Everywhere you look in Washington, D.C. today, there are signs that industrial policy is back. The Biden Administration has characterized its main legislative achievements as an expression of a new industrial strategy. Influential commentators argue that industrial policy is the backbone of a new political-economic paradigm. The presidential election is, among many other things, shaping up to be a fight for the soul of this new paradigm—a fight not over whether we will have industrial policy, but over what kind of industrial policy we should have.

The broad turn makes sense: The neoliberal paradigm has exhausted itself. But what will take its place? What is industrial policy, even? What would it take for the state to successfully undertake industrial policy, given where neoliberalism has left us—with massive private power, weakened public power, and inequality so persistent and extreme that it is undermining democracy itself? And is industrial policy a narrow economic program, or can it be something broader, that helps revitalize and deepen that democracy?

There are no more critical questions in this campaign season, or for the years to come. The skirmishes over the implementation of the Inflation Reduction Act in fact point to deeper conceptual questions that we must answer as new paradigms come into view. Before we know, for example, whether clean energy investments should come with conditions that empower workers, or firms receiving subsidies should have to provide child care or swear off stock buybacks, we need to know what industrial policy is, why we should want it—and then, what its preconditions are.

In a new paper, out soon in the Harvard Law and Policy Review, Joel Michaels and I argue that we should welcome a new industrial policy, if it is a democratic industrial policy. Ours is not a paper of political theory, but it does argue that there is a particular conception of democracy—a substantive, materialist one—that we should embrace here, because it is broadly shared by progressives and important to describing the problems of our politics and political economy today. We also explain that undertaking democratic industrial policy effectively, in a world remade by neoliberalism, means two things in particular: deliberately cultivating greater administrative expertise and power, and at the same time building more countervailing power among workers, communities of color, and others who are structurally marginalized. This kind of industrial policy will strengthen both our democracy and our economy, by increasing our ability to bring forth the economy that we want and by building more organized power among communities that need authority if we are to make good on promises of inclusion that have long been unfulfilled.

So what is industrial policy anyway, and why should we want it? There is a tendency for academics and policy wonks to either dismiss the idea, arguing that all economic policy is industrial policy, or to define it narrowly, as referring only to policies designed to promote an internationally competitive manufacturing sector. The former makes the term useless, and the latter confines its aims too narrowly and technocratically. Industrial policy should be understood more capaciously—as the deliberate attempt to shape different sectors of the economy to meet public aims. This is what we call democratic industrial policy.

The very idea that economic policy should only serve the actuarial metric of efficiency is not particularly democratic. It makes sense if you think the economy is a kind of machine, and that concepts like efficiency are neutral expressions of the natural facts of economic life—ideas that came into favor in the neoliberal period, promoted by prominent scholars in law and economics. But that’s not what an economy is, as the recent scholarship on law and political economy has helped to show. Economies are made, sustained by laws and policy choices that make things like contracts, banking, and modern employment and corporations possible. In a democracy, the economy should meet broadly shared public aims. These aims may include international competitiveness or well-paid manufacturing jobs, but today most jobs are in the services, so jobs policies must extend well beyond manufacturing. If we want health care, home care, and child care, then these too require supportive economic policy. Industrial policy in a democracy might better be called developmental policy: It is the practice of using government to create the economy that people want, by promoting and shaping certain sectors (like care or green energy) and demoting or curbing others (like markets for cryptocurrency). It shapes sectors using many of the ordinary tools of economic policy, including regulation, but also has certain distinctive tools of its own, like contracting, grantmaking, and government-owned enterprises.

Part of the reason industrial policy is rightly back is that new evidence and scholarship have revealed the flaws of the neoliberal paradigm, and shown us that government must play a significant role in an economy like ours, and that it can do so effectively. Today, it is ever clearer that good jobs, a more fair and dynamic economy, and more resilient energy and care infrastructures require a vibrant role for the state. This is not just because the neoliberal “free market” is tilted toward the privileged or lacks both the fluidity and resilience we need (see the recent supply chain disruptions and energy shocks). It is also because the state alone is well situated to finance long-term and high-risk innovation, to overcome coordination problems, and to invest in the development of interdependent goods, like the workforce and roads and subsidies needed to build semiconductor and clean energy sectors. Scholars have also shown that many forms of industrial policy have been quite successful—especially once we recognize that a program may have aims beyond actuarial efficiency, like protecting communities and jobs, and that its success must be evaluated accordingly.

There are also political reasons that industrial policy is likely here to stay. In Congress, the budget reconciliation process makes spending (a key industrial policy tool) easier than other kinds of lawmaking. And while broad-gauged taxes like carbon taxes were celebrated in the neoliberal period, they have turned out to be extremely difficult to implement politically. Spending, grantmaking, and even sector-specific regulation appear more feasible.

If industrial policy should be democratic, what does that imply about how it should be implemented? There are many traditions of democratic thought, but any compelling account of democracy today, we think, will describe it as having at least two important requirements. One is the polity’s collective capacity—for example, in the form of public administration—to achieve common goals. The other is a measure of material equality that enables citizens to make credible claims about what those goals ought to be. Because many in our democracy today face forms of material, structural subordination, a key task for those who would defend democracy is to help build systematic, countervailing, organized power among groups so subordinated. And just as these requirements are key to democracy writ large, they are also key to how a democracy designs its industrial policy.

Decades of neoliberal retrenchment have undermined state capacity and deepened inequality. As a result, policymakers must have two clear aims in mind, always, in designing industrial policy, if it is to be both effective and democratic. First, they must build administrative expertise and capacity—for example, by writing conditions into grants and contract licenses and experimenting with public options and ownership of various kinds. Second, they must build the organized capacity of structurally disadvantaged groups to exercise collective influence, which we call countervailing power. Both are major tasks, and they must be undertaken at the same time. They are in productive tension with each other, and that is part of the point: A more powerful state bureaucracy or industrial policy will not enhance democracy without more credible structures of accountability, particularly for those with the least power in our democracy. We will need their help to influence not just government but also the firms that industrial policy tends to empower.

Our paper describes the kinds of tools we have in mind and points to examples that policymakers can draw upon. To build administrative power, government can, for example, invest in data development and collection (for example, about economic conditions or the structure of important supply chains) and condition contracts on firms’ agreement to share key information with government and workers. There are some serious legal issues to navigate, which we discuss as well. For example, intellectual property law today generally allows those who first capture data or invent things—even on the government’s dime—to control the results. Drawing on and expanding examples like the march-in provisions of the Bayh-Dole Act, which give the National Institutes of Health the authority to make or buy medicines developed via government grants as a condition of that funding, government can flip these entitlements, building its own capacity and authority to act.

We also define a continuum of tools that can facilitate more administrative control, from open market operations on the one end to public ownership on the other. This spectrum has many, many options in between, and this is the heart of good industrial policy: not only creating conditions for those receiving subsidies that protect public aims, but also exploring partial public ownership stakes, like equity shares or patent ownership. Rebuilding an administrative state that has expertise and capacity will require exploring options all along this continuum. We also must attend to how these structural design choices interact with government knowledge and expertise, as we now realize that outsourcing and privatization can not only lead to cream-skimming (when a contractor serves a select cohort, such as healthier patients, but neglects others who are more expensive to serve, because it’s most profitable) but they can also amplify private power while undermining the capacity and expertise of government itself.

A government increasingly empowered in this fashion must also be held accountable. But the traditional tools of administrative law—such as notice-and-comment rule-making, or transparency requirements—were developed with rule-making agencies like the EPA in mind, and are not well designed to discipline government when it is acting as a firm, a contractor, or a grant-maker. Ad hoc negotiations with firms are difficult to discipline with public comment, for example. And disclosure of the reasoning behind the government’s bargaining goals may undermine the government’s negotiating position, implicate business confidences, or elongate the timeline of negotiations.

Countervailing power, we think, is therefore essential both for holding government accountable in this context and for checking and shaping the acts of firms. We identify a host of modifications—from grant or contract terms to federal advisory committees—that can operate to encourage countervailing power. This can be done in three broad ways. First, agencies can seek to empower disadvantaged groups with administration itself (for example, by giving representatives roles on advisory committees, or creating forums for tripartite bargaining). This has already been done in some contexts, including the Federal Reserve regional banks and the Fair Labor Standards Act wage and hour boards. Second, countervailing power can be achieved in the realm of production (for example, by negotiating project labor agreements or community benefit agreements with firms). Third, it can be attained through diffusion of benefits to disempowered communities (for example, by disproportionately sending resources there, as the Biden Administration’s Justice40 initiatives seek to do).

There will, of course, be critically important questions about precisely who requires and can wield countervailing power in a particular context, and how to deal with the risk of factions and capture. But there has been so much attention to the risks associated with community participation in industrial policy that we have lost sight of a far greater risk: that major new industrial policy programs will simply amplify private power in our economy and run roughshod over disempowered groups. We also point out that many forms of empowerment (like those mentioned above) do not function to allow single individuals—the proverbial NIMBY litigator—to block policy, but instead can give voice to collectives that credibly can build support for, and even accelerate, good policy.

Those concerned more broadly about the survival of the administrative state after the recent attacks from the Supreme Court in decisions like Loper Bright Enterprises v. Raimondo, which gives judges much more power to set aside agency regulations, also have reason to look to industrial policy, and to a materialist and power-shaping approach to administration today. As the old administrative law is being taken apart, it’s important to recognize that there were many parts of the administrative it did not fully reach. Core aspects of industrial policy are in this category: Things like contracts and government ownership fall well outside of the realm that rule-making and judicial review are designed to discipline.

Scholars of the administrative state have long been calling for a new paradigm—something better than painfully slow and only ostensibly neutral procedures like notice-and-comment rule-making. These were the hallmarks of the “neutral process” school in administrative law, which too often treated politics like a debating society, and did not attend to material power imbalances in the world. It ignored the fact that some—for example, firms, or white homeowners—have a lot more power and money than others, such as workers, tenants, and unhoused people. And so it tended too often to amplify rather than counteract concentrated power, allowing the privileged to bend processes to their will, by flooding agencies with comments or requests or by strategically using litigation to block popular programs.

It is time to imagine a revitalized administrative state that operates with less reliance on such ostensibly neutral procedural rules, one that recognizes that accountability may also require—or demand—efforts to build countervailing power and administrative power while disciplining corporate power. Percolating in our industrial policy paradigms may also be new paradigms of administrative accountability and administrative law that are oriented less to ostensibly neutral process and more to the sharing of power, in a manner that openly recognizes, rather than assumes away, deep inequality. Given the importance today of shaping our economy and urgently bringing about a more materially fair, equal, and sustainable democracy, we need these now more than ever.