Industrial policy was once so out of fashion that it was jokingly called “the policy that shall not be named.” Now it’s back in a big way. On issues ranging from clean energy to semiconductors to Covid-19, governments are trying to improve the performance of key business sectors. Can they manage to do so without subverting competition and subsidizing special interests?
This article is the second in a ProMarket series about industrial policy. Stay tuned as we publish an article each week this quarter on the topic.
An “industrial policy” is government action that encourages or directly subsidizes the expansion of some economic sectors over others. Dani Rodrik famously suggested that the core question about industrial policy is not “why?” but “how?”
Few mainstream economists believe either that all pricing and resource allocation decisions should be made by the state, or that markets will flourish in a setting with no state contract enforcement, police protection, or dispute adjudication at all. That means, as Rodrik suggests, that some substantial commercial activity will generally be encouraged, with some degree of state direction or steering operating in the background.
Industrial policy is different from regulations that seek to limit or correct “market failures” such as externalities or information asymmetries. In the case of regulation, state action is focused on restoring Pareto optimality to commercial settings, by imposing taxes and subsidies or licensing requirements. Industrial policy, by contrast, is an evaluation of the direction and speed of the overall growth path of different sectors of the economy.
Capitalism automatically has a self-executing industrial policy, described by Joseph Schumpeter as “creative destruction.” Profits are a sign that additional resources devoted to a sector produce greater social welfare; losses are a sign that the opportunity cost of resources being expended in that sector would be better reallocated elsewhere. Industrial policy operates on the margins of these investment decisions, offering subsidies and tax breaks for “good” industries—green energy in the U.S. right now, for example—and increased taxes on industries (such as tobacco products) that the government seeks to limit.
But industrial policy can also be more muscular, and intrusive, when it focuses on manipulating and improving industry structure, and encouraging workers to leave some sectors in favor of another. In the 1980s much of progressive Washington swooned at the genius of the Japanese industrial policy that focused on a few large export industries. In the 2000s the same kind of admiration was heaped on the Chinese government’s “investments” in enormous manufacturing facilities for textiles, plastics, and electronics.
Skeptics about industrial policy often take an Austrian perspective, which doubts the bureaucrats, or anyone else for that matter, would have the information necessary to direct an economy better than the signals of profit and loss, or a Public Choice perspective, which doubts that elected officials and the bureaucrats who report to them have any incentive to direct an “optimal” industrial policy even if they possessed perfect information about the future.
These two objections, bad information and wrong incentives, are not mutually exclusive and in fact reinforce each other. The tendency of politicians to use taxes and subsidies to threaten enemies and reward supporters actually distorts the profit signals in industry, meaning that the information sent out to investors is misleading and biased toward cronyism rather than producing what consumers actually want. Unsurprisingly, the two iconic examples of “good” industrial policy mentioned above have not held up very well: Japan’s spending on industry created a zombie economy that has never recovered. More recently, China’s huge government spending on industry has at best had no positive effect if the analysis accounts for the opportunity cost of the investments.
Of course, there really are market failures, and so it seems tempting to conclude that while politically-chosen industrial policies aren’t perfect, we have to try to improve on the problems that afflict markets, especially in settings where insufficient investment in research and short time horizons would make market outcomes seem weakest and least desirable.
The problem is that—as I argued recently at greater length— “good” industrial policy is literally impossible, either to pass in the first place or to sustain if by lucky accident it were implemented. The reasons are actually spelled out clearly in the literature advocating for the insulation of bureaucrats from political incentives, beginning with A.C. Pigou in 1920. First, elected officials are far from passive, and have shown themselves perfectly capable of rewriting rules and “reforming” bureaucracy for their own benefit, even if this is harmful to the “public good.” Second, if a legislature would prefer a policy other than the socially best policy for electoral reasons, then that same legislature’s members can anticipate the consequences of an institutional arrangement that would result in that less-preferred policy, and would veto the institutional change.
Pigou should be given credit for recognizing that there is a conflict between democracy and industrial policy. In 1920, he wrote in Wealth and Welfare:
The case . . . cannot become more than a prima facie one, until we have considered the qualifications, which governmental agencies may be expected to possess for intervening advantageously in this class of matter…
It is not sufficient to contrast the imperfect adjustments of unfettered private enterprise with the best adjustments that economists in their studies can imagine. For we cannot expect that any State authority will attain, or even whole-heartedly seek, that ideal. Such authorities are liable alike to ignorance, to sectional pressure, and to personal corruption by private interest. A loud-voiced part of their constituents, if organized for votes, may easily outweigh the whole. (pp. 247-248; emphasis added)
Later, in “State Action and Laissez-Faire”, Pigou again sounded a note of caution:
In order to decide whether or not State action is practically desirable, it is not enough to know that a form and degree of it can be conceived, which, if carried through effectively, would benefit the community. We have further to inquire how far, in the particular country in which we are interested and the particular time that concerns us, the government is qualified to select the right form and degree of State action and to carry it through effectively. (1935, p. 124)
Proponents of industrial policy are fully aware that politics, at least democratic politics with elections in which organized groups can be pivotal, is a problem. In fact, for many supporters, politics may be the most important problem. As Robert Reich put it, “It is the claim of industrial policy… that the only alternative to formulating an explicit program for improving the nation’s competitive performance is to cede effective responsibility for policy to groups with back-door political influence.” (Emphasis added)
So advocates of industrial policy do not see their program accommodating democratic politics; their claim is that the “correct” policy should be implemented instead of the outcomes that would be chosen by politics. If experts are allowed a free hand to choose the right taxes, subsidies, and regulations, the results will be a good industrial plan, better than what would be produced by market processes alone. But the surprising thing, at least to outsiders, is learning that the “right” industrial policy is fully understood by proponents to require the suspension of democratic accountability.
The seductive mirage of industrial policy thus connects two dangerous illusions, in a way that seems irresistible yet doomed. First, the idea that we (meaning elite experts, which is not “we” at all) could use central planning and direction to do better than the set of investments implied by the profit/loss test is hard to resist. Second, given that we could do better, any suspension of democracy and accountability to the public seems justified, because experts will only use this power responsibly, and for our own good. But the information problem means that experts can rarely do better, and the incentive problem shows that even if they could the contest to control just who would get to appoint the unaccountable benevolent despots would consume all the resources of the society in conflict.
Read more about our disclosure policy here.
