Paris Climate Agreement passes the cost-benefit test

The Paris Climate Agreement's temperature targets, resulting from a complex political process, raise questions about their economic feasibility. While econometric analyses suggest substantial damages at higher temperatures, these haven't been used to determine the economic benefits of meeting the targets. This study addresses this gap by incorporating recent estimates of warming-induced economic growth impacts (Burke et al., BHM) into a macroeconomic assessment. BHM's work reveals a universal non-linear relationship between warming and economic growth, showing reduced growth beyond a temperature threshold. Previous studies using BHM's estimates have indicated a high social cost of carbon, suggesting stringent emission reduction. However, the implications for optimal policy have only been explored with predetermined warming and growth scenarios. This study aims to compare BHM's estimates to climate change mitigation costs using an integrated assessment model (IAM). IAMs account for the dynamic interactions between the economy and the climate. This comparison helps determine the end-of-century warming level associated with the lowest total costs (damages and mitigation). The optimal warming is determined by the shapes of the mitigation-cost and damage-cost curves. Mitigation costs diverge at present-day warming (especially without negative-emission technologies) and decrease to zero without mitigation. Damage costs, conversely, are zero without warming and increase with rising temperatures. The study investigates whether damages extrapolated from the observed economic growth-temperature relationship exceed the level at which the optimal temperature becomes insensitive to further damage increases. The study utilizes the DICE-2013 IAM due to its transparency and widespread use in policy advising.

The study references several key works on the economic impacts of climate change. Dell et al. (DJO) found a linear relationship between temperature and growth, impacting only poor countries. Burke et al. (BHM) updated this, finding a non-linear, quadratic relationship affecting all countries. Other studies have explored the implications of BHM’s estimates, often finding high social costs of carbon and advocating for stringent mitigation. However, these studies often used predetermined scenarios, rather than integrated assessment models (IAMs) allowing for dynamic interactions between economic and climate factors. Prior work using IAMs like DICE often lacked the updated damage functions incorporating recent estimations of economic impacts. Several studies have attempted to integrate temperature-growth relationships into IAMs like DICE, employing different methods such as adjusting total factor productivity (TFP) or incorporating multiple impact channels. The authors note that existing models often lack comprehensive empirical validation and involve arbitrary assumptions about the underlying mechanisms. The current study aims to advance this literature by consistently integrating the most recent empirical estimates (BHM) into a dynamic IAM to develop a damage function and assess optimal climate policy without making arbitrary assumptions.

The authors utilize the DICE-2013 integrated assessment model, known for its simplicity and transparency, to assess cost-benefit optimality. They develop a novel procedure to incorporate BHM's temperature-induced growth relationship into DICE-2013, while preserving the model's growth characteristics. This involves iteratively adjusting the damage function to reproduce the estimated temperature-induced growth relation. A key aspect is the creation of a scenario emulating a future without climate policy to maintain consistency with BHM's estimates. The study assumes no significant negative-emission technologies in this century. The optimal climate policy is determined by maximizing global welfare, starting in 2020. The optimal policy is compared to a business-as-usual (BAU) scenario with no climate policy. The analysis accounts for uncertainties in climate sensitivity (ECS) by considering three alternative values and conducts extensive robustness tests. These tests include: 1. Bootstrapping the BHM estimates to account for parameter and model specification uncertainties. 2. Sensitivity analysis regarding social preferences for consumption changes, socioeconomic futures (SSPs), and mitigation costs. The damage function is recalibrated using an iterative algorithm that disentangles productivity loss from investment responses while maintaining DICE's growth model feature. The algorithm iteratively adjusts the damage function, generating a temperature-dependent damage function consistent with BHM's growth estimates. This function is then used in policy runs to determine the cost-benefit optimal temperature path, accounting for uncertainties in ECS values, damage estimates, social preferences, and mitigation costs. The study also recalibrates DICE-2013 to start in 2020, using updated GDP and emission data. Sensitivity analyses include the consideration of various socioeconomic pathways (SSPs) and the impact of different assumptions about the availability of negative-emission technologies.

The study's key finding is that limiting warming to 2°C, as targeted by the Paris Agreement, represents the cost-benefit optimal temperature pathway until the end of the century for the base calibration (ECS of 2.9°C). This result remains robust across various uncertainties. Higher ECS values shift the optimal target slightly higher (2.4°C for 4°C ECS), while lower ECS values (2°C) suggest more stringent targets. Compared to a business-as-usual scenario, significant damage reduction is observed (~14% for 4°C ECS, ~10% for 2.9°C ECS, and ~8% for 2°C ECS). The analysis of uncertainty in the damage function, using the cumulative GDP losses in the BAU scenario, shows that even with substantial uncertainty, the median optimal temperature is close to 2°C (slightly above for 2.9°C ECS and well below for 2°C ECS). The results are robust to alternative bootstrapping approaches and various model specifications from both BHM and DJO. Sensitivity analysis regarding social preferences (time preference and inequality aversion) reveals that, except for a few unusual parameter values, the optimal warming remains around 2°C or lower. Furthermore, the analysis shows robustness to alternative socioeconomic assumptions (SSPs). SSP1 and SSP2 suggest lower optimal temperatures, while SSP5 implies a slightly higher optimal temperature (around 2.5°C) due to the difficulty of successful climate policy under a fossil-fuelled development scenario. An additional sensitivity test considering the availability of negative-emission technologies shows that harnessing this potential further reduces optimal end-of-century temperatures. Finally, the study finds that the results are robust to the use of alternative mitigation-cost functions calibrated against a detailed process model.

The findings strongly support the Paris Agreement's 2°C target, demonstrating its economic optimality under various scenarios and uncertainties. The study’s robustness across several key parameters provides strong evidence for the economic benefits of ambitious climate action. The key driver behind this result is the substantial increase in marginal damage costs at higher temperatures combined with the mitigation costs’ universal functional behavior near present-day temperatures. While the model uses GDP as a measure of damages, excluding non-monetary factors like loss of life, and uses a simplified, global model, these factors may require even stronger mitigation than indicated. Future research should investigate regional impacts and integrate more detailed empirical estimates into more complex IAMs, accounting for regional heterogeneity and distributional issues. The analysis considers only gradual warming pathways, and the additional economic costs of temperature variability are likely to further support stringent mitigation efforts. The study addresses some limitations of existing research by using a consistent, up-to-date damage function derived from the BHM study within a dynamic integrated assessment model.

This study demonstrates that the Paris Agreement's 2°C temperature target aligns with economic optimality, even when accounting for significant uncertainties. The robustness of this finding across different climate sensitivities, socioeconomic pathways, and model specifications strengthens the case for ambitious climate action. While limitations remain (simplification of the model and the use of GDP as the primary damage indicator), this analysis provides compelling evidence for the economic benefits of following the Paris Agreement's targets. Future research should focus on improving the integration of regional economic impacts and the complexities of climate-related risks into more detailed IAMs.

The study uses GDP as a primary measure of economic damage, which may underestimate the full costs of climate change by neglecting non-monetary impacts like loss of life and biodiversity. The study uses a simplified one-region global IAM (DICE), which doesn't fully capture regional variations in climate impacts or the distributional effects of mitigation and adaptation measures. The extrapolation of historical temperature-impact relationships into the future might not accurately reflect future societal responses or potential tipping points. The study assumes that significant negative-emission technologies are not available, and the findings may change with advances in this technology. The analysis uses a simplified representation of the carbon cycle in DICE, which may lead to underestimation of the temperature effects and necessary mitigation efforts. The robustness tests account for a range of uncertainty but do not cover all potential factors influencing the outcome.