On its face, the proposed 2021 budget unveiled this week for the U.S. National Science Foundation (NSF) looks like a disaster: An overall reduction of 6.5%, including a 7.8% cut in its research programs.
But a closer look suggests outgoing NSF Director France Córdova has made it as easy as possible for Congress, the final arbiter of federal spending, to lessen the sting. She has crafted her last budget to appeal to both Democratic and Republican lawmakers—and left holes that those legislators will likely want to fill.
Córdova steps down next month at the end of her 6-year term, so she won’t be around when the final decisions are made, probably after the November elections. But she’s staking her agency’s prospects on the bipartisan support NSF has traditionally enjoyed. Here’s how.
Industries of the future
Despite a bottom line that is $537 million below its current level of $8.28 billion, NSF’s proposed 2021 budget gives a big boost to two White House priorities—artificial intelligence (AI) and quantum information science (QIS). Those increases, a near doubling over 2019, should appeal to lawmakers who have already shown that they see these two disciplines as key drivers of U.S. economic growth and national security. For example, a bill introduced last month by Senator Roger Wicker (R–MS), the chairman of the Senate commerce and science committee, calls for a 2-year doubling of research funding in these areas, dubbed “industries of the future.” The Senate’s top Democrat, Charles Schumer (NY), is thinking of proposing something that could be even more generous.
Wicker’s plan, which has attracted Democratic co-sponsors, is not a spending bill. But it could influence the appropriators who write bills funding specific agencies. That’s also the goal of Representative Frank Lucas (R–OK), the top Republican on the House science committee. His recently introduced bill would back a 10-year path to doubling the budgets of several agencies, including NSF, that support research fueling these industrial sectors.
For 2021, six of NSF’s seven research directorates plan to invest a total of $868 million in AI compared with $465 million in 2019. For QIS, the request is for $226 million, up from $106 million. (The comparisons are with 2019 rather than 2020 because NSF’s budget provides no breakdown of 2020 spending levels beyond the total for each of its five top-level accounts.) This year, NSF expects to fund the first six in a new network of AI research institutes at $20 million a pop over 5 years—and its 2021 request would allow it to set up several additional centers.
Not surprisingly, NSF’s computer and information science and engineering directorate (CISE) is managing the biggest share. But its education, biology, and social sciences directorates are also taking part, for example, with an augmented learning institute that will be part of the new network.
NSF’s big ideas
The Democratic-controlled House of Representatives is likely to be suspicious of a narrow focus on AI and QIS. Its members have traditionally opposed attempts to skew NSF’s expansive portfolio in a particular direction, arguing that research in one field often has a serendipitous impact on other, seemingly unrelated, fields.
But Córdova has also found enough money to maintain momentum for an initiative with broader appeal: NSF’s 10 big ideas, her vision for where research is headed. The topics, which include Navigating the New Arctic, the Future of Work, and Windows on the Universe, would require sustained investments across a much wider array of disciplines.
In an administration that has tried to cut NSF’s budget 4 straight years, the new label is another way to win support for additional funding. Along with the breadth of NSF’s big ideas, congressional Democrats tend to like the prominent role many give to environmental researchers and those in the social and behavioral sciences, areas often viewed with skepticism by Republicans.
The phrase “big ideas” is actually an umbrella term covering several activities that, in toto, would receive $432 million in 2021, or $108 million more than in 2019. One of NSF’s big ideas is not actually a research frontier but rather an attempt to speed up the payoff from all types of basic research. It’s done by connecting scientists from many disciplines, a process NSF labels convergence research. NSF has also begun to test that approach in two areas by funding what it calls convergence accelerator (CA) projects.
NSF is now funding CA projects built on Harnessing the Data Revolution and the Future of Work at the Human-Technology Frontier. The 2021 request contains $35 million for a third CA track. Córdova has also proposed spending $230 million on the six research “big ideas,” up from $180 million in 2019.
That growth helps soften the blows to NSF’s six traditional research directorates in the 2021 request. CISE racked up the only increase, at 8%. Mathematics and the physical sciences would drop by 3%, less than the agency-wide 8% decrease for research—but biology, geosciences, and the offices of international and polar research would suffer double-digit cuts.
Not so fast
To win back money from legislators, Córdova’s 2021 request also employs the time-tested strategy of paring activities that Congress will almost certainly want to fund at a higher level. At the top of the list is the established program to stimulate competitive research (EPSCoR), which earmarks money for 27 states and territories to supplement the tiny amount of money they win through NSF’s traditional process of competitive grants.
Its goal is to level the playing field, but 40 years in, most of those states continue to receive EPSCoR funding. Not surprisingly, EPSCoR is hugely popular among legislators from those states. So NSF’s plan to shrink it by 17%, from $190 million this year is unlikely to prevail.
Nor is Congress likely to turn its back on a program that gives scientists the skills to turn their research into a company. Started in 2011 with a $1 million budget, Innovation-Corps (I-Corps) is arguably the most successful legacy of former NSF Director Subra Suresh. It has spawned more than 500 startup companies. The 2021 request would cut its current $38 million budget by 17%.
The 2021 budget proposal also puts NSF’s flagship graduate fellowship research program (GRFP) under the knife, proposing a 20% drop from 2000 annual fellowships to 1600. Begun in 1952, the fellowship program is a pillar of NSF’s investment in training the next generation of scientists, and several labor economists have argued that its success warrants growing it to 3000.
Córdova, however, has repeatedly proposed trimming GRFP since President Donald Trump took office, in some years down to 1500. In 2017 she said it was time to compare its impact with that of other mechanisms for supporting graduate students, notably traineeships and through research grants. And NSF’s current program announcement says the agency plans to make 1600 awards in fiscal years 2020 and 2021.
The Senate has rejected previous attempts to trim the program, however. For example, the final 2020 spending bill approved a few months ago ordered NSF to stay at 2000 slots. An NSF spokesperson says the agency will comply with that language.
A race against space
Despite the Congress-pleasing budget that NSF has constructed, federal legislators may find it harder to restore the overall cuts than in years past, much less give NSF any increase.
That’s because the appropriations process puts NSF in direct competition with NASA, which is one of the few science agencies that would grow significantly in Trump’s 2021 budget request. The president requested a $3 billion boost to help NASA meet his goal of returning astronauts to the Moon by 2024, 4 years ahead of the current schedule.
NASA isn’t the only challenge for NSF. Last summer, Congress and the White House agreed to keep domestic spending basically flat in 2021, at $634 billion, after it had grown by $32 billion in 2020 over the previous year. Trump ignored that agreement and requested only $590 billion in his latest budget.
Congressional Democrats have vowed to hold to the deal. But even the agreed-upon number for domestic programs will force legislators to play a zero-sum game.