I love freakonomics. The Steven and Stephan are awesome people and have a great book and podcast and NY times article. But do remember they look at correlation. Correlation does not equal causality. It's damn entertaining though and does hint enough at possibles causes of things.
For those who don't know correlation, it can be for example, "more people in North Carolina get lung cancer than any other state." One might think that people moving to NC would have a high risk of lung cancer just living there. But maybe the real reason is the social aspect that perhaps more people who live in NC, smoke-- and smoking causes lung cancer. The correlation is they live in NC. The causation is smoking causes lung cancer and more people smoke there than most states.
So like the post above, maybe the more people die earlier when they retire because most people sit around? Maybe people who retire early have depression due to not having the tight social network they had at work or a goal in life. Or maybe many people actually retire early due to a low energy because they have an underlying disease?
It is important to know this for sure and what is really important is to have a plan when retiring. If the plan is to get fat, go on cruises and get drunk every night, maybe retirement is not such a good thing because it might actually make one's life expectancy decrease.
If the plan is to be energetic, do fun things, help people, be inspired, it might actually increase one's longevity since there is evidence that lower stress is healthy.