Order Code RS21737
February 13, 2004
CRS Report for Congress
Received through the CRS Web
NAFTA at Ten: Lessons from Recent Studies
J. F. Hornbeck
Specialist in International Trade and Finance
Foreign Affairs, Defense, and Trade Division
Summary
On January 1, 2004, the North American Free Trade Agreement (NAFTA)
completed its tenth year and most of its provisions are now implemented. NAFTA is
a free trade agreement (FTA) that effectively added Mexico to the U.S.-Canada FTA
completed in 1989. Its anniversary has sparked numerous evaluations, which are
particularly relevant as the United States pursues free trade agreements with multiple
Latin American countries. Most studies found that NAFTA’s effects on the U.S. and
Mexican economies to be modest at most. This report provides an analytical summary
of the economic lessons reached in support of Congress’s role in the trade policy
process. It will be updated as needed.
Introduction
Free trade agreements are supposed to enhance the welfare of participating countries,
so evaluating their effects is a valuable exercise. NAFTA is particularly relevant to the
bilateral trade agreements being considered by the United States today because it was the
first trade agreement in a non-multilateral setting between a developing and two
developed countries. As such, it is important to note that this report focuses on U.S.-
Mexico issues, not because Canada is unimportant, but because the U.S.-Canadian free
trade agreement predates NAFTA, is less controversial in the eyes of most trade critics,
and is less relevant (because it entails trade between two developed economies) to the
pending trade agreements with Latin America. Further, because trade between Canada
and Mexico remains very small, the trilateral trade agenda is still only emerging, although
there is growing interest in analyzing immigration, security, and other issues within this
trilateral framework.
This report evaluates four studies produced by the Congressional Budget Office
(CBO), the World Bank, the Carnegie Endowment for International Peace, and the United
States International Trade Commission (USITC). These assessments of NAFTA, by and
large, are analytical in nature, use established methodologies, caveat their own work to
reflect limitations of the research, and draw on academic rather than special interest
research. The details of their methodologies are not reproduced here, but it is important
to note that they faced similar research challenges. These include: 1) isolating the effects
Congressional Research Service ˜ The Library of Congress

---

CRS-2
of NAFTA from many other economic policies and forces at play; 2) using sufficiently
long time-frames to separate out pre- and post-NAFTA trends and effects; and 3)
comparing NAFTA over time and across countries to provide relative measures of the
importance of any observed or inferred change.1
The lessons outlined below reflect conclusions of the reports, but not all addressed
each of the issues. When analyses overlap, agreements and differences are identified, and
other sources are cited when needed to clarify or expand on a certain theme. Conclusions
regarding NAFTA’s effects are not drawn relative to expectations espoused prior to its
implementation. For political and other reasons, many of the claims made about NAFTA
in the early 1990s, both good and bad, were less than credible. Details below provide a
sense of NAFTA’s economic success or failure relative to its effect on trade, investment,
economic growth, productivity, employment, wages, and immigration.
Trade and Investment Effects
NAFTA is a broad agreement, but improved market access, including tariff
reductions on merchandise trade, was the major U.S. goal. After ten years, most tariffs
have gone to zero, except for some very sensitive (mostly agricultural) goods that have
limited protection for up to 15 years. Clearly, U.S.-Mexico trade and investment have
grown sharply over the past decade. From 1994 to 2003, U.S. exports to Mexico rose
91%, compared to 41% to the world. U.S. imports increased by 179%, compared to 89%
from the world. This surge, however, began prior to NAFTA, so the question is, how
much of the post-1994 growth can be attributed to NAFTA?
NAFTA had a modest effect on U.S.-Mexico trade growth. The CBO,
World Bank, and USITC approached the problem differently, but all found that NAFTA
had a modest effect on U.S.-Mexico trade growth. The CBO model of U.S.-Mexico trade
estimated that 85% of the U.S. export growth and 91% of U.S. import growth would have
occurred without NAFTA. Although the effect was modest, it accelerated over time,
accounting for a 2% marginal growth of U.S. exports and imports in1994 up to11% and
8% marginal growth of U.S. exports and imports in 2001. As a percentage of economic
activity, the increased trade was more pronounced for Mexico than the United States.
Separately, the World Bank makes the point that NAFTA has reinforced existing trends
in trade growth and estimates that Mexico’s global exports would have been 25% lower
without NAFTA.2
1 Lederman, Daneil, William F. Maloney, and Luis Serven. Lessons From NAFTA for Latin
America and the Caribbean Countries: A Summary of Research Findings. The World Bank,
Washington, D.C. December 2003, Audley, John J., Demetrios G. Papademetriou, Sandra
Polaski, and Scott Vaughan. NAFTA’s Promise and Reality: Lessons from Mexico for the
Hemisphere. Carnegie Endowment for International Peace, Washington, D.C. 2004, and United
States International Trade Commission. The Impact of Trade Agreements: Effect of the Tokyo
Round, U.S.-Israel FTA, U.S.-Canada FTA, NAFTA, and the Uruguay Round on the U.S.
Economy. Publication 3621. Washington, D.C. August 2003. The Congressional Budget Office.
The Effects of NAFTA on U.S.-Mexican Trade and GDP. Washington, D.C. May 2003.
2 CBO, pp. 17-19, World Bank, p. v, 298, and 307, and USITC, p. 66.

---

CRS-3
The USITC analyzed tariff preferences by sector to isolate the effects of NAFTA on
U.S.-Mexico trade. It estimated that NAFTA tariff preferences accounted for one-third
of the growth in U.S. import shares from Mexico (higher among textile and apparel
goods) and 13% of growth of U.S. exports to Mexico. The remaining growth in trade
would have occurred anyway, and was influenced more by factors such as the 1994 peso
devaluation and existing preferences provided Mexico under the Generalized System of
Preferences (GSP) and production-sharing (maquiladora) programs. The U.S. benefit
from NAFTA may also be seen in Mexico’s response to the 1994 peso crisis, which was
to raise tariffs against non-NAFTA partners. (In another report, the USITC noted that
there was also an increased variety of goods traded and a rise in the average price of
Mexican imports, suggesting new and upgraded goods being imported from Mexico.)3
There is little evidence of trade diversion. A key concern of trade analysts
is whether an FTA results in trade shifting from nonmember countries to members of a
trade agreement because the tariff preferences have allowed them to become the lower-
cost producers. This has the effect of switching trade from more to less efficient trading
partners. The World Bank Study found no significant evidence of trade diversion in
NAFTA, particularly with respect to textile and apparel producers in neighboring Central
America and the Caribbean.4 This is consistent with a majority of studies done earlier.
NAFTA did not cause the widening U.S. trade deficit with Mexico. From
1994 to 2002, the U.S. trade deficit with Mexico grew from -$1.4 to -$37.1 billion. These
studies found that trade deficits are largely macroeconomic phenomena, in this case
predominantly attributed to the respective business cycles in Mexico and the United
States. Strong U.S. growth in the 1990s combined with Mexico’s deep recession caused
by the December 1994 peso crisis (devaluation) were the main factors cited for the large
deficits. Importantly, none of the studies attributed the peso crisis to NAFTA, but to
structural misalignments in the Mexican economy combined with political events.5
NAFTA helped increase bilateral foreign direct investment (FDI). 6 From
1994 to 2002, U.S. FDI in Mexico rose from $16.1 billion to $58.1 billion, or 259%.
Mexican FDI in the U.S. increased 244% to $7.9 billion, albeit from the much smaller
base of $2.3 billion. FDI in Mexico (mostly U.S.) grew on average from 1.1% of GDP
in 1980-93 to 3.0% in 1994-2001. The World Bank noted that NAFTA was one of
numerous factors directing FDI and estimated that it led to a 40% annual increase in FDI
to Mexico, without diverting FDI from other countries. The CBO and USITC studies
basically agreed, finding that NAFTA’s investment and trade liberalization worked
together to reduce risk and improve profitability, and so observed that NAFTA helped to
increase total investment flows to Mexico.
3 USITC, pp. 302-04, 308, 313. Hillberry, Russell H., and Christine A. McDaniel. A
Decomposition of North American Trade Growth Since NAFTA. USITC. Washington, D.C.
Working Paper No. 2002-12-A.
4 World Bank, pp. xvii-xviii and 298.
5 CBO, p. 20 and USITC, p. 49. The Carnegie Endowment study (Polaski, p. 14) does state that
the tariff cuts, “resulted in a shift from a net trade deficit with the United States before NAFTA
to a substantial net trade surplus in 2002” without citing evidence for it.
6 World Bank, p. v, 155, and 366, CBO, p. 4, USITC, p. 158.

---

CRS-4
Domestic Economic Effects
The various studies reached different conclusions with respect to the macroeconomic
effects of NAFTA. Choice of methodology and depth of research varied and likely
explain many of the differences.
NAFTA slightly increased growth in output and productivity. The CBO
study, which had a limited model for estimating the trade effects on GDP, found that
NAFTA increased annual GDP growth in the United States by no more than .04%, and
for Mexico, no more than .8%. Similarly, the USITC cited other literature suggesting that
U.S. GDP could grow by an additional 0.1% to 0.5% once NAFTA is fully implemented.
The World Bank study, which undertook a more elaborate modeling effort, concluded that
Mexico’s economic performance was similar to the rest of Latin America prior to
NAFTA. After NAFTA went into effect, Mexico’s per capita GDP converged
increasingly toward that of the United States, although a large discrepancy still exists.
Estimates of the rate of convergence suggest that without NAFTA, Mexico’s per capita
GDP growth would have been 4-5% lower by 2002.7 The World Bank also suggested that
NAFTA contributed to “a substantially faster rate of productivity convergence than in
previous years.” For example, Mexican manufacturers were able to halve the time needed
to adapt U.S. technological innovation. The Carnegie study also notes that productivity
rose dramatically after NAFTA in both the United States and Mexico and suggests
NAFTA “likely played a significant role.”8
NAFTA had little or no impact on aggregate employment. NAFTA is at
the heart of a long-standing debate over the employment effects of trade because of fears
that trade with developing countries causes U.S. job losses and that trade deficits equate
to higher unemployment. None of the reports attributed changes in aggregate U.S. or
Mexican employment levels to NAFTA, but the author of the first chapter of the Carnegie
study suggests that changing the assumptions of a USITC model would allow for a net
gain in U.S. employment over the past decade of between zero and 270,000 jobs, a small
increase. For Mexico, it concludes that “the sum of the effects of the trade pact to date
has not been a strong net gain in overall employment.” The second chapter (different
author) argues for zero net growth in U.S. jobs. The USITC study demonstrates, contrary
to some popular opinion, that U.S. trade deficits tend to occur during periods of low
unemployment, and “vice versa.” This evidence supports well-established economic
theory that would suggest both the U.S. trade deficit with Mexico and U.S. employment
levels over the past decade were responding to economic growth, not each other.9
NAFTA contributed to employment shifts among sectors. The Carnegie
report observes a shift in Mexican employment away from agriculture toward services and
manufacturing. It attempts to correlate this shift with NAFTA-induced trade balances in
agriculture (deficit) and manufacturing (surplus) goods, but concludes that it is impossible
7 CBO, p. 22, USITC, p. 32, and World Bank, p. 27-28.
8 World Bank, p. 5 and 28 and Carnegie Endowment, (Polaski, pp. 24 and 33). The USITC
summary of the literature on productivity and trade cited numerous studies supporting the claim
that trade liberalization enhances productivity. pp. 100-13.
9 Carnegie Endowment, (Polaski, p. 20 and 28, Papademetriou, p. 39), USITC, p. 49

---

CRS-5
to establish precisely how much of the jobs shift can be attributed to NAFTA. The World
Bank points to productivity growth in irrigated agricultural lands and the lack of
opportunity in subsistence agriculture as alternative reasons for these shifts. In the United
States, employees who lost jobs because of NAFTA are eligible for NAFTA trade
adjustment assistance. The Carnegie report notes that at the end of 2003, some 525,000
workers have been certified under this program, heavily concentrated in manufacturing,
especially apparel. Over a period of ten years, this represents a small portion of the
aggregate work force, many of whom are already re-employed. Nonetheless, it is among
the most salient adjustment issues related to trade, along with the possible need for a
larger “social safety net.” The USITC and World Bank also show that structural shifts in
both countries had been affecting long-term employment patterns as well.10
NAFTA has had a small effect on real wages. The USITC, in summarizing
the vast literature on the observed rising U.S. income gap between more-skilled and less-
skilled workers, suggests that while estimates varied, trade in general has contributed to
no more than 10-20% of the wage gap. Economists generally consider the wage-gap
problem to be a function of skill-based technological change that causes an employment
bias toward more highly educated or trained workers. Increased trade of intermediate
goods using outsourcing or production-sharing arrangements has also been linked in
recent research. For Mexico, the Carnegie Endowment and the World Bank note that real
wages are lower than when NAFTA began, but conclude that it was not the cause.
Decomposing the trend shows that Mexico experienced a 25% fall in real wages after the
1994 peso crisis. Real wages began a steady recovery in 1997 and are approaching 1994
levels. Interestingly, the World Bank study showed that those Mexican states tied to FDI,
exports, and maquilas had higher and faster-growing wages than other states.11
Immigration patterns were not affected by NAFTA. The long-term trend in
legal and unauthorized worker migration from Mexico to the United States continued and
accelerated after NAFTA was implemented. The trade agreement, however, was found
to be largely irrelevant. The Carnegie Endowment study points to many factors that
outweigh any effect an FTA can have on migration patterns. Mexican workers have been
drawn to the United States by big wage differentials and the high demand for low-skilled
workers. They have been encouraged to leave Mexico by the burgeoning work force that
cannot be absorbed and the strong migration networks in place. Periodic financial crises
(1982, 1986, 1994) exacerbate the problem. These cause huge losses in formal sector
jobs, large declines in real wages relative to those in the United States, and failing
confidence in the Mexican economy, all of which encourage emigration. The World Bank
found that Mexican emigrants had higher levels of education and earned more than non-
emigrants and so migration may have contributed to Mexico’s growing wage gap.12
NAFTA has a minor role in Mexico’s rural-urban migration. The first
chapter of the Carnegie Endowment study argues that the observed trend of migration
10 Carnegie Endowment, (Polaski, p. 20-28), World Bank, p. xv, xxv, 183-85, and USITC, p. 41-
45. On U.S. sectoral issues related to textiles, see: Rhodes, Karl. Global Gain, Local Pain.
Regional Focus. Federal Reserve Bank of Richmond. Winter 2004. pp. 12-18.
11 USITC, p. 67-68, 114, 125, World Bank, pp. vi-vii, Carnegie Endowment, (Polaski, p. 24-25).
12 Carnegie Endowment, (Papademetriou pp. 40-41and 50-53) and World Bank, pp. 174-80.

---

CRS-6
from rural areas of Mexico to urban centers, along with the attendant problems of
unemployment, family disruption, and poverty, is directly the result of agricultural
liberalization linked to NAFTA. This is contradicted by a second chapter (different
author) that argues such migration patterns have been in place since 1960 when over 50%
of the workforce was agricultural compared to 36% in 1980 to less than 25% in 1995.
NAFTA appears at most to have affected at the margin an established trend that
economists long have argued is common in the development process of most countries.13
Outlook
The four studies discussed above point to three broad themes. First, by most
aggregate measurements, NAFTA has had only a modest, but positive, effect on the U.S.
and Mexican economies and tends to reinforce long-term trends already evident by its
inception. This is in keeping with what is widely understood about trade and trade
agreements; they work at the margin of economies and their effect can be easily confused
with much more powerful factors such as long-term structural change and short-term
volatility (e.g. financial crises). Such confusion is seen in the reluctance of many
supporters and opponents of NAFTA to engage in a more nuanced debate on trade.
Second, adjustment problems related to trade liberalization present the greatest
challenge to policy makers. For export firms and sectors, the adjustment is positive and
provides evidence of the winners from trade. For import competing sectors, displacement
can have devastating effects on communities and raises the question of whether to fight
freer trade or attempt to adjust to it as part of the larger global integration process. The
World Bank report argues that trade agreements can do more by improving distorting
rules of origin, taking on the hard tasks of antidumping and countervailing duty measures,
and tackling adjustment problems. The Carnegie Endowment study argues that trade
agreements should address trade-related adjustment issues through longer tariff reduction
schedules, use of special safeguards, removal of agricultural subsidies, and provision for
regionally funded trade adjustment assistance and social safety net programs.14
Third, two studies address a common call for better integration of trade policy into
a country’s overall development program by coordinating and supporting it with domestic
reforms. The Carnegie study argues for more attention to agricultural, environmental,
immigration, tax, and labor rights protection policies, among others. The World Bank
study prioritizes institutional reform (especially rule of law and anti-corruption efforts),
educational development (to promote technology transfer), other innovation supporting
policies, and labor reform that facilitates employment transition among industries and
sectors.15 In the end, these reports do not provide easy answers to trade-related policy
problems, but do attempt to explain how the gains from trade may be enhanced by
understanding and responding better to the adjustment challenges all countries face.16
13 Carnegie Endowment, (Polaski, pp. 17-23 and Papademetriou, pp. 46-47, 51-52).
14 Carnegie Endowment, pp. 7-8 and World Bank, pp. v, xiv, xxiv, xxv.
15 Carnegie Endowment, pp. 7-8 and World Bank, pp. v, viii, ix, xxiv, xxv.
16 Themes developed here can be found in academic literature. See: Chambers, Edward J. and
Peter H. Smith, eds. NAFTA in the New Millennium. University of California Press. 2002.

---