The labor market returns to a for-profit college education☆
Introduction
After several decades of moderate growth and relatively little controversy in for-profit college education, the 2000s ushered in a new era in higher education. The low cost of providing online education, the availability of federal student aid, and the growing number of students seeking skills beyond the high school level, combined to produce enormous growth in the for-profit sector. Since 2000, enrollment in for-profit colleges has more than tripled, climbing to 2.5 million students in 2010 (National Center for Education Statistics (NCES), 2012, Tables 219 and 222).
With this growth has come increased attention from policymakers, the media, the education community, and students themselves. In the past few years, reports of unscrupulous recruiting practices, fraud in federal financial aid programs, low graduation rates, and high student loan default rates (e.g., Goodman, 2010, Lewin, 2010, GAO, 2010; U.S. Senate Committee on Health, Education, Labor, and Pensions, 2012) have led to new policy proposals and an intense debate over the costs and benefits of for-profit higher education. Notably, the Obama administration recently proposed controversial “gainful employment” regulations that would link an institution's eligibility for federal student aid to student debt, default rates, and labor market outcomes (Federal Register, 2014).
Central to the debate over these regulations and the future of the for-profit industry is the question of the quality of a for-profit education. Those arguing in favor of additional regulation believe that for-profit colleges leave students with insurmountable debt and few skills, while proponents argue that these institutions provide valuable job training for underserved students. Both sides rely heavily on anecdotal evidence and simple comparisons of for-profit students with students in other sectors. Without a better understanding of the causal effects of a for-profit education on earnings, it is difficult, if not impossible, to assess the merits of the proposed regulations, other policies affecting for-profit students, and the sector as a whole.
This study provides an assessment of college quality in the for-profit sector by estimating the labor market returns, or earnings gains, to associate's degree programs in for-profit colleges. We focus on associate's degree programs due to data restrictions, but also because of the importance of the for-profit sector in sub-baccalaureate education. For-profit colleges account for a disproportionate share of associate's degree enrollment and degree receipt. For-profits enroll 11% of all postsecondary students, yet they confer 21% of all associate's degrees – about 194,000 each year (National Center for Education Statistics (NCES), 2012, Table 219).1
A long literature on the returns to education has focused on estimating the earnings gains generated by a year of high school or four-year college. Several studies have also assessed the returns to associate's degrees in public community colleges, but for-profit sub-baccalaureate education has received much less attention in the literature.2 Our study fills this gap in the literature, attempting to mitigate a crucial endogeneity problem that plagues studies of earnings gains: both observable and unobservable factors may jointly influence a student's decision to attend a for-profit institution and her subsequent earnings. For example, students who pursue a degree at a for-profit college may be more motivated than students who do not enroll in any postsecondary education, but motivation can also independently influence subsequent labor market success. Such endogeneity issues can bias cross-sectional OLS estimates of the impact of for-profit colleges on employment and earnings.
To address this problem we draw on 14 years of data contained in the restricted-access 1997 panel of the National Longitudinal Survey of Youth (NLSY97) to implement an individual fixed effects approach. Unlike bachelor's degree candidates, students pursuing associate's degrees often work before, during, and after they attend college, allowing us to compare an individual student's earnings after attendance to her earnings before. We compare the before–after earnings gains of for-profit associate's degree students to a falsified before–after for high school graduates without college degrees, generating a difference-in-difference estimate of the returns to for-profit college attendance.
The individual fixed effects control for all time-invariant student characteristics that may bias cross-sectional estimates of returns, but time-varying unobservables and dynamic selection issues may still remain. To address such concerns, we include an extensive set of controls including an individual-specific county of residence trend, an indicator for the years in which individuals are enrolled in college, and an indicator to control for pre-education dips in earnings among for-profit students. We conduct many robustness checks to test the sensitivity of our estimates to the inclusion of various controls, different measures of earnings, alternate samples, and heterogeneous effects.
We find that students enrolled in associate's degree programs in for-profit colleges experience a 10% increase in weekly earnings, conditional on employment, in our baseline fixed effects specification. Dividing by 2.6 to account for the average length of enrollment yields a 4% return per year of education – a result that is incredibly robust to various assumptions and specifications. Most of the increase comes from hourly wages rather than more hours worked per week. For-profit students are also slightly more likely to be employed full-time than our control group of high school graduates, but we find no difference in the likelihood of any employment. Unconditional estimates including observations with zero earnings (thereby conflating employment and earnings effects) suggest higher earnings gains of around 18%, or 7% per year. We also find suggestive evidence that students completing associate's degrees in for-profits earn about 20% more than their counterparts who dropout.
Overall, our results suggest that students enrolling in associate's degree programs in for-profit colleges earn positive returns on the order of 4% (conditional on employment) to 7% (unconditional) per year. These estimates are slightly higher than other estimates of for-profit returns (Turner, 2012), but fall below estimates of the returns to public community colleges (Jepsen et al., in press, Jacobson et al., 2005) and traditional four-year colleges (Oreopolous & Petronijevic, 2013) found in the literature. They also fall below the returns needed to offset the private and social costs of for-profit associate's degree attendance (Cellini, 2012), suggesting that for-profits may not be worthwhile for the average student.
The rest of the paper proceeds as follows: Section 2 reviews the literature on the returns to education and Section 3 provides background on for-profit colleges. Section 4 details our estimation strategy. Section 5 describes the data, Section 6 presents results, and Section 7 concludes.
Section snippets
Related literature
Over the past half-century, a large literature has developed to measure the returns to schooling. Reviews of the literature by Card (1999, chap. 30) and report that one additional year of education results in earnings gains in the range of 6–9%. More recent and better-identified analyses reveal higher returns for college attendance, averaging 10–15% per year (Card, 2001, Goldin and Katz, 2008, Oreopolous and Petronijevic, 2013). The vast majority of the research in this area has focused on high
Background
Our paper estimates the returns for students enrolled in associate's degree programs in for-profit institutions. Associate's degree programs typically require two years of full-time coursework and result in the attainment of an Associate of Arts (AA) or Associate of Science (AS) degree.6 Students may obtain their degree in any number of majors, including traditional liberal arts and science majors like history, psychology, or computer
Empirical methods
Many studies of the returns to education estimate cross-sectional regression models following Mincer (1974). The main concern with such estimates is the potential endogeneity of attendance because individuals more likely to earn higher earnings may also be more likely to attend college, resulting in biased estimates of the returns to education.
To address this endogeneity problem, our research design exploits the panel structure of the NLSY97 data. Unlike high school and traditional four-year
Data
To implement our analysis, we draw on the restricted-access 1997 panel of the National Longitudinal Survey of Youth (NLSY97), a major nationally representative longitudinal survey that tracks a cohort of students through secondary school, college, and beyond. The NLSY97 is based on a panel of 8984 youths who were 12–18 years old when they were first surveyed in 1997.13
Main results
Table 2 presents our first set of findings on log weekly earnings. Specifications (1) and (2) are cross-sectional OLS models common in the literature, while specifications (3) through (5) are our preferred estimates, exploiting within-individual variation by including individual fixed effects.
The first row of Table 2 finds small and statistically insignificant effects of post-education (Postit) in every specification. This result is expected because Postit is falsified for the high school only
Concluding remarks
This study takes a step toward assessing the quality of for-profit postsecondary education, estimating the before–after earnings gains of students pursuing associate's degrees in for-profit institutions. Using an individual fixed effects approach and data from the restricted-access NLSY97, we find that students who enroll in for-profit institutions experience earnings gains of about 10% over individuals who complete high school without any college degree, conditional on employment, in the 3 or
References (49)
- et al.
A review of estimates of the schooling/earnings relationship, with tests for publication bias
Labour Economics
(1999) Choice of for-profit college
Economics of Education Review
(2012)- et al.
Estimating the returns to community college schooling for displaced workers
Journal of Econometrics
(2005) Economic returns to community and four-year college education
Journal of Socio-Economics
(1994)- et al.
Does compulsory school attendance affect schooling and earnings?
Quarterly Journal of Economics
(1991) - et al.
Empirical strategies in labor economics
- et al.
Over-identification tests in earnings functions with fixed effects
Journal of Business and Economic Statistics
(1991) Estimating the effect of training programs on earnings
Review of Economics and Statistics
(1978)- et al.
Using the longitudinal structure of earnings to estimate the effect of training programs
Review of Economics and Statistics
(1985) - et al.
A review of estimates of the schooling/earnings relationship, with tests for publication bias
(1999)
Proprietary schools and their students
Journal of Higher Education
For-profit higher education and community colleges
The causal effect of education on earnings
Estimating the return to schooling: Progress on some persistent econometric problems
Econometrica
Crowded colleges and college crowd-out: The impact of public subsidies on the two-year college market
American Economic Journal: Economic Policy
Financial aid and for-profit colleges: Does aid encourage entry
Journal of Policy Analysis and Management
For-profit higher education: An assessment of costs and benefits
National Tax Journal
College costs and financial constraints: Student borrowing at for-profit institutions
Does federal student aid raise tuition? New evidence on for-profit colleges
American Economic Journal: Economic Policy
Effects of for-profit training on earnings
The for-profit postsecondary school sector: Nimble critters or agile predators?
Journal of Economic Perspectives
Program integrity: Gainful employment
Longitudinal analyses of the effects of trade unions
Journal of Labor Economics
For profit colleges: Undercover testing finds colleges encouraged fraud and engaged in deceptive and questionable marketing practices
Cited by (77)
Returns to different postsecondary investments: Institution type, academic programs, and credentials
2023, Handbook of the Economics of EducationCitation Excerpt :This implicitly identifies the effect of for-profit enrollment relative to no college attendance. Using NLSY97, Cellini and Chaudhary (2014) set up a difference-in-difference analysis where the control group consists of people who never enroll in college compared to for-profit enrollees (before and after enrollment). This means that the estimates of for-profit returns are relative to high school graduates with no college experience.
Addressing nonfinancial barriers to college access and success: Evidence and policy implications
2023, Handbook of the Economics of EducationStudent debt and default: The role of for-profit colleges
2022, Journal of Financial EconomicsCitation Excerpt :In addition, the use of student fixed effects requires one to focus on older students who have sufficient pre-enrollment earnings. Findings from these analyses point to for-profits raising earnings relative to those who do not attend college (Cellini and Chaudhary, 2014) but lowering earnings growth relative to those who attend public institutions (Cellini and Turner, 2019; Liu and Belfield, 2014b). The papers that most credibly overcome the selection biases associated with identifying the causal effect of for-profit enrollment on labor market outcomes are two randomized audit studies (Darolia et al., 2015; Deming et al., 2016).
Post-Schooling off-The-Job training and its benefits
2021, Labour EconomicsCitation Excerpt :In terms of the other characteristics, both groups are quite similar except for those who enroll in Off-JT having larger (parental) household net worth in 1996. As discussed in Cellini and Chaudhary (2014), there are some extreme values in the earnings data in the NLSY97. To deal with outliers, the data is winsorized as in Autor, Autor et al. (2016).
The effects of education on mortality: Evidence using college expansions
2024, Health Economics (United Kingdom)
- ☆
We thank Burt Barnow, Dylan Conger, Hector Cordero-Guzman, Joseph Cordes, Janet Currie, Jesse Cunha, Erin Dunlop, Sue Dynarski, Claudia Goldin, Jonathan Guryan, Wes Hartmann, Caroline Hoxby, Larry Katz, Donald Parsons, Roberto Pedace, Cecilia Rouse, Kenneth Troske, and seminar participants at Harvard University, George Washington University, Naval Postgraduate School, U.S. Treasury, U.S. Government Accountability Office, the Federal Reserve Bank of Atlanta, AEFA, APPAM, SEA, and the NBER Education Program Meeting for helpful comments. Faith Fried, Phil Gross, Megan Hatch, and Aisling Scott provided excellent research assistance. We are grateful for financial support from the Ford Foundation (Grant Number 1095-0464). The views expressed in this paper are the sole responsibility of the authors and do not necessarily represent the official views of the Ford Foundation, Department of Defense, or the U.S. Government. A previous version of this paper was titled “The Labor Market Returns to Private Two-Year College Education.”.