Does climate action bring peace? Assessing the geopolitics of renewables using global investment data

The study examines whether the global shift to renewable energy contributes to peace by reducing conflict and violence. The background is that fossil fuel systems are associated with corruption, institutional capture, and conflict, while renewables are more geographically widespread and potentially decentralizing. The authors note that the literature often conflates peace with absence of violence and that multiple pathways (democracy, development, human security) may link renewables to peace, with effects likely contingent on context (conflict-prone vs. stable settings). They formulate three hypotheses: H1 that increasingly distributed energy systems enhance participatory democracy (energy democracy → democratic peace); H2 that increased renewable deployment fosters economic development (capitalist peace); and H3/H4 that distributed energy systems advance human development and reduce violence, enhancing political stability (human security). The paper aims to provide large-N empirical tests using global investment data to detect early indicators of these hypothesized effects, recognizing temporal lags and equifinality.

The paper synthesizes three strands: (1) Energy democracy: Decentralized renewables purportedly redistribute political and economic power to local actors (prosumers, cooperatives), increasing participation and accountability, which could scale to democratic peace internationally. (2) Capitalist peace/commercial liberalism: Economic development and openness reduce incentives for conflict; renewables may spur growth, diversification (notably in resource-rich economies), job creation, and cross-border investment interdependence, thereby lowering conflict likelihood. (3) Human security: Focus on ‘freedom from fear and want’; access to clean energy can reduce energy poverty, inequalities, and climate vulnerabilities, fostering human development and local stability, with potential spillovers to international peace. The authors emphasize contextual contingencies—conflict settings may moderate these relationships—and note debates on renewables’ own geopolitical risks (e.g., rare earths, uneven transitions, new forms of extractivism).

Design: Country-year panel econometric analysis (1996–2019) with country and year fixed effects, one-year lagged independent variables, standardized unbounded variables, and panel-corrected standard errors to address heteroskedasticity and contemporaneous correlation. Tests implemented in R (plm). Model: Y_ct = β1 inv_ct-1 + β2 controls_ct-1 + μ_c + ε_ct, with Y being democracy, GDP per capita, HDI, or political stability. Independent variables: (i) Concentration of renewable energy (RE) project finance investment per country-year, proxied by a Herfindahl–Hirschman Index (HHI) over project investment sizes (lower HHI = more distributed; higher = more concentrated). (ii) Total RE project finance investment per capita per country-year. RE investment derived from Thomson Reuters SDC Platinum project finance database (8,493 RE project-level observations, 1991–2019; estimated total value US$2.89 trillion), covering solar, wind, hydro, geothermal, biomass; robustness checks exclude hydro (results similar). Dependent variables: - Participatory democracy: V-Dem participatory democracy index (0–1); robustness checked with Polity IV and other V-Dem indices. - Economic development: World Bank GDP per capita (current US$). - Human development: UNDP Human Development Index (0–1). - Political stability/absence of violence: World Bank Worldwide Governance Indicators (−2.5 to 2.5). Controls: World Bank Governance Indicators—regulatory quality; rule of law (as applicable); IMF Financial Development Index (depth, access, efficiency); World Bank fuel exports share of merchandise exports; pertinent dependent variables used as controls across models per Table 2 (e.g., GDP per capita, HDI, pol.stab). Conflict moderator: Dummy (UCDP/PRIO Armed Conflict Dataset v20.1) equals 1 when a country-year experiences intrastate or internationalized intrastate conflict (≥25 battle-related deaths); interaction terms constructed with RE investment variables. Sample and estimation: Main period 1996–2019 (complete controls). Fixed effects preferred over pooled OLS (F tests); time effects necessary (Breusch–Pagan LM); PCSE used (Beck–Katz). Collinearity acceptable (VIF < 8; mostly < 5). Lag t−1 chosen; t−2 robustness yields similar significance. Panel unbalanced due to data availability. No differentiation by RE technology in main models; hydro exclusion robustness performed. HHI acknowledged as a proxy with limitations (e.g., large equal-sized projects may appear as ‘decentralized’).

• Participatory democracy (H1): Distributed RE investment (lower HHI) does not increase participatory democracy. In bivariate/limited-control models, higher concentration (higher HHI) correlates positively with democracy (e.g., coefficient ≈ +0.010 in simple model), contrary to the energy democracy hypothesis; with fuller controls (governance, financial development), RE concentration loses significance while governance and financial development remain strong predictors. No significant interaction with conflict. - Economic development (H2): Total RE investment per capita shows no significant positive association with GDP per capita in any specification. Controls—regulatory quality, political stability, financial development—are strong, significant predictors; fuel export share is negatively associated with GDP per capita. Conflict and interaction with RE investment are not significant. - Human development (H3): Contrary to expectations, higher RE investment concentration (higher HHI) associates positively with HDI in simpler models (≈ +0.005), but loses significance when comprehensive controls are included. Conflict significantly moderates the relationship: the positive effect of concentration on HDI is weaker in conflict settings (significant negative interaction term, e.g., −0.024 to −0.018), indicating context dependence. - Political stability/absence of violence (H4): Higher RE investment concentration is positively associated with political stability and absence of violence across models, including with full controls (coefficients ≈ +0.083 to +0.108; within R² ≈ 0.02). This contradicts the hypothesis that more distributed systems reduce violence; instead, concentrated investment correlates with greater stability. Overall: No evidence that renewables foster peace via democratization or prosperity. Limited evidence suggests concentrated investment correlates with stability and, context-dependently, with human development; the ‘peace through prosperity’ pathway is not supported.

Findings challenge the assumption that decentralized renewables democratize politics and lead to international peace. Distributed systems do not improve national participatory democracy; if anything, concentrated investment aligns with higher democracy in simple models, though this vanishes once governance and financial development are considered. The capitalist peace pathway is unsupported: increased RE investment does not raise GDP per capita. For human security, distributed RE does not generally improve HDI or reduce violence; rather, concentrated investment correlates with greater political stability, and its positive association with HDI weakens in conflict contexts, underscoring context dependence. Possible explanations include: (i) renewables’ share in the energy mix remained relatively small for much of the study period, diluting macro effects; (ii) democratizing and empowerment effects may be localized and not captured by national indices; (iii) reverse causality and selection—countries with better governance and financial systems both attract RE investment and exhibit higher democracy/development; (iv) measurement limits of the HHI proxy and macro indicators. The results suggest alternative mechanisms may connect renewables and peace (e.g., environmental peacebuilding), and that micro- to meso-level processes and longer time horizons merit investigation. Policymakers should be cautious in assuming renewable deployment reduces conflict; benefits for climate and access remain compelling, but peace dividends are not evidenced by these macro-panel findings.

The study finds no empirical support that renewable energy deployment, particularly in distributed forms, brings about peace by enhancing democracy, prosperity, or human security at the country level. Instead, higher concentration of renewable investments is associated with increased political stability and, in some models, with human development—effects that are attenuated in conflict settings. These results call for theoretical refinement and context-sensitive, process-focused research, including qualitative case studies, deviant case analysis, and before–after designs with process tracing, to uncover micro- and meso-level mechanisms and temporal dynamics. Policy linkage of renewables to peace should be approached with caution; while renewables are vital for climate action and may improve lives locally, expecting them to systematically reduce conflict is not supported by the evidence presented.

• Temporal scope and early-stage transition: Much of the hypothesized peace impact likely lies in the future; renewables’ share of the energy mix was modest for much of 1996–2019, potentially muting macro effects. - Causal identification: Large-N fixed-effects panel detects associations with lagged predictors but cannot unpack detailed causal mechanisms; residual endogeneity (e.g., reverse causality, time-varying omitted variables) may remain. - Measurement: HHI as a proxy for decentralization can misclassify systems with many large equal-sized projects; national-level indices (democracy, HDI, stability) may not capture localized empowerment or micro-level effects. - Data constraints: Unbalanced panel; proprietary investment data; no differentiation by RE technology in main models (though hydro exclusion robustness performed). - Context and heterogeneity: Equifinality and context dependence are acknowledged; interactions with conflict captured only by a binary indicator, which may oversimplify conflict intensity and proximity. - Time lags: One-year lag may be too short for institutional or development effects; longer lags could be informative but require additional theoretical justification. - Scope: Focus on two investment measures (concentration, total per capita) may omit other relevant features (e.g., existing energy capital stock, policy frameworks, socio-political cleavages).