A large proportion of university students switch fields of study. This is a consequential decision for labour market outcomes, as previous research documents significant variation in earnings across fields. In this paper, I estimate, for the first time, the impact of switching fields on labour market earnings. Using detailed administrative data that allows me to track high school grades, university enrollment decisions, and earnings, I quantify the causal effect of switching across several starting and ending fields. Since the decision to switch fields may be endogenous, I exploit the richness of covariates in my data and use a matching estimator that allows me to create a credible counterfactual for switchers. I find that switchers gain, on average, nearly $ 5,000 in annual labour earnings when leaving programs with large heterogeneity in earnings changes across initial field of study and gender. My findings have potential policy implications for post-secondary institutions, suggesting that reducing the cost of transferring between fields may be beneficial for students.
We study whether racial disparities in economic opportunity appear at an early age. Using administrative education data linked to tax records, we study the income-achievement gap across different races and find important variation. The income-achievement gap is small for East Asian children while it is close to twice as large for Indigenous children. Sorting by income into schools accounts for a large portion of the variation in the income-achievement gap across all student groups. In addition, our results suggest that the large income-achievement gap for Indigenous students may be rooted in inequality in health outcomes and poor housing conditions. Our findings on income-achievement gaps across race could partially explain the different intergenerational mobility outcomes by race documented by others.
This paper examines how students across the high-school grade distribution match into post-secondary programs with different predicted earnings. Using administrative education data linked to post-secondary enrollment and earnings information, we incorporate two post-secondary pathways overlooked by the matching literature: community college and second degrees. Our results indicate that for male students with low high-school grades, the predicted earnings from doing a community college degree exceeds that of doing a four-year college degree. Furthermore, while women tend to go to lower-earning programs than men conditional on high-school grades, this inequality is significantly lessened when incorporating expected earnings from second program completion.
Previous work has documented that gender gap in Engineering and Technology undergraduate programs often exceeds the gender gap in Science undergraduate programs. This paper uses administrative data on undergraduate program applications from high school students in Ontario to study the gender gap in applications to Engineering, Mathematics and Computer Science (EMCS) programs. I find that the STEM courses taken in the final year of high school have a sizeable impact on the likelihood of EMCS application for both women and men. Moreover, high school STEM courses have differential impacts across genders, as Biology has a larger negative impact on EMCS application likelihood for women. I also find that the gender gap in EMCS application grows as the sample is restricted to students who are the most-prepared for undergraduate STEM programs.