Justifications For Tariffs (Part II)

This article continues the discussion of “why would anyone want to impose tariffs” that was started in this article. The first article discussed a non-standard justification: tariffs as a source of revenue.

The standard justification for tariffs is to better control the development of your domestic economy (and/or to reduce the influence of perfidious foreigners). Although I have limited sympathies for this concept — Canada pursued a policy of developing behind a tariff wall — it is a somewhat dated tool (as I discuss below).

Free Market Orthodoxy: Tariffs Bad

Being against tariffs is a now ancient stance for mainstream economics, going back to David Ricardo and his arguments about comparative advantage. (I explain the concept in the technical appendix.) To summarise: nations should concentrate their efforts producing goods that they are relatively efficient at producing, and then trade for other goods that other countries have relative production advantages.

This is in contradiction to alterative strategies.

1. Mercantilism. The mercantilist strategy is that countries should aim to run trade surpluses so that they can stockpile gold (or in the modern era, reserve assets). The justification for such a strategy is not that it would necessarily lead to a higher standard of living, rather having a gold stockpile might be useful if one finds oneself fighting in Eighteenth Century Imperialist Wars. One also need to keep in mind that mercantilism was taken seriously in an era of monarchs, where the “national treasury” was actually the monarch’s.

2. Protectionism. Protectionism is the modern version of being opposed to free trade, although it comes in different variants. The idea is that tariffs are meant to protect domestic industries from foreigners.

The fact that Ricardo’s arguments about free trade goes back to debate about The Corn Laws has resulted in discussions largely ending up being rote repetition of stale talking points. The debates about the Trump tariffs in 2025 have put me into a position of being somewhat unhappy with partisans of both sides of the debate. Although I can see justifications for wanting to protect industries, tariffs are an eighteenth century tool for the job.

Why Protect Industries?

Manufacturing businesses have economies of scale: the more units of goods you sell, fixed costs per unit sold decline. Furthermore, being larger makes it easier to get your way in negotiations. As such, it is an entirely sensible strategy to want to undercut competitors, drive them out of business, so that one ends up either as a comfy monopolist or oligopolist.

As a result, one might want to protect less-developed domestic industries from foreign competition so that they can master production processes and grow so that they have a chance to compete against larger foreign incumbent firms.

A related angle is job creation: there is considerable mythology about manufacturing jobs being superior to service industry jobs, and so there are arguments that those jobs need to be protected. This explains why American automotive unions supported President Trump’s tariffs.

Skeptical About the Job Story

I am not entirely convinced about the “free trade destroys manufacturing jobs” panic that has motivated both President Trump and many on the economic left. People get misty-eyed about the high paying unionised manufacturing jobs that were a feature of the main “strategic” industries of the post-World War II era, but those are not reflective of all manufacturing jobs. Garment workers in “sweat shops” or those stuck in mindless repetitive assembly operations in low value-added industries are also manufacturing jobs — but those jobs stink. Furthermore, manufacturing quite often literally stinks — courtesy of industrial pollution.

Although my engineering career was mainly in academia (and doing mathematical modelling in a computer lab), I still had contact with the engineering mindset. (I did spent one summer at the large electrical machine plant in Peterborough Ontario.) Although engineering jobs are “high quality” and part of manufacturing, most manufacturing does not need that many engineers — the firm is mainly applying existing technology. In general, the job of research engineers in manufacturing is to get rid of the line employees (that unions and politicians are so fond of). You want to automate repetitive tasks as much as possible — you have less defects, and you can get rid of unpleasant, dangerous jobs. Certain jobs are hard to replace — e.g., sewing shirts — but even those sometimes can be automated.

The manufacturing sector shows a trend increase in productivity. That is just telling us that engineers are collectively doing their jobs. We can get the same output with less workers, and so manufacturing jobs as a percentage of the workforce exhibits a long downward trend. Furthermore, developed countries have offloaded the problems associated with low value-added manufacturing by offshoring them to poorer countries.

Tariffs the Wrong Answer

No matter what one thinks about the merits of protecting domestic industries from foreign competitors, tariffs are no longer the standard way to approach the problem. Tariffs made sense in a world where governmental administrative capacity was limited — you just posted a list of import taxes (bringing us back to the origin of the word tariff) that customs inspectors could easily understand. In the modern era, governmental regulations are far more intrusive and complex. It is a lot easier to snarl foreign competitors with regulations than trying to set a tariff rate. (For example, Canadian provinces obliterated American liquor exports to Canada in 2025 by removing American products from province-run liquor distribution monopolies.) Furthermore, in a world of floating currencies, currency volatilities are considerable. A tariff of 10% could be easily cancelled out by a month’s movement in the currency exchange. The only way for a tariff to have a bite is for it to be quite chunky — e.g., 50%, which would then catch the eye of foreign countries.

The figure above shows the number of Regional Trade Agreements in force (data from the World Trade Organization). As we can see, there was a “hockey stick” in the data in the early 1990s, as regional trade agreements became the favoured mode of regulating international trade.

Instead of just having some oaf pick out tariff rates and post them on the internet, countries negotiate complicated trade treaties that manage some aspects of trade between a group of countries. Since most countries do wish to protect their national champions, the trade treaty ends up being a messy compromise in which some oxen get gored.

You Can’t Get There From Here

President Trump wants to roll the clock back to 1890 and make it possible to just reorient trade policy by changing tariff rates. However, countries are now enmeshed in existing trade agreements. (In fact, in 2025 President Trump violated the trade treaty with Canada and Mexico that he signed in his first term.) If you want to change the rules to protect your domestic industries, your trade partners expect you to follow the rules and renegotiate the treaties. They made compromises in those treaties that they would certainly want to revisit in response to you making changes beneficial to your country.

For example, the Republican White House, American auto unions, as well as the American media covered the auto tariff dispute with Canada (and Mexico) as being “Canadian” firms sending cars to the United States. In fact, most of those “Canadian” cars are produced by subsidiaries of American auto manufacturers that split production across the international borders. In Canada’s case, this is the result of a sequency of treaties going back to the 1965 Auto Pact (Mexican auto integration was later). American automakers were given access to the Canadian auto market in return for locating manufacturing operations in Canada. Attempting to stop “Canadian” auto imports would just mean snarling American automaker supply chains — and probably cutting those American automakers out of the Canadian market. Although such an operation might create net American jobs, it would be at the cost of reducing the economies of scale of the American automakers.

Finally, another problem with the implementation of President Trump’s tariff plan was that there was no coherent strategy behind them. If you want to protect a domestic industry, you apply a tariffs on its outputs, not its inputs. Given the instability of the Trump regime, it is not clear that firms will enter into long-term fixed investments to develop domestic industrial inputs if those input industries would not be viable in the absence of tariffs.

Concluding Remarks

Both the left and right are unhappy with modern trade treaties. However, one needs to accept that international treaties are between two or more countries, and each has its own interests. It is standard for new governments to reform domestic laws essentially unilaterally, that is not possible in international affairs. Although posting new tariffs is a unilateral act, it would most likely be a violation of some existing treaty — and the counter-parties on those treaties may exact some form of revenge.

Correspondingly, I am unconvinced that tariffs will be a major tool of policy after President Trump shuffles off into the history books. We are stuck in a world of gradual trade reform, not unilateral changes.

Technical Appendix: Comparative Advantage

David Ricardo used an example of Britain trading cloth (where it has an advantage in production) for Portuguese wine (where they have a production advantage). We will run through an example of how this results in a total production increase where there are some added constraints on behaviour.

Assume that we are only concerned with cloth and wine, and each country has 1 unit of time to allocate to producing those two commodities. We assume that consumers have a “consumption basket” that has an equal amount of cloth and wine — and production has to conform to that demand.

1. We assume that Country A is less efficient in production, it produces 1 unit of cloth or wine per unit of time. In the absence of trade, this implies that Country A must split its time allocation 50/50, and so produces 0.5 units of cloth and wine each.

2. Country B is more efficient at the production of either, but the advantage varies. It can produce 3 units of wine per unit of time, or 2 units of cloth. In the absence of trade, in order to balance production, it has to allocate 40% of time to wine, and 60% to cloth — which means that it produces 1.2 units of both (1.2 = 3×0.4 = 2×0.6).

The need for balancing of production implies that Country B has to spend more time producing cloth where it has less of a productivity advantage (2 to 1) versus wine (3 to 1 advantage). Total production of both goods is 1.7 (0.5 in Country A, 1.2 in Country B).

If we open up the possibility of trade between the countries, we now have an optimisation problem with two free variables (the wine versus cloth time allocation for each country) facing off against a few constraints (total wine and cloth production are equal, and time allocations cannot be less than 0%/greater than 100%).

Courtesy of mathematical intuition, we see that the optimal allocation involves having Country A (the country with less efficient productivity) setting 100% of its time allocation to producing cloth (producing 1 unit). Country B then sets its time allocation so that it produces 1 more unit of wine than cloth (so that total production between the two countries are equal). This is achieved by allocating 60% of its time to producing wine, so that it produces 1.8 units of wine (=0.6×3) and 0.8 units of cloth (=0.4×2).

With this allocation of production, total production across the two countries is 1.8 units of wine and cloth — versus 1.7 units before. The increased allocation of time by Country B to wine (where it has the highest relative productivity) increases total output across the two countries.

How the extra 0.1 units of production are allocated is a question of political economy. If we assume that neither country’s consumers can be worse off than before, Country A would have consumption between 0.5 and 0.6, and Country B would have consumption between 1.2 and 1.3.

Although the idea that international trade makes economies more efficient is compelling, we need to apply some caveats to the logic. Importantly, the above simplified model takes production possibilities as fixed — which is not the case courtesy of the need for capital goods. Nevertheless, common sense can be applied — there is no way that Canada deciding to replace coffee bean imports with coffee beans grown in greenhouses (assuming that coffee can be grown in greenhouses) would be a sensible economic strategy. The Canadian government might decide to replace imported coffee with ersatz “coffee” in order to avoid domestic interference by The International Coffee Cartel, but that substitution would be at the cost of citizen satisfaction (and they probably would revolt). The realpolitik justification of free (actually managed) trade is that you quite often do not want to be stuck with inferior and overpriced domestic products.