The National Defense Argument

Can we use price theory to construct a national defense argument for subsidies and trade restrictions?

You are reading Economic Forces, a free weekly newsletter on economics, especially price theory, without the politics. Economic Forces arrives weekly in the inboxes of nearly 17,000 subscribers. You can support our newsletter by sharing this free post or becoming a paid subscriber:

---

When it comes to tariffs and industry subsidies, perhaps the most common justification people use for these policies is what is known as the “national defense argument.” A lot of economists seem to find this dubious. Some are inclined to conclude that this is just a narrative of special interest groups — an appeal to patriotism to generate special privileges. This is compounded by the fact that many economists are skeptical of subsidies and tariffs to begin with.

However, with tariffs on steel and aluminum back in the news, and with national defense arguments once again being used as justification, I would like to summarize a theory of aid for industries based on the national defense argument and see how it might (or might not) apply to current circumstances. Let’s see if we can use some price theory to understand the national defense argument.

Subsidies, Tariffs, and National Defense

One issue with national defense arguments is that they often fail to specify what should be subsidized or what should be subject to a tariff. At times, the national defense argument is used to vaguely describe why a particular industry deserves a subsidy or protection from foreign competition. “We need X because it is important for national defense or our national security and therefore it should be subsidized (or protected).” Supposing that this is true, what should be subsidized? The output good? The input good? Should it be a subsidy? Should it be protected by a tariff? Assuming that we take such arguments seriously — and I think there is reason to believe that we should — these are important, practical questions that require answers. Otherwise, how will we know if we should be accepting or dismissive of such arguments? And assuming we find such arguments compelling, how will we know how to implement policy to achieve the goal?

Earl Thompson provided some of those answers by presenting a theory of why countries might provide subsidies or protect industries with tariffs for reasons related to national defense.

Thompson started his analysis by noting that the conditions and operation of particular markets during wartime differ from those during peacetime. In particular, certain resources can be diverted by states to alternative use cases and other goods might be subject to price controls.

Sticking with price controls, suppose that there is a consistent pattern of price controls imposed during wartime. If that is the case, we would expect that the owners of capital used to produce the products that are subject to those price controls would rationally expect to suffer the costs. Anticipation of those costs would imply that the capital owners would not expect to capture the full social value of their capital and therefore would undervalue their capital during peacetime and underinvest in capital.

Thompson argued that under these circumstances, the optimal policy would be offer peacetime subsidies in order to offset the wartime costs — to fix the intertemporal distortion. However, his theory is also specific about what should be subsidized. He argued that although the price controls were often imposed on final-stage goods, the optimal subsidy was not to subsidize the output goods in these industries, but rather to subsidize the capital investment necessary to produce them. A subsidy for the output good would only indirectly lead to greater capital investment. By contrast, a direct subsidy to capital investment could directly offset the incentive to underinvest as a result of wartime distortions.

But what about international trade? Thompson argued that some goods that are subject to wartime price controls are produced (at least to some extent) using imported capital goods. This need not imply that all capital goods for such production are imported, but rather that some fraction is imported. In these situations, he argued that if the transportation costs of such imports tend to be sufficiently high during wartime or other international conflicts, this would require the use of domestically-produced capital goods during wartime. As a result, the optimal policy would be to put a tariff on these capital goods that were imported during peacetime but not during wartime.

What is the Evidence?

Thompson’s theory is really just an application of price theory. Given that there are intertemporal distortions in prices associated with war, national emergencies, and other international conflicts, the optimal policy response is to correct the intertemporal distortion. Wartime costs are offset by peacetime subsidies. That seems like a reasonable theory. It is certainly more coherent than the casual argument and provides much more specific guidance for subsidies and thus some testable hypotheses to determine whether existing subsidies might be explained by this national defense argument.

The value of Thompson’s theory is that it also provides much more specific predictions than just “wartime price controls lead to peacetime subsidies.” In particular, his theory suggests the degree to which any subsidy is given depends on the cost of the price control. Thus, according to his theory, we should distinguish between price controls that are difficult to evade and price controls that are easier to evade. Price controls can be evaded by adjusting on other margins, such as reductions in quality or by using tie-in sales. Other price controls, like some of those used in the U.S. during World War II combined those controls with rationing coupons. In those instances, producers have no incentive to reduce quality or use tie-in sales because of the fixed supply of coupons. As a result, we would expect price controls to be more costly for these goods and thus have larger subsidies.

In addition, as I previously mentioned, one would expect subsidies for capital inputs, not for output goods. But this also implies that capital goods that have multiple use cases — assuming only some output produced using that capital is subject to price controls — should have smaller subsidies than capital goods used for particular types of production.

Although I’m referring to aid as “subsidies,” industry aid could come in a variety of forms. It could come in the form of an investment subsidy, a tariff, or — in the case of some raw materials — a peacetime stockpile. The type of aid that is used would depend on whether the goods could be imported during wartime and how costly it would be to stockpile raw materials.

In interpreting the evidence, Thompson focused on World War II price controls and subsequent industry subsidies and trade protections (for context, he was writing in the late 1970s). However, given the wide scope of those controls, he sought to focus his empirical evidence on price controls that were difficult to avoid.

Goods that were subject to price controls and rationing coupons were most difficult to avoid. These were things like gasoline, dairy products, sugar, and meat. We can therefore predict a pattern of subsidies related to these industries. However, recall that Thompson’s argument implies that the subsidies should be going to capital owners, not the final output producer or retailers. Given the characteristics of the goods listed, one would expect subsidies to be aimed at the production of raw materials. Thus, for example, the peacetime subsidy should be for cattle, not slaughterhouses or dairies; for oil producers, not gas stations.

This is precisely what Thompson found in the peacetime subsidies. These were given to sugar growers, oil producers, and those raising cattle. Specifically, crude oil producers and cattle ranchers were given investment subsidies. Thompson even does some some back-of-the-envelope calculations using black market values of the goods subject to price controls during wartime to get a sense of the degree of undervaluation these estimates imply. His calculations suggest that the order of magnitude of the subsidies is consistent with the theory.

With regards to trade, recall that his theory implies that tariffs should be imposed on things that are imported during peacetime, but too costly to import during wartime. He found that “the set of postwar imports which suffered most from our World War II price controls (crude oil, cattle, sugar, cotton, wool, iron and steel, various agriculture commodities, and products of these materials) is virtually identical to Bergsten’s list of commodities which have been subject to significant U.S. import quotas!”

There is a lot more detail in the paper, but the general conclusion is that the pattern of price controls, subsidies, and trade restrictions fit with the predictions of the model.

What About Now? Is This Relevant?

This week, the Trump administration announced 25% tariffs on steel and 10% tariffs on aluminum. Can we possibly make sense of this using Thompson’s theory?

Well, Thompson’s theory suggests a couple of relevant details. The first is that capital goods that have multiple, competing use cases should be given smaller values of aid (either in subsidies or tariffs) than capital used to produce goods that are subject to difficult-to-avoid price controls. He specifically mentions steel as falling into the former category. In addition, one form of subsidy that the government could offer would be to stockpile raw materials during peacetime so that they can draw down on the stockpile during wartime. However, he also notes that steel would be difficult and costly to stockpile in significant quantities. Finally, his theory implies that tariffs are the appropriate subsidy when these capital goods are imported during peacetime, but not during wartime.

Steel and aluminum seem like candidates for aid using Thompson’s national defense argument. During a time of war, these are goods that are likely to be effected by direct military procurement. Thompson’s theory would imply that they should therefore receive a small level of support. In addition, the U.S. imports approximately 25 to 30 percent of its steel and approximately 30 percent of its aluminum during peacetime. Yet, it isn’t hard to imagine that during a large-scale war, steel and aluminum would be much more difficult and costly to import. Thus, a tariff on steel and aluminum without any further subsidy would seem to be consistent with Thompson’s theory.

Thus, it is at least plausible that these tariffs are driven by this national defense argument.

Of course, there are other issues to consider. One issue is whether one believes that there is validity to Thompson’s theory. Some might be inclined to dismiss it on the grounds that it seems to attribute too much knowledge and foresight to policymakers. After all, can people who write bills to acquire and rename Greenland as “Red, White, and Blueland” really understand the intertemporal distortions created by wartime price controls? Can they also understand that such distortions could be offset by peacetime subsidies or tariffs?

Such knowledge and foresight does seem unlikely. Nonetheless, it is unclear that one needs to credit policymakers with such knowledge or foresight in order for such policies to exist. Durable policies might simply be those that survived a process of experimentation, trial and error, and learning by doing. In addition, one would also need an alternative theory to explain why the pattern of price controls, procurement, subsidies, and trade restrictions occur in the pattern that Thompson predicted and observed.

Another issue is that Thompson’s theory is motivated by the historical prevalence of price controls and military procurement during wartime. Thus, to some extent, the validity of his theory is tied to the continued importance of such tools. Resource constraints are real and mobilizing resources for war is difficult and costly. Nonetheless, it is possible that policymakers could devise other tools. If so, the peacetime subsidies would no longer resolve intertemporal distortions, but rather potentially create distortions.

In addition, even if you accept Thompson’s argument, the magnitude of the subsidies should be tied to the expected future costs of the distortions. This implies that the magnitude of the subsidies depends critically on the probability that the U.S. will find itself at war in any given year. Thompson’s back-of-the-envelope calculations used the historical prevalence of warfare in the U.S. to estimate the present discounted value of the undervaluation of capital. If you think the likelihood of war in the present and the future is lower than it has been in the past, Thompson’s argument would imply that aid to industry should be reduced. The same could be true if one expects similar frequency, but smaller-scale warfare.

Ultimately, the purpose of this post is to provide food for thought. As I said in my previous post on tariffs, my objective isn’t to convince anyone one way or the other. My objective is to understand the world as it is. If that is our objective and if we are going to promote the value of price theory, I do think that we need to take price theoretic arguments like Thompson’s seriously.