Macro America Great Again

Do you smell what Tariff Man is cooking?

get in loser we’re (not) going shopping because we have a trade deficit with the grocery store

Donald Trump, as you might have heard, has won the US presidential election. I don’t like him and think he’s a bad dude with bad ideas, one of which was tariffs: hike up the cost of imports to force companies to produce more in the United States.

Trump’s plan is, loosely, as follows. Stick major trading partners with high exceptionless tariffs: Mexico and Canada at 25%, China at 60%, everyone else at 10%. This is, of course, subject to a lot of course shifts from Trump (he backed out of the Mexico tariff within days), but let’s assume something like that is the base scenario. Is this, like, a good policy? No.

Disclaimer: I won’t talk about the trade deficit. Go read Paul Krugman on that. By itself, it’s irrelevant in the only country that does 100% of its global trade in its own currency¹.

What even is a tariff?

There seems to be a lot of very basic confusion about how tariffs and import taxes work, in general. So let’s start there.

A tariff is a tax on imports. The way it works is, when a company wants to buy products abroad they rent a big boat that picks them up, then they bring them back to the country and have to make some sort of written statement (largely to control, say, food safety, and smuggling) which includes the value. The tariffs are paid at that moment. They’re not paid by the foreign country at the moment of purchasing - they’re paid by the national company at the moment of entry. If China was forced to pay a tax to America, unless it was under military occupation, it would simply refuse. So by definition a tariff is a tax on what the country itself buys from other countries.

The econ 101 impact of tariffs is that they would increase the price of imports and reduce imported quantities, with more detailed models also detailing that tariffs would shift production in really inefficient ways - for example, car companies have really costly practices to import cars and their parts in order to get out of paying tariffs. This means that they are costly and inefficient.

The logic behind this is quite simple: imagine that one country, Spain, can more efficiently produce wine, while another, England, can more efficiently produce wool - but both can produce both. With limited reasons and wanting to maximize output and employment, it stands to reason that England would specialize in wool and Spain in wine, and they would trade. But even if Spain could more efficiently produce wine and wool, it still stands to reason that trade would be beneficial: imagine if Spain is 300% more efficient at wine, but only 1% better at wool. Then, would it be more effective for them to invest in wineries, or in sheep farms? It would still be best to focus on wine, because the marginal gains from extra wine production are still bigger. This is known as comparative advantage and it’s more or less at the core of most a massive chunk of economics - for example, if Michael Jordan was both the world’s best basketball player, earning millions, and the world’s best gardener at minimum wage, nobody would expect him to spend any time gardening.

Comparative advantage dictates that countries should specialize, at least vis a vis trade, on the things they can do best. If output is determined by land, capital, labor, and human capital (i.e. education), as well as technology, then countries can focus on what they do best with the “inputs” (called factor endowments) that they have. So if agriculture takes land and labor, then countries with lots of both have the advantage over countries with few, and countries with lots of capital and little labor would have the advantage on, say, advanced manufacturing. If you consider these factors, as well as the impact of size on trade flows (positive: bigger countries are bigger markets), and the impact of distance (negative: further away is more costly), then you can get a general picture of what countries would trade, why, and how they would benefit. Obviously countries can change how much they have of each factor - they can promote immigration to get more labor, they can change policies to add more capital, or they can invest in education to improve human capital, as well as promote the use and creation of technology.

Most writing on the subject would more or less stop here: Trump tariffs bad, growth lower according to basically all normal forecasts, inflation higher. Something about washer dryers too. But I think the macro arguments are extremely unsophisticated.

For tariff proponents, tariffs cut against comparative advantage, meaning that they would raise prices and lower output - i.e., make the United States poorer. This is a static argument - they focus on what would happen right now. Most of the US’s imports are stuff like fruits and vegetables, crude oil, and “intermediate goods” - parts and pieces for industrial products like cars, and most of them come from close US allies, so at first glance it’s kind of self-evident that they would harm US manufacturing imports, increase prices, and after retaliation, would harm US exports too. Which is why most people writing about the topic say the tariffs are bad - the Peterson Institute for International Economics finds that tariffs would depress incomes for American families, greatly damage the US economy and thrust the global economy into complete chaos. Meanwhile the Tax Foundation also finds evidence for a host of negative effects, the Allianz insurance company finds that they would be inflationary and reduce growth (but they examine a “less bad” scenario), and the Center for American Progress finds (no shock here), that growth would be lower and inflation would be higher.

Tariff proponents, meanwhile, have a somewhat more contrived argument: first, they argue that the impact on inflation would be limited in a vacuum, because inflation is macro and not microeconomic. Secondly, and more controversially, they argue that tariffs would, over the medium run, increase output, by promoting a stronger manufacturing sector. This would happen by increasing investment in fixed capital, and would also utilize a lot of underemployed labor. Who is right is a controversial question (it is, in fact, the purpose of this blog post). So what do I expect to happen?

♫You would never break the chain ♫

The first, most obvious question is whether tariffs do actually have a positive impact on output - that is, whether they would increase or decrease “production”. Most writing on the topic just talks about Trump’s tariffs on washing machines, which did not have any effect on employment on the sector or on output. A separate study of the same set of tariffs, but more broadly, also found a negative effect, mainly because the positive impact on a few sectors was offset by higher costs for all others, as well as retaliation from other countries.

Of course, once again, this applied to only some industries, which meant that there was a lot of wiggle room - so a lot of inefficient strategies to avoid the tariffs and saddle some other industrial sector with the costs. But what would happen if all goods-producing sectors were hit by tariffs at the same time? Well, they would need to get suppliers for the inputs, one way or the other, all at the same time. “Fortunately”, we do have a recent example of this: the COVID-era breakdown of international supply chains.

This is a very thorny subject (particularly because I’ll come back to it when talking about inflation), but overall its easy to consider the subject by imagining the supply chain as a bicycle chain. In the chain, if any link breaks, then the bike won’t go around. Trump’s tariffs, which are proof he would break the chain, would tear out a bunch of links at the same time. Given how complex supply chains have gotten, this is not an especially good idea - it would hit a large number of links across multiple industries all at once. For example, General Motors had 856 “tier 1” suppliers, which had a whopping 18,000 suppliers of their own (“tier 2") - if the rate of expansion continued, “tier 3” would be around 380,000 firms, and tier 4 would be well into the millions. Trump’s trade war would be what’s known as a “systemic supply shock” on value chains, putting global production completely out of whack for an indeterminate time - pandemic-era issues took around 2 years to resolve. Adjustments to supply chains, particularly for intermediate goods, would hterefore play a massive role in the costs of lowered trade, on which little can be said ahead of time, especially without knowing the impact on wages. In this sense, the O-Ring² theory can be somehwat relevant: wages for high-value activities are usually determined by smooth completion of interrelated tasks - therefore, large disruptions like the ones produced by tariffs could also severely disrupt the quality of the domestic advanced services economy. This is especially because quality control and transfer of intangibles are a major reason why firms invest in supply chains - and those assets would have to be rebuilt very quickly. The impacts are, overall, very large, very uncertain, and very hard to predict.

Of course, the “O-Ring” and the like are part of the theory of economic development, which seeks to ask “what if we could change our factor endowments to get a better deal from global trade”. This is, obviously, eminently reasonable, at least as a whole. It should be noted that there is at least some strength behind protecting finished products producers (think: t-shirts), versus intermediate goods (which would be tariffed). One example of recent workable tariffs are electric vehicles in the Inflation Reduction Act: the “Buy American” requirements increased production, but at the cost of decreasing carbon emissions reductions.

The example used by Trump himself are the McKinley tariffs (in fact, President McKinley is known for two things: tariffs, and being assasinated), except recent research puts into question whether they helped at all. Basically, a long running thread in trade economics is that participating in trade means that less productive firms go out of business, which raises aggregate economic output by relocating capital and labor. Tariffs mean that these less productive firms, which are frequently politically connected, remain in business despite being unproductive. American trade policy overall was determined by restrictions on manufacturing competion from the 1860s until the 1930s, and had negative but variable impacts on economic performance. Generally, trade policy responded mainly to political interests, particularly of the scarce factors, and not to strategic economic planning.

Finally, let’s move outside the US: there is something known as the “Infant Industry argument”, positing that industries that are just getting started should be protected in order to help them grow. Leaving aside that textiles and manufacturing are less infants and more grown ass adults who refuse to move out of the basement, the fact is that infant industry arguments have usually been justifications for politically expedient policy, not the other way around, at least until the 20th century. This is because, in recent years, knowledge and technology has taken on a bigger role: it’s not important just to have factories, but also to have staff that knows how to run them. In this sense, agglomeration is becoming increasingly powerful, which means that countries have stronger incentives to protect knowledge-intensive sectors and subsidize the production of knowledge. This is particularly relevant because the main channel linking trade and technology appears to be investment, such that participation in global markets can boost growth - meaning that trade has a heterogenous but broadly positive impact on economic performance through time.

One final item is that, if at this point it’s not obvious that Trump’s plan to do import substitution for t-shirts won’t work in a wealthy and advanced economy, it should become obvious after finding out that it also didn’t work in developing economies. This gets into broader issues in development economics, but basically, back in the 1990s the hot question everyone was asking was why Latin America didn’t develop with protectionist pro-manufacturing policies, but East Asia did. Basically, even though the relative economic performance of both regions reversed between the 19th and 20th centuries, it comes down to, in some part, policy, but also improved external conditions and higher productivity. A large issue is that high tariffs were optimal in the 19th century because other countries had high tariffs, and not retaliating would have put one at a disadvantage even if lower tariffs had benefitted all parties.

So overall, it seems that industrialization mostly would follow factor endowments and institutions: East Asia industrialized with labor-intensive manufacturing, which works for a country with lots of labor and little land, which didn’t work for LatAm because it has lots of land and little labor. This also contributed to income inequality, and particularly to institutional issues - protected sectors became important political constituencies. The main reason import substitution was abandoned by economists was that it just didn’t work, and introduced a complicated element of rent-seeking and technology-blocking that was not easy to undo.

The Price of Peas

Trump’s prospective Treasury Secretary, Scott Bessent, has explained that the tariffs are not inflationary per se, using a seemingly laughable explanation:

… tariffs can’t be inflationary because if the price of one thing goes up, unless you give people more money, then they have less money to spend on the other thing, so there is no inflation. The inflation comes through either increasing the money supply or increasing the government spending, and that’s what happened under Biden

As I’ve said on Bluesky (and also X the Everything App but nobody cares about that one anymore), I think Bessent is essentially right. The first and most obvious thing to mention is that a one-time tariff increase would result in a one-time price increase of some fraction of the tariff. So if tariffs of 25% get imposed on avocado imports, avocado prices will increase around 25% - unless companies either reduce the supply of avocado imports (which would raise prices more), or swallow the cost of the tariff (which would raise prices less). But either way, prices would go up one time. You can see it graphically here: prices go up with the tariff, and then continue on their previous trend after the tariff gets passed through to consumers.

Let’s go to the Quantity Theory of Money: M*V = P*y, the supply of money (M) times the speed at which it changes hands (V) equals the price level (P) times real output (y). To simplify an outrageously convoluted topic (read more here!), if real output increases, then more money does not necessarily mean higher prices, but if it doesn’t, then it does. Unless velocity does something weird, which it shouldn’t unless Trump fucks around with the Fed or something along those lines. The insight here is pretty simple: inflation happens when there’s too much money chasing too few goods, so people try to spend more money than companies can match with physical stuff - that’s what “fixed real output” means.

This, in sum, also means that real aggregate demand (i.e. the total amount of stuff that people, the government, and companies buy), can never exceed real aggregate supply - you can’t buy more quantities of stuff that is for sale. But nominal demand, i.e. in dollar terms, can - because there can be too much money. Meaning, prices going up has two possible sources: too much money (nominal aggregate demand being too high for nominal supply), or too little stuff (the opposite). If we assume some sort of short-term cap on total output, then we can have a fairly clear view of what any given shift in money, velocity, or output would do to prices.

Which one would tariffs hit? Well, import taxes would both raise the price of imports and reduce their amount, meaning they’re a supply shock, because they would shift prices and output in opposite directions³. The former effect is, I think, rather self-evident: companies tend to pass through at least some portion of their cost to consumers. The Great Greedflation Debate of 2021-22 hinged almost entirely on the relationship between profit margins and cost shocks. Looking at 19th century sugar tariffs, around 40% of the increase got passed on to consumers, and Trump’s 2018 tariff hikes on appliances had an elasticity greater than 1 on prices, meaning a 1% tariff increase raised prices of washers and dryers by more than 1%. COVID-Era disruptions caused inflation with long lags in 2021 as well.

However, the effect on output isn’t totally clear beyond the previous section, especially as long as the effect gets passed through in prices - companies could just be less profitable and operate at lower capacity, period. This is where we get on annoying pandemic-era sophistry territory: Team Transitory versus Team Not Transitory. I’ve had a bunch to say about this, but fundamentally the claim from Team Transitory was that inflation was caused by temporarily lowered supply instead of demand, and the opposite claim was that demand exceeded supply regardless of capacity. I thought both had some truth to them, but we can move past Twitter gripes from 3 years ago and onto something actually productive: the “transitory” reduction of capacity can be, in fact, quite lengthy, meaning that inflation might be elevated even with the economy running at normal capacity for some time.

However, on the consumer side, this would manifest differently: the economy contracting and prices increasing would mean different consumer patterns. If we talk about avocados, higher prices and lower supply would mean that consumers would either spend the same dollar amount on fewer avocados, or spend more dollars on their previous, preferred basket, and less on other things. So, under the tenets of Nominal GDP Targeting, Scott Bessent is correct - the tariffs would not be inflationary, because they would simply cause a prolonged recession that raised the prices of a large array of normal goods, but only in the short term (which can, in fact, be quite prolonged depending on the definition).

At any rate

But what should the Fed do? Well, that’s the million dollar quesiton, and one which the Bank of England is also, for whatever reason, grappling with right now. The story is as follows: back in 2021, the BoE decided not to raise rates in response to high inflation, because they perceived it to be deeply pandemic-related, and that rates would act too slowly to impact transitory price hikes without raising unemployment unacceptably. But instead of clearly communicating this, they said… something else, and so had to hire Ben Bernanke to tell them what to do, which would also not solve anything. So the BoE has decided to admit to a mistake it did not make, and ordered a review of models that were not wrong when reviewing its models, and then the review came out and it started an unsolvable argument because either the Bank had made a mistake by ignoring correct models, or it had not made one by following incorrect ones. Pure Borges/Kafka stuff, but either way you see it, the shock is to supply, not to demand. And the thing about supply shocks is that to quote Allan Meltzer :“Money cannot replace oil” - monetary policy can’t fix trade policy.

But “putting the Fed in an impossible situation for no apparent reason” is not a good idea when they have an, in fact quite recent, record of trying to win the last battle. So either the Fed decides to look tough on inflation and tightens when it knows it shouldn’t, or looks weak on prices and risks its reputation but does the objectively correct thing. If you add Trump’s threats to Fed independence… not a pretty situation to be with! Overall, I’d just recommend caution and to suck up a really bad macroeconomic position of more inflation and more unemployment.

Conclusion

So, overall, it seems that tariffs are a bad idea and that, like in past occasions, the opposition is less to any specific trade arrangement and more to trade itself. Even after a long hard look at history, the record is not very strong for tariffs promoting growth or employment at any meaningful scale, and does point to it creating big headaches for policymakers that have no easy answer. People, as a general rule, won’t be happy to spend more on the exact same thing.

Obviously one out for Trump would be to impose a Destination Based Cash Flow Tax, call it a tariff, and rejoice as the United States economy benefits from a good policy idea. But that is… unlikely.

1. It could, however, be viable to use tariffs to force other countries to revalue their currencies.

2. Despite sounding like a sex toy of some kind, it actually references the Challenger disaster

3. Because GDP and unemployment are negatively correlated, this means that a world without any demand disruptions would also, funnily enough, have a positive Phillips Curve.