## Annex I. Data Sources for Calibration (wpiea2025135-print-pdf)

## Source details

**Canonical URL:** [Annex I. Data Sources for Calibration (wpiea2025135-print-pdf)](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025135-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025135-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025135-print-pdf.pdf.json)

---

### Climate and disaster data — calibration guidance
- Use catastrophe modeling datasets with annual average losses and risk curves for alternative return periods to convert infrastructure damages (replacement cost) into percent of capital stock (e.g., FAD capital stock dataset).
- Suggested sources:
  - Global Risk Assessment; Global Infrastructure Risk Model and Resilience Index (GIRI) for multi-hazard risk curves.
  - Aqueduct Floods for single-hazard risk curves.
- Slow-onset impacts (TFP decline) — two approaches:
  - Top-down econometric estimates (Centorrino, Massetti, and Tagklis (forthcoming)).
  - Bottom-up channel aggregation: collect channel-specific biophysical/econometric impacts and weight by input shares (example channel source: Dasgupta et al., Lancet Planet Health 2021).
- Macro-fiscal baseline calibration sources: IMF World Economic Outlook; IMF Fiscal Monitor Database; IMF Country Staff Reports.

### Disaster — damages to capital stock (calibration steps)
- Select scenario from risk curve and record magnitudes and return periods.
- Convert infrastructure damages into percent of capital stock using country total capital stock data.
- Recommended to use catastrophe model outputs (annualized losses, probable maximum losses).

### Macroeconomic baseline data
- Baseline data and projections drawn from WEO, SRs, DSA as country-specific inputs for model calibration.

---

### Section 4: Baseline Macroeconomic and Financing assumptions (Jamaica) — key inputs and context

### Context and macro-financing framing
- SDG incremental spending needs in 2030:
  - LICs: approximately US$0.5 trillion (average incremental spending of 16 percentage points of GDP).
  - EMs: approximately US$2.35 trillion (average incremental spending of 5 percentage points of GDP).
- Financing options: domestic resource mobilization, spending efficiency, public debt-financing, official development assistance, private sector engagement.
- Examples of tax-to-GDP reforms:
  - China 1995–2002: 5 percentage point increase.
  - Georgia 2004–2009: 12 percentage point increase.
- IMF instrument: Medium-Term Revenue Strategy (MTRS).

### Climate exposure, empirical evidence, and adaptation costs
- IPCC AR6 (2022): 3.3 to 3.6 billion people live in settings “highly vulnerable” to climate change.
- World Bank Lifelines (2019):
  - Disasters currently cost low- and middle-income countries US$ 390 billion a year in infrastructure damages.
  - Investing in resilient infrastructure costs 3 percent more but yields a total net benefit of US$ 4.2 trillion.
- Empirical macro-fiscal transmissions:
  - Droughts: deterioration of primary balance and public debt of EMDEs by as much as 1.4 and 1.5 percent of GDP, respectively (Fuje et al., 2023).
  - Disasters under non-pegged exchange rates: nominal and real exchange rate depreciation up to 6 percent two years following the disaster (Nguyen and Nguyen, forthcoming).
  - High temperatures/severe droughts: reduce GDP per capita by 0.2 percentage points (Akyapi et al., 2022).
  - Climate events in seven Central American countries: monthly economic activity decline ~0.5–1 percentage point and remittances increase >3 percentage points (Kim et al., 2022).
- Adaptation cost signals:
  - Annual adaptation costs can exceed 1 percent of GDP for some developing countries and be above 10 percent of GDP for some island states (Aligishiev et al., 2022).
  - Aggarwal et al. (2024): additional 0.4 percent of global GDP required for mitigation and adaptation needs of SDG sectors by 2030.

---

### The C-SDG Model — structure and mechanics

### Core components and production technology
- Lineage: extends SDG-FiT (Benedek et al., 2021) to include fast and slow-onset climate impacts and adaptation policies.
- Endogenous accumulation: human and physical capital; fiscal policy interventions; public debt dynamics.
- Public capital distinction:
  - Two types: resilient (R) and standard (S) public physical capital.
  - Three SDG-relevant assets: water, roads, electricity/power generation.
- Production function: augmented Cobb-Douglas with inputs:
  - Private capital (K), public resilient capital (R), public standard capital (S), human capital (H), sector-specific labor (L), and TFP (A) affected by slow-onset climate change (D_s,s,i).
- Key parameters and definitions:
  - θ (>1 implies resilient > standard productivity; θ = 1 implies perfect substitutes).
  - Sector-specific elasticity of output to public capital: β_j.
  - Labor share set by ILO estimates; capital share = 1 − labor share.

### Capital accumulation, damage, and resilience mechanics
- Depreciation:
  - Resilient capital δR ∈ (0,1); standard capital δS ∈ (0,1).
- Public investment efficiency e ∈ (0,1]; e = 1 denotes full efficiency.
- Infrastructure damage from climate shock: D_ij ∈ (0, 1−δS).
- Damage reduction (resilience effectiveness) parameter: ρ ∈ (0,1]; 0 = maximum damage reduction (no residual risk), 1 = no damage reduction.
- Per shock gross benefit of resilient capital: ∂Damage_effect/∂R = (ρ − 1) D (given ρ < 1, impact of shock declines with resilient investment).
- Relative price of resilient vs standard capital: p ∈ (0,1]; resilient infrastructure more expensive in example framing.

### Public finances and allocations
- Total public financing = tax and non-tax revenues + grants + net borrowing (domestic and external).
- Tax revenue modeled as function of real per capita growth and exogenous elasticity; non-tax revenue follows observed trend.
- Public expenditure split: capital (SDG infrastructure + additional resilient capital cost + reconstruction) and current (PPE subsidies, interest, other current, transfers due to shocks).

### Human capital accumulation and scarring
- H_t = (1 − δ_h) H_{t−1} + ω * ξ_{t−1}.
- ξ_{t−1} depends on public/private education and health spending (h), spending efficiency (ε), share of school-age population (n), and previous cohort human capital (1 − ω) ξ_{t−2}.
- Climate shock D_h reduces transformation of spending into new human capital (analogous to lowering ε).

### Behavioral and modelling scope
- Household savings rates exogenous (reflecting low financial inclusion and hand-to-mouth behavior).
- Framework abstracts from some general equilibrium effects (price dynamics, labor market adjustments); designed for macro-fiscal planning and scenario analysis.

---

### Calibration and country application guidance

### Default parameterization and recommended practices
- Default parameters from cross-country databases and literature; use country-specific estimates where available.
- Users recommended to calibrate climate impact parameters with country-specific fast and slow-onset data (meteorological agencies, sector ministries, CCDRs).
- IMF compiles climate impact studies including FADCP Climate Dataset (Massetti and Tagklis, 2024) with temperature, precipitation, extreme-event risks, and sea-level rise under alternative scenarios for all IMF members.
- Framework flexible for sectoral parameter adjustments (production technology, capital accumulation, climate impacts, adaptation policy).

---

### Benin application — calibration inputs, scenarios, and selected quantitative results

### Vulnerability and financing context
- 2.7°C global warming scenario (current policies): Benin projected to have high exposure to extreme temperatures by 2070.
- Estimated: 98 percent of Benin's landmass will be impacted by severe temperature conditions (IMF Country Report 24/003).
- INFORM Climate Change Vulnerability Index: Benin ranks 163 out of 185 (185 most vulnerable).
- Floods 2010 affected over 830,000 people, or 9.1 percent of the population (EM-DAT).
- 2022 NAP estimates adaptation needs at approximately US$4.2 billion by 2032.
- World Bank CCDR additional financing needs due to climate change:
  - Annual investment of 0.1 percent of GDP until 2030.
  - 0.3 percent of GDP by 2040.
  - 0.8 percent of GDP by 2050.
  - Annual average of 0.3 percent of GDP over 2022–2050.
- IMF support: Resilience and Sustainability Facility (RSF) financing of US$200 million.
- Benin issued €500 million SDG bond in July 2021.

### Flood calibration and scenario design
- Historical flood frequency: exposed to a flood every 2.2 years since 1980; floods can affect as much as 15 percent of total population.
- Precipitation projections:
  - 5.5 percent increase by 2070 under SSP2-4.5 (current policies).
  - 18.1 percent increase under SSP3-7.0 90th percentile.
- World Bank CCDR: extreme floods can decrease GDP by more than 2 percent by 2050 (mean SSP3-7.0) and up to 6 percent (higher bound SSP3-7.0); additional capital losses between 4.9 and 6.5 percent of capital stock.
- Urban capital stock damage rate: ~26 percent for a 25-year return period flood by 2030 under high emissions scenario; translated to national capital stock damage rate of 13 percent assuming urban flood affects infrastructure critical to secondary and tertiary sectors (~50 percent of total capital stock).
- Flood simulation impacts 5.6 percent of the population, hindering human capital accumulation.
- Extreme scenario: 1,000-year return period flood urban damage rate ~37 percent (upper limit for adaptation investment).

### Resilient infrastructure parameter assumptions (Hallegatte et al., 2019)
- Incremental cost of resilient infrastructure for low- and middle-income countries: 3 percent relative to standard infrastructure (p = 0.97).
- Damage reduction from resilient infrastructure: about a factor of 2 (ρ = 0.40).

### Benin simulation scenarios
- Baseline: current policies and growth projections (includes average weather-related shocks in WEO/DSA).
- Counterfactual 1: two extreme floods in 2027 and 2035, each affecting 13 percent of public capital stock and 5.6 percent of population; reconstruction financed by deficit and completed within three years.
- Counterfactual 2 (adaptation policy): 50 percent of new public investment directed to resilient capital until targeted share of resilient infrastructure (19 percent) reached.
- Robustness: “Higher p” (larger relative price of resilient infrastructure); “Higher R” (higher targeted share of resilient infrastructure).
- Fiscal shock during floods:
  - Fiscal revenues decrease by 0.2 p.p.
  - Government increases transfers by 0.1 p.p.
- Scenario 5 (extreme benchmark): government unable to rebuild damaged public capital due to binding limited fiscal space.

### Selected quantitative results (Benin)
- Adaptation policy reduces public capital damages by 13 percent in 2027 and by 19 percent in 2035, lowering GDP impacts of climate shocks.
- Share of resilient capital stock: ~10 percent by 2027; reaches targeted 19 percent by 2035.
- Extreme flood shocks reduce GDP growth by 2.1 p.p. relative to baseline in 2027 and 2035 without adaptation.
- With resilient investment, flood shocks reduce GDP growth by:
  - 2.0 p.p. in 2027.
  - 1.8 p.p. in 2035.
- Adaptation policy reduces GDP growth losses by 0.10 p.p. in 2027 and 0.30 p.p. in 2035.
- Calibrated exercise implies a growth yield of about 16 percent relative to the scenario with extreme flood shocks (reported in paper).
- Fiscal outcomes:
  - By 2040 debt levels under adaptation policies are 1.6 p.p. lower compared to current climate shock scenario; by 2054 they are 0.6 p.p. lower.
  - Fiscal deficit jumps without adaptation: about 3.1 p.p. and 3.0 per year over reconstruction periods (3 years) after 2027 and 2035 shocks.
  - With adaptation: fiscal deficit jumps by 2.7 and 2.5 p.p. per year for those reconstruction episodes (0.4 and 0.5 p.p. lower), implying a fiscal deficit yield of about 18 percent.
- SDG financing effects:
  - Adaptation narrows financing gap by ~US$120 million in 2027 (0.8 percent of GDP) and ~US$320 million in 2035 (1 percent of GDP) relative to climate shock with reconstruction (no resilient infrastructure).
  - By 2040 financing gap with adaptation decreases by 0.04 percent of baseline GDP.
  - Cumulative discounted gains from adaptation policies by 2040 ≈ US$ 400 million, equivalent to nearly 3 percent of 2024 GDP or 24 percent of 2024 tax revenues.
  - No reconstruction scenario: GDP per capita 2.2 percent lower by 2040 relative to reconstruction scenario (no adaptation policies).
  - Extreme climate shocks increase SDG financing needs by 9.65 percent of baseline GDP by 2040 if no reconstruction; reconstruction reduces additional SDG financing needs to 2.47 percent of GDP by 2040.

---

### Jamaica application — exposure, scenarios, and outcomes

### Exposure and historical losses
- Inform Risk index 2023: Jamaica ranks 47th out of 191.
- BAU global emission scenario (RCP 4.5): Jamaica projected to face 1.8°C increase of mean temperature by 2100 relative to 1984–2014 baseline.
- Tourism ≈ 20 percent of GDP; high climate sensitivity.
- Historical disasters:
  - Hurricane Ivan (2004): US$ 580 million (8 percent of GDP).
  - Hurricane Dean (2007): US$ 329 million (3 percent of GDP).
  - Tropical storm Nicole (2010): US$ 239 million (2 percent of GDP).
  - Hurricane Beryl (July 2024): ~US$ 200 million (1.1 percent of GDP).
- One-in-100-year event estimated losses > US$ 1,729 million (~10 percent of GDP) with a 1 percent annual probability of exceedance.

### Modeling calibration and scenarios
- One-in-100-year cyclone event assumed to cause national public capital stock damage of 8 percent of GDP.
- Two extreme cyclone events modeled in 2027 and 2035, each affecting 8 percent of public capital stock; reconstruction over three years financed by higher deficit.
- Adaptation scenario parameters (Hallegatte et al., 2019):
  - Additional cost of resilient infrastructure: 4 percent relative to standard (p = 0.96).
  - Damage reduction: about a factor of 2 for new infrastructure (ρ = 0.44).
- Sensitivity scenarios: “Higher p” (higher relative price) and “Higher R” (higher share); also “no reconstruction” scenario.

### Selected quantitative outcomes (Jamaica)
- Adaptation policy reduces GDP growth shock by 0.2 percentage points by 2035 and reduces reconstruction costs by ~1.2 percent of GDP.
- Share of resilient capital stock: 14 percent in 2027, rising to 28 percent by 2035 in adaptation scenario.
- With government financing, debt levels increase ~10 percent of GDP higher than baseline by 2038 without adaptation; adaptation reduces debt level by 1.3 p.p. by 2038.
- Higher cost of resilient capital leads to 0.2 percentage points lower capital stock by 2054 versus standard adaptation cost scenario.
- Higher targeted resilient share yields marginal additional effectiveness: GDP shock in 2035 reduced by only 0.1 p.p. compared to baseline adaptation, and debt/GDP in 2038 marginally lower by 0.5 p.p.
- SDG financing:
  - Climate shocks increase financing needs by 6.3 percent of GDP if government does not invest in reconstruction.
  - Reconstruction reduces additional SDG finance needs to 1.4 percent of GDP by 2040.
  - Adaptation narrows financing gap to 1.35 percent of GDP by 2040.
  - Adaptation narrows financing gap by ~US$15 million in 2027 (0.4 percent of GDP) and ~US$50 million in 2035 (0.8 percent of GDP).
  - By 2040 financing gap with adaptation decreases by 0.05 percent of baseline GDP.
  - Cumulative discounted gains from adaptation by 2040 ≈ US$ 64 million, equivalent to nearly 2 percent of 2024 GDP.
  - Sensitivity scenarios (higher p and higher R) do not significantly affect the SDG-financing gap.

---

### Model findings, trade-offs, and policy recommendations

### Key quantitative and qualitative insights
- Investing in resilient infrastructure lowers economic damages from climate shocks relative to scenarios with standard infrastructure.
- Net benefits of adaptation depend on:
  - Country climate risk profile, adaptation options, relative prices (p), and damage reduction potential (ρ).
  - Higher frequency and magnitude of extreme events increase net benefits of adaptation.
- Timing matters: early investment in resilience yields larger long-term benefits; building resilience takes time.
- Climate-related spending typically small relative to overall SDG spending on average but can spike sharply in shock years.

### Fiscal and macro trade-offs highlighted
- Reconstruction financed by deficits increases debt and interest payments, potentially crowding out SDG spending.
- Adaptation reduces reconstruction needs and interest payments, improving fiscal space and lowering crowding-out risks.
- Example fiscal effects:
  - Benin: adaptation reduces fiscal deficit jumps by 0.4 and 0.5 p.p. in reconstruction years (fiscal-deficit yield ~18 percent).
  - By 2040 Benin debt under adaptation 1.6 p.p. lower; by 2054 0.6 p.p. lower.
  - DIGNAD model comparison: extreme event decreases growth by 4 p.p.; resilient investment reduces loss by 1 p.p. (growth yield ~25 percent) — different calibration and setup.

### Policy recommendations
- Governments should:
  - Accurately identify exposed public capital.
  - Define target shares of resilient infrastructure to balance marginal returns.
  - Integrate climate action within development agendas, not treat it separately.
  - Prioritize early resilience building to maximize long-term benefits and smooth shock-year financing spikes.
- Financing considerations:
  - Use a mix of domestic resource mobilization, spending efficiency, concessional financing, grants, and private sector engagement.
  - Explore ex-ante disaster risk financing options and donor/concessional support to mitigate deficit-financed reconstruction.

### Suggested model extensions and research priorities
- Improved modelling of private sector risk reduction and recovery behaviors.
- Alternative ex-ante risk financing options and partial donor financing of reconstruction.
- Modeling persistence of damages and longer recovery dynamics; endogenous interest rate responses to climate shocks.
- Testing shock-responsive social transfers and non-structural adaptation options.
- Broader country applications with multiple hazards and slow-onset impacts via fully stochastic simulation to capture macro-fiscal tail risks.

---

### Annex II — sectoral findings for calibration inputs (selected highlights)

### Roads
- Climate-inflicted transport damage costs: €2.5 billion/year (Cost accounting framework).
- Global EAD to road and railway assets: $3.1 to 22 billion (73% caused by surface and river flooding).
- Global EAD ≈ 0.02% of global GDP; in some countries 0.5 to 1% of GDP annually.
- Vietnam: mean additional cost to maintain same road network through 2050: US$10.5 billion.
- Africa repair/maintenance through 2100: US$183.6 billion total (US$2.3 billion per year).
- Ghana: $473 million to maintain/repair existing roads (no adapt, 2020–2100); new road adaptation cumulative cost: $678.47 million.
- South Africa (2050 decade annual costs):
  - Median: USD$116.8 million; Maximum: USD$228.7 million if no adaptation.
  - With proactive adaptation: costs reduced to USD $55.7 million.
- Malawi, Mozambique, Zambia through 2050: $596 million (median scenarios) to maintain/repair roads.

### Energy
- Temperature effects on thermal power output:
  - 1°C increase reduces power output by ~0.45%; nuclear by 0.8%; coal and gas by 0.6%.
- Europe (2031–2060) summer capacity decrease: 6.3–19%.
- United States (2031–2060) summer capacity decrease: 4.4–16%.
- Global EAD on power generator infrastructure: ~ $15 billion (~0.2% of global power-generation infrastructure value).
- Western US vulnerable stations: average summertime generating capacity reduction 1.1–3.0%; up to 7.2–8.8% under ten-year drought.
- Global power-generation vulnerabilities (2040–2069):
  - Hydropower usable capacity reductions for 61–74% of plants.
  - Thermoelectric reductions for 81–86% of plants.
- Global hydropower production change by end-century:
  - RCP8.5: −8% to +5%.
  - RCP4.5: −4% to +4%.

### WASH (Water/Sanitation/Hygiene)
- Global adaptation costs for raw water supply: $12 bn p.a.; 83–90% in developing countries; highest in Sub-Saharan Africa.
- Global adaptation costs small relative to baseline costs ($73 bn p.a.).
- US instream/abstractive water use costs: $136 to $327 billion per year (2007$).
- US national damages range between benefit of $12.3 billion p.a. (15% precipitation increase, modest warming 1.5 °C) and cost of $54.5 billion p.a. (warming 5 °C, no precipitation change).

---

### Annex III — tradeoff between costs and benefits of disaster risk reduction investment

### Key constructs and relationships
- Parameters p and ρ describe cost-benefit trade-off:
  - Investing in standard infrastructure reduces financing need by factor p.
  - Investing in resilient infrastructure reduces financing need by ρ when disaster occurs.
- Opportunity cost of resilient investment each year is 1/p and increases with relative cost of resilient vs standard capital and size of resilient capital stock.

### Damage and frequency formulation
- Damage to infrastructure capital over interval f (f ≥ 1, f = 1 annual shocks) written as: ρ D_R_{j−1} + D_S_{j−1}.
- Gross benefit (absolute) of investing in resilient infrastructure: |(ρ − 1) ∂ / f|.

### Drivers of net benefits
- Net benefits increase with:
  - Higher damage reduction factor ρ.
  - More frequent shocks (lower f).
  - Higher shock intensity D.
  - Smaller cost advantage of standard capital (p closer to 1).
- In current production function formulation resilient and standard public capital are perfect substitutes; trade-off independent of existing stock composition.

---

*Source: wpiea2025135-print-pdf — Annex I. Data Sources for Calibration; Sections and annexes from IMF Working Paper "A Macroeconomic Model for Climate and Sustainable Development: The C-SDG Model", Working Paper No. WP/2025/135*

### Annex I. Data Sources for Calibration ..................................................................................

### Annex I. Data Sources for Calibration

### Annexes listed in the content unit
- Annex I. Data Sources for Calibration ............................................................................................................. 32
- Annex II. Additional Studies on Specific Countries and Sectors ................................................................. 34
- Annex III. Tradeoff Between Costs and Benefits of Disaster Risk Reduction Investment ......................... 38
- References ......................................................................................................................................................... 39

### Figures included in the content unit
- 1. SDG-C Model Structure .................................................................................................................................... 7
- 2. Historical and Simulated Annual Average Temperature (ºC) .......................................................................... 15
- 3. Average Annual Precipitation Change Relative to 1985-2014 (percentage) ................................................... 17
- 4. Real GDP growth and Infrastructure: Adaptation Policy Scenario (Benin) ..................................................... 20
- 5. Public Debt, Reconstruction Costs, and Fiscal Balance: Adaptation Policy Scenario (Benin) ........................ 21
- 6. Additional SDG Spending Needs from Adaptation Policies (Benin) ................................................................ 23
- 7. Average Annual Temperature Change Relative to 1985-2014 (ºC) ................................................................ 24
- 8. Natural Disaster Impacts in Jamaica ............................................................................................................... 25
- 9. Real GDP growth and Infrastructure: Adaptation Policy Scenario (Jamaica) ................................................. 27
- 10. Public Debt, Reconstruction Costs, and Fiscal Balance: Adaptation Policy Scenario .................................. 27
- 11. SDG Spending Financing with Adaptation Policies (Jamaica) ...................................................................... 29

### Tables included in the content unit
- 1. List of Significant Disasters in Benin ............................................................................................................... 16
- 2. Parameters of Production Function (Benin) .................................................................................................... 18
- 3. Exogenous Macro-Fiscal Data and Sources (Benin) ...................................................................................... 20

*Source: wpiea2025135-print-pdf - Annex I. Data Sources for Calibration*

### 4. Baseline Macroeconomic and Financing assumptions (Jamaica) ..........................................................

### 4. Baseline Macroeconomic and Financing assumptions (Jamaica)

### I. Introduction
- Long-term growth and resilience require substantial public and private investment to meet the SDGs, which encompass 17 goals and 169 targets.
- Estimates of additional spending needs (for health, education, water and sanitation, electricity, and roads) in 2030:
  - LICs: approximately US$0.5 trillion (average incremental spending of 16 percentage points of GDP).
  - EMs: approximately US$2.35 trillion (average incremental spending of 5 percentage points of GDP).
- Financing options for incremental spending include: domestic resource mobilization, improvement in spending efficiency, public debt-financing, official development assistance, and private sector engagement.
- Examples of tax-to-GDP increases from reforms:
  - China between 1995-2002: 5 percentage point increase.
  - Georgia between 2004-2009: 12 percentage point increase.
- The medium-term revenue strategy (MTRS) approach is an IMF-supported instrument increasingly adopted to plan for SDG spending.

### II. Climate Change Impacts and Policies
- IPCC AR6 (2022) estimate: 3.3 to 3.6 billion people live in settings “highly vulnerable” to climate change.
- Slow-onset events defined (IPCC): increasing temperature means, desertification, decreasing precipitation, loss of biodiversity, land and forest degradation, glacial retreat and related impacts, ocean acidification, sea level rise and salinization.
- Extreme/fast-onset events defined (IPCC): occurrence of a weather/climate variable above/below a threshold (e.g., floods, storms, cyclones).
- Selected projected sector impacts from literature:
  - Global maize yields by end-of-century (2069–2099): decline of as much as 10 percent (low emissions scenario) and 20 percent (high emissions scenario) (Jägermeyr, 2021).
  - Global maximum catch potential of fishers projected to decrease almost 8 percent by 2050 under the high emissions scenario, leading to revenue decline of approximately 10 percent (Lam et al., 2016).
- World Bank’s Lifelines report (2019):
  - Disasters currently cost low- and middle-income countries US$ 390 billion a year in infrastructure damages.
  - Investing in resilient infrastructure costs 3 percent more but yields a total net benefit of US$ 4.2 trillion.
- Empirical macro-fiscal transmission evidence:
  - Droughts lead to deterioration of primary balance and public debt of EMDEs by as much as 1.4 and 1.5 percent of GDP, respectively (Fuje et al., 2023).
  - Under non-pegged exchange rate regimes, disasters can lead to nominal and real exchange rate depreciation of up to 6 percent two years following the disaster (Nguyen and Nguyen, forthcoming).
  - More frequent high temperatures, severe droughts, and fewer mild temperature days reduce GDP per capita by 0.2 percentage points (Akyapi et al., 2022).
  - Climate events in seven Central American countries resulted in a decline of monthly (year-on-year) economic activity by around 0.5-1 percentage point, while increasing remittances by more than 3 percentage points (Kim et al., 2022).
- Adaptation policy features:
  - Enhances physical, financial, and economic resilience; alleviates socioeconomic impacts of residual damages.
  - Public investment in resilient infrastructure, early warning systems, R&D for adaptation (e.g., new crops), social safety nets, financial inclusion, and ex-ante disaster risk financing.
  - Annual adaptation costs can exceed 1 percent of GDP for some developing countries and be above 10 percent of GDP for some island states (Aligishiev et al., 2022).
  - Aggarwal et al. (2024) estimate an additional 0.4 percent of global GDP required to account for mitigation and adaptation needs of SDG sectors related to human and physical capital development by 2030.
  - Private sector has mobilized considerably less resources for adaptation compared with mitigation (OECD, 2015; Deepali et al., 2024).
  - Domestic resource mobilization and international support (grants, loans) are critical given increased debt and debt distress after COVID-19.

### III. Model Description (C-SDG)
- Purpose: integrate analysis of climate change and sustainable development within a dynamic growth model to assess macro-fiscal trade-offs and synergies, focusing on adaptation and SDG financing.
- Lineage: built on SDG-FiT (Benedek et al., 2021); extended to include fast and slow-onset climate impacts and adaptation policies.
- Core components:
  - Endogenous accumulation of human and physical capital; fiscal policy interventions; public debt dynamics.
  - Two types of public physical capital: standard and resilient (resilient is more costly and less susceptible to climate damage).
  - Three economic sectors: primary, secondary, tertiary.
  - Three SDG-relevant public infrastructure assets: water, roads, electricity/power generation.
- Production function specifics:
  - Augmented Cobb-Douglas with private capital (K), public resilient capital (R), public standard capital (S), human capital (H), sector-specific labor (L), and TFP (A) which can be affected by slow-onset climate change (Ds s,i).
  - R = sum of resilient public capital across i = water, roads, electricity/power; S = sum of standard public capital across same i.
  - Parameter θ (>1 implies productivity of resilient infrastructure > standard); θ = 1 implies perfect substitutes.
  - Elasticity of output to public capital is sector specific (β_j); labor share set by ILO estimates; capital share = 1 - labor share.
- Capital accumulation and damage mechanics:
  - Resilient capital depreciation: δR ∈ (0,1); standard capital depreciation: δS ∈ (0,1).
  - Efficiency of public investment: e ∈ (0,1]; e = 1 denotes full efficiency.
  - Infrastructure damage from climate shock: D_ij ∈ (0, 1−δS).
  - Damage reduction (resilience effectiveness) parameter: ρ ∈ (0,1]; 0 = maximum damage reduction (no residual risk), 1 = no damage reduction.
  - Per shock gross benefit of investing in resilient capital: ∂Damage_effect/∂R = (ρ − 1) D (given ρ < 1, impact of shock declines with resilient investment).
  - Relative price of resilient vs standard capital: p ∈ (0,1]; resilient infrastructure more expensive (example framing: resilient costs $1 while standard costs $p).
- Public finances and allocation:
  - Total public financing = tax and non-tax revenues + grants + net borrowing (domestic and external).
  - Tax revenue determined as function of real per capita growth and exogenous elasticity; non-tax revenue follows observed trend.
  - Public expenditure allocation: capital (SDG infrastructure + additional cost of resilient capital + reconstruction expenditure) and current (PPPs subsidies, interest expense, other current expenditure, transfers due to climate shocks).
- Private capital follows Solow accumulation; can be reduced by climate shocks (sector specific).
- Human capital accumulation and scarring:
  - H_t = (1 − δ_h) H_{t−1} + ω * ξ_{t−1}.
  - ξ_{t−1} depends on public/private education and health spending (h), spending efficiency (ε), share of school-age population (n), and previous cohort human capital (1 − ω) ξ_{t−2}.
  - Climate shock D_h reduces the transformation of spending into new human capital (similar to lowering ε).
- Behavioral assumptions:
  - Household savings rates exogenous (reflect low financial inclusion and high hand-to-mouth consumers in developing countries).
  - Framework abstracts from some general equilibrium effects (price dynamics, labor market adjustments); suitable for macro-fiscal planning and scenario analysis.

### IV. Calibration
- Default parameterization draws on cross-country databases and academic literature; country-specific estimates used where available.
- Users recommended to calibrate climate impact parameters with country-specific fast and slow-onset data (meteorological agencies, sector ministries, CCDRs).
- IMF compiles relevant climate impact studies including FADCP Climate Dataset (Massetti and Tagklis, 2024), with temperature, precipitation, extreme event risks and sea level rise under alternative climate scenarios for all IMF member countries.
- Framework flexible to accommodate sectoral parameter adjustments (production technology, capital accumulation, climate impacts, adaptation policy).

### V. Country Applications (Benin example provided in source)
- Note: the framework is designed to include slow-onset effects, but the presented application focuses on discrete climate disasters.
- Benin context and vulnerability:
  - High vulnerability due to geography, economic structure, limited adaptive capacity (IMF Country Report 24/003).
  - Under a 2.7°C global warming scenario (current policies), Benin projected to have one of the highest exposures to extreme temperatures by 2070.
  - Estimated: 98 percent of Benin's landmass will be impacted by severe temperature conditions (IMF Country Report 24/003).
  - INFORM Climate Change Vulnerability Index: Benin ranks 163 out of 185 (where 185 is the most vulnerable).
  - Floods in 2010 affected over 830,000 people, or 9.1 percent of the population (EM-DAT).
- Financing and policy efforts:
  - 2022 NAP estimates adaptation needs at approximately US$4.2 billion by 2032.
  - World Bank CCDR additional financing needs due to climate change:
    - Annual investment of 0.1 percent of GDP until 2030.
    - 0.3 percent of GDP by 2040.
    - 0.8 percent of GDP by 2050.
    - Annual average of 0.3 percent of GDP over 2022–2050.
  - IMF support: Resilience and Sustainability Facility (RSF) financing of US$200 million.
  - Benin issued €500 million SDG bond in July 2021.
- Calibrating climate shocks for Benin:
  - Focus on floods: since 1980, exposed to a flood every 2.2 years, affecting as much as 15 percent of total population.
  - Precipitation projected to increase by:
    - 5.5 percent by 2070 under SSP2-4.5 (current policies).
    - 18.1 percent under the pessimistic scenario (SSP3-7.0 90th percentile).
  - World Bank CCDR: extreme floods can decrease GDP by more than 2 percent by 2050 (mean SSP3-7.0) and up to 6 percent (higher bound SSP3-7.0); additional capital losses between 4.9 and 6.5 percent of capital stock.
  - For Benin, urban capital stock damage rate estimated at approximately 26 percent for a 25-year return period flood by 2030 under a high emissions scenario (translated to national capital stock damage rate of 13 percent assuming urban flood affects infrastructure critical to secondary and tertiary sectors representing ~50 percent of total capital stock).
  - Flood simulation impacts 5.6 percent of the population, hindering human capital accumulation.
  - Extreme scenario: 1,000-year return period flood under high emissions estimates urban capital stock damage rate of approximately 37 percent (guides upper limit of adaptation investment).
- Parameter assumptions for resilient infrastructure based on Hallegatte et al. (2019) survey:
  - Incremental cost of resilient infrastructure for low- and middle-income countries estimated at 3 percent relative to standard infrastructure (p = 0.97).
  - Damage reduction from resilient infrastructure about a factor of 2 (implying ρ = 0.40).
- Benin simulation scenarios (baseline + five counterfactuals):
  - Baseline: current policies and growth projections (incorporates average weather-related shocks in WEO/DSA).
  - Counterfactual 1: two extreme floods in 2027 and 2035, each affecting 13 percent of public capital stock and 5.6 percent of population; reconstruction costs deficit financed, reconstruction completed within three years.
  - Counterfactual 2 (adaptation policy): 50 percent of new public investment directed toward resilient capital until targeted share of resilient infrastructure (19 percent) reached.
  - Robustness scenarios:
    - “Higher p”: larger relative price of resilient infrastructure.
    - “Higher R”: higher targeted share of resilient infrastructure.
  - Fiscal policy shock assumptions during floods:
    - Fiscal revenues decrease by 0.2 p.p. due to lower GDP and possible tax relief.
    - Government increases transfers to affected population by 0.1 p.p.
  - Scenario 5 (extreme benchmark): government unable to rebuild damaged public capital due to binding limited fiscal space.
- Benin simulation results (selected impacts):
  - Adaptation policy reduces public capital damages by 13 percent in the short-run (2027) and by 19 percent in the medium-run (2035), lowering GDP impacts of climate shocks.
  - By 2027, about 10 percent of the capital stock is resilient; by 2035, the share reaches the targeted level.
  - Extreme flood shocks reduce GDP growth by 2.1 p.p. relative to baseline in 2027 and 2035 in the scenario without adaptation.
  - With investment in resilient infrastructure, flood shocks reduce GDP growth by:
    - 2.0 p.p. in 2027.
    - 1.8 p.p. in 2035.
  - Adaptation policy therefore reduces GDP growth losses by 0.10 p.p. in 2027 and 0.30 p.p. in 2035 (implied by differences above).

*Source: IMF Working Paper — "A Macroeconomic Framework for Long-Term Resilience and Growth" (Chapter/Section: 4. Baseline Macroeconomic and Financing assumptions).*

### 0.34 p.p., respectively in 2027 and 2035, implying a growth yield of about 16 percent relative to the scenario

### IMF WORKING PAPERS A Macroeconomic Framework for Long-Term Resilience and Growth

### Macroeconomic and growth effects of adaptation policies
- Adaptation policies reduce public capital damages and reconstruction costs, lowering public debt levels and interest payments, and improving the fiscal balance.
- Calibration notes and comparative model result:
  - The paper reports a calibrated exercise implying a growth yield of about 16 percent relative to the scenario with the extreme flood shocks.
  - The DIGNAD model (IMF Country Report No. 24/003, page 78) shows an extreme climate event decreases GDP growth by 4 p.p. relative to the baseline and investing in resilient infrastructure reduces GDP growth losses by 1 p.p., implying a growth yield of about 25 percent. The simulation in the DIGNAD model is different in nature and parameter calibration.
- Adaptation policies create a damage reduction effect that maintains infrastructure for longer periods and reduces GDP losses compared to the scenario with standard infrastructure.

### Fiscal impacts, reconstruction, and debt dynamics
- By 2040 debt levels under adaptation policies are 1.6 p.p. lower compared to the current climate shock scenario; by 2054 they are 0.6 p.p. lower.
- Fiscal deficit impacts during reconstruction periods:
  - In a country with limited fiscal space, the overall fiscal deficit jumps by about 3.1 p.p. and 3.0 per year over the reconstruction period (i.e., 3 years) without the adaptation policy to rebuild destroyed public capital after the 2027 and 2035 shocks.
  - With the adaptation policy the fiscal deficit jumps by 2.7 and 2.5 p.p. per year for those reconstruction episodes (0.4 and 0.5 p.p. lower fiscal deficit), respectively, implying a fiscal deficit yield of about 18 percent.
- Consequences:
  - Interest payments are lower under adaptation policies, improving fiscal space and reducing crowding-out of public investment and other (SDG) spending over the medium term.
  - The framework assumes reconstruction of damaged capital can be deficit financed through non-concessional borrowing, avoiding crowding out other SDG spending; a more realistic scenario may involve partial crowding out and partial borrowing, potentially supported by grants and concessional financing from MDBs (left for future research).

### Sensitivity and scenario analysis
- Higher relative price for resilient public infrastructure:
  - If resilient infrastructure is 10 percent more expensive, the effectiveness of the adaptation policy is reduced due to lower public capital accumulation.
  - By 2054, GDP per capita and the public infrastructure stock are 0.04 and 0.1 percent lower, respectively, compared to the scenario without an adaptation policy.
  - Despite lower long-term GDP per capita in this narrow metric, the public debt-to-GDP ratio is 0.5 p.p. lower than in the scenario without an adaptation policy, underscoring multi-metric evaluation (economic stability, fiscal sustainability, resilience).
- Higher targeted share of resilient public infrastructure:
  - A larger share of resilient infrastructure increases the net benefits of the adaptation policy; with the larger share the debt-to-GDP is 2.4 p.p. lower (text truncated at that point in the source).

### Key exogenous macro-fiscal data (Benin) and sources
- Columns correspond to years: Real 2023, 2024, 2034, 2054.
- Real GDP growth rates: 5.8 6.0 6.0 5.7 (Source: WEO (Apr 2024) & SR No. 24/3)
- CPI inflation (percent): 2.8 3.0 2.0 2.0 (Source: WEO (Apr 2024) & SR No. 24/3)
- Revenue and Grants (percent of GDP): 15.2 15.2 17.1 18.9 (Source: WEO (Apr 2024) & SR No. 23/175 (DSA))
- Primary spending (percent of GDP): 17.7 17.4 18.6 20.4 (Source: WEO (Apr 2024) & SR No. 23/175 (DSA))
- Nominal interest rate on public debt (percent): 2.5 2.3 2.6 3.3 (Source: SR No. 23/175 (DSA))
- Overall fiscal balance (percent of GDP): -4.5 -3.7 -2.9 -2.9 (Source: WEO (Apr 2024) & SR No. 24/3)
- Efficiency of government spending (percent): 0.58 0.58 0.58 0.58 (Source: PIMA & C-PIMA (2023))

*Source: IMF staff estimates and calculations as presented in the source content.*

### 0.8 p.p. lower and the growth yield increases to 23 percent by 2040 and 2054, respectively. However, as

### A Macroeconomic Framework for Long-Term Resilience and Growth

### Benin: Impacts, Scenarios, and Adaptation Benefits
- Reconstruction versus no reconstruction:
  - No reconstruction scenario leads to GDP per capita 2.2 percent lower by 2040 relative to the scenario with reconstruction (and no adaptation policies).
- SDG financing gaps and climate shocks:
  - Without external grants, Benin is unable to achieve the SDGs under the baseline scenario.
  - Extreme climate shocks increase SDG financing needs by 9.65 percent of baseline GDP by 2040 if the economy does not invest in reconstruction.
  - Reconstruction reduces additional SDG financing needs to 2.47 percent of GDP by 2040.
- Adaptation (resilient public infrastructure) effects:
  - Adaptation policies narrow the financing gap by approximately US$120 million in 2027 (0.8 percent of GDP) and US$320 million in 2035 (1 percent of GDP) relative to the climate shock scenario with reconstruction (but no resilient public infrastructure).
  - By 2040, the financing gap with adaptation policies decreases by 0.04 percent of baseline GDP.
  - Cumulative discounted gains from adaptation policies by 2040 are estimated at approximately US$ 400 million, equivalent to nearly 3 percent of 2024 GDP or 24 percent of 2024 tax revenues.
  - The US$ 400 million is noted as comparable to two RSF programs (120 percent of quota, equivalent to about US$200 million).
- Trade-offs and policy considerations:
  - Resilient infrastructure is more expensive than standard infrastructure, and GDP per capita will be 0.2 percent lower by 2054 compared to the scenario without adaptation policy due to higher costs.
  - Governments must accurately identify exposed public capital and define the target share of resilient infrastructure to balance marginal returns on public funds.
  - Frequency of climate disasters, intensity, efficiency assumptions of resilient infrastructure, damage reduction factors, and relative prices of resilient infrastructure are critical factors shaping outcomes.

### Jamaica: Exposure, Modeling Assumptions, and Adaptation Outcomes
- Climate exposure and risks:
  - Jamaica ranks 47th out of 191 countries in the 2023 Inform Risk index.
  - Under a BAU global emission scenario (RCP 4.5), Jamaica is projected to face a 1.8°C increase of mean temperature by 2100 relative to the 1984–2014 baseline.
  - Tourism sector is equivalent to some 20 percent of GDP and is highly sensitive to climate effects.
- Historical disaster losses:
  - Hurricanes Ivan (2004) and Dean (2007) caused damages of US$ 580 million and US$ 329 million each (8 and 3 percent of GDP, respectively).
  - Tropical storm Nicole (2010) caused damages of US$ 239 million (2 percent of GDP).
  - Hurricane Beryl (July 2024) caused damages of about US$ 200 million (1.1 percent of GDP).
  - A one in 100 years event is estimated to produce losses exceeding US$ 1,729 million (roughly about 10 percent of GDP) with a one percent annual probability of exceeding that loss.
- Modeling calibration and scenarios:
  - A one in 100-year cyclone event is assumed to lead to national public capital stock damage of 8 percent of GDP.
  - Two extreme cyclone events are modeled in 2027 and 2035, each affecting 8 percent of public capital stock.
  - Reconstruction replaces damaged public infrastructure over a three-year period financed by a higher government deficit.
  - Adaptation scenario uses Hallegatte et al. (2019) estimates: additional cost of resilient infrastructure is 4 percent relative to standard infrastructure (p = 0.96), reducing damage by about a factor of 2 for new infrastructure (ρ = 0.44).
  - Additional sensitivity scenarios: “Higher p” (higher relative price of resilient infrastructure) and “Higher R” (higher share of resilient infrastructure). Also a “no reconstruction” scenario where government cannot finance reconstruction.
- Adaptation impacts on growth, debt, and SDG financing:
  - Adaptation policy reduces the shock to GDP growth by 0.2 percentage points of GDP by 2035 and reduces reconstruction costs by around 1.2 percent of GDP.
  - Share of resilient capital stock: 14 percent in 2027, rising to 28 percent by 2035 in the adaptation scenario.
  - With government access to financing, debt levels increase by around 10 percent of GDP higher than baseline by 2038 without adaptation; adaptation policy reduces the level of debt by 1.3 p.p. by 2038.
  - A higher cost of resilient capital leads to 0.2 percentage points lower capital stock by 2054 relative to the scenario with standard adaptation costs.
  - A higher targeted share of resilient infrastructure marginally increases effectiveness: shock to GDP in 2035 is reduced by only 0.1 pp compared to the adaptation scenario, and debt/GDP in 2038 is marginally lower by 0.5 pp.
  - Jamaica will not reach its SDG goals with extreme climate shocks in the baseline.
  - Climate shocks increase financing needs by 6.3 percent of GDP if the government does not invest in reconstruction.
  - Reconstruction efforts reduce additional SDG finance needs to 1.4 percent of GDP by 2040.
  - Adaptation policies narrow the financing gap to 1.35 percent of GDP by 2040.
  - Adaptation narrows the financing gap by approximately US$15 million in 2027 (0.4 percent of GDP) and US$50 million in 2035 (0.8 percent of GDP) relative to the climate shock scenario with reconstruction (but no resilient public infrastructure).
  - By 2040, the financing gap with adaptation policies decreases by 0.05 percent of baseline GDP.
  - Cumulative discounted gains from adaptation policies by 2040 are estimated at approximately US$ 64 million, equivalent to nearly 2 percent of 2024 GDP.
  - Sensitivity scenarios (higher p and higher R) do not significantly affect the SDG-financing gap.

### The C-SDG Model: Structure, Uses, and Core Findings
- Model features and objectives:
  - The C-SDG model integrates climate change analysis into macroeconomic projections and policies for sustainable development.
  - It incorporates endogenous human and physical capital accumulation, fiscal policy interventions, and public debt dynamics in a fully integrated macro framework.
  - Calibrated with default parameters from the literature and country-specific estimates where available; flexible to incorporate additional country information.
  - Captures damage channels from climate change: lower productivity from slow-onset drivers and increased frequency/severity of climate shocks.
  - Useful to quantify impacts on growth, debt, SDG achievement, and macro-fiscal costs and benefits of adaptation options.
- Key quantitative and qualitative insights:
  - Investing in adaptation policies and resilient infrastructure decreases economic damages of climate shocks relative to scenarios with standard infrastructure.
  - Net benefits of adaptation strategies depend on the country’s climate risk profile, available adaptation options, relative prices, and damage reduction potential.
  - Benefits of adaptation rise with frequency and magnitude of extreme events.
  - Early investment in resilience is important because building resilience takes time; delayed action reduces long-term benefits.
  - Climate spending, compared to overall SDG spending, is small on average but can spike sharply in shock years.

### Policy Recommendations, Trade-Offs, and Future Model Enhancements
- Policy guidance and trade-offs:
  - Adaptation policies can play a critical role in closing SDG financing gaps and accelerating progress toward sustainable development, especially for countries with limited fiscal space.
  - Governments should:
    - Accurately identify exposed public capital.
    - Define target shares of resilient infrastructure to balance trade-offs and marginal returns on public funds.
    - Integrate climate action into broader development agendas rather than treating climate as a separate objective.
    - Prioritize early resilience building to capture long-term benefits and mitigate acute spike needs in shock years.
  - Fiscal implications: reconstruction financed by higher deficits increases debt and interest payments, crowding out SDG spending unless mitigated by resilient investments that lower future reconstruction costs.
- Suggested model extensions and research priorities:
  - Improved representation of private sector risk reduction and recovery behaviors (including fiscal policy to motivate private climate investment).
  - Elaboration of alternative ex-ante risk financing options for response and recovery.
  - Modeling persistence of damages and longer recovery dynamics.
  - Flexibility on the design of optimal adaptation policy, including partial financing of reconstruction via donor finance.
  - Endogenous reaction of interest rates to climate shocks.
  - Further testing of shock-responsive social transfers and non-structural adaptation options (e.g., climate smart agriculture).
  - Application of the framework to a broader set of countries and inclusion of multiple hazards and slow-onset impacts via fully stochastic simulation to better capture macro-fiscal tail risks.

*IMF Working Papers — A Macroeconomic Framework for Long-Term Resilience and Growth*

### Annex I.  Data Sources for Calibration

### Annex I.  Data Sources for Calibration

### Climate Change Aspects — Data and Calibration Guidance
- Purpose: Calibrate macroeconomic and climate change related parameters for country applications of the SDG-C framework; default parameterization draws on comparable cross-country databases and may be adjusted to data availability.
- Disaster — Damages to Capital Stock
  - Calibration approach:
    - Use catastrophe modeling datasets that contain annual average losses and risk curves with probable maximum losses for alternative return periods.
    - For scenario analysis, select a particular disaster scenario from the existing risk curve, noting magnitudes and return periods.
    - Convert total infrastructure damages (valued per replacement costs) into damages as a percentage of capital stock using country total capital stock data (e.g., FAD capital stock dataset).
  - Suggested data sources:
    - Multi-hazard risk curves: Global Risk Assessment; Global Infrastructure Risk Model and Resilience Index (GIRI).
    - Single hazard risk curves: Aqueduct Floods.
- Slow-onset climate change — Decline of Total Factor Productivity
  - Two estimation approaches:
    - Top-down econometric estimates: Centorrino, Massetti, and Tagklis (forthcoming).
    - Bottom-up assessments: identify channels (e.g., labor, land), collect channel-specific impact estimates from biophysical and econometric studies, aggregate reductions in productivity weighting by input shares.
  - Example channel sources:
    - Labor impact: Dasgupta S, van Maanen N, Gosling SN, Piontek F, Otto C, Schleussner C-F. Effects of climate change on combined labour productivity and supply: an empirical, multi-model study. Lancet Planet Health 2021; 5: e455–65.
    - Agricultural impact: IFAD climate-adaptation assessment tool.

### Macroeconomic Aspects — Macro-fiscal Baseline Calibration
- Baseline data sources: IMF, World Economic Outlook; IMF’s Fiscal Monitor Database; and IMF Country Staff Reports.

---

### Annex II.  Additional Studies on Specific Countries and Sectors

### Roads Sector — Findings and Calibration Inputs
- Climate-inflicted damage costs to transports:
  - €2.5 billion/year (Cost accounting framework). Climate drivers: Floods, storms, droughts.
- Global Expected Annual Damages (EAD) to road and railway assets:
  - $3.1 to 22 billion (73% is caused by surface and river flooding).
  - Global EAD are 0.02% relative to global GDP. In some countries reach 0.5 to 1% of GDP annually.
- US adaptation costs for roads (2075 projections):
  - BAU scenario: adaptation costs rise to $6 billion annually in 2075.
  - Policy 3.7 and Policy 4.5 scenarios: remain about $2 billion per year through 2075.
- Vietnam — physical impact on roads:
  - Mean additional cost of maintaining same road network through 2050: US$10.5 billion.
- Africa — repair and maintenance cost through 2100:
  - US$183.6 billion total (US$2.3 billion per year) due to temperature and precipitation changes.
- Ghana — impact on road infrastructure:
  - $473 million to maintain and repair damages caused to existing roads (no adapt, 2020-2100).
  - Design and construction of new road infrastructure adaptation total cumulative cost: $678.47 million.
- South Africa — 2050 decade annual costs:
  - Median and maximum climate scenario cost USD$116.8 million and USD$228.7 million annually if no adaptation.
  - With proactive adaptation: these costs can be reduced to USD $55.7 million.
- Malawi, Mozambique, Zambia — through 2050:
  - $596 million based on median climate scenarios to maintain and repair roads.

### Energy Sector — Findings and Calibration Inputs
- Temperature effect on thermal power output:
  - 1°C increase reduces power output by approximately 0.45%, nuclear power output by 0.8% and coal and gas power output by 0.6% (thermal efficiency loss).
- Europe and United States (2031–2060) — summer average decrease in capacity of power plants:
  - Europe: 6.3–19%.
  - United States: 4.4–16%.
- Global EAD on power generator infrastructure:
  - Around $15 billion, or around 0.2% of the global value of the power generation infrastructure.
- Western United States — vulnerable power stations:
  - Climate change may reduce average summertime generating capacity by 1.1–3.0%, with reductions of up to 7.2–8.8% under a ten-year drought.
- Global power-generation system vulnerability (2040–2069):
  - Reductions in usable capacity for 61–74% of hydropower plants and 81–86% of thermoelectric power plants worldwide.
- Global hydropower production projections:
  - Change in net global hydropower production between −8% and +5% under RCP8.5 and between −4% and +4% under RCP4.5 by the end of the century.

### Water/Sanitation/Hygiene (WASH) — Findings and Calibration Inputs
- Global costs of climate change adaptation for supply of raw industrial and municipal water:
  - Global costs of $12 bn p.a., with 83–90% in developing countries; highest in Sub-Saharan Africa.
  - Globally, adaptation costs are low compared to baseline costs ($73 bn p.a.).
- US costs associated with changes in instream and abstractive water use:
  - Between $136 and $327 billion per year (adjusted to 2007$).
- National damages of climate change on U.S. water resources:
  - Range between a benefit of $12.3 billion per year (associated with a 15 % increase in precipitation and modest warming of 1.5 °C) and a cost of $54.5 billion per year (associated with warming of 5 °C and no change in precipitation nationally).

---

### Annex III.  Tradeoff Between Costs and Benefits of Disaster Risk Reduction Investment
- Key model constructs:
  - Parameters 휌 and p describe the trade-off between benefits and costs of adaptation policies in public infrastructure investment.
  - Investing in standard infrastructure reduces financing need by a factor p; investing in resilient infrastructure reduces financing need by 휌 when a disaster occurs.
  - Opportunity cost of investing in resilient infrastructure in each year is 1/p and is incurred every year; it increases with the relative cost of resilient versus standard capital and the size of resources invested in public resilient capital.
- Damage formulation and frequencies:
  - Damage to infrastructure capital associated with climate shocks is written as 휌D_R_{j−1} + D_S_{j−1} over interval f where f ≥ 1 represents the interval between occurrences of the shock (f = 1 implies annual shocks).
  - Gross benefit (in absolute value) of investing in resilient infrastructure: |(휌−1)휕 / f|.
- Drivers of net benefits:
  - Net benefits of resilient infrastructure investment increase with:
    - Higher damage reduction factor 휌.
    - More frequent climate shocks (lower f).
    - Higher intensity of climate shock (D).
    - Smaller cost advantage of standard capital (p closer to 1).
  - In current production function formulation, resilient and standard public capital are perfect substitutes; trade-off does not depend on existing stock composition.

---

*Source: Annex I. Data Sources for Calibration; Annex II; Annex III — from IMF Working Paper "A Macroeconomic Model for Climate and Sustainable Development: The C-SDG Model", Working Paper No. WP/2025/135*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025135-print-pdf.pdf_