Economic Growth without CO2 Emissions?

by Nikos Tsafos

May 2017

Introduction

We need energy to survive—to light, heat and cool our homes and offices; to grow, preserve and cook food; to move people and goods across continents and oceans; to erect structures and build things; to communicate with one another; to play; to live. But what, exactly, is the relationship between incomes and energy consumption? Can we grow an economy without consuming more energy? More importantly, can we keep raising our living standards without emitting more CO2 into the atmosphere?

Income, energy and CO2 emissions: A global view

Since 1990, world gross domestic product has almost doubled (+93%); energy consumption has increased by 56%, and CO2 emissions from fuel combustion have increased by 58%. In short, incomes have grown faster than energy consumption and CO2 emissions, but rising incomes have still meant rising energy use and emissions.

On a per capita basis, the numbers are a bit different: income is up ~40% since 1990, but energy consumption is just 14% higher—the average person on earth consumed only 14% more energy in 2014 than in 1990. In some respects, that is remarkable given the dramatic advances in living standards across the world and the proliferation of new, energy-consuming machines (like computers and smartphones). Even so, the big picture is hardly reassuring that growth is possible without additional energy use and CO2 emissions.

World population, income, energy consumption and CO2 emissions (1990=100)
Variable 1971 1980 1990 2000 2010 2014
Population 71.2 84.0 100 115.7 131.0 137.3
Gross domestic product 53.0 74.4 100 131.3 173.9 193.2
Gross domestic product per capita 74.4 88.6 100 113.4 132.8 140.7
Energy consumption 63.0 82.1 100 114.4 147.6 156.2
Energy consumption per capita 88.4 97.7 100 98.9 112.7 113.7
CO2 emissions 68.0 86.4 100 112.9 148.5 157.9
CO2 emissions per capita 95.5 102.8 100 97.5 113.4 115.0

Source: International Energy Agency, CO2 Emissions from Fuel Combustion (2016 Edition). GDP is measured in constant 2005 dollars ($2005), energy consumption in tons of oil equivalent (toe), and CO2 emissions are from fuel combustion only and are measured in tons of CO2 equivalent.

Income vs. CO2 emissions: A static view

Richer countries have higher CO2 emissions—but the relationship is not linear. Countries at similar income levels can have vastly different CO2 footprints. Take Switzerland and Qatar: similar incomes (~$75k per capita), but Qatar emits almost 8x more CO2 per capita than Switzerland. Or Portugal and Bahrain: similar incomes (~$22k per capita), but Bahrain emits almost 4x more CO2 per capita than Portugal. Or Panama and Curacao: similar incomes (~$10-12k per capita), but Curacao emits 11x more CO2 per capita than Panama. Or Switzerland and Portugal: Switzerland emits 12% more CO2, but its income is 3.5x higher. Or Luxembourg and Trinidad and Tobago: similar CO2 emissions (~16 tons per capita) but incomes are 6.4x higher in Luxembourg. Income does not predict CO2 emissions.

Source: International Energy Agency, CO2 Emissions from Fuel Combustion (2016 Edition). GDP is measured in constant 2005 dollars ($2005) and CO2 emissions are from fuel combustion only and are measured in tons of CO2 equivalent.

Why not? First, countries use different amounts of energy. Their economies may be different: one might rely on heavy industry, another on services. Or their climate may be different: cold countries use energy for heat, warmer countries use energy for cooling. Or their energy prices may be different: some produce energy, others import it; some tax energy, others subsidize it. Or their capital stock may be different: machines may be newer or the buildings may be older. Or their policies may be different: governments may have energy efficiency mandates or they may not. These differences affect energy use. Malta and Bahrain have similar incomes (~$20k per capita), but Bahrain consumes 5x more energy per capita as Malta. Or Uruguay and Trinidad and Tobago: similar incomes ($14-16K per capita), but Trinidad consumes 9x more energy per capita than Uruguay. Income, alone, is a poor predictor of energy consumption.

Source: International Energy Agency, CO2 Emissions from Fuel Combustion (2016 Edition). GDP is measured in constant 2005 dollars ($2005), energy consumption in tons of oil equivalent (toe), and CO2 emissions are from fuel combustion only and are measured in tons of CO2 equivalent.

Second, countries generate energy in different ways. Fossil fuels such as coal, oil and natural gas emit more CO2 than hydro, nuclear power, wind or solar. Qatar and Iceland consume similar amounts of energy (20 vs. 18 tons of oil equivalent, or toe, per capita), but CO2 emissions are 5x higher in Qatar. Qatar uses oil and gas; Iceland uses hydro and geothermal energy. Australia consumes slightly (7%) more energy per capita than Sweden, but its emissions are over 4x higher. Germany's emissions are 2x higher than France's, even though the two countries consume similar amounts of energy. Not even energy consumption, on its own, predicts CO2 emissions.

Income vs. CO2 emissions: A dynamic view

So far, we have looked only at 2014 data. Maybe countries end up at different places. But can an economy grow without more energy? Or without more CO2 emissions? The evidence is clearly yes, but it is not a given. There are many different paths—in fact, eight distinct paths:

  1. Turkey: GDP, energy consumption and CO2 emissions are closely linked. Not perfectly overlapping, but close.
  2. Australia: Income at some point delinks from energy use and CO2 emissions.
  3. Russia: GDP grows but energy use and emissions do not, largely because of continued efficiency gains.
  4. France: CO2 emissions fall faster than energy use, usually due to increased nuclear power (but also gas for other countries).
  5. United Kingdom: No clear link between GDP, energy and CO2 emissions: incomes grow faster than energy or emissions.
  6. China: Income, energy use and emissions grow, but unequally: incomes are up 10x vs. 1990, energy use 3x, and CO2 emissions 4x.
  7. Iran: Income grows slower than energy use and CO2 emissions—energy-hungry development without commensurate growth.
  8. Bangladesh: Growth in CO2 outpaces energy use or incomes, as countries replace traditional (e.g. biomass) with fossil fuels.

GDP, energy consumption and CO2 emissions

Source: International Energy Agency, CO2 Emissions from Fuel Combustion (2016 Edition). GDP is measured in constant 2005 dollars ($2005), energy consumption in tons of oil equivalent (toe), and CO2 emissions are from fuel combustion only and are measured in tons of CO2 equivalent.

In sum: there is no reason to think that economic growth equals more energy consumption. Or that more energy consumption must equal more CO2 emissions. Or that a pattern we have observed in the past will continue in the future. Consider the following data points. In the United States home energy use per household has declined by 27% since 1972—Americans have gotten bigger houses and more gadgets while consuming 27% less energy than they did 45 years ago. In the United Kingdom, the chemical industry used 84% less energy per unit produced relative to what it did in 1970. In fact, in 2014, the United Kingdom consumed less energy that at any point in the past 50 years—not on a per capita basis, but on an absolute basis.

There is really no reason to think growth cannot occur alongside declines in energy consumption and emissions. We have so many different trajectories around the world and throughout history (you can explore individual countries in the appendix below).

Living standards rising; CO2 emissions, less so

Let's step back and look at the big picture. Most countries continue to grow: in 101 countries, real GDP per capita was either at its highest level or just below its peak (peak occurred in 2008-2013). There are also a few countries where per capita GDP peaked before 1990. That's because of war, revolution or economic stagnation, the last often linked to mismanagement of natural resources. Finally, there are countries where per capita incomes have stagnated, peaking after 1990 but before 2007 (21 countries), many of them in Europe.

Year of peak GDP per capita

Source: International Energy Agency, CO2 Emissions from Fuel Combustion (2016 Edition). GDP is measured in constant 2005 dollars ($2005).

Per capita CO2 emissions show a dramatically different pattern. In 30 countries, they peaked before 1990—mostly in Europe, North America and parts of Africa (largely due to war). In the Former Soviet Union, energy consumption and CO2 emissions have been declining since 1990, often at rapid rates. Per capita emissions are rising mostly in Asia and Latin America, as well as parts of Africa (36 countries in total). In all other places, emissions peaked either recently (25 countries in 2008-2013), or a long time ago.

Year of peak CO2 emissions per capita

Source: International Energy Agency, CO2 Emissions from Fuel Combustion (2016 Edition). CO2 emissions are from fuel combustion only and are measured in tons of CO2 equivalent.

What does all this mean? We know that the link between incomes and CO2 emissions is far from linear. We have so many countries that have grown without using more energy and/or emitting more CO2. Not all countries are equal, of course: the growth in China and India can overwhelm what happens in many other countries. Yet the broad picture is clear: there are so many countries where GDP per capita is growing while emissions are not. There is no reason to think there is a fundamental trade-off between economic growth and reducing CO2 emissions.

Appendix

GDP, energy consumption, CO2 emissions

  • Gross domestic product
  • Energy use
  • CO2 emissions

Source: International Energy Agency, CO2 Emissions from Fuel Combustion (2016 Edition)

Notes

All data from the International Energy Agency, CO2 Emissions from Fuel Combustion (2016 Edition). GDP is measured in constant 2005 dollars ($2005), energy consumption in tons of oil equivalent (toe), and CO2 emissions are from fuel combustion only and are measured in tons of CO2 equivalent. For more information, please visit my GitHub repository.