Introduction
Congress plays a major role in formulating and implementing U.S. trade policy through its Constitutional role in regulating foreign commerce. This role includes providing authority to the President to conclude trade agreements to eliminate and reduce tariff and non-tariff barriers to U.S. trade with other countries. For Members of Congress, trade can be a difficult issue, because decisions that liberalize trade flows have mixed effects on U.S. domestic and foreign interests, both economic and political. Historically, Congress has supported policies to open international markets to U.S. goods, services, and agriculture. It also has weighed carefully the economic impact of trade agreements to asses their benefits and costs. Congress also has enacted targeted adjustment assistance programs to provide training and other assistance to help workers and firms that may be dislocated by greater market opening adjust to new economic opportunities.
Anticipating congressional consideration of a free trade agreement between the United States and South Korea (KORUS FTA), a number of groups offered estimates to quantify potential effects of the FTA on U.S. employment.1 This report discusses the models used to estimate the employment effects of the agreement and their various assumptions, necessary in order for the model to generate results. Invariably, these assumptions determine to some extent the results that are generated and, therefore, limit their representation of the real world economy. The models also are highly sensitive to assumptions used to establish the parameters of the models and are hampered by a lack of data in some areas.
In general, the various economic models provide differing estimates of the magnitude of changes in U.S. employment that could arise from a trade agreement with South Korea, reflecting different assumptions and conditions. Both advocates and opponents of the KORUS FTA cite results of these studies to support their position. Each of these models has strengths and weaknesses and varies in the degree to which it reflects economic reality. These models are discussed at greater length in the analyses that follow. In addition, the potential job effects need to be viewed in terms of their relative share of U.S. employment.
Estimates from the studies range from as many as 280,000 jobs potentially gained to 159,000 jobs potentially lost over the ten-year implementation period of the KORUS-FTA. Middle-range estimates include one often cited by the Obama Administration, which suggests that the KORUS FTA would support 70,000 export jobs. At the other end of the spectrum, other estimates suggested that the United States could lose 345,017 jobs over ten years if the United States did not approve and implement the KORUS-FTA but the European Union (EU) and Canada implemented their FTAs with South Korea. None of these estimates, which represent gains or losses in employment over a ten-year period, account for more than 1.4% of total goods-producing employment, or 0.26% of total employment in the U.S. economy in 2009.
None of these studies, however, draws a direct link between any expected changes in trade from a change in tariffs and subsequent changes in employment. These estimates reflect different methodologies and assumptions, and a partial accounting of the total economic effects of such agreements. For instance, the agreement includes provisions on trade in goods, services, and investment. Due to the complexities involved and a lack of data, nearly all estimates focus only on changes in goods-producing employment. As a result, they do not serve well as an indicator of the overall impact of the agreement on the economy. The estimates selected for analysis are the following.
- U.S. International Trade Commission (USITC): U.S-Korea Free Trade Agreement Potential Economy-wide and Selected Sectoral Effects2
- The University of Michigan: Economic Effects of a Korea-U.S. Free Trade Agreement3
- U.S.-Korea Business Council: Failure to Implement the U.S.-Korea Free Trade Agreement: The Cost for American Workers and Companies4
- Economic Policy Institute (EPI): Trade Policy and Job Loss: U.S. Trade Deals with Colombia and Korea Will be Costly5
In addition, two other estimates, derived by the USITC and released by the Obama Administration and the Majority Staff of the Senate Finance Committee Trade Subcommittee, are reported.
Congress has demonstrated an ongoing interest in concluding trade agreements to eliminate and reduce barriers to U.S. trade in goods, services, and agriculture, and it has assumed a direct role in assessing the impact of trade agreements on the U.S. economy. This report provides background and context for the analyses of the various estimates. It summarizes congressional interest in the issue and then examines the specific estimates—their findings, assumptions, methodologies, and the limitations of those methodologies. Finally, it explains the broader macroeconomic and microeconomic context in which the composition of employment within the economy is affected by a new trade agreement.
Estimating Employment Effects of Trade Agreements
Although discussions of trade agreements often focus on potential employment gains or losses, most economists argue that such employment estimates represent a partial accounting of the total economic effects of trade agreements and, therefore, do not perform well as an indicator of the overall impact of the agreement on the economy. As this report attempts to explain, estimating such employment effects is imprecise and highly sensitive to the assumptions that are used.6 In addition, while trade agreements generally are comprehensive in nature and cover goods, services, and investment, most employment estimates focus narrowly on the goods sector and do not adequately represent the total impact of the agreements. It is difficult to estimate precisely the employment effects associated with liberalizing trade in services and reducing or eliminating barriers to investment flows. Trade in services, in particular, is characterized by a broad array of formal and informal barriers that challenge efforts to translate the barriers into tariff-equivalent values. Negotiations to reduce barriers to trade in services, however, potentially could have a very large and positive effect on the U.S. economy, since the United States is highly competitive in a number of services sectors and U.S. direct investment abroad often spurs exports.
Finally, estimates of employment arising from FTAs, by themselves, do not account for a broad range of benefits for the economy as a whole. For example, FTAs may provide individual consumers and firms with broader economic benefits and yield broad productivity and efficiency gains for the economy over the long run that may enhance employment. While most economic trade models attempt to model consumer and corporate behavior associated with changes in income, the results of the models depend on a number of factors, including the way changes in prices are passed along to consumers. Most estimates generally assume that the changes in tariffs and, therefore, changes in the prices of goods, will be adopted at the time the agreement is signed and then the annual changes in traded goods are aggregated over ten years. In fact, in the KORUS-FTA, some tariffs will drop to zero immediately, while others will be implemented in tariff reduction stages over ten years, with the impact on prices accruing over time. In the models, expected changes in the trade in services generally are treated as exogenous factors and must be specified outside the model itself.
A key assumption used in generating a number of the estimates of the impact of the KORUS-FTA is the level of employment in the U.S. economy during the ten-year period required to phase in the agreement. None of these estimates, however, generate precise estimates of the exact number of jobs to be gained or lost, but generate estimates that reflect the relative magnitude of changes in employment that could be expected over ten years, provided that all other things remain constant. One approach assumes that the U.S. economy operates at a level of full employment over the next decade and holds the level of employment in the economy stable while wage rates are allowed to adjust. A second approach, and one that has been used extensively to derive many of the estimates evaluated in this report, assumes that the economy is marked by a rate of unemployment that reflects current economic conditions, or an economy that is operating at less than full employment. This assumption, more than any other, seems to be the most important factor generating estimates of employment effects under the KORUS-FTA. While the U.S. economy currently is operating at less than full employment, it seems reasonable to assume that over the next ten years the economy will return to its long-term trend that is closer to full employment as the economy recovers from the most severe economic recession in the post-World War II period.
Trade models of the type used in the analysis of free trade agreements are part of a class of economic models referred to as computable general equilibrium models (CGE) that incorporate data on trade and a range of domestic economic variables on nearly 100 countries. As a result of this large number of countries, and the vast amounts of data that are a part of the model, the models can provide important insights into the mechanisms by which changes in tariffs or other parameters can affect a range of countries. For practical reasons, however, the data in the models must be limited, so the models necessarily must sacrifice some level of precision in their estimating abilities. Since such trade models originally were developed with the intent of analyzing the economic effects of multi-country trade agreements such as the Uruguay Round, this lack of precision was not considered to be an important drawback. However, this lack of precision may be an issue when the models are used to estimate the effects of bilateral trade agreements where the overall amount of trade and, therefore, the impact of the agreement, is expected to be less than that of a comprehensive multilateral agreement. In addition, such models do not account for changes in exchange rates, since such effects were considered to be neutral in a large multi-country trade agreement. Movements in exchange rates, however, could have an important impact on trade patterns that involve countries that are parties to a bilateral trade agreement.
With floating exchange rates, movements in those rates can have an important effect on the prices of internationally-traded goods. Figure 1 shows the movement of the internationally trade value of the won relative to the dollar and the Euro and movements in the dollar-Euro relationship, indexed to January 2005. In this presentation, a rise in the index indicates a depreciation in the value of the won relative to the dollar and the euro and the depreciation of the dollar relative to the euro between January 2006 and January 2011. Between January 2008 and January 2009, the won depreciated about 44% against the dollar and 30% against the euro largely due to a flight to the dollar during the financial crisis, a sharp drop in capital inflows to South Korea, and the largest annual drop in South Korean exports on record. After recovering somewhat between January 2009 and mid-2010, the won experienced another bout of depreciation relative to the dollar and the euro as international markets became more cautious as a result of financial problems in Greece, Ireland and other European countries and concerns over political developments in neighboring North Korea. By February 2011, the won remained 19% and 10% below its value in January 2008 relative to the dollar and the euro, respectively. Such changes in the value of the won relative to the dollar and the euro would have a significant impact on the prices of goods to and from South Korea and possibly eclipse the relatively small changes in tariff rates included under the respective FTAs.
|
Figure 1. Value of the Korean Won Relative to the Dollar and the Euro |
Source: Developed by CRS from data published by the Federal Reserve and the European Central Bank. |
Since the KORUS FTA was signed in 2007, a number of estimates of the potential effects of the KORUS FTA have been released that attempt to estimate the effect of the trade agreement on U.S. employment. The estimates range from 280,000 jobs gained to a loss of 159,000 jobs if the United States does adopt the agreement. Two of the estimates are based on standard economic modeling, two others were developed by the USITC, and two represent estimates derived from unique approaches. Table 1 summarizes all six estimates. These estimates will be discussed in greater detail below. The two standard model estimates were undertaken by the USITC and the University of Michigan. Neither of these efforts estimated the total number of U.S. jobs created or lost by the KORUS FTA.7 but focused instead on potential economy-wide trade effects and sectoral shares of those effects. In contrast, two other approaches by the U.S.-Korea Business Council and the Economic Policy Institute, attempted to estimate employment effects of the KORUS FTA indirectly.
Table 1. Estimates on the Employment Effects of KORUS FTA
|
Study/Estimate |
Estimated Trade or Employment Gain or Loss over Total Implementation Period (about 10 years) |
Employment Estimate in Perspective |
|
USITC: U.S.-Korea Free Trade Agreement Potential Economy-Wide and Selected Sectoral Effects, 2007 (for the goods sector only). |
No estimate. Study assumed full employment, so it projected only share shifts among goods-producing industries. Study estimates net export gains of $3.3-$4.0 billion in the goods-producing sectors. |
No employment estimates were made. |
|
University of Michigan: Economic Effects of a Korea-U.S. Free Trade Agreement, 2007 (for the goods and services sector). |
Like the USITC study, this study assumes full employment and therefore estimates share shifts among industries. However, it also includes jobs data by industry for 2006, so it applies estimated share changes to actual jobs data, to estimate job gains and losses by industry. In addition, this study covers services as well as goods. |
85% of the projected job gains are in agricultural industries; 90% of the projected job losses are in six industry sectors: government services, trade and transportation services, and manufacturing production for textiles, apparel, transportation equipment, and metal products. |
|
The White House: Remarks by the President at the Announcement of a U.S.-Korea Free Trade Agreement, 2011 (for the goods sector only). |
Estimates, based on USITC projections of goods trade, that the agreement could support at least 70,000 U.S. export jobs. |
This estimate is equal to 0.4% of total manufactured goods employment in 2010. |
|
Majority Staff of the Senate Trade Subcommittee of the Finance Committee: Updated Assessment of the U.S.-Korea Free Trade Agreement (FTA), 2011 (for the goods sector only). |
Estimates reflect recalculations using the USITC model, at the request of the Subcommittee, to reflect 2009 trade and economic conditions. From this recalculation, the Subcommittee estimates that the agreement has the potential to create 280,000 U.S. goods-producing jobs. |
This estimate is equal to 1.4% of total manufactured goods-producing jobs in the United States in 2010. |
|
U.S.-Korea Business Council: Failure to Implement the U.S.-Korea Free Trade Agreement: The Cost for American Workers and Companies, 2009 (for the goods and services sector). |
Estimates that if European Union and Canada implement their FTAs with Korea and the United States does not, the United States could lose 345,017 goods and services jobs from U.S. trade diversion to other countries. |
Estimated jobs lost are equal to 0.3% of total goods and services jobs supporting exports for 2010. |
|
Economic Policy Institute (EPI): Trade Policy and Jobs Loss: U.S. Trade Deals with Colombia and Korea Will be Costly, 2010 (for the goods sector only). |
Projects that 159,000 U.S. goods-producing jobs could be lost as a result of increased trade deficits with South Korea. |
Estimate represents 0.8% of all goods producing jobs in 2010. |
Trade and Employment in the U.S. Economy
In a dynamic economy such as the United States, jobs are constantly being created and replaced as some economic activities expand, while others contract, reflecting broad macroeconomic developments. In this process, various industries and sectors evolve over time at different speeds, reflecting differences in technological advancement, productivity, and efficiency. Those sectors that are the most successful in developing or incorporating new technological advancements generate greater economic rewards and are capable of attracting greater amounts of capital and labor. In contrast, those sectors or individual firms that lag behind are less capable of attracting capital and labor and confront ever-increasing competitive challenges. Indeed, depending on the overall state of the economy, some sectors may need to relinquish some capital and labor in order for others sectors to grow to avoid economic stagnation. Also, advances in communications and technology have facilitated a global transformation of economic production into sophisticated supply chains that span national borders and defy traditional concepts of trade that potentially could involve a greater share of the labor force in trade-related activities.8 How firms respond to these challenges likely will determine their long-term viability in the market place.
At the plant level, job openings may come from new business openings or from expansions at existing facilities, including those that are used to support increased exports. Job losses may come from voluntary departures, involuntary discharges, or from business closures for any reason, including bankruptcy, personal choice, inability to compete in the domestic market, import competition, or production shifts. The Bureau of Labor Statistics, Business Employment Dynamics (BED) Report tracks gross jobs gained and lost in the economy as a whole and in specific sectors and reveals the dynamics underlying gross changes in employment. As Appendix A shows, 15% of total U.S. employment in 2010 was in the goods-producing industries, while the services-producing industries accounted for 68% of the employed population. The remaining 17% of employment was in the government sector, with employment at the State and Local government level accounting for 15% and the federal government sector accounting for 2%
Jobs Gained and Lost for the Economy as a Whole
Table 2 shows the gross number of jobs gained and lost for the economy as a whole for the goods- and services-producing sectors, and for the number of jobs related to exports, during the period 2006-2009. This process of job turn-over typically affects 18-20% of the jobs in the total economy each year (when percentages of gross jobs gained and gross jobs lost are combined). This process is stronger in the goods-producing sector of the economy, where a number equivalent to 24%-27% of the jobs are typically gained or lost each year, compared with the services sector, where the comparable share is 17%-18%. Such job turn-over is amplified by economic expansions or recessions, as was the case in 2008-2009, when the economy experienced the most severe recession in the post-World War II period.9
The data in Table 2 also indicate that
- 1. Jobs supported by exports represent a small share of the total number of jobs in the economy and are equivalent to about one-third of the total number of jobs gained or lost in the economy as a whole during a year. Jobs supported by exports were equivalent to a small share of the annual turn-over in jobs, ranging from 6% to 8% of the annual turnover of jobs in the economy between 2006 and 2009.
- 2. Within the goods-producing sector, however, such export-related jobs were equivalent to 29% to 35% of total jobs, and within the services sector they averaged about 2%. Similar to overall job gains and losses for the total economy, jobs supporting exports are sensitive to the business cycle.
- 3. Business cycle effects were particularly apparent between 2008 and 2009, when the global economic recession and financial crisis sharply reduced global trade. As a result, jobs in the economy supported by trade fell by about 1.8 million jobs, or from 8% of total employment to 6% of total employment, or by about one-fourth of its share of total employment.
Some would argue that the data on jobs gained and lost throughout the whole economy do not adequately reflect the concern of workers about job losses. According to data published by the Bureau of Labor Statistics, during the last two years of the most recent economic expansion (2005-2007), 3.6 million workers were displaced from jobs they had held for at least three years. Of those workers, by the end of January, 2008, only 67% had been re-employed in full-time jobs and only 22% were in jobs where their salaries matched their previous income level.10
Table 2. Jobs Gained and Lost Annually in the U.S. Economy and Job Turnover Rate, 2006-2009
(Percentages refer to shares of total jobs in that respective sector)
|
2006 |
2007 |
2008 |
2009 |
||||||
|
Sector |
Jobs Measure |
No. of Jobs (000) |
% of Total Jobs in Sector |
No. of Jobs (000) |
% of Total Jobs in Sector |
No. of Jobs (000) |
% of Total Jobs in Sector |
No. of Jobs (000) |
% of Total Jobs in Sector |
|
Total Economy |
total jobs |
136,086 |
100% |
137,598 |
100% |
137,066 |
100% |
131,997 |
100% |
|
gross jobs gained |
14,019 |
10% |
13,441 |
10% |
12,704 |
9% |
10,048 |
8% |
|
|
gross jobs lost |
11,438 |
8% |
11,941 |
9% |
12,609 |
9% |
15,912 |
12% |
|
|
jobs supported by exports |
8,950 |
7% |
9,540 |
7% |
10,293 |
8% |
8,500 |
6% |
|
|
Goods- Producing Sector |
total goods-producing jobs |
22,531 |
100% |
22,233 |
100% |
21,419 |
100% |
18,938 |
100% |
|
gross jobs gained |
2,961 |
13% |
2,677 |
12% |
2,049 |
11% |
1,618 |
9% |
|
|
gross jobs lost |
2,512 |
11% |
2,891 |
13% |
3,041 |
14% |
4,238 |
22% |
|
|
jobs supported by exports |
6,576 |
29% |
6,920 |
31% |
7,525 |
35% |
6,000 |
32% |
|
|
Services- Producing Sector |
total service-producing jobs |
113,556 |
100% |
115,366 |
100% |
115,646 |
100% |
113,059 |
100% |
|
gross jobs gained |
11,059 |
10% |
10,763 |
9% |
10,296 |
9% |
8,430 |
7% |
|
|
gross jobs lost |
8,928 |
8% |
9,048 |
8% |
9,571 |
8% |
11,672 |
10% |
|
|
jobs supported by exports |
2,374 |
2% |
2,619 |
2% |
2,768 |
2% |
2,500 |
2% |
|
Source: BLS Business Employment Dynamics (BED) Reports, various years for job gains and losses, and Bureau of Labor Statistics (BLS): Employment and Earnings Table B-3 for total jobs in the U.S. economy, by sector.
Note: The "% of total jobs" column in shows the "churning" rate, or total employee turnover rate.
Benefits and Costs of International Trade
In the current highly globalized economy, international trade has come to represent a complex set of transactions. Nations not only trade goods and services, but they also trade a broad range of financial products. In addition, liberalized capital flows and floating exchange rates have greatly expanded the amount of capital that flows between countries. Basically, trade represents an exchange of goods or services between two or more willing parties. Such trade allows nations to use their resources more efficiently in order to maximize the total amount of goods and services that are available to their citizens, a common definition of a nation's standard of living. As a result of this maximization process, nations trade because it serves their national interests. In the same way that individuals gain by specializing in activities that use their strongest skills and then trade with others, nations specialize in the production of certain goods and then trade with other nations for the goods they do not produce. Essentially, nations export in order to import goods and services they do not produce, or cannot produce efficiently.
Benefits of Trade
Both the benefits and the costs of trade have become well-publicized. These benefits are categorized as one-time, or static, benefits, which include gains for consumers and gains for producers, and dynamic benefits that accrue over time and can positively affect the long-term rate of growth of a country. While it is not always possible to measure these effects precisely, most economists believe that the net effect of international trade on the national economy as a whole is positive, that is, that the total gains exceed the total costs. By reducing foreign barriers to U.S. exports and services and by removing U.S. barriers to foreign goods and services, trade liberalization helps to strengthen those industries that are the most competitive and productive and to reinforce shifts in labor and capital from less productive endeavors to more productive economic activities.
Impact of Trade on the Composition of Employment in the Economy
An important issue in estimating employment effects from increased trade and trade agreements is the difference between the microeconomic viewpoint and the macroeconomic viewpoint. In a dynamic economy such as the U.S. economy, the composition of jobs is constantly changing as some sectors grow and other sectors decline. This constant churning would continue even in the complete absence of international trade. International trade adds to the myriad of factors that determine the composition of output and jobs within the economy. From the microeconomic perspective, or the viewpoint of the individual firm, competitive pressures from international trade, as well as a broad range of other factors, determine the viability of individual plants or firms. Given these competitive pressures, firms face a number of choices. How such firms choose to respond to the competitive pressures likely determines the overall viability of the firm in the marketplace. As a result of these actions, some firms expand, while others contract.
In contrast, the macroeconomic viewpoint focuses on the net overall direct and indirect economic effects of trade. Direct effects include changes in the composition of employment and production for the economy as a whole. Indirect effects can include secondary and tertiary effects in industrial sectors that may be more difficult to estimate. It is assumed that trade agreements affect production and consumption as a result of the impact of the agreements on the prices of goods. Presumably, trade agreements lower the tariffs on imported goods, which result in lower prices on imported goods and shift domestic production in ways that make the economy more efficient. Consumers benefit directly from the lower prices of imported goods to the extent that they shift their consumption in favor of the lower-priced imported goods. They also benefit from an increase in their real incomes as a result of the lower prices of imported goods. Consumers may benefit further from lower domestic prices to the extent that domestic producers lower their prices in response to the competition from imports.
Throughout this process, however, the total number of jobs in the U.S. economy is not affected, that is, trade agreements do not affect the total number of jobs in a large open economy such as the U.S. economy, but can affect the composition of employment. This result becomes more binding as the economy approaches full employment. Since the KORUS-FTA is expected to be implemented over a 10-year period, it seems reasonable to assume that the U.S. economy will revert to its long-term trend toward full employment, rather than remain at the current level with underutilized resources that reflect the 2008-2009 recession. The total number of jobs in the U.S. economy is determined by such macroeconomic factors as productivity growth, the growth rate of the population, and the pace of technological innovation. Changes in any of these "structural" factors, together with short-term fluctuations in the business cycle and shifts in the relative value of the dollar against other currencies, may overwhelm effects of any industry-wide job "gains" or "losses"11 from trade and trade agreements.
As the U.S. economy shifts the composition of production from labor-intensive, import-competing products to capital-intensive, export-oriented products, more labor than capital is released. Because those displaced from labor intensive jobs may not have the skills to immediately become employed in more capital intensive jobs, labor dislocations in the economy may result.
Adjustment Costs of Trade
Economists have long recognized, however, that the long-term production gains associated with greater specialization in the economy create a wide range of adjustment costs as labor and capital are shifted from less efficient industries and activities into more efficient industries and activities. These adjustment costs are difficult to measure, but they are potentially large over the short run and can entail significant dislocations for some segments of the labor force, for some companies, and for some communities.12 In negotiating trade agreements, governments are most mindful of the adjustment costs involved and, at times, are constrained in their ability to fashion such agreements because of opposition by groups within the economy that would bear heavy costs from trade liberalization. These costs are especially acute for labor groups within the economy that lack advanced education and training skills that provide them with the means necessary to be redeployed in other sectors of the economy.
Issues Involved in Quantifying the Employment Effects of Trade
Quantifying the relationship between international trade and the composition of employment in the economy is problematical and complex. For the United States, international trade is not the primary force creating employment in the economy. While trade agreements with specific countries may well benefit certain sectors of the economy, their effect is dubious as an employment program for the economy as a whole. In addition, changes in exchange rates and in the business cycle can affect the overall state of the economy in ways that can outweigh the effects of trade agreements, given the already highly open state of the U.S. economy. In addition, significant gaps in data, particularly relative to the services sector, hinder the ability to model the effects of trade agreements that lower barriers to trade in services. These gaps are important, because the services sector accounts for 69% of output and 68% of full- and part-time employment in the U.S. economy and increased trade in services offer potentially large gains for the U.S. economy.
In addition, concerns over the impact of a trade agreement on employment often focus on a comparison of labor compensation rates between countries. For instance, some groups in the United States are concerned that U.S. employment could be negatively affected because labor compensation rates, in general, are lower in South Korea. Measures of competitiveness, however, reflect not only the rate of labor compensation, but the rate of compensation relative to the level of productivity. Rates of labor compensation in the United States could be many multiples of that in South Korea, but as long as U.S. workers remain more productive than workers elsewhere, U.S. goods would continue to be competitive in the global marketplace. In general, most estimates of employment effects do not incorporate measures of productivity in their analysis.
Various approaches typically have been used to estimate the employment effects of trade agreements, including the KORUS FTA. In most cases, these approaches share some common features. Similar to other free trade agreements, the KORUS FTA is a comprehensive agreement that includes provisions to: (1) lower tariffs and non-tariff barriers on trade in goods and agriculture; (2) reduce barriers to trade in services; and (3) encourage increased flows of investment. Formal and informal barriers to trade in services, however, are extremely difficult to quantify in monetary terms, and it is equally difficult to estimate the impact of an agreement on potential flows of investment. As a result, most approaches derive their estimate of changes in employment almost exclusively from estimated changes in the value of exports and imports of goods under the agreement. Even these estimates, however, are imprecise and sensitive to the assumptions that are used. Appendix B explains some of the most common assumptions used in economic models.
Estimates of Employment Effects of the KORUS FTA
The USITC Estimate
The USITC13 used a standard approach and analysis by addressing the issue of how U.S. exports to and imports from South Korea would be affected by the FTA with South Korea that lowered tariffs on a set of traded goods. This approach attempted to measure the long-term trade (but not employment) effects of a one-time full implementation of the KORUS FTA after the full implementation period of 10 years on U.S. exports and imports. The ITC used an economic model known as the Global Trade Atlas Project (GTAP), located at Purdue University14 to estimate changes in trade (exports and imports) from changes in tariff rates and tariff rate quotas. The results are expressed as proportional effects (percentage increases or decreases in trade) for various sectors, relative to the projected 2008 economy. According to this estimate, if the KORUS FTA were to go into effect, U.S. merchandise exports to South Korea likely would increase by an estimated $9.7 billion to $10.9 billion over the first decade after enactment, while merchandise imports from South Korea likely would increase by an estimated $6.4 billion to $6.9 billion over the first decade. This would result in an estimated net increase in U.S. exports of $3.3 billion to $4.0 billion during the first decade after enactment.
The University of Michigan Estimate
In a slightly different approach, economists at the University of Michigan used a similar world trade model to estimate the effect of a negotiated reduction in tariff rates on export and import prices and then estimated changes in the volumes of goods exported and imported arising from those change in prices, as indicated in Table 3. They estimated, using 2006 data, that U.S. bilateral exports to Korea would increase by $9.2 billion and that U.S. imports from Korea would increase by $6.9 billion during the first ten years after implementation of the agreement.15 The total number of jobs gained or lost in the economy is assumed to be zero, because the model begins with the assumption of full employment. Given this constraint of no changes in the overall number of jobs gained or lost, the estimates represent expected changes in the composition of employment among sectors in the economy. As a result, the estimates should not be viewed as projections of the exact number of jobs gained or lost by sector, but as estimates of the magnitude of the changes that could occur. Given this caveat, the estimates indicate that 85% of the projected job gains would be in the agricultural sector, and 90% of projected job losses would be in six industrial sectors: government services, trade and transportation services, manufactured textiles and apparel, transportation equipment, metal products, and machinery equipment.
Standard economic theory provides some insight into which sectors may be expected to gain, and which sectors may be expected to lose employment as a result of increased trade. Accordingly, those sectors that represent areas of greatest comparative advantage16 based on the economy's endowment of the factors of production (land, labor, capital, technology, and entrepreneurial ability), are the ones most likely to gain employment. Compared with other countries, the United States has a comparative advantage in capital-intensive products, many services, and agricultural products that characterize U.S. exports. This is in contrast to most developing countries, that have a comparative advantage in more labor-intensive products that characterize U.S. imports.
Table 3. University of Michigan Estimate of Potential Sectoral Employment Effects of the Korea-U.S. Free Trade Agreement
|
Industry |
Projected Change in Employment Relative to 2006 level |
|
|
Change in Number of Workers |
Change in Share of Total Industry Employment (%) |
|
|
Rice |
460 |
1.1 |
|
Wheat |
388 |
0.2 |
|
Other grains |
7,446 |
1.6 |
|
Vegetables and fruits |
1,064 |
0.2 |
|
Oil seeds |
7,151 |
1.7 |
|
Sugar |
67 |
0.1 |
|
Plant-based fibers |
63 |
0.0 |
|
Other crops |
3,390 |
0.3 |
|
Livestock |
1,092 |
0.1 |
|
Mining |
-363 |
-0.0 |
|
Food, beverages, and tobacco |
1,880 |
0.1 |
|
Textiles |
-4,426 |
-0.5 |
|
Wearing Apparel |
-3,482 |
-0.6 |
|
Leather products and footwear |
-171 |
-0.2 |
|
Wood and wood products |
-483 |
-0.0 |
|
Chemicals |
119 |
0.0 |
|
Nonmetallic mineral products |
289 |
0.0 |
|
Metal products |
-1,077 |
-0.0 |
|
Transportation equipment |
-2,287 |
-0.1 |
|
Machinery and equipment |
1,438 |
0.0 |
|
Other manufactures |
-186 |
-0.0 |
|
Construction |
-68 |
-0.0 |
|
Electricity, gas, and wager |
-639 |
-0.0 |
|
Trade and transport |
-5,379 |
-0.0 |
|
Other private services |
-569 |
-0.0 |
|
Government services |
-5,714 |
-0.0 |
|
TOTAL |
0 |
-0.0 |
Source: Economic Effects of a Korea-U.S. Free Trade Agreement, by Kozo Kiyota and Robert M. Stern, University of Michigan, April 2007. Source of data for table is ILO (2006), UNIDO (2006) and World Bank (2006).
Note: Numbers of workers result from estimated share changes in trade projected onto 2006 employment data.
Associated Estimates
Two additional estimates of the employment effects of the KORUS FTA are based on the data developed by the ITC on the number of U.S. jobs supported by exports.
In one estimate, the White House apparently used the ITC data to estimate that an additional 70,000 U.S. workers would be supported by new exports under the KORUS FTA. This estimate appears to be based on data developed by the ITC regarding the amount of additional exports that would be created by lowering tariff rates, combined with the data on the number of U.S. jobs supported by exports. As indicated below, the ITA has stated that the data developed on the average number of jobs in the economy that are supported by exports should "not be used to estimate the net change in employment that might be supported by increases or decreases in total exports, in the exports of selected products, or in the exports to selected countries or regions." Some could argue, however, that the White House estimate honored this caution by stating only that the proposed KORUS FTA "would support" (not create) 70,000 U.S. jobs.17
Another estimate developed at the request of the Majority Staff of the Senate Trade Subcommittee by the ITC using the GTAP model indicated that the KORUS FTA has the potential to create about 280,000 American jobs. The Subcommittee asked the USITC staff to "examine the agreement based on current data and economic conditions." The Trade Subcommittee staff modeled two scenarios with differing assumptions. In the first scenario, the staff assumed that labor and capital in the economy were held fixed, representing an economy close to full employment and full capacity utilization. Since the scenario assumes that labor is fixed, there would be no net change in employment, but labor and capital from other industries would need to be reallocated from other sectors as a result of the KORUS-FTA. This scenario estimates slightly lower percentage increases in output and employment as a result of the FTA, since it assumes that the economy is fully employed.
In the second scenario, it is assumed that wages and capital rentals are held constant, but that there is underutilized labor and capital, similar to an economy that is operating at less than full employment and that it has excess capacity. The results of this scenario indicated that, "in an economy with substantial unemployment and underused capital, the agreement has the potential to expand employment by 0.16%, or approximately 280,000 American jobs."18 In this scenario, labor and capital would be drawn from sectors with unemployed labor and capital and from other sectors. In both scenarios, U.S. exports to South Korea would increase by 4.4% and imports from South Korea would increase by 1.3%.
The differences between the two scenarios offer a clear example of the importance of the assumptions that are used to generate estimates of the KORUS-FTA from the economic models. In this case, the importance of the assumption that the U.S. economy will operate at less than full employment during the ten-year period of full implementation drives the estimate of employment gains. It seems reasonable to assume, however, that during the next ten years the U.S. economy will return to its long-term trend of operating closer to full employment than to remain at the current levels of underutilization as the economy recovers from the most severe recession in the post-World War II period.
Additional Estimates
U.S.-Korea Business Council Estimate19
The U.S.-Korea Business Council estimated the employment effects of the KORUS-FTA using a slightly different approach. This approach does not attempt to model the impact of the KORUS-FTA on the U.S. economy, but addresses a different, but related issue: What would be the impact on U.S. trade with South Korea if the proposed FTAs between South Korea and the Europe Union and Canada were implemented, but the South Korea-U.S. FTA were not implemented? This question was rendered moot when Congress passed the KORUS FTA.
Nevertheless, according to the Council, they also used the GTAP model to derive their result. The Council's approach involved a number of steps and assumptions. Most importantly, the Council's approach assumed that there were underutilized capital and labor in the economy. This assumption is consistent with an assumption that the U.S. economy operates at less than full employment during the duration of the ten-year phase-in period. The Council also assumed that the Canadian and EU FTAs with South Korea would result in a 5% reduction in barriers to trade in services and that there would be a reduction in trade facilitation costs.
The Council's first scenario modeled the impact on U.S. output and trade if the three FTAs involving Canada, the EU, and the United States with South Korea were fully implemented. Next, the second scenario modeled the impact on U.S. output and trade if the Canadian and EU FTAs with South Korea were fully implemented, but the U.S.-South Korea FTA were not implemented. Next, the council derived the difference in the impact on U.S. output and trade by subtracting the data from the first scenario from the second scenario. As a result of this process, the Council estimated that the United States would experience a potential loss of 345,000 goods and services export jobs if the KORUS FTA is not implemented, but the EU and Canadian FTAs with South Korea are implemented. The estimate assumes that exports to South Korea primarily from Canada and the EU would be substituted for a certain share of U.S. exports of goods and services, in an amount equal to $35.1 billion.
This scenario has a number of strengths and weaknesses, although the final estimate appears to be within the same general level of reliability as the other estimates that are analyzed in this report. The attempt to model the impact of not implementing the KORUS-FTA is a departure from the other approaches and highlights the fact that a large number of FTAs are being negotiated by a broad range of countries with potentially far-reaching impact on the patterns of international trade. Since the approach used by the Council does not estimate directly the negative impact on U.S. output and employment of not implementing the KORUS-FTA, this approach can introduce potential distortions into the estimates. For instance, trade agreements involve both trade diversion, or shifting trade among competing countries, and trade creation, or new trade possibilities as a result of changes in trading relationships. Although the model does not capture the possible trade-creating effects of the FTAs, which would offset some of the trade diversion effects, the trade-creating effects likely would be small relative to the trade diversion effects. This estimate is the only one of the current estimates that imposes an assumption of the net impact of a change in the trade in services.
Economic Policy Institute Estimate
The Economic Policy Institute estimated that the KORUS FTA would result in a loss of 159,000 U.S. goods-producing jobs stemming from a projected increase in the trade deficit with South Korea. This estimate was derived using an unconventional approach. The EPI did not derive its estimate directly from estimated changes in trade or the potential employment effects from a free trade agreement between South Korea and the United States. Instead, the EPI derived its estimate by averaging the changes that occurred in U.S. trade with Mexico between two different periods: the two seven-year periods immediately before and after the North American Free Trade Agreement (NAFTA) went into effect in 1993; and the two seven-year periods immediately before and after China joined the World Trade Organization in 2001. It then used the average rate of growth in exports and imports between these periods to estimate percentage changes in U.S. trade with South Korea under the proposed KORUS FTA. This estimate suggested that the KORUS FTA would result in a net increase in imports (a trade deficit) rather than net increase in exports. Next, the EPI study used an estimate of jobs associated with trade to estimate the number of jobs that would be lost in the United States as a result of a projected increase in the U.S. trade deficit with Mexico. The EPI approach made a number of assumptions, including
- 1. Trade proxies. This approach uses a non-trade-weighted average of U.S. trade with Mexico and China as a proxy for estimated U.S. trade with South Korea with no clear explanation of why these two cases match the Korea case. Data on U.S. trade with South Korea are readily available, so it is not clear why these data are not used to estimate potential changes. Also, U.S. trade with South Korea is about one-fourth that of U.S. trade with Mexico, and the United States and South Korea do not share a common border as is the case between the United States and Mexico.
- 2. Exchange rates. There are numerous differences between the movement in exchange rates among the three countries considered in this analysis: China, Mexico, and South Korea. For Mexico, the second seven-year period following adoption of NAFTA coincided with the devaluation of the peso. This boosted U.S. imports from Mexico and likely contributed to a decline in U.S. exports there. Unlike the peso, China's currency is linked primarily to the value of the U.S. dollar. The South Korean won is determined mainly by market forces.
- 3. Trade and jobs. The EPI estimate also uses some form of an estimate of jobs related to trade similar to that developed by the ITA on exports supporting U.S. jobs to estimate the number of U.S. jobs that would be lost as a result of a projected increase in the U.S. trade deficit with South Korea. As previously indicated, the ITA has warned that this approach to estimating jobs gained or lost from trade is not reliable.
Additional Problems Involved in Linking Changes in Production with Changes in Employment
In a joint project, the Department of Commerce (DOC), Bureau of Economic Analysis (BEA), and the Bureau of Labor Statistics (BLS), estimated the average number of jobs supported by exports in the U.S. economy based on the dollar value of output relative to the average number of jobs required to produce that output for each industry.20 As a result of this joint effort, the Commerce Department determined in its 2010 update that on average $166,000 in goods exports, $216,000 in services exports, or a weighted average of $180,000, supported one job in each respective sector.21
At times, some estimates of the employment effects of FTAs have been derived from this DOC report on employment in the economy supported by exports. In some cases, the data have been used in reverse to argue that if a certain number of jobs were supported by a billion dollars of exports, then that same number could be used to argue that a certain number of jobs would be "lost" by a billion dollars of imports, so that any net increase in imports associated with a trade agreement would necessarily result in a loss of employment for the economy. The composition of U.S. imports, however, is fundamentally different from U.S. exports. While some imports and exports represent clearly substitutable items, other imports represent inputs to further processing, or are items that are not available in the economy. The ITA has issued various statements indicating that using the data in this manner is not appropriate. As the ITA has indicated, the employment estimate is not a multiplier and should not be used to estimate jobs that are associated with exports or imports in a multiplier fashion. In addition, the ITA estimates relate to the average number of jobs supported by exports across a broad section of the economy and should not be used in conjunction with trade agreements where the approach should more appropriately focus on estimating the change in the composition of employment that are associated with a change in trade as a result of a trade agreement.
The International Trade Administration (ITA) argues, however, that the job estimates should not be used with projected changes in trade to estimate potential employment effects from trade agreements. It says:22
Averages derived from IO [input-output]23 analysis should not be used as proxies for change. They should not be used to estimate the net change in employment that might be supported by increases or decreases in total exports, in the exports of selected products, or in the exports to selected countries or regions. (Emphasis added.)
The averages are not proxies because the number of jobs supported by exports usually does not change at the same rate as export value. The rate is not the same because other factors, such as prices, resource utilization, business practices, and productivity, do not usually change at the same rate. In addition, the material and service inputs and the labor and capital inputs differ significantly across types of exports. For example the labor requirements for an exported aircraft are significantly different from those of an exported agricultural product or an educational service.
Ideally, estimates of trade changes from tariff reductions would be multiplied by figures which reflect actual changes in employment (based on the mix of goods traded) that would occur at the margin as a result of changes in the volume of goods traded. According to the ITA, though, such data do not exist. The only data that are available reflect the estimated average number of jobs supported across the U.S. economy by a given level of exports. According to the ITA, "As a result, multiplying trade estimates from the computable general equilibrium (CGE) models by employment averages would tend to overestimate the actual number of jobs potentially lost to trade changes." (Emphasis added.)
The ITA also indicated that
In addition, estimates of the average number of jobs associated with exports cannot be adjusted for fluctuations in manufacturing capacity over the course of the business cycle. As explained by the USITC, the more slack there is in the U.S. economy, the more potential there would be for job creation:24
During periods of slack business activity, increased output, such as exports, would tend to increase employment, to lower unemployment, and to increase labor force participation. Conversely, during periods of high business activity, when industry operates at or near full capacity and employment, increased output, including output for exports, tends to raise employment less—if at all—and instead mainly shifts employment to industries that pay higher wages.
Conclusion
The 112th Congress passed legislation implementing the KORUS FTA. Part of the debate surrounding the agreement focused on the potential impact the agreement could have on U.S. employment, particularly employment in certain sectors of the economy. Given the current high rate of unemployment, it is not uncommon for communities or workers to raise concerns over the impact of a new FTA. Indeed, for some communities that already are affected by high rates of unemployment or have experienced plant closings due to foreign competition, the KORUS FTA could pose additional challenges.
An analysis of the available estimates of the potential effects of the KORUS FTA on U.S. employment, however, raises a number of questions concerning the usefulness of the estimates. Economic modeling naturally incorporates various assumptions and entails differing methodologies that can have a profound effect on the estimates that are generated, even when the estimates are derived from the same economic model. Standard models incorporate standard assumptions and approaches that generally are well explained. The mark of a good economic model often is one that uses assumptions and methodologies that seem reasonable and are not geared toward generating any particular result. In contrast, some models use non-standard assumptions and approaches that may be difficult to justify and seem to have been chosen in order to generate pre-determined results. Given the current state of economic modeling and data availability, the most accurate models likely can provide only rough estimates of the magnitude of the potential changes in employment in certain sectors, but cannot offer estimates of the precise size of the shifts in employment.
As this reports indicates, the most important assumption involved in generating estimates of the impact of the KORUS-FTA on employment appears to be the expected level of labor utilization in the U.S. economy over the ten-year phase-in period expected to be required to fully implement the KORUS-FTA. The existence of unutilized labor and capital make it possible for the models to generate long-term estimates of output and employment gains for the fully-implemented KORUS-FTA and losses, in case the agreement had not been adopted. It seems reasonable to assume, however, that the U.S. economy will return to its long-term trend approaching full employment during the ten years following the adoption of the KORUS-FTA. Another major qualification for the estimates may be that they do not account for changes in exchange rates, which can have a wide-ranging effect on the prices of internationally traded goods and may overwhelm changes in prices of goods that arise from changes in tariff rates.
Estimates of employment effects of new FTAs often tend to be highly subjective and can be misleading, because they represent a partial accounting of the total economic effects of new FTAs. In most cases, FTAs are comprehensive agreements that include provisions for goods, services, and investment. With few exceptions, estimates of employment effects of the KORUS FTA focus narrowly on employment effects in the goods sectors and neglect the potential effects in the services and investment areas.25 In addition, the estimates neglect a broad range of benefits for the economy as a whole that potentially can provide consumers with large economic benefits and that can yield broad productivity and efficiency gains for the economy and may enhance employment. As a result, estimates of the employment effects of new FTAs may serve poorly as an indicator of the total impact of a new FTA on the economy as a whole.
As policymakers considered the KORUS FTA and consider other FTAs, they likely will continue to weigh the results of a range of estimates of the employment effects of the FTAs to gauge the impact of the agreements. In this process, policymakers likely would be aided by estimates that clearly state the assumptions that are used and that inform policymakers about the broad implications of such agreements for the economy as a whole. In addition, policymakers likely would benefit from more reliable data on the potential magnitude of such agreements on specific sectors that would allow policymakers to craft programs that could assist those most directly affected by the agreements. As a result, policymakers may benefit from a number of initiatives to improve information and data on the impact of international trade on the economy. These might include
- increased information and data on services in the economy, including the shifting of in-house services from the manufacturing sector to the services sector and the formal and informal barriers to U.S. services posed by major trading partners;
- better data on worker dislocations including the reasons for those closings;
- better understanding of the development of global supply chains and the role they are playing in the U.S. economy.
Appendix A. Employment Population and Shares of the Economy, by Major Industry and Sector, 2010
|
Sector |
Employment (in thousands) |
Share of total private employment (%) |
||
|
Total including agriculture |
|
|
||
|
Total goods producing including agriculture |
|
|
||
|
Mining and logging |
|
|
||
|
Construction |
|
|
||
|
Manufacturing |
|
|
||
|
Agriculture |
|
|
||
|
Services |
|
|
||
|
Trade, Transportation, and Utilities |
|
|
||
|
Information |
|
|
||
|
Financial Activities |
|
|
||
|
Professional and Business Services |
|
|
||
|
Education and Health Services |
|
|
||
|
Leisure and Hospitality |
|
|
||
|
Other Services |
|
|
||
|
Government |
|
|
||
|
Federal |
|
|
||
|
State |
|
|
||
|
Local |
|
|
Appendix B. Standard Assumptions of Economic Models
In order to assess the validity of the various estimates of the employment effects of the KORUS FTA, it is first necessary to understand how the economic models work and the assumptions that are made. The standard approach to estimating the effects on U.S. employment associated with trade agreements follows a number of steps, although most of these steps are not included in the discussion of the estimates. In general, these steps include the following:
- 1. Estimating the impact a change in the tariff rate would have on a change in the prices of goods. For instance, if tariff rates are lowered by 10%, would goods prices also fall by 10%? Are the changes in tariff rates accomplished at once, or are they reduced slowly, according to a set schedule? If tariff rates are adjusted over time according to a set schedule how does that affect the rate at which goods prices adjust?
- 2. Estimating the impact a change in goods prices would have on a change in sales of those goods. This estimate attempts to quantify the responsiveness of consumers to changes in prices. In markets with a number of close substitutes, the consumer response could be strong with consumers buying less of a particular product as its price rose relative to those of similar products. For goods with few substitutes, consumers would be less responsive to changes in prices.
- 3. Estimating the impact a change in sales would have on a change in output or a change in employment. A change in employment associated with a change in sales would depend, at the very least, on the level of plant utilization, the level of productivity, and the availability of labor.
In attempting to estimate these intermediary steps, most models incorporate a number of assumptions to reduce the high level of variability that is intrinsic in such estimates. Often these additional assumptions are not well explained. In general, these basic assumptions are:
- 1. Changes in Tariff Rates Will Translate Directly into Price Changes. Most models assume that any change in the tariff rate will be passed along completely to the change in the prices of traded goods. In most cases, however, tariff rates are not adjusted completely at the time of implementation, but are adjusted over a set schedule that can encompass a decade or longer. Also, there is no guarantee that the prices of traded goods would adjust at exactly the same rate as tariff rates are adjusted. They may be passed along at a differential rate or simply absorbed by the exporter.
- 2. Exchange Rates Will Remain As They Are. In most cases, the estimates do not make any assumptions concerning potential changes in the exchange rate of the dollar as a result of a trade agreement. In addition, the exchange value of currencies of competitors could affect the expected change in exports or imports associated with a trade agreement.
- 3. Consumer Purchases Will Follow Tariff Reductions. Most of the estimates on changes in employment are based on the assumption that reductions in the prices of goods associated with a drop in the rate of tariffs would result in a complete substitution of goods toward the FTA countries and away from other foreign suppliers. In general, there may be some goods within the total basket of traded goods that consumers consider easily substitutable (such as steel and fasteners), but consumer choices often are predicated on more than relative prices (such as brand names and quality) and, therefore, the rate of substitution overall may be quite low. To the extent that consumers do not shift their purchases based on changes in prices, the change in employment would be blunted. In addition, standard economic theory argues that trade agreements generally entail both trade diversion, or the substitution of lower-priced goods among the parties to the agreement for the now higher-priced goods from other suppliers, and trade creation. Trade agreements can create trade by increasing efficiency in production and by increasing the real incomes of consumers that, in turn, leads to a greater level of consumption of both domestic and imported goods. To the extent that a trade agreement creates additional trade, it could have a positive effect on employment.
- 4. Employment or Wages Will Not Change. Another assumption that often is made is whether employment or wages should be held constant. Such an assumption often is necessary in order to generate results from the economic models. By holding employment constant, the model is attempting to estimate the change in the composition of employment that would be associated with a trade agreement. This assumption would allow wages to change, although most models do not attempt to estimate changes in wages. As a result, these models generally attempt to highlight the magnitude of the impact of an agreement on various sectors in the economy, rather than attempting to generate precise estimates of the actual number of jobs that may be gained or lost by individual sectors. In contrast to these assumptions, models that hold wages constant allow employment to change in an effort to estimate changes in the number of jobs in various economic sectors. Since changes in wages often lead to changes in employment, this assumption is questionable. In either case, the estimates usually attempt to aggregate the total cumulative effects over a number of years, often a decade, in order to derive an estimate of the overall change in employment. While such an assumption is a necessary condition for generating results, the assumptions compromise the validity of the results, since the composition of employment in the economy over a decade can change quite noticeably, even in the complete absence of international trade, due to business cycle effects and the constant churning that occurs. Currently none of the estimates of the employment effects of trade agreements incorporate either business cycle or other transformational effects into the estimates.
Key Steps in Converting Changes in Tariffs to Changes in Employment
In general, there are a number of steps involved in converting changes in tariffs into changes in employment. In most cases, these steps are not explicitly explained. The chart below summarizes these three steps.
Table B-1. Key Steps Involved in Converting Changes in Tariff Rates into Changes in Employment
|
Step 1 |
Estimate changes in trade (exports and imports) arising from a change in tariff rates.a |
|
Step 2 |
Link changes in trade (exports or imports) to changes in the industrial sectors that produce those goods.b |
|
Step 3 |
Link changes in industrial production to changes in employment.c |
a. This step assumes that a reduction in the applied rate of tariffs would lead to a similar reduction in the prices of imported goods. As a result of the reduction in prices, it is assumed that consumer demand would shift from other goods in favor of the lower-priced goods, so that there is a one-for-one substitution of goods, or that the process of tariff reduction is a zero-sum game, with gains by the FTA partner comparable to losses by domestic or non-FTA suppliers.
b. Linking traded goods with their respective industries is complicated, because the data collected on exports and imports are entirely different from those collected on the industrial composition of the economy. Whereas exports are recorded as the number of specific goods exported at a certain price, the industrial composition of the economy is organized not by individual commodity, but by productive processes, so that it requires considerable effort to link together the goods produced for export with the industry that produced those goods. However, in recent years, classification "dictionaries" have been developed that can make these translations easier.26
c. In a joint project, the Department of Commerce (DOC), Bureau of Economic Analysis (BEA), and the Bureau of Labor Statistics (BLS), estimated the average number of jobs supported by exports in the U.S. economy based on the dollar value of output relative to the average number of jobs required to produce that output for each industry.27 As a result of this joint effort, the Commerce Department determined in its 2010 update that on average $166,000 in goods exports, $216,000 in services exports, or a weighted average of $180,000, supported one job in each respective sector. As indicated previously, this estimate should not be used to estimate the number of jobs lost as a result of imports, or a trade deficit.