Inernaional Journal of Business, Humaniies and Technology Vol. 1 No.1; July 2011 Expor-led Growh: A Case Sudy of Mexico Kimberly Waihe a, Troy Lorde b, and Brian Francis b a Minisry of Economic Affairs, Warrens, S. Michael b Deparmen of Economics, Universiy of he Wes Indies, Cave Hill Campus, Barbados Absrac Agains he backdrop of expor-oriened policy reforms o Mexico s rade regime in he mid 1980s, he presen sudy underakes a case sudy of economic growh in Mexico. Using an expor-augmened neoclassical producion funcion, he validiy of he Expor-led Growh Hypohesis for Mexico is also esed over he period 1960-2003. Evidence offers suppor for he Hypohesis in he shor run; however, conrary o he Hypohesis, long-run resuls sugges an inverse relaionship beween expors and GDP. A likely explanaion is he high impor conen and diminishing local conen of expors, and weak linkages wih domesic suppliers, hus reducing possible spillover or muliplier benefis. If Mexico is o succeed in is ques o achieve high and seady economic growh, curren incenive schemes ha allow ax-free enry of impored inpus and raw maerials for expor purposes mus be reconsidered. Finally, policies ha promoe echnological innovaion in manufacuring and linkages wih local suppliers are imperaive. Keywords: impor subsiuion; expor-orienaion; rade reform; expor-led growh hypohesis 1. Inroducion In he 1950s o 1970s, he policy of impor subsiuion (IS) formed a major plank upon which many developing counries hoped o achieve significan economic growh. However, from he mid-1970s, here was a considerable shif owards adopion of he sraegy of expor promoion (EP). By he early 1980s he EP sraegy had aained wide consensus among researchers and policymakers o such an exen ha i had become convenional wisdom among mos economiss in he developing world (Balassa, 1985). Advocaes of he expor-led growh hypohesis (ELGH) conend ha Lain American counries which pursued inwardoriened policies under he IS sraegy had underperformed. Several recorded no growh on average, while real income declined beween 1960 and 1990 (Barro and Sala-i-Marin, 1995). Moreover, counries in he region increasingly faced ballooning fiscal deficis, acue inflaion, supply shorages and ulimaely, severe balanceof-paymens crises coupled wih economic recession. Consequenly, he 1980s he los decade in Lain America s economic developmen were marked by a series of aemps o correc hese macroeconomic disequilibria in he face of serious consrains o accessing foreign credi and capial markes. Following he convenional wisdom, Lain American governmens adoped ouward-looking sraegies as inegral componens of heir adjusmen and sabilizaion programs. Mexico was no spared from he need o reform. From he 1940s unil he 1970s, Mexico s economic developmen was based on srong sae inervenion o foser indusrializaion hrough impor subsiuion. This sraegy was iniially quie successful; i ransformed he counry from an agrarian o an urban, semiindusrial sociey. However, during he early 1980s, in he afermah of a dramaic balance-of-paymens crisis, he Mexican governmen implemened a series of economic reforms o shif he economy away from is radiional sae-led developmen sraegy. One major policy shif was he change from IS iniiaives o an expor orienaion. Agains his backdrop, he broad objecive of his paper is o underake a case sudy of Mexico s economic growh under expor-oriened policies. The specific objecive is o es he validiy of he ELGH for Mexico. Sudy of he relaionship beween expor growh and naional income is imporan as expor promoion indusrializaion (EPI) has been idenified as a key o growh and developmen. In addiion, he World Bank and he Inernaional Moneary Fund (IMF) have advocaed EPI wihin he conex of rade liberalizaion. Mexico provides an ineresing case for sudy as i is highly open 1, and a peroleum-exporing economy. The remainder of he paper is srucured as follows: Secion 2 analyses he economic environmen which fosered he need for reform in Mexico and he policies implemened as a resul. 1 By he 1990s Mexico was considered o be one of he mos open economies in he world (OECD 2002). 33
Cenre for Promoing Ideas, USA www.ijbhne.com This secion also looks a he impac of reform on he composiion of expors. Secion 3 provides a selec review of he heoreical and empirical lieraure on he ELGH. Secion 4 describes he empirical model, economeric mehodology and daa sources while Secion 5 presens our resuls and analysis. The final secion conains some concluding remarks. 2. Trade Reform in Mexico 2.1 The Background for Trade Reform in Mexico Indusrial policy in Mexico during he 1940s o 1970s operaed hrough secor-specific program, wih he aim of developing a manufacuring secor capable of producing capial goods and, o some exen, complex inermediae inpus (Ros, 1994). To achieve his goal, ax cus and rade resricions were implemened, wih sric requiremens regarding, for insance, he degree of local conen and ne-expor performance. In addiion, a fundamenal elemen of Mexico s indusrial sraegy was, and sill is, he Maquiladora program. Iniiaed in 1966, is objecive was o simulae he esablishmen of labour-inensive, in-bond expor processing plans (known as maquiladoras), by offering hem ax-free access o impored inpus and machinery, as well as exempion from sales and income axes. These assembly plans ransformed impored pars and componens ino finished or semi-finished producs for expor. The IS sraegy on he whole was iniially quie successful. I ransformed he counry from an agrarian o an urban, semi-indusrial sociey. From 1940 o he mid-1970s, Mexico s real per capia GDP grew a an average annual rae of hree percen. Manufacuring was he driving force of his growh process, wih oupu expanding annually a an average of nearly eigh percen, boosed by domesic demand. During his period, he share of manufacuring in GDP rose from 15 percen o 25 percen. Noneheless, several concerns wih he design and applicaion of he IS sraegy were no sufficienly addressed (Mereno-Brid e al., 2005). The firs concern was he uneven disribuion of he benefis from economic growh. Second, he public secor s dependence on exernal deb was no reduced by fiscal reforms ha would srenghen ax revenues. Third, wih he excepion of a small number of special developmen secoral program, here were few policies in place o efficienly promoe expors. The failure o adequaely address hese concerns proved faal. In he lae 1970s, Mexico s economic expansion los momenum, especially caused by difficulies in subsiuing impors of high-echnology capial goods. In response, he governmen launched an ambiious developmen program in 1977 funded by he vas inflow of oil revenues and by exernal deb. This oil-driven boom was, however, shor-lived. As a resul, fiscal and foreign exchange revenues, increasingly dependen on peroleum expors, became very vulnerable o exernal shocks. In urn, impors of inermediae and capial goods rapidly increased, causing a huge rade defici. Given Mexico s vulnerabiliy o exernal shocks, he collapse of he inernaional oil marke in 1981, coupled wih he rise in Unied Saes (US) ineres raes, severely affeced he economy, riggering win fiscal and foreign exchange crises. These evens ended Mexico s fory-year economic expansion, and were he caalys for a series of economic reforms direced owards posiioning he privae secor and he marke as he pivoal agens of invesmen and indusrializaion (Mereno-Brid e al., 2005). I became increasingly eviden ha Mexico required reforms o diversify is sources of foreign revenues, and o improve he srucure of he producive secor. As oil expors decreased and foreign deb servicing requiremens rose, he need o increase non-oil expors became imperaive. Thus, in he afermah of he mos dramaic balance-of-paymens crisis ha Mexico had faced in decades, he governmen began seing up a series of economic reforms o shif he economy away from is radiional sae-led developmen sraegy. Mexico sared reform of is rade regime in 1985. I eliminaed impor licenses on capial and inermediae goods and reduced ariffs. Specifically, he percenage of impors subjec o licensing dropped from 64.7 percen in 1984 o 10.4 percen by he end of 1985, and he average ariff dropped from 27 percen in 1983 o 23.8 percen in 1985 (Li, 1999). Inspired by he Washingon Consensus, a major policy shif from IS policies o expor-oriened growh sraegies was embraced. Expor orienaion was also accompanied by greaer effors in regional inegraion. As par of a broader policy of privaizaion, deregulaion and liberalizaion, Mexico joined he General Agreemen on Tariffs and Trade (GATT), he predecessor o he World Trade Organizaion (WTO), in 1986. Trade links wih he US and Canada were also reinforced hrough membership in he Norh American Free Trade Agreemen (NAFTA) in 1994. According o Mereno-Brid e al., (2005), NAFTA was seen as a vehicle for achieving wo goals. The firs was o se he Mexican economy on a non-inflaionary, expor-led growh pah, driven by sales of manufacured goods mainly o he US. 34
Inernaional Journal of Business, Humaniies and Technology Vol. 1 No.1; July 2011 The second was o escape from he series of repeaed economic and financial crises ha had plagued he counry from he mid-1970s hrough he mid-1990s 2 by guaraneeing he lock-in of Mexico s macroeconomic reform process. 2.2 The Impac of Trade Reform on he Composiion of Expors in Mexico In he lae 1970s, Mexico was fundamenally an oil-exporing economy. As of 1985 Mexico ranked fifh in he world in erms of peroleum reserves, and was he larges supplier o he US (McCarville and Nnadozie, 1995). Peroleum provided significan revenues unil a decline in inernaional peroleum prices in 1985. Due o declines in prices and volume of expors, he peroleum indusry was adversely affeced, causing revenue loss for he counry; he value of Mexico's oil expors plummeed from US$13 billion in 1985 o less han US$6 billion in 1986. The oil secor's share of oal expor revenue consequenly fell from 78 percen in 1982 o 42 percen in 1987 and he value of Mexico's expors fell from US$24 billion in 1984 o US$16 billion in 1986. Consequenly, oil expors ceased o be a major facor in Mexican growh during he period of low real oil prices from abou 1987 hrough 1999 (Blecker, 2006). In more recen imes, here has been renewed imporance of oil expor revenue as a resul of increasing global prices since 2000. Thus, even wih is limied capaciy o expor oil and oher energy producs, Mexico has benefied from he higher prices of oil expors. The value of Mexico expors of crude oil rose by 27.3 percen beween 2003 and 2004. Virually all of his increase was due o a 25.2 percen increase in he average price of he counry s crude oil expors in 2003-2004. Addiionally, he volume of such expors rose by 1.4 percen over his period (Blecker, 2006). Unforunaely, Mexico s energy policies, since he shor-lived oil boom of he lae 1970s, have failed o creae an efficien energy secor ha could maximize he revenues of high energy prices. According o Blecker (2006), much of he revenue from he sae-owned oil company has been siphoned off for oher uses, including deb service, corrupion, and general governmen revenues, raher han reinvesed in modernizaion of he energy secor or oher infrasrucure needs. Furhermore, he counry s capaciy o produce naural gas has never been fully exploied. Norhern Mexico now impors naural gas from Texas a high prices (parly offseing he benefis of high prices for is oil expors), despie Mexico s subsanial naural gas reserves in he souhern and Gulf regions and possibly in he norh as well. Since 1985, a number of iniiaives have been aken o promoe non-oil expors, such as he easing of requiremens for imporing inermediae and capial goods, he greaer access of credi for exporers, and he reducion of resricions on he use of expor earnings. Trade liberalizaion, crowned by NAFTA, has also been associaed wih Mexico s inserion ino global markes and is rising imporance in non-oil expors, consequenly allowing he economy o be more resilien o possible price changes in he inernaional oil markes and less vulnerable o exernal shocks. Manufacures accouned for more han 50 percen of he counry s oal expors by 1988, and in 2004 heir share exceeded 85 percen, as heir rapid growh more han compensaed for weak performances in he expors of oil, minerals and agriculural commodiies. From 1985 o 1994, Mexico ranked fifh among counries wih he larges increases in heir share of world manufacures expors; during 1994-2001, i moved o second place, behind China. The Maquiladoras also consiued a vial force behind he expor drive in manufacures. In he early 1990s, hey provided more han half of Mexico s oal expors of manufacured goods, and more han 40 percen of Mexico s oal expors. As a resul of he dynamism in non-oil expors, oal expors represened 16 percen of Mexico s real GDP in 1994. By 2000, oal expors had more han doubled, reaching 35.1 percen. Finally, i is imporan o noe ha parallel o he expor boom in manufacures, he Mexican economy has experienced a massive peneraion of impors, mainly manufacured goods, since he 1980s. In 1982, manufacures represened 90 percen of oal impors, measured in consan pesos. By 1994, heir share was 95 percen, a level which sill mainains. As a share of GDP, hey climbed from 10 percen in 1982 o more han 30 percen by he mid-1990s. From 1988 o 2003, impors of manufacures a consan prices grew more han wice as quickly as expors. As such, he rade defici in manufacuring has been widening, puing exra pressure on he overall rade balance. 3. Lieraure Review 2 During he 1980s, Mexico was hi by wo major crises, namely he 1982 deb crisis (he peso was devalued by over 57 percen from he previous year s level) and he collapse of oil prices in 1986 (from US$25.50 per barrel in 1985 o US$12 per barrel in 1986). The peso was furher devalued on December 20, 1994 and he financial crisis ha followed sen inflaion soaring and se off a severe recession in Mexico in 1995. For more deailed analyses of he peso/equila crises, see Lusig (1998), Blecker (1996) and Williamson (1997). 35
Cenre for Promoing Ideas, USA www.ijbhne.com 3.1 Theoreical Review of he Expor-Led Growh Hypohesis The expor-led growh hypohesis posulaes ha expors are he main deerminan of overall economic growh. One of he main argumens in suppor of he hypohesis is ha expor growh may affec oal facor produciviy hrough dynamic spillover effecs on he res of he economy (Feder, 1983). Empirical sudies based on he producion funcion framework include expors because of his spillover effec. In shor, his is learning by doing, or more precisely, learning by exporing (Tyler, 1981; Lucas, 1988). In heory, here are several ways in which expors can poenially cause an increase in produciviy. An expansion in expors may promoe specializaion in he producion of expor producs, which in urn may boos produciviy levels and may cause he general level of skills o rise in he expor secor. This hen leads o a reallocaion of resources from he (relaively) inefficien non-rade secor o he higher producive expor secor. This produciviy change leads o oupu growh. Balassa (1985) assers ha in general, he producion of expor goods is focused on hose economic secors of he economy which are already more efficien. Therefore, expor expansion helps o concenrae invesmen in hese secors, which in urn increase he overall oal produciviy of he economy. These argumens have been reinforced by he endogenous growh lieraure. The new endogenous growh models have made major modificaions o he neoclassical growh heory s approach o handling rade effecs. 3 Endogenous growh heory emphasizes ha expors are likely o increase long-run growh by allowing a higher rae of echnological innovaion and dynamic learning from abroad (Lucas, 1988; Romer, 1986, 1989; Grossman and Helpman, 1991; Edwards, 1992). The suppor for he ELGH is, however, no universal. Moon (1998) argues ha naions characerized as following ouward-oriened policies do no manifes levels of rade noably higher, or expand heir rade a raes higher, han hose regarded as inward-oriened. In addiion, he noes ha i is no apparen ha expor expansion is he principal source of superior macro-economic performance of ouward-oriened naions. Criics poin ou ha he experiences in he Eas and Souheas Asian counries are unique in many ways and no necessarily replicable in oher counries (Buffie 1992). Moreover, he producion and composiion of expors was no lef o he marke bu resuled as much from carefully planned inervenion by heir governmens. Jaffee (1985) also quesions wheher a reliance on expors o lead he economy will resul in susained long-erm economic growh in lesser developed counries (LDCs), due o he volailiy and unpredicabiliy in he world marke. 3.2 Empirical Evidence Early sudies of he ELGH used cross-secion analysis. Among hese sudies, Emery (1968), Kravis (1970), and Krueger (1978) should be menioned. This group of sudies used bivariae correlaion he Spearman rank correlaion es o illusrae he effecs of he ELGH. The general conclusion was ha high levels of economic growh are significanly associaed wih high levels of expor growh. However, resuls were likely due o spurious correlaion since expors are a componen of GDP. This lead some researchers o use oupu ne of expors or alernaive expor variables. The laer research concluded ha here may be a need for a minimum hreshold of developmen before any associaion may exis. Naions which are heavily dependen on agriculural commodiies are less likely o benefi from expors, in comparison wih counries ha have a higher level of developmen and whose expors conain a higher domesic value added (Kohli and Singh, 1989). This concern led o a second group of cross-secion sudies ha esimaed aggregae producion funcions which included expors as an explanaory variable along wih oher proposed economic growh deermining fundamenals, such as labour, capial, and invesmen. In his group of sudies, expors were included in an ad hoc manner in he producion funcion, ogeher wih labour and capial. Many of hese sudies, published in he lae 1970s, found a significan posiive relaionship beween expor performance and he growh of naional income. Balassa (1978) summarized hem, saing ha The evidence is quie conclusive: counries applying ouward oriened developmen sraegies performed beer in erms of expors, economic growh and employmen han counries wih coninued inward orienaion. Several doubs were raised abou he usefulness of resuls from cross-secion sudies. Counries in similar sages of developmen were grouped ogeher. Therefore, he empirical resuls obained are averages ha do no capure he peculiariies of many developing counries. 3 In he neoclassical model, rade and oher ancillary variables affeced he equilibrium level of aggregae oupu bu no is rae of growh. However, criics argue ha his does no explain he origin of growh. 36
Inernaional Journal of Business, Humaniies and Technology Vol. 1 No.1; July 2011 Implicily, by assuming he same producion funcion across differen ypes of economies, hey do no ake ino accoun he level of echnology, which is likely o differ across counries. Moreover, cross-secion analysis ignores he shifs in he relaionship beween variables overime wihin a counry, while expor growh and economic growh is a long-run phenomenon, which canno be sudied by using cross-secion analysis. Finally, sudies which used large samples were limied o specific ypes of LDCs. Mos researchers chose a priori middle-income counries and excluded low-income counries and major oil exporers (Feder, 1983; Kavoussi, 1984). Recen emphasis has been placed on ime series analyses o deermine wheher a long-erm relaionship beween expor and economic growh exiss and he direcion of causaliy. Jung and Marshall (1985), use sandard bivariae Granger causaliy ess o analyze he relaionship beween expor growh and economic growh for 37 developing counries and find evidence o suppor he ELGH in only 4 (Indonesia, Egyp, Cosa Rica, and Ecuador). Using Sims procedure, Chow (1987) invesigaes he causal relaionship beween expor growh and indusrial developmen in eigh newly indusrialized counries (NICs). He finds evidence of a srong bi-direcional causaliy beween expor growh and indusrial developmen for Brazil, Hong Kong, Israel, Korea, Singapore and Taiwan; one-way causaliy from expor growh o oupu growh for Mexico; and no causaliy for Argenina. Chow s resuls are conras o Jung and Marshall, who find no significan causaliy in Brazil or Mexico, and causaliy only from oupu o expors in Korea and Taiwan. The conras in empirical findings of he wo sudies may be parly explained by he fac ha Chow uses oupu of he manufacuring secor as a measure of aggregae oupu as opposed o Jung and Marshall who uilize GDP. The conribuion by Bahmani-Oskooee and Alse (1993) was perhaps he firs o uilize an error correcion mechanism in heir analysis of expors and GDP. They find srong empirical suppor for bidirecional causaliy beween expor and GDP in eigh ou of nine counries; he only excepion was Malaysia. They concluded ha here is evidence ha expor-promoion conribues o economic growh in LDCs and viceversa. A recen sudy on Taiwan by Chang e al. (2000) depars from previous research by including impors as a facor in he relaionship beween expors and GDP. According o Riezman, Whieman and Summers (1996), omiing impors can resul in spurious conclusions regarding he ELGH, because capial goods impors in paricular are necessary inpus for enhancemen of expor and domesic producion. Furhermore, expor growh may relieve he foreign exchange consrain, allowing capial goods o be impored o boos economic growh. Chang e al. find bidirecional feedback beween income and expors, income and impors, and expors and impors respecively. Few sudies have focused on Mexico as an individual case sudy of he ELGH. Using Granger-causaliy ess, McCarville and Nnadozie (1995) repor evidence of a relaionship beween expor growh and GDP growh in Mexico. Thornon (1996) also finds similar resuls. Using coinegraion and Granger-causaliy ess wihin a wo-variable framework, he concluded ha real expors and real GDP in Mexico over 1895 1992 were coinegraed and here was a significan and posiive Granger-causal relaionship running from expors o economic growh. Li (1999) ook a differen approach by examining he relaionship beween economic growh, expors and expor diversificaion in he conex of he ELGH. The empirical findings sugges an imporan role for expors on oupu in he long run, and for expor diversificaion on economic growh in he shor run. Alam (2003) examines he ELGH for Mexico and Brazil, using an augmened producion funcion framework. 4 The main argumen is ha he wide variaions in empirical resuls can be aribued o he fac ha causaliy ess are exremely sensiive o omied variables. Even if expors are found no o cause growh in bivariae models, his same inference does no necessarily hold in he conex of larger economic models ha include oher relevan variables such as capial and labour (Awokuse, 2003). Alam s empirical evidence does no suppor he ELGH. Moreover, he resuls indicae ha impored capial goods appear o be very significan in he growh process of hese wo counries. The resuls in his sudy were found o be robus across esimaion echniques. 4. Empirical Model, Economeric Mehodology and Daa Sources 4.1 Empirical Model Our empirical model sars wih a simple neoclassical producion funcion: Y A K L (1) 4 Oher sudies which have used he expor-augmened producion funcion approach include Van den Berg and Schmid (1994) and Herzer e al. (2006). 37
Cenre for Promoing Ideas, USA www.ijbhne.com where Y denoes he aggregae producion of he economy a ime ; A is he level of oal facor produciviy; K,L are he levels of he capial sock, and he sock of labour, respecively; and α and β are consans beween zero and one ha measure capial and labour s share of income respecively. This paper goes beyond he radiional neoclassical heory of producion by esimaing an augmened Cobb- Douglas funcional form, which includes expors. The inclusion of expors as a hird inpu provides an alernaive procedure o capure oal facor produciviy (TFP) growh. Following Herzer e al. (2006), we assume ha oal facor produciviy can be rewrien as a funcion of expors (X ), impors (M ) 5 and oher exogenous facors (C ) assumed o be uncorrelaed wih X and M. This implies ha our esimaes will be unbiased. The resuling specificaion is: A f ( M, X, C ) M X C (2) Combing (2) wih (1) we obain: Y C K L M X (3) where α, β, δ and γ are he elasiciies of producion wih respec o K, L, M and X. Taking naural logs (Ln) of equaion (3) and expressing i economerically for esimaion purposes we obain: LY a LK LL LM LX e (4) where L sands for Ln; a is a consan parameer; all coefficiens are consan elasiciies; and e is an error erm, which capures he influence of all oher exogenous facors. Some sudies have argued ha i is necessary o separae he economic influence of expors on oupu from he influence incorporaed ino he growh accouning relaionship (Heller and Porer, 1978; Islam, 1998; Herzer e al., 2006). We address his issue by using aggregae oupu ne of expors: LNY a LK LL LM LX e (5) where NY represens Y ne of expors. The signs of he coefficiens α, β, γ are expeced o be posiive, while he sign of δ could be posiive or negaive. 4.2 Economeric Mehodology This sudy empirically examines he ELGH using he mulivariae coinegraion echniques of Johansen (1988) and Johansen and Juselius (1990). This mehod applies he maximum likelihood procedure o deermine he presence of coinegraing vecors in nonsaionary ime series. They have proposed wo es saisics for esing he number of coinegraing vecors: he race (Tr) and he maximum eigenvalue (L-max) saisics. The Schwarz Informaion Crierion (SIC) will be used o selec he number of lags required in he coinegraion es. A necessary precondiion o esing for coinegraion is o inspec he uni roo properies of he variables under consideraion. We es for uni roos using he Augmened Dickey-Fuller (ADF) es by Dickey and Fuller (1979, 1981), he Phillips-Perron (PP) es by Phillips and Perron (1988) and he KPSS es by Kwiakowski e al. (1992). A conclusion on he degree of inegraion is made based on he agreemen of a leas wo of he hree uni roo ess. The nex sep is o esimae he shor-run dynamics by means of Granger-causaliy es from an error correcion model. The objecive is o ascerain wheher expors leads growh, growh leads expors, boh, or neiher hold rue for Mexico. For illusraion purposes, we ouline he procedure for expors and non-expor GDP only: n LNY c [ LK LL LM LX LNY ] ECT u (6) 0 i 1 i 1 i 1 i 1 i 1 i i 1 n LX c [ LK LL LM LX LNY ] ECT e (7) 0 i 1 i 1 i 1 i 1 i 1 i i 1 where is he difference operaor and he ECT is he error correcion erm which represens he lagged error from he coinegraion equaion. Innovaion accouning is used o deermine he dynamic responses of he variables. We use he impulse response funcion o race how economic growh responds over ime o a shock in expors and compare his o responses o shocks from oher variables. If he impulse response funcion shows a sronger and longer reacion of economic growh o a shock in expors han shocks in oher variables, his is suppor for he ELGH. 5 Riezman, Whieman and Summers (1996) poin ou ha he inclusion of impors is crucial in esing his hypohesis o avoid producing spurious causaliy resuls. 38
Inernaional Journal of Business, Humaniies and Technology Vol. 1 No.1; July 2011 Similarly, if he impulse response funcion shows a sronger and longer reacion of expors o a shock in economic growh han o shocks in oher variables, we would find suppor for he hypohesis ha economic growh leads expors. Variance decomposiion provides informaion concerning he relaive imporance of each innovaion owards explaining he behaviour of endogenous variables. In his sudy, variance decomposiion is used o answer he quesion: How much of he variance in forecas error of fuure income growh can be aribued o innovaions in expor growh? If, for example, a shock o expors leads subsequenly o a large change in economic growh, bu ha a shock o economic growh has only a small effec on expors, his is evidence in suppor of he ELGH. 4.3 Daa The required series for NY, K, M and X were aken from he World Bank World Developmen Indicaors online daabase, while L was colleced from he Yearbook of Labour Saisics published by he Inernaional Labour Office (ILO). The variables NY, M and X, represen non-expor real GDP, real impors, and real expors respecively. K, is proxied by real gross fixed capial formaion, while L, is proxied by he oal number of people employed each year. All daa are annual and run from 1960 o 2003, and are logged for esimaion purposes. 5. Resuls and Analysis The ADF, PP and KPSS ess are in agreemen ha a convenional levels of significance, LNY, LK, and LM are inegraed of order 1 or I(1), and LL inegraed of order 2, I(2). However, he ADF and PP es fail o rejec he null of non-saionariy for LX a he level and a firs difference, conradicing he KPSS, which indicaes i is saionary a firs difference. However, using he rule ha a conclusion will be drawn based on agreemen beween a leas wo of he hree ess, we conclude wih furher analysis ha LX is I(2). The linear combinaion of LL and LX can be I(2) or I(1); however, he order of inegraion mus be one for coinegraion o exis. To verify if hese variables are coinegraed, we proceed wih he Johansen es. The resuls are summarized in Table 2. The race and max es boh indicae he presence of a one coinegraing vecor. We hus conclude ha here exiss a long-run relaionship among LNY, LK, LL, LX and LM. Afer normalizing on LNY, we obain he long-run relaion shown in Table 3. The esimaed coinegraing vecor has plausible coefficiens; however, conrary o expecaions, he resuls indicae a negaive relaionship beween expors and non-expor GDP. Tha is, in he long run an increase in expors will lead o a decrease in economic growh. A possible explanaion for his surprising resul is ha alhough Mexico s expors, especially is manufacures, experienced a surge in he lae 1990s, afer adopion of NAFTA and he depreciaion of he peso in 1994-95, he impor conen of expors was high and hus grealy reduced he spillover or muliplier benefis. According o Blecker (2006), he mos dramaic case is he Maquiladoras, since approximaely hree-quarers of he gross value of Maquiladora expors are offse by Maquiladora impors. Thus he ne benefi o he Mexican domesic economy is minimal. As Moreno-Brid e al. (2005) poin ou, around 70 percen of Mexico s expors of manufacures are produced hrough assembly processes involving impored inpus ha ener he counry under preferenial ax schemes such as PITEX, which allows ax-free enry of impored inpus and raw maerials for expor purposes and ALTEX, which allows ax-free enry of emporary inpus o large exporers. Mexico s manufacured expors have also become increasingly characerized by reduced local conen and weak linkages wih domesic suppliers. Is expor indusries are well inegraed ino regional (Norh American) and global producion chains, and more and more disinegraed from he domesic Mexican economy (Blecker, 2006). The resul is ha even if expors growh is high, he ne simulus o he overall growh of he economy and o job creaion is relaively small. Indeed, saisics show ha he value added in he Mexican manufacuring secor has been sagnan, even hough he gross value of expors has grown considerably (UNCTAD, 2002, pp. 77-81). As a resul, he relief ha Mexico obains from he balance-ofpaymens consrain hrough expor growh has been minimal, because much of he foreign exchange earnings are needed o pay for impored inpus. Table 3 also shows he imporance of invesmen and impors as a source of long-erm economic growh: A one percen increase in invesmen leads o an increase of 0.94 percen in non-expor GDP; while a one percen increases in impors leads o a 2.24 percen increase, which accords wih findings by Alam (2003). 6 The significance of he error correcion erm in he LX vecor implies ha real expors in Mexico adjus o changes in labour, non-expor GDP, impors and invesmen. The esimaed error correcion erm has he correc sign and is saisically significan a five percen. 6 This resul is consisen wih Rodrik s (1995) hypohesis ha expor orienaion is he resul of an increase in demand for impored capial goods for many of he foreign exchange consrained developing counries. 39
Cenre for Promoing Ideas, USA www.ijbhne.com The coefficien of 0.11 indicaes ha adjusmen owards long run equilibrium is approximaely 11 percen each year. Given he presence of only one coinegraing relaionship, he error correcion specificaion can be used o es for Granger causaliy. Resuls indicae ha expors Granger-causes non-expor GDP a he 10 percen level (p-value = 0.08). This resul is evidenced by Mexico s expor drive, which is mainly based on he dynamism of manufacured expors. GATT and especially NAFTA led o a rapid growh in manufacures and, as a resul, by 2003 oal expors represened 34.9 percen of Mexico s real GDP. In addiion, since 2000, here has been a renewed imporance of oil revenue as a resul of increased global oil prices. However, resuls do no sugges ha growh leads expors (GLE). Our findings agree wih Thornon (1996) and McCarville and Nnadozie (1995), who found unidirecional causaion running from expors o GDP. Table 3 shows ha in he shor run, invesmen and impors Granger-cause non-expor GDP a he five percen and one percen significance levels respecively. Causaion from impors o income gives credence o he view ha impors provide he capial and echnology o generae increased incomes, produciviy and economic growh. Applied sudies (see for example, Pacheco-López, 2004) reveal ha, in he las fifeen o weny years, he Mexican economy s srucural dependence on impors has increased significanly. His findings indicae ha he long-erm income-elasiciy of demand for impors (essenially manufacured goods) has more han doubled over his period; previously i was valued beween 1.2 percen and 1.5 percen and has now risen o almos hree percen. Pacheco-López furher noes ha i is doubful ha he upurn deeced hus far in Mexico s long-run income elasiciy of impors is permanen. I is more likely o abae, and hen decline o a cerain degree as some effecs of he rade liberalizaion process on he demand for foreign goods and services wear off. Bu if i remains a curren highs, he exernal secor will become a major obsacle in Mexico s sruggle o seer a pah of solid economic growh, away from recurren balance-of-paymens crises. Finally, he evidence indicaes ha labor, impors and expors Granger-cause invesmen. Table 3 also shows causaion running from invesmen, and impors o expors. The laer causal link is subsaniaed by he performance of he Maquiladoras indusry. As noed in Moreno-Brid e al. (2005), Mexico s vigorous expor drive has ranspired concurrenly wih a greaer use of echnological sophisicaion in he producion of some of is manufacured goods sold abroad. The resuls of he impulse response funcions and variance decomposiion lend suppor o he ELGH. Panel (a) of Figure 1 depics he ime pahs of he responses of non-expor GDP o shocks in labour, invesmen, impors and expors; i also illusraes he impac of non-expor GDP on expors as compared o oher variables (Panel (b)). The evidence shows ha he response of non-expor GDP o a shock in expors has a longer and sronger effec han he response of non-expor GDP o he oher variables. This is srong evidence for he ELGH in Mexico. On he oher hand, expors respond o is own innovaion wih he highes inensiy followed by impors. Table 4 shows he forecas error variance decomposiion a a 10-period-ahead forecas horizon. The findings agree wih hose from he impulse response funcion, indicaing ha expors explains approximaely 77.75 percen of he forecas error variance of non-expor GDP, implying i is he mos imporan facor affecing economic growh. On he oher hand, afer 10 periods approximaely 85.67 percen of he forecas error variance of expors is accouned for by is own innovaions, followed by an 11.57 percen response o shocks in impors. 6. Concluding Remarks The presen sudy underook a case sudy of Mexico s economic growh under expor-oriened policies. I also esed wheher he expor-led growh hypohesis holds for Mexico over he period 1960-2003. To overcome he problem of specificaion bias from which previous sudies have suffered, an expor-augmened neoclassical producion funcion was employed. In his framework, oal facor produciviy was assumed o be a funcion of expors and impors of goods and services. The oupu variable of he funcion was defined ne of expors o separae he influence of expors on oupu from ha incorporaed ino he naional income ideniy. The echniques of coinegraion and error correcion, and innovaion accouning analysis were employed. The empirical evidence indicaes ha in he shor run expors Granger-causes non-expor GDP. On he oher hand, he resuls indicae a negaive relaionship beween expors and non-expor GDP. This is explained by he high impor conen and diminishing local conen of expors, and weak linkages wih domesic suppliers, hus reducing possible spillover or muliplier benefis. Hence, even if expors growh is high, he ne simulus o he overall growh of he economy and o job creaion is relaively small. Currenly, he world is headed ino an era of high and rising energy prices. Therefore, Mexico could ry o ake advanage of his siuaion, while avoiding he misakes of he pas. Oil revenues should be re-invesed in he developmen of he energy secor or invesed in oher vial infrasrucure needs; 40
Inernaional Journal of Business, Humaniies and Technology Vol. 1 No.1; July 2011 hey should no be divered o pay for general governmen expendiures or oher uses such as leveraging large inernaional debs, as occurred in he lae 1970s. Mexico could also seek o fully exploi he possibiliy of producing naural gas, as he capaciy is available for hem o do so, hereby increasing he benefis of high prices in oil expors. From a policy perspecive, he Mexican economy can no longer base is place in he global economy on low wages and maquiladoras. Therefore, if Mexico is o succeed in is ques o achieve high and seady economic growh, i urgenly needs o rehink key elemens of is overall sraegy and indusrial policies. In paricular, curren incenive schemes ha allow ax-free enry of impored inpus and raw maerials for expor purposes mus be reconsidered. Finally, policies ha promoe echnological innovaion in manufacuring and linkages wih local suppliers are imperaive. References Alam, M.I. 2003. Manufacured Expors, Capial Good Impors, and Economic Growh: Experience of Mexico and Brazil. Inernaional Economic Journal 17, no.4, 85-105. Awokuse, T.O. 2003. Is he Expor-led Growh Hypohesis valid for Canada? Canadian Journal of Economics 36, no.1, 126-36. Bahmani-Oskooee, M., and J. Alse. 1993. Expor Growh and Economic Growh: An Applicaion of Coinegraion and Error-correcion Modeling. Journal of Developing Areas 27, 535-42. Balassa, B. 1978. Expors and Economic Growh: Furher Evidence. Journal of Developmen Economics 5, no. 2, 181-89. Balassa, B. 1985. Expors, Policy Choices, and Economic Growh in Developing Counries afer he 1973 Oil Shock. Journal of Developmen Economics 4, no. 1, 23-35. Barro, R. J., and X. Sala-i-Marin. 1995. Economic Growh. New York: McGraw-Hill. Blecker, R.A. 2006. Macroeconomic and Srucural Consrains on Expor-led Growh in Mexico. Deparmen of Economics, American Universiy Working Paper no. 2006-03. Bolho, A. 1996. Was Japanese Growh Expor-led? Oxford Economic Papers 48, no. 3, 415-32. Buffie, E.F. 1992. On he Condiion for Expor-led Growh. Canadian Journal of Economics 25 no. 1, 211-25. Chang, T.; W. Fang; W. Liu; and H. Thompson. 2000. Expors, Impors and Income in Taiwan: An Examinaion of he Expor-led Growh Hypohesis. Inernaional Economic Journal 14, no. 2, 151-60. Chow, P.C.Y. 1987. Causaliy beween Expor Growh and Indusrial Developmen: Empirical Evidence from he NICs. Journal of Developmen Economics 26, no. 1, 55-63. Dickey, D.A., and W.A. Fuller. 1979. Disribuions of he Esimaors for Auoregressive Time Series wih a Uni Roo. Journal of American Saisical Associaion 74, no. 366, 427-31. Dickey, D.A., and W.A. Fuller. 1981. Likelihood Raio Saisics for Auoregressive Time Series wih a Uni Roo. Economerica 49, no. 4, 1057-72. Edwards, S. 1992. Trade Orienaion, Disorions and Growh in Developing Counries. Journal of Developmen Economics 39, no. 1, 31-57. Emery, R.F. 1968. The Relaion of Expors and Economic Growh: A Reply. Kyklos 21, no. 4, 757-60. Feder, G. 1983. On Expors and Economic Growh. Journal of Developmen Economics 12, no. 2, 59-73. Granger, C.W.J. 1969. Invesigaing Causal Relaions by Economeric Models and Cross-specral Mehods. Economerica 37, no. 3, 424-38. Grossman, G.M. and E. Helpman. 1991. Innovaion and Growh in The Global Economy. Cambridge, Mass: MIT Press. Heller, P.S., and R.C. Porer. 1978. Expors and Growh: An Empirical Re-invesigaion. Journal of Developmen Economics 5, no. 2, 191-193. Herzer, D., Nowak-Lehmann, F. and Silversovs, B. (2006) Expor-led Growh in Chile: Assessing he Role of Expor Composiion in Produciviy Growh. The Developing Economies 44, no. 3, 306-28. Inernaional Labor Office. Yearbook of Labor Saisics, various issues. Islam, M.N. 1998. Expor Expansion and Economic Growh: Tesing for Coinegraion and Causaliy. Applied Economics 30, no. 3, 415-25. Jaffee, D. 1985. Expor Dependence and Economic Growh: A Reformulaion and Respecificaion. Social Forces 64, no. 1, 102-18. Johansen, S. 1988. Saisical Analysis of Coinegraing Vecors. Journal of Economic Dynamics and Conrol 12, no. 2-3, 231-54. Johansen, S., and K. Juselius. 1990. Maximum Likelihood Esimaion and Inference on Coinegraion wih Applicaions o he Demand for Money. Oxford Bullein of Economics and Saisics 52, no. 2, 169-210. Jung, W.S., and P.J. Marshall. 1985. Expors, Growh and Causaliy in Developing Counries. Journal of Developmen Economics 18, no. 1, 1-12. Kavoussi, R.M. 1984. Expor Expansion and Economic Growh: Furher Empirical Evidence. Journal of Developmen Economics 14, no. 1, 241-50. 41
Cenre for Promoing Ideas, USA www.ijbhne.com Kohli, I., and N. Singh. 1989. Expors and Growh: Criical Minimum Effor and Diminishing Reurns. Journal of Developmen Economics 30 no. 2, 391-400. Kravis, I.B. 1970. Trade as a Handmaiden of Growh: Similariies beween he Nineeenh and Twenieh Cenuries. Economic Journal 80, no. 320, 850-70. Krueger, A.O. 1978. Foreign Trade Regimes and Economic Developmen: Liberalizaion Aemps and Consequences. Cambridge, Mass: Ballinger. Kwiakowski, D.; P.C.B. Phillips; P. Schmid; and Y. Shin. 1992. Tesing he Null Hypohesis of Saionariy agains he Alernaive of a Uni Roo. Journal of Economerics 54, no. 1-3, 159-78. Li, C.A. 1999. Expor Diversificaion and Expor-led Growh in Mexico. Deparmen of Economics, Universiy of Essex Economics Discussion Paper no. 503. Lucas, R. 1988. On he Mechanics of Economic Developmen. Journal of Moneary Economics 22, no. 1, 3-42. Lusig, N. 1998. Mexico: The Remaking of an Economy. Washingon: Brookings Insiuion. McCarville, M., and E. Nnadozie. 1995. Causaliy Tess of Expor-led Growh: The Case of Mexico. Alanic Economic Journal 23, no. 2, 140-45. Moon, B.E. 1998. Expors, Ouward-oriened Developmen, and Economic Growh. Poliical Research Quarerly 51, no. 1, 7-36. Moreno-Brid, J.C.; J.C. Valdivia; and J. Sanamaria. 2005. Mexico: Economic Growh Expors and Indusrial Performance afer NAFTA. hp://www.wilsoncener.org/news/docs/ Mexico_afer_NAFTA_ECLAC.pdf (accessed May 31, 2007). OECD. (2002) Economic survey: Mexico. (Paris). Pacheco-López, P. 2004. The Impac of Trade Liberalizaion on Expors, Impors, The Balance of Paymens and Growh: The Case of Mexico. Working Paper no. 401. Canerbury: Universiy of Ken. Phillips, P., and P. Perron. 1988. Tesing for Uni Roo in Time Series Regression. Biomerika 75, no. 2, 335-46. Riezman, R.G.; C. Whieman; and P.M. Summers. 1996. The Engine of Growh or is Handmaiden? A Time Series Assessmen of Expor-led Growh. Empirical Economics 21, no. 1, 77-110. Rodrik, D. 1995. Geing Inervenions Righ: How Souh Korea and Taiwan Grew Rich. Economic Policy 20, 53-97. Romer, P. 1986. Increasing Reurns and Long Run Growh. Journal of Poliical Economy 94, no. 5, 1002-37. Romer, P. 1989. Wha Deermines he Rae of Growh and Technological Change? World Bank Working Paper no. 279. Ros, J. 1994. Mexico s Trade and Indusrializaion Experience since 1960. In: Trade Policy and Indusrializaion in Turbulen Times, ed. G. K. Helleiner. London: Rouledge. Thornon, J. 1996. Coinegraion, Causaliy and Expor-led Growh in Mexico, 1895-1992. Economics Leers 50, no. 3, 413-16. Tyler, M. 1981. Growh and Expor Expansion in Developing Counries: Some Empirical Evidence. Journal of Developmen Economics 9, no. 1, 121-30. Unied Naions Conference on Trade and Developmen (UNCTAD). 2002. Trade and Developmen Repor, 2002. Geneva: Unied Naions. Williamson, J. 1997. Mexican Policy Toward Foreign Borrowing. In Coming Togeher? Mexico-U.S. Relaions, ed. B. Bosworh, S.M. Collins and N.C. Lusig. Washingon: Brookings Insiuion. 42 Table 1: Uni Roo Tess Variable ADF PP KPSS Conclusion LNY (a) -2.217(1) (a) -1.970 (a) 0.161** I(1) (b) -3.368(0)** (b) -3.278** (b) 0.259 LK (a) -2.246(1) (a) -1.908 (a) 0.158** I(1) (b) -3.548(0)** (b) -3.508** (b) 0.250 LL (a) -1.423(1) (a) 0.206 (a) 0.182** I(2) (b) -1.111(0) (b) -1.231 (b) 0.354* LM (a) -2.214(1) (a) -1.813 (a) 0.143* I(1) (b) -3.457(0)** (b) -3.424** (b) 0.220 LX (a) -2.505(1) (a) -1.853 (a) 0.153** I(2) (b) -2.170(0) (b) -2.170 (b) 0.209 Noes: The saisics for he level of he variable is given in (a) and for he firs difference in (b). *** indicaes saisical significance a he one percen level; ** indicaes saisical significance a he five percen level;* indicaes saisical significance a he en percen level. T he erms in parenheses are he opimal number of lags chosen by he SIC.
Inernaional Journal of Business, Humaniies and Technology Vol. 1 No.1; July 2011 Table 2: Resuls from Johansen s Coinegraion Tes Null Hypohesis Alernaive Hypohesis Tes Saisic 5% Criical Value Trace Tes r = 0 r < 1 70.590** 60.061 r = 1 r < 2 37.860 40.175 r = 2 r < 3 17.664 24.276 r = 3 r < 4 6.605 12.321 r = 4 r < 5 0.652 4.130 Max Eigenvalue Tes r = 0 r = 1 37.730** 30.440 r = 1 r = 2 20.196 24.159 r = 2 r = 3 11.060 17.797 r = 3 r = 4 5.953 11.225 r = 4 r = 5 0.652 4.130 Noes: ** indicaes saisical significance a he five percen level. Table 3: Long-run and Shor-run Resuls LONG-RUN: COINTEGRATING EQUATION LNY 0.94 LK ** 0.03LL 2.24 LM ** 2.10 LX ** [1.92] [-0.26] [3.50] [-5.60] SHORT-RUN: VECTOR ERROR CORRECTION MODEL Independen variable LNY LK LL LM LX LNY - 1.162 (0.281) 0.175 (0.676) 0.666 (0.415) 2.264 (0.132) LK 4.063** (0.044) - 2.070 (0.996) 1.450 (0.229) 3.680* (0.055) LL 0.812 (0.367) 3.179* (0.075) - 1.656 (0.198) 0.677 (0.411) LM 11.872*** (0.001) 12.485*** (0.000) 0.043 (0.835) - 3.675* (0.055) LX 3.086* (0.080) 3.652* (0.056) 0.359 (0.549) 2.277 (0.131) - ECT 0.069 [0.987] 0.138 [1.601] -0.001 [-0.659] 0.107 [1.258] -0.114** [-1.900] Noes: For he coinegraing equaion, -saisics are in parenheses. For he shor-run resuls, ECT is he one period lagged error erm from he coinegraing equaion. T-saisics are in square parenheses [.] and p-values from chi-square ess are in circular parenheses (.). *** indicaes saisical significance a he one percen level; ** indicaes saisical significance a he five percen level; and * indicaes saisical significance a he en percen level. 43
Cenre for Promoing Ideas, USA www.ijbhne.com Table 4: Variance Decomposiion of 10-period-ahead Forecas Error Variance Shock o: Percenage of Variance Explained by Shock LNY LL LK LM LX LNY 13.353 8.911 9.362 4.185 2.708 LL 0.004 75.471 0.059 0.038 0.004 LK 2.206 14.728 0.408 0.341 0.044 LM 6.681 0.540 5.532 13.439 11.568 LX 77.756 0.351 84.639 81.998 85.677 Figure 1: Impulse Response o Cholesky One S.D. Innovaions.35.30.25.20 Response of LNY o Cholesky One S.D. Innovaions LX.40.35.30.25 Response of LX o Cholesky One S.D. Innovaions LX.15 LNY.10 LM.05.00 LL -.05 LK -.10 1 2 3 4 5 6 7 8 9 10 Panel (a).20.15.10 LM.05 LNY.00 LK LL -.05 1 2 3 4 5 6 7 8 9 10 Panel (b) Noes: S.D. means sandard deviaion. Variable labels on each curve indicae ha he response of LNY (Panel (a)) and LX (Panel (b)) respecively were as a resul of a shock o hese variables. 44