Thank you again for this answer.

What I really don't understand here is the transition between :

1 - You can read Kaya's identity as per capita CO2 emissions from human sources is equal to per capita GDP times energy use per GDP unit times CO2 emissions per energy unit production.

and

2 - So, you have three ways of reducing per capita CO2 emissions...

What is the logic behind this ? I mean, if you pose [ CO2 = CO2/pop * pop ] (simplified, but indeed always true by definition), how can you say that one way of reducing the CO2 emissions is reducing the [pop] term ?

I don't know if I'm clear enough...