Measuring peer effects on youth smoking behavior

This paper examines the role of peer effects in smoking behavior using data of middle and high school students in the United States. I present a random utility model that explicitly incorporates complementarity between individual and peer smokings. A Markov process model of smoking interactions between individuals is presented, under the assumption that such interactions occur frequently. I estimate the structural parameters of the model using a steady state distribution that is uniquely determined by the Markov process. The empirical results strongly support the presence of positive peer effects in smoking behavior among young people. Interestingly, peer interactions are found to be stronger within the same gender than across genders. The same result is found for race. Moreover, a multiplier effect is found. The impact of a tax on youth smoking increases by a factor of 1.5 when peer interactions are present.