Trade and wages

What effect does freer trade have on the gap between skilled and unskilled workers' pay?

TRADE is frequently blamed for fuelling wage inequality in rich countries. As tariff barriers fall and imports rise, goods produced by cheap labour in poor countries compete with those made by better-paid workers in wealthy ones. Some factories close, others slim their payrolls to cut costs. At the same time, however, liberal trade causes exports to increase: new factories open, others increase their payrolls to expand production. Working out the net effect on the pay of different workers is a subtle problem.
     A theory first developed in the 1940s is frequently applied to the question. Known as the Stolper-Samuelson theorem, it is a cornerstone of modern thinking about trade. Paul Samuelson and Wolfgang Stolper predicted that removing trade barriers would have different effects in different countries. If a country has lots of skilled labour, its exports will tend to be intensive in skilled labour, and skilled workers will gain from more trade. But if unskilled labour is more abundant, its exports will be intensive in unskilled labour, so workers with skills will lose.
     Rich countries have relatively more skilled labour, so freer trade should raise the wages of skilled workers and lower the wages of the unskilled. Accordingly, many economists argue that trade accounts for at least some of the growing wage disparity between high-and low-skilled workers in America, Britain and other advanced economies.
     The changing distribution of income in rich countries, especially in America and Britain, has been widely noted. Less well known is the fact that similar changes have occurred in some developing countries. Pay differentials between skilled and unskilled workers are often high (see chart); a skilled industrial worker in Thailand, for example, earned around $7,100 in 1995, according to UBS, a Swiss bank, while an unskilled building labourer earned barely one-quarter as much. Since tariffs have been lowered, the gap has widened in some countries. In Mexico, for instance, the difference between a typical university-educated worker's pay and that of an unskilled worker rose by a third between 1987 and 1993.
     What makes these figures intriguing is that, on the face of it, they contradict the evidence from rich countries and refute the Stolper-Samuelson theorem. Why? Because the theory suggests that freer trade should reduce inequality in poorer countries, just as it increases it in rich ones. In effect, the theory says that whatever sort of labour is most abundant in the economy before liberalising is the one that prospers most as a result of trade. Poor countries have lots of unskilled labour. They should therefore export goods that are intensive in unskilled labour, pushing up its wages.
     Adding to the puzzle, if the Stolper-Samuelson theory works anywhere, it should work in the third world. Trade often plays a much more important role in many developing economies than it does in rich ones, and many poor countries have recently undertaken radical trade liberalisations. The average tariff in Latin America, for example, has fallen from 45% in 1985 to 13%. Yet Donald Robbins, of Harvard University, finds that wage disparities increased after trade liberalisation in Chile, Uruguay, Colombia and Costa Rica, as well as Mexico*.
     It would be rash to draw firm conclusions from this handful of countries. (Unfortunately, few others publish the detailed information on individual workers that is required to track wage inequalities.) Nonetheless, Mr Robbins's findings demand an explanation. What has gone wrong with the theory? There are three possible answers:

Other factors drive wages. In every case, much else will have changed during the span of time over which wage inequality is being compared--a change in the trade regime is only one of many differences. In general, growth is going on. This means that countries are accumulating capital and new technologies, a process that will itself have a big effect on the distribution of income. Many other disturbances may play a part. In Latin America, for instance, countries have wrestled with high inflation, deep recessions and huge swings in foreign-capital flows. Any of these shocks could have influenced wages and overshadowed the direct impact of trade.

Trade raises the demand for skills. Traditional trade theory makes two big assumptions--that both rich and poor countries use the same kind of technology, and that capital is internationally immobile. The question is not whether these assumptions are false: they obviously are. The question is how much it matters. By reducing its trade barriers, a country may be able to import more new capital equipment. This will bring new technology, which will raise the demand for higher-skilled workers to use it. Thus freer trade will raise the premium paid to skilled workers in a way not envisaged by Stolper-Samuelson.

The China effect. Trade theory traditionally lumps all developing countries together as having relatively unskilled workforces. But there are substantial differences among them. China, for example, has hundreds of millions of unskilled workers and relatively few with skills. Many middle-income developing countries that compete with China in world trade (Colombia, say, or Chile) have relatively high average skill levels. In such countries, trade liberalisation might be expected to cause the wage gap to rise in the same way as in rich countries.
The evidence on wage inequality in the third world will be greeted by many as "further confirmation" that trade hurts the poor. That would be wrong. Greater inequality may mean only that higher incomes are rising faster than lower ones, not that lower ones are falling. Beyond that, far from confirming anything, the results only deepen a mystery. They challenge the theory which seemed, to some, to explain why trade harmed low-income workers in rich countries. If that theory does not fully explain wage changes in the third world, it may also be deficient when applied to rich countries. The main explanation for widening wage inequality in rich and poor countries may lie not with trade, but elsewhere.

*"Evidence on Trade and Wages in the Developing World", OECD Development Centre Technical Paper No 119. December 1996

The Economist Home Page - Search - Contents - Next article