Abstract

This study estimates the effect of hours worked at a part time job on GPA among 12th grade respondents to the annual 1991–2004 Monitoring the Future surveys. I use two stage least squares (2SLS) with indicators for various levels of unearned income, which are strong predictors of hours worked, as instruments. Results show that GPA increases with additional work hours up to 15 per week and then declines. 2SLS estimates are substantially larger than those from ordinary least squares and robust to exclusion restrictions variations. Working has a small negative impact on educational time, but a much larger quadratic impact, which is negative up to 15–20 hours per week, on time spent watching television and in social activities. Effects are stable across the sample period, larger for females, non-whites and metropolitan area residents, and linearly positive but substantially smaller for students with high future discount rates.