The US government has acted as major contributor to science research since the mid-20th century, both in terms of broad basic research and targeted projects. As industrial policy has gained traction, especially during the Biden Administration, the distinction between industrial and science policy has become increasingly obscure. Hybrid policies like the CHIPS and Science Act have spurred continued debate surrounding role and value of federal funding for science research.
Today on Political Economy, I talk to Tony Mills about American science policy past, present, and future.
Mills is a senior fellow here at AEI and director of the Center for Technology, Science, and Energy. He is also a senior fellow at the Pepperdine School of Public Policy and a scholar associate of the Society of Catholic Scientists.
His new paper, “Recovering Science Policy,” explores the blurred lines between US industrial and science policy in today’s political landscape.
What follows is a lightly edited transcript of our conversation. You can download the episode here, and don’t forget to subscribe to my podcast on iTunes or Stitcher. Tell your friends, leave a review.
Pethokoukis: Before we dig into the paper, I wonder if you could just briefly tell us: What is federal science policy? There didn’t always used to be a science policy, then we had a science policy. I wonder if you give me the quick history of US science policy in 45 seconds.
Mills: Sure. Roughly, I would say federal science policy is the area of public policy concerned with the federal government’s role in science, the broadest level. A little bit more specifically, what federal policies we should have for funding, conducting, directing, or utilizing scientific research. Considered sort of liberally, you might also include in this questions about the role of scientists themselves and scientific expertise in the policymaking process. So the role of scientists as federal advisors, either in the White House or in federal agencies, and the way in which scientific knowledge interfaces with the federal policymaking process. That’s the overall sketch of what science policy is.
As you say, it’s sort of interesting, we haven’t always had science policy. It really appears on the scene, I would say, in the middle of the 20th century, in its modern form, although you could really trace it — and I would trace it — back even farther to the late 19th century.
But in its sort of modern, recognizable form, it’s sort of a creature of the World War II era and the following years. It really was, I would say, one of the prominent areas of public policy in those days, especially in the 1950s and ’60s, and this is kind of an informal history, but I would say sometime in the late ’90s, early 2000s, what we now think of as technology or innovation policy, really kind of supplants science policy. And today we’re actually seeing a little bit of a renaissance of science policy, and you can now find, at think tanks around Washington, including AEI, a kind of discrete area of science policy, which you would’ve found, say, in the 1990s at institutions like AEI and Brookings. You’re starting to see that again now. But there was a stretch of time when technology policy was ascendant and encompassed into it some of — maybe not all of — the kinds of policy questions that you would traditionally have associated with science policy.
I guess when I think science policy, I think government giving money to scientists to think big thoughts, to do very fundamental research on physics, or chemistry, or neuroscience. I think of that as a kind of a basic research science policy. Then there’s other kinds of policy that might be better called technology policy, and then we have something called “industrial policy.”
Can you give me even shorter explanations of the difference between government giving money to some neuroscientist to investigate brain synapses or something, and technology policy, and then even industrial policy?
There’s a kind of straightforward answer to that question, which might pick out areas of research or aspects of economic activity and differentiate between science, and technology policy, and industrial policy, but lurking beneath that, I think, are deeper philosophical questions, which are part of what I’m trying to get at in the paper that you mentioned, which maybe we can get into later.
But roughly, when we think of science policy in its traditional form, we’re talking about things like: What is the appropriate role of the federal government, federal dollars, in scientific research? So we have a National Science Foundation and National Institutes of Health and other federal science agencies whose purpose, in part, and in the case of the NSF, primarily is to use public money to fund scientists to do research.
Most of the time we’re talking about scientists who are domiciled in the academy, they’re professors at universities, and they are conducting research, they submit grants to these agencies, the agencies give them money if they’re lucky, and then they use that to conduct experiments and other kinds of research. The idea, of course, on some level is that this is generating a kind of economic or social return. There’s a kind of social contract implied in that.
We didn’t always do this. In fact, the idea of using public money to fund science, once upon a time, was quite controversial. It was seen as potentially unconstitutional, even. But over the course of a 150-plus-year history, we developed this infrastructure for federal-supported scientific research. And science policy questions would turn on: What kind of research should we fund? How should we fund it? How do we allocate resources? What sorts of mechanisms do we use to decide that? Who should decide? Scientists? Should we rely on a peer-review process for deciding what projects should be funded, or should we do something else? A lot of science policy operates within that realm.
But as you move sort of on a spectrum toward what we sometimes call “applied research” or technology, we get into domains where we’re less concerned with advancing the frontiers of our basic understanding of nature, but of building things and trying to invent new techniques, and technologies, and tools. Here we’re entering more the realm of technology and technology policy, which obviously overlaps, but has its own set of concerns. And as I would say, you move farther down that continuum into particular areas, sectors of the economy, that are reliant on certain technologies or that use them for certain purposes, we’re now getting into a domain that we traditionally call industrial policies.
So a really good illustration of both the differences between, and the way in which these various aspects are often run together, is the CHIPS and Science Act. The CHIPS and Science Act included a lot of different things in it, but, at its core, it was two different legislative efforts. This was a bipartisan bill that was signed by President Biden in 2022, and it had two basic purposes: The major one was to reshore the semiconductor industry, essentially. The other piece was to boost federal research, to increase spending on science and technology, to allocate more money, in particular, for the National Science Foundation. And this came from two different bills that were fused together. I would characterize the Chips part of that bill as pretty solidly, clearly, an instance of industrial policy. We could talk about what that looked like and whether it’s good or bad. And then the other piece, the science part, is very clearly a kind of classic issue of science policy, where you’re talking about taking federal money, giving it to an agency that’s then going to decide how to disperse money to researchers. But within that debate around whether and how to include this science part of the bill, there was this tension between science and technology.
Running through a lot of legislative history very quickly, what you saw was the Senate introduced a version of what was then called the Endless Frontier Act. It was very oriented toward technology. The idea was, there are certain areas of technology that we want to be competitive in. We’d like the federal government to do more to help us be competitive in those areas. The original proposal was $100 billion dollars that was going to go to the NSF, whose budget at the time was something like $7 billion. All of that money was going to go to a new technology directorate that was going to build out these new areas of technology or support them.
The scientific community was actually not all that happy with this bill, sort of surprisingly, you might think here’s a massive check — and the response was actually, that’s not really what the NSF does. We do some of that, but actually our primary mission is to support basic scientific research. There was a kind of negotiation that went on — the House version of the bill reflected the priorities of the scientific community much more than the Senate version, and you had this sort of back and forth, and what you wound up with was a compromise that left a lot of people unhappy. But what I think you could see through that process was not only that these are different areas of policy — It’s one question to ask, how can we as a country be competitive in artificial intelligence or semiconductors? It’s another thing to ask, how do we ensure that the next Einstein is here in America doing great scientific work, discovering new theorems, new fundamental theories of the universe? And they implicate different constituencies, different institutions. Are we talking about private sector companies, large corporations or entrepreneurs, or are we talking about physics professors at universities who are trying to get grants from the NSF to do research on things that may not have any obvious utility in the short term?
There’s obviously going to be some overlap, and especially now when we’re so concerned about geopolitical competition, China getting ahead in certain technologies. Is your concern that the basic science piece is just going to get lost, or swallowed up, or de-emphasized when that’s the thing producing the seed corn for all these other advances?
That’s certainly a part of my concern, but I would say two things, one is a kind of practical concern and one gets to the deeper philosophical issues that we were talking about or alluding to. The practical level is what I was alluding to. So the process by which we got the CHIPS and Science Act was, in a certain sense, it’s just good old-fashioned Washington politics: a lot of money at stake, a lot of constituencies, a lot of people get involved and want a piece of the pie. But on another level, it was a foreseeable and avoidable mess, because the policymakers were not attentive to these kinds of distinctions. They weren’t attentive to the different constituencies involved in the institutions, what their purposes are, their goals, and I think just bagging all of this stuff together can generate sloppy, and sometimes counterproductive, policy.
I just flag one kind of example here, which, Jim, you and I have talked about before, which is the fetish for the moonshot. And this really dates to World War II with the creation of the Office of Scientific Research and Development, which oversaw the Manhattan Project and other really successful essentially weapons programs. And then the post-war years, and we had Apollo, we had these large-scale, I would describe them as engineering projects that were in some way or other directed or facilitated by the state.
The success of those has, since that time, been a blueprint for a lot of folks in Washington to say, “Well, we did this with the moon, we did this with the nuclear bomb, why can’t we do that for . . . ?” pick your favorite problem. And the difficulty is that technical problems are not all the same, and different areas of research and innovation are not all the same, and they don’t necessarily lend themselves to the same kinds of policy interventions. So a moonshot model might work for a given kind of problem and be completely inapt for another kind of problem. So you can have these sorts of practical issues, if we’re not sensitive to the distinctions, but I think there are also deeper questions lurking here, which is what this paper tries to get at.
What’s interesting is, the origins of science policy really turn on the question of the distinction between science and technology, and that debate over science and technology was a part of an even wider-ranging debate about the role of a state, the nature of intellectual inquiry, the possibilities or dangers of central planning. And it’s quite interesting to look at that history now and see how views that — to put it provocatively — once upon a time you would find on the Marxist left, or the left that was Marxist-friendly have almost become come ubiquitous across the political spectrum.
These figures like Vannevar Bush, who oversaw the OSRD during World War II and others that I mentioned in the paper, like the scientist-philosopher Michael Polanyi and the sociologist Edward Shills, what they were worried about was the effort to emulate the Soviet Union, which at the time had explicit programs for central planning. In the Soviet Union — this is an anecdote that Michael Polanyi tells, when he visited the Soviet Union in the 1930s, I believe it was, he was told by a Party official that there’s no such thing as “pure science” in the Soviet Union, because all science ultimately serves the goals of the Party’s five-year plan. And Polanyi was really horrified by this and worried that this would fundamentally undermine the scientific enterprise and the very idea of autonomous intellectual inquiry, which he thought was basic to the liberal social-political order.
And so efforts in Western Europe and the United States to emulate aspects of that Soviet-style planning were seen as dangerous by some of these figures that I’m talking about, not just because it would lead to a bad allocation of resources — although they thought it would — but there was a deeper moral, political issue at stake, which was the need to defend scientific autonomy as something essential to free society. So I think there are broader issues at stake, and we tend not to think about them in the context of sort of dry questions of NSF funding policy, but one of the values of going back and looking at the history is seeing the way in which these figures recognized the political stakes of these seemingly finer philosophical nuances and how they interplayed with practical questions of policy.
Do you think government should be funding basic research? Should it be funding a National Science Foundation? Do you have any qualms about that?
Yes, I do think the government should be funding basic research, at a bare minimum, but I also do have qualms about how the current system operates. Just to stick with the history piece for another minute, one of the things that I think is interesting about looking at this history and the figures involved is that they had complex views. They were nuanced thinkers. So we sometimes think very black and white: The state should do this or the state shouldn’t do it, and it’s sort of a black-and-white question. And here in DC, in the think tank world, there are certain strains of libertarianism that would say, the state just doesn’t have a role here at all. Why would we want to have the state funding science? Can’t the market take care of that?
What’s interesting about figures to me like Bush and Polanya, whom I mentioned, is that you wouldn’t describe these guys as statists, right? Polanyi is often considered one of the intellectual sources of libertarian or broadly classically liberal thought. He was friendly with Hayek, he kind of came out of that world, had connections to the Mont Pelerin Society, not a big statist-type thinker, but it doesn’t seem to have ever really occurred to him that the state had no role to play in supporting science. And Bush was the same way. He recognized — Bush, that is — especially coming out of World War II, that we were entering a completely different world. In the late 19th century up until the middle of the 20th century, the role of the state was quite limited. And what Bush saw was that we weren’t going to go back, and that that wasn’t bad, that there were lots of good things that the state could do. The question was how should it do it and how do we mitigate the downsides of a system of state-managed science and scientific research?
There were lots of concerns in the 1950s and ’60s and ’70s about the limits of that system. Today in policy land, we often look back at that as a golden age. We had the moon landing, we had all this federal funding, and everything was wonderful, but it was a really fraught time and it ended very badly with a profound crisis — economic and legitimation crisis — in the 1970s. And one of the concerns — there were lots of concerns, obviously, with that overall crisis — but one concern in the realm of science was we had created this sort of bureaucratized, state-managed complex that threatened to erode the traditional autonomy of science, that would lead to uncreative, middling scientific research that would become sclerotic, and so on and so forth.
These are a lot of the kinds of concerns that people in science policy have today, and a lot of the most insightful thinkers, in that moment, saw that these were problems that were kind of inherent in the system that had been created, and they want to find ways to counteract them. I think we’re in a similar kind of moment now, where what we’re asking is not, “Should the state be doing this?” because you can’t really find anyone in Washington that wants to eliminate the NSF. The question is, “How can it do it well, while guarding against some of the downsides?” And we can talk more about what those look like in the contemporary scene, if you’d like. But that would be my answer, is that it’s not a simple one, it’s sort of “both-and” in something like the way that Bush and Polanyi recognized as well.
I think that what you just said suggests two questions here as we get near the end: One, should we be spending more on science research in this country by the federal government? Maybe you could tell us how much we actually increase it from the CHIPS and Science Act.
Two: The kind of reforms, and how we go about spending that money, and how we allocate that money. So there’s sort of a funding question I’m giving you, how we allocate that money, and how we actually operate that system so we get the most bang for the buck.
In terms of what happened, the Chips part of the bill did far better than the science part of the bill, is the short version — authorization versus allocation. What the NSF wound up getting was nothing like — certainly nothing like the $100 billion that was initially floated — I t wasn’t even what was initially authorized by Congress. And now we’re in a kind of new moment where all sorts of different forces are conspiring in ways that could really . . . We just had an election, we’ll see what happens, but you could imagine a trajectory where the kind of growth that had seemed like we might get circa 2020 is not really in the offering, but we’ll see.
For my part, I think federal funding is important, and I would like to see the federal government do more, to spend more, particularly in basic scientific research, but not only there, but I would say with two caveats, maybe three:
One is — again, going back to the history —folks like Bush and Polanyi, what they recognized was that money is not the answer to everything, and actually more money can sometimes create more problems. Bush himself, he’s remembered as writing The Endless Frontier about how the search for science is this endless kind of adventure, and some people have caricatured that document as “Endless Expenditure,” it was just giving free license to the government to just keep funding and funding. That wasn’t his view at all. He was really worried that an overextended system could, in his mind, lead to mediocrity and all other sorts of problems.
If we look in more recent history at periods when we had exponential growth, like the doubling of the NIH, which did lots of wonderful things, it does also create real problems. That kind of rapid growth can create difficulties for the institutions that are the beneficiaries of them, particularly when it’s followed by a complete drop-off or a flattening. So there are practical questions about how much money and the effects that that can have on these institutions.
But I think that the bigger point here is: Money’s not sufficient. The idea that you can simply just spend more money and get all the wondrous things you want is a fiction, and it’s a fiction that scientists like to trade in because it gets them more funding. So what I would suggest is tying funding increases to genuine reforms. To do that, the reality is we don’t have a kind of blueprint we can take off the shelf for, “Here’s how we ensure we get flourishing scientific research.” There are lots of different models for thinking about this that are specific to different domains. What I would suggest is a kind of pluralistic — we have a pluralistic system already; I think we should embrace that pluralism, by which I mean we have different kinds of institutions that fund different kinds of research in different domains for different purposes. The NSF funds research, primarily basic research to advance our understanding of nature. The NIH does a lot of that, it also spends a lot of money on clinical research and more applied research. We have the Department of Energy, we have a whole litany of federal agencies that are involved in this, and there is a benefit to pluralism because science and technology are pluralistic domains. I think we should tie the reforms not just to increases in spending, where appropriate, but to the particular domains in which we’re interested in making advances and be somewhat experimental about how it is that we allocate this money.
The traditional mechanisms that we’ve relied on — the main one being the peer-review system — has lots of problems. I’m actually a defender of peer review when used in the academic realm as a way of vetting scientific ideas and publication, but when it comes to making decisions about how to allocate public resources, it’s not clear that reliance on peer review is the only, or the best, mechanism. So that will be just one area where we could experiment with other policy options to try to generate a little bit more competition into the system that we currently have.