Scaling Climate Adaptation Through Development Finance and Technology Diffusion

Executive Summary

The resilience gap is usually described as a shortage of money. It is better understood as two different gaps wearing one name — and the reason neither closes is that both are treated as the first one.

Adaptation is the part of climate policy with the least argument and the least delivery. Almost no one disputes that a country facing more frequent floods should build drainage, or that a farmer facing an erratic monsoon should get a forecast. The economics are, by the standards of public investment, extraordinary. And yet the distance between what vulnerable economies need and what they receive is not narrowing but widening — a gap now running to an order of magnitude, in a decade when the hazard is intensifying.

This report argues that the persistence of that gap is a delivery failure rather than a knowledge failure, and that it has two distinct causes which conventional adaptation policy conflates. Development finance is the instrument that opens the first gate: it changes the price and tenor of capital, which is what determines whether a capital-intensive protective asset can be built at all. Technology diffusion is the instrument that opens the second: it changes the speed at which a proven, cheap intervention reaches the last mile. A programme that supplies concessional finance to a country with no delivery channel buys plans. A programme that supplies technology to a country that cannot finance the infrastructure it depends on buys pilots. Closing the resilience gap requires both gates open at once.

1. The Gap, Measured

Before diagnosing why the gap persists, it is worth establishing its size, its direction of travel, and the losses accumulating on the other side of it.

The authoritative measurement is the UN Environment Programme's annual assessment. Its Adaptation Gap Report 2025 puts the adaptation needs of developing countries at roughly US$310–365 billion a year by 2035. Against that, international public adaptation finance reached those countries at just US$26 billion in 2023 — lower than the US$28 billion of the previous year. Needs are running some twelve to fourteen times ahead of flows, and the two lines are diverging.

The figure makes an uncomfortable point about ambition. The COP30 commitment to at least triple adaptation finance from its roughly US$40 billion 2025 level — toward something near US$120 billion a year by 2035 — would be the largest proportional increase adaptation finance has ever received. It would also still leave roughly two-thirds of assessed need unmet. The gap is not the kind that closes through incremental generosity.

1.1 What the gap costs while it stays open

A financing gap is an abstraction; the losses it fails to prevent are not. Natural hazards already cost low- and middle-income countries on the order of US$390 billion a year, equivalent to one to two per cent of their combined GDP, and the International Finance Corporation's 2026 assessment projects that unmanaged climate risk could put some 43 million jobs across 49 countries at risk by 2050.

The forward-looking figures are larger still, and unevenly distributed in a way that matters for this report's argument. The Swiss Re Institute's climate stress test, summarised by the World Economic Forum, found that global GDP could be some 11 to 14 per cent smaller by mid-century than in a no-warming baseline even on a path consistent with well below 2°C, and 18 per cent smaller under severe warming — with the emerging economies of South and Southeast Asia facing losses two to three times the global average.

2. Two Gates, One Gap

The analytical core: every unit of adaptation must pass a capital gate and a delivery gate to become installed resilience — and which gate binds depends on what is being adapted.

Adaptation is discussed as though it were one commodity purchased with one budget. It is not. Two very different classes of intervention sit under the word, and they fail for opposite reasons.

Hard adaptation is physical, lumpy and long-lived: embankments, drainage, water storage, climate-proofed roads, hardened power and telecommunications. Its costs are almost entirely up front and its benefits accrue over decades, which makes its viability a function of the price and tenor of capital more than of anything else. This is the class that dominates the dollar value of the assessed need, and it is capital-constrained.

Soft adaptation is informational and institutional: multi-hazard early warning, digital climate advisory for farmers, parametric insurance, adaptive social protection, heat action plans. Its unit costs are trivial by comparison — a forecast delivered to a phone costs cents — and its returns are among the highest measured anywhere in development. It is not, in any meaningful sense, blocked by a shortage of money. It is blocked by the absence of channels: the payment rails, the identity systems, the extension networks, the local institutions and the user trust required to get a cheap thing to a hundred million people.

Table 1 — Two classes of adaptation, two binding constraints

Classifications are analytical judgements drawn from the evidence in Sections 3 to 5, intended to show the pattern rather than to score individual programmes. "Both" marks interventions that are genuinely constrained at each gate.

Stating the distinction this way makes a diagnostic error visible. Because the dollar value of the need is dominated by hard adaptation, the whole field is discussed in the vocabulary of finance — billions pledged, gaps in flows, tripling commitments. That vocabulary is correct for the capital gate and close to irrelevant for the delivery gate. Soft adaptation does not need a larger share of a US$26 billion pool; it needs the rails that carried mobile money to hundreds of millions of people in under two decades. Conversely, no amount of digital ingenuity will drain a flooding city.

3. The Capital Gate

Why globally cheap protection is locally unaffordable, and what development finance actually has to change.

Hard adaptation shares an economic signature with clean energy: it is capital-intensive, front-loaded and fuel-free, so its lifetime cost is dominated by the price of money rather than by the price of the thing. That makes the cost of capital, not the engineering, the decisive variable. Emerging-market sovereigns and the projects they host borrow at rates inflated by currency risk, perceived political risk and thin local capital markets — a wedge the IEA commonly puts at 400 to 800 basis points above comparable advanced-economy rates. An embankment with a fifty-year life financed at fifteen per cent is a different proposition from the same embankment financed at four.

Fiscal space compounds the problem. UNCTAD reports developing-country external debt at a record US$11.4 trillion, with net interest payments of US$921 billion in 2024. A government spending more on debt service than on health does not have a spare capital budget for assets whose returns arrive in the 2050s, however good those returns are. The capital gate is therefore not only about the price of new money; it is about whether there is any fiscal room to use it.

3.1 The appraisal problem

There is a second, quieter reason hard adaptation is under-financed, and it is an accounting one. Adaptation is conventionally appraised as insurance: its value is the expected damage avoided, discounted by the probability that the damage occurs. On that basis a protective asset looks like a bet on a bad outcome, which is a difficult case to take to a finance ministry weighing it against a school.

The evidence says the framing is wrong. The World Resources Institute's 2025 study of 320 real adaptation investments across twelve countries found that more than half of the benefits accrued whether or not a climate disaster actually occurred — through higher yields, lower insurance costs, better land values, more reliable services and reduced business interruption. Most of the return, in other words, is not contingent. An asset delivering that profile is not insurance; it is infrastructure with an unusually good risk-adjusted return, and it should be financed like one.

The same logic appears in the design margin. Across the four essential systems the World Bank examined in Lifelines, building resilience in from the design stage adds only about 3 to 5 per cent to upfront capital costs, while the damage and service disruption avoided over an asset's life routinely runs to 50 to 100 per cent of the asset's value. Resilience is cheap when specified at the drawing board and expensive when retrofitted after a loss — which makes the current pipeline of new infrastructure across the emerging world a closing window rather than a standing option.

3.2 What is actually on the table

The instruments that open this gate exist and are, for the first time, being assembled into something resembling an architecture. The IMF's Resilience and Sustainability Trust, capitalised through rechannelled Special Drawing Rights, supplies long-tenor concessional financing tied to climate-policy reform — tenor being precisely what a fifty-year asset needs. The Fund for responding to Loss and Damage addresses the harm that adaptation fails to prevent. Climate-resilient debt clauses suspend repayments when a disaster strikes, keeping a shock from becoming a debt crisis. The Sevilla Commitment, the outcome of the 2025 Financing for Development conference and the first such framework since 2015, set out an agenda spanning debt, multilateral bank reform and domestic resource mobilisation.

What remains missing is scale and, above all, price. These instruments are correctly designed and radically under-capitalised relative to a US$310–365 billion annual need. The negotiating position that has moved them furthest has come from the vulnerable countries themselves — the Bridgetown Initiative and the V20 group of climate-vulnerable economies, which converted individual appeals for finance into a common agenda on debt, bank reform and the cost of capital.

3.3 Frontier economies: where the gate is heaviest

Within the developing world, a subset faces the capital gate in its most closed form. Frontier economies — small, low-income, often single-hazard-exposed, with shallow domestic capital markets and limited or episodic access to international bond markets — combine the highest vulnerability with the least ability to borrow against it. They are too small to interest large project financiers, too risky for commercial tenor, and too fiscally constrained to self-fund. For this group the marginal instrument is not a better bond spread but concessional capital and guarantees, pooled at regional scale so that individually unbankable projects become a portfolio.

4. The Diffusion Gate

Soft adaptation is cheap, proven and largely absent. What made mobile money spread, and why early warning has not spread the same way.

The second gate is the one adaptation policy discusses least and needs most. Consider the shape of the shortfall.

The middle bar is the interesting one. Early warning is the highest-return intervention in the adaptation repertoire — the WMO and the Global Center on Adaptation put its benefit-cost ratio at roughly nine to one — and it is cheap. Half the world's countries still lack an adequate system. That is not a financing outcome. Nine-to-one returns on low-cost assets do not go unbought for want of capital; they go unbought for want of a delivery path.

4.1 What diffusion looks like when it works

The comparison that should discipline adaptation policy is mobile money. It moved from a Kenyan pilot to a mainstream financial system across much of Africa and South Asia in under two decades; by 2025, roughly one in ten adults in developing economies was using a mobile-money account to save, and the World Bank attributed a measurable surge in formal saving to phone-based accounts. India's public digital stack went further: the Unified Payments Interface was processing on the order of 19.5 billion transactions a month by mid-2025, which the IMF assesses as the world's largest real-time retail payments system by volume (Observer Research Foundation).

Nothing about those systems was more technically demanding than delivering a flood warning. What they had was a diffusion architecture: a device already in the user's hand, an identity layer, a settlement rail, a dense network of local agents, a regulatory posture that permitted the service, and a commercial model that paid someone to reach the last customer. Adaptation technology has the device and, increasingly, the identity and payment layers — the same rails now underpin direct benefit transfers used for drought relief — but it has rarely had the agent network, the regulatory clarity or the party paid to reach the last farmer.

4.2 The adaptation services waiting on rails

Three classes of soft adaptation are demonstrably effective and demonstrably under-diffused.

• Digital climate advisory. Forecast and agronomic guidance delivered to smallholders has been associated with yield gains of 10 to 30 per cent alongside reduced drought losses, and the Global Center on Adaptation has published a practical roadmap for scaling it across Africa. Randomised evaluations of digital agricultural extension in India have found meaningful gains in yields and input efficiency at very low cost per farmer.
• Parametric insurance and anticipatory payouts. Where an identity layer and a payment rail already exist, a rainfall index can trigger a transfer to a household before the harvest fails rather than after — converting a slow relief process into an automatic one. The constraint is almost never the actuarial design; it is enrolment, trust and the rail.
• Pay-as-you-go asset finance. Distributed solar reached millions of off-grid households through instalments collected over mobile money, a model IRENA documented as a distribution innovation rather than a technological one. The same mechanism is directly transferable to irrigation pumps, cold storage and efficient cooling — resilience assets that are individually small and collectively decisive.

The pattern across all three is that the binding scarcity is institutional and commercial, not financial or technical. This is why the diffusion gate cannot be opened with a larger adaptation fund. It is opened by treating digital public infrastructure — identity, payments, connectivity, and the data and forecasting services layered on them — as adaptation infrastructure, and by paying for last-mile delivery as an explicit line item rather than assuming it.

5. The Returns Are Not in Doubt

On the evidence now available, adaptation is among the highest-return uses of public capital in development — which is what makes the gap a delivery failure rather than an allocation one.

The economic case has strengthened considerably as the evidence base has moved from modelled estimates to observed investments. The Global Commission on Adaptation established the headline in 2019: US$1.8 trillion invested across five areas — early-warning systems, climate-resilient infrastructure, improved dryland agriculture, mangrove protection and water management — could generate US$7.1 trillion in total net benefits. More recent bottom-up work has found higher returns still.

These numbers should be read with the caution their differing methods require — a benefit-cost ratio computed on 320 implemented projects is not the same object as a global portfolio estimate — but the direction is unambiguous and the range is narrow enough to be decisive. There is no plausible reading of this evidence on which adaptation is a marginal investment.

That conclusion is what forces the report's framing. When an asset class returning four to ten dollars per dollar is funded at eight per cent of assessed need, the explanation cannot be that decision-makers have weighed the returns and declined. It must be that something between the return and the disbursement is broken — which is precisely what the two gates describe.

6. Evidence from Delivery

Four programmes that opened both gates, and what each shows about the mechanism.

The strongest argument for the two-gate framing is that where both gates have in fact been opened, the results are not incremental. They are step changes in mortality and loss.

6.1 Bangladesh: cyclone mortality

Bangladesh built a coastal protection system over five decades that combined hard and soft adaptation deliberately: embankments and a network of raised concrete shelters, paired with a forecasting service and a corps of tens of thousands of trained community volunteers who carry warnings the last mile by bicycle and megaphone. The result is the most dramatic adaptation outcome on record.

Neither half would have worked alone. Shelters without warnings are empty buildings; warnings without shelters tell people to run nowhere. The programme is the clearest available demonstration that the two gates are complements rather than alternatives.

6.2 Ahmedabad: heat, at almost no capital cost

Ahmedabad's 2013 Heat Action Plan, the first in South Asia, combines an early-warning trigger, public cooling infrastructure and reflective cool roofs on public buildings and low-income housing. It has been credited with avoiding roughly 1,190 heat deaths a year, and the model has since been adopted — with mixed implementation — by more than a dozen Indian states. The capital requirement is negligible; the delivery requirement is a functioning municipal institution able to trigger a protocol and reach vulnerable households. It is a pure delivery-gate intervention, and its replication record shows both the promise and the limit: the plan diffuses easily, the institutional capacity to run it does not.

6.3 India: adaptation on public digital rails

India's Aadhaar identity layer and UPI payment system were not built for climate adaptation, but they have become adaptation infrastructure. The same rails that carry direct benefit transfers have been used for drought relief and pandemic support and are being extended to parametric crop insurance and heat-related wage compensation. This is the diffusion gate opening as a by-product: once a country has identity, payments and connectivity at population scale, the marginal cost of adding an adaptation service to them collapses.

6.4 Kenya: the distribution lesson

Kenya's contribution is the distribution model itself. M-PESA made the country the global reference for mobile money, and the same rails subsequently financed pay-as-you-go solar for off-grid households through small mobile instalments. The lesson generalises past energy: the constraint on reaching poor, remote, cash-flow-constrained households is rarely the product and almost always the collection mechanism. Adaptation services that piggyback on an existing collection mechanism scale; those that require a new one usually stall at pilot.

7. An Agenda for Both Gates

Sequenced by which gate each instrument opens, because an instrument aimed at the wrong gate buys plans or pilots rather than protection.

Table 2 — Instruments by the gate they open

7.1 Open the capital gate

• Scale concessional finance, guarantees and currency-hedging instruments explicitly against the cost-of-capital wedge, and report progress in basis points achieved rather than dollars pledged.
• Complete multilateral development bank reform to expand lending headroom, and capitalise the Resilience and Sustainability Trust and the Loss and Damage Fund at a scale proportionate to a US$310–365 billion need.
• Resolve the debt overhang documented by UNCTAD, and make climate-resilient debt clauses standard rather than exceptional.
• Mandate resilience at the design stage for all new infrastructure, capturing it at the 3–5 per cent premium Lifelines identifies while the emerging world's build-out is still ahead of it.

7.2 Open the delivery gate

• Classify digital public infrastructure — identity, payments, connectivity, forecasting — as adaptation infrastructure, and fund it from adaptation budgets where it carries adaptation services.
• Fund last-mile delivery as a named line item: agent networks, extension workers, community volunteers, enrolment campaigns. Bangladesh's volunteer corps is as much a piece of adaptation infrastructure as its embankments.
• Finish Early Warnings for All. A nine-to-one return at low unit cost, half-covered, is the largest unclaimed risk reduction available.
• Invest in the municipal and agency capacity that operates soft adaptation, on the evidence that plans replicate far more readily than the institutions needed to run them.

7.3 Fix the appraisal that governs both

• Require non-contingent benefits — yields, service reliability, insurance costs, avoided business interruption — to be counted in adaptation appraisal, on the WRI finding that these constitute more than half of the return.
• Report adaptation delivery in installed terms — households covered, kilometres protected, warnings received — rather than in commitments, so the measure matches the thing that reduces risk.
• Sequence by gate: for hard adaptation, finance first; for soft adaptation, rails and delivery first. Programmes that ignore which gate binds produce plans and pilots in roughly equal measure.

7.4 Conclusion

The resilience gap is the most solvable problem in climate policy and among the least solved. The technologies are proven, the returns are exceptional and non-contingent, and the political intent is now nearly universal — 172 of 197 countries hold a plan. What is missing is not knowledge, will or evidence but the plumbing between them: capital priced for the places that need protection most, and delivery channels that carry cheap protection the last mile.

Framing that as a single financing shortfall has been costly, because it directs a money instrument at a problem half of which money does not solve. The corrective is not a larger number but a sharper diagnosis: ask, for each intervention, which gate is closed. Where the answer is capital, the work is concessional finance, tenor, fiscal space and design standards. Where the answer is delivery, the work is rails, agents, institutions and enrolment — and the returns available there are the highest and the least claimed.

Bangladesh's cyclone mortality fell four orders of magnitude in fifty years, on a combination of concrete and volunteers that no one would describe as technologically advanced. It is the strongest evidence in this report, and its lesson is narrow and encouraging: when both gates open, the gap closes fast.

References

Every quantitative claim above is attributed inline. The principal sources are collected here.

• UNEPAdaptation Gap Report 2025 — assessed needs of US$310–365bn a year by 2035; flows of US$26bn in 2023; 172 of 197 countries with a national adaptation plan. See also the accompanying release.
• World Resources InstituteClimate Adaptation Investments Yield Massive Returns (June 2025) — 320 investments across twelve countries; benefits above $10 per $1; IRR of 27 per cent; more than half of benefits non-contingent. Full study.
• IFCLow Cost, High Yield: The Adaptation and Resilience Investment Opportunity (2026) — US$390bn a year of hazard losses in low- and middle-income countries; $8.60 of protected asset value per dollar; 43 million jobs at risk across 49 countries by 2050.
• Global Commission on AdaptationAdapt Now: A Global Call for Leadership on Climate Resilience (2019) — US$1.8tn of investment for US$7.1tn in net benefits. Also via WRI.
• World Bank / GFDRRLifelines: The Resilient Infrastructure Opportunity — a 3–5 per cent design-stage premium against avoided losses of 50–100 per cent of asset value.
• WMOEarly Warnings for All — universal coverage targeted for 2027; roughly half met. See also Early warning systems reach new heights, critical gaps jeopardize global progress.
• WMOThe Triple Dividends of Early Warning Systems and Climate Services — a benefit-cost ratio of roughly 9:1.
• IEAWorld Energy Investment 2025 — the 400–800 basis-point cost-of-capital wedge.
• UNCTADA World of Debt — US$11.4tn developing-country external debt; US$921bn of net interest payments in 2024.
• UNFCCCBaku to Belém Roadmap to 1.3T — the COP30 finance framework and the commitment to at least triple adaptation finance. See WRI, COP30: Outcomes and Next Steps.
• IMFResilience and Sustainability Trust — long-tenor concessional financing for resilience.
• UNFCCCFund for responding to Loss and Damage.
• United NationsThe Sevilla Commitment — outcome of the fourth Financing for Development conference, 2025.
• World Economic Forum / Swiss ReThis Is How Climate Change Could Impact the Global Economy — the Swiss Re Institute climate stress test.
• World BankMobile Phone Technology Powers Saving Surge in Developing Economies (2025) — Global Findex evidence on mobile-money saving.
• World BankDigital Public Infrastructure — identity, payment and data layers as development infrastructure.
• Observer Research FoundationDigital Public Infrastructure and the Future of Digital Payments: Lessons from Pix and UPI — UPI transaction volumes.
• Global Center on AdaptationAdaptation Insights: A Practical Roadmap for Scaling Digital Climate Advisory Services in Africa.
• VoxDevHow Digital Extension Helps Indian Farmers Grow More and Lose Less.
• IRENAPay-as-you-go Models: Innovation Landscape Brief.
• The ConversationBangladesh Has Saved Thousands of Lives from a Devastating Cyclone — Here's How.
• Down To EarthAhmedabad Uses Cool Roofs, Misting Bus Stops and Heat Insurance for Climate Resilience.
• World BankRising to the Challenge: Climate Adaptation and Resilience; and Adaptive Social Protection: Building Resilience to Shocks.