Article body

Why is the debate about globalization so rancorous and unproductive? To anti-globalists, it seems obvious that import competition from low-wage countries such as China and India is forcing down wages and living standards in the developed countries of Europe and North America. Economists can’t see what’s in front of their faces due to their fixation on outmoded mathematical models. To economists and other promoters of free-trade, the anti-globalists are simply ignorant; they fail to understand the most basic of economic principles, i.e., wages reflect productivity, so that having low wages in no way confers an “unfair” competitive advantage to developing countries. Moreover, even when trade does create losers, it is possible to mathematically prove that the gainers gain more than the losers lose. Hence, it makes no more sense to protect textile workers from low cost imports than it does to protect typewriter manufacturers from low cost PCs.

Why can’t the two sides of this debate agree? In International Trade and Labor Markets, two leading trade theorists, Carl Davidson and Steven Matusz, propose a simple and persuasive answer. They argue that the two sides are speaking different languages. Their basic premise is that anti-globalists are right to point out that economists often defend free trade with outmoded models that are incapable of addressing their concerns. At the same time, however, the authors argue that it is possible to incorporate the concerns of the anti-globalists in fully articulated general equilibrium models that respect basic economic principles. The authors contend that the key to bridging the gap is to enrich the labour market specification in traditional models of international trade. Traditional models abstract from search and other frictions, and assume all prices are fully flexible. As a result, resources (in particular labour) are always fully employed. Trade does influence labour market outcomes and the distribution of income, but only by influencing market clearing prices. It does not influence employment prospects. This is unfortunate, since the public debate is more about jobs than about wages.

Summarizing their recent research on labour markets and international trade, Davidson and Matusz show how realistic frictions can be incorporated into standard trade models in a way that preserves analytical tractability yet at the same time generates an equilibrium unemployment rate that is responsive in predictable and empirically testable ways to the forces of globalization. They do this by wedding the well known Mortensen-Pissarides search model to a standard international trade model. In this model, it takes time and effort for workers to find new jobs and for firms to hire and train new workers. Most of the analysis centres on a simple two-sector example in which one sector (Sector 1) features low wage jobs that are easy to find, and another sector (Sector 2) features high wage jobs that are more difficult to find. Both jobs are assumed to disappear at their own exogenously specified “job destruction rates”. Once a worker loses his job, he must decide where to start looking for a new job. He knows that Sector 2 has higher wages, but he also knows that it is harder to find a job there. In equilibrium, Sector 2 wages are just high enough to offset the higher search costs.

One of the major conclusions from the analysis is that labour market institutions, to the extent that they influence relative search costs, can become a source of comparative advantage, and thereby influence the pattern of international trade. For example, if an industry’s job destruction rate is relatively high in a particular country, then that country is likely to be an importer of that good. The reason is clear. High job destruction rates (and/or low job finding rates) translate into high autarky wages and prices. More surprisingly, they show that it is quite possible for countries with relatively flexible labour markets to experience a permanent increase in unemployment following an expansion of trade.

Most of the book involves fleshing out the implications of the connection between search frictions and international competitiveness. The first two chapters offer background material. Chapter 1 highlights the discrepancies between the questions being raised in public debates about globalization and the questions capable of being addressed in traditional international trade models. Chapter 2 provides a nice summary of the recent debate between international economists and labour economists about the relative importance of trade to the decline in wages of low skilled workers. Labour economists have tended to assign a significant role to trade, whereas international economists have found much smaller effects, and instead point the finger at skill-biased technological progress. Davidson and Matusz don’t really take sides in this debate. Instead, they show how different conclusions arise from different methodological assumptions. More to the point, they argue that much of this debate misses the mark, since it is based on highly unrealistic models of the labour market.

The heart of the book is Chapter 3, which incorporates search frictions into a one-factor Ricardian trade model. In addition to showing how the structure of labour markets influences the pattern of trade, this chapter also shows how search frictions provide a natural bridge between the predictions of Heckscher-Ohlin and Ricardo-Viner models of trade and income distribution. The Heckscher-Ohlin model predicts that trade benefits a country’s abundant factors and hurts its scarce factors. Ricardo-Viner models predict that trade benefits all factors specific to exporting industries, while hurting all factors specific to import-competing industries, with the effects on mobile factors being ambiguous. With search frictions, factors are neither industry-specific nor perfectly mobile. Currently searching workers are in a sense mobile, since it is assumed they can redirect their search efforts without cost. Currently employed workers are in a sense specific, since they must incur search costs to change jobs. Assuming wage bargaining is efficient, the authors show that Heckscher-Ohlin predictions apply to searching factors, while Ricardo-Viner predictions apply to currently employed workers.

The final two chapters provide empirical evidence and discuss policy implications. Chapter 4 provides surprisingly strong and robust evidence that net exports across U.S. industries are negatively related to job destruction rates. Chapter 5 speaks most directly to recent policy debates. It attempts to go beyond the “compensation principle” by showing that trade not only represents a potential Pareto improvement, but can actually become a Pareto improvement with appropriate policies. These policies call for targeted wage subsidies to movers and targeted employment subsidies to those who choose to remain in the import-competing sector. Interestingly, it turns out not to be very costly in efficiency terms to completely compensate the losers from trade.

This is an excellent book, which will be of interest to both academics and policymakers. It is a Krugmanesque blending of clear and concise prose, simple yet rigorous modeling, and suggestive empirical evidence. My main complaint concerns its assumption that turnover rates are exogenous. I buy the argument that turnover rates influence trade. However, it seems equally clear that trade feeds back and influences turnover rates. The recent work of Sargent and Ljungqvist provides a vivid example. Another drawback is that the compensation analysis is group-focused rather than based on individual labour market experiences. Matching models are nice because they produce deterministic aggregate dynamics. However, this masks a lot of individual randomness and heterogeneity, and it seems unlikely (and undesirable) that policy would not respond to the vagaries of the search process.