9.12 — Trade, Tariffs and Comparative Advantage

Here is a claim that sounds wrong and is true.

Suppose Country A is better than Country B at producing absolutely everything — better at making cloth, better at growing wine, better at everything else. It still pays both countries to trade.

That is David Ricardo's result from 1817, and it is the least intuitive correct idea in economics. Paul Samuelson, asked to name a proposition in the social sciences that was both true and non-obvious, named this one.

Why it works

The insight is that what matters is not how good you are at something, but what you give up to do it.

Take one person: a surgeon who is also the fastest typist in her city. She is better than her secretary at both surgery and typing.

Should she do her own typing?

No — and the reason has nothing to do with skill. An hour spent typing is an hour not spent operating. The cost of her typing is the surgery forgone, and that cost is enormous. The secretary's hour of typing costs only whatever else the secretary would have done, which is far less.

So the surgeon specialises in surgery, the secretary types, and both are better off than if each did everything. The surgeon's absolute superiority at typing is irrelevant.

That opportunity cost is comparative advantage, and Ricardo's contribution was to show it applies to countries.

A worked version. Suppose Portugal can produce a unit of cloth with 90 hours of labour and a unit of wine with 80. England needs 100 hours for cloth and 120 for wine. Portugal is better at both.

Now ask what each gives up. For Portugal, a unit of wine costs 80/90 — about 0.89 — units of cloth. For England, a unit of wine costs 120/100 — 1.2 — units of cloth. So wine is cheaper for Portugal to produce, in terms of cloth forgone, and cloth is cheaper for England in terms of wine forgone.

Portugal makes wine, England makes cloth, they trade, and total output of both rises. Neither country needed to be better at anything in absolute terms — only differently bad.

And this means every country has a comparative advantage in something, necessarily, because the ratios cannot all be the same. A country cannot be uncompetitive at everything.

What the model leaves out

The theory is correct and it is not the whole picture, and the omissions are where the political argument lives.

It is about aggregate gains, and it says nothing about distribution. Trade makes a country better off in total while making specific people worse off. The textile worker whose factory closes does not get a share of the consumer surplus enjoyed by everyone buying cheaper shirts. The gains are diffuse and the losses are concentrated, which is exactly the configuration that produces political backlash — the losers know who they are and organise, the winners barely notice.

It assumes factors move between industries. In the model, workers move from cloth to wine costlessly. In reality a fifty-year-old textile worker in a mill town does not become a software engineer, and the town does not relocate. Recent research — the "China shock" literature by David Autor and colleagues — found that American communities most exposed to Chinese import competition after 2000 suffered persistent unemployment, falling wages and social damage for over a decade. Adjustment was far slower and more painful than the theory assumed.

It is static. Ricardo says specialise in what you are currently good at. If a country's current advantage is in raw materials and agriculture, the model says stay there — and there is no country that got rich by staying there. Chapter 9.13 shows that every successful late developer deliberately built advantages it did not have.

It ignores increasing returns and learning. In many industries, cost falls with cumulative experience, so whoever starts first stays ahead permanently. Which industry a country ends up in is partly historical accident and partly policy, not a fact of nature.

And it assumes full employment and balanced trade, neither of which reliably holds.

The infant industry argument

The main intellectual case for protection, and it is not a fringe position.

The argument: a new industry in a developing country cannot compete initially with established foreign producers who have decades of experience, scale and financing. Protect it temporarily; it learns; then it competes.

It was formulated by Alexander Hamilton in the 1790s for the United States and by Friedrich List in Germany in the 1840s.

It is theoretically sound under specific conditions: learning effects must be real, the protection must be temporary, and the government must be able to withdraw it.

And the last condition is where it usually fails. Protected industries acquire political power and lobby to keep the protection. India's licensing system is the standing illustration (Chapter 6.28): protection intended as temporary became permanent, incumbents were shielded from competition, and quality and productivity stagnated for four decades.

Whereas East Asian protection was different in a specific and decisive way, which Chapter 9.13 develops: it was conditional on export performance. A firm that could not sell abroad lost its support. That single design feature is the difference between protection that builds capability and protection that builds rent-seeking.

The hypocrisy of the rich countries

This should be stated plainly because it is documented and it is systematically omitted from the standard account.

Britain built its industry behind protection. The Calico Acts against Indian textiles (Chapter 6.16), the Navigation Acts requiring trade in British ships, and high tariffs on manufactures — and it converted to free trade in the 1840s, when its industry was dominant and free trade served it.

The United States was among the most protectionist countries in the world for most of its industrialisation, with average tariffs on manufactures frequently above 40 percent between the 1860s and 1930s, and it became a free trade advocate after 1945 when its industry was unchallenged.

Germany, France, Japan and South Korea all used extensive protection and industrial policy during their catch-up phases.

The economist Ha-Joon Chang's phrase for this is "kicking away the ladder" — climbing up by one method and then telling those below to use another.

And India's own history is the sharpest case. Chapter 6.17 showed Britain protecting its market against superior Indian textiles and then imposing free trade on India when the advantage reversed, including the countervailing excise on Indian mill cloth in 1894.

The honest conclusion is not that protection is good. It is that the historical record does not support the claim that free trade is how countries develop, and that the policy advice given to developing countries has frequently differed from the policy the advisers' own countries used.

What the evidence actually shows

Trade raises aggregate income. This is about as well established as anything in economics. Open economies have grown faster than closed ones on average, and the collapse of trade in the 1930s (Chapter 9.9) made the Depression worse everywhere.

Trade lowers prices for consumers, which is a real gain and falls disproportionately on the poor, who spend a larger share of income on traded goods like clothing and food.

Trade destroys specific jobs and creates others elsewhere, and the two do not happen in the same place or to the same people.

Trade is not the main cause of manufacturing job loss in rich countries. Automation is. US manufacturing output rose substantially while employment fell, which means output per worker rose — that is technology, not imports. Trade accelerated the process and concentrated it geographically.

And the political consequence is the finding that matters most. The regions most exposed to import competition showed measurable shifts toward populist and protectionist politics. A society that captures the aggregate gains from trade and does nothing for the concentrated losers will eventually lose the political consent for trade. Chapter 12.8 traces exactly that from 2008 to the present.

The institutions

GATT, from 1947, then the World Trade Organization from 1995.

Its principles: most favoured nation — a concession to one member must be extended to all; national treatment — imported goods treated the same as domestic once inside; binding tariff commitments; and a dispute settlement system with enforceable rulings, which was genuinely novel.

Its record. Average tariffs on manufactures fell from around 40 percent in 1947 to a few percent, and world trade grew enormously.

And it is currently paralysed. The Appellate Body, the WTO's supreme court, ceased functioning in 2019 because the United States blocked the appointment of new judges — a veto exercised by a single member because appointments require consensus. The dispute system that made the rules enforceable no longer works.

India's position has consistently been that developing countries need policy space, and it has fought hard on agricultural subsidies and on public food stockholding — defending the right to hold buffer stocks and to procure at support prices, which is essential to its food security system (Chapter 6.28) and which some other members regard as a trade-distorting subsidy.

The current turn

Tariffs are rising again, and the reasons are three and they are different from each other.

Security. Semiconductors, rare earths, pharmaceutical ingredients and batteries have become national security categories rather than ordinary goods, and countries are subsidising domestic production regardless of cost. Chapter 12.10 covers the geopolitics.

Resilience. Chapter 4.6's efficiency-fragility trade, relearned during the 2020 pandemic when single-source supply chains failed.

And politics — the constituency created by the concentrated losses above.

The honest assessment. Some of this is a rational correction to an over-optimised system. Some of it is protection dressed as security, which will raise costs and protect incumbents in the usual way. And distinguishing the two requires asking whether the support is conditional on performance, which is the test Chapter 9.13 supplies.

Where this shows up in your life

The price of nearly everything you own. A phone assembled from components made in a dozen countries costs a fraction of what a domestically produced equivalent would.

India's specific position. A large domestic market, a comparative advantage in services rather than manufacturing, historically high tariffs, and an ongoing argument about whether protection can build manufacturing capability or will repeat the Licence Raj. The production-linked incentive schemes launched from 2020 are explicitly conditional on output, which is a design lesson taken from East Asia, and their results are not yet in.

What the next page covers

Four countries took populations poorer than most of Africa in 1950 and made them rich within a single lifetime. Chapter 9.13 covers how Japan, South Korea, Taiwan and China actually did it — what each did in what order, what land reform and universal primary education had to do with it, how industrial policy was made to work when it fails almost everywhere else, what it cost politically, and which parts of it can be copied.