Last week, after a surprise attempt by the Trump Administration to take back $1.2 billion of committed funds from NIH, HHS Secretary Tom Price responded:

Price repeatedly suggested reducing the amount the NIH pays universities to cover “overhead” costs, like lab equipment and utilities. That would let the agency direct more of its funds to actual research, even if the overall budget were reduced, he said.

“I was struck by one thing at NIH,” Price said, “and that is that about 30 percent of the grant money that goes out is used for indirect expenses, which as you know means that that money goes for something other than the research that’s being done.”

This led to much internet discussion about how certain places suck down YUGE! and BIGLY amounts of overheads, while other institutions (usually said discussants’) were frugal and noble. If universities had to compete on indirect costs, the major research gluttons would lose and the money would be distributed into the heartland, where good, pious (in a sciencey way) researchers and their administrators would spend the money wisely.

Before I get to the main point–that nobody is leaving money on the table–let’s discuss two other things. First, as Potnia Theron notes, this will just result in a lot of category shifting; related to that, there are things in the indirect category that are legitimate expenses (e.g., personal computers). Second, to all the overhead warriors, you do realize that neither scientists nor universities are getting that money? Instead, it’s going to pay for rich people’s tax cuts.

OK, so onto indirect costs. I went to NIH Exporter and downloaded the FY 2016 RePORTER Project Data (you can too!). Then I started looking at what percentage of institutions’ NIH funding was spent on direct versus indirect costs (keep in mind the caveat I mentioned above). I didn’t break this down by grant type, since the goal was simply to identify the percentage spent on indirect costs across the entire institution (besides, low overhead grants ‘subsidize’ high overhead grants). I also exclude grants that had no overheads, since these were small in size and uncommon, and when I checked a few, it wasn’t clear why there weren’t any overheads (these could be errors).

So let’s look at some institutions. Harvard (and I lumped together Harvard institutions, which administratively probably isn’t correct, but in the popular mind, they’re indistinguishable) has 72.7% of NIH funds spent on direct costs, and 27.3% spent on indirect costs. Now let’s look at some heartlandy stuff. University of North Dakota has a 75.4:24.6 direct-indirect cost split (go Fighting Hawks!), and North Dakota State (go Bison!) has a 72.2:27.8 split. ND State is basically the same as Harvard.

University of Nebraska (go ‘Huskers!) has a 71.8:28.2 split, while the Alaska universities had a 72.7:26.3 split. Stanford, not surprisingly, as it’s always played fast and loose with the rules, has a 68.7:31.3 split.

To examine the Boston effect, Northeastern University has a 74.8:25.2 split, better than many of the public universities. The Broad Institute (which is TAKIN ALL THE MUNEEZ!!) has a 78.6:21.4 split–very efficient! Entities with LLC in their names had a 73.6:26.4 split, while entities with “INC.” in their names had a 77.7:22.3 split.

When I first wrote about this, a faculty member at Ohio University worked himself into high dudgeon over coastal indirect rates. In 2016, Ohio University had a split of 68.9:31.1–worse than Harvard.

Whether overhead rates should be this high to begin with is another question (though this loss of funds would drive up tuition rates, even at large public universities by hundreds of dollars). But there aren’t significant disparities in indirect costs. Your humble heartland institution is just as greedy as your coastal elite one (sometimes even a bit more).

No one is leaving money on the table.