I.e. a total of $85,000 or so per year. If we subtract about $16,000 that Duke lists as non-tuition expenses in its cost of attendance, then the amount is about $69,000 per year (Duke’s tuition is $47,488).
On the other hand, http://www.startribune.com/local/157976625.html says that the University of Minnesota spends $9,625 to $16,049 per undergraduate student (average $12,055), $23,454 per graduate student, and $39,022 per professional school student, per year.
Even if we allocate the separately listed research money proportionally (i.e. adding the $870 million research cost to the $1.1 billion that was used to get the costs above), the undergraduate cost is about $22,000 and the graduate cost is about $43,000.
What accounts for such a huge variation in costs? How does Duke spend more than three times the amount per student as the University of Minnesota?
As for the difference in expenditures/student, there’s a lot of “overhead” (like the cost attributed to recruiting professors) that’s Duke’s provost included that do not directly go towards instruction. Among them, according to the story:
$8,000 for building and maintenance
$14,000 for admin & academic salaries
$7,000 for staff salaries of deans and faculties
$5,000 for miscellaneous
$20,000 for campus wide financial aid
It’s one way to look at things and I emailed the story to my H when it came out.
I think it would be moderately easy to calculate expenditures per student at a liberal arts college, and it might be interesting to see what the per capita expenditures are at different schools.
No idea how you’d do it at all at a major research institution like UCLA, where many professors rarely teach undergraduates, grad students teach some classes but also get their tuition covered and earn stipends, and various institutes and hospital facilities have almost no undergraduate influence, and yet those facilities burden the campus with a lot of costs. Allocating the systemwide administration costs is even messier. That’s a pretty ugly cost accounting problem.
I looked quickly at Amherst College, because its financial statements are easy to find and it’s about as pure an undergraduate college as there is. The last two years, expenses per student have been just shy of $100,000. That includes, however, about $26,000, in both years for things like “auxiliary activities,” “other,” and “revision to science center plans” that may or may not have a direct relationship to undergraduate education.
Swarthmore, in 2012-2013, claimed $84,300 cost per student.
Much more important, you really cannot subtract those overhead/General and Administrative expenses (as your financial manipulations do); they are both fundamental and crucial to every universities' operations (obviously, Duke included).
Post #1 provides an excellent summary of those VITAL overhead/G&A costs; clearly, expenses such as building maintenance, non-faculty academic salaries, other staff compensation, and aggregate non-endowment Financial Aid are critical (thank you very much @AttorneyMother for providing this).
My essential "bottom line points" are:
a. Even students who pay "full fright" receive an annually subsidy of of $20K to $25K;
b. I don't agree with your calculations and, therefore, I entirely non-concur with your 3X conclusion. However, I suggest your overall premise is incorrect: fundamentally, Duke is a market-based -- not a cost-based -- enterprise. While public higher education may necessarily be largely cost-based, later this month Duke will deny about 92 percent of its '19 RD applicants. Many to these approximately 25,000 candidates would be delighted to pay full tuition and fees, in exchange for the opportunity to attend undergraduate school at Duke (remembering, of course, that half of Duke's undergraduates receive no-repayment required, need- or merit-based FA). Obviously, annually tens-of-thousands of students (and their families) find the VALUE of a Duke education to be worth the costs incurred.
Differences in FASB and GASB indirect cost allocation can make comparability a real chore.
My personal opinion is that pricing based on instructional function costs as currently applied effectively subsidizes research at most research universities. This is based on the difference I observe in how faculty allocate their time vs. how personnel costs are allocated in the budget. There is inherent difficulty in society’s overlapping values for the higher ed system.
Professor X spends a hundred hours analyzing a pre-publication professional journal article and conducting concomitant research. He’s not an hourly employee and he does not have to charge his time to specific, detailed cost accounts. However, were he required to do so, there is no way X could even approximately estimate – in the LONG TERM – how much of that hundred hours would/might be instructional and how much would/might be research. Even more difficult, X’s administrative assistant’s compensation – and overhead/G&A expenses – should be directly tied to X’s work. Therefore, if X’s efforts cannot be accurately allocated, obviously the admin assistant’s work also can’t be. The same problem exists for ALL overhead/G&A costs . . . to illustrate ludicrously, the custodian in X’s building would also have to divide his efforts between “instruction” and “research.” This is why aggregate overhead/G&A accounts/pools exist.
In her model, highly selective colleges–like Duke–are investing ever greater sums in the education of the most able students. This investment pays off for the colleges, not so much in tuition, but in the long-term contributions by alumni.
^ ^ ^
That’s likely, and not overlooked by universities’ leadership. However, I’d respectfully suggest there is also a MAJOR national – and global – societal payoff.
Hoxby’s point is well taken, but I think her measured ‘student oriented resources’ are too coarse and too deeply grounded in institutional values to deliver much meaning for either individual students or society. The benefits from these supposed subsidies are hard to quantify and vary widely from student to student.
Duke’s S:F ratio is 7:1; its average FT faculty salary is $118K. So, Duke’s expenditure on FT faculty salaries is $16,857 per student.
Minnesota’s S:F ratio is 17:1; its average FT faculty salary is $93K. So, Minnesota’s expenditure on FT faculty salaries is only $5,471 per student.
It appears that Duke spends just over 3X as much per student on full-time faculty salaries as Minnesota does. This closely reflects (although it does not nearly account for) the overall cost-per-student difference that ucbalumnus estimated above. I would think the difference in financial aid accounts for another large chunk of the overall difference.
“How does Duke spend more than three times the amount per student as the University of Minnesota?”
A Yugo will get you to your destination in the exact same time as a Mercedes. But a Yugo and a Mercedes are not the same product, price or customer experience.
Why does Duke spend so much more money? Because it can.
Giant universities do benefit from an economy of scale to an extent (this is even more true when comparing RU’s to LACs). At giant publics, that means many opportunities but (unless you are in special programs), little hand-holding; it’s a more sink-or-swim setting with a big bureaucracy that often could work against you as for you. A kid who is enterprising and driven can definitely get much more than their money’s worth from such a school. Obviously, many kids don’t.
In 2013-14 (per CDS data) Duke spent $111,321,829 on need-based scholarship/grant aid for 6726 full-time undergraduates, of whom 2620 received n-b s/g aid. So on average, Duke spent $16,551 per FT undergraduate or $42,489 per recipient on n-b s/g aid.
In 2013-14 (per CDS data) Minnesota spent $113,108,036 on n-b scholarship/grant aid for 28332 full-time undergraduates, of whom 12306 received s/g aid. So on average, Minnesota spent $3,992 per FT undergraduate or $9,191 per recipient on n-b s/g aid.
Apparently, Duke spent more than 4X what Minnesota spent on need-based scholarship/grant aid per full time undergraduate student. Duke spends about $13K per undergraduate more than Minnesota on need-based s/g aid, in addition to about $11K more per undergraduate on FT faculty salaries.
If the total difference in expenditures is about $47K ($69K - $22K per ucbalumnus), then these two expenditures alone (faculty salaries and need-based aid) would account for more than half the total difference in expenditures per student.
Does FA not affect educational quality? Good FA allows a college to enroll the best applicants from all over the country somewhat regardless of ability to pay. It allows a college to enroll a more socioeconomically diverse class of students than it otherwise would. So I would expect FA to have a greater impact on the learning environment than some other things that seem to be driving up expenditures these days (such as bigger administrative staffs or fancier student amenities.)
On the other hand, a college may be able to get a more socioeconomically diverse class of students just by keeping sticker prices low. So yes, it would be interesting to adjust for the effects of aid and then compare the remaining differences in spending.
@ucbalumnus (re your initial pst to this tread): What accounts for such a huge variation in costs? How does Duke spend more than three times the amount per student as the University of Minnesota?
I presume the foregoing analyses answer your questions? I wonder if faculty and student recruiting isn’t enhanced by the 3X and 4X greater allocations for salaries and undergarduate financial aid (respectively)? Generally in life, one gets what one pays for.
