This article is published in the March 2016 issue.

President’s FY 2017 Budget Request: A Disappointment for Computer Science

On February 9, President Obama released his final Budget Request to Congress, a $4.1 trillion request for fiscal year 2017 (FY17) that some in the science community have called “aspirational,” which might be a nice way of saying disappointingly unrealistic.

Before getting into details, it’s worth pointing out that the president has been a tremendous champion for federal investments in science throughout his two terms. His administration has launched a large number of new initiatives on brain science, big data, robotics, clean energy, advanced manufacturing, strategic computing, cybersecurity, smart communities, and more that have brought new funding and new energy to federally supported science.

That noted, the president’s FY17 budget calls for a funding increase of 4 percent to federal R&D, including a 6 percent increase for basic and applied research. The National Science Foundation (NSF), which supports 82 percent of all fundamental computer science research in U.S. universities, appears to be one of the big beneficiaries of the increased investments, with its budget growing by 6.7 percent under the president’s plan. NSF’s Computer and Information Science and Engineering (CISE) directorate, where most of the foundation’s computer science research support originates, would be slated for a 6.3 percent budget increase under this plan.

These would appear to be more-than-respectable increases, particularly as they come while under a difficult budget agreement, which the president reached with Congress at the end of last year, that would cap discretionary spending growth to just 2 percent in FY 2017.

However, all is not what it seems. The president, hamstrung by the tight discretionary funding caps (and a congressional majority not interested in revisiting them), is asking for new mandatory spending to make up the bulk of his requested increases at NSF and elsewhere in the budget. For example, of the $500 million in requested increase at NSF, the president is asking Congress for $400 million in a new “one-time” mandatory funding stream.

Mandatory spending is funding decided by statute or formula, as with Social Security, Medicare, Medicaid, and federal food stamp programs. It is spending that is essentially on autopilot—Congress doesn’t need to reach agreement on it every year. Creating a new mandatory funding stream for NSF, or any other discretionary program, would require special legislation outside the normal appropriations process and the approval of Congress with a Republican majority increasingly inclined to cut federal spending, not create new sources. The likelihood of Congress approving the President’s new mandatory funding streams is probably near zero.

So, when you remove the requested increases in the president’s budget that rely on mandatory spending, the remaining investments look pretty underwhelming. Overall funding at NSF would be up just 1.3 percent, compared to FY 2016. Funding in the CISE directorate would be up just 0.3 percent, compared to the FY 2016 level, a level that doesn’t even keep pace with inflation.

An examination of three specific agency budgets might add some clarity:

National Science Foundation

No other science agency, with the possible exception of the National Institutes of Health (NIH), is as affected by the use of this new mandatory spending tactic as NSF.

At NSF, the research directorates are included under the Research & Related Activities (RRA) budget line. The FY 2017 top-line number for RRA is $6.4 billion, which represents a $392 million increase (6.5 percent) relative to FY 2016. However, of that increase, $346 million is funded using mandatory funding. If we just look at discretionary funding, as the appropriators will, RRA would receive a bump of only $46 million, or a 0.8 percent increase, relative to FY 2016. That’s a significant difference.

Drilling into the directorates of RRA, CISE’s FY 2017 increase is almost completely funded under the new mandatory line. CISE would only receive a ~$2 million increase in discretionary funding, or just 0.2 percent above FY 2016. This is about on par with the other research directorates.

Department of Defense

The Department of Defense’s Science and Technology (DOD S&T) program is made up of three accounts: 6.1 (basic research), 6.2 (applied research), and 6.3 (advanced research). These accounts are made up of individual accounts for each of the three services (Army, Navy, and Air Force), as well as a Defense Wide (DW) account. DOD S&T overall would receive a 4.1 percent cut under the president’s request, going from $13 billion in FY 2016 to $12.5 billion in FY 2017.

DOD 6.1 basic research would see a significant cut under the president’s plan. The account would see a 9 percent cut, going from $2.3 billion in the FY 2016 Omnibus to $2.1 billion in the FY 2017 budget request. 6.2 applied research fares slightly better, receiving a 3.6 percent cut, going from $5 billion in FY 2016 to $4.8 billion in FY 2017. DOD 6.3 advanced technology development receives a 2.6 percent cut, going from $5.7 billion in FY 2016 to $5.6 billion in the FY 2017 request.

DARPA would see a 3.7 percent increase, going from $2.87 million in FY 2016 to $2.97 million in FY 2017.

This is a pretty disappointing DOD S&T budget request, but it is not unexpected. The leadership at DOD knows that defense S&T is a Congressional priority, so they use a little gamesmanship and remove money from S&T, expecting Congress to put it back during the budget process, and use it to fund other areas in the request that are not Congressional priorities, in the hope that some of that money will stick during appropriations. The obvious problem with this strategy is the chance that Congress won’t put those funds back into S&T. And while the good news is that this request is unlikely to be passed as is, it is still a difficult place to start the process.

Department of Energy

The two key parts of the Department of Energy (DOE) for the computing community are the Office of Science (SC), home of most of the agency’s basic research support, and ARPA-E, or the Advanced Research Projects Agency-Energy. For SC, the president’s FY 2017 plan includes a very healthy increase of 6.1 percent, using mostly discretionary funding, increasing to $5.67 billion (its FY 2016 budget is $5.35 billion). ARPA-E would see a huge increase of 71.8 percent, growing from $262 million in FY 2016 to $500 million in FY 2017, but the great bulk of that increase comes from mandatory funding.

Within the SC budget, about one-third of the requested increase ($100 million) comes from additional mandatory funding, but it is only within the “University Grants” spending line. As the name implies, this is for funding for “competitive merit-based review of proposals solicited from and provided by the university community.” If you remove that $100 million from the $325 million DOE SC is slated to receive in the president’s plan, the office would still receive a healthy 4.2 percent bump, or $225 million over FY16.

There is some reprogramming within the computing lines of the office budget that’s worth noting. The Advanced Scientific Computing Research (ASCR) program within the Office of Science, where most of the computing research at the agency is located, would see a healthy increase of 6.8 percent (or $42 million more than FY16) in the president’s plan. The majority of that increase is slated to go into a new program line focused on the exascale computing program. As a result, funding is “reprogrammed” from the mathematical, computational, and computer sciences research program lines in ASCR and transferred to the new exascale line. Whether this reprogramming will change the character of the work from research to more development-oriented work related to exascale, or whether this is just a reclassification of research which is already focused on exascale problems under a new program line, remains to be seen, but it’s something that bears watching. The bottom line, though: It appears exascale gets almost all of the increase in ASCR’s budget, and the rest of the program is flat funded.

With regard to ARPA-E, the great bulk of its massive requested budget increase, of $209 million over FY 2016, would be funded using mandatory spending. The goal, according to Secretary of Energy Ernie Moniz, is to put the agency on a path to a $1 billion budget within 5 years, as recommended in the National Academies Rising Above the Gathering Storm report (which is where the idea for ARPA-E originated). However, mandatory spending accounts for $150 million of that planned $209 million increase, so the chances are somewhat slim that the agency will grow as the secretary hopes and the president’s plan calls for. Additionally, large increases in the ARPA-E budget request have been almost a tradition since the agency was founded, with Congress rarely approving the president’s request. Removing the mandatory spending from the request, the administration is still calling for a large 21.5 percent increase for ARPA-E, or $56 million, which would bring the agency’s budget to $318 million (it received $262 million in the FY16 Omnibus).

Our Final Analysis

It is not hard to understand why the administration sought to get a little creative in this request given the cap constraints. But this approach—taking an end-run around the discretionary budget caps by designating new “mandatory spending”—is somewhat troubling for the signals it sends to Congress and, in particular, the appropriators. What they likely see is not that the president has justified a new way to pay for science, but instead that he wouldn’t prioritize science investments under the discretionary caps. It appears that in this budget, other programs were more deserving of the discretionary funds.

We knew, given the tight caps for FY17, that this budget cycle would be challenging for science. It’s unfortunate that we start even further in the hole with this budget request. The good news, if we can call it that, is that the budget will most likely not be passed until after the November presidential election. Depending on who wins in the fall, it could radically change what a final budget looks like. The community will want to keep a close eye on how things play out, so be sure to check the CRA Policy Blog ( for new updates.