Measuring global monetary damages from particulate matter and carbon dioxide emissions to track sustainable growth

The study addresses how to measure sustainable economic growth by integrating natural capital into national accounts. Traditional national income and product accounts (NIPAs) such as GDP omit the value of natural resources in situ and external costs from pollution, limiting policymakers’ ability to assess sustainability. The authors argue for monetization of environmental impacts to allow deduction of environmental costs from market activity (yielding net growth), enable aggregation across disparate environmental media into unified indices, and capture evolving societal willingness-to-pay (WTP) for environmental quality and health risk reductions as incomes rise. Adopting the sustainable growth criterion of non-negative comprehensive capital formation (including man-made and natural capital), the paper focuses on two key determinants of global sustainable growth: monetary damages from ambient PM2.5 exposure (primarily via increased mortality risk) and from CO2 emissions (valued via the Social Cost of Carbon reflecting future global damages). The objective is to construct globally comprehensive, country-year monetary damage accounts for PM2.5 and CO2 for 165 countries over 1998–2018 and integrate them with GDP to form environmentally adjusted value added (EVA), enabling reassessment of growth and pollution damage intensity patterns over time and across the development spectrum.

The paper builds on several strands: (1) Extensions to NIPAs and green accounting that argue for incorporating nonmarket environmental values into national accounts. (2) Prior environmental accounting efforts that monetized pollution damages in the U.S. economy and showed sizable PM2.5 and CO2 damages relative to output. (3) The Environmental Kuznets Curve (EKC) literature, which traditionally relates physical pollution measures to income; the authors extend this by plotting monetary damages versus income, noting how WTP/VSL and the rising SCC shift damage peaks relative to emission peaks. (4) SCC development, where recent advances suggest higher SCC values and rising SCC over time as atmospheric stocks accumulate. (5) VSL transfer literature supporting income-elastic VSLs that are higher in low- and middle-income settings and lower in high-income settings, which the authors adopt for international valuation. Collectively, these literatures motivate monetized, globally comparable environmental-economic accounts and inform key parameters (SCC, VSL elasticities) used in the analysis.

Scope and data: The study computes monetary damages from PM2.5 exposure and CO2 emissions for 165 countries annually from 1998–2018 and integrates them with GDP to form EVA. Real GDP (constant 2010 USD) is from the World Bank. CO2 damages: Country-year CO2 emissions are from the Emissions Database for Global Atmospheric Research (EDGAR). Monetary damages are computed using the global Social Cost of Carbon (SCC) from the U.S. Interagency Working Group (IWG): central estimate $36 (2007$) in 2015 at a 3% discount rate. Annual SCC values for the study years are derived via growth rates between IWG benchmarks. SCC varies over time but not across countries (reflecting globally mixed impacts). Country-year CO2 damages D_CO2(i,t) = Emissions(i,t) × SCC(t). Sensitivity analyses include a higher SCC (e.g., ~$185/tCO2 in 2020$ from Rennert et al.) to evaluate impacts on pollution intensity and EVA. PM2.5 damages: Population-weighted annual average PM2.5 concentrations by country-year are from satellite/model/monitor fusion datasets (Hammer et al., 2020; van Donkelaar et al., 2016). Baseline mortality risks (by cause and age) and population are from GBD 2019. The integrated exposure-response (IER) function maps PM2.5 concentration to relative risk for six causes (stroke, lower respiratory infections, diabetes, COPD, ischemic heart disease, lung cancer) by age group using cause/age-specific parameters and a minimum risk exposure level. Pollution-attributable fractions (PAF) are computed from IER, then multiplied by cause- and age-specific baseline mortality rates and exposed population to estimate excess mortality ΔM(i,t). Monetary damages are ΔM(i,t) × VSL(i,t). VSL parameterization: U.S. EPA VSL is $7.4 million (2006$) as the anchor. Country VSLs are transferred using income elasticities that vary by income level (e.g., ~0.8 for high-income; ~1.2 for low and middle income) and updated over time with country-specific real income growth. Within-country cross-sectional VSL variation is not modeled. Sensitivities include: lower anchor VSL ($2.8M, 2000$), and an alternative elasticity specification with unit elasticity (1% income → 1% VSL change) across all countries. Integration into accounts: Gross External Damages (GED) = PM2.5 damages + CO2 damages. Environmentally Adjusted Value Added (EVA) = GDP − GED. All monetary values are harmonized to constant prices (e.g., VSL in 2006$, SCC in 2007$ converted to 2010$ using CPI inflators). Results are analyzed globally, by World Bank income groups (using 2018 classifications), and for the eight largest economies. The study also estimates Environmental Kuznets Curves by regressing per-capita physical measures (PM2.5 concentration; CO2 tons) and per-capita monetary damages against per-capita income (quadratic functional forms), using all country-year observations. Sensitivity and decomposition: Decomposition analysis separates PM2.5 damage changes into contributions from income-driven VSL changes (holding concentrations fixed) and from concentration changes (holding income/VSL fixed), applied to the eight largest economies. Parametric sensitivity scenarios vary the VSL anchor/elasticity and SCC values to assess robustness of GED/EVA and pollution intensity trends.

- Global trend reversal around the Great Recession: From the late 1990s to 2008, EVA growth exceeded GDP growth (global economy became less pollution-damage intensive). The spread in growth rates peaked near 0.4% around 2000 (≈$200B on $50T GDP). After the Great Recession, GDP growth outpaced EVA, indicating rising pollution damage intensity. - Structural drivers: Post-2008, the global output share of developing economies rose (from ~23% in 1998 to 36% in 2018). Upper-middle income economies (notably China) and lower-middle income economies (including India, Bangladesh, Vietnam) experienced increased pollution damage intensity alongside higher investment (capital formation) shares of GDP. High-income countries’ pollution damage intensity fell (≈10% to 7% of GDP, 1998–2018) amid long-standing air pollution controls and relatively higher consumption shares. - Country specifics among top eight economies: In Western economies, PM2.5 damages declined while CO2 damages rose; U.S. CO2 GED increased by nearly 50% versus smaller increases in Europe (reflecting earlier, stronger CO2 policies in Europe). In China, PM2.5 and CO2 damages rose rapidly until ~2007; PM2.5 damages growth slowed following air quality measures around 2008 and intensified after 2014’s “war on pollution,” while CO2 damages continued to rise. In India, PM2.5 and CO2 damages increased exponentially through 2018, with major national air pollution policy only announced in 2019. - EKC with monetary damages vs physical measures: PM2.5 concentrations peak at very low incomes, but PM2.5 monetary damages peak near ~$45,000 per capita. For CO2, per-capita emissions peak just above ~$60,000, whereas monetary damages peak near ~$80,000. In 2018, only two countries (Luxembourg, Norway) were at or above ~$80,000, implying most countries have not reached peak per-capita pollution damages. Combined PM2.5+CO2 damages peak just under ~$3,000 per capita at incomes just under ~$50,000. - Income-driven WTP dominates PM2.5 damages in China and India: Decomposition shows rising real incomes (hence higher VSL) are the principal drivers of increasing PM2.5 damages; concentration increases play a smaller role (especially in China). In developed countries, declining concentrations offset rising VSL, yielding net PM2.5 damage declines. - Composition of GED: PM2.5 damages comprise the majority of GED across all income groups and especially in high-income countries (≈80–90%), driving EVA dynamics. Whether CO2 eventually dominates depends on future air quality policy stringency/enforcement in China and India and development paths in low-income countries. - Policy and development stages: Three stages emerge—(1) developed, regulated economies with declining pollution intensity; (2) rapidly growing, capital-intensive economies introducing air pollution controls (e.g., China) that temper PM2.5 damages while CO2 damages rise; (3) rapidly growing economies without strong emission constraints (e.g., India) where GED rises exponentially.

The findings show that monetizing environmental damages materially alters assessments of sustainable growth relative to tracking physical indicators alone. Incorporating PM2.5 and CO2 damages into EVA reveals that the global economy’s pollution damage intensity fell before the Great Recession but rose afterward, driven by shifting output shares toward capital-intensive, middle-income economies. Decomposition highlights that in major developing economies, rising incomes (and thus WTP/VSL) are central to increasing PM2.5 damages, a dimension missed by physical-only metrics. The EKC analysis underscores that monetary damages peak at substantially higher income levels than physical emissions/concentrations, implying most countries have not yet reached peak per-capita damages and may experience rising damages as they develop. These results inform policymakers that integrating monetized environmental accounts into NIPAs can better track net growth and guide policy—especially the timing and stringency of air pollution and climate regulations. Early, stringent air quality policies can decouple PM2.5 damages from income growth, while climate policies must broaden and strengthen to address CO2 damages that peak later along the development path.

The paper demonstrates a globally consistent framework to monetize damages from PM2.5 exposure and CO2 emissions for 165 countries (1998–2018) and to integrate them with GDP to form EVA for assessing sustainable growth. Three core contributions emerge: (1) a documented global reversal from declining to rising pollution damage intensity after the Great Recession; (2) evidence that changes in WTP/VSL associated with income growth are a dominant driver of rising PM2.5 damages in China and India; and (3) EKC results showing that monetary damages peak at higher income levels than physical measures, with combined damages peaking below $50,000 per-capita income—levels not yet reached by most countries. Policy implications include the need for earlier and stronger environmental regulations, particularly climate policies, and the adoption of monetized environmental-economic accounts within NIPAs. Future research should expand coverage to additional environmental dimensions (e.g., water pollution, ecosystem services, non-CO2 GHGs), refine SCC and VSL parameterizations with improved data and equity considerations, and extend analyses beyond 2018 with post-pandemic data to continue tracking sustainability using comprehensive, monetized accounts.

- Scope of damages: The accounts are not comprehensive; they omit water pollution, biodiversity and ecosystem services, in situ natural resource values, and most non-CO2 GHGs (except a global methane illustration in SI showing methane <10% of GED and not altering main trends). - Time coverage: Analysis is truncated at 2018 due to limited post-2019 cause- and country-specific mortality data; using pre-COVID mortality rates beyond 2019 would add uncertainty. CO2 damages were extended to 2021 only for exploratory checks. - Valuation assumptions: VSL transfer relies on income elasticities and a U.S. anchor; alternative anchors/elasticities change levels and trends. Applying a uniform high-income VSL globally can yield negative adjusted output for some countries and is not adopted. - SCC uncertainty and uniformity: SCC is applied uniformly across countries and varies only over time; alternative SCC estimates (e.g., higher values) materially increase GED but do not reverse qualitative trends. - PM2.5 damage modeling: Uses IER-based mortality-only valuation; morbidity and productivity losses are excluded (likely smaller and partly captured in GDP). PM2.5 damages are valued based on local concentrations rather than emission origin due to data constraints. - Aggregation and EKC estimation: Fewer observations at very high incomes increase uncertainty in EKC peaks; combining physical measures across pollutants is infeasible due to unit differences, hence reliance on monetization. - Data and parameter uncertainties: Satellite-derived PM2.5, baseline mortality rates, population data, CPI adjustments, and investment/consumption share indicators carry measurement uncertainty that can affect levels but are unlikely to overturn qualitative patterns.