Are individuals more likely to smoke when they are surrounded by smokers? In this paper, we examine the evidence for peer effects in smoking. We address the endogeneity of peers by looking at the impact of workplace smoking bans on spousal and peer group smoking. Using these bans as an instrument, we find that individuals whose spouses smoke are 40 percent more likely to smoke themselves. We also find evidence for the existence of a social multiplier in that the impact of smoking bans and individual income becomes stronger at higher levels of aggregation. This social multiplier could explain the large time series drop in smoking among some demographic groups.