The "fossil fuels cause 1 in 5 deaths" myth

By Alex Epstein

The widely-publicized claim that fossil fuels cause 1 in 5 deaths is the worst kind of pseudoscience. It ignores fossil fuels' life-extending benefits and wildly overstates their negative side-effects. In reality, fossil fuels lengthen 5 out of 5 lives.

Quick summary

  1. A recent study by Harvard and UK universities is being circulated as proof that fossil fuels are responsible for 1 in 5 deaths--over 8 million deaths annually.

    And yet the places where fossil fuel use has grown most in recent decades have had skyrocketing life expectancies.1

  2. Since 1980, India's fossil fuel use has increased by 700% and China's has increased by 600%. Did this lead to many more premature deaths given "fossil fuels cause 1 in 5 deaths"? No, India's life expectancy increased by almost 16 years and China's increased by almost 10!2
  3. The strong correlation between increasing fossil fuel use and increasing life expectancy is not coincidental--it is causal. Billions of people have brought themselves out of poverty by using uniquely cost-effective fossil fuels to power factories, farms, vehicles, and appliances.
  4. The developing world overwhelmingly uses fossil fuels because that is by far the lowest-cost way for them to get reliable energy. Unreliable solar and wind can’t come close. That’s why China and India have hundreds of new coal plants under construction.3
  5. Any study about fossil fuels' side-effects must prominently mention their life-extending benefits. But modern studies, funded by governments with anti-fossil fuel agendas, systematically ignore fossil fuels' benefits and wildly overstate side-effects.
  6. In my book The Moral Case for Fossil Fuels, I discuss the "false attribution" and "no threshold" fallacies used to wildly overstate the side-effects of fossil fuels. Both are used in the recent Harvard "study"--which should properly be called a "speculation."
  7. If the fossil fuel byproduct PM2.5 were as harmful as the Harvard study's model speculation claims, the subway systems in the NY and NJ would be death traps--as they have PM2.5 concentrations often many times above the “healthy” limits according to the EPA.4
  8. As evidence of the agenda typical in government-funded pseudoscience, note that the authors of the Harvard "scientific" study make sure to advocate "current efforts to cut greenhouse gas emissions and invest in alternative sources of energy"--i.e., their political views.5
  9. Two good tests for any "study" about fossil fuels' side-effects: 1) Does it acknowledge the irreplaceable, life-extending benefits of fossil fuels? 2) Does it clearly differentiate between what is demonstrated science and what is wild speculation? Most "studies" fail both tests.
  10. Fossil fuels are powering the development that is making the world a better and better place to live for billions of people--a place where climate-related deaths are at record lows. Fossil fuels lengthen 5 of 5 lives. That is what the real science says.6


  1. Vohra et al (2021) - Global mortality from outdoor fine particle pollution generated by fossil fuel combustion: Results from GEOS-Chem

  2. BP Statistical Review of World Energy 2020
    World Bank Data - Life expectancy at birth, total (years) - India, China)

  3. Combined, China and India have 284 new coal plants (over 188 GW of capacity) under construction or in the planning phase as of January 2021.
    Global Coal Plant Tracker by Global Energy Monitor

  4. - New York City/New Jersey subway PM2.5 levels 77 times greater than EPA standards; No bodies found
    Luglio et al. (2021) - PM2.5 Concentration and Composition in Subway Systems in the Northeastern United States
    U.S. Environmental Protection Agency - NAAQS Table

  5. Vohra et al (2021) - Global mortality from outdoor fine particle pollution generated by fossil fuel combustion: Results from GEOS-Chem

  6. For every million people on earth, annual deaths from climate-related causes (extreme temperature, drought, flood, storms, wildfires) declined 98%--from an average of 247 per year during the 1920s to 2.5 in per year during the 2010s.

    Data on disaster deaths come from EM-DAT, CRED / UCLouvain, Brussels, Belgium – (D. Guha-Sapir).

    Population estimates for the 1920s from the Maddison Database 2010 come from the Groningen Growth and Development Centre, Faculty of Economics and Business at University of Groningen. For years not shown population is assumed to have grown at a steady rate.

    Population estimates for the 2010s come from World Bank Data.