systems.
  - Should be complemented with energy efficiency labelling and minimum energy performance standards.
- Feebates for forest carbon storage (Box 3 concept):
  - Fees and rebates applied to landowners per ton of stored carbon relative to a baseline.
  - Basic formula (as presented): Y_t = τ_t^CARBON * (CAR_t^i − CAR_t,BASE^i)
    - CAR_t^i = tons of stored carbon on property for landowner i at time t.
    - CAR_t,BASE^i = baseline level of carbon storage attributed to that landowner at time t.
    - τ_t^CARBON = payment per ton of stored carbon.
    - Y_t = landowner total payment at time t (or subsidy if Y_t < 0).
  - Key design issues:
    - Baselines could be based on REDD+ inventories; payments could be annual or biennial; coverage could focus on agricultural/forestry border areas.
    - Critical to ensure property rights are well defined; only 25 percent of total land surface of Madagascar is registered (mostly urban areas).
    - Payments should be annual (‘rental’) rather than large upfront payments to address non-permanence of changes.
    - CO2 price should align with forestry emissions objectives in Madagascar’s NDC and be phased in gradually.
    - Exemptions: timber harvested for wood products could be exempt due to delayed carbon emissions.

### Electricity tariffs, VAT exemption, and efficiency
- Electricity tariffs have not matched cost-recovery prices:
  - Average cost of electricity nearly doubled between 2015 and 2019 due to rising dependence on expensive thermal energy and operational inefficiencies.
  - Average tariffs remained stagnant, creating large transfers from the government to bridge the gap.
  - VAT exemption for the first 100 kWh per month for all consumers contributes to pricing inefficiency.
  - Government retains a social tariff for vulnerable households; eliminating the VAT exemption and underpricing of average tariffs would likely not have adverse distributional implications because poor households generally have limited access to grid electricity in Madagascar.
- World Bank estimate:
  - Without tariff increases, JIRAMA would continue to incur losses and require government transfers even after significant efficiency improvements.

### Recommended tax package and fiscal impact
- Sample policy package proposals:
  - Raise the diesel excise by 275 MGA/liter.
  - Raise the road maintenance levy by 154 MGA/liter.
  - Eliminate the VAT exemption for the first 100 kWh of monthly household electricity use by 2023.
  - Adopt an automatic fuel pricing mechanism by 2024 to erase gap between market prices and pump prices for gasoline and diesel.
  - Increase average electricity tariffs by 15 percent annually in line with JIRAMA’s recovery plan.
    - Raising average tariff from $0.15 per kWh to $0.22 per kWh by 2025, with a social tariff retained for vulnerable households.
- Fiscal and mitigation impacts:
  - These measures would raise up to 1.7 percent of GDP in revenue by 2025.
  - Could contribute to energy CO2 reductions of up to 5 percent relative to baseline in 2030.
  - Note: estimated fiscal impacts modelled under the BAU scenario; attainment of NDC RE share would lower tariff increases needed for cost recovery.

### Distributional impacts and recommended revenue recycling
- Price increases projected:
  - Normalizing excise taxes on gasoline and diesel, moving to automatic fuel pricing, and raising electricity tariffs would increase prices by 10, 44 and 36 percent, respectively.
- Household impacts:
  - Direct burden from higher energy prices = 1.6 percent of household consumption, borne mostly by the richest households.
  - Indirect costs (higher transportation costs increasing food prices) burden bottom four quintiles relatively more.
  - Overall consumption changes:
    - Poorest quintile: loss of 1.1 percent of consumption.
    - Richest quintile: loss of 2.4 percent of consumption.
- Compensation and revenue recycling:
  - Only 10 percent of the 1.7 percent of GDP in raised revenues would be needed to fully compensate the bottom 4 quintiles for their welfare loss in 2025.
  - The remaining 1.5 percent of GDP could support other development goals: improving access to electricity, strengthening social spending, investing in resilience.
  - Example compensation scenarios (using revenues from the mitigation policies excluding potential savings from automatic fuel adjustment, about 1.2 percent of GDP):
    - Universal lump sum transfer, expanding Human Development Cash Transfer (HDCT) to a universal grant per child below age 5, a grant per child below age 15, or providing a non-contributory old-age pension would all reduce inequality and poverty.
    - Grant per child below age 5: reduces Gini from .405 to .381 and reduces poverty rate by 1 percent.
  - Distributional outcomes:
    - Universal lump sum or grant per child below age 5 benefits most households across income quintiles.
    - Grant per child below age 15 benefits households in the bottom three quintiles; top quintile would experience a maximum loss of around 2 percent of pre-tax consumption under that scheme.
    - Grant per child below age 5 raises consumption by 14 percent for the poorest quintile and by 8 percent for the second poorest.
- Policy synergies:
  - Social programs funded by revenues can reduce reliance on forest resources, aiding mitigation and adaptation (resilience to cyclones and natural disasters).

### Disaster risk and fiscal exposure (relevant fiscal context)
- Annual average fiscal costs of natural disasters estimated at about 1 percent of GDP:
  - Catastrophe probabilistic models (SWIO RAFI) suggest annual average losses (AALs) of over $100mn or about 1 percent of GDP from tropical cyclone, floods, and earthquakes.
  - SWIO RAFI estimates a 10 percent probability that damages could exceed $230mn in any given year.
  - Historical extreme cyclones (Chedza/Enawo/Ava) produced damages ranging between 1-4 percent of GDP.
- Fiscal risk disclosure and assessment:
  - In-house quantifiable assessments of disaster risks are not available.
  - Disaster risks are not systematically disclosed in budget documents; annual fiscal risk statements (FRSs) have only occasionally discussed natural disaster risks.
  - Recommendation: include analysis of disaster risks in annual FRSs to better manage fiscal risks and encourage adequate provisioning.

*Source: IMF staff analysis (chapter content).*

### 33. Madagascar has fiscal mechanisms in place to respond to shocks including from

### 33. Madagascar has fiscal mechanisms in place to respond to shocks including from natural disasters

### Fiscal mechanisms and buffers
- Budgetary contingency
  - The Organic Budget Law provides for an annual budget line to self-insure against unforeseen expenses, including those from disasters, but this line is not activated in practice.
- Advance decrees and budget re-allocations
  - Additional funds can be approved through advance decrees, up to 1 percent of the approved budget, without provoking a supplementary budget.
  - Intra- and inter-ministerial re-allocations can be made, up to 10 percent of approved appropriations, after the government notifies the National Assembly including in a written report.
  - Advance decrees have often been used in response to disasters but are constrained by the need to find offsetting revenue or expenditure measures; in practice, new funds often arrive from donors or insurance mechanisms.
- Imprest accounts
  - Renewable imprests: up to 15mn MGA for urgent minor operating payments.
  - Exceptional single imprests: up to amount of open credit lines for one-off larger payments, subject to ex-post audit.
  - Several “régies d’avance” were utilized in 2020 to pay for emergency expenditures in response to tropical rainfalls and Covid.
- Supplementary budgets
  - Additional spending exceeding the above may be approved by the legislative assembly as supplementary budgets; used after the 2017 cyclone Enawo and the 2020-21 Covid episodes, although the process can take several months.
- National Disaster Fund (NDF)
  - Annual transfers from the budget of 2bn MGA or around 0.003 percent of GDP recently put in place.
  - The NDF is extra-budgetary, to be managed by the BNGRC and intended as a basket fund for external donors.
  - A first best would have been to integrate the NDF into the budget; strong PFM systems with clear accumulation and utilization rules and publishing the NDF’s operations along with official budget documents would help reduce fiscal risks and improve transparency.
- FX reserves as buffer
  - As of end-2021, FX reserves reached $2.2bn or about 6 months of imports. These buffers could be used to respond to shocks including from disasters if needed.

### Social protection and shock-responsive systems
- Low social spending and program characteristics
  - Social spending is around 0.3 percent of GDP.
  - Health insurance covers around 15 percent of the population (mostly private).
  - Pension assets are less than 2 percent of GDP; public system covers 3 percent of the workforce and private system covers 10 percent.
  - No unemployment insurance.
- Cash transfer programs
  - Main programs: Human Development Cash Transfer (HDCT) and Productive Safety Net Program (PSNP).
  - Current coverage around 2.5 million beneficiaries and a target of 4 million beneficiaries (or 20 percent of the population under the poverty line).
  - Payments are around $100/household/year or about 20 percent of the consumption basket of the poor; payments increasingly made via mobile money services.
  - Beneficiary identification: initially community-based pre-selection, geographic criteria and proxy-means-testing; an integrated beneficiary registry became operational in 2020.
- Shock-responsive features and constraints
  - Safety nets can be scaled-up vertically and horizontally following the government’s 2020 manual on shock-responsive systems.
  - Cash and in-kind transfers were scaled-up in response to the El Niño drought in 2016, cyclone Enawo in 2017 and cyclone Belna in 2019.
  - Expansion constrained by availability of additional external financing and limited own-budget allocations.
  - Recommendation: expand the beneficiary registry to a wider social registry; the envisaged centralized digital identification database supported by a WB project would help verify identities and reduce misappropriations of public funds.

### Risk transfer and contingent financing
- African Risk Capacity (ARC)
  - Madagascar purchased drought insurance in 2019, with coverage up to $2.5mn and annual insurance premia of around $0.5mn/year mostly financed through AfDB-funded ADRiFi through 2022; Government pays full premium in 2023.
  - Madagascar received a $2.1mn payout in July 2020 in response to extreme drought in the south.
  - Madagascar purchased ARC’s tropical cyclone parametric insurance with an annual premium of $2mn fully financed through ADRiFi for the 2021-22 cyclone season; received a $10.7mn payout in March 2022 in response to TC Batsirai.
- World Bank and other ex-ante instruments
  - WB Cat-DDO: country limit for IDA clients is $250mn.
    - The 2019 DPO: joint operation by WB for $50mn and AFD for €25mn; WB and AFD disbursed $15mn and €3mn, respectively, in response to the 2020 floods.
  - WB Immediate Response Mechanism (IRM) and CERCs
    - IRM can mobilize up to 5 percent of undisbursed IDA projects to emergency response, subject to declaration of emergency and preliminary damage assessment; Madagascar mobilized $13mn through IRM after cyclone Enawo in 2017.
    - CERCs allow re-allocation of undisbursed IPF funds to emergency response; example: $12mn CERC activated in 2021 for drought in Southern Madagascar.
    - In response to Covid, several CERCs were activated for a total of $123mn for the Multisectoral Emergency Plan.
    - Currently, 13 out of 21 ongoing WB IPFs have CERCs embedded, with an undisbursed total amount of around $1bn.
- Ex-post instruments from development partners
  - IMF emergency financing:
    - PRGF augmentation of $32mn after the 2000 cyclones.
    - ECF augmentation of $42.4mn after the 2017 cyclone Enawo.
    - In response to Covid, Madagascar received $338mn (or 100 percent of quota) under two 2020 RCFs and benefited from debt relief under CCRT tranches for a total of $30mn over 2020-22.
    - Madagascar can have increased access to RCF/RFI resources under the large natural disaster window in case of severe disaster shocks.
  - WB Crisis Response Window (CRW):
    - Madagascar accessed $20mn after the 2016 El Niño drought and $50mn in 2021 to address the food security crisis in Southern Madagascar.
    - Discussions ongoing for a $250mn CRW in response to the 2022 cyclones.
  - AfDB:
    - Madagascar received about $30mn in emergency humanitarian aid following previous disasters; latest was an AfDB 2020 grant of $0.7mn in response to drought and Covid in the South.

### Insurance of public assets and private market coverage
- Public assets
  - Public buildings and infrastructure assets are not insured; the government acts as its own insurer.
  - New insurance law pending implementation decrees: construction insurance to become mandatory (excluding the state) to cover damages from defective material or construction, but with no requirement to insure against disaster damages.
  - Recommendation: adhere to building code and develop a centralized registry of public assets to incentivize larger consolidated policies and facilitate post-disaster damage assessments.
- Private insurance market and penetration
  - Five licensed insurance companies; two large SOEs hold more than 70 percent of market share.
  - As of end-2018, total assets correspond to about 2 percent of GDP with premiums at about 0.7 per cent of GDP.
  - Access to insurance services low at 3 percent of the adult population.
  - Cyclone and crop insurance are limited but available; cyclone insurance available as optional extension on an indemnity basis.
  - Index-based agricultural insurance piloted with payouts dependent on drought index thresholds and capital insured.
  - Transfer of supervisory role of insurance companies from MEF to CSBF noted as welcome.

### Risk layering and DRFI strategy
- Current assessment
  - External contingent financing and insurance typically available only after extreme events; domestic resources used for smaller events.
  - Figure 13 assessment (based on WB Risk Layering Framework) indicates some instruments are absent/weak (red), partially available/utilized (yellow), and others available/well utilized (green).
- Recommendations
  - Finalize an overarching National Disaster Risk Financing and Insurance (DRFI) Strategy (currently under preparation with AfDB-ADRiFi technical assistance) to coordinate instruments, provide centralized guidance on disaster risk assessments, and outline financing and insurance instruments in a risk layering framework.

### Financial sector preparedness and digital financial services
- Financial sector vulnerabilities
  - Large disaster shocks can pose risks to financial stability due to credit concentration in disaster-exposed sectors such as commerce, transportation, manufacturing and agriculture.
  - Natural disaster stress tests under the 2015 FSAP suggested banks may incur losses up to 40 percent of their 2014-15 net income.
  - Recommendations from FSAP: tighten prudential limits on large exposures and supervision of concentration risks; create a deposit insurance scheme to protect small depositors.
- Central Bank and regulatory measures
  - In response to Covid, the CBM temporarily relaxed reserve requirement rules, encouraged loan repayment deferrals, and extended credit lines to SMEs at reduced interest rates; CSBF simplified customer due diligence for mobile accounts. These measures were later unwound but similar temporary measures could be introduced in response to large natural disaster shocks.
- Role of digital financial services
  - Mobile money outlets have much higher penetration compared to traditional financial services; mobile payments promoted financial inclusion and efficiency and were increasingly used to deliver cash transfers during Covid.

### Adaptation planning and measures (overview)
- Limitations in adaptation strategy
  - Adaptation strategy hindered by lack of sufficient information, analysis, and integration with development strategies.
  - Investments often determined by fragmented external financing initiatives without overall prioritization framework.
- Recommended planning enhancements
  - Recognize and articulate the relationship between adaptation and development strategies in the NAP and sectoral plans.
  - Base sectoral policies on systematic climate impact analysis and develop methodologies to quantify costs and benefits of potential adaptation measures.
  - Broaden current adaptation analysis (which mainly addresses current extreme events) to include forward-looking climate risks and slow-moving effects such as steadily increasing temperature and rainfall variation, and their impacts on agriculture, fisheries, tourism, public health, and energy.
  - Improve knowledge of donor landscape and capacity to meet financing requirements; facilitate access to private sector finance by enabling risk assessment and transparency of the policy environment.
- Priority adaptation sectors
  - Agriculture, water, ecosystems, and infrastructure identified as sectors where adaptation measures would provide greater resilience and additional economic and environmental benefits; complement with informational, regulatory, and operational measures with minimal fiscal cost.

*Source: IMF staff assessment (content unit: 1mdgea2022004 - 33. Madagascar has fiscal mechanisms in place to respond to shocks including from)*

### 46. Adaptation measures should shield agricultural productivity from erosion by

### 46. Adaptation measures should shield agricultural productivity from erosion by climate change.

### Agriculture: impacts, gaps, and priority investments
- Main impact channels:
  - Short run: erosion of capital inputs (floods and tropical cyclones damage agricultural infrastructure and land).
  - Long run: loss of productivity (drought and extreme heat reduce productivity; steadily increasing temperatures, more extreme rainfall episodes, and more dry days accelerate productivity decline).
- Observed loss:
  - Productivity has already fallen by USD 95 per worker over the past 20 years (World Bank, 2022).
- Current plans and costs:
  - The NAP and the sectoral investment plan prioritize agricultural research and disaster-risk reduction systems for farmers.
  - These investments cost up to 157 million USD, or 1.1 percent of nominal GDP in 2019 annually.
- Gaps:
  - Agricultural plans largely focus on disaster emergency responses rather than pre-disaster resilience.
  - Strategies to address slow-moving effects of climate change are missing.
  - Lack of granular risk mappings for existing agricultural infrastructure limits targeted investments and efficient allocation of financial resources.
  - Only a small subset of proposed adaptation measures adopted (e.g., a water pipeline project in the south with an irrigation component and some research in climate-resilient crop varieties).
- Potential gains from scaling investments:
  - A 2019 IFPRI study: an additional US$1.97 billion per year in agricultural research and development aimed at climate resilience in emerging and developing economies could:
    - improve food supply by six to seven percentage points between 2030 and 2050;
    - reduce hunger by 21-24 percentage points.
- Other recommended investments:
  - Expand irrigation systems and improve water use efficiency, rainwater harvesting, soil water holding capacity, and resilience of transport and electricity infrastructure.

### Water and Sanitation: vulnerabilities, recommended actions, and costs
- Vulnerabilities:
  - Water and sanitation infrastructure is poor and outdated (Weiskopf et al. 2021) and would not withstand more intense floods and storms.
  - Stormwater overflow contaminates drinking water, harming human health.
- NAP recommendations and costs:
  - NAP recommends expanding and climate-proofing water infrastructure.
  - Together with other strategic recommendations, these adaptation measures are expected to cost $25 million over ten years, according to the NAP. (No specific estimation for resilience-enhanced water infrastructure alone.)
- Implementation gap:
  - Reconstruction and rehabilitation plans have not incorporated resilience; resilience to floods and tropical cyclones has not been built into construction, rehabilitation, and reconstruction processes.
- Cost-effectiveness:
  - In Low-middle Income Countries (LMICs), resilient investments could increase investment costs by 1.1 or 2.2 percentage points, implying an additional investment up to $16.5 million annually for the next 10 years in Madagascar for resilience-building.
  - Such investments would halve the disaster damages from climate change (Hallegatte, et al., 2019).

### Ecosystem: restore and grow natural capital
- Impacts:
  - Slow-moving climate change has reduced volume and productivity of natural capital in terrestrial and marine ecosystems (tree growth/reproduction affected; coral bleaching, seagrass loss, fisheries loss, ocean acidification).
  - These changes have directly harmed agriculture, fisheries, and tourism.
- NAP actions and costs:
  - NAP highlights restoration and creation of new conservation zones, coral reef and mangrove conservation, and fisheries research.
  - NAP’s ecosystem resilience strategy embodied in two national programs costing in total USD 90 million over 10 years.
- Returns:
  - Benefits of creating and managing protected areas average from USD35 to USD53 per hectare each year (Cooke et al., 2021).
  - Management costs are typically 20 percent of those benefits.
  - Cost of creating new protected areas estimated in the range of USD 27 and USD 101 per hectare; such costs could be recovered economically in about three years.
- Note:
  - It is virtually impossible to isolate the incremental cost of adaptation in the ecosystem sector, but investments in restoration and growth of natural capital are themselves good adaptation strategies.

### Infrastructure and buildings: grey and ecosystem-based interventions
- Recommended approach:
  - For long-lived assets, combine grey (retrofitting, upgraded standards) and ecosystem-based interventions to reduce vulnerability.
  - Coastal cities: address coastal flooding, storm surge, saltwater intrusion, aggravated by coastal erosion and mangrove loss.
  - Inland cities: address wetland losses and soil erosion.
- Current programmatic costs:
  - Current adaptation plans total programmatic costs range from USD110 million – USD 170 million annually, or 1% of nominal GDP in 2019.
  - Main investments: hydropower infrastructure (reservoir capacity enhancement, powerhouse relocation, sediment control) and transport infrastructure (road surface enhancement, wind-proofing bridges, measures against landslides).
  - Resilience standards for road infrastructure and building codes exist but rollout is constrained by local capacity.
- Cost-benefit evidence:
  - Resilient transport infrastructure in LMICs is on average 5 percent more costly than standard infrastructure but can reduce damage from disasters by 50 percent.
  - Application to the four tropical storms in 2022:
    - Total damage in the infrastructure sector (mainly roads and railways) likely around USD170 million (GRADE, February 2022).
    - If roads and railways had been more resilient, damage would have been reduced by USD85 million.
    - Adaptation investment could cost between USD 25million and USD 115million (depending on cost assumptions — LMIC average 5 percent; Madagascar has assumed 25 percent).
    - Under these assumptions, adaptation may have paid off from one year’s climate events alone.
- Specific investment examples and metrics (Table 4 excerpts):
  - Electricity transmission and distribution:
    - Construction with higher resilience standard; regular power system maintenance.
    - Cost increase: 30% in LMICs.
    - Benefit: Reduction of damage risk by 50% or 60% for new infrastructure assets.
  - Hydropower:
    - Measures: conveyance efficiency, reservoir storage, turbine capacity.
    - Cost increase example: 2.7%, total cost increase of the seven hydropower projects by the Program for Infrastructure Development in Africa to enhance resilience.
    - Benefit/cost ratio: 4.03 considering the seven projects altogether.
  - Transport:
    - Construction with higher resilience standard; technical and low-cost maintenance.
    - Cost increase: 5.5% in LMICs.
    - Benefit: reduce damage to new infrastructure by 50%.

### Broader economic and environmental benefits of adaptation investments
- Adaptation investments generate economic, development, and environmental benefits that accrue even without climate change.
- Quantification challenges:
  - Environmental benefits are difficult to quantify but matter for project finance and monitoring (GCA, 2019; Heubaum et al., 2022).
- Evidence from case studies (Table 5 highlights):
  - Benefit-cost ratios (disaster damage reduction only) and full triple-dividend results show substantially higher returns when economic and environmental co-benefits are included:
    - Kunshan, China (Forestation initiatives): disaster-only BCR 5.7 to 1; economic/environment BCR 43.8 to 1; total BCR 49.5 to 1.
    - Princes Park, Australia (Coastal protection): disaster-only BCR 0.6 to 1; economic/environment BCR 1.3 to 1; total BCR 1.9 to 1.
    - Felixstowe, UK (Water-saving irrigation): disaster-only BCR 15.1 to 1; economic/environment BCR 16.7 to 1; total BCR 31.8 to 1.
    - Ningxia, China (Water-saving irrigation): disaster-only BCR 0.06 to 1; economic/environment BCR 5.6 to 1; total BCR 5.6 to 1.
  - These studies show benefit-cost ratios 1.1 to 93 times higher than from damage reduction alone (Heubaum et al., 2022).

### Non-fiscal public programs and regulatory measures
- Existing non-fiscal measures in Madagascar include risk zoning, community-based early warning, disaster response protocols, weather forecasts and climate services used to revise agriculture calendars seasonally, and emergency response training in school curricula.
- Recommendations to strengthen non-fiscal measures:
  - Accelerate adoption of resilience standards for building and infrastructure (standards developed but not fully rolled out due to local capacity constraints).
  - Develop operational guidelines to facilitate private sector adaptation (tourism, conservation, adoption of innovations such as sea surface temperature control, architecture for heat insulation).
- Table 6 (summary):
  - For tropical cyclone and flooding: planning (risk zoning), regulatory (building code, infrastructure standards), educational/outreach (community-based early warning), operational guidance (drainage design manual) — implemented by CPGU, MAEP, DGM, MID, MTP, BNGRC, MSANP.
  - For drought: planning (water resource mapping, water level monitoring), regulatory (conservation measures) — some measures adopted (e.g., weather forecasting by DGM), some with no information on adoption.
  - For extreme heat: planning (hot spots mapping, identifying cooling centers), educational/outreach and operational guidelines in fisheries and tourism — some measures adopted by MSANP.

### Public investment management (PIM) and climate integration
- Assessment framed by IMF Climate PIMA conceptual framework; summarized in Table 7 heat map.
- Institutional and implementation findings:
  - C1. Climate-aware planning: Efforts to incorporate climate change in building code and land use rules, but no real mainstreaming of climate change into planning strategies.
  - C2. Coordination across public sector: Limited coordination mechanisms for the State; none with respect to climate-related investment for PCs and SNGs.
  - C3. Project appraisal & selection: No methodology related to climate change for appraisal and selection of projects.
  - C4. Budgeting & Portfolio Management: No clear identification of investment spending related to climate change and lack of ex post climate audit and asset management.
  - C5. Risk Management: Some elements exist, but scope to progress in assessment of fiscal risks.
- Additional context:
  - National and sectoral public investment strategies only partly incorporate climate objectives; IEM 2019-2023 includes environmental pillars but falls short of full mainstreaming.
  - Strong efforts to incorporate climate resiliency into land use regulations and building codes (series of decrees published in late 2019), supported by donors (WB PPCR and AFD), but implementation is constrained by lack of capacity at local government level.

*Source: 1mdgea2022004 - 46. Adaptation measures should shield agricultural productivity from erosion by climate change (IMF PDF chapter).*

### 68. However, there is no centralized guidance or support for government agencies on

### 68. However, there is no centralized guidance or support for government agencies on how to plan public investment in the context of climate change

### Coordination Across the Public Sector
- Multiple services involved in PIM: DGEP at MoF (planning and guidelines for project evaluation and selection), Direction of Economic Cooperation and General Directorate in charge of Presidential Projects within the Presidency. These entities:
  - Lack specialized expertise in climate change.
  - Do not provide specific guidance on climate-related investments.
- Recommendation:
  - DGEP, building on its economic expertise and with help from the Ministry of Environment, could start work on Climate Resilient Planning Guidelines to:
    - Screen projects for climate risks.
    - Provide training and support to line ministries to integrate climate change in investment plans.
- Institutional gaps:
  - The Public Investment Management Strategy (adopted in 2017 and currently under review) does not refer to climate change.
  - Decision-making remains largely ad hoc and responsive to political impulses and donor priorities rather than an integrated climate approach.
  - Suggested way forward: integrate into the revised PIM Strategy a mechanism to ensure sectoral investment strategies and projects adequately reflect climate change concerns, possibly involving both DGEP and MEDD.

### Subnational Governments (SNGs) and Climate Integration
- Administrative structure: Madagascar divided into 23 administrative regions, 119 districts and 1,695 municipalities (communes).
- Legal and fiscal realities:
  - Communes are mandated (Organic Law 2014-018 of August 14, 2014) to provide local public services (primary education, primary health care), but this is largely an unfunded mandate.
  - Investment by communes represented only 2 percent of total public investment in 2017.
- Existing prescriptive framework:
  - Decrees 2019-1931 and 2019-1932 prescribe that SNGs integrate climate resiliency concerns into investment planning.
- Support and limitations:
  - 2019 guide (gradual roll-out to regions) on integrating CC concerns into regional development plans supported by WB PPCR.
  - AFD project started in 2019 to integrate climate concerns into town planning.
  - Low own resources for SNGs cause heavy reliance on grants via the Local Development Fund (FDL), whose allocation criteria do not factor vulnerability to climate change.

### Public Corporations (PCs) Oversight
- Significance:
  - Combined balance sheet of majority-owned PCs represented 11 percent of GDP in 2018.
  - Key PCs relevant to climate: JIRAMA (electricity generation and distribution) and Air Madagascar (airline).
- Oversight gaps:
  - Within MoF the DGT is tasked with oversight of PCs; there are no strategy or guidelines for the State as shareholder.
- Recommendation:
  - Develop a State shareholder strategy and incorporate climate change aspects in governance of PCs to ensure contribution to CC goals.
- Illustrative example (France):
  - APE defined guidelines for PCs to respect Paris Agreement goals, including measuring direct and indirect carbon impact, setting GHG targets, defining action plans, and annual reporting to Boards and APE.
  - Supplementary Budget Law 2020-935 conditioned COVID-related bailouts on explicit GHG reduction commitments (e.g., Air France bailout conditioned on tougher emission reduction targets and termination of some domestic flights).

### Project Appraisal and Selection
- Current practice:
  - Appraisal and selection of large infrastructure projects lack a standard climate-related analysis and standard methodology.
  - Office National pour l’Environnement (ONE) conducts environmental impact studies for major projects, but these occur late in the process after principle decisions and do not include specific climate change analysis.
- Draft reforms:
  - Latest draft public investment strategy requires appraisal of projects to align with National Development Plan objectives (which includes a pillar on environmental emergence and sustainable development), but does not provide a clear climate methodology.
  - June 2020 draft of the PIM guide details institutions, processes and coordination for appraisal and selection of externally financed public investment, but contains no reference to a clear methodology assessing climate change impact during appraisal.
- Externally financed projects:
  - Development partners identify and appraise public investments using their internal systems; they transmit files to sectoral ministries and the MoF Debt Division when financed through a loan.
  - For investments that might trigger environmental and climate risks, development partners (e.g., World Bank) systematically undertake environment and climate impact assessments.
- Public Private Partnerships (PPP):
  - PPP legal framework (Law 2015-039; decree 2017-149; decree 2014-150) does not include specific references to climate considerations or established risk sharing guidelines.
  - Risk sharing is negotiated in contracts ad hoc; climate change and adaptation measures are often treated as force majeure.
  - Examples: aviation PPP included a rendez-vous clause; energy PPP (e.g., Volobe hydroelectric dam) included modalities for calculating indemnities in case of natural disaster.

### Budgeting and Portfolio Management
- Capital budget:
  - According to the 2022 budget law, the capital budget is 5,257 billion Ariary (MGA), amounting to 36.4 percent of the total expenditure and 9.2 percent of GDP.
  - Externally financed public investment projects account for 68 percent of the capital budget.
- Externally funded climate-related projects (last ten years):
  - 94 projects with commitments totaling USD 2.8 billion and disbursements of USD 991 million.
  - Sector breakdown (Table 8):
    - Rural development: Number of projects 48; Total Commitments (millions of USD) 900.3; Total Disbursements (millions of USD) 322.2; Percent of total disbursement 33%
    - Environment: Number of projects 23; Total Commitments (millions of USD) 159.9; Total Disbursements (millions of USD) 88; Percent of total disbursement 9%
    - Energy: Number of projects 11; Total Commitments (millions of USD) 480.2; Total Disbursements (millions of USD) 215; Percent of total disbursement 22%
    - Road infrastructure: Number of projects 6; Total Commitments (millions of USD) 838.7; Total Disbursements (millions of USD) 169.5; Percent of total disbursement 17%
    - Fishing: Number of projects 1; Total Commitments (millions of USD) 74; Total Disbursements (millions of USD) 17.4; Percent of total disbursement 2%
    - Others: Number of projects 5; Total Commitments (millions of USD) 303.1; Total Disbursements (millions of USD) 178.9; Percent of total disbursement 18%
    - Total: Number of projects 94; Total Commitments (millions of USD) 2,756.2; Total Disbursements (millions of USD) 991; Percent of total disbursement 100%
  - Source: Directorate of Economic Cooperation at the Presidency
- Budget identification and tracking gaps:
  - Public investment expenditures related to climate change are not clearly identified in the budget or related documents, and are not properly monitored or reported.
  - FMIS composed of two modules (budget preparation and budget execution) that do not fully integrate preparation and execution data.
  - PIP code (alphanumeric) is not always assigned to individual projects and not followed at execution stage, preventing comprehensive monitoring and tracking of climate-related investment spending.
  - Recommendation: implement a tagging system to address identification and tracking issues.
- Ex post reviews:
  - No systematic ex post reviews of impacts of public investment projects on adaptation or mitigation, except when externally financed.
  - No legal and regulatory framework defining methodology for ex-post evaluation of public investment projects.
  - Ex post reviews including climate impact are carried out for some externally financed projects (e.g., World Bank requirements).
- Asset maintenance and climate risks:
  - Climate-related risks are not systematically reflected in methodologies for estimating maintenance needs of infrastructure assets.
  - Progress in road sector: standards for road infrastructure resistant to flooding and geological phenomena (NIRIPG – Normes sur les infrastructures résistantes aux inondations et phénomènes géologiques) adopted and published formally in January 2022, covering elements related to maintenance.
  - Recommendation: identify infrastructure assets most vulnerable to climate change and define a methodology for asset management and maintenance as a priority.

### Risk Management
- Disaster risk management:
  - Madagascar has a well-established and published disaster risk management strategy with ad hoc financing and insurance mechanisms.
  - Integration into sectoral plans through PPCR under CPGU (Prime Minister’s Office for Prevention and Management of Emergencies).
  - Ex ante financing mechanisms exist but are not yet captured in an overall disaster risk financing strategy, which is currently under preparation.
- Fiscal risk analysis gaps:
  - Madagascar does not conduct an analysis of fiscal risks linked to the impact of climate disasters on public infrastructure.
  - The last two fiscal risk statements (FRS) appended to the draft budget law did not contain elements on disaster risks or impact on public infrastructure.
  - In the FRS appended to the 2020 budget, some elements on cost and impact of recent climate disasters were presented but with no specific mention of infrastructure.
  - Recommendation: include a systematic analysis of fiscal risks related to climate, notably affecting infrastructure, in the FRS on a yearly basis.

### Public Financial Management (PFM) and Climate Mainstreaming
- Country commitments:
  - Madagascar is a member of the Coalition of Finance Ministers for Climate Action and has committed to the “Helsinki Principles”, including Principle 4 urging MoFs to mainstream climate concerns into PFM and PIM.
- Capacity and legal context:
  - Limited progress in mainstreaming climate into PFM due to broader PFM framework and capability limitations.
  - Progress contingent on enhanced dialogue and collaboration between MoF and Ministry of Environment and eventual capture in the legal framework.
- Strategic planning and fiscal framework:
  - Climate considerations somewhat reflected in Government strategic plan and sector strategies, but not in medium-term fiscal plans (MTFF / MTBF).
  - No specific and systematic effort to align MTFF / MTBF with climate objectives.
- Budget preparation:
  - Limited focus on climate in budget preparation, except for major externally funded projects.
  - Budget preparation circular does not highlight climate concerns; no general requirement for climate impact assessment of budget measures.
  - No specific tagging of climate expenditure; current budget program structure does not allow identification of climate-relevant spending.
  - Ongoing reform of budget classification with IMF support could allow insertion of a code to track climate-relevant expenditure.
- Climate Budget Tagging (CBT) example (Ghana):
  - Ghana developed in 2018 a CBT tool (CLIMATRONIC); groups climate-relevant expenditures into high/medium/low relevance weighted 100%, 50%, 20%.
  - MoF leads process and captures categories in Chart of Accounts; CBT tracked climate-related spending estimated at an average of 4% over 2015-2020.
- Transparency recommendations:
  - Create and gradually enrich a dedicated annex to the annual budget on climate change (example: Bangladesh) to include information from CBT, climate-related contingent liabilities, and feed into citizen budget.
- Budget execution and accounting:
  - Climate considerations not explicitly reflected in budget execution and accounting.
  - Major scope for improvement in mechanisms for budget execution and accounting (2021 PEFA assessment).
  - Public accounts produced with significant delay (as of April 2022 the 2020 Compte Général de l’Administration des Finances yet to be published) and do not provide information on climate spending.
- Control and audit:
  - Climate not explicitly and systematically reflected in internal controls or internal and external audit.
  - Cour des Comptes receiving USAID support to expand into performance audits but has not carried out performance audits of climate projects or programs.

### Macro-Fiscal Implications of Climate Policy
- Current projections:
  - Effects of climate (natural disasters) and climate change (higher frequency/intensity of disasters and gradual permanent reduction in productivity) are not included in authorities’ baseline macro-fiscal projections.
- Implications:
  - Inclusion of climate impacts results in an unsustainable debt trajectory and/or a permanently lower growth path.
  - Investments in resilient infrastructure and adaptation projects alone are insufficient to restore debt sustainability, even when public investment efficiency is enhanced.
- Policy context:
  - Adaptation investments should be pursued within the broader context of efforts to reach the SDGs, including investments in human capital to durably increase potential growth and investments in resilient infrastructure.

*Source: 1mdgea2022004 - 68. However, there is no centralized guidance or support for government agencies on*

### 87. Madagascar’s exposure to recurrent natural disasters

### 1mdgea2022004 - 87. Madagascar’s exposure to recurrent natural disasters

### Exposure, frequency, and direct impacts
- Madagascar has been historically hit by a tropical cyclone every 0.9 years on average; droughts and floods take place every 2.7 and 3 years respectively (EM-DAT, 2022).
- A typical tropical cyclone could become 18.4% more damaging by 2050 (Acevelo, 2016).
- Typical disaster direct impacts and sectoral patterns:
  - A typical tropical cyclone produces 1.8% of GDP in total losses, of which 57% in damages to physical assets (authorities’ assessments of Ava and Enawo).
  - An average flood produces 2.4% of GDP in total losses (EM-DAT, 2022).
  - An average drought may reduce yearly real growth by 0.3 percentage points, with persistent impacts in following years (IMF estimates on Sub-Saharan Africa).
- Recent event damage estimation caveat: African Risk Capacity estimates of cyclone Batsirai (2022) point to total damages of up to US$2.1bn (14% of GDP), while World Bank estimates point to damages of about 4.5% of GDP — indicating publicly available figures may be underestimated.

### Channels from climate shocks to growth, fiscal outcomes, and debt
- Growth channels (three complementary channels):
  - Immediate GDP losses and damages to physical assets and productivity (see typical loss figures above).
  - Permanent scarring effects due to long recovery times; reconstruction of public infrastructure estimated to take approximately 4 years to fully rebuild.
  - Gradual permanent reduction in total factor productivity reflecting slow-moving effects of climate change; literature points to a loss in GDP per capita of up to 1.5% in GDP per capita by 2050 based on historical trends (Kahn et al, 2019).
- Fiscal and debt channels (four channels):
  - Fall in government revenues due to lower economic activity and tax relief measures.
  - Higher emergency current and capital spending; authorities estimated social spending needs alone after cyclone Batsirai (2022) at 1.3% of GDP.
  - Lower non-emergency growth-enhancing spending (capital spending cuts), with donors partially shifting financing toward post-disaster programs.
  - Financing constraints: higher deficits raise financing needs and may reduce concessional financing and donor appetite, increasing financing costs.

### Baseline macroeconomic assumptions (no disaster materialization)
- Growth stabilizes at a potential level of 5 percent in 2030, after slightly higher growth (around 5.1% in 2022-2026 on average).
- Primary deficits: 0.2% of GDP on average in 2022-2026.
- Public investment scale-up in the medium term: +0.5% of GDP by 2030.
- Revenue mobilization: +2.4% of GDP by 2040.
- Madagascar should remain at a moderate level of debt distress under these assumptions, with concessional funding and grants covering the bulk of financing needs.

### DIGNAD model setup and key calibrated parameters
- Model: Debt, Investment, Growth, and Natural Disasters (DIGNAD) calibrated to Madagascar using historical data and the ECF macroeconomic framework.
- Assumptions on disaster frequency increases (annual increases):
  - Tropical cyclones: 1.45%
  - Floods: 4.6%
  - Droughts: 1.2%
- Model treats two types of capital: standard capital and resilient (adaptation) capital (resilient capital less prone to destruction and higher returns).
- Selected calibrated parameters and initial values (in percent):
  - Return to standard infrastructure: 25
  - Return to adaptation infrastructure: 30
  - Public investment to GDP ratio: 5.2
  - Grants to GDP ratio: 1.4
  - Consumption tax (VAT) rate: 20
  - Public domestic debt to GDP ratio: 13.2
  - Public external concessional debt to GDP ratio: 22.3
  - Public external commercial debt to GDP ratio: 3.5
  - Remittances to GDP ratio: 2.6
  - Imports to GDP ratio: 17.6
  - Real interest rate on domestic debt: 0
  - Real interest rate on external commercial debt: 0.98
  - Depreciation rate of standard public infrastructure: 7.5
  - Depreciation rate of resilient public infrastructure: 3
  - Public investment efficiency: 31%
- Public investment efficiency is subject to an efficiency factor (IMF’s Public Investment Efficiency database value: 31%).

### Climate-adjusted baseline results and debt dynamics
- Two illustrative outcomes under climate-adjusted baseline (includes both baseline climate impacts and climate change increases in intensity/frequency and productivity losses):
  - Unconstrained financing (country can borrow gross financing needs in full each year; fully covers reconstruction and humanitarian spending):
    - Debt-to-GDP would increase by 28 percentage points relative to a no-climate change baseline by 2040 and become unsustainable.
    - Debt expected to exceed 70% of GDP in 2034 and reach 84% of GDP in 2040.
    - Overall deficits expected to remain at -5.4% of GDP on average after 2024.
    - Total cumulative GDP losses amount to 4.4% of GDP by 2040.
    - Trend real GDP growth lower by 0.18 percentage points relative to a no-climate change scenario.
  - Financing-constrained scenario (country faces a financing constraint when no longer assessed as moderate risk of debt distress; indicative debt ceiling ≈ 65% of GDP):
    - Debt stabilizes around 65% by 2040, but at the price of higher cumulative GDP losses.
    - When debt-to-GDP moves closer to 65%, authorities still cover reconstruction and humanitarian needs in full but strongly limit debt raising and implement fiscal adjustment mostly via public investment cuts and partly via revenue mobilization.
    - This contractionary scenario yields durable cumulative GDP losses of 5.2% by 2040 and an average yearly GDP growth loss of 0.22 percentage points.
- Contribution decomposition to total real GDP loss by 2040:
  - Climate shocks (floods, cyclones, droughts): about 63 percent of total real GDP loss.
  - Climate change shocks (increase in intensity/frequency and productivity loss): about 37 percent.

### Policy responses and modeled intervention scenarios
- Four priority policy areas identified:
  - Climate mitigation measures that improve fiscal and distributional outcomes.
  - Adaptation investments that are more resilient to natural disasters and shield the rest of the economy.
  - PFM measures to improve the efficiency of public investment.
  - Risk management measures to facilitate the authorities’ fiscal response to a natural disaster.
- These responses should be combined with SDG-oriented human capital investments (health and education) and ambitious revenue mobilization; mitigation revenues could be recycled into SDG-compatible programs and transfers and complemented by donor grants.
- Three modeled policy scenarios:
  - Scenario 1: adaptation investments consisting of (i) building back better after disasters (rebuilding standard infrastructure as resilient infrastructure) and (ii) a resilient investment scale-up plan amounting to 1.5% of GDP annually.
  - Scenario 2: Scenario 1 plus gradual improvement in public investment efficiency from 31% up to 63% (average Sub-Saharan African level) over the next 10 years.
  - Scenario 3: SDG investments in human capital and resilient infrastructure amounting to 4.8% of GDP per year for 10 years (i.e., until 2033), funded partly by mitigation measures and mobilization of grants.

*Source: Madagascar’s authorities and IMF staff calculations.*

### section III.)

### Section III.

### Summary of policy measures and modeling assumptions
- Adaptation investments
  - Adaptation investment is assumed to be 25 percent more costly for the government than standard investment. 1/
  - Returns: returns on standard investments in literature estimated at 25%, while returns on adaptation investments are assumed to be 30%. 2/
  - Adaptation investments yield: higher resilience to disasters; higher costs vs. standard investments; higher economic returns on adaptation investments; shielding effect on the rest of the economy.
- Risk management measures
  - Budget measures (contingency buffers, automatic stabilizers) and insurance products to mitigate revenue loss, absorb spending needs after a shock and preserve growth. Outcomes include: no loss of public investment efficiency in the aftermath of a shock; insurance to cover for losses beyond average frequency/impact.
- Public financial management / public investment management
  - Public investment efficiency (PIE) baseline at 31%; average in LIDCs is 61%.
  - Public investment reforms are modeled as a gradual 10-year increase of efficiency from 31% to 61%. 3/
  - Expected outcome: gradual improvement of public investment efficiency to average Sub-Saharan Africa level; higher build-up of capital stock.
- Sustainable Development Goals (SDGs) oriented investments
  - Ambitious spending plan in human capital (health and education) and in resilient infrastructure (water, roads, energy) to reach SDGs within 10 years. 4/
  - SDG costing: total annual public investment needs to reach the SDGs in 5 sectors (health, education, water, roads, energy) by 2033 could amount to c. 4.8% of GDP over the next 10 years on top of existing public investments, including additional costs of building resilient infrastructure. 4/
  - Human capital investments raise productivity and potential growth; IMF SDG fiscal financing tool implies human capital investments of the needed magnitude could boost potential growth by 1.5% by 2040. 5/
  - Resilient infrastructure in water, energy, roads: same benefits as adaptation investments and higher cost than standard investment. Focus on renewable energy for decarbonization targets.
- Non-fiscal measures for adaptation
  - Regulatory standards (risk zoning, early-warning systems, disaster response protocols). Standards for adaptation infrastructure reflected in higher unit costs; public programs not modeled.

### Scenario analysis and fiscal outcomes
- Financing mix assumptions for Policy scenarios 1 and 2:
  - 67.1% of total financing needs met through concessional debt.
  - 24.6% through grants.
  - 8.3% through external commercial debt.
- Policy 3 financing assumption:
  - Assumes additional grants in support of SDG investment (about 1 percent of GDP per year), with the remaining financing covered by the same mix as in Policy 1 and 2.
- Scenario outcomes and key quantitative findings:
  - Policy scenario 1 (scaling up adaptation investments alone)
    - Debt reaches 98% of GDP.
    - Adaptation capital remains limited by 2040 and provides only limited protection against disasters of increasing frequency and intensity.
    - Conclusion: insufficient to restore a viable debt trajectory given public investment inefficiency and higher cost of resilient infrastructure.
  - Policy scenario 2 (public investment efficiency reforms + adaptation)
    - Total debt-to-GDP ratio is reduced by 4 percentage points compared to scenario 1, due to lower damages from natural disasters and higher potential GDP.
    - Debt, however, remains unsustainable.
  - Policy scenario 3 (sustainable growth approach centered on SDGs)
    - Puts debt on a downward trajectory in the medium term.
    - More ambitious investments in resilient infrastructure combined with public investment efficiency measures limit losses from natural disasters and boost growth.
    - Human capital investments durably boost productivity by 17% by 2040 and strengthen economic recovery after a shock.
    - Ability to mobilize revenues from mitigation measures and additional grants from donors in support of human capital spending keeps deficits in check.
- Policy implication:
  - Implementing the first-best policy scenario (policy scenario 3) requires substantially higher support from donors. Madagascar needs a credible package of high-quality climate measures centered on durable resilience and higher spending efficiency to build the case for higher donor support.

### Mainstreaming climate into macro-fiscal policy — recommended measures
- Reflect natural disasters in annual and medium-term budgets by including explicit assumptions on climate scenario and its expected impacts, adequate buffers and well-costed response measures (short-term).
- Prepare a well-costed document reflecting a package of climate measures — including adaptation, public investment efficiency, mitigation and SDG-oriented measures — with associated implementation plans to anchor future reform efforts; build on existing policy documents (e.g., NDC, NAP, PFM reform strategy) (short-term).
- Develop technical capacity to assess the impact of natural disasters on macroeconomic variables and include climate shocks in all macroeconomic projections (short- to medium-term).
  - Develop a database of past economic impacts of natural disasters.
  - Conduct systematic economic impact assessments (sectoral GDP impact, physical capital losses).

### Annex I — assessment of adaptation policies: objectives, policies, and gaps
- National Adaptation Plan strategic areas:
  - Strengthen governance and integration of adaptation.
  - Implement a program of priority sectoral actions.
  - Financing adaptation to climate change.
- Sectoral preparedness:
  - Contingency plans to cope with Tropical Cyclones and Droughts.
- Institutional arrangements and knowledge mechanisms:
  - Sectoral, participatory approach; national entities provide technical support upon request.
  - Ten structuring programs in NAP: targeted sectoral development (4/10), improved information systems for extreme events (4/10).
  - Sectors are responsible for collecting and reporting disaster impact data, but capacity is lacking.
  - Adaptation policy-making not effectively informed by evaluation of past policies.
- Funding mechanism gaps:
  - Most adaptation interventions are project-based and driven by development partners.
  - Limited private sector involvement.
- Strategy gaps:
  - National plan does not fully distinguish adaptation and mitigation; mitigation included among the ten structuring programs.
  - Sector plans do not clarify relationship between adaptation and development.
  - Insufficient recognition of slow-moving climate stresses.
  - Lack of prioritization mechanism for efficient selection of adaptation interventions.
- Knowledge gaps:
  - Sectoral ministries lack systematic impact data from past disasters and projections of future climate stresses.
- Finance gaps:
  - Sectoral entities lack capacity to navigate international adaptation finance landscape, limiting access to international funding mechanisms.
  - No strategy to unlock private finance for adaptation.

### Annex III — DIGNAD model: structure, channels, and modeling notes
- Model overview
  - DIGNAD = Debt, Investment, Growth, and Natural Disasters model: dynamic small open economy model simulating impacts of natural disasters and trade-offs between debt accumulation, public/private investment, and growth.
  - Two types of public capital: standard physical capital (vulnerable to disasters) and adaptation capital (more resilient).
  - Government financing includes external concessional loans and grants.
- Three channels used to simulate climate impacts
  - (i) Permanent decrease in total factor productivity (TFP) due to changing temperature patterns.
    - Modeled as a gradual and anticipated fall in factor productivities in tradable and non-tradable sectors; calibrated to match mid-range estimates of GDP losses in Kahn et al. (2019).
  - (ii) Scarring from multiple and coinciding natural disasters.
    - Three disaster types (tropical cyclones, floods, droughts) occur at historical frequencies over the last 43 years.
    - Disasters affect the economy via: (i) damages to public capital; (ii) damages to private capital; (iii) temporary productivity loss.
    - Assumes economy gradually recovers to steady state after a disaster absent another disaster.
  - (iii) Increasing frequency and intensity of natural disasters.
    - Assumes linearity: increasing intensity of a disaster by factor β produces macroeconomic responses higher by factor β.
- Modeling and results notes
  - The largest driver of lower potential GDP under the climate baseline are tropical cyclones, which account for 1.3 percentage points of the total real GDP loss in 2040.
  - The contribution of increased intensity and frequency of natural disasters rises and becomes by 2040 the second largest driver behind the lower potential GDP projection.
  - Hodrick-Prescott filtering used to separate long-run adjustment from short-term fluctuations; smoothing parameter λ=100.
  - Note on limitations: the linearity assumption may lead to underestimation of economic costs since damage functions are typically non-linear. 65

*Source: IMF staff section III. from 1mdgea2022004 - section III.*

### Annex VI. The Role of Social Protection in Energy Subsidy

### Annex VI. The Role of Social Protection in Energy Subsidy Reforms

### Main recommendation
- Expand social transfers using budgetary savings from the energy pricing reforms to mitigate the impact of higher energy prices on the poor.
- Empirical finding cited: A study of 25 fuel subsidy reform episodes in the MENA region found that all reforms that used cash transfers successfully removed fuel subsidies, whereas only 17 percent of reforms without cash transfers succeeded.

### Country examples and lessons learned

- Iran (2010)
  - Reform action: Increased the price of petroleum products by between 230 and 840 percent.
  - Compensation design: Government opened bank accounts for heads of households prior to removal of subsidies and started making bi-monthly cash transfers to 70 to 80 percent of all citizens.
  - Financing: Used about 80 percent of the revenue from price increases for transfers.
  - Impact: The equal transfers were large enough to lift virtually every Iranian out of poverty as the compensation represented a larger share of poor household’s income than it did for the middle class.
  - Takeaway: Successful reforms are well planned with a clear reform strategy, including by expanding financial inclusion. In cases where targeting is administratively difficult, providing all households with equal transfers could still achieve redistributive effects and build support for the reform. Targeting of these transfers can then be improved over time.

- Kenya (2001–2008)
  - Reform action: Electricity subsidies declined from 1.5 percent of GDP in 2001 to zero in 2008 as part of a broader package to address supply problems.
  - Compensation/redirected subsidies: Subsidized connection costs to expand coverage to poor households and those in remote and rural areas; rural electrification program increased connections from 650,000 in 2003 to 2 million as of 2014, with a fund for connection fee payments financed by donors.
  - Takeaway: Subsidies can be redirected to achieve other development goals such as improving access to electricity. Electricity pricing reforms accompanied by a broader reform package to improve operational efficiency of the electricity company would increase the likelihood of success.

- Ghana (2005)
  - Reform action: Increased the price of petroleum products by 50 percent on average.
  - Compensation measures: Introduced programs including elimination of fees for state-run primary and secondary schools; an increase in public-transport buses; a price ceiling on public-transport fares; and more funding for health care in poor areas.
  - Takeaway: While targeted cash transfers or near-cash transfers are typically preferred, they might not be immediately feasible due to capacity constraints. Other programs can be utilized while administrative capacity is developed to scale up cash transfer programs. Efficient and visible reallocation of resources saved through subsidy removal to programs with immediate benefits to the most vulnerable helps secure successful reform.

### Practical implications for Madagascar
- Prioritize using budgetary savings from energy-pricing reform to expand social transfers to protect poor households.
- Where administrative targeting is limited, consider broad-based or universal transfers initially to build support and then refine targeting over time.
- Consider redirecting subsidy savings toward programs that expand access (for example, subsidizing connection costs) and toward visible, immediate benefits for vulnerable populations while building capacity for cash-transfer systems.
- Combine price reforms with broader sectoral reforms (for example, improving operational efficiency of electricity providers) to increase reform success.

*Source: Annex VI. The Role of Social Protection in Energy Subsidy Reforms (from the provided IMF content).*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2022/english/1mdgea2022004.pdf_