- If we truly believe smokers are rational and fully informed, graphic label warnings are pointless.
- If smokers are truly addicted, graphic warning labels cannot be expected to make them quit.
- Caution is in order whenever we are tempted to devise policies aimed at protecting adults from themselves.
A recent FDA regulation on graphic warning labels for cigarettes has resurrected a huge debate among health economists over whether adult smoking is a rational choice or merely the consequence of an addiction. The vast majority of smokers begin smoking before age 18, so it is argued that as adults, such smokers smoke principally to avoid the unpleasant effects of withdrawal rather than because the pleasures of smoking outweighs its very real and well-known risks to life and health. Nine health economists have written a cogent summary of the various strands of evidence in support of this view (pp. 10-11).
I concur with these economists (and my Duke colleague Don Taylor evidently agrees) that the truth is likely between these two extreme views. Interestingly, if we truly believe smokers are rational and fully informed, graphic label warnings are pointless since they will offer smokers no information they don’t already know. Conversely, if smokers are truly addicted, graphic warning labels cannot be expected to make them quit.
A book I co-authored with Don Taylor and other Duke colleagues, The Price of Smoking, found that that the total private and social costs of smoking amounted to $40 per pack (year 2000 dollars). But $33 of this total represent private losses to smokers themselves, primarily in the form of lost life expectancy. As a rough approximations, smokers lose 1 minute of life expectancy for every minute spent smoking. The addiction model essentially says that smokers irrationally impose massive future costs on themselves: “their decision to continue to smoke are time inconsistent, satisfying their short‐run desire for immediate gratification rather than their long‐run desire for good health, then later regretting these decisions” (p. 12).
To a non-smoker, it may seem bizarre that any smoker could conceivably get pleasure equivalent to $33 a pack, i.e. $1.65 per cigarette, from smoking. But we offered two thought experiments to help clarify this smoking decision (p. 248):
Thought Experiment #1: We state in chapter 1 that each pack of cigarettes costs two hours of life expectancy. Since each cigarette consumed probably takes about six minutes to smoke, this is equivalent to saying that smokers give up one minute of life for every minute spent smoking. This does not even take into account discounting: were we to do so properly, smokers probably give up 10-20 seconds of life per minute spent smoking. But leaving discounting aside, this formulation implies that smokers enjoy each minute spent smoking twice as much as each minute spent not smoking. A priori, this seems plausible on its face. That is, we presumably all can imagine certain activities so pleasurable that it might be worth trading some quantity of life to achieve higher quality of life. It becomes even more plausible once discounting is taken into account since it effectively implies that for smokers, a life in which smoking was denied might be worth only two-thirds to five-sixths as valuable as a life in which smoking was permitted.
Thought Experiment #2: The average smoker evidently smokes about 400 packs a year, which implies giving up 800 hours of future life expectancy, or a little more than two hours a day. A minimum of one-sixth and possibly upwards of one-third or more of adults watch two or more hours of TV a day.
Suppose we were doing a study on the private and social costs of TV watching. Leaving aside the multitaskers who might do other thing while the TV is on, assume that each hour of TV watching is completely unproductive  and therefore the cumulative loss of time to such activity can be monetized as if one had literally lost that many hours/years of life as a consequence. Thus, if we were to monetize the private costs of TV watching, we would derive aggregate figures very similar to smoking, that is, measured in hundreds of billions of dollars, and if we were to measure costs over the life cycle we would get figures much higher than for smoking, since people don’t die prematurely from watching TV and since we treat each hour watched as if the person had “died” for that interval; hence we would be accumulating hours lost even at age 18, and so on. So suppose we end up with an average lifetime figure of $200,000 attributable to watching TV. Would we compare that to the cost of a house and express bafflement that anyone could possibly give up so much of their precious lifetime in pursuit of this activity? The point is, we could take any lifetime activity that consumed about two hours daily–be it eating, reading, or any other hobby/activity–and it would have roughly the same social cost as smoking, that is, of eye-popping size, but does that necessarily mean that we have to conclude that humans are irrational or addicted to engage in them?
Let me be clear: I myself have never smoked . I’m not arguing that taking up smoking is a good idea. But there’s plenty of activities I would never consider either because they’re too risky (e.g., 1 in 61 Everest climbers dies, but the risk of death for someone my age is 25%!) or simply not my cup of tea. A 30-year-old Everest climber with 50 years of life expectancy faces a 2.2% risk of death meaning that they are giving up 1.1 years of life expectancy to make that climb. Since it typically takes about 30 days to complete an Everest climb, this implies such climbers are giving up 11 days of life for each day spent climbing Everest. Should we prohibit such individuals from engaging in such apparently reckless behavior? Should we levy a stiff climbing tax or expose them to graphic warning labels so that they better appreciate the risks they are incurring?
If we measure risk in terms of minutes of lost life expectancy per minute of activity, there’s other very routine activities that many Americans engage in that are considerably more risky than smoking, including drinking milk and eating peanut butter . Are these “risky” activities next in line for FDA regulation? I sure hope not.
The hallmark of a free society is that people get to make their own mistakes and learn from that experience. There’s a role for government to play to protect those who cannot protect themselves. So policies that discourage those under 18 from taking up smoking surely have some merit. But I think a considerable degree of caution is in order whenever we are tempted to devise policies aimed at protecting people from themselves. That kind of paternalism doesn’t square with the vision of our Founders and it doesn’t bode well for peaceful relations among the adult citizenry (nearly one-fifth of whom smoke). Should we be certain that smokers don’t foul the air of non-smokers? Sure. But we should think twice about tut-tutting those who choose to engage (in the privacy of their own homes or outdoors where it does no one harm) in a habit that gives them pleasure and may be substantially less risky than other adult activities such as climbing that for some reason have never invited the public opprobrium that smoking has. The world just might be a better place if we all minded our own business.
 Discounting recognizes that time and money in the future are not as valuable as today, in part because people cannot be certain they will even live long enough to make use of that time and money. Because of the time value of money, people implicitly discount the future (some much more than others) meaning that a dollar 10 years from now is worth less than having a dollar today. If I am willing to lend you $100 today if and only if you pay me 5% annual interest, that means that $100 today is equivalent to $163 in the future ($100 x 1.05^10); conversely $100 in the future is worth only $61 today ($100 / (1.05^10)). This same concept applies to people’s valuation of their time. Thus, a 24-year-old who smokes 1 cigarette today is giving up 6 minutes of life expectancy 50 years from now. Assuming a 6% discount rate, those 6 minutes are only worth a half minute today (6 / (1.05^50)).
 “Completely unproductive” here means that it is done purely for its own pleasure, not as an investment in knowledge that perhaps might have a future pay-off in job performance or the quality of one’s social interactions.
 Well, there was that time at age 5 when my big sister and I tried to smoke one of Granny’s cigarettes, but my point is that I never became a “smoker.”
 Bernard Cohen shows that both drinking 1 pint of milk a day and eating 1 tablespoon of peanut butter are associated with 1 day of lost life expectancy (as with the 1 minute of life expectancy lost per minute of smoking, these figures are undiscounted). To calculate the figures in my chart, I generously assumed people drink 2 glasses of milk a day and that each takes 1 minute to consume. I have assumed it takes 1 minute to consume a tablespoon of peanut butter. Readers can adjust my figures accordingly if they think I have ascribed too much or too little time to these activities. Note that these elevated risks are related to carcinogens that occur naturally in these foods.