Do Welfare States Adapt?

• In Event: Flood Risk, Insurance, and Housing Finance

Abstract

Introduction/Background
Climate change is introducing eco-social risks—such as extreme heat, flooding, and vector-borne disease—that threaten core welfare domains like income maintenance, housing stability, and health security. While scholarship argues the welfare state must evolve into an ecowelfare state to manage these risks, we know little about whether established welfare states are actually integrating them into their strategic planning. Drawing on discursive institutionalism, this study treats countries' National Adaptation Plans (NAPs) as instruments of pre-distribution: upstream institutional signals that determine, before crises strike, whether eco-social risks are met with collective social guarantees or left to private responsibility.

Purpose/Research Question
The study asks: (1) To what extent are climate hazards explicitly linked to welfare-state instruments versus technical infrastructure or market-based mechanisms? (2) Do these response patterns vary across European and U.S. welfare regimes, reflecting traditional path-dependent logics?

Data/Methods
The study analyzes the NAPs of 30 countries (EU-27, Norway, UK, and US), yielding a corpus of roughly 24,000 vulnerability and strategy sentences. It introduces a consensus-based Natural Language Processing (NLP) pipeline triangulating three classifiers—a lexical TF-IDF model, a fine-tuned ClimateBERT transformer, and dense sentence embeddings—to measure the coupling between seven climate hazards and three instrument families (welfare, technical, financial). Findings are validated through zero-shot LLaMA-3 sensitivity analysis and a structured qualitative audit of eight purposively selected cases. Multi-method clustering (HAC, NMF) identifies cross-national policy profiles.

Results/Findings
Three findings emerge. First, adaptation planning discourse converges across welfare regime types: 24 of 30 countries fall into a single dominant instrument cluster, and the traditional Three Worlds typology is not reproduced. Slovakia is the lone consistent outlier; the UK and US form a small sub-grouping driven by coverage depth rather than a residual instrument logic. Second, this shared portfolio is structurally tilted: welfare instruments appear in 29 of 30 countries but are paired less frequently and less densely with core hazards (flood, heat, drought, storm) than market and technical instruments—subordination, not absence. Third, the most significant cross-national cleavage is hazard selectivity: which secondary risks (wildfire, cold extremes, land instability) are recognized as public problems, rather than how they respond to core risks. Hazard recognition operates as an upstream act of political boundary-drawing.

Conclusion/Implications
These findings suggest that the ecowelfare state has not yet been institutionalized as a primary mode of climate adaptation. By routing hazards through technical and financial instruments rather than welfare-based ones, NAPs perform an ex-ante privatization of climate risk that operates upstream of any explicit retrenchment decision.The U.S. case sharpens this implication: the U.S. plan couples hazards to a broad instrument portfolio comparable to European peers but was developed under the Biden Administration and may not survive the current political environment, a test of whether the materiality of climate can sustain adaptation commitments independent of political regime. For policymakers, these results highlight the need to explicitly integrate social policy institutions into climate adaptation frameworks to achieve equitable resilience.

Author

• Christopher Taylor Brown, University of California, Berkeley
  Presenting Author