Research Spending at Academic Medical Centers

Apr 03 2015 Published by under Uncategorized

Have you heard an administrator say 'we lose money on research?'  That sentiment often irritates faculty members for at least two reasons. First, the use of the term "lose" implies that the purpose of doing research at a university, medical center, or research institute is to make money at it rather than because it is a key component of their reason for existence. Wouldn't the term "spend" be more appropriate? Second, it can sound like the implication is that they would rather be out of the research business because it is costing them so much money.

The American Associate of Medical Colleges (AAMC) recently released a short report that provides some insights into how the cost accounting to reach conclusions about these costs. These data are based on surveys of 46 academic medical centers. The bottom line conclusion is that a collection of academic medical centers spend 53 cents for every dollar of sponsored research support that they receive.


These four largest categories include unrecovered indirect (facilities and administrative) costs on sponsored research, facilities and administrative costs on research supported internally, salary support for research effort not supported by sponsored research, and the costs of start-up packages. Additional categories include bridge funding, costs for salaries over the salary cap, and cost sharing either voluntarily committed or volunteered by the institution.

It is estimated that these costs average $111,151,553 with a 95% confidence interval of $21,143,786 per medical school. I am not sure how to interpret this average (aside from the significant figure issues) since the direct cost support of the 46 schools participating the study ranged from $26 million to $751 million, but this does give some sense of the distribution.

Do these figures make sense? If you were in charge, how would you try to allocate these institutional resources?

49 responses so far

  • DJMH says:

    Not feeling a lot of sympathy for the "over the salary cap cost share" category, but that's a pretty small percent of the total.

    What's the difference between Unrecovered F&A and Departmental F&A subsidies? Those are the most opaque costs, since as categories they seem to encompass everything from HVAC to deanlets.

    From the report, F&A as a whole is: "Institutional reimbursement for costs that are not directly related to an individual research project but are essential to support the research endeavor. These charges include costs of buildings (operations, maintenance, and depreciation); equipment depreciation; information systems; environmental health and safety; grant management; and many other support costs"

    but I couldn't figure out the difference between the first and second bars on the chart.

    • datahound says:

      My reading the report indicates that the first bar corresponds to unreimbursed indirect costs on sponsored research (i.e. the difference between the indirect costs paid from grants and the costs covered by the facilities and administrative costs for this level of research activity). The second bar corresponds to indirect costs on research for which the direct research costs are paid by the institution.

  • Established PI says:

    I am very glad you posted this. I have always wondered about the claim that it costs the institution money whenever we get a grant, since administrators are always rending their garments whenever we lose a faculty member who brings in big grants. If big research programs means big losses, then what's the fuss when they leave, other than loss of prestige?

    My question for you is whether there is creative accounting in the largest two categories. The loss, particularly in the first category, is relative to the amount that the university declares it should be getting. What is that amount based on? Are they based on legitimate costs for research or additional university expenses that get folded into that category? Does it include fancy facilities that do not benefit researchers?

    Another large category is "additional salary support." In addition to doing research, basic scientists at medical schools teach medical and graduate students, as well as serve on myriad committees for education, training and hiring. This is what the institutional contribution to our salaries is supposed to cover. If this what the salary component in the graph refers to, or is it salary for investigators who fall below the target recovery percentage?

    • datahound says:

      This sort of accounting is always challenging to understand. Part of the issue relates to the fixed versus marginal costs. If an well-funded investigator leaves an institution, the institution loses the indirect costs reimbursement associated with these grants while a substantial portion of the indirect costs (building costs, salaries for administrators under the administrative cost cap of 26%, etc.) do not go away. This is a net loss for the institution. It is a bit like sharing an apartment with four roommates. If one moves out and stops paying his/her portion of the rent, it costs the group money. Some of the costs may go down (e.g. less water consumption) but others won't change much.

      The facilities and administration (indirect cost) rate is negotiated based on documentation of real costs. See . It is unlikely that other expenses get folded in. As you may know, Stanford got caught including indefensible items back in the early 1990s and everyone took notice. See .

      According to the report, additional salary support is support for research that is funded by the institution. In principle, this should not include salary for other activities such as those you note. This, of course, depends on effort reporting processes.

  • David Russler-Germain says:

    This is very interesting. I certainly have heard of administrators (from several different institutions) claim "for every grant our professors receive, it costs our institution money."

    As you touch on, is this insinuating that successful grant applications are net financial losses, in that the incoming grant money all goes to the research and other expenses (i.e. not to university coffers), and then more capital outflow occurs when the university has to help pay for the research in the ways the AAMC explored in this report?

    If so, then I strongly agree with the point / issue that we need to consider that the goal of doing scientific research as an academic institution is not to find an optimal "level" of research output that maximizes "profit" for the host institution. The goal is, in my view, to produce the most and best research output in a fashion that remains fiscally responsible and aligns with the institutional mission.

    This reminds me of a great post by The Incidental Economist:

    tl;dr: A 2009 study from Tufts looked at drug development costs. By factoring in “opportunity costs of capital” this number was calculated to be $800M/drug, not the actual $400M "out of pocket costs" for developing a drug found in the raw data. The explanation for this discrepancy is that the pharmaceutical industry and the Tufts researchers argued that companies could have spent that $400M on research or, for instance, put it in the bank, where they'd have earned interest. That interest income was “lost” by developing drugs (requiring spending the money) instead of investing the "out of pocket costs" dollars. Bizarre, and slightly dishonest, eh?

    Back to the AAMC study, and in a way related to the Tufts/drug development cost topic, I'd be very interested to know where the money comes from that goes towards these "research expenditures". I'm neither disputing that this money is getting spent by the institutions, nor that it doesn't go towards worthwhile things (that's a different debate). I am curious, though, how much of the "availability" (so-to-speak) of this capital comes from the very having of research grants in the first place.

    For example, how much of this money comes from private donations? How much comes from clinical revenue generated by physician-scientist PIs whose laboratories bring in the grants? How much comes from revenues from patents / licenses / fees that are (perhaps often indirectly) derived from research activity? Controversially, how much comes from undergraduate tuition dollars? As only a student myself, I can't say I have a good guess to the answers to these questions.

    I'm sure these questions would be incredibly hard to answer given the complexity of university accounting, but I think knowing the "income" source for these "research expenditures" is exceedingly important. If the university wouldn't have some of the money going to these expenditures if not for the very research activity supported by the grants, then the lens through which to interpret this is very different. It's also important to consider the alternative uses for these dollars. If the money was collected, for example, by a division's own clinical activities to support research within that division (i.e. some start-up packages, from my understanding), then that is what the money is there for in the first place (and thus calling it a "expense" with the slant of "look how generous the institution is, having been burdened by new hires / successful grants!" isn't very genuine).

    Overall, university budgets are fascinating. I wish we could all know more about them.

    • datahound says:

      Lots of good questions...I wish I could find a good example of a simple balance sheet for a university (Total dollars in one column and total dollar out in another). This basic information would be useful in developing an understanding. As you note, this would account for true out of pocket costs and into pocket revenues.

      The sort of assignment of costs done in this report is much more complicated since most dollars are fungible (can be moved around and assigned to different categories depending on context). This is subject both to unintentional (challenging in clearly defining accounting rules, particular across institutions) and intentional distortions.

  • DJMH says:

    I'm now noticing an even more confusing part of the definitions, namely that the "research dollar" basis of this analysis appears to be direct, not total costs? Page 3,
    "These institutions each received between $26 million and $751 million in external funding (total direct costs) for medical research in 2013. The average medical school investment was an additional $0.53 for each dollar of sponsored research received."

    If "dollar of sponsored research" here means what it sounds like from the first sentence, dollar of *Direct costs*, then this 53% additional is...what, on top of existing IDC? It seems to me that it's a bit disingenuous to show how much the med school is contributing to F&A, without noting that the NIH probably provided another $0.55 on the dollar to that exact category as well.

    • datahound says:

      Interesting point. I checked in the report and with sources at AAMC and it is 53 cents on every dollar, direct and indirect.

  • Comradde PhysioProffe says:

    Interesting. First, our Dean never refers to "losing" money on research. Rather, he uses the language of "investing" money in research, because that is an integral part of our mission as a medical school. Second, I am pretty surprised at the 53 cents per dollar number, as our Dean tells us that we invest about 10 cents per dollar. So either our medical school is vastly more efficient than others, or there is something very different about our accounting and the accounting in the survey. I suspect that it's the latter.

  • Dave says:

    Something isn't right here. There is no way this can be correct, but I've had too many Monks Cafe brews to give a shit about this piece of medical school propaganda.

    • datahound says:

      What strikes you as not correct? I agree that this is put out by AAMC to support their position, but I would be interested in knowing what strikes you about it. Bear in mind that these figures undoubtably vary substantially between institutions.

  • Neuro-conservative says:

    I agree with Dave & CPP. These numbers are being manipulated for a political effect.

    Whether or not they read this blog and Drugmonkey, the AAMC leaders smell the changes in the air and realize that some pressure might fall on then to ante up more $$. This is their proactive defense.

    Look at what they do, not what they say. When the opportunity arises to expand research, hire new faculty, build new buildings -- they seize it every time.

    • DJMH says:

      Look at what they do, not what they say. When the opportunity arises to expand research, hire new faculty, build new buildings -- they seize it every time.


    • datahound says:

      I am not sure what they are saying and what they are doing are all that different. The chart indicates that they spend 5 cents for every dollar of sponsored research for start-up packages (i.e. recruiting new faculty) compared with a bit over 2 cents for bridge funding. Thus, they are spending a substantial amount to hire (a relatively small number of) new faculty while spending half as much to help bridge the large number of faculty who are struggling in the present environment.

      Similarly, the 14 cents of unrecovered facilities and administrative costs is likely largely administrative costs (since the administrative component of indirect cost rates for universities is capped at 26%). This is likely mostly administrative staff who work on regulation-related issues.

  • mytchondria says:

    I think my beer googles are messing up the font on your graph.

  • UCProf says:

    I guess this isn't a problem at medical schools, but when you throw the Schools of Arts/Humanities/etc into the mix they clobber us over the head with these stats.

    At our university, the Humanities/Arts/etc faculty questions why the university is heavily subsidizing the Medicine/Science/Engineering research, but offers so little to the Humanities/Arts/etc.

    The problem with indirect costs is that they are mostly fixed costs, with a small incremental cost. For research funded by small organizations (companies), it makes sense to add on a percentage, but for federally funded research it would make more sense for the federal government to pay a large annual fixed fee. Doing so, would also eliminate a tremendous amount of internal activity at my university that is geared toward which internal organization controls indirect cost reimbursements. (For instance, I have a dual appointment in science/medicine, and can submit a grant through either school. The indirect costs will go into different Dean's pockets, depending on how I route the grant proposal internally.)

  • Roger says:

    Admin: We're losing money? Quick, let's create a new VP position and fill it fast!

  • AcademicLurker says:


    When I was at a medical school I once submitted a joint NIH proposal with some folks in the department of mathematics in the college of A&S. That still stands as the most aggravating grant submission experience of my career due to all the wrangling between A&S and Medicine over how the indirects would be split.

  • Dave says:

    @Datahound: the reason for my suspicion is very simple. These numbers IN NO WAY reflect what I am seeing on the ground. Sure, my experience may be unique somehow, but I am inclined to believe that I am at a very, very typical standalone state medical school. I could go into endless details, but I just can't be arsed and......well.....whats the point?

    If you looked carefully at the 'investment in research' they are claiming, I'm convinced most of the $0.50 probably goes straight back to the administration. Even the 'additional salary support of research effort' is vague enough that it could mean salary for literally anyone in a medical school.

    The 50 cents number is obviously inflated. The only question for me is how and why?

    • datahound says:

      The largest category is un-reimbursed facilities and administrative costs. This is likely largely administrative costs that exceed the 26% administrative cost rate cap. These administrative costs (I would guess) are largely administrative salaries for compliance-related tasks and, thus, these funds to go largely to the administration.

      My guess is that these numbers are accurate in an accounting sense, but investigators do not feel much benefit (and may feel harm) from a substantial portion of the costs. On the other hand, many of the costs are required to comply with the rules and regulations that are in place for sponsored research.

  • Dave says:

    And the new hires are probably concentrated in a few prestigious schools, and/or reflect new Chair hiring more than Assist Prof TT jobs.

    Can we get a list of the schools that participated in this?

    • datahound says:

      Unfortunately, it is very difficult to get this information from AAMC.

      • Dave says:

        Why is that?

        • datahound says:

          The AAMC obtains data voluntarily from their members with various conditions about anonymity. AAMC is concerned that they would not get the data that they do if their members felt that it would be harmful to them.

          Given my feelings about the benefits of transparency, I am frustrated by this policy, but some data are better than no data, in my opinion.

          You could ask your institution for their accounting along these lines as see what you get (although I understand that asking may be unproductive or have other undesirable consequences).

  • UCProf says:

    Datahound, do you know the reasoning behind why NIH grant/contracts do not compete on total costs?

    I'm not sure if it is tradition or policy, but when our study section reviews proposals we comment on the budget, but only direct costs never total costs.

    That is one way the Federal government could put pressure on universities to control indirect costs better.

  • Datahound says:

    I can give you the policy position. The purpose of the study section is to judge the merit of the science as proposed and to comment on whether the budget is appropriate for conducting that proposed research. This information is provided to the advisory councils and institute staff to make funding recommendations. Thus, study sections are not, as a matter of policy, expected to comment on the total cost of research.

    At the institute/funding decision level, total costs (in my experience) are rarely taken into account although for some program announcement/RFAs, there is a limit on the total costs for a project.

    I agree that this could be used (potentially) to get more bang for the buck. There would be a major outcry from free-standing hospitals (e.g. MGH) and free-standing research institute (e.g. Scripps) because their indirect cost rates tend to be substantially higher than average (in part because they are not subject to the cap on administrative costs in the indirect cost rates).

    • DJMH says:

      in part because they are not subject to the cap on administrative costs in the indirect cost rates

      Whoa, I did *not* realize that. So, when we complain about deanletting, we may not be completely out of line?

      I have long felt that it would be appropriate not for study sections to comment on the budget per se, but for NIH to distribute the same way NSF does, with the budget being for total costs, not direct. So a $400K modular would yield a $258K direct at a place with 55% indirects, and $211K at a place with 90% indirects. This is the only way I can imagine to get the faculty to pressure administration to bring down IDC.

    • UCProf says:

      There would be a major outcry from free-standing hospitals (e.g. MGH) and free-standing research institute (e.g. Scripps) because their indirect cost rates tend to be substantially higher than average (in part because they are not subject to the cap on administrative costs in the indirect cost rates).

      Is this difference in administrative cost reimbursement abused? For instance, the Broad Institute has an indirect rate of 73%, while MIT has a rate of 56%. Most of the PI's at the Broad Institute are MIT professors. This structure sounds like something every big center could set up.

  • UCProf says:

    Here's a specific example of audited financials from Scripps Research Institute, which seems to operate almost like a free standing medical school, but without any clinical revenue:

    It looks like they had $377 million in revenue in 2014 (down about 3% from 2013). When you total up their expenses, it was $386 million in 2014. So, they lost $9 million in 2014.

    Near the end of the long pdf file, they list "SOURCES OF GRANTS AND CONTRACTS REVENUE" by year. In 2014, they had $257.8 million from NIH funding. This amount is down about 10% from their high mark in 2011.

    So, a place that is charging 89.5% indirect costs can't be profitable. It looks like they needed another 4 percentage points added on to their indirect costs to be in the black.

    • drugmonkey says:

      Given the news of a possible USC bailout about a year ago, Scripps may not be the best general example. Clearly their business model is unsustainable.

  • Established PI says:

    Thanks for clarifying a lot of confusing issues regarding research costs. Now I understand that when the CFO tells us that our research costs them money, the added cost to them for each grant I bring in is pretty small. It is the cost of having recruited faculty, and for supporting faculty with inadequate grant support, that are the big ticket items.

    As several commenters noted, there don't seem to be incentives for cost-effectiveness; institutions can hand over the bill and negotiate a higher IDC rate because they happen to have higher adminstrative costs or built more (or more expensive) buildings. While I agree that the study section is not the place to scrutinize this, wouldn't it make sense to have this reviewed by an external advisory panel?

    Is all the information on the basis for IDC rates for each institution publicly available? If so, where can one find it?

    • datahound says:

      First, as noted by DJMH, the big ticket items are largely indirect costs. You are correct that the marginal cost of a new grant is quite low so that helps the institution, but research activity overall costs the institution.

      Second, the institution cannot just hand over the bill. Indirect cost rate negotiations are quite tough and involved. Furthermore, the administrative component at universities has been capped at 26% for some time. The biggest anomaly that I know of is that free-standing hospitals and, I believe, free-standing research institutes are not subject to the OMB circular with this cap (Circular A-21) but rather a separate set of rules that does not include the cap. From my earlier analysis ( ), such institutions have significantly higher indirect cost rates (MGH 74%, Brigham and Womens 76.5%, Scripps 89.5%, Salk 90%, Cold Spring Harbor 88.5%). The variation between university-based institutions is much smaller (most between 50 and 60% with a few outliers).

      To my knowledge, I do not think the details for indirect cost rates are publicly available. I do know of one detailed description (from the University of Cincinnati) at .

      I do not know these any useful materials could be obtained through FOIA.

  • DJMH says:

    It is the cost of having recruited faculty, and for supporting faculty with inadequate grant support, that are the big ticket items.

    No, they're not. It's the unreimbursed F&A (the left two columns) that are the big issues. If this is indeed because of federally mandated regulatory bookkeeping, then the universities are pretty helpless to bring those costs down, short of petitioning the government to reduce regulatory load (hah!) If some of it is because of decisions to build out the campus under the assumption that more grants would flow in, then it's more the fault of the institution.

    Is all the information on the basis for IDC rates for each institution publicly available? If so, where can one find it?

    There used to be a great repository that DM linked to, but it looks like they've put it behind a paywall. Other than that, I don't know.

  • DJMH says:

    @EPI, I wrote a quick macro that operates on downloaded data from RePorter. Since the grant data now includes direct, indirect, and total, it's simple to calculate IDC rate per institution.

    The only catch is that IDCs can be reduced on any given grant if, for example, the PI buys a lot of equipment. But as far as I know they can't be increased for any reason (DH? true?) so my macro just takes the max calculated IDC per institution. I don't know if that's accurate or not but it matches from the few that I've spot-checked.

  • Established PI says:

    Thanks @DH and @DJMH for further clarification and information.

    I still wonder why there isn't more transparency in this process. DH is in a far better position than I to know what aspects of the information on which IDC is based may be highly sensitive that would somehow hurt an institution if it were released (I don't mean the Stanford debacle). However, if the taxpayer is footing the bill, don't they (we) deserve a bit more information?

    • datahound says:

      Given my feelings about transparency is general, you should not be surprised that I agree that both the public and the institutions would benefit from more transparency. I imagine that the general view is that release of information could somehow be harmful has prevailed. My view is that the lack of transparency leads many to believe that something untoward is being hidden. I do not know of any sensitive information that is, in general, included in these negotiations.

  • E rook says:

    If the institutions (MRU, state/private U, standalones, etc.) can be considered non-profits organizations, then it seems perfectly reasonable that they operate on negative budgets most years. The accounting gets sketchy when they are associated with hospitals or univs that have all the trappings of for-profit entities. I think that the administrations are doing exactly what they say they're doing with these IDCs. The regulatory burden has certainly increased and the administrative burden of handling more grant apps has probably increased too. As far as building new buildings, they have an incentive to remain as competitive as possible with respect to winning grants/contracts and attracting the best talent (to win the grants and do the research), it's a long term investment....with risk, but they must be crunching the numbers to assess the risk of that the capital improvements will pay off, and doing fundraising from philanthropic foundations to pay for much of the up front costs.

    But the charged language from a dean, "it costs me money for you to carry ou [our] research [mission]," reflects poor leadership and probably a toxic environment. These are non-profit entities whose mission statements include carrying out research, and if the leadership sees a problem with that, then we need a new leader, not a new mission.

    • UCProf says:

      As usual, the University of California is a leader in this trend:

      While the overall number of employees has increased 51% since 1991, we have managed to increase the number of managers by 252%.

      • E-Rook says:

        Odd that this happened at a time when UC was slashing its budget by a 1/3 during the Yudof days. Lost retirement "matching" contributions for faculty, an effective freeze in full-time equivalent ("TT") lines (not replacing retires), and forced furlough for the staff. It's as though a group of individuals has implemented policies whose effects are to secure their power/resources/stability. I'm shocked.

    • DJMH says:

      Wow. Just, wow.

  • SaG says:

    If schools are "losing money by investing in" research then why does the answer seem to be to expand their research enterprise (i.e., build new buildings) to attract more NIH dollars?

    That is the paradoxical answer I regular hear from Deans and such.."We are losing money on research so we have to expand to attract more NIH dollars." Wait doesn't that mean you will lose even more money?

    • datahound says:

      I agree that this has been a response in the past which I, too, found paradoxical. On the other had, there are other incentives (mission, prestige, etc.) that have favored construction. It would be interesting to know how much construction at academic medical centers at present. My guess is not much.

  • UCProf says:

    DH, thanks for posting this. I've spent an unhelathy amount of time looking into the finances of research places over the past week.

    I have a hobby of investing. So, I've spent plenty of time poring over the finances of for-profit companies.

    One problem that I sense is that universities just don't have the feedback that a for-profit company has. I can tell a well run company. Everyone knows. They have consistent profits. They control costs. Of course, poorly run companies also exist. Those managers don't last long.

    I have no idea which universities are well run and which are poorly run from a financial point of view. Presidents of universities (and deans) seem to focus, and are lauded, more for raising money than for cutting costs.

    For a non-profit, minimizing unnecessary costs and directing those savings to the mission (research and/or education) is absolutely key for a manager.

    I'd like to see some metrics on that. Who are the super-star financial officers at universities? I'm not sure what those metrics are. I'd say minimizing overhead, but there are incentives to recover that (the OMB A-21/A-122 recovery).

Leave a Reply