Climate change is projected to increase the frequency and severity of natural disasters in many countries. Emissions reductions will not fully ameliorate this, at least in the medium term. Therefore, climate policy attention is increasingly turning towards adaptation: investments to limit the damages from natural disasters when they occur. For example, coastal protections to limit flooding from cyclones, or early warning systems to prevent wildfire casualties. Such adaptive investments are however costly, and many disaster-prone economies face high costs of borrowing. This paper studies the implications of sovereign risk for adaptation and therefore the economic losses from natural disasters. Bringing together model and data, I find that sovereign risk can substantially restrict adaptation investment, increasing the costs of climate change for exposed economies.
The ‘Adaptation Trap’ is outlined in the model. Economies with high costs of borrowing are constrained and therefore spend less on adaptation. Due to this lower investment in adaptation these economies then suffer larger damages from natural disasters when they arise. This lack of resilience increases the probability of sovereign default following a disaster, which in turn increases borrowing costs further. The model therefore predicts a negative feedback loop between high costs of borrowing, low adaptation, and high losses from natural disasters. Economies with high existing debt and significant climate exposure become locked in to low adaptation and a high likelihood of climate-related fiscal crises in the future.
I validate the predictions of this model using a newly constructed dataset of public adaptation expenditures in Latin America and the Caribbean. Between 2014 and 2025 these governments spent on average 0.3% of GDP on adaptation. Countries spend more on adaptation when sovereign credit ratings improve: a one standard deviation improvement in credit rating is associated with an additional $250 million spent on adaptation per year. Moreover, hurricanes cause the probability of default to increase, especially for those governments who have spent less on adaptation.
This adaptation trap is quantitatively important. On average, Caribbean economies invest 13% less in adaptation than they would have in the absence of default risk. As a result, the GDP loss per hurricane in this region is 10% larger. As climate change intensifies this wedge grows. By the end of the century the model predicts that the GDP loss per hurricane in the Caribbean will be 13% larger than it would have been absent default risk.
These substantial losses bring into focus the potential gains from improvements in adaptation financing. The paper evaluates two instruments. First, a long-term interest-free loan of the kind offered through the IMF’s Resilience and Sustainability Trust. A loan of 10% of GDP leads governments to increase adaptation investment by 5%, reducing hurricane losses by 40% of the maximum achievable reduction from eliminating default risk entirely. An ‘adaptation bond’ which reduces coupon payments once a government reaches a pre-specified adaptation target increases adaptation investment by 10%, closing 80% of the gap in the GDP loss per hurricane between the baseline model and the counterfactual without default risk. The potential gains from making debt adaptation-contingent are therefore large. By allowing governments fiscal space to invest in adaptation, such a bond reduces climate-related losses and default probabilities.