One of the most common refrains from politicians of both parties who want to justify their new programs or higher spending is that it will create jobs. Then there are those who argue that the goal of industrial policy is the creation of “good jobs.” I wish they would stop so we could have a conversation about what industrial policy is about and whether it is the best way to achieve whatever goal it is supposed to achieve.

With that in mind, I thought these comments from Noah Smith and Larry Summers were interesting.

During a Good Fellows podcast a few months ago, Larry Summers made a comment  that is worth keeping in mind (The whole discussion is very good and worth listening to).

Look, I think this is something we need to think very hard about. I believe that the best generals are the ones who hate wars most, and I believe the best industrial policy experts are the ones who hate it most. The problem is that the only people who talk about industrial policy are the ones who love it and are looking for a reason to do it. Anyone who can […] talk for five minutes about why we need industrial strategy or resilience without saying the words “it will create US jobs” I have substantial time for but whenever it comes with a whole set of talk about US jobs then it becomes much more problematic.

What Summers is getting at is that industrial policy should be about using the government to achieve what arguably can’t be achieved through market forces. National security is one of those things. But national security will be best promoted if the goal is, well, national security. It will not be effectively promoted if it is used as a Trojan horse for achieving goals – such as job creation – other than national security. As soon as you start talking about job creation, you get distracted from what you claim you are trying to achieve.

Noah Smith also suggests that the first goal of industrial policy isn’t and shouldn’t be job creation. He writes that industrial policy’s goal is climate strategy in the case of the Inflation Reduction Act,  and national security (among other things) in the case of the Chips Act. He also believes that that industrial policy as implemented in these two bills could and will lead to a manufacturing boom. However, he explains that even if the manufacturing boom happens, we should not expect a job boom too. Here’s Smith:

So far, the critics of industrial policy tend to spend a lot of their time rebutting the notion that we can bring back the golden age of widespread “good jobs” in factories….

And they’re right. Industrial policy will not turn us back into a nation of factory workers. Every country in the world, including China, sees its share of manufacturing employment fall as it becomes rich — both because consumers start demanding more services as they get richer, and because rich countries can only stay competitive in manufacturing by using lots of automation.

If we do manage to engineer a manufacturing boom, most of the actual production work will be done by robots, because we are a rich country with very high labor costs and lots of abundant capital and technology. Automated manufacturing is what we specialize in, not labor-intensive manufacturing. The latter is for countries like India and Tanzania. America needs to build with robots

This is an important point overlooked by industrial policy fangirls on the right and the left. I would add another point. Industrial policy is meant to reallocate resources in different directions than what profit-motivated actors and market forces would achieve (translation in practice: produce less output using more resources). That doesn’t add up with the belief that “incentivizing” those who invest their money to do something they wouldn’t do without specially created government privileges (such as protective tariffs, tax credits, loans and subsidies) somehow will end up creating growth and jobs.

Smith does correctly note that industrial policy could create some particular jobs in the subsidized industries. As he explains, labor is needed to build factories, which when operating require workers. That said, we should also account for the fact that subsidies given to some industries necessarily shift resources – capital and labor – away from other, nonsubsidized industries. For instance, Dean Baker once explained the unfairness of the ExIm Bank noting—“by diverting capital to the winners picked by the Ex‐​Im Bank, we are raising the price of capital for other firms.” If I believed that the government has a way to identify profitable ventures better than market actors, I could buy that on net we would be better off. But I don’t harbor such a belief.

I have to say, though, that I am more skeptical than Smith that industrial policy will create a sustainable manufacturing boom. In part it is because the government forced reallocation of resources once again means producing less output using more resources. Either the government incentives are incentivizing companies to do something they believed wasn’t in their interest before or the government payouts are going to companies to do things they would have done anyway. Neither is the best use of resources. I have found those two things to be true, no matter what government favoritism programs I have studied in the past.

Talking about companies now getting subsidies for something they would have done absent the subsidies, I found this interesting. In a recent paper by Réka Juhász, Nathan Lane, Emily Oehlsen, and Verónica Pérez found that industrial policy “is highly correlated with an industry’s revealed comparative advantage.” Translation: The market discovers ‘winners,’ and then politicians encourage them to expand inefficiently and – more and more – also saddles these firms with politically fashionable strings such as childcare and Buy American requirements.

Free-market advocates recognize that exceptions exist to a policy of free trade or the unhindered functioning of the market. National security is the main one that comes to mind. But it must be a real national-security reason, not a fabricated one like the Trump steel tariffs. We also recognize that if real market failures exist, and persist over time, they could justify government interventions. In this case too we must be clear about what market failure means. It certainly doesn’t mean more than merely that the outcome produced by the market happens to be permanently different than what a few people believe should be the appropriate outcome.

The most important thing, however, is that government interventions that disrupt market allocations almost always means less efficiency (recessions are a different situation). These costs (which include future taxes and distorted capital markets) might be justified if industrial policy is to address a real national-security threat or some other genuine problem that we deem incredibly important. But if so, stop claiming that the goal is job creation.


Veronique de Rugy is a Senior research fellow at the Mercatus Center and syndicated columnist at Creators.