Debt burden, aid, and financial resources
College affordability data lives in several different places.
The College Scorecard tells us about student debt, loan payments, earnings, and net price. The Common Data Set tells us how colleges use non-need merit aid. IPEDS tells us about institutional finances, including endowment and instructional spending.
This page joins those sources to look at them together.
The starting point is debt burden: the share of median earnings that would go toward federal student-loan payments each year. We then calculate an alignment gap for each school: roughly how much less debt its graduates would need to carry, expressed per year of college, to reach the median debt burden in this group of schools.
The gap is not a recommended tuition price or a claim that lowering price by exactly that amount would produce the same reduction in borrowing. It is a common yardstick that lets us compare the size of the debt burden with other financial measures from the college.
The vertical axis shows the alignment gap in dollars per year. The horizontal axis shows estimated non-need merit aid per full-time first-year student, using the Common Data Set. Because most values are spread over a wide range, the axis uses a log scale. Schools reporting no non-need merit aid are shown separately at $0.
The diagonal line is where the two amounts are equal. Schools to the lower-right of the line report more non-need merit aid per first-year student than the size of their alignment gap. Schools to the upper-left report less. Color shows endowment per undergraduate, adding another measure of the institution's financial resources.
These amounts are useful for comparison, but they are not interchangeable dollar for dollar. Merit aid is generally a tuition discount rather than a cash expenditure, and changing a college's aid policy would not necessarily produce an equal change in student borrowing. Hover over any dot to see the school.
| School | Gap/yr | Merit spend / first-year | Merit share | Avg merit grant | Net price |
|---|---|---|---|---|---|
| Quincy University | $1,646 | $27,587 | 100% | $27,587 | $20,359 |
| DePauw University | $653 | $17,132 | 43% | $39,482 | $22,264 |
| Wabash College | $703 | $14,368 | 37% | $39,105 | $24,336 |
| Beloit College | $1,830 | $13,476 | 29% | $45,857 | $21,526 |
| Gettysburg College | $567 | $12,082 | 31% | $38,359 | $31,490 |
| The College of Wooster | $1,470 | $11,435 | 29% | $39,386 | $23,458 |
| Pratt Institute | $1,806 | $10,083 | 46% | $21,747 | $52,659 |
| School | Gap/yr | Merit spend / first-year |
|---|---|---|
| University of the Incarnate Word | $1,845 | $0 |
| Kentucky State University | $3,339 | $330 |
| Hollins University | $3,286 | $2,973 |
| Baker College | $3,152 | $932 |
| North Carolina A & T State University | $2,908 | $729 |
| Point Park University | $2,786 | $2,474 |
Now add financial resources
The merit-aid comparison only works for schools with usable CDS H2A data. Many colleges do not publish that field consistently.
A second view uses a larger group of 375 schools and asks a different question: how does graduate debt burden compare with the financial resources of the institution? Here we plot the same alignment gap against endowment per undergraduate.
The vertical axis again shows the alignment gap. A value above zero means the school's debt burden is above the 4.40% median. A value below zero means it is below the median. The horizontal axis shows endowment per undergraduate on a log scale. The vertical divider marks the sample median of about $64,000 per undergraduate.
Color shows instructional spending per student as a share of average net price. This adds some context about the relationship between what students pay and what the institution reports spending on instruction. Endowment per student is only a broad measure of financial capacity. Endowments contain restricted funds, and two colleges with the same endowment per student may have very different obligations and spending policies. The chart is intended to show financial context, not available cash. Hover over any dot to see the school.
An example: Bard and Grinnell
Bard and Grinnell illustrate why putting these datasets together can be useful.
Their published median SAT scores are similar: 1510 for Bard and 1490 for Grinnell. But the financial picture in the joined data is very different.
Grinnell's average net price is $17,648, compared with $34,649at Bard. Grinnell's graduate debt burden is 3.5%, compared with 6.6% at Bard. And Grinnell reports about $1.54 million in endowment per undergraduate, compared with about $69,500 at Bard.
Those numbers do not make Bard and Grinnell equivalent colleges, nor do they prove that the difference in endowment caused the difference in price or debt. Admissions alone make that clear: Grinnell admitted 1,416 of 9,758 applicants in 2024–25, or 14.5%.
The comparison simply shows what becomes visible when data that normally sits in separate systems is placed side by side. Two colleges can enroll students with similar published test profiles while having very different prices, graduate debt burdens, and institutional resources.
Neither Bard nor Grinnell appears in the first chart because neither has a usable H2A merit-aid row in this dataset. That is why the second chart uses the broader endowment comparison.
For a student or family, that is the practical point of the exercise: sticker price, financial aid, debt, earnings, and a college's financial resources are usually presented separately. Looking at them together gives a more complete picture.
How the alignment gap is calculated
The calculation starts with debt burden.
debt burden = median monthly federal loan payment × 12 ÷ median earnings 10 years after enrollment
A school with annual loan payments of $3,000 and median earnings of $60,000 would therefore have a debt burden of 5%.
The median debt burden in the 375-school comparison on this page is 4.40%.
For schools above that level, the alignment gap estimates how much lower median completer debt would need to be for the school to reach a 4.40% burden at its existing earnings level. We divide that amount by four so it can be read as an annual figure.
alignment gap = completer debt × (1 − median burden ÷ school burden) ÷ 4
A positive gap means debt burden is above the sample median. A zero or negative gap means it is at or below the median.
Again, this is a comparison measure. It should not be read as a prediction that reducing annual net price by the same number would reduce student debt one-for-one.
For the first chart, we also estimate non-need merit aid per full-time first-year student:
merit aid per first-year student = share receiving non-need merit aid × average non-need merit grant
For example, if 40% of first-year students receive non-need merit aid and the average grant among recipients is $20,000, the measure used here is $8,000 per first-year student.
Both charts use the same 4.40% median debt burden so a school does not have a different benchmark depending on which chart you are viewing. The smaller merit-aid sample has a very similar median of 4.50%.
What is being joined
Panel A combines three sources. College Scorecard supplies debt, loan payments, earnings, and net price. Common Data Set H2A supplies non-need merit-aid information for full-time first-year students. IPEDS supplies endowment information.
After data-quality checks, 223 schools have enough usable information for this view.
Panel B does not require H2A merit-aid data, so it can include many schools that are missing from Panel A. It uses College Scorecard debt, earnings, and net price; IPEDS endowment and instructional spending; and a recent CDS SAT midpoint to define the 375-school comparison group.
The two samples overlap, but neither is simply a subset of the other.
Data limits
This page combines institutional data from different systems and different years, so the numbers should not be read as if they describe one group of students moving through college at the same time.
The College Scorecard figures used here are from 2022-23. Earnings data describe federally aided students from an earlier cohort, roughly a decade after they first enrolled. The CDS data are from 2024–25 or 2025–26. The join is by institution, not by individual student or graduating class.
CDS H2A also has limits. It covers full-time first-year students and specifically reports non-need aid. Some awards that combine need and merit may not appear in the measure. It is best understood as a view of one part of a school's recruiting and tuition-discount strategy, not its total financial-aid budget.
Federal student debt has limits as a measure of affordability as well. Median debt in this 375-school sample ranges from $10,000 to $30,964, and $27,000—the federal aggregate borrowing limit for many dependent undergraduates—is the most common value in the sample. Families may pay college costs with income, savings, parent borrowing, private loans, grants, or other resources that are not captured by this debt number.
That is why this page does not use federal debt as a substitute for price. Net price remains the price measure. Debt burden tells us something different: how large federal loan payments are relative to later earnings.
Data-quality checks
Panel A begins with 488 possible merit-profile rows. We keep 223 after requiring usable CDS merit-aid data and the Scorecard and IPEDS fields needed for the comparison.
Rows are removed when the merit profile is too incomplete, required values are missing, or published values fall outside reasonable ranges. That includes 3 reported merit-aid shares outside 0–100% and 1 average grant above $80,000. A legitimate published value of $0 is kept.
5 schools in the final sample report no non-need merit aid. Four already have debt burden at or below the median. The University of the Incarnate Word has a positive alignment gap.
The checks matter because joining datasets can make a bad source value look much more meaningful than it really is. The filters used for this page are part of the reproducible recipe rather than manual exclusions made after looking at the chart.
Pull the data yourself
The underlying data and calculations are public. The queries below reproduce the source rows used for the two panels. Panel B contains more rows than the API's 1,000-row response limit, so retrieve it in pages.
# Panel A — CDS H2A merit × Scorecard (488 rows; under the 1,000-row cap) curl 'https://api.collegedata.fyi/rest/v1/school_merit_profile?select=school_id,school_name,canonical_year,merit_profile_quality,non_need_aid_share_first_year_ft,avg_non_need_grant_first_year_ft,avg_need_grant_first_year_ft,earnings_10yr_median,median_debt_completers,median_debt_monthly_payment,avg_net_price&limit=1000' \ -H 'apikey: <anon key>' \ -H 'Authorization: Bearer <anon key>' # Panel B — Scorecard + IPEDS endowment (2,158 matching rows) # PostgREST max-rows is 1,000. limit=5000 and Range: 0-4999 are both capped # at 1,000 with HTTP 206 / Content-Range: 0-999/2158 — no error body. # Page with offset (or Range: 0-999, then 1000-1999, then 2000-2999): curl 'https://api.collegedata.fyi/rest/v1/scorecard_summary?select=ipeds_id,earnings_10yr_median,median_debt_monthly_payment,median_debt_completers,avg_net_price,endowment_end,instructional_expenditure_fte,enrollment&earnings_10yr_median=gt.0&median_debt_monthly_payment=gt.0&avg_net_price=gt.0&endowment_end=gt.0&limit=1000&offset=0' \ -H 'apikey: <anon key>' \ -H 'Authorization: Bearer <anon key>' curl 'https://api.collegedata.fyi/rest/v1/scorecard_summary?select=ipeds_id,earnings_10yr_median,median_debt_monthly_payment,median_debt_completers,avg_net_price,endowment_end,instructional_expenditure_fte,enrollment&earnings_10yr_median=gt.0&median_debt_monthly_payment=gt.0&avg_net_price=gt.0&endowment_end=gt.0&limit=1000&offset=1000' \ -H 'apikey: <anon key>' \ -H 'Authorization: Bearer <anon key>' curl 'https://api.collegedata.fyi/rest/v1/scorecard_summary?select=ipeds_id,earnings_10yr_median,median_debt_monthly_payment,median_debt_completers,avg_net_price,endowment_end,instructional_expenditure_fte,enrollment&earnings_10yr_median=gt.0&median_debt_monthly_payment=gt.0&avg_net_price=gt.0&endowment_end=gt.0&limit=1000&offset=2000' \ -H 'apikey: <anon key>' \ -H 'Authorization: Bearer <anon key>'
For Panel A, keep non-need aid shares between 0 and 1, average non-need grants between $0 and $80,000, and merit profiles rated strong or partial. Keep non_need_aid_share_first_year_ft in [0, 1] and avg_non_need_grant_first_year_ft in [0, 80000]. A published $0 grant is kept.
Panel A joins endowment from scorecard_summary and undergraduate enrollment from school_browser_rows.undergrad_enrollment_scorecard.
Panel B joins undergraduate enrollment from institution_directory.undergraduate_enrollment, using Scorecard enrollment as a fallback, and keeps schools with a recent CDS SAT composite midpoint.
Both panels then use the same calculations:
burden = monthly payment × 12 ÷ earnings
gap = completer debt × (1 − median burden ÷ burden) ÷ 4
merit per first-year = share receiving non-need aid × average non-need grant
The anonymous API key is available on the API page. To rebuild the checked-in dataset: python3 tools/scorecard/build_alignment_gap_recipe.py