The Last Mile of Adaptation

The technology to move disaster money before a storm arrives already exists, and it works. In Fiji, a parametric insurance scheme has paid cyclone claims within days rather than months, releasing part of the payout before landfall so families can act while there is still time. For a premium of US$2,500 to US$5,000, cooperatives of farmers, fishers and market vendors hold cover of US$25,000 to US$50,000, and the model has since spread to Papua New Guinea, Samoa, Solomon Islands, Vanuatu, Tonga and Kiribati. Yet across the humanitarian system, money arranged before a shock rather than scrambled after one came to just 2.7 percent of crisis financing in 2021. The barrier to earlier action is no longer knowing what to build.

It is a question of who is responsible, who holds the budget, and on whose calendar the money moves. Across income levels, the duty to zone floodplains, clear drainage, run local warning and order an evacuation has devolved to municipal and district governments. The finance and the hazard data have not followed them down. The office now accountable for the last mile of climate defense is frequently not the office holding the money or the exposure maps.

Read through these numbers, the familiar story of local “failure to prepare” dissolves into two problems that happen to share a victim. The first is a gap between authority and responsibility: the tier now charged with protection often lacks the engineers, the models and the standing protocols to act on a warning. The second is a gap between funding and implementation: the money exists somewhere in the system but arrives on an annual, national, post-disaster cycle that outlasts the window in which action would have mattered. Only the second is a money problem, and treating both as one is why a decade of larger climate pledges has changed so little on the ground.

The capacity gap will not close with transfers. A district that cannot read a flood model, or lacks a protocol to act on a forecast, does not need a larger grant; it needs the engineer and the protocol. Early-warning coverage has expanded quickly: 119 countries now run multi-hazard systems, up 113 percent since 2015. But national coverage says little about whether a district can act, and the least-protected places are the poorest. A warning that reaches a capital and stops there is a statistic, not a defense.

The funding gap, unlike the capacity gap, is close to solved in principle. Instruments that pay against a trigger rather than a damage assessment — parametric insurance, contingent credit lines, forecast-based financing — compress the lag from months to days and can release money before impact. The Caribbean’s sovereign risk pool has done this for governments for two decades; Fiji has now done it for cooperatives. What keeps these at the margin of crisis budgets is not doubt about whether they work. It is the annual, nation-to-nation architecture of public finance into which a trigger-based, sub-national instrument must be fitted, and the political discomfort of paying a premium for a disaster that may not come.

A fiscal conservative would answer that the fix is simply to devolve the money and let localities decide. But money without the capacity to spend it well produces waste, not resilience, which is why the two gaps have to be closed in sequence rather than swapped for each other. The harder obstacle is political. Pre-arranged finance asks a government to spend on a disaster that has not happened, and sometimes to pay for a cyclone that never arrives; democracies reward the visible rescue, not the flood that quietly did not drown a neighborhood. Until the budget and the data sit with the office that now carries the duty, and a mayor is credited for the emergency that never came, the last mile of adaptation will keep being run by whoever happens to be standing in the water when it rises.