Panel Paper: Does Education Demand Respond to Labor Demand? Evidence from the Community College Sector

Thursday, November 7, 2019
Plaza Building: Concourse Level, Governor's Square 16 (Sheraton Denver Downtown)

*Names in bold indicate Presenter

Riley K. Acton, Michigan State University


America’s community colleges educate nearly one-third of undergraduate students and maintain close links to their local economies, in part by offering vocational programs that train students for relevant local occupations. In recent years, a combination of changing labor market structures following the Great Recession and increasing evidence that labor market returns to community college credentials vary substantially by program (Bahr et al., 2015; Stevens et al., 2018) have prompted an interest in policies that steer community college students toward high-return programs. For example, various states are engaged in efforts to track and publicize program-specific earnings metrics while others have begun tying state appropriations to colleges' ability to produce credentials in high-demand areas.

Implicit in these well-intentioned policies is the assumption that community college students, or potential students, do not sufficiently respond to changes in labor market opportunities on their own. Yet, while there is some evidence that overall community college enrollment responds to local labor market conditions (e.g. Betts & McFarland, 1995; Foote & Grosz, 2018), there is surprisingly little research on the extent to which labor market opportunities for specific occupations affect students' decisions to enroll in closely related educational programs. This paper seeks to establish the relationship between changes in occupational labor demand and students’ education choices by using detailed administrative data on students in Michigan and introducing a novel method of measuring occupation-specific labor demand shocks. Specifically, I match all mass layoffs and plant closings occurring in Michigan between 2001 and 2017 to their respective NAICS industries and rely on the distribution of occupations across industries to estimate, and assess students’ responses to, occupation-specific layoffs in a given county and year.

Consistent with a model of education choice where layoffs reduce students’ expected future earnings, I find that students are deterred from entering programs in which closely-related occupations have recently experienced a local labor demand shock. In particular, county level enrollment in a program decreases by 1-5% for every additional layoff per 10,000 working-age residents in the county. This effect is driven by particularly strong among students residing in the most rural counties in the state. I then find that students shift their enrollment to other types of vocationally oriented programs. By further using data on occupational skill composition and exploiting the close links between community college programs and occupations, I document that students primarily shift enrollment into programs that require similar skills.

Full Paper: