WILL NEW TRADE GAINS MAKE US RICH?
AN ASSESSMENT OF THE PROSPECTIVE GAINS FROM NEW TRADE AGREEMENTS

by Dean Baker and Mark Weisbrot

October 3, 2001

 Executive Summary

            This paper examines the likely effects of future trade agreements on economic growth and the living standards of most Americans. Standard models show that trade liberalization should lead to some gains in income, although these gains are considerably smaller than is often claimed by political figures. Economists also generally accept that trade liberalization has been one of the factors increasing inequality, redistributing wage income from workers without college degrees to workers with college and advanced degrees, in addition to shifting income from wages generally to profits.

            For the vast majority of workers—the three quarters of the labor force who lack college degrees—the negative distributional effects of trade over the last two decades almost certainly outweighed the positive growth effects, causing them a net loss of real income. This is true even under the assumption that trade was a relatively minor factor in the upward redistribution of income over the last two decades, and accepting the more optimistic assessments of the impact of trade on growth.

            * Using a low estimate of the impact of trade on wage inequality from Princeton economics professor Paul Krugman, three-fourths of the labor force has seen a net reduction in hourly wages, attributable to expanded trade between 1.6 percent and 2.4 percent.

            * Using a higher estimate of the impact of trade on wage inequality by William Cline of the Institute for International Economics, the net reduction in hourly wages for these workers is between 9.4 percent and 10.1 percent.

            * If Cline's estimate is adjusted to take account of indirect ways in which trade may lower wages—such as weakening unions' bargaining power—trade may have reduced the hourly wages of three-fourths of the labor force by between 12.2 percent and 12.6 percent.

Introduction

            In order to press their case for new trade agreements, some government officials have occasionally made questionable assertions. For example, Trade Representative Robert B. Zoelleck recently claimed that “between 1990 and 2000, exports of goods and services have accounted for one-fifth of U.S. economic growth.”[1] While this is literally true, it ignores the fact that imports have actually grown far more rapidly than exports, as the trade deficit soared from $71.4 billion in 1990 to $399.1 billion in 2000. Since imports are subtracted from GDP, in this simple accounting framework trade has been a serious drag on growth over the last decade.[2]

The previous administration also made claims about the gains from trade that were not always accurate. For example, in 1994, the President's Council of Economic Advisors reported that the Uruguay Round of GATT would lead to an increase in annual GDP of $100 to $200 billion by 2004 or 0.9 to 1.7 percent of projected GDP.[3] Five years later, this figure had been revised down considerably. In a paper published in 1999, the Council of Economic Advisors estimated the gains from the Uruguay Round at between 0.4-0.6 percent of GDP, less than half the gains that it had projected five years earlier.[4]

It is not just political figures who have been cavalier in touting the benefits of trade liberalization. In a recent paper, Harvard economist Jeffrey Frankel suggested that the gains from a major round of trade liberalization would be approximately 1 percent of world GDP (Frankel 2000, p. 34). At another point, the same paper cites studies indicating that Europe alone could increase its GDP by 7 percent if it carried through with liberalization of trade in agriculture and elsewhere (Frankel 2000, p. 10). Since Europe's economy is approximately one quarter of the world's economy, the gains estimated for Europe alone would be equal to approximately 1.75 percent of world GDP, an amount that is 75 percent larger than this study estimated as the gains accruing to the whole world. These two numbers are inconsistent; if the figure for Europe is plausible, then the estimate of worldwide gains is far too low. Alternatively, if the estimate of worldwide gains is plausible, then the gains estimated for Europe are hugely exaggerated.

            Economic theory predicts that trade liberalization will lead to economic gains. The basic argument is relatively simple: consumers will have the opportunity to buy cheaper goods if tariffs, quotas and other trade barriers are eliminated. The elimination of trade protection can hurt producers in the industries affected—for example the textile industry will lose jobs if trade barriers are eliminated—but it can generally be shown, given the important assumption of full employment, that the gains to consumers will exceed the losses to producers. Because of this result, economists generally view the removal of trade barriers as desirable.

            However, the fact that removing trade barriers can lead to increased output does not imply that it is necessarily a desirable policy. Reducing military spending leads to large economic gains in standard models, yet few people would consider eliminating the Defense Department a desirable policy. In the case of trade liberalization, the standard models imply that trade liberalization leads to a redistribution of income. Specifically, most of the forms of trade liberalization currently being considered would redistribute income from workers to corporations, and from lower wage workers to higher wage workers. It is entirely possible that for most workers, the lost wages due to the upward redistribution caused by trade liberalization outweigh the overall gains to the economy. This is especially true if trade liberalization results in long-term unemployment for workers in the affected industries; but even under the assumption of full employment, it is very possible for trade liberalization to cause a net loss of income for the majority of the labor force.

            If policy makers and the public at large are to have a reasonable basis for assessing the merits of future trade agreements, then it is essential that they have plausible estimates of the potential gains from liberalization, which can be balanced against the potential costs to the groups that are harmed. This paper is an effort to place a range on the size of these gains and losses. The next section examines a recent study by the International Trade Commission (a United States government agency that examines the impact of trade policy), which sought to quantify the economic gains from eliminating all identifiable trade barriers. The second section examines some of the distributional issues raised by this study and in other research. The third section examines the relative size of the gains from trade liberalization, with the losses to less educated workers that result from worsening income distribution.

How Large are the Gains?

            In the midst of much heated rhetoric, there have been some serious efforts to model the impact of trade liberalization. For example, a recent study by the International Trade Commission (ITC) estimated the gains to the United States from eliminating the remaining tariffs and quotas on imported goods.[5] This study used standard economic methods to carefully quantify the impact of eliminating these barriers. It also explicitly laid out the assumptions used in the model: most importantly, that there is no transitional unemployment resulting from the loss of protection.

            The study’s findings are informative. It found that the net welfare gains to the economy from eliminating all tariff and quotas, using 1996 prices and output, would be $14.9 billion annually (ITC, p 15). This would be equivalent to approximately $19 billion annually, if the gains were scaled up to the current size of the economy. While these gains are not trivial, they are less than 0.2 percent of GDP.

It is important to note some of the factors that caused the ITC estimate of gains to be even this large. The ITC estimated $3.6 billion of the gains from trade liberalization, or nearly 25 percent of the total, were attributable to the elimination of the rents associated with import quotas (p 14).[6] In standard trade models, import quotas lead to large costs to the importing nations, and large benefits to the producers in exporting nations. This is because they raise the price of imports to consumers—as tariffs do—but unlike tariffs, quotas allow foreign producers to pocket the price increase in the form of higher profit margins.

For example, suppose the United States places a quota on the number of foreign produced cars that can be sold in the United States. Since the supply of foreign cars is restricted, this move will raise the price, for example, from $20,000 to $22,000 per car. The foreign car producers can now make an additional $2,000 on every car they sell in the United States. They can keep this money as extra profit. By contrast, suppose the United States imposed a tariff of 10 percent on imported cars. While the tariff also raises the price of imported cars by 10 percent to $22,000, in this case the additional $2,000 goes to the United States government in the form of tariff revenue. Tariffs do not provide extra profits to foreign producers in the same way as do quotas. 

In standard trade models, such as the one used by the ITC, exporting nations will typically benefit when other nations apply quotas to their exports. While the quotas restrict the volume of exports, they allow exporters to charge a premium on each unit they are allowed to sell. In these models, the extraordinary profits that result from quota restrictions typically will far more than offset any profit reduction from selling fewer units.[7]

The structure of standard trade models implies that, in trade negotiations, nations should attempt to have import quotas placed on the goods they export, in order to increase the quota rents that their exporters earn.[8] In reality, nations usually push to have the barriers to their exports removed. In the context of this model, this means that the benefits from eliminating quota rents in the United States—which account for nearly one quarter of the total gains estimated from trade liberalization— may be offset by the elimination of quota rents earned by U.S. exporters in other nations. Since import barriers—including quota-type barriers—are generally higher in other countries than in the United States, it is entirely possible, that in a full model of multi-lateral trade liberalization, the loss of quota rents earned by U.S. exporters will more than offset the welfare gains to the United States by eliminating quota barriers here.[9] Even if the loss of quota rents earned elsewhere were assumed to be just equal to the welfare gains of eliminating quota rents in the United States, it would reduce the estimate of the gains from trade liberalization by approximately $4.6 billion to $14.4 billion annually.[10]

  There is a second important reason for believing that the ITC estimate overstates the gains from trade liberalization—given the assumptions of the model. This reason stems from its treatment of lost tax revenue. The elimination of all tariffs would result in a loss of tax revenue of approximately $20 billion a year. The ITC model assumed that this tax revenue is offset by a “lump-sum” tax of the same size. Lump-sum taxes are a purely hypothetical construction. In effect, lump-sum taxes imply that the government just pulls the revenue away from the economy—without affecting economic activity.

In reality, the government must impose specific taxes—income taxes, payroll taxes, sales taxes etc.—in order to obtain revenue. All taxes actually in use result in some economic distortions. Most economists put the range of the distortions at between 10-20 percent of the revenue raised by the tax. This means that the tax increases needed to make up for the lost tariff revenue would reduce the ITC's estimate of welfare gains by between $2-4 billion.  This would leave a net gain from trade liberalization of between $15 and $17 billion annually. If the loss of quota rents from U.S. exports offsets the gains from eliminating quota rents in the United States, then the ITC's estimate of the gains from trade liberalization would be further reduced to between $10.4 billion to $12.4 billion annually.  This places the gains from trade liberalization at between 0.10 and 0.12 percent of GDP. 

As noted earlier, an explicit assumption of standard trade models, like the ITC model, is that there is no transitional unemployment (ITC 1999, p7). Implicitly, this model assumes that workers who leave the industries that are losing jobs due to increased imports are immediately re-employed in the sectors where their skills can be best utilized.[11] In reality, many of these workers will be unemployed for at least some period of time, before they find another job.[12] Such spells of unemployment can significantly reduce the economic gains from trade liberalization, especially in the first years after barriers are reduced.

The ITC estimated that the job losses due to increased imports would be equivalent to 130,000 full-time positions. This job loss is equal to approximately 0.11 percent of the economy's total employment. If a significant percentage of these job losers take a long time to find new employment, or leave the labor force altogether without finding new jobs, then much of the gains from trade liberalization will disappear. (For example, since the percentage of the workforce facing displacement is approximately equal to the percentage of GDP for projected gains from trade, it follows that if one-third of these workers remained unemployed, the gains from trade would be reduced by approximately one-third.) This will be especially likely if there is a multiplier effect due to job losses—for example, store closings in a town where an apparel factory was the major employer. On the other side, the impact of job losses will be mitigated to the extent that positions are eliminated through attrition.

            There is one final point worth noting about the effect of unemployment on the economic gains estimated in the ITC model. Insofar as there are government transfer payments associated with unemployment caused by trade liberalization, such as unemployment benefits, food stamps, or other means-tested benefits, they come with an economic cost. Tax revenue must be raised to cover the cost of such benefits, and the economic distortions caused by higher taxes must be subtracted from the estimated gains from trade. This cost can be especially large if the upward redistribution of income caused by trade liberalization leads to large increases in payments for programs such as Medicaid or the earned income tax credit. The cost of these programs may rise not only because of workers being displaced by imports, but also as a result of an upward redistribution of income due to trade. This issue will be considered more directly in the next section.   

Trade and Income Distribution 

            A point on which nearly all economists agree is that trade has been one of the factors that has increased income inequality in the last two decades. This is both a prediction of trade theory, and an empirical finding in a large body of research. The prediction of trade theory—that in an industrialized country like the United States, trade should increase corporate profits and the income of highly educated workers at the expense of less educated workers—has been accepted by economists for more than fifty years.[13] A large body of empirical work supports this theoretical prediction (e.g. Krugman 1995; Cline 1997; Schmitt and Mishel 1996). The only real issue for most economists is the exact size of the effect of trade on inequality.

The findings of the ITC report are consistent with other work on trade and inequality. The report found that eliminating the trade barriers it examined would benefit corporations more than workers. For example, it estimated that eliminating the barriers to textile and apparel imports, which accounts for almost 70 percent of the total welfare gains in this study, would increase capital income by 0.14 percent, while increasing labor income by just 0.06 percent (ITC, p36 fn). It did not estimate the impact of trade liberalization on wage inequality because it did not distinguish between different types of labor. But if top quintile of workers—who account for close to half of all wage income—experienced the same rate of income gain as corporations, then it would mean that the bulk of wage earners were losers from trade liberalization in textiles and apparel.[14] This is exactly the sort of situation that could result from trade liberalization more generally.

It is important to recognize that the ITC study, like other research on this topic, does not attempt to measure indirect effects that trade liberalization could have income distribution. Most obviously this indirect effect can take the form of threats, where employers threaten to move their operations abroad unless workers make wage concessions. Bronfenbrenner (2000) found that in more than half of all unionization drives, employers threatened to close down all or part of their operations. The rate was much higher (68 percent) in mobile industries such as manufacturing, communications, and wholesale/distribution. Of the campaigns in which such threats were used, 18 percent of employers directly threatened to move to another country if the union were to win representation.

There is no easy way to measure the extent to which such threats might have lowered the wages of manufacturing workers, or less skilled workers more generally,[15] but one piece of evidence of the effectiveness of such threats is the decline in the unionization rate among manufacturing workers. The unionization rate in the manufacturing sector fell by 46.8 percent from 1983 to 2000, the sharpest rate of decline in any industry. (The first year for which consistent data exists is 1983.) This decline took place even though the manufacturing sector experienced virtually no job growth over this period.[16] By comparison, unionization rates in the construction industry fell by 33.5 percent over the same period, even though it started at an almost identical level.

Of course trade has not been the only factor depressing unionization rates. Greater hostility to unions from employers, and a less union-friendly National Labor Relations Board have also hurt unionization efforts. But it is striking that the unionization rate has fallen so much more in manufacturing than elsewhere. Even the transportation and communications industries, which were deregulated during this period, have not seen as large a decline in their unionization rates as manufacturing. Given the survey evidence found by Bronfenbrenner, and the trends in industry unionization rates over the last two decades, it is reasonable to believe that trade has weakened workers' bargaining power in manufacturing, leading to downward pressure on wages in ways that would not be picked up in standard economic models. 

Taking the Final Score--Gains and Losses from Trade 

As noted above, given the assumptions in standard trade models—most importantly that displaced workers are quickly re-employed—trade liberalization can increase aggregate output. However, the standard model also predicts that trade will increase income inequality—shifting income from wages to profits, and from low-wage workers to high-wage workers. This means that even if trade leads to gains for the economy as a whole, the upward redistribution of income that it causes can mean that most people lose from expanded trade. Whether most people are winners or losers will depend on the relative size of these two effects.

Many of the claims made for the gains from trade liberalization have little foundation. Even among those who have tried to seriously examine the issue, there is a wide range of estimates of the size of the impact of liberalization. At the low end, models along the lines used by the ITC indicate that gains from past trade liberalization may have increased annual GDP on the order of 0.1-0.2 percent over the last two decades. At the high end, some estimates—generally far less careful or plausible—have placed the gains as high as 1.0 percent of annual GDP.

By comparison, at the low end economists such as Krugman (1995) have attributed approximately 10 percent of the increase in wage inequality to trade. It is important to note that Krugman, like other economists who have done research in this area, makes no effort to calculate any indirect effects of trade liberalization on wages, such as the impact that a threat to move jobs abroad can have on wage negotiations or unionization effort. For this reason, this estimate would almost certainly underestimate the impact of trade liberalization on wage inequality. At the high end, Cline (1997) attributed 39 percent of the increase in wage inequality over the period from 1973-1993 to trade liberalization (7 percentage points out of a net increase in inequality of 18 percent).[17] This study also made no effort to incorporate any indirect effects of trade on wage inequality. If Cline's estimate of the direct effects is accurate, then it is plausible that the indirect effects could easily mean that trade is responsible for 50 percent, or more, of the increase in wage inequality in the last two decades. For purposes of this discussion, Krugman's 10 percent estimate will be taken as a lower bound of the impact on wage inequality, while the 50 percent figure, as an upward adjustment of Cline's estimate, will be taken as an upper bound.

The income shifts that need to be explained are the shift from labor in general to capital, and from non-college educated workers to college-educated workers. The first shift led to a 2.6 percent drop in labor compensation, as the capital share of corporate income increased from 17.7 percent in 1979 to 19.8 percent in 1999, the profit peak of the last business cycle.[18] Over the last two decades, the ratio of wages for college educated workers to the wages of workers without college degrees rose from 1.36 in 1979 to 1.67 in 1997 (Mishel, Bernstein, and Schmitt 2001, table 2.42). As a result of this shift, wages for workers without college degrees fell by 4.2 percent from 1979 to 1999, a time in which average hourly compensation rose by 17.9 percent.[19] This means that wages for workers without college degrees would have been on average 23.1 percent higher in 1999, if there had been no increase in wage inequality. If there also had been no shift from labor income to capital income, this would have raised their wages by an additional 2.6 percent. This means that wages for workers with less than college degrees would have been 26.3 percent higher in 1999 if there had been no unfavorable shifts in income distribution over the prior two decades.[20]

The net benefit from trade for the three quarters of the labor force who lack college degrees depends on the extent to which trade liberalization is responsible for growing inequality compared with the extent to which trade has expanded the total amount of income, by increasing growth. The numbers here suggest that there is little doubt that most workers have been losers from trade liberalization over the last two decades. The high-end estimates of the gains from trade liberalization imply that it has increased GDP by 1.0 percent over the last two decades. By contrast, even a low-end estimate of the impact of trade on inequality (10 percent of the total effect) would place the losses at 2.6 percent of wages for workers without college degrees. This means that the net loss due to trade liberalization for workers without college degrees has been 1.6 percent of their wages, as shown in table 1.

            Of course less favorable assumptions (for proponents of trade liberalization) imply larger losses. If trade is directly or indirectly responsible for half of the increase in wage inequality over the last two decades, then it implies that it has cost workers without college degrees an amount equal to 13.1 percent of their current wages. The net loss would still be equal to 12.1 percent of their wages, even assuming a large effect of trade on growth. For a worker earning $25,000 a year, this loss would be slightly over $3,000 per year.   

Table 1—Net Effect of Trade Liberalization on Wages of Workers Without College Degrees

            Given the size of the upward redistribution of income that has actually taken place over the last two decades, if trade is responsible for even a portion of this redistribution—even a portion so small as virtually all economists would acknowledge—the bulk of the work force must have experienced a net loss of income from the trade liberalization that has taken place over this period. The potential gains from liberalized trade are far too small to offset the losses due to greater inequality.

Conclusion--No Reason to Rush Trade Deals

This paper has briefly evaluated the likely benefits from further trade liberalization by examining the estimates from the International Trade Commission's study of the issue. The ITC study estimated that expanded trade would produce gains in the range of 0.1-0.2 percent of GDP. The ITC study, like other theoretical and empirical work in the area, indicates that trade liberalization will lead to an upward redistribution of income, as profits increase by a much higher percentage than do wages. While the ITC study did not examine this issue, there is a large body of research that indicates that trade liberalization will also lead to an upward redistribution of income among workers, from those without college degrees to workers with college and advanced degrees.

Over the last two decades, there has been a well-documented shift in wages, as the three quarters of the labor force without college degrees has seen declining real wages, even as the average wage has continued to grow at a modest pace. There also has been a redistribution from wage income to capital income over this same period. The effect of these two shifts has been to reduce the average hourly wage for workers without college degrees by more than 26 percent. If even a small fraction of this decline is attributable to trade liberalization, for most workers the loss due to increasing wage inequality and the redistribution from wages to profits will vastly outweigh the modest increments to growth resulting from trade. At the high end, these losses could exceed $3,000 annually for a worker earning $25,000 per year.

It is important to realize that there are no obvious losses to delaying trade liberalization. In other words, standard models would not predict any smaller gains from liberalizing trade in five years than at present. The fact that other nations may move ahead with trade liberalization in the meantime should not affect the potential gains to the United States, if it decides to liberalize at some future point. This is worth noting, since there continues to be a great deal of uncertainty around many issues related to trade. The range of estimates of the potential gains from trade is quite wide. Also, the full extent of the impact of trade liberalization on inequality is still not well understood.

In addition, the impact of other aspects of recent trade agreements is barely understood at all. Most noteworthy in this respect is TRIPS, and other efforts to apply U.S.-type patent and copyright laws to developing nations.[21] There has been virtually no research undertaken that tries to quantify the economic losses that developing nations will incur as a result of granting patent and copyright monopolies to producers of pharmaceuticals and other products. Since these forms of protection raise the price of products to people in developing nations by several hundred percent above the competitive market price, there is reason to believe that such costs would be substantial. At the very least, there should be some effort to quantify these costs before pursuing trade agreements that impose such costs on developing nations.

References

Brofenbrenner, K. 2000. “Uneasy Terrain: The Impact of Capital Mobility on Workers, Wages, and Union Organizing.”  Paper submitted to the U.S. Trade Deficit Review Commission.  Ithaca, NY: Cornell University, School of Industrial and Labor Relations.

Cline, W. 1997. Trade and Income Distribution. Washington, D.C.: Institute for International Economics.

Council of Economic Advisors, 1999. "America's Interest in the World Trade Organization: An Economic Assessment." Washington D.C.: The President's Council of Economic Advisors.

Economic Report of the President, 1994. Washington, D.C.: U.S. Government Printing Office.

Frankel, J. 2000. "Assessing the Efficiency Gains from Further Trade Liberalization." Harvard University.

Hufbauer, G. and K. Elliot, 1994. Measuring the Costs of Protection in the United States. Washington, D.C.: Institute for International Economics.

Krugman, P. 1995. "Growing World Trade: Causes and Consequences." Brookings Papers on Economic Activity.

Mishel, L. J. Bernstein, and J. Schmitt, 2001. The State of Working America, 2000-01. Ithaca, NY: Cornell University Press.

Schmitt, J. and L. Mishel, 1996. "Did International Trade Lower Less-Skilled Wages During the 1980s? Standard Trade Theory and Evidence." Washington, D.C.: Economic Policy Institute.

Stoeckel, A., D Pearce, and G. Banks, 1990. Western Trade Blocs. Game, Set or Match For Asia-Pacific and the World Economy? Sydney, Australia: Centre for International Economics.

Stolper, Wolfgang and Paul A. Samuelson. 1941.  Protection and Real Wages. Review of Economic Studies 9, no.1 (November): 58-73

United States International Trade Commission, 1999.  The Economic Effects of Significant U.S. Import Restraints.  www.usitc.gov/wais/reports/arc/w3201.htm. 

United States Trade Representative, 2001. Response to Freedom of Information Act Request from Patrick Woodall, Washington, D.C.: Public Citizen's Global Trade Watch.