The Mexican trade liberalization process and its net effects on employment:

Transcription

1 The Mexica trade liberalizatio process ad its et effects o employmet: Rosario Cervates, Gerardo Fujii 1 Abstract: This paper examies the direct ad idirect effects of trade liberalizatio o employmet i Mexico. First, we estimate the et effect that export ad import growth has had o employmet; ad secod, iasmuch as Mexico is a coutry relatively abudat i uskilled labor, we seek to verify whether the most dyamic sectors i terms of employmet creatio have bee those expected by trade reform promoters, i.e., tradable-goods sectors such as maufacturig, where uskilled labor is used more itesively. We fid that, betwee 1988 ad 2004, the et balace of job creatio by foreig trade was positive ad icreasig util Uskilledlabor-itesive maufactures ad o-tradable goods ad services (i a idirect way) sectors have bee the mai cotributors to employmet growth associated with foreig trade. The et effect of foreig trade o employmet teds to be egative i skilled labor-itesive maufacturig sectors. Resume: E este trabajo se aaliza los efectos directos e idirectos de la liberalizació comercial e el empleo, e México. Primero, estimamos el efecto eto que el crecimieto de las exportacioes y las importacioes ha teido e el empleo; y, e segudo lugar, dado que México es u país relativamete abudate e trabajo o calificado, buscamos verificar si los sectores más diámicos e térmios de creació de empleo ha sido aquéllos esperados por los promotores de la reforma comercial: sectores de biees trasables tales como las maufacturas itesivas e trabajo o calificado. Ecotramos que, etre 1988 y 2004, el balace eto de creació de empleo asociado al comercio iteracioal fue positivo y creciete hasta el año Las maufacturas itesivas y los biees y servicios o trasables (de maera idirecta) ha sido los sectores pricipales e cotribuir al crecimieto del empleo asociado al comercio exterior. El efecto eto del comercio exterior e el empleo tiede a ser egativo e los sectores maufactureros itesivos e trabajo calificado. Keywords: Trade Liberalizatio; Trade ad Labor Market Iteractios; Employmet; Mexico. jel Classificatio: F13, F16. 1 Faculty Profesors, Ecoomics-Admiistrative Scieces, Uiversity Ceter, Uiversity of Guadalajara, ad Faculty of Ecoomics, Natioal Autoomous Uiversity of Mexico. s: mariac@cucea.udg.mx ad fujii@servidor.uam.mx. This paper has bee supported by the Natioal Autoomous Uiversity of Mexico with the assistace of the Departmet of Posgraduate Studies ad the Departmet of Academic Persoel, through PAPIIT Project No. IN

2 82 EcooQuatum Vol. 9. Núm. 2 Recepció: 31/10/2011 Aceptació: 17/05/2012 Itroductio Free trade assumes that if you throw me out of work i oe directio you re-employ them i aother. As soo as that lik i the chai is broke the whole of the free trade argumet breaks dow. J. M. Keyes Although it is well kow that the structural reform programs followed by Lati America ecoomies did ot result i higher ecoomic growth rates, a decrease i volatility, or a lower exposure to risk (BID, 1997; Stalligs ad Peres, 2000; Frech- Davis, 2005), the effects that this policy had o employmet ad its log-term structure are less kow. Oe of the mai traits of the ecoomic reforms was trade liberalizatio. I Mexico, this process took place rapidly. As a result, Mexica export ad import coefficiets experieced accelerated growth. It was expected that, give the relative stregth of resources i the Mexica ecoomy, trade liberalizatio would geerate a icrease i the employmet of uskilled labor, which would be partially offset by a declie i skilled labor. Therefore, the et effect for the coutry would be positive, ad would additioally lead to a reductio i differetial wages betwee these two groups of workers. Most iteratioal trade theories suggest that free trade will lead to a patter of specializatio i productio. Sice Adam Smith to the most recet theories, free trade affects labor markets through a process of resource reallocatio. I the static models of these theories, productio will icrease i some sectors ad decrease i others ad employmet demad will follow this patter. Surveys that compile studies regardig the relatioship betwee trade ad employmet, amog others, are Hoekma ad Witers (2005), Wood (1994) ad Cervates (2008). There are several studies attemptig to establish the effect of trade liberalizatio o employmet i Mexico. Research has bee coducted o the effect of the policy chage o employmet geeratio through growth i the export sector ad the chages i demad by type of employmet ad their impact o wage dispersio. I geeral, betwee 1970 ad 1992, export growth did ot icrease labor demad (Dussel, 1995, 2003). Relyig o ecoometric techiques, Dussel (1995) shows that, i terms of employmet creatio, the exports coefficiet is statistically o-sigificat for those sectors of the ecoomy displayig the highest employmet growth rates. However, based o the iput-output aalysis, Ruiz-Nápoles (2004) fids that the share of total employmet geerated by exports betwee 1980 ad 2000 has actually icreased. I the same way, it has bee established that free trade caused a higher relative growth i the demad for skilled labor, which cotributed to a growig wage gap (Cragg ad Epelbaum, 1996; Revega, 1997; ad Feestra ad Haso, 1997). I this paper, we

3 The mexica trade liberalizatio process will show that whe estimatig the foreig trade factor compositio (icludig the idirect effects o labor demad), opposite balaces are revealed: the uskilled labor embodied i exports is more sigificat both i absolute ad i relative terms whereas there is more skilled labor embodied i imports. Hece, by relyig upo a alterative approach, we coclude that the wage gap icrease is ot likely to be associated with trade liberalizatio ad ot ecessarily with FDI. Accordig to the Natioal Accout System, i Mexico, the maquiladora idustry 2 uses up to three times more productio workers tha the o-maquiladora maufacturig sectors (Fujii ad Cervates, 2010). Hece, the aim of this paper is to preset a estimate of the effect that export ad import growth, owig to trade liberalizatio, has had o employmet i Mexico. Regardig the employmet geerated by exports, this method eables a estimate of the direct as well as idirect effects of job creatio resultig from foreig trade. To the best of our kowledge, this is the first time that iput-output aalysis is utilized for the Mexica case i order to estimate the flows ad balaces of labor demad for differet idustrial sectors, as well as accordig to the labor itesity of uskilled or skilled labor. The estimate is based o data from the Mexica System of Natioal Accouts ad o State Matrix Iput-Output Matrices prepared by Cosultoría Iteracioal Especializada, S. A. The paper is orgaized as follows: sectio oe outlies a backgroud of the evolutio of foreig trade i Mexico sice trade liberalizatio; the secod sectio describes the iput-output methodology used to calculate the employmet cotet of exports ad imports; the third sectio presets the results of these estimates. I the last sectio, a coclusio icludig a brief summary of our fidigs is preseted. Dyamics of foreig trade sice trade liberalizatio Before we describe the evolutio of the Mexica trade variables, briefly, we will explai how free trade could affect the structure of labor markets. Sice Adam Smith, the free trade theory has evolved from very restrictive assumptios to more realistic models. I the ew trade theory, it is possible for a coutry to egage i trade eve though the coutry has ot comparative advatage, due to icreasig returs i some maufactures. Regardless of the reaso why coutries trade ad why it is possible for them to gai from trade, most theories predict some level of specializatio i productio. This specializatio i the short ru follows after a chage i the structure of the labor market, measured as the share of employmet by sector. The latter sice some sectors will demad relatively more labor tha others i order to satisfy their expadig demad ad because some sectors will lose participatio i the domestic market. Noetheless, sice the aalysis covers a period of thirtee years, it must be recogized that there are several factors that affect labor demad i the log ru. Such 2 The Maquiladora idustry or maquila cosists of foreig owed assembly plats that re-export materials ad equipmet that have bee exported from the U.S. or other coutries to them (Truett ad Truett, 1984: 45-46).

4 84 EcooQuatum Vol. 9. Núm. 2 factors iclude iter alia, techological chage, a icrease i labor productivity, ad chages i the structure of domestic demad. I fact, there are may studies that fid that i developed ecoomies, the relative fall i uskilled labor demad is explaied by techological chage ad ot by the trade with developig coutries (Lawrece, 1996; Morriso Paul ad Siegel, 2001; Boyle ad MacCormack, 2002). O the other had, Wood (1994), with a factor cotet of trade methodology, estimates that trade would reduce the employmet of uskilled workers i developed coutries ad icrease the employmet of these i the developig ecoomies. Util the early 1980s, Mexica idustrial policy aimed at promotig the coutry s idustrializatio. As a result, such policy was import-substitutio orieted: the Mexica market was protected from foreig competitio. The begiig of the trade liberalizatio process i Mexico ca be dated back to 1987, the year of Mexico s accessio to GATT. Such accessio meat that official prices ad import liceses were substituted for taxes o imports which were oly temporary, i.e., they were to be reduced later o. The process of tariffs reductio was a rather quick oe. Cosequetly, i 1980 the average import tariff was 22.8 percet whereas, by 1988, it came dow to 10.2 percet. Similarly, i 1988 the highest tariff was cut dow from 100 to 20 percet. I additio, the umber of products subject to import quotas fell from 1,200 to 325. Moreover, by 1996, the weighed average import tax was reduced to 3 percet, whereas the umber of tariff levels fell from 11 i 1986 to 5 i the subsequet years. I the same way, owig to the North America Free Trade Agreemet (NAFTA), which started off i 1994, 41 percet of Mexica imports from the Uited States ad Caada were tariff-free by such year (Dussel, 2000: 86; Clavijo, 2000: Cuadro A 26). As the Mexica govermet opeed up the ecoomy to imports, it also set i place a umber of programs desiged to promote exports, particularly those associated to the maufacturig sector. Betwee 1992 ad 2006, Mexico s total exports experieced a remarkable growth: from 46.2 to 250 billio dollars. From 1989 to 2006, the average aual growth rate of exports was 9.6%; from 1989 to 1993, 5.8%; ad from 1994 to 2006, 11.1%. This led to a icrease, betwee 1988 ad 2006, i the trade opeess coefficiet, from 25.5 to 85.7% 3. As Table 1 shows, sice 1994, whe NAFTA came ito effect, the trade opeess coefficiet has icreased dramatically. The dyamism of exports coicided with a chage i the structure of exported goods. By the late 1980s, the share of maufacturig exports was 71% of total exports (1988). Maufacturig exports have bee supported by several govermet iitiatives. For istace, the Programa de la Idustria Maquiladora para la Exportació (Maquiladora Idustry Exportig Program), created i 1965, was later o complemeted by PITEX (Programa para la Importació Temporal para Producir Artículos para Exportació). PITEX allows Mexica producers who import iputs i order to produce exports to brig such iputs ito the coutry tariff-free; they do ot pay value-added tax either. The objective of these programs was to icorporate Mexico ito the iteratioal 3 These coefficiets were estimated usig exports, imports ad GDP aual series i costat Mexica pesos. Whe we estimate trade coefficiets usig curret values i US dollars (or values i Mexica pesos usig real exchage rates), the proportio of foreig trade to GDP seems sigificatly lower.

5 The mexica trade liberalizatio process Table 1 Mexico. Trade opeess coefficiet, (% of GDP) Exports (X)/GDP Imports(M)/GDP (X+M)/GDP Total variatio (X+M)/GDP Aual average variatio (X+M)/GDP Source: Authors calculatios based o INEGI, Ecoomic Iformatio Bak. Joit Ecoomic Idicators. Supply ad demad of goods ad services productio etwork by takig advatage of NAFTA ad of the atural Mexico-USA viciity. I order to ecourage ivestmets i maufacturig ad maquila activities, Mexico emphasized its large labor force ad its correspodet low cost, amog other advatages for busiesses. As it is well kow, labor cost gaps, both i wage ad o-wage terms, are likely to result i the fragmetatio of the productio process accordig to the advatages each coutry has regardig the process itself. I this way, whe the product is to be assembled, a phase of productio which ormally relies o uskilled labor, it teds to be the case that this particular activity takes place i lowwage coutries whereas the desig ad egieerig stages ted to be developed i those coutries where labor is well qualified (ad wages are higher) 4. Some sectors i the maquiladora -such as the electroic idustry- are clearly part of this fragmetatio, sice they import semi-maufactured iputs, ad most of the tasks performed i the coutry cosist of assembly. I additio, the Mexica govermet set i place the Programa para Empresas Altamete Exportadoras (Program for Heavy Exportig Firms), which grats admiistrative support ad assistace to this eterprises, such as the quick reimbursemet of ay value-added tax they have paid ad have the right to claim for, as well as speedig up ispectios at customs check poits (Dussel, 2000: 87). The performace of the maquiladora export idustry has bee remarkable. Actually, by the late 1980s, almost half of the Mexica maufacturig exports origiated i that idustry. Maquila exports also represeted 33 percet of total exports, whereas the o-maquila idustry cotributed with 38 percet of such total. Toward 2006, the maquila share over total exports icreased to 45 per cet while the o-maquila oe decreased to 36 percet. Figure 1 shows the growth of maufacturig exports, both maquila ad o-maquila, from 1992 to 2006, which is the last year for which separate data for maquila ad o-maquila maufacturig are available. From 1998, the value of maquila exports has bee higher tha that of the rest of the maufacturig exports. I this way, the maquila idustry became the most dyamic sector of the etire Mexica maufacturig export idustry. I fact, i 2006, Mexica maufacturig exports amouted to 203 billio US dollars, out of which 112 billio (well over 50 percet of 4 Grossma ad Rossi-Hasberg (2008) ad Baldwi (2006), amog others, state that off - shorig could happe for o-productio activities.

6 86 EcooQuatum Vol. 9. Núm. 2 the total) correspoded to maquila exports. Hece, Mexico has become a coutry that specializes i the stages of the productio process which are highly itesive i the use of low-skilled labor ad, simultaeously, a importat exporter: Mexico is a major exporter amog developig coutries of maufactured goods, such as textiles ad clothig, automobiles ad automotive parts, ad electrical ad electroics goods, which have bee very importat i iteratioal productio etworks (UNCTAD, 2007: 71). Figure 1 Maufacturig ad maquila exports (billios of dollars) Total Maquila exports No-maquila exports Bilios of dollars Source: Authors calculatios based o INEGI, Ecoomic Iformatio Bak. It is reasoable to assume that the strikig growth of foreig trade as a share of the ecoomy ad the chage i the structure of exports must have had a substatial effect o employmet. O the oe had, export growth creates jobs i direct ad idirect ways while, o the other had, imports, isofar as they replace domestic productio, are either job-destroyig or they evice the potetial to geerate employmet to the extet that they are replaced by domestic products. Employmet shares throughout the ecoomy chaged sigificatly i the period. I this way, the participatio i total employmet of the sectors producig tradable goods fell from 41.3 per cet i 1988 to 33.1 i This is basically a result of a fall i agricultural employmet, whose share dropped from 28 to 21 per cet. The share of employmet i the maufacturig sector also fell: i 1988, it represeted 13.4 per cet of total employmet; by 2004 it was oly 11.4 per cet. Simultaeously, the compositio of maufacturig employmet chaged i regards to the levels of qualificatio it actually demads. I this way, whereas the employmet share of highly-skilled labor itesive maufactures fell from 5.9 to 4.2 percet, that belogig to the low-skilled labor itesive maufactures icreased from 6.4 to 7.2 percet. This higher share for the low-skilled labor sectors is etirely explaied by the behavior of the maquila idustry: its employmet share icreased from 1.5 per cet of the total to 3.5 percet. At the same time, the share of other low-skilled labor itesive maufactures dropped.

7 The mexica trade liberalizatio process Iput-output aalysis applied to the estimate of labor cotet of foreig trade Iput-output methodology eables us to estimate total employmet (direct ad idirect) associated with the productio of a certai good. So the aalysis preseted i this paper ca be related to Vaek s (1968) predictio i the sese that we assume that the factor cotet of Mexica trade should reflect the relative factor edowmets of the coutry. As Trefler ad Zhu (2010) show, the iput-output matrices are useful tools to estimate the factor cotet of trade for short periods. Nevertheless, whe applyig such methodology to the chages i foreig trade patters i Mexico some limitatios arise. Natioal Accouts classificatios are made by type of product, meaig that small, medium, ad large eterprises that are heterogeeous i forms of productio ad productivity levels are classified i the same sector. Therefore, the observed averages i techical coefficiets of atioal iputs ad employmet may ot be represetative if eterprises respod i differet ways to chages i trade policy. However, the major limitatio of the methodology proposed i this study lies o the assumptios uderlyig the estimates of employmet associated with imports. First, it is ecessary to assume that imported goods could be produced iterally usig the same combiatio of iputs ad factors as those already used i domestic productio. Moreover, i order to estimate the job-destructio effect of imports, it ca be assumed that imports completely replace domestic productio, or that oly some of them are job-destroyig as a result of the chage i trade policy. Thus, there are two types of problems: 1) how would the goods imported be produced i Mexico? ad 2) what is the proportio of imports that could realistically be produced domestically? Assumig that imports destroy the same umber of jobs that they create i their coutry of origi is ot reasoable. This is the case because oe of the basis of iteratioal trade is productivity differeces. Therefore, we will retai the assumptio that imported goods would be produced i Mexico accordig to the techology ad forms of productio already foud i the coutry. However, although it is extremely urealistic to assume that all of the products curretly imported could be produced i Mexico, i this study we preset oly oe type of the estimates to measure the impact of import growth o employmet ad its mobility amog sectors. We calculate the total labor cotet of imports, direct ad idirect, with the aim of verifyig whether there is a tedecy to import more skilled labor-itesive goods after trade liberalizatio. More precise estimates could be obtaied either by assumig that the growth i import coefficiets has oly a job destructio effect or from a deeper study of the productio capabilities of the Mexica ecoomy. This sectio presets the methodology used to calculate the employmet cotet of exports ad of imports. From iput-output aalysis we ca calculate the labor cotet of foreig trade. However, certai assumptios are ecessary. First, we assume there are o sigificat differeces i the forms of productio of exported goods compared with their productio for iteral cosumptio; ad secod, we suppose that imported goods (fial or itermediate products) are perfect substitutes for atioal productio, i.e., that

8 88 EcooQuatum Vol. 9. Núm. 2 they could be produced with a idetical combiatio of iputs ad factors. The output associated with the volume of exports ad imports writte i matrix otatio is: (1) xe = ^I - A h -1 f d d e d d (2) xm = ^I - A h -1 f m d where x e ad x d m are total productio vectors (direct ad idirect) associated with exports, f e, ad with the imported goods, f m, ad where ^I - A d h -1 is the Leotief iverse matrix, with I as the idetity matrix of dimesio, rxr, where r is the umber of ecoomic braches or sectors ad A d is the techical coefficiet matrix, which represets the proportio of domestic iputs eeded to produced a uit of a give good. The total labor cotet of exports ad imports is obtaied by multiplyig the labor coefficiet vector by the gross output associated with tradable goods. (3) e = my t (4) m = my t (5) m = / x where m is the vector for employmet by sector; is the vector whose elemets are give by the umber of workers i each sector; x is the gross product by sector; ad fially Yt e ad Yt m, which are the diagoal matrices of the gross value of exports ad imports, i.e., each sector icludes a estimate of the cotet of traded iputs ad ot oly the value added. To separate direct from total employmet, the labor coefficiet vector is multiplied by the diagoal matrices of the value of exports ad imports, E t ad M t. (6) le = me t (7) lm = mm t e m Employmet creatio ad destructio due to foreig trade Relyig o the method outlied above, i this sectio we preset the estimate of the effects o employmet derived from growth i foreig trade betwee 1988 ad 2004, 5 as well as its distributio by ecoomic sector. First, we submit estimatio by ecoomic braches of direct ad idirect employmet geerated by the maufacturig idustry 5 Sice the last iput-output matrix estimated for the Mexica ecoomy -based o the Natioal Accout System classificatio- is for the year 2000, we are uable to estimate more recet levels of employmet associated with trade flows.

9 The mexica trade liberalizatio process exports accordig to their labor itesity (sectio Labor cotet of exports). The, we itroduce the total (direct ad idirect) labor cotet of imports (sectio Labor cotet of imports). I sectio Balace betwee labor cotet of imports ad exports, we preset the balace betwee the labor cotet of exports ad that of imports: this is the et result of the effects of foreig trade o employmet ad its compositio owig to the process of trade liberalizatio. Labor cotet of exports Table 2 shows the direct ad idirect labor cotet of exports from 1988 to 2004, estimated usig the method previously discussed. Key results are as follows: a) The umber of paid jobs associated with exports, both directly ad idirectly, grew remarkably betwee 1988 ad 2004: from 2 millio to almost 5 millio. Hece, the share of total employmet attached to exports icreased from 8.8 to 16.1 per cet. b) Most of the additioal 2.94 millio jobs derived from export growth are explaied by maufacturig exports, whose cotet of labor icreased by 1.4 millio, from 697,000 jobs i the first year to 2.1 millio i c) The greater part of this expasio resulted from the growth of employmet i the uskilled-labor-itesive export maufacturig sector, 6 whose umber of jobs icreased from 538,000 i 1998 to 1.7 millio i The employmet share of uskilled-labor-itesive maufacturig exports, relative to total employmet associated with exports, icreased from 27% to 34%. d) Employmet i these activities reached its maximum level i 2000, whe it amouted to 1.9 millio jobs. Give the weight of employmet associated with exports of the products of these activities, relative to total jobs derived from exports, 2000 saw the maximum level of total employmet derived from export activity. e) Direct ad idirect employmet derived from skilled labor-itesive maufacturig exports icreased from 159,000 jobs i 1988 to 408,000 i However, give that employmet grew much faster i other export sectors, particularly uskilledlabor-itesive maufacturig, i relative terms the proportio of skilled labor to total employmet associated with exports remaied at 8%. f) Fially, exports growth has also affected employmet o two other o-qualifiedlabor itesive sectors: the umber of paid jobs uder the Other goods ad services category icreased from half a millio i 1988 to 1.6 millio i I the same period, agricultural employmet expaded from 641,000 to 1.2 millio posts. I Table 2, the total labor cotet of exports is preseted for both direct ad idirect employmet. 7 The most sigificat fidigs are the followig: 6 The uskilled-labor-itesive sector icludes all the maufacturig braches with a ratio of more tha three productio workers per oproductio worker. 7 The direct employmet i oe sector represets the cotet of employmet required to produce a give amout of exports i the same sector. O the other had, the idirect employmet i oe sector represets the cotet of employmet required to produce the iputs for the total amout of exports. Thus, i Table 2 the row of idirect employmet for the maquila idustry is empty because there is ot a sigle sector demadig i-

10 90 EcooQuatum Vol. 9. Núm. 2 a) The direct employmet cotet of exports is three times that of idirect employmet i almost all of the years cosidered. This reflects the relatively weak productio likages of export activities, particularly those of maquila idustry. I additio, i , the ratio of direct over idirect employmet geerated by exports icreased from 3.1 to 3.4, which implies that employmet likages associated to exports are turig weaker. This is particularly relevat for the maquiladora idustry, sice 90 per cet of the iputs it utilizes are imported, which implies that this idustry s idirect effects o employmet are almost trivial. I this way, the patter sketched throughout the period shows a slight icrease i direct employmet per idirect employmet. Nevertheless, it must be highlighted that durig , the ratio of direct over idirect employmet was about 2.6, which suggests that, somehow, the 1995 fiacial crisis might have geerated a process of import substitutio regardig itermediate iputs. b) The demad of idirect employmet has grow at almost the same rate as that of direct employmet. The average aual growth of direct employmet associated with exports, icludig maquila, is 5.8%; of direct employmet without maquila, 5.4%; ad of idirect employmet, 5.2%. c) The greater part of direct employmet geerated by exports derives from maufacturig exports, particularly those that are uskilled-labor-itesive, ad withi those, by maquila idustry exports, which o average use up to three times more productio workers per oproductio worker tha the rest of Mexica maufacturig. Figure 2 shows that maquila idustry suffered a relative loss i terms of the geeratio of employmet by exports, especially aroud the years whe NAFTA took effect; however, o average, more tha 65% of the employmet i uskilled-labor-itesive sectors associated with exports was attributable to maquila. d) Other importat ad expadig sectors for direct employmet geeratio are agricultural exports ad other goods ad services exports. 8 e) The growth of idirect employmet is maily cocetrated i uskilled-laboritesive sectors. O average, the aual growth rate of direct labor cotet of exports of uskilled-labor-itesive maufactures is greater tha 4.5%, while idirect employmet i these sectors grew at a average aual rate exceedig 14%. However, this huge icrease is partly explaied by oly oe fact: the idirect demad of 27,000 jobs for o-maquila exports i 1988 had rise to 256,000 by Give this expasio, the idirect labor cotet of exports i uskilled-laboritesive sectors rose from 14% of total employmet geerated by labor-itesive sectors to 42% (excludig employmet geerated by maquila idustry). Other export activities with importat idirect effects o employmet are agriculture ad those grouped i the category of other goods ad services. puts to them; furthermore, we are assumig that the maquila idustry does ot demad atioal iputs either, so the growth of maquila exports does ot affect the idirect employmet i ay other sector. 8 Sectors that export or import goods ad services i a direct way from the sector other goods ad service are: Electricity, Gas ad Water, Professioal Services, Recreatio Services ad Other Services

11 The mexica trade liberalizatio process Figure 2 Maquila employmet as proportio of labor cotet of exports (percet) Uskilled-labour itesive Maufacturig products Total Source: Authors calculatios based o INEGI, System of Natioal Accouts, ad Iput-Output matrices Table 2 Direct, Idirect ad Total Employmet o Exports (aual average; thousads of paid jobs) Total employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive No Maquila Maquila Skilled-labor itesive Other goods ad services Total Direct employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive No Maquila Maquila Skilled-labor itesive

12 92 EcooQuatum Vol. 9. Núm Other goods ad services Total Idirect employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive No Maquila Maquila Skilled-labor itesive Other goods ad services Total Source: Authors calculatios based o INEGI, System of Natioal Accouts, ad Iput-Output matrices Labor cotet of imports Table 3 shows total, direct ad idirect labor cotet of total imports assumig that those goods were produced i Mexico with the same combiatio of iputs ad factors as similar domestically-produced goods. This allows us to establish whether imported goods are maily skilled-labor-itesive, i accordace with the productio capacity ad techology of the coutry. The fact that this estimate is based o the techology used i Mexico to produce these goods results i a very high labor cotet of imports, owig to the low labor productivity i some sectors of the ecoomy. This is particularly strikig for the agricultural sector, as we will show below. Give their acceleratig growth, the labor cotet of imports icreased substatially i the period of study, from 1.35 millio to 3.85 millio paid jobs. The greatest absolute icreases occurred betwee 1996 ad 2000, i.e., i the years that followed the comig ito effect of NAFTA. I spite of the fact that the value of imports i the primary sector ever exceeded 4% of total imported goods durig the period, the low productivity i the agricultural ad livestock sector i Mexico, which has its couterpart i the high level of employmet cocetrated i these activities, meat that a large proportio of the jobs associated with imports was cocetrated i this sector, averagig 37% of the total. The total labor cotet i maufacturig imports grew from 529 to 1,640 thousad jobs. I 1988, skilled-labor-itesive maufactures represeted the secod-highest share of employmet associated with imports ad, to a lesser extet, the direct ad idirect labor cotet of uskilled-labor-itesive sectors was also sigificat. By the ed of the period the share of employmet was slightly higher i the uskilled-labor-itesive sectors. Whe separatig the idirect effect of imports, 9 it ca be appreciated that, betwee 1988 ad 2004, their average direct labor cotet was close to 60 per cet of the total labor associated to them. Similarly, durig the 1995 crash, the share of direct employmet 9 I order to separate the direct ad idirect effects of imports o employmet, we used the same method as i the case of exports, i.e., the direct employmet, that is estimated by multiplyig the employmet coefficiet by the value of imports, is subtracted from the total labor cotet of imports.

13 The mexica trade liberalizatio process i the total employmet liked to imports decreased. Such situatio ca be explaied by the devaluatio the Mexica peso suffered which, i tur, ecouraged the substitutio of imports, especially of fial goods. By usig the same techical coefficiets of atioal iputs, the methodology allows us to ifer that imported goods ted to be relatively less labor-itesive, i accordace with the forms of productio that exist i Mexico. Thus, eve if goods that have t bee classified as skilled-labor-itesive are imported, there is a tedecy to import goods that cotai relatively less labor. The methodology thus reveals that i a idirect way, import growth sigificatly affected the tertiary sector: the umber of jobs associated with it icreased by more tha a factor of I 1988, total imports represeted 207,000 idirect jobs i the other goods ad services sector; i 2004, the umber was close to 704,000. Marketig ad trasport services are icluded i this sector, however, the ecoomic brach most affected idirectly by import growth was professioal services, where idirect employmet grew more tha tefold i the period, icreasig from 7,000 to 73,000 jobs. There is a low effect of imports o direct employmet i the tertiary sector because Mexico imports oly small values of services i the electricity, gas, water, professioal services ad etertaimet services braches. Actually, the share of imports belogig to the latter represet less tha oe percet of total imports. Table 3 Direct, idirect, ad total employmet associated with imports (aual average; thousads of paid jobs) Total employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive Skilled-labor itesive Other goods ad services Total Direct employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive Skilled-labor itesive Other goods ad services Total Although the methodology used does ot allow us to kow the origi of the iputs cotaied i imports, calculatio of the estimate of idirect employmet associated with imports shows that if the volume of imported maufactured goods icreases sigificatly, this will affect the productive chais.

14 94 EcooQuatum Vol. 9. Núm Idirect employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive Skilled-labor itesive Other goods ad services Total Source: authors based o INEGI, System of Natioal Accouts ad Iput-Output matrices Balace betwee labor cotet of imports ad exports The balace of the effect of foreig trade o labor demad is summarized i Table 4. a) Throughout the period aalyzed, the balace was positive. After 1995, whe the coutry s trade liberalizatio coefficiet icreased dramatically, foreig trade had a importat positive et effect o employmet. By the ed of the period, i 2001 ad 2002, this surplus had falle sharply, but it recovered i 2003 ad b) Of all the categories of tradable goods represeted i Table 4, maufactured goods caused the greatest positive balace i the employmet geerated. I the rest of the post-nafta period, with a egative employmet balace betwee 1990 ad 1994, et job creatio derived from foreig trade i maufactured goods averaged more tha 430,000 jobs. c) Such balace is explaied by the employmet effect derived from foreig trade i uskilled-labor-itesive maufactures, which was positive ad icreasig, i some years approachig oe millio jobs. However, toward the ed of the studied period, this balace teded to fall, i part because of the displacemet of Mexica uskilled-labor-itesive maufacturig exports to the Uited States caused by the growth of Chiese exports (Bracho, 2003). d) The employmet balace of skilled-labor-itesive maufactures had bee egative but icreasig sice However, its magitude was cosiderably less tha the positive balace of employmet derived from foreig trade i uskilled-laboritesive maufactures. e) The positive et effect of employmet derived from foreig trade i maufactures was sigificatly offset by the egative balace of agricultural employmet, which exceeded 400,000 jobs i the last years of the period studied, explaiig most of the drop i the balace of employmet derived from foreig trade. f) Table 4 shows as well that the positive balace of total labor cotet preset i the Other goods ad services sector augmeted importatly, a result that is maily explaied by the icrease i the direct employmet balace registered i such sector itself.

15 The mexica trade liberalizatio process Table 4 Balace betwee labor cotet of imports ad exports (aual average; thousads of paid jobs) Total employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive Skilled-labor itesive Other goods ad services Total Direct employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive Skilled-labor itesive Other goods ad services Total Idirect employmet Primary Products Agriculture, livestock, etc Maufacturig Products Uskilled-labor itesive Skilled-labor itesive Other goods ad services Total Source: Authors calculatios based o INEGI, System of Natioal Accouts, ad Iput-Output matrices It has bee show that trade opeess has had a et positive effect o employmet creatio i Mexico. However, these results must be assessed i the cotext of the challeges the coutry faces i terms of employmet eeds. We believe that the most pressig issue i the Mexica labor market is the quality of the jobs beig geerated. Actually, data from the Mexica Natioal Istitute for Geography ad Statistics, from the UN Ecoomic Commissio for Lati America ad from the Iteratioal Labor Orgaizatio reveals that more tha 25 per cet of workig Mexicas are employed i the udergroud ecoomy (2008); 46 per cet of urba employmet correspods to low productivity activities, which are i tur liked to microbusiesses ad low-skilled idepedet work (2006); the salaries of over 80% of wage-earers do ot exceed five times the miimum wage (the Mexica miimum wage lies i the viciity of 140 US dollars per moth (2009)); ad 45 per cet of wage-earers have o health isurace at

16 96 EcooQuatum Vol. 9. Núm. 2 all (2009). As it has bee discussed, most of the jobs geerated by trade liberalizatio are low-skilled-labor itesive. Hece, they ted to be of low quality. Coclusios Usig the iput-output aalysis, i this paper we have foud that the ecoomic orietatio toward a patter i which exports bega to display a growig importace for global demad had a sigificat effect o employmet, particularly i the period O the other had, trade liberalizatio led to a parallel growth i imports; however, the et effect of foreig trade o employmet has bee positive. Cofirmig Vaek s predictio, trade liberalizatio led to a restructurig of employmet towards a greater cocetratio i uskilled-labor-itesive maufacturig sectors, i.e., i activities which do ot ecessarily geerate good quality employmet. Nevertheless the positive effect o employmet, we are far from achievig a full employmet situatio for the Mexica ecoomy. So, ay ecoomic policy should cosider: a) how much employmet is ot beig geerated due to the icreasig imports of itermediate iputs; b) that if most of the employmet demad is for uskilled labor the most likely this employmet would have a low impact i the domestic market; ad c) that Mexico ow has bee displaced by Chia as the secod major exporter to the U.S. market. That is, we cosider it is crucial to desig a idustrial policy that favors the import substitutio of some iputs (those with a high level of value added), ad/or promotes a icrease i labor productivity i order to restore some of our competitiveess. As a cosequece, Mexica exportig sectors could geerate more employmet (due to the icrease of idirect effects of exports), better quality of this idirect employmet, ad/or a icrease i the wage rate. Refereces Baldwi, R. (2006). Globalisatio: the great ubudlig(s), Paper prepared for Fiish Prime Miister s Office for EU Presidecy ( September BID (1997). América Latia tras ua década de reformas, Washigto, D.C. Boyle, G. ad P. McCormack (2002). Trade ad techological explaatios for chages i sectoral labour demad i OECD ecoomies, Applied Ecoomics, 2002, 34, Bracho, G. (2003). Mexico s Foreig Trade Strategy i Trouble: The Impact of Chia, Workig Paper, Cetre for Mexica Studies, Uiversity of Oxford (December). Cervates, R. (2008). Apertura comercial y empleo e México, PhD Thesis, Faculty of Ecoomics, Natioal Autoomous Uiversity of Mexico. Clavijo, F. (comp.) (2000). Reformas ecoómicas e México , México: FCE, cuadro A 26.

17 The mexica trade liberalizatio process Cragg, M. I., Epelbaum, M. (1996). Why Has Wage Dispersio Grow i Mexico? Is the Icidece of Reforms or the Growig Demad for Skill?, Joural of Developmet Ecoomics, 51: Dussel, P. E. (1995). Recet Developmets i Mexica Employmet ad the Impact of NAFTA, Iteratioal Cotributios to Labor Studies, 5: Dussel, P. E. (2000). Polarizig Mexico. The Impact of Liberalizatio Strategy, Lye Rieer Publishers. Dussel, P. E. (2003). Características de las actividades geeradoras de empleo e la ecoomía mexicaa ( ), Ivestigació Ecoómica, Vol. LXIII, No. 243: Feestra, R., Haso, G. (1997). Foreig direct ivestmet ad relative wages: Evidece from Mexico s maquiladoras Joural of Iteratioal Ecoomics. May Tomo 42, Nº 3-4; pg. 371, Amsterdam, Holada. Ffrech-Davis, R. (2005). Reformas para América Latia después del fudametalismo eoliberal, Siglo XXI Editores Argetia, Bueos Aires. Fujii, G., Cervates, R. (2010). Liberalizació comercial y empleo e México, Revista de Ecoomía Mudial, 26, 2010, , Huelva, España. Grossma, G. M., Rossi-Hasberg, E. (2008). Tradig Tasks: A Simple Theory of Offshorig America Ecoomic Review. December Vol. 98, No. 5: Hoekma, B. ad A. Witers (2005). Trade ad Employmet: Stylized Facts ad Research Fidigs, World Bak Policy Research Workig Paper INEGI, System of Natioal Accouts, Baco de iformació ecoómica. dgcesyp.iegi.org.mx/cgi-wi/bdieitsi.exe/nivr15#arbol Lawrece, Robert Z. (1996). Sigle World Divided Natios? Iteratioal Trade ad OECD Labor Markets. Brookigs Istitutio Press. Morriso Paul, C. J. ad Siegel, D. S. (2001). The Impacts of Techology, Trade ad Outsourcig o Employmet ad Labor Compositio. The Scadiavia Joural of Ecoomics, 103: Revega, A. (1997). Employmet ad Wage Effects of Trade Liberalizatio: The Case of Mexica Maufacturig, Joural of Labor Ecoomics, 15, o.s3:s20-s43. Ruiz-Nápoles, P. (2004). Exports, Growth, ad Employmet i Mexico, , Joural of Post Keyesia Ecoomics, Vol. 27, No.1: Stalligs, B., Peres, W. (2000). Crecimieto empleo y equidad. El impacto de las reformas ecoómicas e América Latia y el Caribe, FCE, Satiago. Trefler, D., Zhu, S. C. (2010). The Structure of Factor Cotet Predictios, Joural of Iteratioal Ecoomics, 82, Truett, B. D., Truett, L. J. (1984). The Maquiladoras: Prospects for Mexico, Joural of Ecoomic Developmet, December, 9 (2), UNCTAD (2007). Trade ad Developmet Report, 2007, New York ad Geeva. Vaek, J. (1968). The Factor Proportios Theory: The N-Factor Case, Kyklos, 21, , October. Wood, A. (1994). North-South Trade, Employmet ad Iequality: Chagig Fortues i a Skill Drive World. Oxford: Claredo Press.