Abstract
Based on the 2003 Trends in International Mathematics and Science Study, the authors find that negative associations between student employment and academic achievement are stronger in some countries than in others – differences likely to result from country-specific work opportunities and needs. Turning to the 2004 Educational Longitudinal Survey of the United States for causality, they observe a curvilinear association between employment and math proficiency: working up to ten hours per week has a modest positive effect, 10–19 hours has no effect, and 20 hours or more has a substantial negative effect. The possible endogeneity of work-hours is then tested with instrumental variables.