The high cost of funded research in colleges and universities

A college or university that does research ends up spending considerable resources of its own even when most of its research is “funded”. How and why is this the case, and where does the institution look to find the resources needed to cover this unfunded research cost? Undergraduate tuition seems like one likely source.

 

These internal research expenditures fall into two categories, which I will call “open” and “hidden”. As these terms may suggest, the first is a set of costs that are reported nationally by the NSF and consequently appear in numerous reports put out by the institutions themselves. On the other hand, hidden costs are well known, but seldom openly discussed even thought they contribute very significantly to the institutional cost of research .

OPEN COSTS

Funding for research in higher education comes from many sources as described in the NSF Higher Education Research and Development (HERD) Survey for 2015:

Higher Education R&D Expenditures by source (in Millions of current dollars )

	Federal	State	Business	Nonprofit	All other	TOTAL EXTERNAL	Institutional	TOTAL RESEARCH
2010	37,475	3,852	3,198	3,740	1,048	49,313	11,940	61,253
2011	40,767	3,829	3,179	3,854	1,037	52,666	12,610	65,276
2012	40,140	3,695	3,271	4,023	968	52,097	13,633	65,730
2013	39,445	3,653	3,506	3,889	1,537	52,030	14,985	67,015
2014	37,923	3,869	3,725	3,980	1,905	51,402	15,754	67,156
4 year % change	1.20%	0.44%	16.48%	6.42%	81.77%	4.24%	31.94%	9.64%

These data include medical schools, which contribute about $23B to the Total Research column in 2014. Medical schools are organized and financed very differently from the rest of the university; conclusions drawn from these overall numbers should therefore be understood with that caveat.

“Institutional” is the sum spent by universities and colleges from their own resources on research. Note that while the Federal support has oscillated somewhat but overall flat, the Institutional component has grown steadily as a percentage of the Federal:

 

Institutional as a % of Federal
2010	31.86%
2011	30.93%
2012	33.96%
2013	37.99%
2014	41.54%

 

That is, by 2014 the universities contribution to research from their own resources was almost 42% of what the Federal government contributed.

A more complete picture of the institutional cost of “externally funded” research is:

 

Institutional as % of External
2010	24.21%
2011	23.94%
2012	26.17%
2013	28.80%
2014	30.65%

Or, for every $1,000 an institution gets from external funders for research, on average the institution ends up spending $306 of its own internal resources.

Obviously, the institutional cost of externally funded research is growing steadily over time. As the HERD survey shows, the total increase in Institutional support from internal funds over this 4 year period was about $3.8 B, or almost 32%. However, most of the components of institutional support arguably are necessary to enable the institution to attract the externally funded research - a cost of doing business.

The 2015 HERD survey breaks down the components of Institutional support over time: 

 

Totals for the numbers in this graph are slightly different from those shown in the table above. The table above included data from all 891 institutions that had research expenditures of more than $150K; this graph is based on data from the 645 institutions that had research expenditures of more than $1M.

“Direct funding for R&D” includes both University Research and (most of) Departmental Research contributions. “Unrecovered indirect costs” describe shortfalls in indirect cost recovery occurring when research grants are accepted that provide indirect costs less than the federally negotiated rates. Since the federal rate is obtained by dividing all allowed indirect research expenditures by the total organized research expenditures, this shortfall describes unreimbursed real expenditures which must be covered by university funds. This term does NOT refer to costs the universities felt were real but which the federal negotiators decided to reject. HERD data instructions are clear:”R&D does not include unrecovered indirect costs that exceed your institution’s federally negotiated Facilities and Administration (F&A) rate”. Thus this number underestimates what most university CFO’s would say is the real shortfall in recovery.

Just to keep definitions clear, OMB Circular A21 defines University Research as university funded research that is awarded similarly to federal research, i.e. there is a specifically budgeted university program that solicits proposals which describe a research program and its desired outcome, are reviewed by some (usually) internal panel, and awarded to some fraction of applicants. At the end of the grant period, financial and programmatic reports are prepared. A21 defines Departmental Research as all other internally funded research programs. Thus Departmental Research covers cost sharing (separately broken out in the table above); start-up packages, bridge funding, and seed funding for faculty; tuition aid to students working on sponsored projects; funds for visiting researchers; and everything else related to research that is not funded elsewhere.

Most elements of Departmental Research, e.g cost sharing and start-up packages for new faculty, are the “seed corn” of funded research. As such, very prudent trimming is possible, but significant cut backs are likely to lead over time to drops in external research funding.

Whatever the value of institutional Direct Funding of R&D, it is obviously the driver of the rapidly growing overall cost of Institutional R&D. Over the 4 years covered by the report, Direct Funding grew by an impressive 54.8%.

HIDDEN COSTS

The HERD instructions emphasizes by exclusion what is certainly the largest component of the hidden institutional cost of research at research universities: “R&D does not include estimates of the time budgeted for instruction that is spent on research.” In other words, course release for successful researchers cannot be charged to R&D (thus showing up in these data), but rather sits unobtrusively on the university budget as part of instructional personnel costs.

Due to the vagaries of university budgeting, it is difficult to get a hard number that captures this cost, but one can make a hand-waving estimate to get the order of magnitude of this type of cost of at research universities. In general, a reasonable estimate is that tenure line faculty at research universities are expected to spend during the academic year roughly 50% time for teaching, 50% time for research. Thus roughly ½ of the tenure-line faculty salary costs actually are attributable to research. A mid-size research university has of the order of $400M/year in “faculty salary”. If we make a conservative estimate that half of that salary is for tenure line faculty, the rest for non-tenure line faculty, then the research cost hidden in tenure line faculty salary at this mythical university is $100M/year. There are about 200 research universities in the US, so the total research component of faculty salary is of the order of $20B/year – of the order of magnitude of the Institutional R&D expenditures.

Adding this salary component to the numbers above for Institutionally Funded R&D Expenditures suggests that universities and colleges are spending more than $30B per year of their own funds on research, distinct from and in addition to the slightly more than $51B in external research funds that they expend.

This poses the obvious question: where do the colleges and universities get the $30+B per year that they spend on research that is not coming from research funding from government, industrial, or foundation sources?

HOW TO PAY THE INSTITUTIONAL COSTS OF RESEARCH?

Philanthropy and endowment income play a major role in many institutions in providing some portion of the needed research funds. However, in almost all institutions the greatest source of fungible revenue is tuition (or state support in lieu of tuition). What better place to look for support of research?

A21 opened the door to this coupling of research and education by defining (“for the purposes of this document”) Departmental Research to be a component of educational spending. Of course, A21 was focused on separating externally funded research from everything else in the college budget in making that definition, not making a judgment about the educational role of Department Research . However,in 2002, the National Association of College and University Business Officers (NACUBO) picked up on this A21 definition and officially defined Departmental Research as being a necessary educational expense in their calculation of  a greatly inflated “cost of delivering undergraduate education” to be used to placate parents and legislators upset about rapidly rising tuition.  This definition allowed Department Research to be moved into the Education budgetary category along with faculty salaries. The combination deftly intermixed research costs (both Departmental Research and faculty release time) and actual educational costs and made them one – and used this as a base to calculate a” cost of  undergraduate education”. Thus rationale for moving student tuition dollars to cover costs of faculty research was officially established.

IS THIS REALLY THE APPROPRIATE WAY TO PAY FOR RESEARCH?

But can faculty research costs be separated from educational costs? I would argue that the answer to that is a strong " that depends." The institutional costs of research described above are in most cases the cost of attracting and retaining research-active faculty. So the underlying question really is whether this set of research active faculty provides a "value-added" to the educational process that justifies mingling the costs.

For graduate doctoral education, research-active faculty are essential, and their research forms a core component of the graduate education process. Here, differentiation between research costs and educational costs essentially is meaningless. Once one leaves doctoral education, however, answers become less well defined.

Regarding undergraduate education, NACUBO stated: 

Departmental research is vital and has a direct impact on the value and quality of instruction provided to students. Any arbitrary attempt to distinguish between departmental research and instruction ignores the fact that the integration of research and education is a major strength of the nation’s colleges and universities and directly benefits undergraduates

A strong statement indeed, but one not strongly supported by data. To be honest, we in higher education have a very strong aversion to studying our outcomes, so there are few data that strongly contradict that statement as well. What this is, therefore, is an aspirational statement describing what we strongly hope is a truth, but very well may be incorrect.

What data do show clearly, is that student learning is tied more closely to the teaching methods used than to research or seniority characteristics of faculty. Unfortunately, faculty reward systems generally reward research over teaching, which leads to a situation in which many research-active faculty find little reward in learning new and better teaching approaches. Beyond this, however, is the strongly held belief that faculty who do not do research somehow become "stale" in their teaching, while research faculty do not. I know of no studies that actually address this issue, but based on long observation would guess that it is untrue for most undergraduate required courses, possibly true for some advanced specialized undergraduate courses.  In any case, lack of data again. All in all, I think it is fair to say that it is not possible at present to make a convincing, data-based argument that use of research active faculty generally leads to better undergraduate student learning outcomes than are obtained by using an equivalently degreed but not research focused faculty. (Research does strongly suggest that there are learning benefits obtained with a faculty that is full-time versus part- time, but that is a separate organizational issue.)

In Cost allocation in the research university and what it tells us, I discussed several other direct and indirect ways in which faculty research may benefit undergraduate students. For example, institutional reputation is driven primarily by faculty research prowess rather than educational effectiveness. Some analyses suggest that graduates of high reputation institutions "do better" in life - although there are also analyses suggesting that is not the case.

 At the end of the day, it is impossible to tell how much undergraduate tuition actually is used to cover faculty research costs. What we do not have in higher education is cost accounting - which would tell us the cost of the various functions of higher education. Thus it is impossible to make a judgement of how justified any amount of such cost shifting is - we can't weigh costs and benefits. What is clear is that there is considerable pressure on higher education to find the necessary $30+B each year to fund research programs, and that undergraduate tuition provides a very big and very tempting source - pre-justified by NACUBO.

CORRELATION IS NOT CAUSATION - BUT IT MAY MAKE YOU THINK

The above sections suggest that it is plausible that increasing research costs may be putting upward pressure on tuition levels. Going back in NSF reports, one finds that total research expenditures grew from $14,976 M in 1989 to $67,156 M in 2014 (current dollars). This corresponds to an average annual rate of growth of about 6.2%. Over that same period, inflation averaged 2.7% annually, leading to a real annual rate of growth of research expenditures of about 3.5%. It is curious that, as I showed in A return to the elephant of college pricing, published average real tuition marched upwards at roughly 3.2% a year during the period 1983-2013. So there is a correlation between increasing research costs and increasing tuition, at least over these time periods  Lots of apples and oranges there, but curious nonetheless. 

AND THE PROBLEM IS

The underlying problem is that higher education has decided to price funded research at less than its actual cost. This research underpricing arguably leads to a great deal of the upward push on tuition that has occurred over the past decades. Since reputations of universities are linked directly to their research expertise and external research funding - not their education - there is a continuing pressure to do ever-more funded research. As colleges and universities seek to move up the reputational ladder, they increase emphasis on funded research,which leads to a need for more internal research spending , which in turn increases the pressure to increase tuition.

This present relationship between education, research and reputation was set in place in a time when resources were less constrained. Over time,however, this coupling  of research, education and reputation with its internal feedback loop has produced an enormously expensive product. Very high expenditures in the name of "education" are required in order to balance the product budget, but perversely the loop leads to de-emphasis of education in favor of research. Unfortunately, these "educational" costs that are loaded with research costs have risen so high that both governments and individuals are finding it extremely difficult to pay them, and are rebelling and demanding to know if the outcome justifies the price. And justifying the price when the outcome actually includes a significant purposely hidden component of faculty research is particularly difficult. At the very least, justification will require some actual studies that relate educational outcomes to a wide variety of variables including research status of faculty - studies we have thus far generally tried to avoid.  

It may be time to come clean, and publicly acknowledge that the costs of doing funded research are much higher than the price payed by the sponsors. Parents, students, and governments deserve to know that educational dollars are being spent on effective education, and that faculty are being rewarded primarily for the quality of learning that goes on in their classes.  A wide ranging discussion of research funding, how large it should be and how it is allocated must be part of that process. In several countries, such discussions have led to decisions to funnel research funding to a smaller number of institutions in order to maximize research impact. Methods of determining educational effectiveness must be developed, so that reputations of most colleges and universities can be based on student learning rather than faculty research.

What we have is an enormously successful system of colleges and universities that has feedback mechanisms that lead in many institutions to ever increasing costs and ever sharper focus on high quality research at the expense of other parts of the mission. Eventually such a system will run out of resources and will not meet societal expectations based on its entire mission. One can argue that we have reached that point in many parts of our system of higher education. 

 

 

 

 

Comments

[Ed Hackett]

Understood, Lloyd, and agreed: because so much has been added atop research responsibilities, I am thinking things are even worse than depicted in your article. Best wishes, Ed

[Ed Hackett]

Provocative and sensible argument...have you considered splicing in the argument about increasing amounts of research time required for administrative responsibilities and the increasing numbers of proposals submitted (requiring preparation, review, and often panel review) caused by lowered success rates (and higher research and research funding expectations)? These things then add to the cost (in faculty time) of doing research, and the whole the weighs on the university budget, as you describe. Then take the whole and go visit Dan Sarewitz's recent
article in The New Atlantis titled "Saving Science." I guess I am less optimistic than Dan: my thinking along these lines led to a talk titled "Is Science Possible?"

Best wishes,

Ed
Lloyd responds: thanks, Ed, for the good input. No doubt researchers are forced to devote evermore time to preparing proposals, etc..However, the arguments in this post are based on how much of a faculty member's time is spent on research - not how effectively that time is spent. Your good point that much of the research time is spent on non-productive efforts just makes it even more inappropriate that significant amounts of "educational" funding in used to subsidize those efforts. Thanks for pointing me to Sarewitz's very interesting article. He takes Stokes' Pasteur's Quadrant a needed extra step.