## wpiea2021096-print-pdf

## Source details

**Canonical URL:** [wpiea2021096-print-pdf](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021096-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021096-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021096-print-pdf.pdf.json)

---

### The model (DIGNAD / DIG)
- Framework: DIGNAD model (Marto, Papageorgiou, and Klyuev (2017)) extending the DIG model of Buffie and others (2012) to simulate natural-disaster impacts and adaptation.
- Type: real, dynamic, two-sector small open economy general equilibrium growth model with traded and non-traded sectors.
- Public capital: used as an input alongside private capital and labor; public capital is productive and can crowd in private investment.
- Key mechanisms captured:
  - Investment-growth nexus
  - Fiscal adjustment (government budget constraint and financing options)
  - Private sector response (investment, consumption behavior)
- Natural-disaster features: two forms of public capital—standard and adaptation infrastructure—with differing durability, returns, and damage susceptibility.

### A. Investment-Growth Nexus (model specifics)
- Production function (Cobb-Douglas):
  - 푦푡 = 퐴푡 (푘푡푔)^휓(푘푡)^훼(푙푡)^{1−훼}
- Parameter 휓 determines rate of return to installed public capital.
- Public investment raises marginal productivity of private capital and labor → can crowd in private investment.

### B. Fiscal Adjustment
- Government budget constraint includes:
  - Real exchange rate 푠푡; changes in external commercial ∆푎푡, external concessional ∆푑푡, and domestic ∆푏푡 borrowing with real interest rates 푟푎,푡, 푟푑,푡, 푟푡.
  - Government spending: public investment 푖푔,푡 and transfers/consumption spending 푔푡.
  - Revenues: grants 풜푡, natural resource royalties ℛ푡, and taxes 휏푗푡 on consumption or factor incomes.
- Grants and concessional borrowing treated as exogenously given; fiscal gap arises if insufficient.
- Financing options: external commercial and domestic borrowing, or tax/transfers adjustments to stabilize debt.
- Fiscal rules for tax instruments:
  - Non-linear rule: 휏푗푡 = Min(휏푗^R, 휏푗푡^B)
  - Adjustment dynamics for 휏푗푡^B with parameters 휆휏, 휆T > 0 and debt-stabilizing tax values; rules also apply to current expenditures/transfers with floors.
- Feasibility constraints and ceilings on tax adjustments can be imposed.

### C. Private Sector Response
- Private investment subject to real frictions (adjustment costs) and limited international capital market access.
- Heterogeneous consumers: asset-holders (can smooth consumption) and hand-to-mouth consumers.
- Distortionary taxes and hand-to-mouth consumers break Ricardian equivalence.
- Crowding effects:
  - Public investment increases public capital → raises marginal product of private capital → crowding in.
  - Domestic financing of public investment, tax increases, or transfer reductions → crowding out of private investment and consumption.
  - Long run: crowding in if projects are good; transition may see crowding out, especially if foreign financing is limited.

### D. Building Resilience to Natural Disasters
- Two public capital types: standard and adaptation infrastructure.
  - Adaptation infrastructure: greater durability, smaller damages after disasters, higher rate of return (climate-proofing allows more intensive use).
  - Can be complementary (seawalls, breakwater retrofitting, climate proofing) or substitutes (climate resilient infrastructure).
- Natural disaster channels:
  - (i) damages to private capital
  - (ii) damages to public capital
  - (iii) temporary productivity loss
  - (iv) decline in public investment efficiency during reconstruction
  - (v) loss in credit worthiness
- Reconstruction financing: domestic, concessional, external commercial debt, and/or donor grants; fiscal adjustments (taxes/transfers) needed to stabilize debt ratios.

### E. Capturing the Maldivian Context (key facts)
- Tradable sector captures tourism; non-tradable captures the rest.
- COVID-19 shock effects:
  - about 30 percent contraction in GDP
  - total publicly guaranteed debt rose from 78 percent of GDP (end-2019) to about 138 percent of GDP (end-2020Q3)
- Limited fiscal space constrains large investment scale-ups, especially for adaptation infrastructure, unless revenue increases or donor funds (grants or concessional debt) are available.
- Experiments consider investment options and financing designed to prevent accumulation of additional government debt.

### Calibration: selected initial values and parameters (annual frequency)
- Steady-state public standard investment to GDP 푖푧퐵,퐵: 11.9 percent
- Public adaptation investment to GDP 푖푧푎,퐵: 0.0 percent
- Public domestic debt to GDP 푏퐵: 37.5 percent
- Public concessional debt to GDP 푑퐵: 5.7 percent
- Public external (commercial) debt to GDP 푑푐,퐵: 12.0 percent
- Real interest rate on public domestic debt 푟퐵: 9.3 percent
- Real interest rate on public concessional debt 푟푑,퐵: 0.0 percent
- Real interest rate on public external debt 푟푑푐,퐵: 2.5 percent
- Disaster fund savings to GDP 푝푝퐵: 0.0
- Consumption tax (VAT) rate 휏퐵푐: 6.0 percent
- Private external debt to GDP 푏퐵∗: 12.3 percent
- Depreciation rate of standard public infrastructure 훿푧퐵: 7.5 (%)
- Depreciation rate of adaptation public infrastructure 훿푧푎: 3.0 (%)
- Depreciation rate of private capital 훿푘: 5.0 (%)
- Trend per capita growth rate G: 5.0 (%)
- Additional calibration notes:
  - Initial cumulative public and publicly guaranteed debt amounted to 55 percent of GDP in 2019 (model uses IMF and authorities’ debt data).
  - Ratio of financially constrained consumers set to 15 percent.
  - Assumed efficiency of public investment: 60 percent (standard and adaptation), aligning with IMF Public Investment Management Assessments for small developing countries.
  - Gross return on standard infrastructure at initial steady state: 25 percent; gross return on adaptation infrastructure set at twice that value.
  - Adaptation infrastructure depreciation implying an additional lifespan of about 75 years versus standard infrastructure.

### IV. Simulation design and key experiments
- Objective: analyze interaction of investment options and disaster shocks on macro outcomes; assess donor financing trade-offs.
- Investment plan for experiments:
  - No adaptation infrastructure in initial steady state.
  - Authorities use a budget envelope of 1 percent of GDP for 5 consecutive years to build either standard or adaptation infrastructure.
  - A natural disaster occurs immediately after the 5-year investment plan is completed (during the fifth year after investment started).
- Main questions:
  - Gains from adaptation infrastructure when a natural disaster occurs?
  - Outcomes if initial investment is financed by international grants rather than raising taxes?
  - Post-disaster reconstruction financing needs depending on prior investment type, and implications for international donors?
- Scenarios considered:
  1. Investment in standard infrastructure (financed within the budget envelope).
  2. Investment in adaptation infrastructure financed by higher taxation.
  3. Investment in adaptation infrastructure financed by international grants.
- Post-disaster reconstruction assumption: financed by increased taxation (Maldives assumed to have no fiscal space for deficit-financed spending); model computes taxes necessary to keep budget balanced without new debt issuance.
- Natural disaster shock implemented:
  - Destroys 10 percent of both private and public capital.
  - Triggers a symmetric productivity drop in both tradable and non-tradable sectors.
  - Shock magnitude illustrative and within range of shocks observed in disruptive floods in island countries.

### IV. Simulation results (illustrative findings)
- Reporting convention: all reported responses are percentage or percentage point deviations from initial steady state.

Impact (single stylized disaster destroying 10 percent of capital)
- GDP impact:
  - GDP declines about 1.2 percent with adaptation infrastructure capital.
  - GDP declines about 3 percent with standard infrastructure.
- Consumption and private investment on impact:
  - Both fall initially by about 15 percent.
- Public investment response:
  - Public investment spikes up to replace lost capital, driving the recovery.
  - In aggregate, total investment falls only marginally on impact as public effort steps up.
- Complementarities and reconstruction:
  - Public investment in standard capital can be partly intertwined with adaptation capital, requiring increases in standard capital investment during post-reconstruction.
  - Adaptation capital reinforces the quality of standard capital.
- Nonlinearities and initial conditions:
  - Gains from adaptation infrastructure after disaster shocks depend on shock size and initial conditions, such as magnitude and duration of ex-ante investment.
  - Private investment in the tradable sector leads in speed non-tradable investment during post-disaster phases.
- Recovery horizon:
  - The economy takes about three-four years to fully recover.
- Public debt to GDP:
  - Public debt to GDP increases by 1 percentage point more in the case of damages to standard capital (relative to adaptation capital case).
  - The worsening of the public financial position relative to GDP is primarily due to the denominator falling while the numerator remains constant at steady state (assumption: tax revenue increases to finance reconstruction).
- Tax pressure and household effects:
  - If the economy has only standard public infrastructure, the tax pressure would have to be twice as large (6-7 additional percentage points of GDP) than the case with adaptation capital.
  - Consumption of both Ricardian and non-Ricardian households suffers a larger hit, especially in the short run, mostly due to lower disposable income.
  - Footnote example: If the burden were to fall only on consumption taxes, the government would have to raise the average tax rate by about 18 percentage points in the short term and by about 2 percentage points over the medium term.
- Interaction with high initial debt:
  - Dynamic responses of GDP and other variables virtually unaltered if a disaster occurs when debt-to-GDP is already high; however, the percent point change in debt-to-GDP would be larger purely due to the higher initial debt level, constraining fiscal space.
- Role of grant financing:
  - Grant-financing can reduce pre- and post-disaster cost or eliminate it if it covers the investment plan in full.
- Pre-disaster financing effects:
  - Tax-financed investment in standard and adaptation infrastructure causes sacrifices in GDP, private consumption, and investment also in the pre-disaster period when public investment is ramping up.
- Robustness:
  - Results are robust within a reasonable range of parameters, including intertemporal discount rate, capital depreciation rates, and parameters governing infrastructure resilience.

Effects of a sequence of disaster shocks (annual small shocks)
- Motivation:
  - Frequent disasters can prevent full recovery; effects accumulate permanently on macroeconomic outcomes.
- Calibration approach:
  - Damage function calibrated so damage is a certain percent of GDP based on literature estimates.
  - Average economic damages from floods in a small sub-tropical economy placed near 0.3 percent of GDP yearly.
- Sequence simulation setup:
  - Investment scale-up program completed in year zero; sequence of shocks begins in year 1.
  - Shocks deliver damages of 0.3 percent on impact (yearly) in terms of GDP losses under the standard capital scenario.
- Ten-year outcomes:
  - Standard capital scenario: real GDP about 1 percent lower after 10 years of annual disasters of this size.
  - Adaptation capital scenario: output loss less than half—about 0.4 percent lower GDP after 10 years of consecutive shocks.
- Public investment for reconstruction after 10 years:
  - Standard infrastructure: level of investment dedicated to replenish damaged capital will be 2 percent higher than baseline.
  - Adaptation infrastructure: level will be 1.5 percent higher than baseline.
- Tax revenue to finance reconstruction after 10 years:
  - Standard infrastructure case: tax revenue would have to be almost 2 percentage points of GDP higher than baseline.
  - Adaptation infrastructure case: increase would have to be 1 percentage point of GDP higher than baseline.
- Private consumption:
  - Losses in private consumption are double in the standard capital economy compared with the adaptation capital economy.

Gains from adaptation under worsened climate conditions
- Climate projection context:
  - In South Asia, rainfalls expected to increase by 10-20 percent by the end of the 21st century (IPCC, 2018).
- Simulation approach:
  - Economy hit annually by natural disaster shocks; shock size increased by factor equal to projected increases in rainfalls.
  - Two alternative increases studied:
    - (i) 15 percent (intermediate value).
    - (ii) 30 percent (more severe, accounts also for sea-level rise and increased salinization).
- Cumulative output gains from adaptation (adaptation minus standard):
  - Historical climate conditions (baseline shocks of 0.3 percent of GDP yearly): output gain from investing in adaptation exceeds 0.6 percentage points over a ten-year horizon.
  - Projected +15 percent shock size: output gain increases by 22 percent to 0.8 percentage points.
  - Projected +30 percent shock size: output gain doubles relative to baseline change, reaching 0.94 percentage points.

Trade-offs for international donors
- Donor dilemma:
  - Cooperate to finance initial adaptation infrastructure (higher up-front cost) with prospects of much smaller post-disaster disbursements, or wait and finance reconstruction after disasters with potentially twice-as-large disbursements in standard infrastructure case.
- Cost differential:
  - Resilient structures could be 25 percent higher in cost than equivalent standard structures.
- Net present value (NPV) comparison over ten years (discounted to 2019 USD) for investment envelope of 0.5 percentage point of GDP (values in millions of 2019 U.S. dollars):
  - Discount rate 1%: Fiscal savings (A) = 21.1; Extra spending (B) = 7.5; Net savings (A)-(B) = 13.6.
  - Discount rate 3%: Fiscal savings (A) = 18.7; Extra spending (B) = 7.9; Net savings (A)-(B) = 10.8.
  - Discount rate 5%: Fiscal savings (A) = 16.7; Extra spending (B) = 8.4; Net savings (A)-(B) = 8.3.
- Example interpretation at 3% discount rate:
  - Ten-year saving could be as large as 18.7 million of 2019 USD.
  - Extra spending to build adaptation amounts to 7.9 million of 2019 USD.
  - Net saving is almost 11 million USD.
- Conclusion for donors:
  - Cumulated discounted fiscal savings from smaller damages with adaptation infrastructure are more than double the extra spending required to build it under historical climate conditions.
  - Worsened climate conditions imply higher savings from resilient infrastructure, making ex-ante donor support more attractive.
  - Ex-ante donor support tied to resilience investment can reduce moral hazard problems compared to ex-post intervention.

### Conclusions (summary)
- Model and calibration:
  - Employed the Debt-Investment-Growth (DIG) model extended for natural disasters (DIGNAD) and calibrated to the Maldives.
- Key findings:
  - Building infrastructure using adaptation technologies yields a dividend: adaptation capital sustains less damage and cuts GDP losses by more than half when hit by a natural disaster shock.
  - For sequences of annual, smaller-magnitude disasters calibrated to Maldivian floods (about yearly), cumulated output loss under adaptation capital is less than half that under standard capital in the long run.
  - Under worsened climate conditions, cumulative output gain from investing in adaptation can increase up to a factor of two.
- International cooperation:
  - Given high fiscal costs of adapting to worsening climate conditions, international community likely needs to step up cooperation and finance resilience building prior to disasters.
  - Cooperation should also include capacity building in public finance management given large investments required.
- Limitations and uncertainty:
  - Quantitative simulations rely on parameters set equal to cross-country averages because of Maldivian data limitations (e.g., productivity of infrastructure investment and resilience to natural disasters).
  - Assumptions about worsened climate conditions are surrounded by significant uncertainty.
  - Simulations use illustrative scenarios of infrastructure spending scale-ups; results may be refined as more data on authorities’ infrastructure projects become available.

*Source: wpiea2021096-print-pdf - conclusions are presented in Section V.*

### conclusions are presented in Section V.

### wpiea2021096-print-pdf - conclusions are presented in Section V.

### The model (DIGNAD / DIG)
- Framework: DIGNAD model (Marto, Papageorgiou, and Klyuev (2017)) extending the DIG model of Buffie and others (2012) to simulate natural-disaster impacts and adaptation.
- Type: real, dynamic, two-sector small open economy general equilibrium growth model with traded and non-traded sectors.
- Public capital: used as an input alongside private capital and labor; public capital is productive and can crowd in private investment.
- Key mechanisms captured:
  - Investment-growth nexus
  - Fiscal adjustment (government budget constraint and financing options)
  - Private sector response (investment, consumption behavior)
- Natural-disaster features: two forms of public capital—standard and adaptation infrastructure—with differing durability, returns, and damage susceptibility.

### A. Investment-Growth Nexus
- Production function: Cobb-Douglas form with public capital 푘푡푔, private capital 푘푡, labor 푙푡, and total factor productivity 퐴푡:
  - 푦푡 = 퐴푡 (푘푡푔)^휓(푘푡)^훼(푙푡)^{1−훼}
- Parameter 휓 determines rate of return to installed public capital.
- Public investment raises marginal productivity of private capital and labor → can crowd in private investment.

### B. Fiscal Adjustment
- Government budget constraint includes:
  - Real exchange rate 푠푡; changes in external commercial ∆푎푡, external concessional ∆푑푡, and domestic ∆푏푡 borrowing with real interest rates 푟푎,푡, 푟푑,푡, 푟푡.
  - Government spending: public investment 푖푔,푡 and transfers/consumption spending 푔푡.
  - Revenues: grants 풜푡, natural resource royalties ℛ푡, and taxes 휏푗푡 on consumption or factor incomes.
- Grants and concessional borrowing treated as exogenously given; fiscal gap arises if insufficient.
- Financing options: external commercial and domestic borrowing, or tax/transfers adjustments to stabilize debt.
- Fiscal rules for tax instruments:
  - Non-linear rule: 휏푗푡 = Min(휏푗^R, 휏푗푡^B)
  - Adjustment dynamics for 휏푗푡^B (equation (5)) with parameters 휆휏, 휆T > 0 and debt-stabilizing tax values; rules also apply to current expenditures/transfers with floors.
- Feasibility constraints and ceilings on tax adjustments can be imposed.

### C. Private Sector Response
- Private investment subject to real frictions (adjustment costs) and limited international capital market access.
- Heterogeneous consumers: asset-holders (can smooth consumption) and hand-to-mouth consumers.
- Distortionary taxes and hand-to-mouth consumers break Ricardian equivalence.
- Crowding effects:
  - Public investment increases public capital → raises marginal product of private capital → crowding in.
  - Domestic financing of public investment, tax increases, or transfer reductions → crowding out of private investment and consumption.
  - Long run: crowding in if projects are good; transition may see crowding out, especially if foreign financing is limited.

### D. Building Resilience to Natural Disasters
- Two public capital types: standard and adaptation infrastructure.
  - Adaptation infrastructure: greater durability, smaller damages after disasters, higher rate of return (climate-proofing allows more intensive use).
  - Can be complementary (seawalls, breakwater retrofitting, climate proofing) or substitutes (climate resilient infrastructure).
- Natural disaster channels:
  - (i) damages to private capital
  - (ii) damages to public capital
  - (iii) temporary productivity loss
  - (iv) decline in public investment efficiency during reconstruction
  - (v) loss in credit worthiness
- Reconstruction financing: domestic, concessional, external commercial debt, and/or donor grants; fiscal adjustments (taxes/transfers) needed to stabilize debt ratios.

### E. Capturing the Maldivian Context
- Tradable sector captures tourism; non-tradable captures the rest.
- COVID-19 shock effects:
  - about 30 percent contraction in GDP
  - total publicly guaranteed debt rose from 78 percent of GDP (end-2019) to about 138 percent of GDP (end-2020Q3)
- Limited fiscal space constrains large investment scale-ups, especially for adaptation infrastructure, unless revenue increases or donor funds (grants or concessional debt) are available.
- Experiments consider investment options and financing designed to prevent accumulation of additional government debt.

### Calibration: selected initial values and parameters (annual frequency)
- Steady-state public standard investment to GDP 푖푧퐵,퐵: 11.9 percent
- Public adaptation investment to GDP 푖푧푎,퐵: 0.0 percent
- Public domestic debt to GDP 푏퐵: 37.5 percent
- Public concessional debt to GDP 푑퐵: 5.7 percent
- Public external (commercial) debt to GDP 푑푐,퐵: 12.0 percent
- Real interest rate on public domestic debt 푟퐵: 9.3 percent
- Real interest rate on public concessional debt 푟푑,퐵: 0.0 percent
- Real interest rate on public external debt 푟푑푐,퐵: 2.5 percent
- Disaster fund savings to GDP 푝푝퐵: 0.0
- Consumption tax (VAT) rate 휏퐵푐: 6.0 percent
- Private external debt to GDP 푏퐵∗: 12.3 percent
- Depreciation rate of standard public infrastructure 훿푧퐵: 7.5 (%) 
- Depreciation rate of adaptation public infrastructure 훿푧푎: 3.0 (%) 
- Depreciation rate of private capital 훿푘: 5.0 (%) 
- Trend per capita growth rate G: 5.0 (%)
- Additional calibration notes:
  - Initial cumulative public and publicly guaranteed debt amounted to 55 percent of GDP in 2019 (model uses IMF and authorities’ debt data).
  - Ratio of financially constrained consumers set to 15 percent.
  - Assumed efficiency of public investment: 60 percent (standard and adaptation), aligning with IMF Public Investment Management Assessments for small developing countries.
  - Gross return on standard infrastructure at initial steady state: 25 percent; gross return on adaptation infrastructure set at twice that value.
  - Adaptation infrastructure depreciation implying an additional lifespan of about 75 years versus standard infrastructure.

### IV. Simulation design and key experiments
- Objective: analyze interaction of investment options and disaster shocks on macro outcomes; assess donor financing trade-offs.
- Investment plan for experiments:
  - No adaptation infrastructure in initial steady state.
  - Authorities use a budget envelope of 1 percent of GDP for 5 consecutive years to build either standard or adaptation infrastructure.
  - A natural disaster occurs immediately after the 5-year investment plan is completed (during the fifth year after investment started).
- Main questions addressed:
  - What are the gains from adaptation infrastructure when a natural disaster occurs?
  - How do outcomes differ if initial investment is financed by international grants rather than raising taxes?
  - What are post-disaster reconstruction financing needs depending on prior investment type, and implications for international donors?
- Scenarios considered:
  1. Investment in standard infrastructure (financed within the budget envelope).
  2. Investment in adaptation infrastructure financed by higher taxation.
  3. Investment in adaptation infrastructure financed by international grants.
- Post-disaster reconstruction financing assumption: financed by increased taxation (Maldives assumed to have no fiscal space for deficit-financed spending); model computes taxes necessary to keep budget balanced without new debt issuance.
- Natural disaster shock implemented:
  - Destroys 10 percent of both private and public capital.
  - Triggers a symmetric productivity drop in both tradable and non-tradable sectors.
  - Shock magnitude illustrative and within range of shocks observed in disruptive floods in island countries.

### IV. Simulation results (illustrative findings)
- All reported responses are percentage or percentage point deviations from initial steady state.
- During the accumulation (investment) phase:
  - Real GDP and private investment increase when the government steps up infrastructure investment.
  - Investing in adaptation capital raises the marginal product of private capital more than investing in standard infrastructure.
  - Adaptation investment crowds in a larger fraction of private investment prior to the disaster due to:
    - Lower depreciation rate (greater durability)
    - Higher gross return (climate-proofing increases average usage)
- Simulation setup allows direct comparison per “taxpayer dollar” spent by using the same 1 percent of GDP budget envelope for standard versus adaptation investment.
- Further analyses in the source examine outcomes under worsened climate conditions and donor pre-disaster versus post-disaster financial interventions (details in Section IV.D and subsequent subsections).

*Source: wpiea2021096-print-pdf - conclusions are presented in Section V.*

### Section III).

### Section III)

### Impact of a stylized disaster shock on selected macroeconomic variables
- Scenario: natural disaster destroys 10 percent of capital in all sectors.
- GDP impact:
  - GDP declines about 1.2 percent with adaptation infrastructure capital.
  - GDP declines about 3 percent with standard infrastructure.
- Consumption and private investment on impact:
  - Both fall initially by about 15 percent.
- Public investment response:
  - Public investment spikes up to replace lost capital, driving the recovery.
  - In aggregate, total investment falls only marginally on impact as public effort steps up.
- Complementarities and reconstruction:
  - Public investment in standard capital can be partly intertwined with adaptation capital, requiring increases in standard capital investment during post-reconstruction.
  - Adaptation capital reinforces the quality of standard capital (example: climate-proofing roads).
- Nonlinearities and initial conditions:
  - Gains from adaptation infrastructure after disaster shocks depend on shock size and initial conditions, such as magnitude and duration of ex-ante investment.
  - Private investment in the tradable sector leads in speed non-tradable investment during post-disaster phases.
- Recovery horizon:
  - The economy takes about three-four years to fully recover.
- Public debt to GDP:
  - Public debt to GDP increases by 1 percentage point more in the case of damages to standard capital (relative to adaptation capital case).
  - The worsening of the public financial position relative to GDP is primarily due to the denominator falling while the numerator remains constant at steady state (assumption: tax revenue increases to finance reconstruction).
- Tax pressure and household effects:
  - If the economy has only standard public infrastructure, the tax pressure would have to be twice as large (6-7 additional percentage points of GDP) than the case with adaptation capital.
  - Consumption of both Ricardian and non-Ricardian households suffers a larger hit, especially in the short run, mostly due to lower disposable income.
  - Footnote: If the burden were to fall only on consumption taxes, the government would have to raise the average tax rate by about 18 percentage points in the short term and by about 2 percentage points over the medium term.
- Interaction with high initial debt:
  - The model calibrated to Maldivian historical averages shows dynamic responses of GDP and other variables virtually unaltered if a disaster occurs when debt-to-GDP is already high; however, the percent point change in debt-to-GDP would be larger purely due to the higher initial debt level, constraining fiscal space.
- Role of grant financing:
  - Grant-financing can reduce pre- and post-disaster cost or eliminate it if it covers the investment plan in full.
- Pre-disaster financing effects:
  - Tax-financed investment in standard and adaptation infrastructure causes sacrifices in GDP, private consumption, and investment also in the pre-disaster period when public investment is ramping up.
- Robustness:
  - Results are robust within a reasonable range of parameters, including intertemporal discount rate, capital depreciation rates, and parameters governing infrastructure resilience.

### Effects of a sequence of disaster shocks
- Motivation:
  - Many disaster-prone countries experience disasters more frequently and may not fully recover before another shock occurs; effects accumulate permanently on macroeconomic outcomes.
- Calibration approach:
  - Damage function calibrated so damage is a certain percent of GDP based on literature estimates (Bayoumi et al., 2020).
  - Average economic damages from floods in a small sub-tropical economy placed near 0.3 percent of GDP yearly.
- Sequence simulation setup:
  - Investment scale-up program completed in year zero; sequence of shocks begins in year 1.
  - Shocks deliver damages of 0.3 percent on impact (yearly) in terms of GDP losses under the standard capital scenario.
  - Cumulative impact calculated considering overlapping impacts of sequential shocks while the economy is on a transition path.
  - All percentage deviations relative to a baseline without disaster shocks.
- Ten-year outcomes:
  - Standard capital scenario (blue line): real GDP about 1 percent lower after 10 years of annual disasters of this size.
  - Adaptation capital scenario (red line): output loss less than half—about 0.4 percent lower GDP after 10 years of consecutive shocks.
- Public investment for reconstruction after 10 years:
  - In case of standard infrastructure: level of investment dedicated to replenish damaged capital will be 2 percent higher than baseline.
  - In case of adaptation infrastructure: level will be 1.5 percent higher than baseline.
- Tax revenue to finance reconstruction after 10 years:
  - Standard infrastructure case: tax revenue would have to be almost 2 percentage points of GDP higher than baseline.
  - Adaptation infrastructure case: increase would have to be 1 percentage point of GDP higher than baseline.
- Private consumption:
  - Losses in private consumption are double in the standard capital economy compared with the adaptation capital economy.

### Gains from adaptation under worsened climate conditions
- Climate projection context:
  - In South Asia, rainfalls expected to increase by 10-20 percent by the end of the 21st century (IPCC, 2018).
- Simulation approach:
  - Economy hit annually by natural disaster shocks; shock size increased by factor equal to projected increases in rainfalls.
  - Two alternative increases studied:
    - (i) 15 percent (intermediate value).
    - (ii) 30 percent (more severe, accounts also for sea-level rise and increased salinization).
- Cumulative output gains from adaptation (difference: adaptation minus standard):
  - Under historical climate conditions (baseline shocks of 0.3 percent of GDP yearly): output gain from investing in adaptation exceeds 0.6 percentage points over a ten-year horizon.
  - Under projected +15 percent shock size: output gain increases by 22 percent to 0.8 percentage points.
  - Under projected +30 percent shock size: output gain doubles relative to baseline change, reaching 0.94 percentage points.

### Trade-offs for international donors
- Donor dilemma:
  - Cooperate to finance initial investment in adaptation infrastructure (higher up-front cost) with prospects of much smaller post-disaster disbursements, or wait and finance reconstruction after disasters with potentially twice-as-large disbursements in standard infrastructure case.
- Cost differential:
  - Resilient structures could be 25 percent higher in cost than equivalent standard structures (Cantelmo, Melina and Papageorgiou, 2019).
- Net present value (NPV) comparison over ten years (discounted to 2019 USD) for a small investment envelope of 0.5 percentage point of GDP:
  - Table 2 results (all values in millions of 2019 U.S. dollars):
    - Discount rate 1%: Fiscal savings (A) = 21.1; Extra spending (B) = 7.5; Net savings (A)-(B) = 13.6.
    - Discount rate 3%: Fiscal savings (A) = 18.7; Extra spending (B) = 7.9; Net savings (A)-(B) = 10.8.
    - Discount rate 5%: Fiscal savings (A) = 16.7; Extra spending (B) = 8.4; Net savings (A)-(B) = 8.3.
  - Example interpretation at 3% discount rate:
    - Ten-year saving could be as large as 18.7 million of 2019 USD.
    - Extra spending to build adaptation amounts to 7.9 million of 2019 USD.
    - Net saving is almost 11 million USD.
  - Conclusion: Cumulated discounted fiscal savings from smaller damages with adaptation infrastructure are more than double the extra spending required to build it under historical climate conditions.
- Climate change caveat:
  - These calculations represent a lower bound for future costs; worsened climate conditions imply higher savings from resilient infrastructure, making ex-ante donor support more attractive.
- Additional benefits of ex-ante support:
  - Ex-ante donor support tied to resilience investment can reduce moral hazard problems compared to ex-post intervention.
- Policy implication for donors:
  - Donors face a trade-off between higher initial costs to fund adaptation infrastructure and potentially larger, recurring reconstruction costs if waiting until after disasters; ex-ante financing of resilience can be financially convenient and reduce future fiscal burdens.

### Conclusions
- Model and calibration:
  - Employed the Debt-Investment-Growth (DIG) model extended for natural disasters (DIGNAD) and calibrated to the Maldives.
- Key findings:
  - Building infrastructure using adaptation technologies yields a dividend: adaptation capital sustains less damage and cuts GDP losses by more than half when hit by a natural disaster shock.
  - For sequences of annual, smaller-magnitude disasters calibrated to Maldivian floods (about yearly), cumulated output loss under adaptation capital is less than half that under standard capital in the long run.
  - Under worsened climate conditions, cumulative output gain from investing in adaptation can increase up to a factor of two.
- International cooperation:
  - Given high fiscal costs of adapting to worsening climate conditions, international community likely needs to step up cooperation and finance resilience building prior to disasters.
  - Cooperation should also include capacity building in public finance management given large investments required.
- Limitations and uncertainty:
  - Quantitative simulations rely on parameters set equal to cross-country averages because of Maldivian data limitations (e.g., productivity of infrastructure investment and resilience to natural disasters).
  - Assumptions about worsened climate conditions are surrounded by significant uncertainty.
  - Simulations use illustrative scenarios of infrastructure spending scale-ups; results may be refined as more data on authorities’ infrastructure projects become available.

*Source: wpiea2021096-print-pdf - Section III).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021096-print-pdf.pdf_