Abstract
We contribute to the literature on peer effects of smoking by studying a husband's decision to quit smoking as a function of his wife's smoking. We relax some of the strong assumptions made in previous studies. For example, we study a known peer group and allow for observed or unobserved time invariant exogenous characteristics that may influence both the agent's and his peers' smoking habits. Using data from the Panel Survey of Income Dynamics (PSID), we find a significant peer group effect. The point estimate suggests that the probability that a husband quits smoking increases by 24 percentage units if his wife quits smoking. Because we still make some strong assumptions, this estimate should be interpreted with caution.