Local labor - market effects of NAFTA in Mexico: Evidence from Mexican commuting zones

T ach enbe g, Danielle Wo king Pape Local labo - ma ke e ec s o NAFTA in Mexico: E idence om Mexican commu ing zones IDB Wo king Pape Se ies, No. IDB-WP-1078 P o ided in Coope a ion wi h: In e -Ame ican De elopmen Bank (IDB), Washing on, DC Sugges ed Ci a ion: T ach enbe g, Danielle (2019) : Local labo - ma ke e ec s o NAFTA in Mexico: E idence om Mexican commu ing zones, IDB Wo king Pape Se ies, No. IDB-WP-1078, In e - Ame ican De elopmen Bank (IDB), Washing on, DC, h ps://doi.o g/10.18235/0002013 This Ve sion is a ailable a : h ps://hdl.handle.ne /10419/208212 S anda d-Nu zungsbedingungen: Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen Zwecken und zum P i a geb auch gespeiche und kopie we den. Sie dü en die Dokumen e nich ü ö en liche ode komme zielle Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich machen, e eiben ode ande wei ig nu zen. So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen (insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en, gel en abweichend on diesen Nu zungsbedingungen die in de do genann en Lizenz gewäh en Nu zungs ech e. Te ms o use: Documen s in EconS o may be sa ed and copied o you pe sonal and schola ly pu poses. You a e no o copy documen s o public o comme cial pu poses, o exhibi he documen s publicly, o make hem publicly a ailable on he in e ne , o o dis ibu e o o he wise use he documen s in public. I he documen s ha e been made a ailable unde an Open Con en Licence (especially C ea i e Commons Licences), you may exe cise u he usage igh s as speci ied in he indica ed licence. h ps://c ea i ecommons.o g/licenses/by-nc-nd/3.0/igo/legalcode Local Labo -Ma ke E ec s o NAFTA in Mexico: E idence om Mexican Commu ing Zones Danielle T ach enbe g IDB WORKING PAPER SERIES Nº IDB-WP-01078 In e -Ame ican De elopmen Bank In eg a ion and T ade Sec o No embe 2019 Local Labo -Ma ke E ec s o NAFTA in Mexico: E idence om Mexican Commu ing Zones Danielle T ach enbe g In e -Ame ican De elopmen Bank In eg a ion and T ade Sec o No embe 2019 Ca aloging-in-Publica ion da a p o ided by he In e -Ame ican De elopmen Bank Felipe He e a Lib a y T ach enbe g, Danielle. Local labo -ma ke e ec s o NAFTA in Mexico: e idence om Mexican Commu ing Zones / Danielle T ach enbe g. p. cm. — (IDB Wo king Pape Se ies ; 1078) Includes bibliog aphic e e ences. 1. Labo ma ke -Mexico. 2. Job c ea ion-Mexico. 3. Wages-Mexico. 4. Impo s- Mexico. 5. F ee ade-No h Ame ica. 6. In e na ional ade. 7. No h Ame ica-Comme cial ea ies. I. In e -Ame ican De elopmen Bank. In eg a ion and T ade Sec o . II. Ti le. III. Se ies. IDB-WP-1078 JEL Codes: F14, F16, F66, J23 Keywo ds: In e na ional ade, NAFTA, impo compe i ion, local labo ma ke s h p://www.iadb.o g Copy igh © 2019 In e -Ame ican De elopmen Bank. This wo k is licensed unde a C ea i e Commons IGO 3.0 A ibu ion- NonComme cial -NoDe i a i es (CC-IGO BY-NC-ND 3.0 IGO) license (h p://c ea i ecommons.o g/licenses/by-nc- nd/3.0/igo/legalcode ) and may be ep oduced wi h a ibu ion o he IDB and o any non-comme cial pu pose, as p o ided below. No de i a i e wo k is allowed. Any dispu e ela ed o he use o he wo ks o he IDB ha canno be se led amicably shall be submi ed o a bi a ion pu suan o he UNCITRAL ules. The use o he IDB’s name o any pu pose o he han o a ibu ion, and he use o IDB’s logo shall be subjec o a sepa a e w i en license ag eemen be wee n he IDB and he use and is no au ho ized as pa o his CC-IGO license. Following a pee e iew p ocess, and wi h p e ious w i en consen by he In e -Ame ican De elopmen Bank (IDB), a e ised e sion o his wo k may also be ep oduced in any acade mic jou nal, including hose indexed by he Ame ican Economic Associa ion’s EconLi , p o ided ha he IDB is c edi ed and ha he au ho (s) ecei e no income om he publica ion. The e o e, he es ic ion o ecei e income om such publica ion shall o nly ex end o he publica ion’s au ho (s). Wi h ega d o such es ic ion, in case o any inconsis ency be ween he C ea i e Commons IGO 3.0 A ibu ion -NonComme cial- NoDe i a i es license and hese s a emen s, he la e shall p e ail. No e ha link p o ided abo e includes addi ional e ms and condi ions o he license. The opinions exp essed in his publica ion a e hose o he au ho s and do no necessa ily e lec he iews o he In e -Ame ican De elopmen Bank, i s Boa d o Di ec o s, o he coun i es hey ep esen . Local Labo -Ma ke E ec s o NAFTA in Mexico: E idence om Mexican Commu ing Zones Danielle T ach enbe g1 Abs ac This pape es ima es he e ec s o NAFTA on labo and wages in Mexico using a local labo -ma ke s app oach. While NAFTA o e ed g ea e expo oppo uni ies o Mexican i ms ha may aise employmen , i also opened he doo o inc eased impo compe i ion ha may dampen employmen gains. This pape inds ha in he i s decade o i s exis ence, NAFTA had a ne posi i e impac on domes ic employmen o 870,000 wo ke s, an inc ease o 13.7%. P oduc ion wo ke s in Mexican gained signi ican ly, wi h employmen inc easing by 32.8%. The impac s o NAFTA a ied by egion, wi h employmen gains acc uing mos ly o s a es in he no heas , no hwes , and cen al eas egions o Mexico, which aded mo e wi h NAFTA coun ies. Key wo ds: In e na ional ade, NAFTA, impo compe i ion, local labo ma ke s JEL: F14, F16, F66, J23 1 In eg a ion and T ade Sec o , In e -Ame ican De elopmen Bank. I am g a e ul o commen s and sugges ions om Juan Blyde, Lo enzo Caliendo, Kun Li, Mau icio Mesqui a Mo ei a, Emanuel O nelas, Ma isol Rod iguez Cha uc, Nina Pa cnik, and pa icipan s a IDB wo kshops. Camila Co es Rod iguez p o ided assis ance wi h da a. The iews exp essed in his pape a e hose o he au ho and should no be a ibu ed o he In e -Ame ican De elopmen Bank, i s boa d o di ec o s, o any o i s membe coun ies. Email: danielle @iadb.o g. 1. INTRODUCTION When NAFTA en e ed in o o ce on Janua y 1, 1994, economis s widely expec ed he ag eemen o ha e a posi i e impac on employmen in Mexico, pa icula ly in manu ac u ing indus ies. Rela i e o he Uni ed S a es and Canada, Mexico specialized in unskilled-labo -in ensi e manu ac u ing indus ies. The implemen a ion o NAFTA inc eased he abili y o all h ee coun ies o engage in egional p oduc ion-sha ing, wi h Mexico se ing as he egion’s sou ce o unskilled-labo -in ensi e in e media e inpu s and a cen e o p ocessing and assembly o inal goods o be expo ed o he no h. Coupled wi h s eep a i declines in manu ac u ing indus ies in all h ee coun ies ha lowe ed he cos o goods c ossing bo de s, employmen o wo ke s in ol ed in he p oduc ion o manu ac u ing goods in Mexico s ood o gain. Al hough employmen o manu ac u ing wo ke s was expec ed o inc ease, he po en ial o nega i e impac s on labo ma ke s om impo compe i ion s ill emained. In he yea s be o e he implemen a ion o NAFTA, Mexico had unde aken a pe iod o sweeping ade libe aliza ion a e yea s o pu suing impo subs i u ion indus ializa ion policies ha p o ec ed domes ic indus ies and kep i s ex e nal sec o ela i ely closed. Impo compe i ion impac ed labo ma ke s, including by widening wage gaps be ween skilled and unskilled wo ke s and c ea ing downwa d p essu e on employmen (Hanson and Ha ison, 1999; Re enga, 1997). NAFTA migh ha e had a simila nega i e impac on wo ke s in indus ies exposed o impo compe i ion om he Uni ed S a es and Canada, dampening any employmen gains om he ag eemen . Fu he mo e, he employmen impac o NAFTA could a y ac oss egions, indus ies, occupa ions, o wo ke skill le els, c ea ing winne s and lose s wi hin Mexico. In his pape , I analyze how NAFTA impac ed employmen in Mexico. I conside wo channels h ough which his may ha e happened: expanded oppo uni ies o expo o NAFTA coun ies and inc eased impo compe i ion om NAFTA coun ies. I u ilize a ia ion in he composi ion o indus ies a he local labo -ma ke le el o calcula e he exposu e o labo ma ke s o impo s om and expo s o NAFTA coun ies. I es ima e he employmen impac o NAFTA ade lows, ollowing he app oach in Au o , Do n, and Hanson (2013). I ind ha NAFTA esul ed in a ne posi i e impac on employmen in Mexico. In he i s decade o i s exis ence, NAFTA inc eased o al employmen in he o mal sec o by 870,000 wo ke s. This numbe masks he e ogeneous impac s ac oss di e en g oups o wo ke s and egions. P oduc ion wo ke s expe ienced a la ge ne inc ease in employmen o 940,000 wo ke s du ing he i s decade o NAFTA. Con ac employmen inc eased sligh ly, while nonp oduc ion wo ke s expe ienced small ne losses. Employmen in egions in he no heas , no hwes , and cen al eas o he coun y gained signi ican ly om NAFTA. 2. BACKGROUND NAFTA u he libe alized Mexico’s ade egime, wi h la ge declines in a i s among Mexico, Canada, and he Uni ed S a es. Figu e 1 shows he a e age applied a i s on NAFTA ade. The blue line shows he simple a e age a i on Mexican impo s applied by he Uni ed S a es and Canada, which d opped om a ound 6% o nea 0% in he i s decade o NAFTA. Mexican a i s applied o impo s om he Uni ed S a es and Canada expe ienced a mo e ex eme decline om o e 13% o nea 0% du ing his pe iod. FIGURE 1. MEXICO-NAFTA TARIFFS, 1993–2013 Sou ce: Au ho ’s calcula ions using an un eleased Wo ld Bank da ase T ade be ween Mexico and NAFTA pa ne s g ew d ama ically upon NAFTA’s implemen a ion. Figu e 2 shows he ise in Mexico’s expo s o and impo s om he Uni ed S a es and Canada. Al hough ade wi h he Uni ed S a es and Canada was ela i ely balanced in he ea ly 1990s, Mexico has enjoyed a ade su plus wi h i s NAFTA pa ne s since NAFTA’s implemen a ion. Mexico’s ade su plus widened in he 2000s, sugges ing he ag eemen boos ed Mexico’s expo oppo uni ies e en i he ag eemen led o heigh ened impo compe i ion. FIGURE 2. MEXICO-NAFTA TRADE, 1991–2013 Sou ce: Au ho ’s calcula ions using da a om UN Com ade. Figu e 3 shows ou di e en measu es o employmen in he o mal sec o in Mexico in i e-yea in e als om 1993 o 2013, paid wo ke s, p oduc ion, nonp oduc ion, and con ac wo ke s. Employmen o paid wo ke s—which comp ise all p oduc ion and nonp oduc ion wo ke s—inc eased om app oxima ely 6 million o almos 11 million wo ke s om 1993 o 2013, an inc ease o 79%. Du ing he same pe iod, he wo king-age popula ion ( he numbe o people be ween ages 15 and 64) inc eased by only 45%. The ise in paid wo ke s is d i en by a d ama ic inc ease in p oduc ion wo ke s in he 1990s, coinciding wi h he implemen a ion o NAFTA. Du ing he same pe iod, he epo ed da a shows a decline in nonp oduc ion wo ke s om 1993 o 1998, al hough his may be due o di e ences in da a collec ion be ween he 1993 and 1998 censuses. 2 A e 1998, nonp oduc ion wo ke s expe ience only a modes inc ease in employmen in Mexico’s o mal sec o , inc easing om 1.4 million in 1998 o 1.6 million in 2013. Figu e 3 also shows ha con ac employmen inc eases signi ican ly in Mexico du ing his ime, going om less han 500,000 o o e 3.1 million wo ke s om 1993 o 2013. 3 Wo ke s in he con ac sec o a e no co e ed by labo egula ions go e ning he minimum wage and social secu i y. The g ow h in he con ac sec o migh explain he sluggish g ow h o nonp oduc ion wo ke s du ing his pe iod. As con ac employmen becomes mo e common, i ms migh be mo e likely o hi e less cos ly con ac wo ke s in nonp oduc ion jobs. F om igu e 3, i is clea ha Mexico’s o mal sec o enjoyed subs an ial employmen g ow h in he pos -NAFTA pe iod, pa icula ly in p oduc ion employmen . This pape explo es o wha ex en NAFTA was esponsible o his inc ease in p oduc ion wo ke s and whe he impo compe i ion om NAFTA coun ies esul ed in any losses o employmen . FIGURE 3. LABOR MEASURES, 1993–2013 Sou ce: Au ho ’s calcula ions using economic census da a om he Ins i u o Nacional de Es adís ica y Geog a ia (INEGI). 2 Addi ionally, because no explici da a on con ac wo ke s we e collec ed o 1993, i is possible ha he la ge numbe o nonp oduc ion wo ke s in 1993 e lec s bo h nonp oduc ion and con ac wo ke s. 3 Employmen o con ac wo ke s in 1993 is ex apola ed om la e yea s because o missing da a in ha yea . 3. LITERATURE REVIEW A numbe o pape s ha e examined employmen and wage ou comes in Mexico in he con ex o i s ade libe aliza ion o he 1980s, inding ha bo h employmen and wages adjus ed o inc eased impo compe i ion. Re enga (1997) no es ha be ween 1985 and 1988, Mexico’s maximum a i d opped om 100% o 20% while he a e age a i dec eased om 25% o 12%. Hanson and Ha ison (1999) ind ha because a i s we e ini ially high in unskilled-labo - in ensi e indus ies, a i educ ions in he 1980s had a disp opo iona ely nega i e e ec on he ea nings o unskilled wo ke s and widened he wage gap be ween skilled and unskilled wo ke s. Simila ly, Re enga (1997) inds ha libe aliza ion was associa ed wi h declines in wages a he i m and indus y le el in manu ac u ing sec o s, which a e in ensi e in unskilled labo , because wage p emiums declined when lowe a i s educed en s, disp opo iona ely a ec ing unskilled wo ke s who did no ha e enough ba gaining powe o cap u e a la ge sha e o en s. Re enga (1997) also shows ha ade libe aliza ion ha esul ed in lowe indus y demand pu downwa d p essu e on employmen a he i m le el in Mexico. On he o he hand, Esqui el and Rod iguez-Lopez (2003) conclude ha ade libe aliza ion om 1988 o 1994 led o a educ ion in he gap be ween skilled and unskilled wo ke s, as ce ain ade models (i.e. Hecksche Ohlin) p edic , bu ha his educ ion was o e shadowed by an inc ease in he wage gap due o echnological change. The e a e a numbe o empi ical pape s analyzing he impac o NAFTA on Mexico, gene ally inding a mos ly posi i e impac on ou comes such as employmen and wel a e wi h some e idence ha NAFTA also inc eased wage inequali y, pa icula ly be ween egions. In a gene al equilib ium analysis, Caliendo and Pa o (2015) ind ha NAFTA a i educ ions esul ed in an inc ease in wel a e in Mexico, wi h ade in in e media e goods, an impo an ace o in a-NAFTA ade, being a pa icula ly impo an ac o in he inc ease in wel a e. Chiquia (2008) and Hanson (2003) bo h ind NAFTA inc eased wage gaps be ween egions in Mexico, d i en by a egion’s access o in e na ional ade oppo uni ies. Chiquia (2008) also inds ha e en hough wage gaps a e ising be ween egions, wage p emiums o skilled wo ke s in egions mos exposed o in e na ional ade decline as a esul o NAFTA. Esqui el and Rod iguez- Lopez (2003) a ibu e he ise in skill p emium in he pos -NAFTA pe iod o echnological changes, inding ha ade libe aliza ion om 1994 o 2000 had no impac on he wage gap be ween skilled and unskilled wo ke s. Mos ele an o his pape is Chiquea , Co a ubias, and Salcedo (2017), who use he local labo -ma ke s app oach o Au o , Do n, and Hanson (2013) o analyze he impac o NAFTA and Chinese impo compe i ion on Mexican labo ma ke s. The au ho s use me o a eas in Mexico as he local labo -ma ke uni o analysis and ocus on he impac o Mexico’s expo s o he Uni ed S a es ins ead o conside ing impo s om and expo s o bo h NAFTA coun ies. They ind ha NAFTA dec eased unemploymen and inc eased wages in Mexico, wi h gains in he me o a eas along he US–Mexico bo de ha a e mo e exposed o ade and in manu ac u ing indus ies such as au os. Se e al ela ed pape s also employ a local labo -ma ke s app oach. Hakobyan and McLa en (2018) examine he employmen impac o Mexican NAFTA a i educ ions bu ocus hei labo -ma ke analysis on blue-colla wo ke s in he Uni ed S a es who a e di e en ially exposed o Mexican a i educ ions depending on hei indus y o employmen . The au ho s ind US wo ke s in indus ies o loca ions exposed o la ge Mexican a i educ ions show slowe wage g ow h compa ed o o he wo ke s, which hey sugges migh be due o o sho ing. Blyde e al. (2018) u ilize a local labo -ma ke s app oach o examine he e ec o Chinese impo compe i ion on Mexican employmen , inding a ne educ ion in Mexican employmen . This pape builds on he exis ing li e a u e in se e al ways. Fi s , I use he de ailed commu ing zones om Blyde e al. (2018) as he local labo -ma ke uni o analysis, allowing me o e 600 obse a ions in a c oss-sec ional analysis. Second, I conside he employmen impac o Mexico’s impo s om and expo s o he Uni ed S a es and Canada. I calcula e he ne impac o NAFTA and quan i y na ional and s a e employmen changes esul ing om NAFTA. While he e a e se e al pape s es ima ing he impac o NAFTA on wages and inequali y, he e a e ela i ely ew es ima es o he employmen impac o NAFTA on Mexico. Wi h he s a e-le el esul s, I am able o analyze he he e ogeneous impac o NAFTA on Mexican egions, highligh ing majo disc epancies ac oss egions due o his o ical and geog aphical ac o s. In addi ion o p o iding no el es ima es o he employmen impac o NAFTA, his pape TABLE 4. IMPACT OF NAFTA ON WAGE OUTCOMES, 2SLS ESTIMATION, 2003–1993 Table 4 p esen s he 2SLS esul s o o al wages paid pe wo king-age popula ion. The esul s display simila bu no iden ical pa e ns as OLS. Inc eased expo exposu e esul s in a ise in wages pe inc ease in wo king-age popula ion o p oduc ion wo ke s, while impo compe i ion dec eases wages o p oduc ion and nonp oduc ion wo ke s, al hough he e ec is only signi ican when he wo g oups a e conside ed oge he in all paid wo ke s. The esul s could be consis en wi h declining wage gaps be ween skilled (nonp oduc ion) and unskilled (p oduc ion) wo ke s, bu he p ecision o he eg essions is no high enough o a de ini i e in e p e a ion. C. NAFTA’s impac on na ional employmen The esul s p esen ed abo e show ha expo oppo uni ies om NAFTA had a posi i e impac on ela i e employmen among commu ing zones while impo compe i ion educed ela i e employmen in some ins ances. I is possible o calcula e he ne impac on employmen in Mexico du ing he i s decade o NAFTA by compu ing he ne impac o each commu ing zone and summing o e all commu ing zones, as in equa ion 6. The ne impac on a commu ing zone is calcula ed as he sum o he e ec s om impo and exposu e, mul iplied by he wo king-age popula ion. Es ima es a e calcula ed using he coe icien s in column (1) o able 3. 4 The es ima ed 𝛽s a e di ided by 1,000 o ma ch he coe icien ’s uni s wi h he le -hand side a iables. The na ional and s a e-le el es ima es p esen ed below ely on he assump ion ha NAFTA changed he absolu e le el o employmen a he han a ec ing only he ela i e le els be ween commu ing zones, which is a easonable assump ion gi en he employmen ends seen in igu e 3. 𝑁𝑒𝑡𝐸𝑓𝑓𝑒𝑐𝑡 = ∑𝑖[(𝛽1∗ Δ𝐼𝑖𝑡 𝑁𝐴𝐹𝑇𝐴 + 𝛽2∗ Δ𝐸𝑖𝑡 𝑁𝐴𝐹𝑇𝐴) ∗ 𝑃𝑜𝑝] (6) Al hough no e e y commu ing zone expe iences a ne posi i e impac , he economy-wide e ec is a ne gain in o al employmen o 870,000 wo ke s, seen in column (1) o able 5, which p esen s he expo , impo , and ne e ec s on employmen o a ious g oups. Fo all wo ke s, expanded expo oppo uni ies gene a ed 1.3 million jobs in he i s 4 Es ima es om column (1) we e chosen o calcula e he ne impac on o al employmen a he han he impac on only p oduc ion, nonp oduc ion, paid, o con ac wo ke s. While he coe icien on he expo exposu e e m in column (1) is signi ican , he coe icien on he impo exposu e e m is no signi ican , po en ially limi ing i s in e p e a i e powe in an exe cise such as his. Howe e , columns (3) h ough (5) allow us o see he esul s o he g oups comp ising o al employmen in column (1), highligh ing he di e en ial impac s o impo exposu e on p oduc ion, nonp oduc ion, and con ac wo ke s. A e e iewing he coe icien s in columns (3) h ough (5), i appea s ha 1.54 in column (1) is a easonable es ima e o he a e age e ec on all wo ke s, e en hough i is no s a is ically signi ican . Con idence in e als o 95% we e calcula ed a ound all esul s using es ima ed coe icien s. decade o NAFTA, while impo compe i ion esul ed in a loss o 456,000 jobs, esul ing in he ne gain o 870,000. The ne gain ep esen s a 13.7% change in employmen om p e-NAFTA le els in 1993 wi h a lowe bound o 10.1% and uppe bound o 16.7%. Table 6 gi es he pe cen age change in employmen and a lowe and uppe bound on he pe cen age change in employmen , calcula ed om he 95% con idence in e als on he coe icien s in able 3. 5 TABLE 5. NATIONAL CHANGE IN EMPLOYMENT DUE TO NAFTA, THOUSAND WORKERS, 1993–2003 Ne e ec Expo e ec Impo e ec All wo ke s 870 1,325 -456 P oduc ion wo ke s 940 1,453 -513 Nonp oduc ion wo ke s -144 -18 -125 Con ac wo ke s 70 -122 192 Sou ce: Au ho ’s calcula ions. The employmen changes o o al employmen mask he e ogenei y in e ec s among he di e en g oups o wo ke s. F om he second ow o able 5, i is clea ha he o al employmen change is d i en by inc eases in he employmen o p oduc ion wo ke s, who expe ienced a la ge ne gain in employmen o 940,000 jobs, a 32.7% inc ease. Expo s o NAFTA coun ies esul ed in a gain o o e 1.4 million p oduc ion jobs, while impo s dec eased p oduc ion employmen by only 513,000. Nonp oduc ion wo ke s, on he o he hand, expe ienced a ne decline in employmen o 144,000 jobs, a dec ease o 4.7%, due o NAFTA. Bo h expanded expo oppo uni ies and impo compe i ion esul ed in employmen declines o nonp oduc ion wo ke s. The decline in nonp oduc ion employmen could be due o se e al ac o s. P oduc ion-sha ing a angemen s migh ha e esul ed in esou ces and in e na ional in es men shi ing o he p oduc ion o manu ac u ing goods, making i mo e di icul o o he indus ies o emain compe i i e o o wo ke s in nonp oduc ion jobs o emain in demand. Addi ionally, a i educ ions and non a i concessions in NAFTA would ha e allowed indus ies om he Uni ed S a es and Canada g ea e access o Mexican ma ke s, possibly o cing some local se ice indus ies ou o business. Finally, i is possible ha some o mally employed nonp oduc ion wo ke s emained in employmen bu as con ac wo ke s. The las ow o able 5 shows ha con ac wo ke s expe ienced a ne gain o 70,000 wo ke s due o NAFTA, an inc ease o 19.3%. Cu iously, expanded expo oppo uni ies o NAFTA coun ies esul ed in a decline in con ac employmen , while impo compe i ion esul ed in a ne inc ease, sugges ing impo compe i ion may ha e o ced Mexican i ms o educe cos s by hi ing con ac wo ke s, simila ly o Blyde e al.’s (2018) inding o a posi i e impac on in o mal sec o wo ke s in an analysis o Chinese impo compe i ion on Mexican wo ke s. TABLE 6. NATIONAL CHANGE IN EMPLOYMENT DUE TO NAFTA, PERCENTAGE CHANGE, 1993–2003 Ne e ec Lowe bound Uppe bound All wo ke s 13.7% 10.7% 16.7% P oduc ion wo ke s 32.7% 26.5% 39.0% Nonp oduc ion wo ke s -4.7% -6.9% -2.4% Con ac wo ke s 19.3% 2.5% 36.1% Sou ce: Au ho ’s calcula ions. D. NAFTA’s impac on employmen by s a e In his sec ion, Mexico’s ne gain in o al employmen o 870,000 wo ke s is b oken down by s a e. Table 7 shows he changes in o al employmen due o NAFTA in e e y s a e as a pe cen age and le el change. De ailed impo and expo e ec s by s a e a e in appendix able 10. S a es a e g ouped acco ding o egion. Lowe and uppe bounds on 5 Uppe and lowe bounds calcula ed om implied le el uppe and lowe bounds o ne employmen changes. The ne employmen change is he sum o wo no mally dis ibu ed andom a iables ( he impo and expo e ec s). The sum o wo no mal dis ibu ions is also no mal wi h a iance equal o he sum o bo h a iances. Uppe and lowe bounds in ne employmen changes we e calcula ed a ound he ne e ec using a s anda d e o de i ed om he sum o bo h a iances. he pe cen age change in employmen a e gi en. I is clea ha he impac o NAFTA on o al employmen is he e ogeneous among s a es. S a es in he no he n and cen al eas egions o Mexico expe ience he la ges e ec s on employmen , while employmen in s a es in he sou h and cen al wes is no a ec ed g ea ly by NAFTA, wi h some excep ions. TABLE 7. NET CHANGE IN TOTAL EMPLOYMENT DUE TO NAFTA, BY STATE, 1993–2003 Region S a e Pe cen age change Lowe bound Uppe bound Le el change *No heas Chihuahua 16.6 14.8 18.4 54,109 Coahuila 11.2 9.1 13.3 17,947 Nue o León 13.2 11.5 15.0 61,473 Tamaulipas 18.7 16.2 21.2 45,081 *No hwes Baja Cali o nia 19.2 17.1 21.3 47,202 Baja Cali o nia Su 2.0 -1.1 5.0 615 Du ango 4.5 1.8 7.1 6,715 Sinaloa 3.5 -0.1 7.2 4,517 Sono a 22.6 20.0 25.3 37,866 *Cen al Eas Fede al Dis ic 12.7 10.7 14.7 235,038 Hidalgo 9.7 4.2 15.3 7,840 México 52.0 47.5 56.5 71,502 Mo elos 47.2 43.8 50.6 38,308 Puebla 16.0 11.9 20.1 38,290 Tlaxcala 5.5 0.6 10.5 1,654 *Cen al Wes Aguascalien es 8.7 6.3 11.0 6,995 Colima 1.6 -2.2 5.4 366 Guanajua o 7.5 4.3 10.8 19,666 Jalisco 21.3 18.7 23.9 102,000 Michoacán 1.3 -4.8 7.3 1,570 Naya i 1.0 -4.6 6.6 308 Que é a o 7.6 4.7 10.4 7,644 San Luis Po osí 7.0 3.1 11.0 8,089 Zaca ecas -0.3 -6.5 5.9 -113 *Sou h Campeche 0.9 -4.2 6.0 273 Chiapas 0.8 -8.4 10.0 762 Gue e o 7.6 2.0 13.1 7,364 Oaxaca 36.8 27.9 45.6 26,294 Quin ana Roo 0.4 -1.6 2.5 251 Tabasco 1.8 -3.5 7.1 878 Ve ac uz 7.1 1.9 12.3 17,345 Yuca án 1.6 -1.9 5.1 1,960 Sou ce: Au ho ’s calcula ions. This pa e n is explained by egional di e ences in he concen a ion o indus ies pa icipa ing in in e na ional ade, which has i s oo s in his o ical and geog aphical ac o s. Figu e 4 shows a hea map o Mexico wi h he pe cen age change in o al ade (expo s plus impo s) wi h NAFTA coun ies om 1993 o 2003. The s a es whose pa icipa ion in NAFTA ade inc eased by he mos a e in he no he n egions nea he Uni ed S a es bo de o clus e ed a ound Mexico Ci y in he cen al pa o he coun y. In Chihuahua, o example, o al ade inc eased by 144%, on pa wi h o he bo de s a es. Nea Mexico Ci y, changes in o al ade wi h NAFTA pa ne s we e as high as o e 300% in Hidalgo and Mo elos, wi h a 143% inc ease in o al ade in Mexico Ci y i sel . By con as , s a es in he sou h saw much lowe inc eases in NAFTA ade, anging om 92% in Yuca án o jus 4% in Quin ana Roo. FIGURE 4. PERCENTAGE CHANGE IN TOTAL TRADE WITH NAFTA COUNTRIES, BY STATE, 1993–2003 Sou ce: Au ho ’s calcula ions. FIGURE 5. PERCENTAGE CHANGE IN TOTAL EMPLOYMENT DUE TO NAFTA, BY STATE, 1993–2003 Sou ce: Au ho ’s calcula ions. Figu e 5 p esen s ano he hea map o Mexico wi h each s a e’s pe cen age change in employmen om column (1) o able 7. A simila pa e n eme ges as in igu e 4. The da kes a eas on he map again a e in he no he n bo de s a es and su ounding Mexico Ci y in he cen al pa o he coun y. Fo example, he s a es o México and Mo elos nea Mexico Ci y expe ience inc eases in employmen o o e 50%. Along he bo de , changes in employmen due o NAFTA ange om 12% in Coahuila o 25% in Sono a. S a es in he sou h see ba ely any change in employmen . NAFTA inc eased employmen by only 0–2% in Tabasco, Chiapas, Campeche, Yuca án, and Quin ano Roo. The s a e o Oaxaca in sou he n Mexico is one excep ion. Employmen inc eased by 36.8%, wi h 75% o he gain o 26,294 wo ke s acc uing o one commu ing zone, Tehuan epec, which is a la ge ag icul u al p oduce and likely bene i ed om educ ions in ag icul u al a i s in he Uni ed S a es and Canada. The egional dispa i y in he impac s o NAFTA is consis en wi h indings in Baylis, Ga duno-Ri e a and Pi as (2012), Chiquia (2008), Hanson (2003) and Chiquea , Co a ubias, and Salcedo (2017), all o which ind ha ade libe aliza ion had a di e en ial impac on egions based on a egion’s exposu e o in e na ional ade. The high pa icipa ion in in e na ional ade o s a es in he no heas and cen al eas a e due o his o ical ac o s in luencing he loca ion o manu ac u ing i ms in hese egions. An In e -Ame ican De elopmen Bank (IDB) epo inds ha ade libe aliza ion esul ed in a deconcen a ion o i ms away om he cen al eas s a es su ounding Mexico Ci y, whe e i ms his o ically loca ed unde impo subs i u ion indus ializa ion due o a ocus on he domes ic ma ke ha made i ad an ageous o be nea cus ome s, supplie s, and wo ke s. The no heas e n s a es bene i ed om his eloca ion o expo -o ien ed i ms due o exis ing anspo ne wo ks and lowe domes ic anspo cos s han o he egions. The IDB epo inds ha anspo cos s a e pa icula ly high in sou he n s a es below Puebla and in he no hwes e n s a es o Baja Cali o nia, Baja Cali o nia Su , Sinaloa, and Du ango. While no hwes e n s a es may be able o o e come high anspo cos s due o p oximi y o he Uni ed S a es, s a es in he sou h con inue o ha e a low concen a ion o i ms pa icipa ing in impo ing o expo ing. The small numbe o i ms engaged in in e na ional ade a ec s he deg ee o which egions expe ienced a meaning ul employmen impac om NAFTA, explaining why NAFTA gene a ed la ge employmen gains in some s a es bu ba ely impac ed employmen in o he s. 6 6. CONCLUSION This pape analyzes he impac o NAFTA on employmen and wages in Mexico using a local labo -ma ke s app oach ha u ilizes a ia ion in he indus ies loca ed in a commu ing zone o de e mine he exposu e o he commu ing zone o expanded expo oppo uni ies o and impo compe i ion om NAFTA coun ies. Expanded expo oppo uni ies gene ally aise employmen , while impo compe i ion lowe s employmen in mos ins ances. Resul s con i m p e-NAFTA p edic ions ha he ag eemen would help p oduc ion wo ke s and boos o e all employmen in Mexico. This pape inds ha Mexico expe ienced a ne gain in employmen o 870,000 wo ke s due o NAFTA, an inc ease o 13.7%. In pa icula , p oduc ion wo ke s gained signi ican ly, wi h employmen inc easing by 32.7%. Employmen o con ac wo ke s also g ew signi ican ly as an adjus men mechanism o inc eased impo compe i ion. Nonp oduc ion wo ke s expe ienced small declines in employmen due o NAFTA. The employmen impac o NAFTA a ies ac oss s a es. NAFTA had a la ge impac on employmen in s a es in he no heas and cen al eas egions because i ms pa icipa ing in in e na ional ade a e concen a ed in hese s a es and he e o e we e mo e easily able o ake ad an age o a i declines, al hough hey we e also mo e exposed o impo compe i ion. S a es in he sou h and some s a es in he cen al wes expe ienced only e y small impac s because i ms in hese a eas a e less likely o pa icipa e in in e na ional ade han i ms in he no h and cen al eas . REFERENCES Au o , Da id H., Da id Do n, and Go don H. Hanson. 2013. “The China Synd ome: Local Labo Ma ke E ec s o Impo Compe i ion in he Uni ed S a es.” Ame ican Economic Re iew 103(6): 2121–2168. Baylis, Ka hy, Ra ael Ga duno-Ri e a, and Gian anco Pi as. 2012. “The Dis ibu ional E ec s o Na a in Mexico: E idence om a Panel o Municipali ies.” Regional Science and U ban Economics 42: 286–302. 6 This pape does no a emp o analyze any spillo e o gene al equilib ium e ec s ha may ha e u he impac ed s a es in he sou h. Blyde, Juan, Ma ías Busso, Vale ia Faggioni, and Da io Rome o. 2018. “Impo Compe i ion and Labo Ma ke Adjus men s in Mexico.” Caliendo, Lo enzo, and Fe nando Pa o. 2015. “Es ima es o he T ade and Wel a e E ec s o NAFTA.” Re iew o Economic S udies 82(1): 1–44. Chiquia , Daniel. 2008. “Globaliza ion, Regional Wage Di e en ials and he S olpe - Samuelson Theo em: E idence om Mexico.” Jou nal o In e na ional Economics 74: 70–93. Chiquia , Daniel, En ique Co a ubias, and Alejand ina Salcedo. 2017. “Labo Ma ke Consequences o T ade Openness and Compe i ion in Fo eign Ma ke s.” Banco de México Wo king Pape 2017–01. Esqui el, Ge a do, and José An onio Rod íguez-López. 2003. “Technology, T ade, and Wage Inequali y Be o e and A e NAFTA.” Jou nal o De elopmen Economics 72: 543–565. Hakobyan, Shushanik, and John McLa en. 2018. “Local-Labo -Ma ke E ec s o NAFTA: The O he Shoe D ops.” Hanson, Go don H. 2003. “Wha Has Happened o Wages in Mexico Since NAFTA? Implica ions o Hemisphe ic F ee T ade.” NBER Wo king Pape 9563. Hanson, Go don H., and Ann Ha ison, “T ade Libe aliza ion and Wage Inequali y in Mexico.” Indus ial and Labo Rela ions Re iew 52(2): 271–288. In e -Ame ican De elopmen Bank. 2013. “Too Fa o Expo . Domes ic T anspo Cos s and Regional Expo Dispa i ies in La in Ame ica and he Ca ibbean.” Re enga, Ana. 1997. “Employmen and Wage E ec s o T ade Libe aliza ion: The Case o Mexican Manu ac u ing.” Jou nal o Labo Economics 15(S3): S20–S43. APPENDICES A. Robus ness TABLE A.1. IMPACT OF NAFTA ON LABOR OUTCOMES, 2SLS ESTIMATION, 2003–1993, EXCLUDING THE FEDERAL DISTRICT TABLE A.2. IMPACT OF NAFTA ON LABOR OUTCOMES, 2SLS ESTIMATION, 2003–1993, EXCLUDING NORTHEAST AND NORTHWEST B. Addi ional s a e employmen esul s TABLE B.1. IMPORT AND EXPORT EFFECTS OF NAFTA, BY STATE, 1993–2003 (NUMBER OF WORKERS) Region S a e Ne e ec Impo e ec Expo e ec *No heas Chihuahua 54,109 -22,412 76,521 Coahuila 17,947 -7,240 25,187 Nue o León 61,473 -33,976 95,449 Tamaulipas 45,081 -21,266 66,346 *No hwes Baja Cali o nia 47,202 -31,507 78,710 Baja Cali o nia Su 615 -872 1,487 Du ango 6,715 -8,374 15,089 Sinaloa 4,517 -1,707 6,224 Sono a 37,866 -11,214 49,080 *Cen al Eas Fede al Dis ic 235,038 -142,009 377,046 Hidalgo 7,840 -15,728 23,568 México 71,502 -19,098 90,600 Mo elos 38,308 -5,060 43,368 Puebla 38,290 -20,544 58,834 Tlaxcala 1,654 -5,058 6,712 *Cen al Wes Aguascalien es 6,995 -4,148 11,143 Colima 366 67 298 Guanajua o 19,666 -16,944 36,611 Jalisco 102,000 -26,709 128,709 Michoacán 1,570 -7,213 8,783 Naya i 308 -232 540 Que é a o 7,644 -14,246 21,891 San Luis Po osí 8,089 -5,244 13,333 Zaca ecas -113 -1,216 1,103 *Sou h Campeche 273 -182 455 Chiapas 762 -2,169 2,931 Gue e o 7,364 -2,952 10,316 Oaxaca 26,294 -13,670 39,964 Quin ana Roo 251 70 181 Tabasco 878 -36 914 Ve ac uz 17,345 -12,834 30,178 Yuca án 1,960 -1,833 3,793