Key Details
The IMF working paper Recurring Natural Disasters and Macroeconomic Scarring: A Risk-Adjusted DIGNAD Framework compares growth and debt paths with recurring disaster losses, then tests what changes when governments borrow to invest in resilient or conventional infrastructure.
Coverage: Model simulations span 33 disaster-prone economies over 2025–2058. India is not in the sample.
Disaster exposure: Estimated expected annual losses range from below 0.2% to nearly 10% of GDP across the countries studied.
Adaptation benefit: Additional resilient-infrastructure investment averts a median of about 25% of the projected output loss relative to a scenario with recurring disasters but no additional adaptation.
Debt cost: When the added investment is financed through borrowing, median public debt is approximately 5–6 percentage points of GDP higher by 2058 than in that baseline.
Like-for-like comparison: For the same additional investment spending, resilient infrastructure generally produces better growth and debt outcomes than conventional infrastructure exposed to repeated damage.
Uneven results: The gains vary with disaster exposure, public-investment efficiency, existing infrastructure, financing costs and fiscal capacity.
Repeated Damage Changes the Economic Baseline
Disaster analysis often asks what would happen if a country faced one severe event. The authors ask a different question: what happens to growth, revenues and debt when expected losses recur year after year? Repeated damage to public and private capital, alongside reduced productivity, can gradually lower output. A smaller economy then generates less revenue while reconstruction continues to require financing.
This means a country’s debt outlook can worsen without one exceptional catastrophe. Treating recurring losses only as occasional shocks may therefore produce an overly favourable starting point for assessing growth and public finances.
The Model Gives Expected Annual Losses a Place in Fiscal Projections
The authors extend the IMF’s Debt, Investment, Growth, and Natural Disasters (DIGNAD) model by incorporating Annual Average Loss: an estimate of expected yearly disaster damage drawn from a range of possible events. This creates a risk-adjusted projection against which investment choices can be compared.
The simulations represent that expected loss as a recurring annual shock. They are illustrative scenarios, not forecasts of when particular disasters will strike, and they do not replace stress tests for rare, catastrophic events.
Resilience Protects Output, but Financing Still Matters
In the simulations, borrowing to build resilient infrastructure reduces future damage more effectively than borrowing the same amount for conventional infrastructure. Standard assets add productive capacity, but can be damaged repeatedly; resilient assets better preserve that capacity.
The advantage does not erase the financing cost. Across the sample, adaptation averts about a quarter of the median projected output loss while leaving higher debt than the no-additional-investment baseline. Country outcomes differ sharply: a well-targeted project delivered efficiently may offer substantial protection, whereas high borrowing costs or limited implementation capacity can restrict the payoff.
What Is Annual Average Loss (AAL)?
Annual Average Loss (AAL) estimates the damage expected per year on average from the full range of possible disaster events, including infrequent severe ones. It does not mean a disaster of that size occurs every year. The paper uses AAL to make recurring risk visible in long-term economic and fiscal scenarios.
Policy Relevance
The paper offers a method for India to consider, not an estimate of India’s disaster losses or adaptation returns. For national and state fiscal planning, the useful distinction is between testing a single severe event and asking how repeated floods, cyclones, droughts or other hazards could affect the baseline used to appraise infrastructure and debt.
It also changes the comparison for public investment. Where an asset faces repeated hazard exposure, assessing only its upfront cost misses the potential value of avoided damage and reconstruction. But resilience spending should still be assessed against financing terms and delivery capacity; the paper does not show that every adaptation project pays for itself or that additional borrowing is always sustainable.
Follow the Full Working Paper Here: Recurring Natural Disasters and Macroeconomic Scarring: A Risk-Adjusted DIGNAD Framework