Many labor representatives in the United States and Canada fear that a North American Free Trade Agreement (NAFTA) will lead to a redirection of investment toward low-wage regions of the continent, consequently reducing wages and employment opportunities in higher-wage areas. These concerns seem to contradict several traditional conclusions of economic theory, and the results of quantitative economic models have suggested that a NAFTA could be beneficial for workers in all three countries. These modeling efforts, however, have been based on highly restrictive assumptions regarding international investment flows, the functioning of labor markets, and other features of the economy. More realistic trade and investment models are needed in order to fairly judge the risks and opportunities posed to labor by continental integration. Some suggestions for future research are provided.
Get full access to this article
View all access options for this article.
References
1.
1. Although not necessarily all individuals in all countries; see later discussion.
Gregory Schoepfle and Jorge Perez-Lopez, “U.S. Employment Impact of TSUS 806.30 and 807.00 Provisions and Mexican Maquiladoras” (Economic Discussion Paper #29, Bureau of International Labor Affairs, 1988). Estimates of employment impacts vary widely according to the assumptions used by investigators, but much evidence suggests that the negative impact on U.S. labor of the large trade deficits of the 1980s has been quite considerable.
4.
4. One model of this type has been used to illustrate the wisdom of Canada's National Policy, which used high tariffs to promote the early development of Canadian industry, and without which the production of increasing-returns manufactures would have become concentrated in the United States; see Paul Krugman, Geography and Trade (Cambridge: MIT Press, 1991).
5.
5. Traditional theory holds that free trade would nevertheless hold the potential for an overall increase in welfare if compensation were paid to the owners of the scarce factor in each country. This welfare result is of theoretical interest only, however, since no such compensation has been proposed, nor could it be practically paid in a real economy. For analyzing the impact of a NAFTA on labor, therefore, more concrete measures of economic well-being—such as GDP, employment, and real wages—are more relevant.
6.
6. Drusilla K. Brown, Alan V. Deardorff, and Robert M. Stern, “A North American Free Trade Agreement: Analytical Issues and a Computational Assessment,”World Economy, 16:1, 11-29 (1992).
7.
7. All models, of course, recognize that there will be temporary industry-specific dislocations following a trade agreement. Their contention is that these transitional losses would be outweighed by a presumed increase in long-run aggregate economic prospects; what is in question is whether or not this long-run aggregate gain can indeed be expected.
8.
8. This terms-of-trade result is actually a win-lose effect, since it harms living standards in the rest of the world.
9.
9. Ray C. Fair, Specification, Estimation, and Analysis of Macroeconometric Models (Cambridge: MIT Press, 1984), p. 18.
10.
10. U.S. firms alone currently account for over two-thirds of foreign direct investment in Mexico.
11.
11. After Paul S. Armington, “A Theory of Demand for Products Distinguished by Place of Production,”IMF Staff Papers, 16:159-176 (1969). Certain models assume that products are differentiated by firm rather than by nation, but these models, by prohibiting international capital flows, ensure that firm-specific product varieties are also nation-specific, with the same effect as the traditional Armington assumption.
12.
12. In technical terms, Mexican imports are assigned a very low share parameter in multitiered Armington consumer demand functions.
13.
13. Note that one CGE model of the North American auto industry, by Linda Hunter et al., abandons this assumption and finds that U.S. and Canadian auto output falls after a NAFTA. See Hunteret al., “Trade Liberalization in a Multinational-Dominated Industry,” in Economy-Wide Modeling. This is in sharp contrast to other CGE models, which generally expect large gains in U.S. auto output, thanks largely to the Armington assumption.
14.
14. A more technical review of these and other modeling assumptions is contained in James Stanford, “CGE Models of North American Free Trade: A Critique of Methods and Assumptions” (Testimony to U.S. International Trade Commission Investigation #332-317, Apr. 1992).
15.
15. The original model was constructed by Raul Hinojosa-Ojeda and Robert McCleery. As with other models, assumptions regarding migration and labor market segmentation are important to the results reported by Spriggs.
16.
16. The model's full-employment feature ensures that displaced workers find alternative employment but with a consequent fall in wages.
17.
17. Interindustry Economic Research Fund, Industrial Effects of a Free Trade Agreement between Mexico and the USA (Washington, DC: Department of Commerce, 1990).
18.
18. Mexican goods do not compete directly with American products in this model; rather, a NAFTA, by lowering the cost of Mexican sales to the United States, is felt only via a very small decline in the average price of all U.S. imports. The penetration of Mexican-made products into U.S. markets following a NAFTA is thus strongly and artificially restricted.
19.
19. North American Free Trade: Issues and Recommendations (Washington, DC: Institute for International Economics, 1992).
20.
20. Note that all but one of these case histories involved the unilateral liberalization of trade and investment rules, not bilateral free trade with a more developed economy; thus the trade balance effects of a NAFTA will likely be much better for Mexico—and thus worse for the United States—than is implied by these other cases. Note also that much of the presumed benefit for the United States of Mexican reforms has already occurred and cannot be credited to a NAFTA.
21.
21. There is a grain of truth in this approach, but to ignore the impact of a trade deficit on output in a heavily indebted nation such as Mexico is unjustified.
22.
22. “Effect of the North American Free Trade Agreement on Investment, Employment and Wages in Mexico and the U.S.” (Working Paper, Skidmore College, Feb. 1992).
23.
23. The passage of the FTA went virtually unnoticed in the United States, compared to the controversy that has surrounded a NAFTA. American commentators on both sides of the debate often mistakenly conclude that this was because the similarity between Canada and the United States reduced the risk of NAFTA-type adjustment costs. In fact, the FTA was fiercely controversial in Canada for many of the same reasons as the NAFTA in the United States: a fear that lower wages, taxes, and labor standards in the United States—particularly in right-to-work states—would divert investment, production, and employment away from Canada.
24.
24. In the Cox-Harris model, which allows capital flows to and from a perfectly competitive world capital market, this improvement in productivity—and hence profitability—actually led to a huge capital inflow, dramatically increasing the stock of foreign investment in Canada. Ironically, then, the elimination of tariff barriers was presumed to lead to a great increase in a stock of foreign capital that was motivated by the initial existence of those barriers.
25.
25. “Wage Effects of a U.S.-Mexican Free Trade Agreement” (Working paper, Department of Economics, University of California, Los Angeles, Apr. 1992).
26.
26. So named because it attempts to capture the impact of socioeconomic structures on income distribution and hence factor prices.
27.
27. Of course, trade imbalances cannot be financed forever, but they can persist long enough to structurally damage a country's economic competitiveness. It may take a generation, for example, until foreign creditors no longer accept Treasury bills, and have exhausted the supply of desirable U.S. real estate properties, to offset their persistent trade surpluses with the U.S.
28.
28. Models incorporating these alternative assumptions can certainly be constructed, although for various reasons they have not been widely applied to the issue of a NAFTA. The author's present research involves the construction of comparative CGE and econometric models of North American integration. See also Lance Taylor, ed., Socially Relevant Policy Analysis (Cambridge: MIT Press, 1990) for other examples of structuralist CGE models.
