Does Replacing Student Loans with Need-Based Gift Aid Change College Outcomes for Low-Income Students?

This study asks whether substituting student loans with need-based gift aid alters postsecondary outcomes for low-income students at Michigan State University. Using a regression discontinuity design around the Spartan Advantage Program’s income eligibility threshold, the author analyzes administrative data from 2000–2020 covering demographics, academic records, and financial aid packages. Over four years, recipients received about $16,465 more in gift aid and $12,470 less in loans than near-threshold nonrecipients. Results show no significant effects on persistence or graduation. However, aid composition shifted academic decisions: recipients were 9–13 percentage points less likely to remain in STEM and showed improved alignment between majors and academic preparation .

Full Citation and Link to Article

Walker, Seth. “Financial Aid Package Composition and Postsecondary Outcomes for Low-Income Students: Evidence from the Spartan Advantage Program.” Education Finance and Policy, 2025. https://doi.org/10.1162/edfp.a.27

Extended Summary

Central Research Question

This paper investigates how the composition of financial aid packages—specifically, replacing loans with need-based gift aid—affects postsecondary outcomes for low-income students. The study focuses on Michigan State University’s Spartan Advantage (SPAD) program, which substitutes loans with grants for students at or below the federal poverty line. The central question is whether this compositional change influences persistence, graduation, and major choices for financially disadvantaged students. Unlike past work emphasizing increases in total aid, this study isolates the impact of aid type, asking: does reducing reliance on loans while holding total aid roughly constant alter students’ academic trajectories?

Previous Literature

A large body of research has examined how financial aid affects access, persistence, and degree completion. Studies on Pell Grants, state merit programs, and institutional grants consistently find that increased gift aid improves persistence and graduation (Bettinger 2004; Dynarski 2005; Castleman and Long 2016). Gift aid is thought to reduce credit constraints, improve student well-being, and lower dropout risk. In contrast, the evidence on loans is mixed: some studies show that debt burdens influence major choice and career trajectories (Rothstein and Rouse 2011), with higher debt steering students toward high-earning majors and jobs.

Several “no-loan” policies at selective universities have provided natural experiments. Rothstein and Rouse (2011) found that additional debt of $10,000 increased starting salaries by about $2,000 annually, suggesting that debt pressure influences career paths. More recent work (Hampole 2023) shows that no-loan policies increase the likelihood of pursuing “high-earning” majors, particularly among students from the lowest socioeconomic tercile. Other studies of programs such as the Wisconsin Scholar Grant and Florida Student Access Grant found strong positive effects on persistence and, in some cases, higher STEM completion (Goldrick-Rab et al. 2012; Castleman et al. 2018).

The contribution of this paper is twofold. First, it focuses on aid composition rather than total aid levels. Second, it studies a broad-access public university and students at or below the poverty line, a population underrepresented in previous research but central to future policy debates.

Data

The author uses administrative data from Michigan State University covering full-time, dependent, in-state students who first enrolled between 2000 and 2020. The dataset contains:

• Admissions data: demographics (race, gender, parental education, first-generation status), high school GPA, ACT/SAT scores, and intended major.

• Academic records: credits attempted and completed, GPA, major changes, and graduation outcomes.

• Financial aid files: detailed itemized aid, including Pell Grants, SPAD awards, loans, work-study, and other institutional or federal aid.

• Earnings data: major-specific labor market earnings from the U.S. Census Bureau’s Post-Secondary Employment Outcomes (PSEO) dataset, matched by classification of instructional program (CIP) codes.

The sample allows the author to compare outcomes for students narrowly below and above the SPAD eligibility threshold (federal poverty line). Because the data extend through 2024, the study can observe long-term outcomes for most cohorts.

Methods

The analysis employs two quasi-experimental strategies:

1. Regression Discontinuity (RD) Design

   Eligibility for SPAD is determined by adjusted gross income (AGI) relative to the federal poverty line. Students below the line are eligible, while those just above are not. Because students cannot easily manipulate AGI and appear balanced on observable traits, this setting produces a natural experiment. The RD design estimates the causal impact of SPAD receipt on outcomes such as persistence, graduation, and financial aid composition. Due to imperfect take-up (not all eligible students receive SPAD, and some just above the line do), the design is “fuzzy” RD, implemented with two-stage least squares.

2. Difference-in-Discontinuities (Diff.-in-Disc.) Design

   For major choice outcomes, the author compares students’ intended majors at admission (chosen before financial aid packages were announced) to their graduation majors. This design accounts for pre-existing differences across the poverty line and isolates how aid composition changes major trajectories.

The study also conducts subgroup analyses (by gender, race, and first-generation status) and robustness checks using alternative bandwidths, placebo thresholds, and difference-in-differences models.

Findings/Size Effects

Financial Aid Composition

• SPAD significantly shifted aid composition without dramatically increasing total aid.

• Over four years, recipients received about $16,465 more in gift aid and $12,470 less in loans compared to non-recipients near the threshold.

• This represented a 30% increase in cumulative gift aid and a 45% reduction in loans relative to peers.

• Total aid rose modestly (about 5.5% over four years), indicating that the program mostly restructured aid rather than expanded it.

Persistence and Graduation

• No statistically significant effects were found on persistence or graduation outcomes.

• Point estimates suggested minor positive effects (e.g., +2 percentage points persistence into the second year, +4–5 pp graduation rates within six years), but none were significant.

• High baseline graduation rates among low-income MSU students (82% within six years) may have left limited scope for improvement.

• Interpretation: aid composition alone, unlike total aid increases, does not strongly affect degree attainment.

• 
  

Major Choice

• The most significant impacts appeared in students’ academic trajectories.

• At admission, SPAD-eligible students were slightly more likely to intend STEM majors and to “overmatch” (choose majors requiring higher ACT scores than their own).

• By graduation, these patterns reversed: SPAD recipients were 9–13 percentage points less likely than non-recipients to remain in STEM.

• Recipients were also 1.4–1.6 ACT points less likely to graduate in overmatched majors, suggesting improved alignment between academic preparation and chosen field.

• Importantly, SPAD recipients were not more likely to move into initially lower-earning majors, but rather into fields better reflecting academic fit.

  
  

Heterogeneity

• White/Asian and non-first-generation students often saw both loan substitution and increases in total aid; these groups showed the largest shifts in majors and mismatch reduction.

• First-generation and non-White/non-Asian students mainly experienced loan-for-grant substitution, but still demonstrated modest improvements in major fit.

• Male students showed little measurable response, while female students exhibited significant reductions in overmatch (≈2 ACT points).

Robustness Checks

• Findings were generally consistent across bandwidths, though major-related effects were strongest with broader samples ($18,000 bandwidth).

• Placebo thresholds showed no significant effects, supporting causal interpretation.

Conclusion

This study demonstrates that altering the composition of financial aid—replacing loans with grants—can shape students’ academic decisions, even if persistence and graduation remain unaffected. For low-income students at Michigan State, receiving substantially more gift aid and fewer loans did not increase the likelihood of completing a degree, but it did change how students navigated majors. Specifically, SPAD recipients were less likely to remain in STEM fields and more likely to choose majors aligned with their academic preparation, suggesting that reduced financial pressure allowed them to prioritize personal and academic fit over perceived financial returns.

The null findings on persistence and graduation underscore the importance of total aid levels in determining degree completion. Aid composition alone appears insufficient to move the needle when students already receive enough aid to cover attendance costs. However, the results highlight that aid type matters for shaping academic paths and potentially long-term career satisfaction.

Policy implications are twofold. First, no-loan programs may not boost graduation rates but can influence field-of-study decisions with downstream effects on earnings, occupational fit, and equity across fields. Second, since SPAD targets students at or below the poverty line, it shows how deeply disadvantaged students may respond differently from the more advantaged populations studied at selective universities. For these students, the substitution of loans with grants may reduce debt aversion and provide the flexibility to reconsider major choices without fear of future repayment burdens.

Overall, the paper provides evidence that financial aid design influences more than just access and completion. By focusing on aid composition rather than total amounts, this study opens a new line of inquiry into how the structure of aid packages shapes not just whether students graduate, but how they navigate their academic and career trajectories.