Growth, sustainability, and the measurement of global gross product

Aniruddh Mohan, Akshay Thyagarajan, Nicholas Muller 31 July 2020



The relationship between economic development and environmental impact is at the forefront of policy debates in both economics and environmental science. There is a considerable amount of research that explores this nexus. Often, this relationship is framed in terms of an Environmental Kuznets Curve or EKC (Grossman and Krueger 1995, Barbier 1997, Andreoni and Levinson 2001, Harbaugh et al. 2002, Brock and Taylor 2010). The EKC posits that, at low levels of per capita income, societies incur little environmental degradation because consumption of natural resources and material goods are at or near subsistence levels. Then, as income rises, nations consume more, imposing greater cost on local, regional, and global environmental systems. The EKC framework then predicts that at some income level, the efficiency of production improves, institutions strengthen, and environmental impacts abate. For several pollutants, especially those discharged into the air, strong empirical support exists for an EKC (Grossman and Krueger 1995, Brock and Taylor 2010).

This in turn relates strongly to current debates about the sustainability of global economic development. Specifically, in the literature one finds definitions of sustainability as falling emissions of pollution coupled with positive rates of real, per capita growth. The empirical support for an EKC suggests that societies become sustainable as rising per capita incomes lead to eventual mitigation of environmental harms.

Our recent research contests the definition of sustainability based on physical emissions (Mohan et al. 2020). Together with our co-authors, we argue that debates about sustainability should be recast from a focus on physical tonnage of emissions, to a measure of monetary damage. In short, what matters to society is not the amount of pollution emitted, but the adverse effects of such pollution. Our approach finds its conceptual foundation in a decades-old literature on environmental accounting (Nordhaus and Tobin 1973, Bartelmus 2009, Muller et al. 2011). Essentially, pollution impacts are an unpriced cost associated with production and consumption of market goods and services. Though some environmental policies exist in many countries, what we monetise are damages from remaining emissions, given current policy requirements. Using the literature as our guide, we deduct environmental pollution damage from the national income and product accounts (measures like gross domestic product, or GDP) to estimate environmentally adjusted value added, or EVA (Abraham and Mackie 2006, Nordhaus 2006, Muller 2014). This net measure more accurately depicts national welfare than a measure like GDP.1 There are also even more recent papers that argue for expanded conceptualisations of the national accounts that include measures of sustainability and well-being (Balasubramanian 2019, Dunga et al. 2019).

This synthesis of pollution damage and GDP facilitates two novel exercises found in our recent study. First, we estimate an EKC for pollution damage. Recall that this stands in sharp contrast to prior work, which tests for an EKC in terms of physical emissions. Second, we tabulate annual growth rates for 163 countries around the world from 1998 to 2018 using officially reported GDP and our proposed measure, EVA. We first discuss our results regarding the damage-based EKC before proceeding to the global growth exercises.

Growth, pollution, and the environmental Kuznets curve

Our paper uses a model of economic growth that dates back to the seminal work of Nobel Laureate Robert Solow (Solow 1956). Solow (1956) used this model to demonstrate the importance of what he referred to as balanced, or equilibrium, growth. This occurs when income per capita ceases to grow in real terms – when savings exactly offset capital depreciation.

We use this model to demonstrate an important divergence between an EKC based on emissions and our definition based on damage. Take the case of the pollutant CO2. The damage from each additional ton of CO2 emitted (marginal damage) increases as the global stock of CO2 increases. Therefore, even as CO2 emissions fall, CO2 damages may be increasing due to the increase in the marginal damage of each ton emitted. The damage-based EKC will therefore peak at a higher income level than the emission-based EKC. Why does this matter? If emissions fall as a country approaches its balanced growth path, its economy may appear to be growing sustainably, if policymakers focus on emissions. In contrast, if damages are still rising in balanced growth, then damages could continue to rise without limit. We argue that this cannot satisfy a reasonable definition of sustainability. Therefore, policymakers should focus on damages, not emissions.

Is this merely an academic concern? We use the case of China, perhaps top-of-mind when it comes to concerns about sustainable development, to demonstrate the importance of this distinction. In China, average levels of ambient fine particulate matter (PM2.5) peaked in 2011. Emissions of CO2 have slowed considerably. From these data, one might reasonably conclude that China is approaching sustainable growth. Yet, monetary damages from these two pollutants in China have increased 50% since 2011. Over this period, damages averaged an annual growth rate of 5%. At this rate, real per capita damages double in 14 years. This cannot be sustainable. These comparisons for the world's second largest economy highlight the importance of an assessment of sustainability based on monetary damage rather than physical units of pollution.

For all countries (see Figure 1), we find that CO2 emissions reach the maximum level of 15 tons per capita at an income level of about $70,000. Monetary damages, in contrast, rise until $130,000 per capita – a level which most countries have not attained to date.

Figure 1 Environmental Kuznets curves for emissions and monetary damages

Notes: This figure displays quadratic fit for the EKC curve showing emissions peaking earlier than damages. The blue line in the plot is CO2 emissions/capita. The red line is Gross External Damage (GED)/capita as measured on the secondary Y axis on the right.

Global growth accounting

Our second set of empirical exercises tabulate growth in officially measured GDP and EVA, which is GDP less pollution damage. We cover 163 countries over 21 years to provide a spatially comprehensive assessment of global environmental performance. Globally, EVA exceeded GDP growth from 1998 to the Great Recession. After 2010, EVA and GDP have grown at essentially the same rate.

Drilling down further (see Figure 2), high income countries exhibit rates of EVA growth that exceed GDP growth, though the difference has attenuated in recent years. In Western Europe, the differences were as much as 1.0%, annually, in the late 1990s. In the US, the difference was about 0.3% at that time. This somewhat counterintuitive result occurs because pollution damages fell, while GDP grew, in most high-income countries over the last 20 years. Stricter environmental rules, movements away from coal-based power generation, and off-shoring of heavy manufacturing likely each contribute to this trend.

In contrast, EVA growth fell short of GDP among upper-middle income countries. This result largely hinges on China. From the year 2000 until 2011, EVA growth was as much as 0.5% slower than GDP. While real GDP per capita increased more than five-fold from 1998 to 2018, pollution damage increased by a factor of nearly seven in real terms. This rapid increase in places a drag on growth when measured using EVA. It escapes metrics like GDP.

Figure 2 Environmentally adjusted value-added GDP growth rates

Notes: This figure shows 5 year rolling averages of EVA-GDP Growth Rates for the world's eight largest economies by GDP (1999-2018)

India is also an important case to consider. Since the Indian economy began to grow rapidly in 2005, EVA growth fell short of GDP by as much as 0.5%, annually. Over the full 21-year period, real GDP per capita has doubled. Pollution damages are five times higher. Ambient PM2.5 and CO2 emissions increased without bound (through 2018). India appears poised to repeat the Chinese experience.

Finally, the analysis shows that very low-income countries have exhibited very little change in the pollution damage intensity of their GDP. As a result, EVA growth for such nations closely tracks officially reported GDP.

The relationship between economic growth and environmental damage is a crucial determinant of national and global sustainability. For years, this discussion relied on measures of physical emissions as the key metric of sustainability. However, recent developments in the natural and social sciences facilitate estimation of monetary pollution damage, as well as the inclusion of such damages into official estimates of national product to produce more welfare-based measures of economic growth. Our research demonstrates the potential for this perspective to significantly influence policymakers’ conclusions about whether the global economy is growing sustainably.


Abraham, K G and C Mackie (2006), "A framework for nonmarket accounting", in D Jorgenson, J S Landefeld, and W D Nordhaus (eds.), A new architecture for the US national accounts, University of Chicago Press: 161-192.

Andreoni, J and A Levinson (2001), "The simple analytics of the environmental Kuznets curve", Journal of public economics 80(2): 269-286.

Balasubramanian, S (2019), “Wellbeing measurements, Easterlin’s paradox and new growth models: A perspective through gross national happiness”,, 17 February.

Barbier, E B (1997), "Introduction to the environmental Kuznets curve special issue", Environment and Development Economics: 369-381.

Bartelmus, P (2009), "The cost of natural capital consumption: Accounting for a sustainable world economy", Ecological Economics 68(6): 1850-1857.

Brock, W A and M S Taylor (2010), "The green Solow model", Journal of Economic Growth 15(2): 127-153.

Dunga, Y, N Hardie, S Kelly and J Lawson (2019), “Social capitalism: Incorporating sustainability factors into macroeconomic analysis”,, 25 March.

Grossman, G M and A B Krueger (1995), "Economic growth and the environment", The Quarterly Journal of Economics 110(2): 353-377.

Harbaugh, W T, A Levinson and D Molloy Wilson (2002), "Reexamining the empirical evidence for an environmental Kuznets curve", Review of Economics and Statistics 84(3): 541-551.

Mohan, A, N Z Muller, A Thyagarajan, R V Martin, M S Hammer and A van Donkelaar (2020), “The Growth of Nations Revisited: Global Environmental Accounting from 1998 to 2018”, NBER Working Paper No. w27398.

Muller, N Z, R Mendelsohn and W Nordhaus (2011), "Environmental accounting for pollution in the United States economy", American Economic Review 101(5): 1649-75.

Muller, N Z (2014), "Boosting GDP growth by accounting for the environment", Science 345(6199): 873-874.

Nordhaus, W D and J Tobin (1973), "Is growth obsolete?", in M Moss (ed.), The measurement of economic and social performance, NBER Book Series Studies in Income and Wealth, 509-564.

Nordhaus, W D (2006), "Principles of national accounting for nonmarket accounts", in D Jorgenson, J S Landefeld, and W D Nordhaus (eds.), A new architecture for the US national accounts, University of Chicago Press. 143-160.

Solow, R M (1956), "A contribution to the theory of economic growth", The Quarterly Journal of Economics 70(1): 65-94.


1 Crucially, other measures of environmental services along with the value of leisure time, home production, and other non-market activities would be required to measure national welfare.



Topics:  Development Environment

Tags:  Sustainability, environmental Kuznets curve, emissions, growth accounting

PhD student, Carnegie Mellon University

Energy Industry Analyst, Federal Energy Regulatory Commission

Lester and Judith Lave Associate Professor of Economics, Public Policy and Engineering, Carnegie Mellon University


CEPR Policy Research