Do smokers think of those around them?

We’ve written a lot on this blog about taxing cigarettes. The usual arguments focus on health costs and potential ‘internalities’. When you look at the calculated costs of these things you usually get a number much smaller than the tax rate on a pack of cigarettes. Today it was suggested to me that cigarettes are nonetheless undertaxed in NZ. That’s because those calculations don’t include the statistical value of lives lost from smoking. If the value that people place on a life were taken into account at around $4m/life the cost of smoking would be far higher than the tax imposed. It would then take into account the harm to all those around the smoker who would be devastated at the loss of their life.

I haven’t had a chance to think it over in detail so I’m interested to know what you think about the argument. While I don’t have any numbers to back anything up, my initial thoughts are:

  • Smokers’ lives aren’t usually all that much shorter than anyone else’s. It’s not like car accidents which claim people of all ages. That makes it difficult to apply the same number, unless we have an age adjusted figure available.
  • Smokers do care about those around them so they’ll take into account the feelings of others when they make the decision to smoke. Some of the cost to those around them is thus internalised.
  • People close to the smoker have an opportunity to directly bargain with them over their cigarette consumption. Perhaps government intervention isn’t needed to solve this problem.
  • VSL calculations are usually done using willingness to pay to avoid harm. That approach puts a number on how much someone values their own life. Does that $4m figure represent the valuation of one’s own or someone else’s life? I imagine willingness-to-pay to avert the death of others is significantly lower. In fact, given people’s willingness to pay to save the lives of blameless and starving African children, I imagine they aren’t going to pay a whole lot to save someone with a pack-a-day habit.

I don’t know how big this number is, but it does sound kinda significant even given my reservations. So why isn’t it included in the calculations of people like Viscusi, Gruber and Koszegi? I’m sure there’s an obvious answer, but I don’t have it so I’m hoping your collective wisdom can help me out here 🙂

  • Pingback: Aussie Internet Dating Sites Examine Emergence in Cyber-dating | Uk Swingers()

  • It’s an internality: the folks who care about the person are linked to him via explicit or implicit contracts: marriage, family, co-workers, etc.

    All of the VSL measures are just on how much you value your own life: the numbers are backed out from how much folks are willing to pay to avoid small risks to themselves. I’m not sure I’ve ever seen a number on how much a person is valued by others. To the extent that the smoker cares about losses to those around him should he die, he can always pick up life insurance.

    Then there’d be the tricky methodological point: if we start admitting these kinds of things as policy-relevant externalities, what do we do in cases where the willingness-to-pay weighted sum argues that the person should be killed or forced to smoke lots more to induce early death? Now, I’m a pacifist so I’d deplore it, but I can certainly imagine a net positive externality generated from an early death for folks like Chris Kahui. What’s your policy conclusion in that case?

    Instead, I’d argue as follows:
    1. The folks with the strongest interest are linked through contracts such that it’s an internality, not an externality
    2. The folks not so-linked may well have a notional willingness to pay but don’t seem to exercise it: it’s then only potentially Pareto relevant rather than actually Pareto relevant (Buchanan & Stubblebine, 1962, which remains the single best treatment of externality, ever.)

  • Thanks for your comment, Eric. It all became a little clearer to me once I’d left the office last night and I mostly agree with you. The thing I can’t quite dismiss is that fact that transaction costs might prevent us from reaching optimal contracts so there could be some inefficiency.

    People directly linked have their welfare considered by the smoker and can contract, but transaction costs might prevent them from reaching a first best outcome. Similarly, those not directly linked might have a small willingness-to-pay that isn’t exercised because the cost of exercising it would be greater than the benefit.

    I don’t think either of these things necessarily warrants policy intervention, but it suggests to me that VSL could be relevant to a welfare analysis.

  • ben

    Rauparaha, your own death is a private cost. The presumption must be that smokers willingly trade off longevity for (their idea of) quality. Smokers overestimate the health effects of smoking and more than pay their way in health costs through cigarette taxes, so taxes on smoking are already too high. How about less intervention?

    People directly linked have their welfare considered by the smoker and can contract, but transaction costs might prevent them from reaching a first best outcome.

    What transaction costs? The cost of sharing with your best mate/sibling/parent that you don’t like their filthy habit? And compared to what – the transaction costs of government intervention?

  • @ben

    your own death is a private cost

    Well, not really. Reduced lifespan has private costs, but you don’t bear any cost following your death. That cost is borne by those who suffer emotional pain from your passing. The question here is whether that is fully accounted for in your decision to smoke prior to your death.

    What transaction costs?

    It’s not easy to get friends to reduce their smoking or form implicit contracts with them. You have to use up social capital to do it, which is costly to you. Mentioning to them that you don’t like their habit is unlikely to be enough to achieve efficiency. Nobody is suggesting more government intervention here but, if I were, the existence of a regulatory mechanism for smoking suggests that adjusting the tax wouldn’t be hugely costly.

  • @Ben: Transactions costs are for folks you’ve never met but who value your existence. It would be difficult for them all to get together, realize that they collectively value your existence by a lot (though individually, not all that greatly) and pool funds to pay you to smoke less. I admit the theoretical possibility of this; I just can’t see that it’s of any empirical relevance. It’s gotta be a tiny effect relative to the internalities.

    @Ben and @ Rauparaha: I don’t think the transactions costs are necessarily all that high if there really is a pent-up demand out there. Imagine a transactions-costs entrepreneur who sets up an anti-smoking charity where donations can be targeted. Donors make pledges to the charity promising $x if they can convince S to stop smoking. The charity adds up all of the donations for each S, sending emails to each S noting the collective sum of pledges and promising to send along the money if S stops smoking for some period of time. Yeah, yeah, lots of rough edges to work out, and of course the charity would take a cut. But it’s not impracticable. If nobody’s established such a charity, it’s because the expected cut isn’t big enough to make it worthwhile: actual willingness to pay doesn’t make it worthwhile. And so it’s only potentially-Pareto relevant, not actually Pareto relevant.

  • @rauparaha 10:22: You’re counting transactions costs on friends? Yikes. I’d say that you’re already in an implicit contract with friends: you’re imposing costs on them that they’re deciding to overlook ’cause they prefer hanging out with you to changing you on those margins where you’re annoying, and vice versa.

    I can’t see why you’d presume that your mentioning your preferences is insufficient to generate efficiency: if the friend’s preference for smoking is more intense than your preference that your friend not smoke, then that he doesn’t change behaviour is a signal of an efficient outcome, not an inefficient one. And it’s just hard to believe that an external party values someone’s life more than does the internal party.

  • @Eric Crampton
    I agree, there’s a net benefit to hanging out with friends, but that doesn’t imply efficiency. The smoker obviously cares about their friend’s opinions and feelings, but not as much as the friend does. The friend might be willing to pay to reduce the smoker’s consumption of cigarettes, but that’s not really socially acceptable.

    I’m not going to pretend that I think these are serious policy issues, but that doesn’t mean it doesn’t exist 😛

  • S and N both care about their own and the other’s life and about the other’s preferences. S smokes; N doesn’t. S has weighed the pleasures of smoking against the long term health costs and has chosen a private optimum. Is it really plausible that N’s personal value on S’s life is sufficiently high that the global optimum would have S make substantial changes in his behaviour? If N’s valuation is that high, is it really plausible that discomfort about talking about such things is sufficient to prevent the transaction? It’s going to be a pretty narrow range where N’s valuation is sufficiently high to be of consequence for S but sufficiently low as to be trumped by discomfort about talking about it. Existence: sure. I just don’t buy that the range is wide enough to worry about.

  • ben

    your own death is a private cost

    Well, not really. Reduced lifespan has private costs, but you don’t bear any cost following your death.

    Eh? Of course your own death is a cost to you. It is your own welfare that is foregone by the end of your existence. Why else do people so strongly resist it, and demand payment for bearing even a small increases in mortal risk?

  • @Eric Crampton

    Existence: sure. I just don’t buy that the range is wide enough to worry about.

    I agree with that.


    It is your own welfare that is foregone by the end of your existence.

    That’s what I said: “Reduced lifespan has private costs.” So there is a cost to you of a premature death but it’s not the only cost of your death. Following your death there are costs that are borne by those who were close to you.

  • Life insurance…

  • moz

    You seem to be assuming that the cost imposed by smokers on non-smokers who are not their friends is zero. I’d argue that in many cases smokers have zero net friends – the people who like them are offset by those who dislike them. I’d cheerfully pay for the ammunition if I was allowed to shoot them, just to reduce the cost to me of enduring their stench.

    I’d also like to see the tax cover the cost of preventing every cigarette butt making it into the stormwater system. They are toxic and not biodegradeable, and they kill a lot of wildlife. If you include the cost of damage to the larger ecosystem the tax is far, far too low.

  • @moz
    We’re just talking about the social cost of a smoker’s premature death, rather than the full social cost of smoking. We haven’t assumed those things because we haven’t addressed them in this post.

    I haven’t seen any literature on the environmental cost of cigarette butts. Do you know of any which quantify the costs?

  • @ben

    Is that not fairly summarised as, “if transactions costs are important enough to prevent friends sharing opinions on smoking…”?

    I guess it’s just semantics. Telling someone you don’t like it will induce them to take your feelings into account when making their decision. However, they are unlikely to value your opinion as highly as you do. Hence, there are probably further reductions in smoking that the friend would be willing to pay for beyond those induced by sharing their opinions. That’s why I don’t think it’s an entirely accurate summary of what I meant, but that’s probably because I was a bit unclear the first time in what I said.