On Wednesday the White House Office of Science and Technology Policy (OSTP) released a set of science and technology highlights from the Trump Administration's first year. It includes the following claims about last year's science budget:
President Trump demonstrated his commitment to the importance of Federal scientific exploration by requesting $151.2 billion for Federal R&D investment in the FY 2018 budget – a 2% increase over FY 2017. Furthermore, the President’s FY 2018 Budget Request would result in the highest percent of the budget for the conduct of R&D since the FY 2014 Budget Request.
These claims are an attempt to cast last year's science budget in a favorable light – and are mostly untrue. Here's a quick fact-check.
The FY 2018 Budget Would Not Actually Have Increased R&D Spending
The numbers OSTP are using can be found here, in the Analytical Perspectives section of last year's budget, released in May. The second table in that document provides the numbers OSTP is using: $151.2 billion for R&D under its broadest definition in FY 2018, and a two percent increase over FY 2017. Straightforward, right? Unfortunately, no.
The first problem lies in the point of comparison. You may recall that last year's appropriations wrapped up very late. Congress didn't approve an omnibus until May, the same month the White House released its full budget. Given this challenging timing, the White House made a perfectly reasonable decision in their budget materials: for lack of a better alternative, they simply assumed a full-year continuing resolution for FY 2017, and used that as their baseline for comparison. This assumption meant flat or slightly declining R&D across most agencies and accounts in FY 2017 with a bottom line of $148 billion for R&D, as can be seen in the official document.
However, as we now know, legislators did not simply adopt a full-year continuing resolution last year for FY 2017. In reality, they passed an actual spending omnibus that increased science and technology spending, including a $2 billion increase for NIH and more modest increases for R&D programs at NASA, NOAA, the Department of Energy, the U.S. Geological Survey, and other agencies. All of these increases were missed by the White House's initial assumption, and in fact ran generally counter to the White House's preferences for domestic spending at the time.
In hindsight, the Administration's updated, broadest estimates of R&D for FY 2017 come out to $155 billion: a much higher number than either their original assumption for FY 2017, or what they proposed for FY 2018. The "two percent increase" is thus only true if one continues to pretend Congress didn't substantially increase the R&D budget in FY 2017, which it did.
"Research" and "Development" Are Not The Same Thing
Perhaps more importantly, however, is how last year's budget actually distributed the R&D dollars. If one looks a little deeper into last year's figures, it's clear the real focus was on the "D" in "R&D," for development – especially development at the Department of Defense. Basic and applied research, on the other hand, were slated for unprecedented reductions from the start (see graph at right).
It's important to understand that "research" and "development," as they're counted by federal budgeteers, can refer to very different activities. Research is about creating new knowledge and applying it in novel ways. It tends to be riskier, and more long-term, and more oriented toward universities and government labs. When one is thinking about, say, creating fundamentally new materials, or advancing computer science, or studying and curing cancer, one is usually talking about scientific research rather than development. These activities mostly fall in the domain of agencies like NIH, the National Science Foundation, and others in the civilian realm. For these agencies, last year's budget was the toughest in decades.
On the other hand, the DOD development enterprise is a different animal. Here's a table that breaks down last year's DOD R&D budget by funding account. The high-dollar accounts labeled "6.4" through "6.7" – the ones slated for big plus-ups in last year's budget – fund nearer-term testing and evaluation of mature DOD technologies in operational environments as they move into production: think hundreds of millions for new fighter jets or combat ships. Those activities are important from a national security perspective, but they're not intended to produce scientific discoveries, and they don't include radically innovative agencies like DARPA or the Office of Naval Research. In fact, some have argued against including DOD development and evaluation programs in the same category as other DOD research programs, because their purposes are so different. Indeed, a big chunk of that development spending is now no longer being counted as R&D, for good reason (more below).
The bottom line: the FY 2018 budget request dramatically reduced funding for scientific exploration, while increasing funding for other activities that are not so exploratory.
Would the FY 2018 Budget Have Increased R&D as a Share of Federal Spending?
This claim by OSTP seems particularly baffling at first blush. According to the official data tables (available by clicking around in here), last year's budget would have left R&D outlays at only 2.9 percent of the federal budget in FY 2018, while earlier budgets would have kept R&D around 3.5 percent or so. So what gives?
One possible explanation: there was a major accounting change in last year's budget, and about $34 billion in DOD development activities, known as "operational systems development" or the "6.7" account, are no longer officially counted as R&D. One can see the effects of this re-definition of "R&D" on the bottom line in table 2 here. This is a good change and makes the federal R&D budget more reflective of science and technology, but it also means the R&D budget is over $30 billion smaller than it used to be. If OSTP was using the old definition to get total R&D, then it certainly could have meant an increase in R&D relative to the rest of the budget in FY 2018, rather than the apparent "decline" in the official figures. A confusing explanation to be sure.
At any rate, R&D has been declining as a share of the federal budget for decades, so credit to them for at least thinking about how to turn it around.