Data Fail: The Divergence Between Rosy . - Public Citizen

Public CitizenData FailMexico under NAFTA, China’s WTO accession and Korea since the FTA. In an economy that isstill far from recovering from the impact of the Great Recession, this can be a serious omission.The fact that the model rules out the possibility of increasing trade deficits and job losses meansthat is unlikely to pick up its full effect on increased inequality. That is to say the mechanics ofthe model assume that when the output of the less competitive sectors of an economy decrease(relative to the baseline), employment will simply shift from contracting to expanding sectors,where wages are assumed to increase.Implicit in the assumption that the trade balance does not change is the assumption of flexibleexchange rates. This assumption clearly does not correspond to reality since currencymanipulation is a significant problem among some of the TPP countries. For instance, the U.S.Department of Treasury recently included TPP nation Japan on its new Monitoring List in itssemi-annual report on “Foreign Exchange Policies of Major Trading Partners of the UnitedStates.”10 Despite previous economic studies showing that foreign currency manipulation hascost the United States an estimated one to five million jobs,11 the TPP does not have enforceablerules governing currency manipulation, and the USITC model assumes away the issue.Finally, the output of any model is also greatly affected by the data put into it. Perhaps the mostcontroversial issue in this regard is how “non-tariff barriers” (NTB) are considered. What aninternational bank may consider an NTB may be what a policymaker or consumer considers animportant safeguard to avoid costly financial crises. What an international agribusiness companyconsiders an NTB may be relied up by public health officials and consumers to avoid food-borneillness outbreaks and associated costs. Increasingly, trade agreement modeling includesguesstimations of gains that will be created by removing NTBs. As well, models may or may notconsider how trade pact investment rules could affect decisions about where to invest inproduction and whether a pact will alter foreign direct investment trends. Also at issue is howintellectual property provisions are included in the model, given longer monopolies may increasesome U.S. firms’ profitability but also may cost governments and consumers more for medicinesand access to information.Different assumptions can result in diametrically opposed outcomes. With respect to the TPP,consider the stark differences between the findings of the Peterson Institute for InternationalEconomics and economists at Tufts University. The Peterson Institute used a CGE model withassumptions similar to those employed by the USITC in past studies and found the pact wouldresult in a modest increase in gross domestic product, but not impact overall U.S. employment.12Using an economic model that allows for the possibility of less than full employment and risingincome inequality, called the United Nations Global Policy Model, Tufts University economistsconcluded that the TPP would reduce U.S. growth rates and lead to 448,000 American jobslost.13 The Tufts findings spotlight just how drastically the assumptions baked into a model affectthe outcomes: The Tufts economists actually employed the Peterson Institute trade flowsimulation data! That is to say that they plugged the Peterson findings on import and exportlevels at full TPP implementation derived from one set of unrealistic assumptions into a modelthat applies more realistic assumptions about how trade flow changes affect growth andemployment and got the opposite outcomes with respect to those measures that the PetersonCGE model produced.May 20162

Public CitizenData FailThe Divergence between Rosy USITC Projectionsand U.S. Trade Agreements’ Actual OutcomesThe USITC’s past trade agreement assessment studies have proved to be widely off mark. Thesystematic disconnect between USITC projections and actual outcomes is evident whenreviewing the USITC studies for the three most impactful trade agreements prior to the proposedTPP: NAFTA, the U.S. bilateral agreement with China that set the terms of its WTO entry andrelated PNTR vote, and the U.S.-Korea FTA.NAFTAIn early 1993, the USITC published a study on the economic effects of Mexico joining thealready existing 1988 FTA between the United States and Canada.14 The NAFTA USITC reportprojected that the U.S.-Mexico trade balance in goods and services would improve because U.S.exports to Mexico would increase from a low of 5.2 percent to a high of 27.1 percent, and U.S.imports of Mexican goods and services would increase from a low of 3.4 percent to no higherthan 15.4 percent.15 In sum, the study concluded that a small U.S. trade surplus with Mexico atthe time of the pact would increase as a result of a fully realized NAFTA.The NAFTA USITC study estimated minimal changes to U.S.-Canada trade flows given many ofthe tariff cuts and non-tariff provisions included in NAFTA were also included in the U.S.Canada FTA. The report also estimated that NAFTA would deliver an increase of less than 1percent in aggregate employment for the United States.NAFTA’s actual outcomes were the opposite of the USITC projections.16 Contrary to the USITCprojections, goods imports from Mexico grew at nearly double the rate of U.S. goods exports toMexico. Before NAFTA, the United States had a trade surplus in goods of 2.6 billion withMexico. In 2015, the U.S. trade deficit in goods with Mexico was almost 106 billion. Evenwhen accounting for goods and services, the U.S.-Mexico deficit was 57 billion in 2015 and nota surplus as the USITC projected. The U.S. trade deficit with Mexico is only beaten by the U.S.deficit with China and is one of the largest trade deficits between any two countries in the world.U.S.-Mexico Trade1993 - Baseline 2.6 billion goods surplus(services data not available)USITC Projection17 10.6 billion goods andservices surplus2015 - Actual 57 billion goods andservices deficitSource: U.S. International Trade Commission, “Interactive Tariff and Trade Dataweb,” for goods data andU.S. Bureau of Economic Analysis for goods and services data.The USITC projections for U.S.-Canada trade were also off. The United States also had a tradedeficit in goods with Canada of 30 billion in 1993. By 2015, the U.S. goods trade deficit withCanada more than doubled to 61 billion.May 20163

Public CitizenData FailThe USITC also thoroughly miscalculated NAFTA’s effects on American jobs. Rather thandelivering a modest increase in American jobs, NAFTA’s investor protections reduced the riskfor U.S. firms to relocate production to Mexico to take advantage of its lower wages and weakerenvironmental standards. The increase under NAFTA in the U.S. trade deficit with Mexico andCanada equated to an estimated net loss of one million U.S. jobs by 2004, according to theEconomic Policy Institute (EPI), which calculated the net balance between jobs created and jobslost from NAFTA trade flows.18 EPI calculates that the ballooning trade deficit with Mexicoalone destroyed about seven hundred thousand net U.S. jobs between NAFTA’s implementationand 2010.19This toll has likely grown since 2010,as the non-oil U.S. trade deficit withMexico has risen further.20 Whileinterests seeking to minimize therelevance of the NAFTA trade deficitoften attribute oil as the primaryreason for the large and growing U.S.deficit with Mexico, the share of thedeficit attributable to oil began todecline abruptly in 2010. By 2015,the United States had a trade surplusof 5 billion with Mexico in oil, asour overall goods trade deficitincreased to 112 billion.USITC NAFTAProjection:NAFTA Reality:The U.S.-Mexicotrade balance in goodsand services wouldimprove and a smallnumber of newAmerican jobs wouldbe created.A 2.6 billion goods tradesurplus with Mexicoexploded into a 106 billiondeficit. The 2015 goods andservices combined deficitwas 57 billion and 845,000Americans were certified asNAFTA job-loss casualtiesunder one narrow program.Another way to capture the reality of American job loss caused by NAFTA is to review the morethan 845,000 U.S. workers nationwide in the manufacturing sector that have been certified forTrade Adjustment Assistance (TAA) since NAFTA because they lost their jobs due to importsfrom Canada and Mexico or the relocation of production to those countries.21 These numbersrepresent a significant undercount, as TAA only covers a subset of jobs lost to trade. The scopeof this program has changed over time. Until 2011 the program did not cover any service-sector,agricultural or fisheries jobs. Prior to 2002, only actual manufacturing jobs were counted in thetally of industrial-sector job loss. Thus, for instance, if an auto assembly factory employing 4,000workers was closed and production was moved to Mexico, TAA did not list 4,000 jobs lost.Workers in office, engineering, maintenance, shipping and distribution and custodial jobs at theclosed facility were not covered. And, to even be considered, workers, a union or a company hadto take the initiative to apply. Thus, the TAA numbers significantly undercount NAFTA job loss.The U.S. Department of Commerce initially maintained a website listing jobs gained underNAFTA. But when NAFTA job loss TAA certifications zoomed above 200,000 within the pact’sfirst years and the job creation list remained below 15,000, the U.S. government’s NAFTA jobcreation list abruptly disappeared.The USITC study also got NAFTA’s effect on U.S. wages wrong. While American workerproductivity has almost doubled since NAFTA, U.S. median wages have remained almost flat.Many economists have concluded that NAFTA has contributed to downward pressure on U.S.wages as a form of labor arbitrage with lower-wage Mexican workers pushed down U.S. wagesMay 20164

Public CitizenData Failfor the 63 percent of Americans without a college degree.22 This in turn had contributed togrowing income inequality.23 According to the U.S. Bureau of Labor Statistics, approximatelysix out of every 10 manufacturing workers who lost a job to trade and found new employmentexperienced wage reductions, most of more than 20 percent.24 More broadly, as increasingnumbers of workers displaced from manufacturing jobs joined the glut of workers competing fornon-offshorable, low-skill jobs in sectors such as hospitality and food service, real wages havealso fallen in these growing sectors as well since NAFTA.25The USITC report also got it wrong with respect to particular industry sectors.AGRICULTURE: The USITC projected that the United States would not gain or lose much inoverall agricultural production. In reality however, the U.S. agricultural trade balance withMexico worsened from a surplus of 1.4 billion in 1993 to a deficit of 3.2 billion in 2015.Agricultural trade flows under NAFTA vividly demonstrate the limitations of the USITC model.Not only were entire U.S. agricultural sectors, such as tomatoes, effectively wiped out, but thecowboy nation developed a significant trade deficit in beef with its NAFTA partners.The USITC concluded that NAFTA would result in little or no impact on meat imports into theUnited States because of already low U.S. tariff rates, and that if anything, U.S. exports of meatto Mexico would increase. The report projected that U.S. beef exports to Mexico would increasein the long-term by 16 percent or more. In reality, American cattle producers experienced theopposite outcome from NAFTA. In 1993, the United States exported 39,000 metric tons of beefand veal to Mexico and imported only 13,000 metric tons. By 2015, the United States importedmore than 30,000 metric tons of beef and veal from Mexico more than it exported to Mexico.26Billions of USD (adjusted for inflation)U.S. Trade Deficits with Canada and Mexico inBeef and Live CattleU.S. DeficitU.S. ExportsU.S. Imports543210-1 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015-2-3-4Source: U.S. International Trade Commission, “Interactive Tariff and Trade Dataweb,” 2016.AUTOS: U.S. manufactured goods also experienced worsening deficits under NAFTA. TheUSITC NAFTA report estimated that U.S. auto production would decrease slightly, but withoutdramatic impact on production or employment levels. The report partly based these findings onMay 20165

Public CitizenData Failoutside testimony. In autos, the three largest automakers in the United States ‒ Ford, GeneralMotors and Chrysler ‒ promised: “Rather than closing U.S. production plants to transferproduction to Mexico, U.S. Big Three automakers are generally planning to undertake at least apartial restructuring of their Mexican operations, although this strategy will vary by firm.”Prior to NAFTA, the United States had a trade deficit of 7.1 billion in passenger vehicles withNAFTA. In 2015, that deficit had grown to nearly 35 billion. Rather than reducing the tradedeficit, NAFTA increased it. Since 1994, the United States has had an average vehicle deficitwith Mexico of 1.2 million vehicles.A final lesson of the USITC NAFTA report is the perils of not including currency values inprojection models. (In this regard, NAFTA itself provides a stark warning of the threats posed toU.S. workers and firms of trade pacts that, like the TPP, fail to include enforceable currencyprovisions.) The USITC projections were premised on the relative values of the U.S. dollar andthe Mexico peso remaining constant. But literally weeks after NAFTA was approved by the U.S.Congress, Mexico devalued the peso 50 percent relative to the U.S. dollar. The effect of thedevaluation was to eliminate the benefit of most of the tariff cuts on U.S. goods entering Mexico.Indeed, many U.S. products became more expensive to import into Mexico post-devaluation thanthey had been pre-NAFTA tariff cuts. Yet, the USITC model, by not considering what many inCongress had warned was an imminent devaluation, assumed new market access for U.S. goodsthat was never forthcoming and underestimated the flood of Mexican imports into the UnitedStates spurred by the devaluation that made their dollar-denominated prices artificially low.China PNTRIn 1999, the United States signed a bilateral agreement with China regarding the terms forChina’s accession into the WTO.27 This agreement was the predicate for a congressional vote togrant China Permanent Most Favored Nation status, which was renamed Permanent NormalTrade Relations by those promoting the pact. Using 1998 as a base year, the USITC reportestimated that U.S. exports of goods and services to China would increase by 10.1 percent or 2.7 billion while imports from China into the United States would increase nearly seven percentor 4.4 billion. In sum, the USITC projected that China’s accession to the WTO under the termsof the bilateral and Congress’ approval of China PNTR would result in a small U.S. bilateraltrade deficit with China increase from 113 billion to 120 billion. The USITC report on China’sWTO accession includes repeated references to the agency not having sufficient time to projectthe outcomes for various sectors, and even U.S. employment effects.Yet even on the scope the report covered, again, the USITC totally miscalculated the outcomesIndeed, the USITC projections seems altogether unrelated to the actual outcomes of U.S.-Chinatrade since the bilateral agreement and China’s WTO accession.In 2000, the U.S. trade deficit in goods and services with China was 113 billion.28 By 2015 itwas 340 billion, which is larger than the economies of most nations, including Denmark, NewZealand and Malaysia. U.S. exports to China increased from 29.7 billion in 2000 to 163 billionin 2015. U.S. imports from China were already 143 billion in 1999. Today, Chinese importshave increased to half a trillion dollars at 502.6 billion.29May 20166

Public CitizenData FailU.S-China Trade in Goods and Services2000 - BaselineUSITC Projection2015 – Actual 113 billion deficit 120 billion deficit 340 billion deficitBillions of USD (adjusted for inflation)U.S. Goods and Services Trade Deficit with ChinaU.S. DeficitU.S. ExportsU.S. Imports6004002000-200-400Source: U.S. Bureau of Economic AnalysisMANUFACTURED GOODS: The USITC also again got the sectoral projections it did includewrong. The report projected that capital-intensive exports to China such as manufactured goodswould increase by more than 300 million a year after textile and apparel quota restrictions wereeliminated in 2005. The United States exported 17.8 billion in manufactured goods to Chinaand imported 134.2 billion in 2000. The U.S. had a trade deficit of 116 billion. Since then,U.S. exports have increased to 81.8 billion while U.S. imports of Chinese manufactured goodsreached 468 billion in 2015.IRON AND STEEL: The USITC report also estimated that U.S. exports of iron and steel wouldincrease by 5.1 percent. The report does not project changes in import levels. In reality, U.S.exports of iron and steel increased by 1.1 billion or 239 percent. The USITC report did nothowever anticipate that U.S. imports from China of iron and steel would increase by 12.3billion or by nearly 300 percent. The U.S. trade deficit with China in steel and iron products hasworsened by nearly 7.9 billion, increasing from 2.7 billion in 2000 to 10.7 billion in 2015. InNovember 2015, nine steel associations wrote a joint letter insisting that China’s“overwhelmingly state-owned and state-supported steel industry” is the root problem of the 700million metric tons of excess steel capacity in the world today, which is making it difficult forprivate sector firms in the U.S. to compete.”30TEXTILES AND APPAREL: Due to time constraints that the USITC faced when doing thisstudy, the USITC mentions that it would be unable to provide “estimates of changes in U.S.textile and apparel production, employment, imports, and exports.” Nonetheless, the USITCprovided an estimate for changes in total U.S. export and import levels to the world as a result ofMay 20167

Public CitizenData Failgranting China ascension into the WTO. The bilateral trade deficit of textiles and apparel in theUnited States was estimated to gradually worsen over time. Although exports would increase by 2.4 billion by 2010, imports would outpace this growth by increasing 6.1 billion. The USITChowever emphasized that “the estimated changes in U.S. imports may be overstated,” whichprovided some leeway for government officials to defend this policy decision. By 2010, U.S.exports of textiles and apparel to the world decreased by 37 percent. Imports increased by 14.6billion.Although the USITC report notes that the agency was unable to estimate the impact on U.S.employment as a result of China joining the WTO due to time constraints, EPI released a studyin 2014 that found that the goods trade deficit with China has resulted in 3.2 million eliminatedor displaced jobs in the United States.31U.S.-Korea FTAThe U.S.-Korea FTA was completed in 2007, but did not enter into force until 2012. In its initial2007 assessment, the USITC estimated that the agreement would eventually reduce the U.S.trade deficit of goods with Korea by increasing U.S. goods exports from 30 to 33 percent. TheUSITC projected that U.S. goods imports from Korea would grow much more slowly ‒ from 1432to 15 percent.In 2011, before the agreement went into effect, U.S. exports of goods to Korea were 43.7billion and imports from Korea into the United States were 59.3 billion. In 2015, U.S. exportsof goods to Korea were 41.2 billion and imports from Korea into the United States were 69.7billion. The U.S. trade deficit with Korea in goods has increased by more than 12.9

113 billion deficit 120 billion deficit 340 billion deficit U.S.-Korea FTA: Trade in Goods 2011 - Baseline USITC Projection 2015 - Actual 15.6 billion deficit 10.6 billion deficit 28.5 billion deficit Source: U.S. Bureau of Economic Analysis. The divergence between the USITC’s rosy projections and trade pacts’ negative outcomes has