Disasters linked to natural hazards, including geophysical or weather-driven events such as floods, droughts or wildfires, are occurring with increasing frequency. Beyond their devastating human toll, they also disrupt economies and impose significant costs on public finances, from emergency spending and infrastructure repair to lost tax revenues. As extreme weather events become more frequent and severe, these fiscal pressures are set to grow.
Yet most governments lack the tools to track and report these costs systematically. Most governments capture only the most visible spending, immediate relief and reconstruction, while broader fiscal impacts such as lost revenues, increased welfare transfers, and disruptions to public services remain largely invisible.
This is because the institutional frameworks that governments have put in place to track, consolidate and report disaster-related spending remain underdeveloped. Formal definitions of what counts as a fiscal cost are often absent or incomplete. Co-ordination among ministries, agencies and levels of government is weak. And the systems used to capture spending are fragmented, sometimes relying on manual data collection, rather than integrated, government-wide tracking.
This data gap has real consequences for how governments respond and plan. By routinely underestimating the true cost of disasters, governments cannot provision appropriately for future events. The data gap also distorts the picture on infrastructure spending. Without reliable data on disaster costs, governments cannot make the case for investing in resilience, nor measure whether that investment is paying off.
Five actions can close the data gap: appoint a central body to co-ordinate and consolidate disaster cost data; establish clear definitions of what counts as a fiscal cost; embed tracking in existing budget management systems; report disaster costs consistently in budgets and financial statements; and use data to inform resilience investment.
How governments account for disasters
Natural hazards are among the most disruptive forces that governments face. Floods inundate homes and destroy infrastructure. Droughts devastate agriculture and strain water systems. Wildfires consume forests, homes and public assets. Earthquakes and cyclones can set back decades of public investment in a matter of hours. While these events have always posed challenges for societies, the evidence is clear that their frequency and severity are increasing. Over the past four decades, the number of recorded disaster events linked to natural hazards nearly doubled (UNDRR, 2020[1]).
The economic consequences have been widely studied. Disasters destroy private and public capital, disrupt supply chains, reduce productivity and force households and businesses to divert resources from growth to recovery. Global economic losses from disasters linked to natural hazards have been estimated at USD 280 billion on average over the last ten years (Haegeli et al., 2026[2]). But aggregate economic losses, while striking, tell only part of the story. The specific impact of disasters linked to natural hazards on government finances is less well understood and far less systematically measured.
The fiscal costs of disasters go well beyond emergency spending
When a disaster strikes, governments are expected to act.1 They deploy emergency services, fund relief operations, repair damaged public infrastructure, compensate affected households and businesses, and support subnational governments overwhelmed by the scale of the response. These immediate, visible expenditures are what most people associate with the fiscal cost of a disaster, but they represent only the tip of the iceberg.
Disasters also reduce government revenues. Economic disruption leads to lower corporate and personal income tax receipts. Businesses close temporarily or permanently. Tax deferrals are granted to affected communities. On the spending side, automatic stabilisers kick in. For instance, unemployment benefits rise and social transfers increase, adding further pressure to public budgets that were not designed with disaster costs in mind.
Beyond the immediate event, governments often face longer-term fiscal consequences. Public assets such as roads, bridges, schools, and hospitals may be damaged or destroyed, generating write-downs in government balance sheets and triggering multi-year reconstruction programmes. Contingent liabilities may crystallise, particularly where governments act as insurers of last resort for households and firms that lack adequate private coverage (OECD/The World Bank, 2019[3]). Finally, the operational disruptions suffered by public agencies and state-owned enterprises can impose costs that ripple through the public sector for months or years.
Most governments see only part of the picture
Despite the breadth and depth of these fiscal impacts, most governments track only a fraction of them. Survey evidence collected through the 2025 OECD Survey on Accounting and Reporting for the Fiscal Costs of Natural Disasters (hereinafter referred to as the “OECD Survey”) from 29 survey respondents, including 26 OECD Members, shows that the fiscal costs most consistently captured are those most closely tied to formal funding and compensation mechanisms: the reconstruction of public infrastructure, emergency relief operations, compensation to citizens or transfers to sub-national governments (see Figure 1). These are visible, administratively manageable, and often channelled through established budget programmes or dedicated funds.
Beyond this core, fiscal visibility drops sharply. Support provided through public agencies or state-owned enterprises are reported less systematically even though they can represent a substantial share of total disaster costs following major events. Indirect fiscal impacts, such as revenue shortfalls or the additional spending triggered by automatic stabilisers, are tracked by only a very small number of countries.
The result is a systematic and significant undercount of the true fiscal cost of disasters linked to natural hazards across OECD and non-Member governments. This is not primarily a technical failure. It reflects the absence of common definitions, weak co-ordination across institutions, and budget and accounting systems that were not designed to tag and consolidate disaster-related spending across the whole of government.
Types of expenses accounted for as direct fiscal costs of disasters during 2020-2024, percentage of OECD Survey respondents
Source: 2025 OECD Survey on Accounting and Reporting for the Fiscal Costs of Natural Disasters.
Government financial systems were not built with disasters in mind
A large part of the problem lies in how government financial systems are structured. In most countries, public spending is recorded by its economic nature (salaries, grants, capital expenditure) or by the ministry or programme through which it flows. This makes it very difficult to answer a simple but essential question: how much did this disaster cost the government, in total, across all agencies and all types of spending?
When a flood hits, the transport ministry records road repairs, the housing ministry records emergency shelter costs, the social affairs ministry records increased benefit payments, and the finance ministry manages transfers to regional governments. Each of these entries is captured somewhere in the system. But because they are recorded under different budget lines, in different ministries, using different classifications, nobody adds them up. The total fiscal cost of the flood – the number that would actually be useful for planning and accountability – is simply not calculated or published.
There is a wider set of challenges that are linked to accounting for disaster impacts. Firstly, the lack of consistency in defining what constitutes a specific disaster. Secondly, even for single-hazard events, it can be difficult to establish clear temporal and spatial boundaries, i.e. when a disaster starts and ends (particularly in cases with slow-onset hazards or prolonged recovery phases).2
In the private sector, it is standard practice for companies to track costs by project, event or client, allowing managers to see, at a glance, the total cost associated with a particular activity. Most government financial systems do not work this way. Spending is recorded, but the link to a specific disaster event is not. A few countries are beginning to introduce tagging, which involves flagging each disaster-related expense so that these can be aggregated. Poland has taken a practical step in this direction by introducing tagging of its state budget to track all disaster-related expenditure.
The alternative is to collect data manually, which requires asking each ministry and agency, after the event, how much they spent on disaster response and recovery. This is time-consuming, incomplete, and heavily dependent on the willingness of institutions to co-operate and report consistently. It is also difficult to sustain as a systematic practice: it tends to happen for the largest and most politically visible disasters, and not at all for smaller but recurrent events whose cumulative fiscal cost may be equally significant.
More fundamentally, governments that have invested in accrual-based accounting systems are better placed to capture the full range of disaster-related fiscal impacts, including asset write-downs, provisions for future liabilities, and contingent commitments, that cash-based systems miss entirely. Accrual accounting, when well designed, can reveal fiscal costs that would otherwise remain invisible until they crystallise into actual expenditure (Moretti and Youngberry, 2018[4]). Yet even in countries with mature accrual frameworks, the link between accounting entries and disaster events is rarely made explicit, limiting the usefulness of the data for fiscal risk management (Moretti, Boucher and Giannini, 2021[5]).
Formal frameworks are the exception, not the rule
Compounding these challenges is the absence, in most countries, of agreed rules about what should be counted as a fiscal cost of a specific disaster in the first place. Without a shared definition, different ministries will inevitably draw the boundary differently. One agency may count only direct emergency expenditure; another may include longer-term reconstruction; a third may exclude costs borne by public enterprises. The result is data that cannot be meaningfully compared or aggregated.
The OECD Survey reveals that comprehensive formal guidance on how to define and record disaster fiscal costs remains rare. A significant minority of countries (ten out of twenty-nine surveyed) operate without any such guidance at all. Where guidance exists, it tends to focus on determining whether particular claimants or programmes are eligible for disaster funding. This eligibility-based guidance plays an important operational role: it helps governments respond quickly and equitably in the aftermath of an event. But it was not designed to produce a coherent picture of the government's overall fiscal exposure, and it leaves significant gaps in the data that ministries of finance need for fiscal planning and risk management.
Spain illustrates both the challenge and the potential for innovation. Despite frequent and costly floods, wildfires and storms, no single institution consolidates the full fiscal picture. In response, Spain's independent fiscal institution AIReF has used artificial intelligence to extract structured data on disaster-related fiscal costs from published government decisions including weekly Council of Ministers summaries, producing a comprehensive estimate of disaster costs and other exceptional events.
Even less information on the fiscal costs of disasters is made public
Even where governments do collect data on disaster costs internally, what reaches the public domain is typically even more limited. Around three-quarters of OECD Survey respondents publish some information on the fiscal costs of disasters (see Figure 2). Australia offers one of the more complete examples of systematic disclosure among OECD countries. The federal budget explicitly identifies major disaster events such as cyclones or floods and quantifies their fiscal impacts within multi-year forward estimates, providing a picture that extends beyond immediate relief spending. Spain provides another example of more systematic disclosure: the Spanish central government's General State Account now includes an environmental sustainability supplement that consolidates the budgetary impact of disasters across the whole of government.
However, these disclosures are often partial, inconsistent and difficult to use. In many cases, information is published only for the largest and most politically salient disasters, not as a matter of routine. It is frequently scattered across multiple documents rather than consolidated in a single, accessible place. And it tends to cover only the most immediate and direct costs, omitting the broader fiscal picture that would be needed to assess the government's true exposure.
This matters because transparency is not just a good governance principle: it is a practical requirement for effective disaster risk management. When parliaments cannot see the full cost of disasters, they cannot scrutinise whether public money is being spent wisely on response and recovery. When citizens cannot access clear information, they cannot hold governments to account for how disaster risk is being managed. And when financial markets cannot assess a government's total disaster-related fiscal exposure, they must price in uncertainty, which can translate into higher borrowing costs, particularly for countries already facing fiscal pressures.
The disclosure gap isthus a problem with direct consequences for trust, accountability and fiscal credibility.
The gaps described above in systems, definitions, and disclosure have direct and tangible consequences for how governments plan their finances, manage risk, make investment decisions, and maintain the trust of citizens and financial markets. As extreme weather events become more frequent and severe, the cost of this data gap will only grow.
Governments cannot plan or provision for what they cannot measure
Sound fiscal management depends on reliable information. When governments systematically undercount the fiscal costs of disasters, the consequences cascade through the entire budget cycle.
In the short term, the absence of reliable cost data makes it harder to anticipate the immediate budgetary impact of a disaster. Rather than being prepared, governments are forced to react through supplementary budgets and emergency reallocations of resources. This reactive posture is not only inefficient; it can also crowd out other planned expenditures, distort budget execution, and undermine fiscal credibility.
Over the medium term, the absence of consistent historical data makes it harder to calibrate the budgetary instruments that governments use to manage disaster risk (OECD/The World Bank, 2019[3]). Contingency reserves, disaster funds and budget provisions all need to be sized based on past experience. Without reliable data on the frequency, severity and fiscal impact of past events, these instruments risk being either underfunded, leaving governments dangerously exposed when the next disaster strikes, or overfunded, tying up public resources that could be deployed elsewhere (OECD, 2022[6]).
This challenge is particularly acute for governments that have taken on implicit insurance obligations towards households and firms – obligations that are rarely costed, capped or disclosed, yet can dwarf the direct costs of disaster response when a major event strikes (Radu, 2022[7]).
Over the long term, the absence of systematic disaster cost data weakens governments' ability to assess their own fiscal sustainability. Disasters are increasingly a structural feature of public finances, not exceptional events. Yet, in most countries, they remain absent from long-term fiscal projections, debt sustainability assessments and macro-fiscal scenario modelling (Salmon-Genel, 2025[8]). The OECD tool for calculating the fiscal impacts of decarbonisation and environmental challenges shows what is possible when governments invest in better fiscal data: granular, forward-looking analysis of fiscal pressures that can directly inform budget planning, debt management and long-term fiscal strategy.
The data gap distorts investment decisions
Perhaps the most significant consequence of the data gap concerns public investment in disaster prevention and resilience. Spending on flood defences, earthquake-resistant infrastructure, early warning systems and other preventive measures can appear, in conventional budget analysis, as a pure cost. It increases expenditure today without an obvious and immediate return.
But this framing is incomplete and potentially misleading. The true return on resilience investment is the fiscal damage that is avoided in the future, such as the emergency spending that does not happen, the infrastructure that does not need to be rebuilt, the revenue losses that do not materialise. When governments lack reliable data on what disasters cost, they cannot quantify this avoided cost. They cannot demonstrate to finance ministries, legislators or the public that investing in resilience is fiscally rational. And they cannot measure, after the fact, whether their investments in prevention are paying off.
The result is a systematic bias against resilience spending in public investment decisions. Governments end up spending more on disaster response, because they have no choice when disasters strike, and less on prevention, because they cannot make a credible fiscal case for it (OECD/The World Bank, 2019[3]).3 This is a self-reinforcing trap: the weaker the data, the harder it is to justify prevention spending; the less that is spent on prevention, the higher the eventual fiscal cost of disasters. Closing the data gap is therefore not just a matter of good bookkeeping. It is a prerequisite for smarter, more cost-effective public investment and for breaking this cycle.
Transparency, accountability and market confidence
Beyond planning and investment, the data gap has important implications for public trust and accountability. Disasters linked to natural hazards represent one of the largest and fastest-growing sources of unplanned public expenditure in many OECD countries (OECD/The World Bank, 2019[3]). Parliaments and citizens have a legitimate interest in knowing how much disasters cost, how those costs are being financed, and what governments are doing to reduce their exposure over time.
While most OECD Survey respondents have some form of disclosure, in many cases these disclosures are fragmented, spread across agency reports, supplementary budgets and ad hoc announcements. This makes it difficult for oversight bodies to form a coherent view of the government's total disaster-related spending or to assess whether public resources are being deployed effectively.
Financial markets are also paying closer attention. Credit rating agencies are increasingly incorporating natural hazard exposures and extreme weather risk into their sovereign assessments (Cappiello et al., 2025[9]). Governments with high hazard exposure, high public debt, or constrained access to capital markets face particular scrutiny.
Comprehensive, credible and consistent disclosure of disaster fiscal costs, and of the frameworks governments have put in place to manage them, can help maintain citizen and investor confidence and, over time, improve borrowing conditions.
Types of disclosures used for fiscal costs of disasters, percentage of OECD Survey respondents
Note: Figure shows data from 29 OECD Survey respondents, including 26 OECD countries.
Source: 2025 OECD Survey on Accounting and Reporting for the Fiscal Costs of Natural Disasters.
Designate a central body, typically the ministry of finance or budget authority, to take ownership of disaster fiscal data: setting definitions, requesting data from across government, and producing a consolidated whole-of-government picture after each significant event. Central co-ordination is the most effective single step governments can take to improve data, in co-operation with other relevant stakeholders including the national disaster risk management authorities and national statistical office.
Establish formal definitions of what counts as a disaster fiscal cost, covering the full range of impacts: emergency spending, reconstruction, revenue losses, transfers to subnational governments, and longer-term liabilities. Definitions should be binding, developed centrally, and agreed with line ministries to ensure they are applied consistently across government.
Introduce disaster cost tracking in existing financial management systems, allowing individual spending transactions to be flagged as disaster-related and aggregated across ministries in real time. This is a modest technical change that can deliver significant improvements in fiscal visibility without requiring new systems or restructured budget classifications.
Integrate disclosures on disaster fiscal costs into regular budget documents or financial statements, ideally reported at whole-of-government level and presented in a way that allows year-on-year comparisons. Over time, develop this information with historical data, forward-looking scenarios and an explanation of the instruments in place to manage these fiscal risks.
Use data to inform resilience investment. Reliable disaster cost data allows governments to quantify the fiscal damage that prevention spending avoids, making the case to finance ministries and legislators that investing in resilience is not a cost but a fiscally rational strategy for reducing future spending pressures.
OECD 2025 Survey on accounting and reporting for fiscal costs of natural disasters in https://data-explorer.oecd.org/s/5fi
OECD tool for calculating the fiscal impacts of decarbonisation and environmental challenges: technical document
References
[9] Cappiello, L. et al. (2025), From words to deeds – incorporating climate risks into sovereign credit ratings, https://www.ecb.europa.eu/press/research-publications/resbull/2025/html/ecb.rb250730~ebfb33d43c.en.html.
[2] Haegeli, J. et al. (2026), Nat Cat sigma 1/2026: Natural catastrophes in 2025: the persistent rise of wildfire and storm risk, Swiss Re Institute, https://www.swissre.com/institute/research/sigma-research/sigma-2026-01-natcat-2025-wildfire-storm-risk.html.
[5] Moretti, D., D. Boucher and F. Giannini (2021), “Managing fiscal risks: Lessons from case studies of selected OECD countries”, OECD Journal on Budgeting, Vol. 2021/1, https://doi.org/10.1787/7db1d712-en.
[4] Moretti, D. and T. Youngberry (2018), “Getting added value out of accruals reforms”, OECD Journal on Budgeting, Vol. 2018/1, https://doi.org/10.1787/budget-18-5j8l804hpvmt (accessed on 23 July 2025).
[10] OECD (2023), Report on the implementation of the Recommendation on the Governance of Critical Risks.
[6] OECD (2022), Building Financial Resilience to Climate Impacts: A Framework for Governments to Manage the Risks of Losses and Damages, OECD Publishing, Paris, https://doi.org/10.1787/9e2e1412-en.
[3] OECD/The World Bank (2019), Fiscal Resilience to Natural Disasters: Lessons from Country Experiences, OECD Publishing, Paris, https://doi.org/10.1787/27a4198a-en.
[7] Radu, D. (2022), “Disaster Risk Financing: Limiting the Fiscal Cost of Climate-Related Disaster”, Discussion Paper 174, https://doi.org/10.2765/270746.
[8] Salmon-Genel, M. (2025), “Assessing the Macro-Fiscal Risks from Climate Change: Concepts, Methodological Approaches, and Insights from Country Practices”, European Economy Discussion Papers, European Commission, https://doi.org/10.2765/8918557.
[1] UNDRR (2020), The human cost of disasters: an overview of the last 20 years (2000-2019), https://www.undrr.org/publication/documents-and-publications/human-cost-disasters-overview-last-20-years-2000-2019.
Data presented in this policy brief are based on responses to the 2025 OECD Survey on Accounting and Reporting for the Fiscal Costs of Natural Disasters (hereafter referred to as the “OECD Survey”). The OECD Survey was distributed to all 38 OECD Member countries, the eight OECD accession candidate countries and selected economies. Responses were received from 26 OECD Member countries, two OECD accession countries (Brazil and Bulgaria) and one non-Member economy (Chinese Taipei). The respondent OECD Member countries were: Australia, Austria, Belgium, Chile, Costa Rica, Czechia, Denmark, Estonia, Germany, Greece, Hungary, Ireland, Japan, Korea, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Norway, New Zealand, Poland, Portugal, Sweden, Switzerland and the United States.
The OECD Survey was administered between February and June 2025. The data are based on self-reporting by governments and reflect respondents’ own assessments of current practices and procedures. As part of the data validation process, the OECD Secretariat conducted consistency checks of the responses received.
Contact
Delphine Moretti (✉ delphine.moretti@oecd.org)
Thibaut Rosenzweig and Giorgia Ponti have contributed to the development of this Policy Brief.
← 1. This policy brief focuses specifically on spending incurred during and after disaster events for response, relief, and recovery. Government spending on disasters can also cover disaster risk management and risk reduction.
← 2. Note: Figure shows data from 29 OECD Survey respondents, including 26 OECD countries. Four OECD Survey respondents indicated not applicable.
← 3. The draft Global Disaster-Related Statistics Framework (G-DRSF), submitted to the UN Statistical Commission in 2026, directly flags this gap: the absence of commonly applied event typologies and unique disaster identifiers limits comparability across datasets and hinders the aggregation of impacts over time and across sectors.
← 4. Even where risk information exists, it is always not translated into investment decisions (OECD, 2023[10]). The gap between knowing the risk and acting on it in budget and capability planning is therefore as important as the data gap itself.