Non-Completion, Student Debt, and Financial Well-Being: Evidence from the Survey of Household Economics and Decisionmaking

Introduction

As the price of college and student loan debt has grown relative to inflation in recent decades, many have questioned the return on investment offered by a college degree.1 Recent research tends to highlight the importance of risk when examining the question of “is college worth the investment?” (Hendricks and Leukhina, 2018; Akers, 2021). These works emphasize that a college degree can both be a financial boon for the average student (amounting to roughly $1 million in additional lifetime earnings), but also a risky investment which will not pay off for many students. Arguably the single-biggest determinant of the downside risk associated with attending college is the substantial likelihood of non-completion (Toutkoushian, Shafiq, and Trivette, 2013; Webber, 2016).2

In this note, we explore the relationship between non-completion in higher education and future financial well-being. Utilizing a unique set of questions in the Survey of Household Economics and Decisionmaking (SHED), we focus on a sub-population of particular interest to policymakers and colleges: individuals who borrowed for college (at any level) but failed to complete their degree. This sub-population faces the long-term financial costs of student loan borrowing without incurring the financial benefits associated with degree completion and, therefore, may face additional risks of falling behind.

The financial returns to college are disproportionately bundled with the diploma (Flores-Lagunes and Light, 2010). This means the implicit and explicit costs of attending college can be particularly burdensome for those who fail to graduate. With only about six in ten attendees of four-year institutions completing a degree, and even lower rates at two-year schools,3 the scale of the non-completion problem is large. Both the government (e.g. Denning, Marx, and Turner, 2019) and schools (Evans et. al, 2020; Goldrick-Rab et. al, 2023) have invested significant resources in attempts to increase graduation rates. The overall take-away message from this literature is that there is no free-lunch when it comes to increasing completion rates; light-touch/cheap interventions tend not to move the needle much (Oreopoulos, 2021).

Yet financial outcomes for borrowers who do not complete a degree are historically understudied due to a lack of data identifying this group. Most financial/labor market surveys do not ask about student loan borrowing, and those that do rarely capture information on which levels of education borrowers financed using student loans. In contrast, the SHED asks respondents both their highest level of education and which educational programs they funded by taking out student loans, which allows us to identify those who borrowed using student loans for a higher level of education than they completed.

We find strong negative effects of non-completion among student loan borrowers on a variety of measures of financial well-being and indicators of sentiment/regret about one’s education. These negative effects exist when comparing dropouts to graduates and when holding the level of education constant and looking at the association between excess borrowing and financial outcomes. These results underscore the importance of identifying and investing in interventions which increase completion. The link between non-completion and receipt of SNAP benefits further underscores potential budgetary effects from reducing the number of programs with low completion rates.

Data and Methodology

This note uses data from the Survey of Household Economics and Decisionmaking (SHED), conducted annually by the Federal Reserve since 2013. The SHED is a nationally representative survey of roughly 12,000 adults which seeks to provide insights into the state of consumer/household finances not available in other surveys.

Our sample period spans the years 2017 to 2019. 2017 is chosen as the starting year because the relevant questions on current college enrollment needed to identify non-completion were not identical in prior survey waves. More recent surveys are not included because the pandemic student loan payment pause that began in 2020 muddles any conclusions or comparisons that can be made to prior/future years without such a pause.4

We leverage a unique set of questions asked in the SHED to identify students who borrowed for a level of education which they did not complete. The SHED separately asks respondents their highest completed level of education and the highest level of education for which they borrowed. By combining these responses with current enrollment status, we are able to identify the sub-population this note focuses on.

We estimate a series of regression models which take the following form:

Yi=α+βNoncomplete+Xγ+ε

Y denotes various financial well-being and attitudinal outcomes (overall financial well-being, the ability to cover an unexpected $400 expense, homeownership status, SNAP receipt, beliefs about the benefits vs. costs of their undergraduate education, and specific regrets about education choices). Noncomplete is an indicator for whether the respondent borrowed for a level of education that they did not complete. X is a vector of control variables including age, race/ethnicity, gender, level of education, level of education borrowed for, and year fixed effects.

In addition to estimating Equation (1) for each dependent variable, we can also alter the implicit comparison being made by limiting the sample to a specific level of education. We fix the highest degree completed, which compares (for instance) individuals who borrowed for graduate school but did not complete their graduate education with individuals holding an undergraduate degree that never attended graduate school. This comparison focuses more on the role that additional student loan debt plays in financial well-being. Alternatively, we can fix the highest level of education an individual borrowed for. This variation focuses more on how financial well-being is associated with the failure to obtain a particular credential.

It is important to note that the variation used in this note is neither experimental nor quasi-experimental. In other words, it is not random which students drop out of college. Moreover, whatever factors lead these students to drop out are likely closely related to labor market and financial success. For instance, if a student drops out because they experience a large negative financial shock (e.g. their car breaks down, they are laid off from their job), at least part of the estimated non-completion effect will actually be causally due to this shock. The results presented below should thus be interpreted descriptively, and are meant to illustrate a specific population which is struggling and could be efficiently targeted by policymakers.

Results

Table 1 presents summary statistics for our full sample of student loan borrowers, as well as a breakdown by completion status.

Or if you are ready to learn how Jobspeaker can help you show value for the tuition students pay, speak with us today! 

 

author avatar
wp-admin

You May Also Like…

2024 Human Progress Report

2024 Human Progress Report

Our findings indicate that upskilling increasingly mirrors the importance of formal education. With the rapid evolution and obsolescence of skills,
individuals globally voice the need for continuousskill acquisition.

read more

Company

About

Terms of Use

Privacy Policy

Help + Tips

FAQs

Blog

How it Works

Contact + Press

Contact

Press Kit

News & Stories

Do Not Sell Request

Follow Us For the Latest Updates

Yes! We have a plug-and-play job board – but we offer so much more!

Ability to manage and scale quality hands-on learning experiences.

Finally, a way to make curriculum mapping easy! Jobspeaker maps your curriculum to skills employers demand. Then Skills Mapping™ guides learners to programs that teach the skills needed for their dream job.

Automatically built from a student's completed coursework. It showcases what a student has learned, earned, and how it relates to employment.

Our products bridge the education-to-workforce talent pipeline through an AI data-driven platform matching in-demand skills to career pathways.