SCC Watch: New Damage Estimates, Weitzman Discounting, and the Vulnerability Assumption

A useful SCC week does not give one number. It tells you which assumptions are doing the work. The newest batch points to three live arguments: whether temperature shocks should be modeled as persistent growth damage, whether discounting should give more weight to patient preferences in the far future, and whether richer societies become less vulnerable to climate damage as they develop.

The high-damage side of the literature is getting more operational. Burke et al. do not just report another SCC estimate; they build an accounting framework that separates historical damages already incurred, future damages from past emissions, and future damages from present emissions. That decomposition matters because a settlement for harms already observed does not extinguish the future damage stream from the same tonne of CO2. Under their 2% discount-rate example, future damages from a 1990 emissions pulse are ten times the historical damages through 2020. 1

The same paper is also a reminder that damage-function choice is not a technical footnote. Their high SCC comes from a temperature-GDP damage function with lagged growth effects. If temperature shocks reduce the growth path, damages compound; if they mostly affect the level of output and economies rebound, estimates fall sharply. The paper itself notes that allowing growth rebounds cuts SCC estimates by half, while faster adaptation could reduce future damages substantially. 1

Kopits et al. are useful because they force the macro-damage literature through a common U.S.-specific modeling frame. Their synthesis still says recent studies have made progress on dynamic temperature effects, but the harmonized central range is not in the same territory as the highest global loss-and-damage numbers: $10-$64/tCO2 for 2030 emissions under a 2% near-term Ramsey rate, or $31-$85/tCO2 after adding heat- and cold-related mortality. 3

That does not mean "low SCC wins." The paper is explicit that many nonmarket impacts remain omitted, and that macro-econometric GDP studies miss damages not visible in historical annual country-level average temperature shocks. It means the more careful debate is about which damages are represented, which are missing, and whether macro GDP estimates can be combined with endpoint-specific damage models without double counting. 3

The Weitzman-premium paper makes the pro-low-effective-discounting case through heterogeneity, not by simply choosing a low rate. If society contains both impatient and patient preferences, high-discount-rate views get less influence over very distant damages because those views put almost no weight on the far future. Patient preferences therefore dominate the long horizon. Dong, Tol and Wang estimate this effect with a large preference dataset and find a very large increase in SCC relative to using average preferences. 2

The opposing position is not just "care less about the future." It asks whether private short-run preference data should govern public, intergenerational policy; whether a representative social planner should aggregate preferences this way; and whether uncertainty about discount rates should be treated as normative disagreement, empirical uncertainty, or both. The paper itself isolates preference heterogeneity in a globally aggregated IAM, leaving income, vulnerability and distributional heterogeneity for other work. 2

Vulnerability-to-income is the sleeper variable. If economic development makes societies less vulnerable, SCC estimates fall; if development increases the value of exposed assets and services, they rise. Tol's Schelling paper frames this as an old but still unresolved question, and gives a concrete magnitude: relative to constant vulnerability, SCC is $58/tC higher if vulnerability increases with growth and $126/tC lower if vulnerability decreases with growth. 5

This is also where the empirical-damage and discounting debates meet. A high SCC from persistent growth damages can still be moderated by strong adaptation or declining vulnerability. A low SCC from faster assumed growth can be challenged if richer economies become more asset-dense and exposed. The next papers that deserve attention are not only new point estimates; they are papers that identify which vulnerability channel is being measured.

The IMF has a May 2026 working paper, "When Forecasts Meet Reality: Assessing Climate Damage Functions," by Johannes Emmerling, Paul Waidelich, Matthieu Bellon and Emanuele Massetti. The accessible primary page identifies the paper, publication date, DOI, and keywords around validation, forecasts, WEO forecasts and damage functions, but the scrape did not expose enough substantive text to summarize its findings fairly. 7

That is still a useful watchlist item. If the full paper is accessible next round, it belongs near the top because out-of-sample validation is exactly what this literature needs: not another preferred damage curve, but evidence on which curves forecast realized macro outcomes better.