When a government is considering an industrial policy that subsidizes a particular industry (e.g., the United States provides subsidies) Domestic semiconductor manufacturing), what is the best way? I’m generally not a fan of such subsidies, but it is possible to make an economic case for certain industries that provide social and individual benefits (e.g., reducing carbon emissions or strengthening domestic supply chains for agricultural products critical to national security). Some form of government subsidy may be required for production. But in what form?
I stumbled upon some examples of problems that can arise in “The Political Economy of Industrial Policy” by Réka Juhász and Nathan Lane (Journal of Economic PerspectivesFall 2024, 38:4, 27–54). They describe a situation in which appropriate forms of subsidies are not available for companies that fail rather than those that succeed in producing the desired product. The mechanism would be a “repayable advance” from the government, with successful companies having to repay the advance, but those failing to compete in the market not having to repay the advance. They write:
Consider a green industrial policy in which public agencies subsidize risky projects that, if successful, will generate both private and societal benefits. How should institutions design conditional grants? Meunier and Ponssard (2024) show that rewards for success are optimal when companies and public institutions have symmetric information about the probability of project success. However, under asymmetric information, where only the company knows the probability of success, it must be compensated for failure(!). This is because it mitigates the windfall profits that arise when agencies subsidize projects that could be funded without the subsidy. These insights speak directly to the experience of the French Agency for Ecological Transition (ADEME), a public agency that monitors innovative activities for the energy transition funded by the Investment for the Future program. Initially, ADEME used flat subsidies, but evidence of windfall revenues quickly emerged for some projects. So the agency introduced ‘repayable advances’ – subsidies that had to be repaid in case of success – or subsidies for failure.
Article cited by Juhász and Lane Guy Meunier and Jean-Pierre Ponssard presented a paper entitled “Green industrial policy, information asymmetry and reimbursable development”. (Journal of Public Economic TheoryFebruary 2024, e12668). This article will require a lot of math for beginning readers. but they
One risk of government subsidies to successful companies is that they reward companies that would already be successful in producing the desired product. In fact, it can even reward companies for projects that are already funded and underway. In effect, they reward companies that would already be profitable with higher profits. In contrast, repayable advances aim to encourage more companies to produce the desired product. That’s because repayable advances (and forgivable advances) mean your losses due to failure are reduced.
On the other hand, the risk with redeemable advances is that they are paid out to companies that are not successful. It is conceivable that a company may not be able to exert optimal efforts for success, even though it knows it has some protection from loss. Additionally, politicians may find it easier to explain subsidy successes that also provide photo opportunities than subsidy failures.
In a particular setting of the Meunier-Ponsard model, the key difference between subsidizing winning and losing firms is determined by what information is available. For example, consider a situation where there are several companies ready to invest to produce a desired product, but neither the government nor the companies themselves have a clear idea of who will be most successful. In this setting, it may make sense to promise additional rewards to the winners. But now consider an alternative situation where the government doesn’t know which companies will succeed, but the companies themselves have a pretty good idea. In such circumstances, businesses that are likely to succeed are more likely to keep going without subsidies, and the government is only rewarding what would have happened anyway.
At a broader level, the key point is that providing a broad overall justification for industry subsidies is only the first step. The actual design of a particular policy can be critical to the incentives created. Meunier and Ponsard wrote:
More generally, our analysis is based on Rodrik (2014) is an organization responsible for monitoring green policies. Let’s take a quick look at three of them. Discipline: Clarify the agency’s preliminary goals and establish evaluation protocols for what should or can be observed, making this a critical aspect of the contracting process. Embeddedness: Introduce reasonably simple conditional incentives for the duration of the project, but be aware of the risks of manipulation. Games: This plays an important role in our analysis. We formalize this problem and show how repayable advances can be used to mitigate its impact in some circumstances.
Of course, these guidelines for ensuring that industry subsidies are not wasteful or ineffective apply to all types of industry subsidies, not just green energy. For those who want to know more, Dani Rodrik’s (2014) thesis is “Green Industrial Policy”. (Oxford Review of Economic Policy30(3), 469–491).
