Missing women has been an important problem for last few decades and one key reason of missing women is son preference. This research provides a causal evidence on how women’s labour market opportunity affects son preference. We define son preference as an indicator for preferring more sons than daughters and compare son preference for women with more access to jobs to that for similar women with lower access. To exploit an exogenous variation in job access measure we consider both road distance based spatial exposure and birth-cohort specific time-varying exposure to export jobs. Since it is possible that roads are placed in areas with higher/lower son preference, road distance between household and job locations can be endogenous. To address this endogeneity, we construct an instrument based on two natural barriers for road placements: straight-line distances between households and job locations and river-widths along those straight-lines. We find that women with higher job opportunity prefer fewer sons. More specifically, a ten percent increase in job access measure reduces probability of preferring sons by 0.009 percentage point. We also find the results robust to alternative specifications of job access measure and not driven by potential concerns like migration, ultrasound, retrospective preference or occupational mobility. This research implies that availability of paid employment for women may reduce the missing women problem.