Drunken Parties and Exotic Trips–Indirect Costs Must Be Slashed

Let’s begin by asking a simple question: what are indirect costs?

For a research project, there are some costs that are directly attributable to that project such as the salaries and benefits of research personnel, relevant travel, and supplies and equipment used in the research. These are called direct costs, and they are itemized individually in grant applications.

But some costs of doing research aren’t directly attributable. Things like the offices or laboratories used, the utilities to keep the water and power on in those spaces, and common administrative support like IT and HR are usually not attributable to any particular project or person. Since these costs can’t be itemized in grant applications, they are instead funded as a percentage of the grant. For example, if a university has a 50 percent indirect rate, a research grant of $100,000 would also receive $50,000 of additional funding for indirect costs. These costs go by several names, including indirect costs, overhead, and facilities and administrative (F&A) costs.

Reimbursement for legitimate indirect costs is appropriate since research cannot be done without incurring some indirect costs. But there is a virtually unlimited supply of unnecessary indirect costs that could be included, so the task is to set indirect rates at an appropriate level.

What are the benefits of cutting unnecessary indirect costs?

Lower indirect rates could be used to either get more research bang for every dollar spent or to lower the cost of funding any given amount of research. For instance, the grant above spent $150,000, but only $100,000 of that funded actual research. If the indirect rate is reduced from 50 percent to 25 percent, then that same $150,000 of total spending would now fund 20 percent more research—i.e., $120,000. Alternatively, the same amount of research could be funded for less total cost with lower indirect rates—i.e., funding the original $100,000 of research costs above would only cost $125,000 with a 25 percent indirect rate rather than $150,000 with a 50 percent indirect rate.

Are indirect rates too high? According to the Chronicle of Higher Education, the average indirect rate for NIH grants was around 27–28 percent. But some universities have negotiated much higher rates. For example, the NIH announcement noted that Harvard’s rate is 69 percent, Yale’s is 67.5 percent, and John Hopkins’ is 63.7 percent.

Just about everyone who isn’t employed by a university administrative office or paid to advocate on their behalf thinks these rates are too high. There are several convincing reasons for this belief.

First, universities have been caught inflating their indirect costs in the past. For example, in the 1980s and 1990s, Stanford University was including costs for “flowers for the its president’s home and the depreciation of a yacht” in their indirect cost formula. Director of Science Programs at the National Association of Scholars, J. Scott Turner, highlighted more recent abuse, including one researcher using indirect funding for “holidays to exotic locations, drunken parties peppered with outlandish sexual behaviors, frequent visits to gay bars, drugs, and humiliating demands made to subordinates.”

Second, you couldn’t pick worse watchdogs. Indirect rates are currently negotiated between each university and either the Department of Health and Human Services or the Office of Naval Research—the Navy was a big early funder of research. It would be hard to find two industries that demonstrate less concern for cost-effectiveness than health care and the military. Call me a cynic, but I have little faith that an industry where it is acceptable to charge $15 for a single Tylenol pill, or one that pays $10,000 for a toilet seat cover, suddenly turn into hard-driving bean counters when it comes to making sure universities aren’t overcounting their overhead. Note that in the Stanford scandal mentioned above, Stanford’s settlement with the government included a finding of “no wrongdoing by the university.”

Third, other research organizations have lower indirect costs. A Government Accountability Office report documented that nonprofit research organizations, small businesses, and industry all have lower indirect cost rates than universities. For example, the average indirect rate on National Science Foundation grants to universities was 27 percent in 2016, but just 14 percent for grants to industry.

Fourth, during World War 2, high indirect rates of around 50 percent were deliberately set at excessive levels, in part, to cajole universities into accepting federal research funding. The average rate is about half that now, which supports the notion that the 50 percent rate was “overly generous.” But some schools—especially those that get the most research dollars—have much higher rates, up to 69 percent. Although it is possible that 50 percent in 1941 was too generous while 69 percent in 2025 is legit, it is quite unlikely.

Fifth, indirect rates don’t change the way they should when research budgets change. As Jay Greene notes:

Sixth, the rates are set based on what the university spends, not what it should cost. Universities have a well-deserved reputation for spending wastefully. For example, a public California college was planning to build a new dorm that would cost $330,000 per room. Such profligate spending likely applies to much of the spending eligible for inclusion in indirect costs as well. If a $100 million dollar building would be perfectly sufficient, but the university opts for a gold plated $500 million one instead, it can base its indirect rate on $500 million. Similarly, if one compliance officer could do the work, but a university decides to hire the president’s spouse to oversee the compliance officer, then that unnecessary salary is included in indirect rates too.

Seventh, indirect rates can function as a slush fund by letting universities double dip when philanthropists fund a new building. If a donor finances the construction costs of a new building, the university can still include depreciation of the building as an indirect cost. Since the donor already paid for the building, the depreciation costs that are included in the indirect rate function as a slush fund for the university. Some colleges require donors to contribute to an endowment to finance operations and maintenance as well when creating a new named building. When that’s the case, taxpayers—partially—pay for the same expense three times. First, the donor gets to deduct their donation to the university from their taxes. Second, the university often doesn’t pay any tax on the endowment’s earnings, and the handful that do pay only 1.4 percent, compared to the typical 15 percent or 20 percent capital gains rate. Third, the university gets to include those costs when calculating its indirect rate.

Is the new rate appropriate?

If we’re going to pick an arbitrary number, a 15 percent cap is probably in the right ballpark. We’ve used caps before, sometimes as low as 8 percent. Meanwhile, the European Union caps rates at 25 percent and Japan caps them at 30 percent. We’re less bureaucratic, which would imply our rates should be lower. On the other hand, our universities are generally better, especially in the research arena, so maybe we shouldn’t be using those rates as benchmarks.

Most of the opposition is rather unpersuasive. For instance, Christopher Newfield argues that indirect rates are too low, writing that “sponsored research, almost all in science and engineering, loses massive amounts of money even at current recovery rates.” He tries to prove this by showing that universities also fund some research on their own, which he interprets as the university needing to subsidize federally funded research because indirect rates are too low. The reality is that the purpose of a university is to discover and disseminate knowledge, and universities, therefore, use some of their many other sources of revenue—tuition, state funding, philanthropy, commercial income—to fund research with no implications whatsoever for whether indirect rates are too high or too low. Ironically, Newfield is a co-editor of a book titled “The Limits of the Numerical: The Abuses and Uses of Quantification.”

But some opposition is more persuasive. The best opposition I’ve come across so far comes from Stuart Buck, who argues that indirect costs have crept up over time due to an increase in regulations. He cites an organization that documents 270 changes in regulations affecting federally funding research since 1991, with 62 percent being enacted since 2014. He argues that “if we want to reduce indirect costs more seriously, we also need to reduce the federal regulatory burden that causes universities to hire so many administrators in the first place.”

My only response to this is the following: your terms are acceptable. It is a libertarian dream to lower government spending—in the form of lower indirect rates—and government interference with research—in the form of excessive regulatory burden.

But I do worry that without accompanying deregulatory changes, a cap of 15 percent is too low, and I worry even more that letting politicians in DC pick an arbitrary number will not generally lead to the right number being chosen. I’d be more comfortable with more decentralized options, and there are several to choose from. Jay Greene argues that federal agencies should pay no more for overhead than the university charges to other funders of research. Richard Vedder argued that indirect rates should be another line item in the budget of proposals.

Another option would be to cap indirect rates by field rather than by institution. For example, chemistry research is likely more expensive to conduct due to extra safety and disposal requirements compared to, say, economics research, which often requires little more than an office and a computer. Thus, indirect rates for chemistry research should be higher than rates for economics research. These rates could be set in a decentralized fashion by taking the median of costs among past grants. For example, if the last three NIH grants provided funding for human resources support of $50, $100, and $400 per professor, then going forward, all grants would be reimbursed at $100, repeating the exercise for the other components of indirect costs relevant for research in that field.

What are the prospects of the 15 percent cap sticking? In the current iteration, not good. Courts have already paused implementation of the cap while they evaluate its legality. The main legal obstacle is that the first Trump administration also wanted to lower indirect rates, and Congress went the other way, placing more restrictions on NIH’s ability to cap rates, restrictions that may kill this effort.

But all of this could just be drawing more attention to indirect costs to pave the way for inclusion in upcoming reconciliation bills. Congressional hearings that provided a forensic accounting of Harvard’s rate—69 percent—Yale’s—67.5 percent—and John Hopkins—63.7 percent—could expose scandals like the Stanford one in the 1990s that would make it politically possible to dramatically cut indirect rates. The savings could be substantial. The federal government spends about $60 billion per year on research at universities, with around $12 billion of that being indirect cost payments. Lowering the rate to 15 percent could save about $5 billion a year, or $50 billion over ten years.