Climate econometric models indicate solar geoengineering would reduce inter-country income inequality

The paper investigates how solar geoengineering (reflecting solar radiation to cool the planet) might affect global and distributional socio-economic outcomes, particularly inter-country income inequality, relative to greenhouse gas (GHG)-driven climate change. Motivated by the challenges of costly mitigation and free-rider incentives, the authors seek to evaluate solar geoengineering impacts on equal footing with conventional climate change scenarios using state-of-the-art macroeconomic climate impact models. They note that conventional reliance on temperature as a proxy for damages is complicated under geoengineering because correlations between temperature and other variables (e.g., precipitation, ocean pH) differ from GHG-driven changes. The study applies empirically estimated climate-economy relationships, linking mean annual temperature and precipitation to GDP per capita growth, to stylized future climate scenarios to assess effects on global growth and inter-country inequality. The introduction also acknowledges scope limits (focus on annual-mean temperature and precipitation) and ongoing debates about the reliability of macroeconomic climate impact estimation, which the authors address through sensitivity analyses across multiple econometric specifications.

The paper situates its contribution within literature on solar geoengineering’s physical climate effects (e.g., GeoMIP results showing potential for cooling but different precipitation responses), and macroeconometric assessments of climate impacts (Dell et al., Burke et al.). It highlights that while a decade of research has improved understanding of solar geoengineering’s climate impacts, impacts assessment still lags behind evaluations of GHG-driven change due to differing climate variable correlations and a historical focus on temperature proxies. Prior studies suggest non-linear temperature effects on economic outcomes and limited robust evidence for precipitation effects on growth. The paper also references concerns about hydrological changes under geoengineering and notes debates over econometric methods (e.g., panel models, fixed effects, adaptation). This work extends the climate-econometric framework to solar geoengineering to compare distributional outcomes across scenarios.

• Climate scenarios: Four illustrative scenarios are constructed using widely used climate projections.
  1. No climate change: present-day climate held constant.
  2. RCP8.5: high-emissions scenario; grid-cell projections (2081–2100 vs. 1986–2005) from CMIP5 ensemble means for temperature and precipitation; interpolated linearly from 2010 to 2100; aggregated to country-level population-weighted means (population distribution circa 2000).
  3. Geoengineering-stabilized RCP8.5: solar geoengineering offsets RCP8.5 warming to keep global mean temperature at 2010 levels throughout the century.
  4. Geoengineering-mirrored RCP8.5: solar geoengineering cools global mean temperature at the same rate RCP8.5 would warm it. Solar geoengineering responses use the GeoMIP G1 experiment ensemble mean (12 models), with country-level population-weighted aggregation; also sensitivity runs per individual model.
• Socioeconomic baselines: Shared Socioeconomic Pathways (SSPs) provide country-level projections of population and baseline GDP/capita growth (no-climate-change). Main text emphasizes SSP3 (high challenges to mitigation and adaptation), with results for all SSPs in Supplementary Materials.
• Econometric impact models: Historical climate-economy relationships are estimated following Dell et al. and Burke et al., using interannual and cross-country variation (1960–2010) in annual mean temperature and precipitation for 165 countries to link these climate variables to GDP per capita growth. Multiple specifications are estimated to capture uncertainty: linear vs. quadratic climate terms; pooled vs. separate functions for rich and poor countries; contemporaneous vs. up-to-5-year lags; growth vs. level effects; different fixed effects (year, region-year) and country time trends (quadratic or none). Temperature emerges as consistently significant; precipitation generally not.
• Projections: For each country and year (2010–2099), baseline growth from the chosen SSP is adjusted by the estimated impact function according to deviations of climate variables from 2010 baselines under each scenario, iteratively generating a GDP/capita trajectory which is then applied to initial 2010 GDP/capita.
• Uncertainty: Bootstrap of impact function estimation (N=1000, sampling countries with replacement) to produce median and 95% confidence intervals. Sensitivity to climate model uncertainty assessed by using each GeoMIP model individually and by varying precipitation responses while holding temperature response at the ensemble mean.

• Temperature-driven impacts dominate: Across models, temperature significantly affects economic outcomes; precipitation effects are limited, making projected outcomes mainly temperature-driven.
• Global GDP effects:
  • Using Burke et al.-style models, geoengineering that stabilizes global temperature mitigates warming harms and modestly increases global GDP relative to RCP8.5 and even to no-climate-change in some specifications due to more zonally uniform temperatures.
  • Geoengineering-mirrored (cooling at RCP8.5’s warming rate) yields substantial global GDP increases driven by rapid growth in warmer developing nations currently above the temperature optimum; however, magnitudes are model-dependent.
• Country-level outcomes (SSP3, Burke-style model):
  • No climate change and Geoengineering-stabilized RCP8.5: No country is poorer in 2100 than in 2010.
  • RCP8.5: 43% of countries are poorer at century’s end and 76% are relatively poorer than under SSP3 alone.
  • Geoengineering-mirrored RCP8.5: 11% of countries are poorer at century’s end and 32% are relatively poorer than under SSP3 alone.
• Inequality reduction:
  • Lorenz curves (2099, SSP3) show RCP8.5 eliminates expected income convergence; geoengineering-stabilized restores convergence; geoengineering-mirrored enhances convergence overall but may entail relative losses for the lowest-wealth quartile (e.g., poorest country in 2100 under geo-mirrored is Mongolia at $316/capita vs. $860/capita in 2010).
  • Gini coefficients decrease with solar geoengineering across all econometric models; geo-mirrored scenario yields the lowest Gini. RCP8.5 consistently increases inequality and the share of countries with poor growth.
• Robustness:
  • The reduction in inter-country income inequality with solar geoengineering is consistent across econometric specifications, SSPs, and GeoMIP climate models.
  • Varying only precipitation responses across GeoMIP models has little effect on economic outcomes; uncertainty in temperature responses is more consequential.
• Climate response characteristics:
  • Solar geoengineering reduces global precipitation more per degree of cooling than GHGs increase it per degree of warming; uniformly applied forcing tends to overcool equatorial regions relative to poles and reduces the equator-to-pole temperature gradient.

The study addresses whether solar geoengineering would exacerbate or alleviate global income disparities by applying the same macroeconomic climate-econometric tools used for GHG-driven climate damage assessments. Findings indicate that harms from warming and benefits from cooling accrue disproportionately to warmer, poorer, more populous countries; thus, solar geoengineering consistently reduces inter-country income inequality across models and scenarios. While total global GDP effects are model-dependent, inequality outcomes are robust. The analysis challenges prevalent concerns that solar geoengineering favors developed countries or yields large residual economic harms; instead, outcomes suggest greater equity in economic terms relative to both RCP8.5 and even a no-climate-change counterfactual. However, extreme geoengineering scenarios with the greatest aggregate gains may be politically infeasible and can entail relative losses for the lowest-wealth quartile, raising governance and ethical considerations. The results also indicate that, contrary to common concerns, uncertainty about precipitation responses under geoengineering is less important for economic outcomes than temperature responses, given the limited role of precipitation in empirically estimated growth impacts. The authors emphasize that if macroeconomic climate-econometric models are considered valid for projecting GHG-driven damages, they should also be applied to solar geoengineering to compare policy options consistently; if skepticism arises, it should apply to both contexts.

The paper extends climate-econometric impact modeling to solar geoengineering and finds that, across diverse model specifications, socioeconomic pathways, and climate models, solar geoengineering reduces inter-country income inequality and can mitigate or reverse warming-related economic harms. Aggregate global GDP benefits are possible—especially under aggressive cooling—but their magnitudes depend on econometric assumptions and may involve trade-offs for the poorest countries. The study underscores the need to evaluate solar geoengineering with the same empirical tools used for climate damage estimation to place policy options on equal footing. Future research directions include: assessing alternative, more realistic deployment strategies and governance scenarios; integrating additional impact channels (e.g., extremes, sea-level rise, ocean acidification, UV changes) into macroeconomic assessments; resolving methodological debates in climate econometrics (e.g., role of precipitation, fixed effects, adaptation) to improve robustness; and examining political feasibility and ethical implications of heterogeneous impacts.

• Impact channels limited to annual-mean temperature and precipitation; extremes, variability, and sea-level rise are not directly captured by the empirical models and are only reflected to the extent they co-vary with means.
• Important non-temperature impacts unaddressed by solar geoengineering (e.g., ocean acidification, CO2 fertilization effects, ground-level UV changes) are excluded; preliminary evidence suggests their economic effects may be smaller than temperature-driven impacts but remain uncertain.
• Empirical climate-econometric methods are debated; projections depend on model specification (growth vs. level effects, linear vs. non-linear responses, fixed effects, lag structures) and their validity for future and geoengineered climates.
• Precipitation-temperature relationships under future GHG and geoengineered climates may decouple from historical correlations, potentially challenging identification in fixed-effects panel models.
• Stylized geoengineering scenarios (global stabilization or mirrored cooling) may be politically and legally infeasible; real-world deployment could be heterogeneous or unilateral, affecting outcomes.
• Country-level focus does not capture within-country inequality or distributional impacts across communities.