Has Global Trade Fueled US Wage Inequality? A Survey of Experts
A familiar framework for examining the relationship between international trade and wages is the Heckscher-Ohlin model (H-O) developed by the early 20th century Swedish economists Eli Heckscher and Bertil Ohlin. H-O says that a country like the United States with abundant skilled labor will export goods that are intensive in skilled labor, while developing countries with large endowments of unskilled labor will export unskilled-labor-intensive products. The H-O model and related theorems are widely taught in standard international economics courses and help explain why trade between the United States and developing countries can lead to a decrease in the relative wage of US unskilled workers and thereby an increase in wage inequality.
This is the textbook theory. Not surprisingly, the striking rise in inequality in the 1980s motivated several economists to explain skill-related US wage inequality using the H-O model. Before running his own model, William R. Cline (1997) reviewed the literature published in the 1990s that examined US wage inequality in the 1980s comprehensively, including articles by both labor and trade economists (table 1). While methodologies and results are mixed, the late 1990s consensus differed from the standard textbook theory: The consensus seemed to hold that technological changes that place a premium on skilled workers—not international trade—constituted the dominant factor contributing to increased US inequality in the 1980s. International trade was characterized as a less important factor, along with immigration and weaker trade unions. As Cline (1997) summarized, “international influences contributed about 20 percent of the rising wage inequality in the 1980s.” In 2008, drawing from almost 30 studies on the trade and wages, Bivens (2008) concluded that the contribution of trade to rising inequality was 10 to 40 percent in the 1980s and early 1990s. However, very few studies supported the high end of this range.
|Table 1 Estimates of the contribution of trade to rising US wage inequality during the 1980s|
|Author||Portion of change due to trade||Remarks|
|Labor economists 1: Emphasis on skill-biased technological change|
|Bound and Johnson 1992||minimal||Technological change (dominant); changes in the structure of product demand was not a major factor, implying that trade has minimal impact|
|Mincer 1991||small||Technological change (primary influence)|
|Juhn, Murphy, and Pierce 1993||inconclusive||Rising demand for skills is the dominant cause|
|Berman, Bound, and Griliches 1994||minimal||Technological change (main influence)|
|Freeman 1991||inconclusive||Falling union density: 15-40 percent|
|Burtless 1995||small||Interprets wage trends as showing neutral impact of trade|
|Labor economists 2: More emphasis on the role of trade, immigration, and other international forces, while also acknowledging a role for technological change|
|Katz and Murphy 1992||modest||Trade reduced demand for high school dropouts by 0.6-4 percent and increased demand for college graduates by 0.6-1.5 percent; but technological change was the main influence|
|Murphy and Welch 1991||moderate||US shift to large trade deficit reduced unskilled-intensive durable industries and boosted skilled-nontradable services|
|Borjas, Freeman, and Katz 1992||20-40 percent||Trade and immigration increased the relative supply of unskilled workers. These forces were responsible for nearly 20 percent of the college/high school wage differential in mid 1980s and 40 percent of the other workers/high school dropouts wage differential|
|Borjas and Ramey 1994||very high||Tests indicate cointegration between the ratio of durable trade balance as a percentage of GDP and wage differential(s); authors question the importance of technological change|
|Borjas and Ramey 1993||10 percent||Immigration: 20 percent|
|Karoly and Klerman 1994||55-141 percent||Regional tests show 55 - 141 percent of inequality rise due to rising durable goods imports; an important effect of changing unionization (4 - 17 percent)|
|Borjas 1994||inconclusive||Immigration explains 1/3 of wage inequality|
|Bernard and Jensen 1994||not available||Between-plant demand shifts, especially to exporting plants, account for much of the rising nonproduction worker share of the wage pie|
|Trade economists 1: Emphasis on skill-biased technological change|
|Lawrence and Slaughter 1993||none||Stolper-Samuelson effects rejected|
|Krugman and Lawrence 1994||minimal||-|
|Sachs and Shatz 1994||very minimal||Increased net imports (1978-90) reduced relative unskilled/skilled manufacturing employment by 5 percent|
|Cooper 1994||10 percent||Decline in textile, apparel, and leather manufacturing as impact of trade upon unskilled|
|Baldwin and Cain 1994||9-14 percent||-|
|Krugman 1995a||10 percent||-|
|Trade economists 2: More emphasis on the role of trade, immigration, and other international forces, while also acknowledging a role for technological change|
|Leamer 1992||major||1972-85 trade reduced US unskilled wages by $1,000 and raised skilled wages by $6,000|
|Leamer 1994||20 percent||Changes in relative factor supply and ongoing technological change are dominant factors|
|Leamer 1995||major||Globalization: most powerful force|
|Feenstra and Hanson 1995||15-33 percent||Globalization explains 15-33 percent of wage inequality|
|Development economists 1|
|Wood 1994||-||Imports of manufactures from LDCs less-developed countries depress the demand for unskilled labor in the North by 22 percent of the unskilled labor force|
|Source: Table 2.3 of Cline (1997), simplified by the authors.|
More recent studies of US wage inequality have drawn conclusions similar to those earlier summarized by Cline (1997). Bivens (2007), building on a general equilibrium model developed by Krugman (1995), found that in 2006, US trade with less developed countries raised the relative wage of skilled workers by 6.9 percent compared to a no-trade scenario, up from 4.8 percent in 1995. Katz (2008) pointed out that this change in the relative wage of skilled workers could explain only 15 to 19 percent of the rise in the college wage premium between 1980 and 2006. Edwards and Lawrence (2013) likewise concluded that the H-O model explains a minor part of rising US wage inequality in recent years.
Contrary to the large amount of research done within the H-O framework using US data from the 1980s, few empirical studies have used H-O to examine more recent data. One possible reason, noted by Lawrence (2008) and Edwards and Lawrence (2013), is that, while rising US inequality in the 1980s is clear, whatever the classification of wages by skill, the story was less pervasive in the 1990s and mixed after 2000. Meanwhile, the H-O model builds on extremely strong assumptions, of which the three most unrealistic assumptions call for homogeneous and fully mobile factors, identical technology in all countries, and perfect substitution between foreign and domestic products. Due to these model limitations, economists turned to other theories, as well as new empirical tools, to examine the possible relationship between trade and wages.
A more recent line of theoretical trade literature—led by Krugman (1979, 1980), Melitz (2002), and Bernard et al. (2003)—modeled heterogeneous firms in a framework of monopolistic competition with differentiated goods and economies of scale. Meanwhile, recent empirical research at disaggregated levels studied the effect of trade on local labor markets and found that the impact of US trade with developing countries on US wage inequality is evident in specific industries, regions, and occupations (table 2). Many of these empirical studies also found a significant impact of trade on local employment in terms of job losses. This quick survey, however, does not cover employment effects but only focuses on changes in wages.
|Table 2 Recent findings on the effect of trade on US local labor markets|
|Ebenstein et al. 2009||1984-2002||A 10% increase in occupational exposure to import competition --> 2.9% decline in real wages for workers in all occupations covered in the sample.|
|But no statistically significant effect of international trade on worker wages when import competition is measured at the industry level.|
|Ebenstein et al. 2015||1983-2008||This study updates Ebenstein et al. (2009) by extending the period to 1983-2008.|
|A 10% increase in occupational exposure to import competition --> 2.7% decline in real wages for workers in all occupations covered in the sample.|
|But no statistically significant effect of international trade on worker wages when import competition is measured at the industry level.|
|For 2000-2008, a 10% increase in occupational import penetration from China --> 5.6% decrease in wages of affected US workers.|
|Hakobyan and McLaren 2016||1990-2000||A high-school dropout employed in industries affected by NAFTA tariff reductions would have had a 17 percentage point higher wage growth if he had been employed in an unprotected industry.|
|A high-school dropout employed in locations specializing in industries most vulnerable to NAFTA tariff reductions would have had an 8 percentage point higher wage growth if he had been employed in a location with few protected industries.|
|Autor, Dorn, and Hanson 2013||1990-2007||A $1,000 per worker increase in a commuting zones (CZ) exposure to Chinese imports during a decade --> reduces mean weekly earnings of all workers by 0.76 log points. For an average weekly wage of $706.58 (calculated based on OECD average annual wage database), this means a drop of $5.40 per week.|
|But the increase in import exposure to Chinese products does not affect mean manufacturing wages in CZs.|
|However, a $1,000 increase in Chinese import exposure per worker --> reduces mean nonmanufacturing wages by 0.76 log points.|
|Feenstra, Ma, and Xu 2017||1990-2007||Augmented Autor et al. (2013) by controlling for changes in local housing prices.|
|No depressing effect of increasing Chinese import exposure on local labor market wages.|
|Rothwell 2017||1990-2007||Reexamined Autor et al. (2013) results.|
|A $1,000 per worker increase in a CZ's exposure to Chinese imports during a decade over 1990-2000 --> increases manufacturing mean weekly earnings by 1.6 log points. For an average weekly earning of $497.71 (calculated based on Average Hourly Earnings of Production and Nonsupervisory Employees: Manufacturing reported by FRED), this means an increase of $8.00.|
|A $1,000 per worker increase in a CZ's exposure to Chinese imports during a decade over 2000-2007 --> increases manufacturing mean weekly earnings by 1.2 log points. For an average weekly earning of $643.14 (calculated based on Average Hourly Earnings of Production and Nonsupervisory Employees: Manufacturing reported by FRED), this means an increase of $7.70.|
|No depressing effect on non-manufacturing wages.|
|NAFTA = North American Free Trade Agreement; OECD = Organization for Economic Cooperation and Development; CZ = communting zone; FRED = Federal Reserve Economic Data
Source: Compiled by authors.
Ebenstein et al. (2009) used data from the Current Population Surveys during 1984–2002 to examine the impact of trade on local wages. The authors did not find that imports depress industry-wide wages but did find a negative effect at the occupation level: A 1 percent increase in occupational exposure to import penetration leads to a 0.3 percent decrease in real wages of those workers. The finding applies to all occupations. Results in a follow up study by Ebenstein et al. (2015) arrived at similar conclusions. Hakobyan and McLaren (2016) analyzed the impact of the North American Free Trade Agreement (NAFTA) on local labor market wages, using US Census data from 1990 to 2000. Empirical results indicate that NAFTA lowered the rate of wage growth for blue-collar workers in the most vulnerable industries and localities that lost protection against Mexican imports as a consequence of NAFTA tariff reductions. Autor, Dorn, and Hanson (2013) found that increasing Chinese import exposure from 1990 to 2007 did not exert a downward pressure on US local manufacturing wages but did decrease earnings in nonmanufacturing sectors significantly. Contrary to Autor et al. (2013), two recent studies by Feenstra et al. (2017) and Rothwell (2017) both found that the China shock did not exert a negative effect on US local labor market wages.
Autor, David, David Dorn, and Gordon Hanson. 2013. The China Syndrome: Local Labor Market Effects of Import Competition in the United State.” NBER Working Paper No. 18054 (May). Cambridge, MA: National Bureau of Economic Research.
Bernard, Andrew, Jonathan Eaton, Bradford Jensen, and Samuel Kortum. 2003. Plants and Productivity in International Trade. American Economic Review 93, no. 4: 1268–90.
Bivens, Josh. 2007. Globalization, American Wages, and Inequality: Past, Present, and Future. Economic Policy Institute Working Paper 279.
Bivens, Josh. 2008. Everybody Wins, Except for Most of Us: What Economics Teaches About Globalization. Washington: Economic Policy Institute.
Cline, William. 1997. Trade and Income Distribution. Washington: Peterson Institute for International Economics.
Ebenstein, Avraham, Ann Harrison, Margaret McMillan, and Shannon Phillips. 2009. Estimating the Impact of Trade and Offshoring on American Workers Using the Current Population Surveys. NBER Working Paper No.15107. Cambridge, MA: National Bureau of Economic Research.
Ebenstein, Avraham, Ann Harrison, and Margaret McMillan. 2015. Why Are American Workers Getting Poorer? China, Trade and Offshoring. NBER Working Paper No. 21027. Cambridge, MA: National Bureau of Economic Research.
Edwards, Lawrence, and Robert Lawrence. 2013. Rising Tide: Is Growth in Emerging Economies Good for the United States? Washington: Peterson Institute for International Economics.
Feenstra, Robert, Hong Ma, and Yuan Xu. 2017. The China Syndrome: Local Labor Market Effects of Import Competition in the United States: Comment.
Furman, Jason, and Peter Orszag. 2015. A Firm-Level Perspective on the Role of Rents in the Rise in Inequality. Presentation at “A Just Society ” Centennial Event in Honor of Joseph Stiglitz at Columbia University (October 16).
Hakobyan, Shushanik, and John McLaren. 2016. Looking for Local Labor Market Effects of NAFTA. Review of Economics and Statistics 98, no. 4, October 2016.
Katz, Lawrence. 2008. Comment on Trade and Wages, Reconsidered. Brookings Papers on Economic Activity 1 (Spring): 103–37.
Krugman, Paul. 1979. Increasing Returns, Monopolistic Competition, and International Trade. Journal of International Economics 9, no. 4: 469–79.
Krugman, Paul. 1980. Scale Economies, Product Differentiation, and the Pattern of Trade. American Economic Review 70, no. 5: 950–59.
Krugman, Paul. 1995. Growing World Trade: Causes and Consequences. Brookings Papers on Economic Activity, Volume 1.
Krugman, Paul. 2008. Trade and Wages, Reconsidered. Brookings Papers on Economic Activity 1 (Spring): 103–37.
Krugman, Paul, Maurice Obstfeld, and Marc Melitz. 2011. International Economics: Theory and Policy, 9th ed. Boston: Pearson Addison-Wesley.
Lawrence, Robert. 2008. Blue-Collar Blues: Is Trade to Blame for Rising US Income Inequality? Policy Analyses in International Economics. Washington: Peterson Institute for International Economics.
Leamer, Edward. 1996. In Search of Stolper-Samuelson Effects on U.S. Wages. NBER Working Paper 5427. Cambridge, MA: National Bureau of Economic Research.
Melitz, Marc. 2002. The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity. NBER Working Paper 8881. Cambridge, MA: National Bureau of Economic Research.
Rothwell, Jonathan. 2017. Cutting the Losses: Reassessing the Costs of Import Competition to Workers and Communities. Draft Working Paper. Gallup, Inc.
Thierfelder, Karen, and Sherman Robinson. 2003. Trade and Tradability: Exports, Imports, and Factor Markets in the Salter-Swan Model. The Economic Record 79, no. 244: 103–111.
 For a detailed exposition of the model, see chapter 5 of Krugman, Obstfeld, and Melitz (2011).
 Most of the studies reviewed by Bivens (2008) were earlier covered by Cline (1997). The high end of Bivens’ assessment, 40 percent, came from two sources: Cline’s (1997) Trade and Income Distribution Equilibrium (TIDE) Model, which indicated that the contribution of increased trade (resulting from falling transportation and communication costs, as well as falling protection) to the rise in the ratio of skilled to unskilled wages in the United States during 1973–93 was one-third; and Leamer’s (1996) study, not covered in Cline’s (1997) literature review, which reported that, in the 1970s, trade effects contributed 40 percent to the decline in real wages of unskilled workers.
 Krugman (2008), however, speculated that because US trade with developing countries increased dramatically since the 1990s, trade now could be a more important contributor to US wage inequality than it was in the 1980s.
 Thierfelder and Robinson (2003) built a model that considered imperfect substitutability between imports and domestic goods. They concluded that changes in commodity prices, factor endowments, and the balance of trade all affect relative factor prices. When imports and domestic goods are poor substitutes, as is the case of international trade with developing countries, the sign of the Stolper-Samuelson effect can be reversed.
 Lawrence (2008) points out that while the skill-premium inequality decreased, another source of inequality came from the rising share of the super-rich in recent years. Hufbauer and Moran explored this aspect of inequality and found trade to play a very small role; see "Does Foreign Trade and Investment Reduce Average US Wages and Increase Inequality?" Part 1 and Part 2 for more detail. Another view on sources of rising inequality focuses on differences among firms rather than differences among individuals. Some superstar firms generate supernormal returns and share the returns with their employees, thereby increasing wage inequality (Furman and Orszag 2015).
 Hakobyan and McLaren (2016) found that roughly 1.3 million American workers experienced slower wage growth of 5 percent or more during the 1990s due to import competition from Mexico. The figure of 1.3 million comes from a calculation supplied privately by professors McLaren and Hakobyan to the authors.
 Autor et al. (2013) estimated an employment reduction of roughly 2 million workers due to the China shock.
 Feenstra et al. (2017) estimated that the net reduction in US employment due to increasing exposure to Chinese imports was about 0.8 million workers, or roughly half of Autor et al.’s estimation.