University Overhead Under Siege

March 22nd, 2013

Mallinckrodt Building, the Home to Chemistry and Chemical Biology at Harvard

There was a fascinating article in The Boston Globe this week on an attempt by the Obama administration to curtail the high rates of research funding doled out to elite universities as overhead. Overhead, formally called “indirect costs”, is typically awarded as a percentage of the funding for a research grant. The purpose of indirect costs is to compensate the institution for expenses that are difficult to ascribe to any single project—e.g., the cost of electricity, water, administration, maintenance, building/instrument depreciation, and more. Each institution negotiates its overhead rate (as a percentage) with the government, and some of the most prestigious universities have huge rates. For instance, Harvard has a rate that stands at 69%, while the national average is 52%. These funds simply enter the universities’ general budgets; once awarded, there is no requirement that they go towards funding expenses associated with research.

Last year, the Obama administration attempted to curtail high overhead rates by setting a single rate for all universities, but the largest schools were successful in protecting their lions’ share by lobbying against the proposed changes. From the article:

Demands for a fairer system have been issued for years, the inequities cited in numerous reports. Vedder and other critics contend the varying overhead reimbursement rates contribute to a widening gap between rich and poor schools. And smaller, less prestigious schools around the country are tired of being shortchanged, particularly when the prime beneficiaries are extremely wealthy schools like Harvard, which has a $30.7 billion endowment.

Bin Guo, a cancer researcher at North Dakota State University who wrote to the government in support of an overhaul, said he supports a flat rate set between 40 and 50 percent. North Dakota State receives a 44.5 percent overhead rate.

“This will eliminate the waste of federal funds at many large institutions that spend the indirect cost on nonresearch-related programs,” Guo said. “If the NIH can save this extra money, they can support more research grants.”

Harvard received $656 million last year in federal research funding, and about $175 million of that was overhead compensation. Harvard maintains that its rate is high because the type of research that is done there requires high-tech facilities and equipment, in addition to being located in a region with high labor and construction costs.

It was interesting to learn that after World War II, when the U.S. government started funding university research in earnest, the rate for overhead was 8%. It rose to 20% by 1965, when the government started allowing individual institutions to negotiate their own rates. Some schools’ rates rose past 90% before coming back down when government audits revealed inappropriate spending (like on sports tickets).

It certainly makes sense for the government to pay indirect costs, but I don’t see a good reason why a school like Harvard should get much more than another school in the same geographic area. Harvard will already secure more in indirect costs by virtue of their professors’ winning more grants. Of course, the main reason for Harvard’s sky-high rate is probably that the school paid for a better team of negotiators and budget planners. It’s another example of the rich getting richer.

When I transitioned from college (at NYU) to grad school (at Harvard), I was amazed at how much more administration there was at my new school. At Harvard, every professor had an administrative assistant (some had more than one), while at NYU, such assistants were rare. At Harvard, the department’s mail room was staffed by two people (not counting the guys in shipping). Harvard’s financial office had an army of people, and each major instrument facility (NMR, MS, X-Ray) had a dedicated staff member. At NYU, these staffers serviced multiple instruments. I was also amazed that professors at Harvard only had to teach one course per year, and when doing so, could lean heavily on an army of teaching assistants. In hindsight, one wonders how much of this excess (relative to smaller R1 schools) is necessary or worthwhile.

23 Responses to “University Overhead Under Siege”

  1. John Spevacek Says:

    Try getting a grant as a corporation. SBIR’s (Small Business Innovation Research) only allow for 17% overhead if I recall correctly.

    You know some politician is going to take this information and make political hay with it, and it won’t be looking good for future research funding.

  2. Postdoc Says:

    I don’t disagree with the general assertion that Universities skim too much off the top of grants, but I do have a beef with the specific examples you give of waste. Having done my grad work at a top 10 institution and now doing a postdoc at another top 10 university, the level of efficiency in a lab and university is astounding when effective support staff are hired. I think that if having an assistant or a dedicated instrument manager means that tasks get done more efficiently and properly, then it is worth the cost. Everyone loses time (money) when an instrument isn’t properly managed.

  3. @countthebricks Says:

    Yes in this little math problem 69% is higher than 44.5%. Harvard has much more facilities to manage than ND State. There is a higher cost of living and working in Cambridge than in Fargo, so everyone who works for those facilities requires a higher salary. Electricity, water, transportation and phone service all cost more in Cambridge than in Fargo. We are not all the same, our research is not all the same, our lifestyles are not all the same. Producing Noble Prize worthy research costs a little bit more money. Harvard has trained 75 Nobel winners in its nearly 400 years of existence. MIT has 71. Of course they are going to require a higher percentage than ND State. I hate that the government picks up 69 cents of every dollar of funding that Harvard affiliated researchers bring in, but that is the price that is paid to have the ability to focus on your work instead of “which of my grad students is going to make sure the LC/MS is running properly”.

    This is 1) capitalism and 2) simple economics. 1) I have the best product (research) and should have the most consumers 2) Demand for these services is high (tons of good research going on) so the price will be high (69% vs. 44.5%).

    This is not an example of “the rich getting richer”. This is an example of someone supplying an excellent product and someone else willing to pay top dollar for it.

  4. Paul Says:

    @Postdoc: So why shouldn’t the scientists at the tier 2 school get to enjoy this same efficiency?

  5. andre Says:

    @countthebricks: Of course Harvard also has the endowment, donors, industry connections that ND State does not. These are the benefits of being successful in your research that are dictated by economics and capitalism. They don’t need the extra overhead to fund their research, but because they have the money to pay for the lawyers and such to negotiate these deals (not to mention lobby the government for more benefits for them), it IS the rich getting richer.

    Also, with your argument, couldn’t you just as easily say that because ND State is not getting the extra money, they do not have X number of Nobel Prize winners. Would research be better at these institutions if their students didn’t have to “make sure the LC/MS is running properly”?

    Harvard gets more money because they do better research because they get money. Chicken or egg?

  6. Postdoc Says:


    That wasn’t the point I was trying to make. In fact my grad school DIDN’T have the kind of support staff Chembark was complaining about and my postdoc DOES. After having seen both situations in action I can say the price is absolutely worth it. Sorry I didn’t make my point more clear. I think tier 2 schools should enjoy the same efficiency.

  7. Paul Says:

    My points are, first, that outside of regional cost-of-living adjustments, overhead rates should not be that different among schools. You could use the same numerical adjustments already applied to the general pay scale. Second, at many places, I think there is fat to cut. Smaller schools that have smaller overhead rates, smaller (absolute) funding, and smaller endowments have already cut this fat and more. In other words, I think a small school could do more with 2% more in IDC than Harvard would lose with 2% less in IDC (if you normalize to account for the difference in absolute funding).

  8. andre Says:

    @Postdoc: I think you need to provide a little more info if you want to claim that the price is “worth it”. What is the cost of this new employee (the support staff) versus the time (and associated cost) of a grad student doing the same job? Grad student labor is dirt cheap compared to hiring a full time employee. These things are luxuries, not efficiencies.

  9. Postdoc Says:


    Haha. You sound like a PI. Of course my analysis is only qualitative, but I still stand by my point. If you have a technician who really knows the instrument, not only is research time maximized by a constantly working machine, but the quality of research has the potential to increase if the best capabilities of that instrument are utilized.

    Of course graduate students time is less expensive (aka dirt cheap to many who don’t value the time of grad students) but it isn’t free. I’ve spent two weeks working on fixing an instrument with another colleague in order to make repairs. Was that really more efficient than paying a few thousand dollars to send in the machine to be fixed? Probably not. I doubt these experiences are unique to me.

    If a large piece of shared equipment is down for 2 weeks because the manager is working on other things, that has the potential to set back dozens of researchers.

    I think that this is a bit of a divergence from the original topic, but your comment Andre strikes a cord. In my opinion, the mentality by many professors that the time of a graduate student is “dirt cheap” is directly related to the high level of unhappiness and mental anxiety that is felt by graduate students. When students are treated as having little value, they notice, and the overall health of a department suffers (in my opinion).

  10. Paul Says:

    Perhaps one more thing to consider is that many professors at the elite schools have “extracurricular” activities that pull them away from campus, and thus, require administrative staff to fill in duties that would typically be taken care of by professors at smaller schools. These activities include giving talks, consulting gigs, and starting companies…which often result in personal compensation in the form of honoraria, consulting fees, and stock options. How much of this money finds its way back to the school? Should funding agencies subsidize these activities by paying for the extra support staff?

  11. Hap Says:

    1) Aren’t instrument grants supposed to cover the cost of maintaining and fixing them? If they don’t, then a higher rate of overhead would be reasonable for a higher rate of instrumentation (because funders are expecting them to be used but aren’t really paying for them); if they do, however, then lots of instrumentation shouldn’t justify a high rate of overhead (because the funders already paid for it).

    2) I think lots of schools are expecting research to fund other things (re my previous comment) – otherwise, universities with a primarily educational mission but dwindling state funding (*cough*OSU*cough*) wouldn’t be interesting in becoming “research universities”. Hence, the federal government would be funding other state functions with the money that isn’t coming from the states. I’m not sure that’s good – if education of students is the goal, you would rather be funding teaching universities and not research ones.

    3) The graduate research report seemed ultimately to advocate a shrinking of the research apparatus. Given that the federal government is less likely to be able to fund research, someone is going to have to figure out what we really want from it. If research is more efficient in an environment with lots of support staff and instruments (and high overhead), then fewer grants will be able to be given. This would imply that research functions would be concentrated at fewer schools. Since lots of congresspeople are counting on that funding for their schools (and would not vote for it if their constituents didn’t get any of it) and are also counting on ancillary business income (from startups, etc.), concentrating research will be difficult to pull off unless one can show that 1) the added overhead is going into research, not general administration and 2) added overhead makes research more productive. If that case can’t be made, then the additional overhead is not sustainable (because in the absence of overhead decreases, research will have to concentrate without any intention because there will be fewer grants).

  12. Postdoc Says:

    I totally agree Paul. Maybe some of that value (tough to say money) comes back to the school in the form of prestige, but I’d probably say the vast majority goes in Prof. X’s pocket.

  13. AJO Says:

    I am from ND State here in Fargo, and I can say with certainty that we grad students in Fargo make less than the ones at Harvard. This is because the cost of living is higher. These wages come out of our PI’s grants, so the money is already there in both institutions.

    If Harvard is really doing the better science (could be), they should accept more students, since those of us up here have to fight tooth and nail to try to get on RAs, since there is no money (no grants being awarded).

    I know people who will have to TA for their entire graduate career. Can Harvard say that? It seems that we deserve a fair shot at being able to produce world-class research, just like Harvard does. We have an NMR and a hood, we should be able to do the exact same reactions as are done in MA.

  14. Mike Gilson Says:

    It is incorrect that SBIRs allow only 17% overhead. These are negotiated rates, but I have personally seen rates commensurate with university rates.

  15. David Says:

    Following up on Mike Gilson, NIH SBIRs and STTRs allow overhead rates at 40% without negotiation.

    Alongside this, arguments about cost of living make little sense. Overhead is calculated as Direct Costs * Overhead rate. Your direct costs should be in line with the cost of living in an area as they reflect the salaries of the people working on the grant. More expensive areas will have higher direct costs, and thus higher indirect costs at the same rate. I know my grad school in Boston paid grad students ~40% more than my undergrad in saint louis; admittedly housing in Boston is greater than 40% more than in Saint Louis, but whatever.

    High overhead rates do impact small businesses trying to subcontract on STTRs to universities, particularly under the old Phase I suggested limit of $100k, the professor’s salary on the University side can eat up almost all the work the university is supposed to perform, providing almost no value to the small business in terms of research assistants to perform work on the grant. I regularly run into this trying to do business with BU and Harvard.

  16. Troggy Says:

    Whether or not it is “fair” that top tier schools get a highly disproportionate amount of money, it seems counter productive trying to “take away” money from what are probably our most productive schools. Especially when this money apparently covers admin assistants and instrument techs. If you have a highly skilled individual making a large salary, it makes complete sense to minimize the time that they must spend on tasks that someone else could do. This is the case regardless of employment sector. Furthermore, in the case of instrument techs, these people are both cheaper and more skilled than profs. (at their specific fields), and their hiring should be encouraged.

    Finally, the argument that this money allows profs to “give talks, consult, and start companies” makes it even more essential to continue providing it. Talks educate people outside of the home institution (effectively spreading the money around) and the other two activities directly promote industry and employment – something that we can all agree is needed in the american economy.

    I think it is easy to lose sight of the multiple roles of the American academic system, one of which is to provide an anchor for American industry. If concentration of resources does that (and this discussion makes it seem like it does), then “fairness” doesn’t enter in to it.


  17. Hap Says:

    No, but it probably does matter to the people who fund it, because it affects them politically (they either fund their political opponents or, at minimum, they don’t have money they can help to use themselves). In addition, you’d like to know whether 1) the products at elite schools is better (hard to define – better science, more effective training, more effective at starting businesses and creating jobs?) and 2) whether enhanced overhead and administration generates effective yields proportionate to or greater than the added cost. In a zero-sum funding environment, these things are useful to know.

    It would also be a question whether the success of schools at getting money is auto-catalytic – whether the higher output (assumption) is a cause or an effect of the higher overhead.

  18. Evan Bayh Says:

    Can’t speak for all off-campus activities, but the “starting companies” example definitely brings $$$ (including equity) back to the university. Of course, institutional policies vary, but it’s typically split up between PI, PI’s lab, department, and university.

  19. bad wolf Says:

    At least you can see where the money doesn’t go–tuition has jumped up at least as much at the same places. So where does all this money really go? Wasn’t there a study a year or two ago that said faculty employment had been flat in academia for the last two or three decades, while university administrative employment has multiplied several-fold?

  20. martin Says:

    this is an interesting topic. I do think “top 10” institutions, in general, produce better research (I was at a top 20 and now a top 10 and i do see a difference). However, I also think top universities are more wasteful (out of topic, they buy tons of chemicals that are not needed) and maybe not as many administrative assistants are needed.

  21. Evan Bayh Says:

    My $0.02 — I think a lot of the money is going to university administrators, but NOT the kinds that are helpful to research.

    Yes, top programs often have administrative assistants for professors and dedicated mass spec staff, etc, and they’re great. Good research support staff are worth every penny, in my experience. The problem in terms of money and sheer getting-things-done is that the university-level administration has gotten bloated – more and more layers of deans and whatnot layered on top of individual departments. So many deans, etc. are drawing 6-figure salaries in jobs that didn’t exist 15 years ago, and they are not contributing to anything except increasing red tape.

  22. bad wolf Says:

    “Excited to share that I’ll be starting a new job as a professor of chemistry at Saint Louis University in August. The @bracherlab exists!”

    The Cubs win the pennant! The Cubs win the pennant!

    Congratulations, looking forward to your new developments!

  23. Man Schaus Says:

    I believe that top 10 institutes due produce better research because the professors are researchers first and teachers second. Which is why they have the professors on payroll to begin with. I disagree with this approach and believe that there should be dedicated researcher positions and that there should be a clear division between reasearch and teaching.

Leave a Reply