May 12, 2008

To Smoke or not to Smoke - Act Two

Now it is a month that I do not touch a single cigarette. The reason why I waited so long for the second part of this little investigation is the fact that I am currently in the posh seminar hotel of the Swiss National Bank. Why could this be a reason for delaying the posting of this article? Well, I refer you to part one and my understanding of commitment, or short: here I have the perfect environment of falling back into my old habit (i.e. nothing to do besides studying and no “outside control”) so I need a new commitment mechanism….

So here the second part of my little investigation into why I smoked and why I stopped to do so. The starting point for my investigation was the data of the “Nationmaster”, where I produced the following correlation:

The correlation is around 0.78. Though the result is based on a macro relationship, Cutler and Glaeser (2008) find a similar relation in a micro panel in their recent paper “Social Interaction and Smoking”. Now, what am I supposed to learn form this? The fact that I am doing a PhD in Economics, and have been running through the education system of various countries since nearly 20 years, finally shows some pay-off!?!. Mhh, I do not know exactly why this should be the case, since I do not believe that more educated people are more aware of the consequences of smoking. I would rather guess that more educated people have jobs that are more stressing and hence find it more difficult to stop smoking. A resolution may be that more educated people tend to behave more often in a rational manner, though I have no proof for this hypothesis. The authors argue that it might be due to the fact that “less educated groups might be more response to peer influences”, or in the case of young people, plainly speaking group pressure.

Another interesting scatter-plot is the following. This observation is in line with the paper above and the ideas on social interaction and smoking in the sense that if people who are more independent are also less affected by social benefits from the action of others, one would expect these individuals to smoke less.

Next, I would like to address the question which country was most successful in reducing the prevalence of smokers. Unfortunately, I could not find sufficiently long series’ for many countries. However, for the below countries there was sufficient information from the OECD Health data (Note that the rate is normalized to 100 in 1988 for all countries to focus on the relative change in the rates).

The three Skandinavian countries, led by Denmark belong to the best performing countries (together with the US). One has to wonder sometimes why Denmark is such a good country. It manages to half the rate of smokers within 15 years. Copenhagen, the capital, is frequently voted one of the cities in the world with the highest quality of life and Danish people tend also to be the happiest in the world, which according to Bo Honore is due to the European-soccer-cup-surprise-victory against Germany in the final in ‘92. Greece, France and the Netherlands seem to have performed much worse (not in terms of soccer though, except for the Dutch who won the last championship when again…???). Maybe the recent changes to forbid smoking in bars and restaurants in Ireland and since this year also in France may affect these numbers in the future. Given the available data it is to early to talk about the effects of the bans on smoking in bars and restaurants in various countries.

However, there is an another finding in the paper by Cutler and Glaeser. Accordingly, being married to a smoker increases the likelihood to smoke. Well this is not surprising and as the authors reckon most likely a biased result as of sorting, i.e. it is more easy to be together with (or ignore the fact that someone is) a smoker if you smoke yourself. Nevertheless, the more interesting result of the authors is that if a worker has quit smoking as a response to a smoking ban at work, the likelihood that the spouse quits as well is roughly 37%. Interestingly, men react more on their wife quitting (50%) then women reacting on their husband quitting. Now this does not surprise me! Anyways the funny thing about these findings is the policy implication: Make the women quit and you get the men to quit as well. So policies should be geared at women!

Personally, I consider the importance of policy rather at the prevention level (more than three-quarters of smokers begin smoking before the age of 19). First it is cheaper. Second, it starts at the root of the problem and third it seems more effective to me since most smokers started smoking early in their lifetime.

According to my humble point of view an important advance in avoiding higher smoking rates under young people would be the following: Parents should credible commit to offer their children a predefined amount of money (or a world trip or whatever valuable thing that children at that age cannot obtain) if their child surpasses the age of 18 without having smoked (or only “randomly” smoked). Why is this the best strategy? Because, 1) it prevents instead of “curing”, 2) it is based on “equality” and 3) it minimizes social costs. Consider, instead the strategy which is more often observed: Parents pay their children a certain amount of money if the child stops smoking for at least a year. The problem with this policy is that the child that did never smoke may end up ex post worse-off since it will not be offered the “bet”.

So from all these unstructured random arguments the conclusion that I draw for myself is the following: The best way is to prevent people from smoking in the first place since this is much easier than making them stop afterwards. In particular in my case no policy is responsible for me stopping to smoke but only my own will. However, I realized in the last week that there is a high probability that I fall back since I feel from time to time the strong will to smoke. Had I never started smoking, I would never feel this temptation.


Pierre-Louis said...

people from higher educated countries are way more aware of the dangers of developing countries, thes still think it is not serious, as they think you can avod getting aids by eating roots...and this is a culture thing, meaning it is explaind by the country's culture (avergae level of educatin) and not a person's education...

Sebastian said...

Sorry, but I disagree strongly with this argument: First: If I take out the developing countries from the sample (first graph) the slope remains significant and negative. So the logic applies to developed countries as well. Are Greek less aware of the danger of smoking than Americans? I do not think so. It has to do with culture however, but then again what is "culture" in this case?

Playing the card of "the others are sillier" does not convince me at all as an argument. Any better one?

Pierre-Louis said...

Are Greek less aware of the danger of smoking than Americans?

the answer to this question is yes, of course.

in america kids are bombarded by ads and teacheras and parnets about how disgusting and dangerous it is to greece it is still socially accpeted and widespread, so people dont believe that it is that harmful...

Sebastian said...

I do not mean to have the last word but I am still not convinced. I see that "culture" has a big role to play, but not education. Telling me ten times how bad smoking is does not make me understand the problem of smoking better than telling it twice to me (i.e. after a certain level of education more education does not have any more impact. To me most developed countries have achieved this minimum level of education).
To me it seems rather that the level of education is highly correlated with something else (e.g. more money being spent on anti-smoking initiatives that erode the "cultural image of smoking") which in turn explains the different smoking rates. And here we would again be in the realm of IV, of which you are an expert now.

cosi said...

I do not think that education is the main driver behind the decision to smoke or not to smoke. At the micro level, I think that the opportunity cost of smoking is higher for people who are busier/wealthier, who have high productivity and cannot really afford to waste time on smoking. These people will of course have high educational attainements, so if you correlate smoking to education you are spuriously picking up the correlation between smoking and wealth. I am tempted to think that this reasoning also holds at the macro level, I would regress numbers of smokers on education controlling for per capita gdp to check it.