One reason adult smokers don't quit is that they simply place lower value on their health and longevity than do non-smokers. So say Khwaja, Sloan and Wang in the latest Journal of Law and Economics.
Khwaja and coauthors run some stated-preference surveys on smokers. I'm pretty skeptical of contingent value methodology in general, but this seems to be one of the more carefully conducted ones. After a presentation on the effects of chronic obstructive pulmonary disease, survey respondents were asked to choose between different living locations which involved tradeoffs between cost of living and risk of contracting chronic obstructive pulmonary disease; they then were asked whether they'd be willing to have an operation to cure said disease where the surgery carried a known risk of death. They found that value of life or cost of disease measures taken by these methods varied systematically across smokers and non-smokers. So long as any problems in using this methodology can wash out as a fixed effect across cohorts, the difference across cohorts is useful.
Where advocates of high tobacco taxation often allege hyperbolic discounting by smokers, with taxes then helping smokers to solve self-control issues, this study shows that observed smoking behaviour can be explained without recourse to hyperbolic discounting. Smokers simply don't value their health as much as the anti-tobbac lobbies might wish them to.
I'd had a bit of fun a couple of years ago with the New Zealand anti-tobacco folks over in the New Zealand Medical Journal. Cost-benefit analyses written on behalf of these groups make great fodder when lecturing on things to watch for in a motivated cost-benefit analysis.