Welfare Resilience Assessment Framework
A Framework for Welfare Resilience in the Age of AI
Using this content? Please see our Content Permissions policy.
This paper introduces the Welfare Resilience Assessment Framework, a diagnostic framework for welfare policymakers. It serves as a structured method to assess how prepared a country's welfare system is for AI-driven economic disruption. We set out four dimensions of welfare-relevant risk, the policy levers available to address risks from each dimension, and the enabling conditions that determine whether those levers function in practice. Our aim is to help governments identify where their systems are prepared and where the gaps lie, and we illustrate the framework via country case studies.
Published August 18th 2026
A substantial body of research now estimates which occupations and tasks are most exposed to AI, including work by the IMF, the ILO, the World Bank and a range of academic groups. However, exposure and preparedness are separate questions. A country with high measured exposure may be institutionally well equipped to respond, while a country with modest exposure may be poorly prepared because its social protection system lacks coverage, fiscal headroom, or delivery reach.
Welfare systems have generally been designed around cyclical fluctuations, and AI-driven change may place pressure of a different kind: on the reallocation of workers between occupations, on the fiscal base that funds social provision, and on the administrative systems through which benefits reach households. The timing, pace, and scale of that pressure remain uncertain. A welfare system's readiness to manage those pressures is still a different question from whether it functions well in normal times.
The Welfare Resilience Assessment Framework is a diagnostic tool to answer the question: how prepared is a national welfare system for the economic disruption AI is likely to produce?
The framework: risks, policy levers, enablers
Risk
Policy Lever
Enabler
Acute shocks to household income and access to essential services
Stabilise incomes and employment during short-term shocks
Eg. Unemployment insurance, Short-time work
Data, institutional, and fiscal infrastructure essential for functioning of the policy levers
Eg. Real-time earnings data, direct payment infrastructure
Pressures on workers and households as occupational structures and sectoral composition shift.
Active labour-market and reallocation-support levers
Eg. Public employment services or wage insurance
Data, institutional, and fiscal infrastructure essential for functioning of the policy levers
Eg. Vacancy and skills data, accredited training providers
Long-run shifts in the labour share of income, employment geography, and welfare provision fiscal base.
Redistributive & place-based interventions
Eg. Universal basic services and social wealth fund mechanisms
Data, institutional, and fiscal infrastructure essential for functioning of the policy levers
Eg. Revenue-raising capacity, independent fiscal institutions
Implementation constraints that determine whether welfare policy reaches eligible populations.
Improve coverage, inclusion, and execution
Eg. Administrative reform or algorithmic oversight
Data, institutional, and fiscal infrastructure essential for functioning of the policy levers
Eg. Interoperable registries, digital ID, legal basis for portability
The framework diagnoses preparedness through a three-step chain:
Identifying the risk comes first. Not all AI-related economic pressures test the welfare system in the same way. A sudden wave of layoffs in a formal sector tests whether income support can reach workers quickly. Where gradual occupational restructuring is the issue, the question shifts to whether retraining systems exist and whether they work for the workers most affected. Long-run shifts in the labour share of income raise a different concern: whether the redistributive architecture remains fit for a changed economy.
Mapping the policy levers is the second step. Each risk dimension has a range of instruments that can address it, but the right instrument depends on the risk's character and the country's institutional context. The framework asks which policy levers exist for each risk and what tradeoffs they carry. An unemployment insurance system that covers formal workers well may provide no protection at all in an informality-heavy labour market. A retraining programme with strong enrolment numbers may not reach the workers most at risk of displacement.
Assessing the enablers is the third step. Enablers are the operational conditions that determine whether a policy instrument can actually function. Examples include data systems that allow earnings to be observed and benefits to be targeted; fiscal headroom to sustain commitments during a prolonged shock; legal authority to administer new instruments or coordinate across agencies; and delivery infrastructure that can reach intended recipients. A country may have the right policy on paper and lack one of these conditions, making the policies ineffective.
Risks
What pressures does the welfare system face?
Policy levers
What instruments can respond to these pressures?
Enablers
Can those instruments actually function effectively?
The chain in practice
Consider a country with a growing share of gig/platform workers (ride-hailing drivers, delivery workers, online freelancers) who are outside formal employment categories.
The risk: these workers face income shocks that arrive quickly and irregularly, but sit outside the categories formal income-support systems are typically designed to reach.
The policy levers: options include extending unemployment insurance to non-standard workers, attaching portable benefits to workers rather than employers, or introducing in-work benefits indexed to fluctuating earnings.
The enabler question: each lever requires different operational conditions, such as earnings visibility for targeting, a legal framework for benefit portability, or a payment system capable of variable disbursement.
The framework’s output is a clear picture of which levers can move now and which depend on reform first.
Four dimensions of risk
The framework organises welfare-relevant risks across four dimensions. The first three are defined by time horizon. The fourth sits beneath all three, and it determines whether policy responses to any of them actually reach people.
Short-term
Disruptive
Sudden income loss, unstable hours, rapid restructuring
Medium-term
Transitional
Reallocation pressures as tasks and occupations shift
Long-term
Structural
Shifts in the labour–capital split and the fiscal base
Cross-cutting
Access-and-delivery
The implementation conditions (data systems, payment rails, registry coverage, legal frameworks) that determine whether welfare policy reaches people at all.
Operates across all three time horizons.
Disruptive risks are short-term income shocks that may arrive faster than the welfare system can respond. Examples include sudden job loss, unstable hours, and rapid firm restructuring. The diagnostic question here is whether income support reaches affected workers quickly enough to prevent hardship, and whether it reaches the workers actually affected.
Transitional risks arise over the medium term as occupational structures shift and tasks are automated or reorganised. Workers may find employment again, but not in the roles they left. The relevant question is whether retraining, employment services, and transitional income support are adequate to support that adjustment, and whether they reach workers across sectors and income levels.
Structural risks develop over the long run. As the share of income going to labour falls relative to capital, the geographic concentration of economic activity shifts, and the payroll-based revenue model that funds many welfare systems comes under pressure. These risks require redistribution and structural redesign rather than stabilisation or reallocation.
Access-and-delivery risks are a cross-cutting set of implementation constraints, not a fourth time horizon. They determine whether welfare policy reaches eligible citizens at all. Identity coverage gaps, fragmented social registries, earnings invisible to the state, and bias/opacity in algorithms driving benefit systems can stop well-designed instruments from working.
Each risk dimension corresponds to a distinct welfare-state capacity.
Time Horizon
Welfare-state capacity
What it does
Short-term
Coping capacity
Stabilises incomes and access to services during acute shocks
Medium-term
Adaptive capacity
Supports workers through reallocation as occupational structures shift
Long-term
Transformative capacity
Addresses long-run distributional shifts and fiscal base changes
Cross-cutting
Systemic capacity
Sustains the delivery infrastructure that makes welfare policy function in practice
How to read the case study
Each case study applies the framework to one country. The structure is the same across all of them, so once you have read one, the rest follow the same path. Read the Welfare Resilience Snapshot first for the diagnosis, then work through the sections for the analysis behind it.
Welfare Resilience Snapshot
A one-page summary that represents the main takeaways. It has four parts: the country's dominant risk dimension, its binding enabling constraints, its capacity across all four dimensions, and what can be done now versus what depends on enabling reform. The rest of the document provides the evidence and analysis for it.
Capacity Assessments
Each capacity assessment contains four sections, one per risk dimension, addressing the different welfare-state capacities — coping (disruptive risks), adaptive (transitional risks), transformative (structural risks), and systemic (access-and-delivery risks). Each section follows the same order: the core risks the country faces in that dimension, the policy levers available to address them, an enabler assessment of data, fiscal, institutional, and delivery conditions, and a key takeaway. Each takeaway names the most pressing risk in that dimension, the policy that is implementable now to address that risk, and the policies blocked by binding constraints.
Indicator dashboards
This section highlights the vulnerability and resilience of the welfare system via two dashboards: an exposure dashboard for structural and labour-market conditions, and a resilience dashboard for institutional and administrative systems. Each dashboard gives indicator values, the signal each one sends, and its source, followed by a short written reading. Do not read the indicator values/signals as rankings or scores; read them as cues drawing attention to the more critical focus areas. The analysis also marks where the signals are based on partial data.
Diagnosis and conclusion
This section is the synthesis containing the country's overall risk profile, its strongest and most fragile capacities, and a sequencing recommendation across near-, medium-, and longer-term priorities.
References
This section includes a full reference list. It may also include a programme-by-programme annex of the social-protection and labour-market instruments cited and a glossary of key terms.
