Statistics can be used to 'prove' virtually anything. There are liars, damned liars, and statisticians.
I am not arguing that smoking, or being fat are necessarily good for you, just that at some point, insurance companies can find an excuse to not insure anyone, or at least raise their rates--especially if enough 'studies' are conducted.
The time honored method has been to insure most, risk assess the members of the pool and charge accordingly, and to give premium discounts to those who are less of a risk.
I think anyone who feels obesity should be 'punished' should dig up the BMI numbers and check where they are. When I was 18, I was 6'1" and weighed 240, but there was no way anyone who looked at me would have considered me 'fat'. The BMI chart (ancestor of today's formula) would have had me morbidly obese.
You're not trying to argue that people who smoke don't get lung cancer at higher rates than non-smokers are you?