"Sin taxes" have been all the rage for some time, with the idea being that people respond to hikes in the prices of various deleterious substances by reducing their unhealthy behaviors. Taxes on goods such as tobacco and alcohol are often justified on the grounds that people's unhealthy behaviors have social costs: one man's boozing and smoking adversely affects those around him.
Conventional wisdom suggests that, on average, such taxes work. However, taxes may work differently for different people. As far as health promotion, one would expect sin taxes to be most powerful if they can effectively help change among individuals who have more severe and intractable problems with smoking and drinking.
Some of my colleagues at Yale (including two of my advisers, Jason Fletcher and Jody Sindelar) have produced some interesting work looking into the effectiveness of tobacco and alcohol taxes on different types of smokers and drinkers, respectively. To explore heterogeneity among smokers and drinkers, they use a latent class method (finite mixture models - discussed two posts ago) to identify different groups of people. What the find is striking: for tobacco use among adolescents and alcohol use among adults, taxes are least effective among groups that tend to smoke or drink more heavily and who generally have the least willpower to quit. The authors make use of very detailed data on people's behaviors, preferences and outlook on life to build these interesting stories.
The results suggest that, among those least likely to quit on their own, taxes appear to have little power in inducing behavior change. Basically, from a public health standpoint, if one is interested in reducing unhealthy behaviors in these populations, policies other than/in addition to taxes will likely be required.
1 comment:
I like that paper and also the finite mixture models in general. The problem is that I don't think we can ever really know who is in the latent groups, but it makes sense in this application.
Post a Comment