Globalization of Labor Markets and Inequality

In thinking about the impact of the globalization of labor markets on inequality, economists naturally turn to the Hecksher-Ohlin model. While in many contexts the effects of trade and migration can presumably be well understood through this model, in other contexts it may be less appropriate. I argue below that considering models other than the Hecksher-Ohlin model may suggest very different implications regarding the effects of the globalization of labor markets on inequality. I first discuss Kremer and Maskin, which looked at the impact of trade on inequality in poor countries, and then discuss Kremer and Watt, which examines the impact of immigration on inequality in rich countries.¹

Trade

Consider a standard Hecksher-Ohlin model with two complementary factors of production: skilled labor and unskilled labor. Suppose there is a rich country with a high ratio of skilled to unskilled labor and a poor country with a low ratio of skilled to unskilled labor. Under the Hecksher-Ohlin model, allowing trade will equalize factor prices, and therefore skilled wages will rise in the rich country and fall in the poor country. Unskilled wages will fall in the rich country and rise in the poor country. This result of trade increasing inequality in the rich country resonates with many in the United States and in Europe.

However, the model also implies that trade will decrease inequality in the poor country. Many, however, are not convinced that this implication of the model is consistent with the experience of poor countries. People opposed to globalization often see it as benefiting elites in the rich world and the poor world at the expense of the poor in each world. The moderate left often argues that for poor countries to take advantage of globalization they need to educate their labor force, an argument that seems difficult to reconcile with the standard Hecksher-Ohlin model.²

Kremer and Maskin argue that there are empirical problems with the Hecksher-Ohlin model's predictions regarding the effects of trade in poor countries and then propose a model that could account for some of these effects.

Empirical Problems with the Hecksher-Ohlin Model

There are at least two problems with the standard Hecksher-Ohlin model. First, the standard model would predict that when factor endowments are most different, trade should be greatest. However, there is not much trade between the rich countries and the poorest countries.

Second, there is little evidence that globalization decreases inequality in developing countries, while there is some evidence that globalization in fact increases inequality in these countries. The evidence is from episodes of trade liberalization and also cross-country regressions. Several studies suggest that wage inequality increased when Mexico joined the General Agreement on Tariffs and Trade (GATT) in 1985 and embarked on a broad liberalization of trade and foreign investment.³

Time series studies that found that wage inequality increased after globalization in developing countries include work on Argentina, Chile, Colombia, Costa Rica, and Uruguay.⁴ Wood surveyed the literature and concluded that increased openness was associated with reduced wage inequality in the Asian Tiger economies in the 1970s and 1980s but with increased inequality in Latin America in the 1990s.⁵

Lindert and Williamson argued that the limited evidence available for other countries indicated that liberalization tends to be followed by increases in inequality, but causality is doubtful, particularly since in several large countries (India, China, Russia, and Indonesia) liberalization had been only partial.⁶

Most of the recent panel data literature either found that trade liberalization was positively associated with inequality in poor countries or found no strong association. Barro's survey of inequality and growth in a panel of countries found that the relationship between openness and inequality was positive for low-income countries and negative for high-income countries, with the turning point occurring at a GDP per capita level of $13,000.⁷ Kapstein and Milanovic found a similar result, with a turning point around $6,000.⁸ Lundberg and Squire found in panel data (which included developed and developing countries) that the income share of the lowest two quintiles was...