1 Introduction

The paper reassesses the macroeconomic impact of climate change, arguing that estimates based on local temperature substantially understate aggregate damages. Using global mean temperature as the primary climate metric, the authors estimate that a permanent 1°C increase reduces world GDP by over 20% in the long run, compared with conventional estimates of 1–3%. Global temperature better captures extreme climatic events and correlated shocks across countries.

2 Global temperature and economic growth

Two datasets are constructed. The ‘BU’ sample covers 43 countries from 1860–2019 using Barro-Ursúa data and NOAA global temperature. The ‘PWT’ sample spans 173 countries from 1960–2019 using Penn World Tables and Berkeley Earth data, complemented by ISIMIP indicators of heat, drought, wind and precipitation extremes.

Temperature shocks are defined as filtered innovations orthogonal to long-run trends. In the PWT sample, a 1°C global temperature shock reduces world GDP per capita by 14% at its peak after six years (95% confidence interval: 6–22%). In the BU sample, the peak decline reaches 18% after five years (confidence interval: 6–30%). Effects are statistically significant over multiple years and display delayed amplification rather than purely contemporaneous impacts.

3 Global temperature tn the panel of countries

Panel estimates confirm that global temperature shocks reduce GDP per capita across countries. Impacts are broad-based and more uniform than local temperature effects. Regional impulse responses show sizeable medium-run output declines across Europe, North America, Latin America, South Asia, Sub-Saharan Africa and other regions. Identification tests support a causal interpretation after controlling for confounding global factors.

4 Global vs. local temperature

Replicating conventional approaches, a permanent 1°C rise in local temperature reduces GDP by 2–3%, consistent with prior studies. However, global and local temperature shocks are weakly correlated.

Joint estimations show that global shocks produce substantially larger GDP losses than local shocks. When incorporating extreme events linked to warming, aggregated local impacts exceed 6% at peak but explain only around half of the global effect. Ocean temperature shocks drive most aggregate impacts, reflecting their role in generating extreme weather. The findings indicate that bottom-up aggregation of local damages understates total macroeconomic costs.

5 A model of climate change across the world

Reduced-form estimates are embedded in a neoclassical growth model similar to DICE. Climate change affects total factor productivity through structurally estimated damage functions calibrated to global temperature shocks.

Under a scenario where temperature rises 3°C above preindustrial levels by 2100 (2°C additional warming from 2024), output in 2050 is 28% below a no-warming baseline. By 2100, output is 53% lower, reflecting cumulative productivity losses of around 40%. Investment initially increases due to forward-looking saving behaviour, then declines as output falls.

6 The macroeconomic impact of climate change

The model implies a present-value welfare loss exceeding 30% under business-as-usual warming. The estimated Social Cost of Carbon exceeds USD 1,200 per tonne of carbon dioxide. These values are substantially higher than models calibrated to local temperature damages.

Because global emissions affect domestic welfare through global temperature, the results indicate that unilateral decarbonisation can be cost-effective for large economies. Adaptation is not modelled explicitly, and distributional effects are not fully captured.

7 Conclusion

Macroeconomic damages from climate change are estimated to be an order of magnitude larger than previously reported. Global temperature, particularly ocean-driven warming and associated extremes, is the primary determinant of aggregate economic losses, with significant implications for climate policy.