Size distribution of national CO2 emissions

- Examine size distributions of CO2 emissions.
- The data cover over 200 countries/territories for the period 2000-2010.
- Composites of Pareto and lognormal distributions provide a good fit for CO2 emissions.
- The results show that large countries' CO2 emissions follow Zipf's law.

We examine the size distribution of national carbon dioxide (CO2) emissions on a sample of 210 countries and territories for the period 2000-2010. We employ lognormal, double Pareto-lognormal, lognormal-upper tail Pareto, and Pareto tails-lognormal distributions to estimate CO2 size distribution. The analysis demonstrates that the lognormal-Pareto composite distributions generally fit the size distribution of CO2 emissions better than the lognormal distribution. The parametric analysis reveals that the upper-tail of CO2 emissions is characterized by Zipf's law. The power law in the upper-tail implies that large countries emit much of the CO2. When controlling for population, Zipf's law no longer holds, but per capita CO2 emissions still exhibit a strong Pareto upper tail.