Why Tariffs On More Countries Can Be Better

Today, I am going to explain why tariffs on more countries can be better! Not to worry, I still think free trade is the best policy (modulo some special cases discussed in Modern Principles) but I am going to show that a uniform tariff can be better than a selective tariff. My example comes from a nice tweet from Malaysia expert Apurva Sanghi, modified for the US context.

Suppose the U.S. can import Hyundai Sonatas from Korea and Toyota Camrys from Japan, and consumers view the two cars as perfect substitutes. We compare three scenarios:

A) Free trade
B) 10% tariff on both countries (uniform tariff)
C) 10% tariff on Korea only (selective tariff)

The surprising result: B can be better than C, even though C is, in one sense, closer to free trade (the “best” policy) than B as it tariffs fewer countries. To focus on the key points I will assume 50 car buyers and no change in the number of buyers when tariffs change (so I will ignore the standard deadweight loss from reduced quantities).

Assumptions

Korea (Hyundai Sonata): $40k pre-tariff
Japan (Toyota Camry): $43k pre-tariff
50 buyers; perfect substitutes

A) Baseline (free trade)

Everyone buys Sonatas from Korea at $40k.
U.S. tariff revenue: $0.

B) Uniform 10% tariff on all countries

Sonata: $40k → $44k
Camry: $43k → $47.3k

Consumers buy Sonatas from Korea (lowest-cost source preserved).

Per car: consumers pay +$4k; government gets +$4k.
Totals (50 cars):
Consumer loss: $200k
Tariff revenue: $200k
Total losses (national): ≈ $0 (consumers transfer $ to government; ignoring DWL from Q changes).

C) Selective 10% tariff (on Korea only)

Sonata (Korea): $40k → $44k
Camry (Japan): $43k (untaxed)

Buyers switch to Camry’s from Japan (trade diversion).

Totals (50 cars):
Consumer loss: 150k (consumers now pay $43k vs $40k in the no tariff baseline → $3k more per car)
Tariff revenue: $0 (imports shifted to untaxed Japan).
Net (national): –$150k.
Global efficiency: production cost rises from $40k → $43k per car → $150k real resource loss.

Total losses: 300k (consumer loss + real resource loss)

The total losses under scenario C in which just some countries are tariffed are larger than in scenario B in which all countries are tariffed! What’s going on? Under a uniform tariff, the lowest-cost supplier still wins. Tariffs create distortions, but a uniform tariff at least preserves efficient sourcing and generates government revenue. Under a selective tariff, sourcing can shift to a higher-cost supplier purely because of the tariff. That’s trade diversion —bad for efficiency which means some combination of consumers and government must lose.

Here’s an analogy. Forget tariffs for a moment and imagine taxing GM but not Ford. That could make Ford win sales even if GM can produce the same car more cheaply — an obvious waste. The same logic applies in international trade.

In general, as Brian Albrecht argues, tariffs are a costly way to raise revenue. Selective tariffs are especially inefficient and wasteful. Sad to say, the U.S. tariff system today is highly selective — wildly different rates on different countries and times. Trade diversion isn’t a necessary consequence of selective tariffs but our current high and chaotic tariff structure makes it all but inevitable. Thus selective tariffs mean standard deadweight losses will compound with large-scale trade diversion and inefficiency, raising losses above headline numbers. Finally, the selective structure invites rent seeking, as firms and industries lobby for favorable treatment — adding yet another layer of economic waste.