Illegal Drug Use and Employment
利用1984和1988年美国青年纵向调查数据，通过工具变量法处理药物使用与就业的内生性，发现吸食大麻和可卡因显著降低就业概率，且标准Probit估计存在严重偏误。
This article investigates the relationship between employment and the use of marijuana and cocaine for males in National Longitudinal Survey of Youth data from 1984 and 1988. Previous studies yielding mixed evidence may have inadequately accounted for the simultaneity between drug consumption and employment. I implement an instrumental variable procedure that identifies drug use with variables that are empirically unrelated to employment, including the regional cocaine price and a state marijuana decriminalization indicator. Results indicate that the use of each drug substantially reduces the likelihood of employment. Exogeneity tests reveal that standard probit estimates are severely biased toward zero.