Research In Brief: Can Climate Change Be More Cost-Effective than the Best GHD Interventions?

There has been significant debate regarding the effectiveness of donating to climate versus global health and development (GHD) interventions, with some arguing that high-cost-effectiveness GHD interventions (e.g., seasonal malaria chemoprevention) offer a more compelling case due to their strong evidence, near-term benefits, and deep pool of high-impact donation opportunities at the margin.

However, our research finds that there are conditions under which climate change donations may look more attractive than GHD donations. We built three models to compare the two cause areas, calculating a simplified cost-per-life-saved (CPLS; $/life) and a more complex social return on investment (SROI) model that incorporates both mortality and economic outcomes.

In this brief, we report results for philanthropic donations to hits-based climate change funds (e.g., Giving Green or Founders Pledge Climate Fund), where we expect most efforts will likely fail, but a small number may yield transformative impacts. This is because donations to higher certainty climate causes (e.g., investing in carbon capture and storage) were orders of magnitude less cost-effective than the best GHD donation opportunities.

Our conclusions are:

• GHD interventions save more lives per philanthropic dollar donated
• Climate has a lower SROI than GHD, even when incorporating economic damages from sea level rise, energy, and agriculture impacts
• Climate’s SROI becomes competitive with GHD when accounting for high-risk, lower-certainty effects like tipping points and endogenous growth

GHD interventions save many more lives per philanthropic dollar donated

In an initial model, we calculate the cost per life saved for climate philanthropy when accounting only for heat-related deaths and fossil fuel-driven air pollution deaths.^[1] The resultant median CPLS of ~$10,000 is more expensive than that of the most cost-effective GHD opportunities (~$1600 and $2550 CPLS for adult and under-5 causes, respectively).

Interim conclusion: Climate cannot compete with GHD when only valuing lives saved.

Climate has a lower SROI than GHD, even when incorporating economic damages from sea level rise, energy, and agriculture impacts

Climate change will cause economic damages across a wide range of natural systems through pathways such as sea level rise, increased energy use for cooling, and reductions in agricultural crop yields. These economic damages are modeled and valued alongside the health impacts using Integrated Assessment Models (IAMs), which calculate a single numerical value for the “social cost of carbon” (SCC): the estimated economic cost of the damages caused by emitting one additional tonne of carbon dioxide equivalent (CO₂e) into the atmosphere, expressed in US$.

We combine our data on air pollution with the SCC of the GIVE model ($195), which accounts for temperature-related deaths, sea level rise, energy costs, and agricultural system damages. This is then translated into an ethically-weighted social cost of carbon using a valuation approach where all lives are valued the same, regardless of when or where they are lost, and economic damages are valued using a logarithmic utility function^[2]. Figure 1 displays the distribution of the SROI estimates and their proximity to the Coefficient Giving benchmark for highly cost-effective giving.

Fig 1: Distribution of SROI for hits-based climate donations compared to the CG bar

Interim conclusion: Our median estimate for the SROI of donating to hits-based climate funds falls short of the Coefficient Giving benchmark for highly cost-effective giving.

Climate’s SROI becomes competitive with GHD when accounting for high-risk, lower-certainty effects like tipping points and endogenous growth

The GIVE model only includes economic damages for impact pathways that are deemed to have robust evidence. However, there are two notable areas of omission, based on uncertainty around their impact magnitude: tipping points and endogenous growth effects.

Tipping points: The GIVE model has a smooth damage function; it assumes that damages increase linearly with temperature. However, evidence suggests that sustained exposure to global average temperatures above 1.5 °C may trigger non-linear and potentially irreversible shifts in some of Earth’s natural systems. E.g., self-perpetuating thawing of Arctic permafrost or bleaching of coral reefs. The Model for Economic Tipping point Analysis (META) model estimates damages from eight of these nonlinear pathways, ultimately concluding that the risk of tipping points should increase SCC estimates by ~25%, though this is likely a conservative estimate.^[3]

Endogenous growth effects: Standard IAMs, including GIVE, tend to assume that any climate impacts on economic growth take the form of “level effects,” For example, if a climate shock like a severe drought occurs, the model registers a drop in economic output for that specific period. Crucially, however, it assumes that the underlying engine of economic growth continues at its original rate from that newly lowered baseline. In contrast, much macroeconomic literature argues that climate change may affect the rate of growth, leading to so-called endogenous growth effects. The likely magnitude of such endogenous growth impacts remains hotly debated; however, eminent economists, such as Stern and Stiglitz,^[4] have labeled it “absurd” that standard IAMs include no reference to this impact pathway.

Applying a lower bound SCC multiplier of 2x for endogenous growth impacts brings our median SROI in line with Coefficient Giving benchmarks for cost-effective GHD giving. Applying a mid-upper range estimate of 5x pushes the SROI well above this bar. Figure 2 summarizes the relative contributions of heat deaths, air pollution deaths, GIVE economic damages, tipping point, and endogenous growth damages to the overall SROI estimate.^[5]

Fig 2: Relative contribution of health and economic damages to our SROI calculation, in CG units

Conclusions

Philanthropic donations to hits-based climate funds, such as Giving Green or the Founders Pledge Climate Fund, can rival the most cost-effective GHD opportunities when accounting for avoided damages from tipping points and endogenous growth effects. While these factors have higher uncertainty than those economic damages included in the GIVE model, we note that applying relatively conservative estimates of these damages from the literature brings climate opportunities just above the Coefficient Giving benchmark for high-impact GHD interventions.

1. For climate interventions, the CPLS was calculated as the abatement cost ($/tonne CO2e abated) divided by the mortality cost (lives lost/tonne CO2e emitted). The abatement cost was estimated from EA climate funds like Giving Green and Founders Pledge ($0.50-3.50/tonne). The mortality cost was calculated using data on heat-related deaths from the RFF-Berkeley Greenhouse Gas Impact Value Estimator (GIVE) model and air pollution deaths using data from Vohra (2021), Lelieveld (2019), and McDuffie (2021). For GHD interventions, we back-calculated the CPLS from Coefficient Giving’s under-5 and adult bars for philanthropic giving ($2550 and $1600, respectively). We do not discount the value of lives over time. ↑
2. This creates an SCC estimate of ~CG$1600/tCO₂e. Dividing these benefits by the median cost to abate one tCO₂e (~$2.00) produces an estimate that the social return on investment derived from reducing global CO₂e concentrations by one tonne is ~800x. ↑
3. A forthcoming report by Rethink Priorities indicates that the mortality impacts through the year 2100 from destabilizing Earth systems (such as albedo, the AMOC, and glaciers/ice sheets), as well as warming-induced emissions, are likely to warrant tens of millions of dollars in research and development at the CG bar for highly cost-effective interventions. ↑
4. Stern, N., & Stiglitz, J. E. (2021). The social cost of carbon, risk, distribution, market failures: An alternative approach (Working Paper No. 28472). National Bureau of Economic Research. https://perma.cc/8PL8-DPCB ↑
5. The ~800x SROI estimate derived from an ethical weighting of the GIVE model outputs is multiplied by 1.25 to account for tipping point effects, totaling ~1000x. For endogenous growth effects, the lower bound multiplier of 2x increases this to 2000x (just on the CG bar), and the mid-upper bound 5x increases it to 5,000x. ↑