Environmental governance is critical for mitigating human displacement due to weather-related disasters

Disaster displacement—defined by the Nansen Initiative as situations where people are forced to leave their homes to avoid the impacts of an immediate and foreseeable natural hazard—is a major humanitarian and environmental challenge. Disasters displace roughly three times more people annually than conflicts and violence; between 2008 and 2018, around 263 million internal displacements occurred, generating severe economic and social insecurity. Weather-related events (storms, floods, wildfires) accounted for 98% of recorded disaster displacement in 2020. Many disaster-prone regions lack the resources to reduce or adapt to risks, and climate change is expected to increase the frequency and intensity of such events, heightening displacement risk. Because most disaster-related movements are internal, national and local actors (governments, NGOs, communities) have opportunities to mitigate risk and support resilience for internally displaced persons. The Internal Displacement Monitoring Centre highlights three ingredients for durable solutions: political commitment, strengthened capacity, and improved evidence. International frameworks (Paris Agreement, UNFCCC Warsaw International Mechanism on Loss and Damage, Sendai Framework, SDGs) promote disaster risk reduction and climate adaptation, but integration of displacement into national policies often occurs reactively, and preparedness remains inadequate in many countries. This study tests whether stronger environmental governance reduces disaster displacement. Using the ND-GAIN readiness index (governance, economic, and social readiness) as a proxy for environmental governance capacity across state, market, and social actors, the study assembles a global panel (92 countries, 2010–2020) to examine bidirectional links between governance and displacement. The work addresses a gap in global, humanitarian-focused quantitative analyses (beyond fatalities and economic losses) to inform adaptation and resilience strategies under climate change.

A growing literature links climate change, disasters, and internal displacement. Kam et al. estimate that each degree of global warming increases displacement risk by 50% and flood displacement by 150% by century’s end. Silva Rodríguez and San Miguel identify water-related disasters as the main environmental driver of internal migration in the Americas. Displacement pressures are most acute in the Asia-Pacific and sub-Saharan Africa (over 85% of disaster-induced displacement). In Bangladesh, Saha and Ahmed document negative correlations between socioeconomic conditions and internal displacement, consistent with findings by Lim and Khan in Southeast Asia; South Asia’s climate-sensitive livelihoods and extreme events heighten vulnerability. De Sherbinin et al. analyze displacement drivers from floods, sea-level rise, and droughts in Latin America. Brant and Mistral show drought-driven displacement via impacts on agriculture in Africa. Bordern/Bordreron et al. detail the complex interplay of sociodemographic, economic, and political factors shaping environmental mobility in Africa. Collectively, studies underscore the multifactorial nature of displacement, with vulnerability and adaptive capacity embedded in social, political, and economic systems of environmental governance. However, prior work is often local or regional; this study contributes a global perspective linking environmental governance (ND-GAIN readiness) and displacement across countries.

Conceptual framework: Environmental governance encompasses policies, norms, institutions, decision-making, and behaviors for disaster risk reduction, adaptation, and resilience. It involves state (legislative, executive, judicial), social (civil society, NGOs, humanitarian actors), and market (private sector, insurance) actors. Effective governance relies on state capacity/government effectiveness, horizontal/vertical coordination, and inclusive networks with transparency and accountability. Readiness in ND-GAIN is used as a proxy for governance capacity across these actors: governance readiness (regulatory quality, rule of law, public management), economic readiness (investment climate for adaptation, access to capital, Doing Business indicators), and social readiness (socioeconomic conditions, inequality, ICT, education, innovation). Data: Annual country-level data come from the Global Internal Displacement Database (GIDD/IDMC) and the ND-GAIN Country Index. The study builds a panel of 92 countries for 2010–2020, excluding countries with multiple data gaps. Analyses focus exclusively on weather-related disasters (flood, storm, drought). To assess robustness to outsized historical catastrophes, a supplementary analysis excludes events deemed “great” catastrophes by Munich Re (homeless >200,000). For each country-year, the percentage of the population displaced by weather-related disasters is computed (all-scale and small-scale subsets). ND-GAIN readiness and its components (governance, economic, social) are scaled 0–100 (higher is more ready). Descriptives (Table 1) indicate, across all countries: mean displaced (all-scale) 0.27% (SD 0.79; min 0.00; max 9.26); mean displaced (small-scale) 0.13% (SD 0.43; max 8.29); mean ND-GAIN readiness 36.95 (SD 12.56; range 11.48–82.32); economic readiness mean 37.68 (SD 16.68; range 0.00–88.06); governance readiness mean 42.65 (SD 14.48; range 11.00–89.68); social readiness mean 34.50 (SD 13.99; range 11.96–85.65). Readiness varies by income level, highest in high-income countries and lowest in low-income countries. Time trends show social readiness increasing globally, governance readiness rising to 2014 then declining, and economic readiness driving most fluctuations; there is an overall negative correlation between readiness and displacement. Econometric approach: To ensure stationarity, panel unit root tests are conducted: Levin-Lin-Chu (LLC), Im-Pesaran-Shin (IPS), and ADF. All variables are level-stationary (Table 2). Granger non-causality tests are then applied in a heterogeneous panel framework (following a heterogeneous linear dynamic panel approach as in Joudi et al.). Two equations are estimated: (i) readiness on lags of readiness and displacement; (ii) displacement on lags of displacement and readiness. Optimal lag length (up to 3) is selected via BIC. Analyses are run for: (a) all weather-related disasters (all scales); (b) small-scale weather-related disasters; and (c) stratified by three income-groupings (higher-income: high + upper-middle; middle-income: upper-middle + lower-middle; lower-income: lower-middle + low) to accommodate shifting classifications and ensure adequate group sizes. Estimation uses Stata 17. Data and code are available on GitHub (links provided).

Descriptive patterns: Most displacement arises from large-scale hydro-meteorological events (floods, typhoons), predominantly in developing, populous countries. Regions with low readiness (e.g., sub-Saharan Africa) exhibit high displacement rates. Several large economies (Russia, China, India, United States, Canada, Brazil) show declines in economic readiness in 2020 versus 2010. Social readiness increased steadily; governance readiness rose to 2014 then declined. All-scale disasters (Table 3): • Displacement Granger-causes higher readiness. Notable coefficients include: ND-GAIN readiness Beta.L1 = 0.503***; Beta.L3 = 0.809**. Economic readiness Beta.L1 = 1.401**; Beta.L3 = 1.788*. Governance readiness Beta.L1 = 0.342**; Beta.L2 = 0.392***. Social readiness Beta.L1 = 0.112***; Beta.L2 = 0.064*. Interpretation: higher past displacement predicts improvements in readiness—economic in the longer run and governance/social in the short run—reflecting post-disaster investment, institutional strengthening, and social capacity building. • Readiness Granger-causes lower displacement. ND-GAIN readiness Beta.L2 = −0.057***; economic readiness Beta.L2 = −0.009***; governance readiness Beta.L2 = −0.048**. Social readiness effects are negative but not statistically significant. Interpretation: more effective environmental governance—especially governance and economic readiness—predicts reduced displacement two years later. Small-scale disasters (Table 4): • Displacement effects on readiness are mixed: economic readiness shows negative feedback (Beta.L1 = −0.598*; Beta.L3 = −0.370**), consistent with disruptive effects of frequent small events; governance readiness shows positive coefficients at lags (Beta.L2 = 0.821***; Beta.L3 = 0.253**); social readiness positive at short lags (L1, L2). • Readiness to displacement: governance readiness exhibits significant negative effects at some lags, while economic and social readiness generally do not Granger-cause small-scale displacement. By income groups (Tables 5–6; text synthesis): • Higher-income group: Displacement predicts increases in economic and governance readiness; readiness predicts lower displacement primarily via governance readiness (significantly negative governance coefficient), underscoring the importance of capable institutions. • Middle-income group: Readiness predicts lower displacement with significantly negative coefficients, mainly through economic and governance readiness. • Lower-income group: Negative feedback coefficients are observed across all three readiness components (economic, governance, social), highlighting the need for comprehensive strategies addressing development, institutions, and social conditions. • For small-scale events, displacement tends to depress economic readiness across all income groups; governance readiness consistently helps reduce displacement across groups; small-scale events may exacerbate inequalities in lower-income settings. Overall, results indicate a robust, bidirectional link: past displacement can catalyze improvements in readiness, while stronger readiness—especially governance and economic—mitigates future displacement.

The findings directly address the research question by demonstrating a two-way dynamic between environmental governance (ND-GAIN readiness) and disaster-induced internal displacement. Mechanistically, improved governance readiness (aligned with WGI dimensions of government effectiveness, regulatory quality, and rule of law) can reduce displacement through better-prepared institutions, coordinated response, and effective policy implementation across sectors. Enhanced economic readiness—capturing investment climate and access to finance (Doing Business indicators)—supports adaptation financing, resilient infrastructure, and emergency planning, thereby lowering displacement risk, particularly in middle- and low-income countries. Social readiness improvements (reducing inequality, strengthening ICT, education, and innovation) bolster community resilience and adaptive capacity, with pronounced effects in lower-income contexts where vulnerabilities are greatest. The positive effect of past displacement on subsequent readiness suggests that disaster experience spurs policy learning, investment, and institutional strengthening, translating short-term shocks into long-run capacity gains. Income stratification reveals threshold effects: in higher-income countries, economic readiness may have reached levels where marginal gains yield limited additional displacement reduction, making governance quality the key lever. In middle- and lower-income countries, both economic and governance readiness are central, and in the lowest-income settings, social readiness is also essential, indicating that comprehensive, multisectoral governance strategies are required. For small-scale events, the disruptive, frequent nature can strain economic conditions, especially in lower-income countries, while capable governance remains a consistently protective factor. These insights emphasize prioritizing governance effectiveness, enabling adaptation investment, and addressing social inequalities to reduce displacement across diverse income contexts.

This study provides global, empirical evidence of a strong, bidirectional relationship between environmental governance (ND-GAIN readiness) and internal displacement from weather-related disasters across 92 countries (2010–2020). For all-scale events, displacement predicts subsequent improvements in readiness (economic in the long run; governance and social in the short run), while higher readiness—especially governance and economic—reduces displacement with a two-year lag. Stratified analyses show governance readiness as the most consistently important factor across income groups, with economic readiness particularly salient in middle- and low-income countries and social readiness additionally important in lower-income settings. Policy implications include: (1) strengthening governance effectiveness (institutional stability, regulatory quality, rule of law) to manage extreme events; (2) enhancing economic readiness to finance adaptation (resilient infrastructure, emergency planning, targeted support to affected businesses and individuals), especially in middle- and low-income countries; and (3) advancing social readiness via reducing inequalities, improving ICT, education, and innovation, and empowering communities and civil society. Given transboundary climate risks, global analyses complement local studies by revealing patterns that can guide international cooperation and investment. Sustained, coordinated action by state, market, and social actors is essential to reduce displacement risk and build long-term resilience and sustainable development.

The analysis is limited to countries and years with sufficient data availability, resulting in a 92-country panel for 2010–2020 and the exclusion of countries with multiple data gaps. The study focuses exclusively on weather-related disasters, which may limit generalizability to geophysical events. Sensitivity analyses exclude “great” catastrophes (homeless >200,000) to mitigate distortion from outsized events, which may affect comparability. Income-group stratification combines categories to accommodate changing classifications over time, potentially masking heterogeneity within groups. Methodologically, results rely on panel Granger non-causality tests (after confirming stationarity), which identify predictive temporal relationships based on lag structure rather than structural causation.