« Back to Results
Pennsylvania Convention Center, 202-B
Hosted By:
American Economic Association
Preliminary results based on the degree combination fixed effects approach show substantial differences across graduate fields in labor market returns. For example, law and especially medicine have high returns that operate primarily through change in occupation. The return to an MBA is essentially zero for engineering and business majors but substantial for economics and political science majors. The extent to which the return to an MBA is within versus across occupation depends upon the undergraduate major. A masters in education has a modest positive return for education and psychology/social work majors, but a negative return for engineering majors. Broadly speaking, OLS performs poorly. It overstates (understates) the returns to advanced degrees that are typically obtained by individuals with high-paying (low-paying) undergraduate majors that lead to high-paying (low-paying) occupations. For example, OLS estimates substantially overstate the average return to an MBA. Prior to attending graduate school, people with low paying undergraduate degrees earned more and were more likely to work in a relatively high paying occupation before getting an MBA. By the same token, OLS estimates tend to understate the return to a masters in education, which is typically obtained by individuals who had chosen to work in the low paying education field prior to graduate school
College Major Choice
Paper Session
Sunday, Jan. 7, 2018 10:15 AM - 12:15 PM
- Chair: Amanda Griffith, Wake Forest University
The Labor Market Returns to College Major and Advanced Degree
Abstract
We study the labor market return to particular fields of study, focusing on graduate degrees. We begin with regression models in which undergraduate field and graduate field enter additively, with 20 undergraduate majors and 20 graduate degree types. Then we consider combinations of undergraduate and advanced degrees, such as a BA in education followed by an MBA. To address selection and omitted variables bias, we work with panel data constructed by merging information on individuals whom we can be track in the “restricted use” versions of multiple NSF surveys. In addition to OLS regression, we use two fixed effects specifications. The first includes individual fixed effects. The second includes fixed effects for undergraduate field/graduate field combinations obtained by the last time we observe an individual. A major concern of our analysis is the fact that wages depend on occupation, and that field of study and occupation choice are driven by preferences as well as by wages. The panel data allows us to study the interplay among undergraduate field, graduate field, and occupation choice.Preliminary results based on the degree combination fixed effects approach show substantial differences across graduate fields in labor market returns. For example, law and especially medicine have high returns that operate primarily through change in occupation. The return to an MBA is essentially zero for engineering and business majors but substantial for economics and political science majors. The extent to which the return to an MBA is within versus across occupation depends upon the undergraduate major. A masters in education has a modest positive return for education and psychology/social work majors, but a negative return for engineering majors. Broadly speaking, OLS performs poorly. It overstates (understates) the returns to advanced degrees that are typically obtained by individuals with high-paying (low-paying) undergraduate majors that lead to high-paying (low-paying) occupations. For example, OLS estimates substantially overstate the average return to an MBA. Prior to attending graduate school, people with low paying undergraduate degrees earned more and were more likely to work in a relatively high paying occupation before getting an MBA. By the same token, OLS estimates tend to understate the return to a masters in education, which is typically obtained by individuals who had chosen to work in the low paying education field prior to graduate school
Understanding Determinants of Major Selection in Higher Education
Abstract
This paper studies the undergraduate application and registration behavior of secondary school students to understand the determinants of student sorting across majors. We find that students' application decisions are influenced by the application, admission, and matriculation outcomes of older cohorts within their high school. From a policy perspective, is important to understand the determinants of college and major selection in order to more effectively target interventions for directing students into high-return undergraduate programs. We study application and registration decisions in the context of Ontario, the most populous province in Canada. We use microdata from the Ontario Universities' Application Centre on the universe of all Ontario secondary school students applying to Ontario universities from 1995 to 2012. Students in this context apply for admission to university-major specific choices. We look at how the probability of a student applying to college programs is influenced by the application and enrollment decisions of students in previous cohorts in the same high school. In general, causal peer effects are difficult to identify primarily because of endogenous peer group formation, correlated unobservables, and simultaneity. We deal with simultaneity by focusing on effects across different cohorts, looking at how students applying to college in one year, influence students that apply in subsequent years. Using information from other cohorts mitigates identification concerns related to reverse causality. To deal with correlated unobservables, we use a rich set of fixed-effects to control for unobserved differences across schools and types of majors. We use idiosyncratic variation from year-to-year differences in admission outcomes and we compare the probability of application to a specific program across students that had different levels of exposure to information about that program. We classify majors into 18 categories according to the general major field, with some majors being more specific, such as engineering, nursing or psychology, and other categories being more general, such as unspecified arts or sciences. The detailed microdata allow us to estimate a rich model of school-major choice using the ranking of preferences in the students’ applications. We incorporate perceived ability from the students’ high school grades and class rank, as well as immigration status. Our findings indicate that there is an information transmission mechanism across cohorts within a high school, with students becoming more likely to apply to a program after someone from their high school has applied to, been admitted to, or registered in that program.JEL Classifications
- I2 - Education and Research Institutions