Stock market behavior of individual investors is highly correlated with stock market behavior of their co-workers. For example, a ten percentage point increase in the fraction of co-workers that purchase stocks in a given month is associated with a two percentage point increase in the likelihood of individuals making a purchase. The high correlation exists even after taking controlling for individual socio-demographic characteristics and for time, stock, zip code, and plant fixed effects. Using data on family relations and on residential zip code, we show that the high correlation is not driven by peer effects at the family or zip code level. Moreover, workplace peer effects appear to be strong relative to geographical peer effects.