The North American Free Trade Agree-ment unites the United States, Mexico andCanada—three nations with a combinedpopulation of 426 million, total output ofmore than $13 trillion and regional trade of$700 billion in goods and services. Because of the North American mar-ket’s sheer size, NAFTA has been repeated-ly dissected. Most studies have sought todetermine whether the pact fulfilled propo-nents’ predictions of increased trade, lowerprices and higher incomes or led to whatcritics warned would be a “giant suckingsound” of U.S. jobs going to Mexico. On balance, researchers have foundNAFTA a slight positive for the U.S. as awhole. For example, a 1996 study estimat-ed that NAFTA had increased U.S. exportsby $5 billion, or 12 percent, a figure pro-jected to grow as more of NAFTA’s phased-in trade liberalization took effect.1A lesser volume of research focuseson what NAFTA has meant to state andlocal economies, although theory and com-mon sense suggest trade deals might havedifferent impacts within countries. States’industrial mixes and workforces vary wide-ly, leading to comparative advantages thatinfluence the composition and destinationof exports. Geography is another key fac-tor. Firms may operate in one state ratherthan another to take advantage of proximityto newly opened markets. The results ofnational studies of NAFTA’s effects may notapply uniformly to all states.Texas is one of the more interestinglenses through which to assess NAFTA. Thestate lies near the center of NAFTA’s eco-nomic space—about equidistant fromMexico City and Toronto, with a 1,200-milefrontier with Mexico and networks of high-ways and rail lines that lead to some of theworld’s busiest border crossings. Texaspolitical and business leaders strongly sup-ported NAFTA’s ratification, an indicationthat many presumed it would benefit thestate’s economy.Has NAFTA been good for Texas?Merely counting the truckloads passingthrough border checkpoints in the LowerRio Grande Valley, Laredo and El Pasowould make it seem so. A more definitiveanswer, though, involves distilling NAFTA’sinfluence from factors responsible for over-all increases in Texas exports over the pastdecade or so. NAFTA can’t be deemed a success forTexas if rising exports to Mexico merelyrepresent sales diverted from markets else-where in the world. Trade theory suggeststhat overall economic effects of NAFTA andother preferential trade agreements dependon trade creation net of trade diversion(see box). A fresh look at the issue, using indus-try-level export data, shows that NAFTAdid indeed increase Texas’ sales toDid NAFTA Spur TexasExports?By Anil KumarFEDERAL RESERVE BANK OF DALLAS • MARCH/APRIL 2006 SouthwestEconomy3Trade Creation Versus Trade DiversionPreferential trade agreements impose lower tariffs on trade in goods and services among theirmember countries. Even with expansion of the multinational World Trade Organization in recent years,nations have found these regional deals increasingly attractive, concluding more than 180 pacts since1990.Two types of preferential deals are common. Free trade areas, such as NAFTA, reduce tariffs ongoods from member countries but allow each nation to set its own duties for nonmembers. Customsunions, such as the European Union, agree to impose a common tariff wall on imports from nonmem-ber countries. In economic terms, they’re similar, so the following discussion applies to both.These preferential agreements would normally violate the WTO’s most favored nation rules, whichrequire each member to extend to other members the lowest tariff applicable on all goods and services.In other words, there should be no discrimination or preference in tariffs. To allow the existence of freetrade agreements and customs unions, WTO rules exempt them from the most favored nation rule if theymandate complete tariff elimination among member countries and if tariffs to nonmembers are no higherthan they were before.Both theory and experience suggest that free trade increases economic welfare. Does the proposi-tion hold for preferential deals as well? Jacob Viner provided the answer in his classic 1950 book, The Customs Union Issue. It introducedtwo important concepts—trade creation, which denotes new imports and exports, and trade diversion,which means a mere shifting of sources from one country to another. Viner argued that only trade dealsthat lead to net trade creation would improve economic welfare. If net trade diversion occurs primarilyby shifting production from a low-cost nonmember country to a high-cost member country, it will hurtoverall economic welfare.Texas is one of the more interesting lenses through which to assess NAFTA.Mexico—and to Canada as well. Perhapsmore interesting, NAFTA also helped raiseTexas exports to Asia, Europe and LatinAmerica, making a strong case for nettrade creation.Before and After NAFTANAFTA went into effect Jan. 1, 1994.In general, it mandated eliminating tradebarriers by 2008. For many products, theagreement did away with tariffs and otherrestraints immediately. Agriculture andapparel were the main sectors scheduledto be liberalized over a longer period. Pre-NAFTA Mexico had the more pro-tected economy, so it committed to largertariff cuts than the U.S. and Canada.Average Mexican duties on U.S. goods fellfrom 12 percent in 1993 to 1.3 percent in2001, while U.S. tariffs on Mexican goodsdeclined from 2.1 percent to 0.2 percent.2The effect of NAFTA on U.S.–Canada traderestraints was minimal because the twocountries operated under a free-tradeagreement that took effect in 1989.Trade has increased by leaps andbounds in the NAFTA years. U.S. exportsto Mexico rose from $42 billion in 1993 to$111 billion in 2004, while imports fromMexico increased from $40 billion to $156billion. Over the same period, U.S. sales toCanada grew from $100 billion to $189 bil-lion, while imports from Canada to the U.S.climbed from $111 billion to $256 billion. During the first six years of NAFTA,Texas gained ground in many foreign mar-kets, allowing the state to grow faster thanthe nation in overall exports (Chart 1A).Texas exports to Mexico also increased—but not by any more than the nation as awhole. From 1994 to 2000, the growth ofTexas shipments across the Rio Grandemirrored that of U.S. exports, just as it didin the five years prior to NAFTA’s takingeffect (Chart 1B).3Indeed, both Texas andU.S. exports to Mexico grew steadily beforeand after NAFTA, except for a