The full price of a graduate has never been published

Every piece of the number is already published somewhere

The spending starts before the first application arrives. UPCEA and Search Influence put the mean marketing cost of one enrolled student at $2,849 and the median at $1,173, with undergraduates at $1,505 and graduate students at $3,804. Their 2024 survey carries a vendor label, meaning it was run by organizations that sell in the market they measured, and its figures generally exclude staff salaries. Ruffalo Noel Levitz (RNL), in a 2023 benchmark carrying the same label, reported $4,825 to enroll one graduate business student with admissions salaries included, so the two studies price different baskets and resist direct comparison.

NACUBO, the association of college and university business officers, put the first-time undergraduate discount rate at private nonprofit institutions at an estimated 57.1 percent in its 2025 Tuition Discounting Study; the discount rate is the share of sticker price given back as institutional aid. On the public side, SHEEO, the State Higher Education Executive Officers association, publishes the State Higher Education Finance report, SHEF. The fiscal 2025 edition counts $19,443 in total education revenue for each full-time-equivalent student, the unit that converts part-time enrollment into full-time terms. Net tuition, meaning the tuition an institution keeps after its own aid, ran $10,505 per full-time-equivalent student in the four-year sector. In its 2012 analysis, the Delta Cost Project put education and related spending, meaning instruction, student services and a prorated share of administration and operations, at $55,800 per completed degree or certificate.

Each figure is published, and each carries its own denominator. Marketing cost is per enrolled student. Revenue is per full-time-equivalent student. Teaching cost is per completion. Combining them requires records that follow one student from first inquiry to diploma, and the education research record contains no study that has done it.

The components span five sources, three denominators, two salary treatments and publication years from 2012 to 2025.

The closest published figure prices the teaching and the services around it

The Delta Cost Project, housed at the American Institutes for Research, computes cost per completion: total spending on education and related activities divided by the number of degrees and certificates awarded. Its own definition sets the boundary. Education and related spending covers the core academic mission and excludes sponsored research, public service and auxiliary enterprises such as housing. Recruitment spending is typically embedded within the student-services line, in Delta's own words, and the measure never separates it out or prices acquisition. Because the measure counts spending, the tuition discount that enrolled the student appears nowhere inside it. Delta's definition also states that the measure reflects institutional spending and says nothing about what students pay to attend.

In fiscal 2013 the figure ran from $40,600 at community colleges to $115,800 at private research universities, with public bachelor's institutions at $66,800 and public research universities at $65,500.

Inside the average completion, $12,800 paid for credits that never became a credential

Producing the average degree or certificate cost $55,800, in the Delta Cost Project's 2012 accounting, which followed the federal cohort that began college in 2003-04 and matched its six-year outcomes to institutional spending from 2008-09. Spending on the students who finished accounts for $43,000 of that. The remaining $12,800 paid for credits earned by students who left before any credential, which works out to 23 cents of every completion dollar, a share computed for this piece from Delta's published pair.

At public two-year colleges, 48.8 percent of student credit hours attached to no degree or certificate, and 32.7 percent of education and related spending went with them. The spending share was 12.9 percent at public four-year institutions and 9.4 percent at private four-year institutions.

Across the whole cohort, 35 percent had left with no credential and were no longer enrolled six years later.

The average completed degree carried $43,000 in institutional spending, against $18,000 for the average student who left. Leaving late narrows that gap almost to nothing: students who departed after more than 36 months were 9 percent of dropouts and 22 percent of attrition spending. Students who left after more than two years carried roughly $34,000 each, and those who left after four years about $47,000.

A 20 percent reduction in attrition would cut spending per credential by $2,500 and add about 260,000 credentials a year, an increase of 6 percent.

Dividing by graduates turns $26,203 into $118,382

Cost per student divides spending by everyone enrolled. Cost per graduate divides the same spending by the students who finish. The cleanest demonstration in the record comes from the control group of a randomized trial. The nonprofit policy research organization MDRC ran the cost analysis of CUNY's Accelerated Study in Associate Programs, following 896 low-income students, randomized in 2010 across three campuses, all with developmental education needs, meaning placement into remedial coursework below college level. Students in the control group received the colleges' regular services. Over three years the colleges spent $26,203 per control-group student. In that window 22.1 percent of the group earned an associate degree. The study's cost per degree for the group is $118,382, or 4.5 times the per-student spending.

Across the published record, cost per graduate runs three to four times cost per student. The gap is the spending on students who left without a degree, which stays in the numerator while the graduation rate sets the denominator at 22.1 percent.

The education research index holds no study that connects the pieces

On August 21, 2026, queries ran against ERIC, the federal government's public index of education research. The query "cost per graduate" returned 24 records, and none of them connects recruitment spending to completion. The nearest matches are a 1993 study of one Florida community college, which priced instructional cost per associate graduate at $10,248 and $13,888 with no recruitment spending and no discount, and a 2018 peer-reviewed benefit-cost analysis of a CUNY program, which computed program cost per degree only. The query "cost per completion" returned 2 records, both instructional-productivity analyses. The query "student acquisition cost" or "cost to recruit" returned 4 records, and the only documents among them that compute a cost to recruit are vendor benchmarking reports. The query that combined tuition discounting, attrition and recruitment returned 0 records.

The office that would match the records has fewer than five people

A president leaves a board meeting and asks the director of institutional research what a graduate costs, all in, from first contact to diploma. The director has published material within reach. The recruitment benchmarks are two vendor studies with different salary treatments. The discount rate is in NACUBO's annual study, and the institution's own version is in the aid ledger. The completion cost is in Delta Cost Project tables built from the Integrated Postsecondary Education Data System, the federal data collection every college files into; her office submits those filings every year. No document on the desk carries the joined number. Producing it in-house means lining up admissions spending, aid records and registrar completion records for each student, one at a time.

In the Association for Institutional Research's national survey, covering 1,261 offices as they existed in 2015, 80 percent of four-year institutional research offices and 87 percent of two-year offices reported professional staffs of fewer than five.

The same association, with EDUCAUSE and NACUBO, issued a joint statement on analytics in 2019, updated in 2022, warning that data silos keep analytics from solving institutional problems.

The recruitment benchmarks, the discount study and the teaching tables are in print, each with its own denominator and year. The closest available answer to the president's question remains Delta's: $55,800 for the average completion, of which the credits that never counted toward a credential cost $12,800.

Methodology

What is published and what is computed here.

• All dollar figures, percentages and counts are carried as published by the sources above, with two computed exceptions, both labeled in the text and figures. First: $12,800 divided by $55,800 is 22.9 percent, rounded in the text to 23 cents of every completion dollar. Second: $118,382 divided by $26,203 is 4.52, rounded in the text to 4.5 times.
• The $118,382 is MDRC's published figure. A reader who divides the rounded published inputs ($26,203 by 22.1 percent) gets $118,566; the study's own figure reflects unrounded internal values. The piece carries the published figure.
• The piece computes no joined total across funnel components. The components use different denominators (per enrolled student, per full-time-equivalent student, per completion), different sectors and different years, and joining them defensibly requires student-level records.
• "Cost per graduate runs three to four times cost per student" is carried from the research corpus's characterization of the published record; the control-group demonstration in this piece runs 4.5 times.

The ERIC search, documented.

All queries ran against the ERIC database API on August 21, 2026.

• "cost per graduate": 24 records. None connects recruitment spending to completion. Nearest: Brann 1993 (instructional cost per AA/AS graduate at one Florida community college, $10,248 and $13,888, no recruitment or discount); Levin and Garcia 2018 (peer-reviewed ASAP benefit-cost analysis, program cost per degree only).
• "cost per completion": 2 records (Jenkins and Belfield 2014; Jenkins and Rodriguez 2013), both instructional-productivity analyses without recruitment spending or discounting.
• "student acquisition cost" OR "cost to recruit": 4 records; the only documents computing an acquisition cost are the RNL vendor benchmarking reports (ED608071, ED623769).
• "tuition discounting" AND attrition AND recruitment: 0 records.

Supplementary checks the same session: a Third Way site search for "cost per degree" surfaced no publication computing an institutional cost per graduate; general web searches surfaced Delta, Complete College America, RNL and aggregator blogs, and no full-funnel study.

The Google Scholar check.

The check ran on August 22, 2026, and found no such study. Google Scholar returned no articles for "cost per graduate" searched together with recruitment, attrition and the tuition discount, and no articles for "cost per graduate" with attrition and "recruitment spending". The broadest neighboring query, "cost per graduate" plus recruitment, returned about 677 results, led by military training-cost models, cost analyses of open universities abroad and the CUNY ASAP studies already cited here. Companion searches of the OpenAlex and Crossref scholarly indexes the same day surfaced no study joining recruitment spending, the tuition discount and attrition into one cost per graduate. The closest item found anywhere is a 2021 single-college dissertation that prices new-student matriculation through activity-based costing and stops at enrollment.

Exclusions and cautions carried from the factbase.

• No EDUCAUSE Core Data Service figure appears in this piece. The EDUCAUSE sites blocked automated fetching during verification, and the pipeline rule is that no CDS number ships without hand verification.
• Figures excluded because they trace to no primary source: the $5,460 student acquisition or replacement cost (stated in its source report with no citation); "colleges spend up to one-third of institutional costs on students who will never graduate" (the traceable version is Delta's sector-specific shares, cited here); "it is 3 to 5 times cheaper to retain a student than to recruit one" (no primary source exists); the $10,000 to $21,000 per-dropout replacement figures (untraceable vendor circulation).
• Vendor caveats: the RNL 2023 figure rests on 32 of more than 200 respondents and includes prorated admissions salaries; the UPCEA 2024 figures generally exclude salaries. The two are never presented as comparable.
• Vintages: Delta's 2012 attrition figures come from the BPS:04/09 cohort (2003-04 first-time beginners followed six years) joined with IPEDS 2008-09 expenditures; the sector table is fiscal 2013 in 2013 dollars; NACUBO 2025 covers private nonprofit institutions; SHEEO figures are FY2025, and the $10,505 net tuition figure is the four-year public sector, with all-sector companions differing.
• The scene in the final section is role-shaped and generic by design; every artifact named in it (the vendor benchmarks, the NACUBO study, the Delta tables, IPEDS filings) is a real published document or filing, and no institution-specific fact is asserted.