IMMIGRATION VERSUS OUTSOURCING: A DEVELOPING COUNTRY S VIEW

Transcription

1 JURNAL ECNMC EVEPMENT 109 Volume 37, Number 2, Jue 2012 MMGRATN VERSUS UTSURCNG: A EVEPNG CUNTRY S VEW SMNTN AS a, AJTAVA RAYCHAUHUR b AN SAKAT SNHA RY c a Rabidra Bharati Uiversity, dia b,c Jadavpur Uiversity, dia This paper provides a comparative study betwee temporary immigratio policy ad product outsourcig process, from the low-icome developig coutry s poit of view, which is supply side costraied by the availability of skilled labour. A two-coutry geeral equilibrium model establishes a iverse relatioship betwee temporary immigratio quota ad product outsourcig. Though temporary immigratio quota ehaces world welfare ad the developed coutry welfare, its impact o welfare level of the developig coutry is ucertai. the empirical part, a pael data aalysis shows that real cosumptio level of a set of developig coutries icreases with a icrease i product outsourcig, give a iverse relatioship betwee product outsourcig ad temporary immigratio policy. Keywords: mmigratio, utsourcig, termediate put, Cosumptio, Pael Aalysis JEL classificatio: C23, E20, 2, 12, NTRUCTN Covetioally iteratioal trade betwee a developed ad a developig coutry, from a theoretical perspective, is based o the differece i relative factor edowmets amog them. teratioal trade helps utilise the abudat factor. teratioally tradig coutries gai from relative factor abudace ad lower relative price of factor iputs. or istace, a relative labour abudat developig coutry use the abudat factor ad the gais from trade follow from lower relative factor price of labour. recet times, trade i itermediate goods ad factor iputs have assumed sigificace alog with covetioal merchadise trade i fial goods. or a relatively capital abudat coutry, with regards to trade i factors iputs, the policy optios are betwee import of low-priced foreig labour ad export of domestic capital. t has bee theoretically We are grateful to a aoymous referee for costructive commets ad suggestios.

2 110 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY established, a la Ramaswami (1968), that importatio of low-priced foreig labour is a better optio tha exportatio of domestic capital for capital abudat developed coutries. Now the questio is, What is the choice for a labour abudat coutry: exportatio of domestic labour or importatio of foreig capital? This paper ivestigates ito the relatioship betwee exports of domestic labour, that is, temporary migratio, ad product outsourcig from the perspective of a developig coutry. Migratio ad outsourcig are the two maifestatios of trade i factor iputs: while migratio icludes cross-border movemet of atural persos, iteratioal outsourcig 1 ivolves cross-border movemet of capital. Both immigratio ad iteratioal outsourcig icrease cost-efficiecy of the global productio system (Murat ad Paba, 2003). The labour-itesive part of the productio process is reallocated from the developed coutry to the labour abudat developig coutry; the developig coutries, egaged i such outsourcig cotract, become a part of the globally itegrated productio chai. 2 or istace, a growig outsourcig demad is directed towards dia from differet capital abudat coutries icludig the USA. There are two types of outsourcig; product outsourcig ad service outsourcig. Globally while product outsourcig takes place i egieerig, automobile, textile, ad chemical idustries (ubey, 2003), 3 service outsourcig takes place i commuicatio, iformatio techology ad iformatio techology eabled sectors. The developig coutries, through outsourcig ot oly expad their market for itermediate products, but also beefit from employmet geeratio, profit maximisatio, quality up-gradatio, capacity utilisatio skill-improvisatio ad techology trasfer. mmigratio is aother way by which the pool of high-skilled labour i developig coutries is beig used by developed coutries. Temporary immigratio refers to the temporary movemet of persos from oe coutry to aother to provide o-site services, whe immigrats do ot get the right to dwell permaetly i the host coutry. eveloped coutries ecourage such temporary immigratio o accout of lower wages of skilled workers from developig coutries. Employers i developed coutries also 1 utsourcig ca also be betwee formal ad iformal sectors withi the domestic boudary of a developig coutry. Maiti ad Marjit (2011) develop a theoretical model showig that the choice of iformal sector subcotractig ad i-house R& ivestmet appears to be alterative optios to the firm to bypass expesive labour i the formal sector give that the formal sector wage is higher tha that of iformal sector. The paper theoretically shows that R& expediture ad labour productivity of formal sector deped o the wage rate of the iformal sector. 2 Sayal (1983) provides a framework to aalyse a coutry s specialisatio patter alog the vertical productio spectrum of a good. Sayal ad Joes (1982) further show that trade i middle products icreases the availability of iputs i the productio of fial cosumptio goods. 3 ubey (2003) provides evidece usig Cofederatio of dia dustry (C) sources that dia is estimated to receive $ 10 billio worth outsourcig order i maufacturig sector by 2007, which is expected to icrease by US $ 50 billios by 2015.

3 MMGRATN VERSUS UTSURCNG 111 prefer temporary immigrats over permaet oes o accout of the legthy process ivolved i permaet immigratio. Temporary immigratio process ot oly provides flexibility i the productio process, but also allows such migrats from developig coutries to ear higher wages tha their couterparts i the home coutry. 4 Most of the existig literature discusses immigratio ad outsourcig idepedetly, studyig separately the impact of these two o labour market. Cohe-Golder ad Paserma (2004), Amuri, ttaviao ad Peri (2008), ustma, abbri ad Presto (2005), Jea ad Jimeez (2007), Loghi, Nijkamp ad Poot (2006), Robertso (2008), Tu (2010) ad Wadsworth ad e Coulo (2008) show that immigratio, i geeral, has o adverse effect o wage ad employmet level of developed coutries i the log ru. However, as Cohe-Golder ad Paserma (2004) ad Jea ad Jimeez (2007) show, there ca be some short term adverse effects of immigratio o the host coutry s wage level. Loghi, Nijkamp ad Poot (2006) fid that loss of employmet opportuities i host coutries o accout of large immigratio iflux is egligible, but the itesity of such egative effect varies across coutries ad geder. There were may other factors like aticompetitive product market regulatios, striget employmet protectio legislatio ad high average replacemet rate of uemploymet beefits, which aggravate such effects (Jea ad Jimeez, 2007). The results of these studies imply that host coutry workers ad migrat workers are imperfect substitutes. cotrast, Amuri, ttaviao ad Peri (2008) show that as ew ad old immigrats are perfect substitute to each other i the cotext of West Germay, immigratio leads to reductio i employmet opportuity ad wage level of previous immigrats. Wadsworth ad e Coulo (2008) highlight that relaxatio of immigratio quota, by the developed coutries, reduces the iteratioal wage differece. There is a debate o the labour market effect of outsourcig. 5 Ardt (2002), Badyopadhyay, Marjit ad Yag (2010) ad Koskela ad Stebacka (2007) show that despite labour market imperfectios outsourcig icreases the aggregate employmet of the host coutry. the other had, Egger ad Kreickemeir (2008) ad Hijze, Görg ad Hie (2004) show that iteratioal outsourcig reduces the demad for host coutry workers. specific, Hijze, Görg ad Hie (2004) fid that outsourcig reduces the demad for low-skilled host coutry workers, whereas Koskela ad Stebacka (2007) show that outsourcig icreases the demad for low skilled workers but reduces that for high skilled workers i developed coutries. Badyopadhyay, Marjit ad Yag 4 This H-1B visa issued by U.S. Govermet ca be take as a measure of temporary immigrats (Jase ad Piermartii, 2005; Badyopadhyay ad Wall, 2005). dia is oe of the importat source coutries of the H-1B migrats. 2000, accordig to U.S cesus, dia set 124,647 H-1B migrats to the U.S. Amog the developig coutries, Chia ad Pakista are also two major seders of the H-1B migrats to the US. 5 There exist a wide rage of studies delvig ito other aspects of impact of outsourcig. or istace, Li (2008) empirically shows that outsourcig arrows the techological gap betwee firms of host coutries ad developig coutries.

4 112 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY (2010) theoretically show that a outsourcig tax has two opposite effects o host coutry employmet level. While substitutio effect of outsourcig tax icreases domestic employmet, its output effect o accout of loss i iteratioal competitiveess reduces the domestic employmet. However, they foud that output effect was stroger tha substitutio effect i most of the cases. Eve though most of the above-metioed studies aalyse the labour market effect of outsourcig from the perspective of developed coutries, Arora ad Chakrabarti (2004), Egma (2007) ad Khalifa ad Megova (2010) discuss the pheomeo of outsourcig from a developig coutry perspective. While Egma (2007) discusses various supply costraits to outsourcig of Busiess Process Services (BPS) ad formatio Techology Services (TS) i four emergig ecoomies like Chia, Czech Republic, dia ad Philippies, Arora ad Chakraborty (2004) show that outsourcig raises the wage differece betwee skilled ad uskilled workers i dia maufacturig. Khalifa ad Megova (2010) theoretically explai that there is a threshold level of skill abudace i developig coutries, beyod which outsourcig icreases wage iequality. However, empirical estimates for 29 developig coutries over the time period show variatios i result: while outsourcig reduces wage iequality below the threshold level of skill abudace, it has o effect o wage iequality above that level. The existig literature shows that temporary immigratio is a substitute of outsourcig (Badyopadhyay ad Wall, 2005; Egma, 2007; Murat ad Paba, 2003; ad Navaretti, Bertola ad Sembeelli, 2008). A icrease i the temporary immigratio quota implyig the reductio i the outsourcig demad by a developed coutry s govermet has a positive impact o her ow atioal icome (Badyopadhyay ad Wall, 2005). Beig the complemetary factor to domestic labour, temporary immigratio icreases the domestic wage level ad domestic employmet level of the developed coutry (Joes, 2005). Temporary immigratio policy is thus a better optio for a developed coutry. However, the outcome ca be differet for ay developig coutry. While temporary immigratio geerates oly labour demad, outsourcig geerates demad for labour as well as other factor iputs i a developig coutry. the other had, i a developed coutry, with higher productivity of immigrats i the presece of good quality ifrastructure, better workig eviromet, etc., the outsourcig demad declies. As a result, derived labour demad by the outsourcig firms i the developig coutry decreases. A icrease i temporary immigratio quota is thus expected to have a idirect egative impact o the labour market of a developig coutry. the basis of the above issues that emerge from the existig literature, this paper aalyses the impact of immigratio ad outsourcig o welfare of developig coutries. this cotext we should metio that Badyopadhyay ad Wall (2005) discuss about the relatioship betwee service outsourcig ad temporary immigratio, 6 while this 6 Khalifa ad Megova (2010) also cosider service outsourcig measured by the amout of directed

5 MMGRATN VERSUS UTSURCNG 113 paper aalyses the relatioship betwee product outsourcig ad temporary immigratio. fact, i this paper the mai emphasis is o developig coutry, while Badyopadhyay ad Wall (2005) focus o developed coutry. Theoretical as well as empirical aalyses have bee doe i the paper. The results show that temporary immigratio policy of the developed coutries sigificatly reduces the product outsourcig. The product outsourcig is better for the low-icome coutries tha the temporary migratio. Secod ad third sectios of the paper cotai theoretical ad empirical aalyses respectively. The last part summarises the major fidigs of this paper with some policy implicatios. this paper, immigratio ad temporary immigratio are syoymous, ad outsourcig ad product outsourcig are also used iterchageably to idicate the same meaig Model escriptio 2. THERETCAL ANALYSS A simple two coutry geeral equilibrium model, without ay skill variatio ad techological differeces across the coutry, is adopted to illustrate the immigratio versus outsourcig debate. The model uses a modified versio (as highlighted i the appropriate places) of a geeral equilibrium structure proposed i Badyopadhyay ad Wall (2005). Usually geeral equilibrium framework provides wider scope to aalyse welfare by cosiderig the demad ad the supply side forces ad their joit iteractio o all factor iputs as well as fial cosumptio good markets. This paper ot oly addresses the welfare issue with respect to the migratio versus outsourcig debate from the aggregate perspective (world output level), but also from the distributive perspective (allocatio of world output betwee developed ad developig coutries i terms of their respective atioal icomes). The world ecoomy is assumed to be divided ito a developed ecoomy (kow as foreig coutry) ad a developig ecoomy (kow as domestic coutry). Here the productio fuctio is of eoclassical type, exhibitig positive but dimiishig margial productivity of factor iputs ad costat returs to scale. The firm i the foreig coutry produces sigle fial cosumptio good (Q) by usig foreig labour (N) ad a itermediate iput (). The eoclassical type productio process for the fial cosumptio good is give by, Q Q(, N), (1) from the developed to the developig coutry.

6 114 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY where Q, Q N 0 ; Q, Q NN 0 ; 7 Q N Q N 0. The foreig firm produces part or whole of her itermediate iput requiremet ( ) usig foreig capital ( K ) ad temporary immigrat workers () from the domestic coutry. The foreig firm s productio fuctio for the itermediate iput is give by, ( K, ), (2) where K, 0 ; KK, 0 ; K K 0. Alteratively, the foreig firm ca outsource the part or the whole of her itermediate iput requiremet ( ) to the domestic coutry, 8 where the represetative domestic firm produces it usig domestic capital ( K ) ad domestic labour ( L ). omestic firm s productio fuctio of the outsourced itermediate iput is give by, K, L ), (3) ( where K, 0 ; KK, 0 ; K K 0. Productio of total itermediate good () is give by,. (4) Apparetly there is o techical differece i producig ad across the coutries. The temporary immigrats have higher productivity tha the domestic workers (so ) o accout of larger stock of capital ad better quality ifrastructure i the foreig coutry. As a result, i the competitive eviromet, immigrats ear higher wage tha the domestic workers, employed i the domestic outsourcig sector. This wage differece esures migratio from domestic to the foreig coutry. The profit fuctio for the represetative foreig firm is give by, Q( ( K, ), N) r K P W N W, k N 7 Here subscript deotes the margial product of that factor iput. 8 The itermediate good is either outsourced or produced by immigrat workers. As a result, developed coutry workers are idirectly complemet to the immigrat workers ad outsourcig. This is ulike Badyopadhyay ad Wall (2005), where developed coutry workers, immigrats ad outsourced workers are substitutes to each other.

7 MMGRATN VERSUS UTSURCNG 115 where P price of outsourced product, WN wage of the foreig worker, rk ret of the foreig capital ad wage of the immigrats. To simplify the model, fial W cosumptio good is treated as uméraire. Therefore all the prices are expressed i terms of fial cosumptio good (Q). Hece, the optimisatio problem ca be stated as Maximise P Q( ( K, ), N) r K P W N W, Q k N where, N, K, 0. The profit maximisatio coditios lead to Q P, (5) Q r, (6) K k Q W, (7) QN W N. (8) These coditios give the foreig firm s optimum demad for itermediate iput d d d ( ), foreig capital ( K ), foreig labour ( N ), ad temporary immigrat workers d ( ).The factor prices are determied assumig full employmet ad perfect competitio, ad the retur (price) to a factor is the value of the margial product of that factor i productio. oreig labour supply is fixed at N ad i equilibrium d N N. (9) Similarly, the equilibrium coditio for foreig capital i short ru is give by, d K K. (10) the other had, i the immigratio market, give social ad political factors related to temporary immigratio quota, the foreig coutry ofte restricts the umber of temporary immigrats. this model temporary immigratio quota is cosidered to be exogeous beig decided at. The equilibrium coditio i the immigratio market is

8 116 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY d. (11) the itermediate goods market, there are two sources of supply the foreig coutry firm beig producig a part of her itermediate iput requiremet ( ) ad the rest is outsourced from the domestic coutry firm. The foreig firm s supply of itermediate good is give by S ( K, ). (12) The represetative domestic firm, ulike Badyopadhyay ad Wall (2005), produces oly the outsourced itermediate iput ad o fial cosumptio good. All available domestic factor iputs are ivolved i the productio of outsourced itermediate iput. Suppose K ad L are the total capital ad labour edowmet of the domestic coutry respectively. the presece of high wage i the foreig coutry, some domestic workers (say ) migrate temporarily to the foreig coutry. The supply of itermediate goods by the domestic coutry firm is as, S ( K, L ). (13) Equatios (12) ad (13) give the total supply of the itermediate iput ( ). S S S ( K,, K, L ). 9 (14) The equilibrium coditio for the itermediate iput market is, d S. (15) Equatios (12), (13), (14) ad (15) give the equilibrium, fuctio of L, K,, ad K., ad P as the The profit fuctio for the represetative domestic firm is, P wl rk, 9 The supply of the itermediate iput is idepedet of P, rather is determied by the capital edowmet of both the coutries ad the labour allocatio of the domestic coutry (sice the possibility of the foreig workers, to be egaged i the productio of the itermediate iput is ruled out).

9 MMGRATN VERSUS UTSURCNG 117 where w wage of the domestic workers, ivolved i the domestic outsourcig sector ad r domestic retal. Hece the optimisatio problem of the domestic coutry firm is Maximise P ( K, L ) wl rk. ( K, 0 ) L Profit maximisatio coditios lead to, P w, (16) P K r. (17) These equilibrium coditios, alog with full employmet assumptio, give the d d optimum demad for domestic labour ( L ) ad domestic capital ( K ) of the domestic outsourced sector. This is give by L d L, (18) d K K. (19) Equatios (9), (10), (11), (18) ad (19) give the equilibrium W N, ad r as fuctios of L, K,, K, ad N. Equilibrium fial cosumptio good productio ca thus be restated as, Q Q ( L, K,, K, N ) Welfare of dividual Coutries r k, W, Real or actual cosumptio level of ay ecoomy ca be cosidered as a good represetative measure of the welfare level of that ecoomy at the aggregate level. With o savig ad taxatio, the atioal cosumptio is equivalet to the atioal icome, which cosists of the aggregate icome of that coutry s factor iputs either employed withi the coutry or outside the coutry. The atioal icome of the foreig coutry (Y ) is, w Y N k W N r K. Uder perfect competitio, productio fuctio exhibitig CRS ad applyig product exhaustio theorem we get,

10 118 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY Y N W N P W. (20) The atioal cosumptio of the foreig coutry is give by, C N Y W N P W. (21) the other had, atioal icome of the domestic coutry ( Y ), which icludes icome of the emigrats i additio to the icome of workers ad capital employed i the domestic coutry firm, is give by Y r K w ( L ) W. Uder competitio, productio fuctio exhibitig CRS ad applyig product exhaustio theorem we get, Y P W. (22) The atioal cosumptio of the domestic coutry is give by, C Y P W. (23) Global welfare level is measured by the total productio of fial cosumptio good ( Q ), sice the total fial cosumptio good produced is cosumed by the labour employed i both the coutries. Q is distributed betwee the developed ad developig ecoomies as their respective atioal icome or atioal cosumptio (i the absece of savig ad taxatio). Q Y Y. (24) or, Q C C. 10 (25) The solutio set ca be iterpreted as the reduced form equatios, sice all edogeous variables are expressed i terms of exogeous variables,, K, ad N. L, K, 10 a simple two-ecoomy model, it ca be iterpreted that the total world welfare is the sum of the idividual coutry s welfare.

11 MMGRATN VERSUS UTSURCNG Comparative Static Cosider a chage i the immigratio policy of the developed coutry. This chage will affect all the factor iput markets, fial cosumptio good market, ad welfare level of both the ecoomies i the two coutry geeral equilibrium model. Propositio 1: A icrease i the temporary immigratio quota icreases total productio as well as the demad for itermediate iput ( ), but reduces the equilibrium price of the itermediate iput ( P ). With a icrease i the temporary immigratio quota, domestic workers migrate to the foreig coutry, sice W w. As a result the productio of the outsourcig sector is reduced by amout per emigrat, sice this is supply determied equilibrium. This reduces the supply of the outsourced itermediate iput by the domestic coutry. So there exists a iverse relatio betwee immigratio quota ad the outsourced itermediate iput. the preset paper supply side costraits determie the iverse relatioship betwee immigratio ad outsourcig, while Badyopadhyay ad Wall (2005) show that demad side factors determie this iverse relatioship. 11 the other had, productio of the itermediate iput i the foreig coutry is icreased by amout per emigrat. As, so there is a et icrease i productivity by ( ) amout per emigrat. Therefore total productio of the itermediate iput is icreased by ( ) per emigrat. As a result, the equilibrium demad for itermediate iput icreases by ( ) d 12, sice there is full-employmet i the itermediate iput market. Therefore, d / d 0. While the demad for the itermediate iput remais uchaged, the supply of the itermediate iput icreases o accout of a icrease i. Therefore the equilibrium price of the itermediate iput falls. ifferetiatig Equatio (5) with respect to gives, dp / d Q ( ). Sice Q 0, ad ( ) 0, dp / d 0. Propositio 2: A icrease i the temporary immigratio quota reduces the wage of the immigrats. 11 Badyopadhyay ad Wall (2005) model is demad determied, while our model is supply determied. 12 d = chage i the immigratio quota.

12 120 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY The supply of the temporary immigrat workers icreases with the relaxatio of immigratio policy of the foreig govermet. the other had, a icrease i the immigratio quota reduces the equilibrium price of the itermediate iput. The value of margial physical product for immigrat workers 13 (or demad for immigrat workers) falls resultig i the fall of the equilibrium wage of the immigrat. ifferetiatig Equatio (8) with respect to we obtai, dw N / d Q ( ). N Sice ( ) 0 ad Q 0, dw / d N 0. N Propositio 3: Eve though a icrease i temporary immigratio quota icreases foreig worker s wage, the impact o equilibrium retal of foreig capital is ambiguous. A icrease i immigratio quota raises the equilibrium demad for the itermediate iput i the productio of fial cosumptio good. This icreases the productivity of foreig workers i the fial cosumptio good productio, due to the complemetary relatioship betwee two factor iputs. Therefore the value of the margial physical product of foreig workers (or the demad for the foreig workers) icreases. Give the fixed supply of the foreig workers at N, the equilibrium foreig wage icreases. This result cotradicts with Badyopadhyay ad Wall (2005) which show that foreig wage decreases with a icrease i immigratio quota. ifferetiatig Equatio (7) with respect to gives, dw / d Q { } Q. Sice ( ) 0 ad Q, 0, dw / d 0. oreig capital is oly used i the productio of itermediate iput by the foreig firm. With a icrease i the temporary immigratio quota, the equilibrium price of itermediate iput falls. This fall i price reduces the value of margial physical product of foreig capital. 14 the other had, productivity of foreig capital icreases o accout of the icreasig use of complemetary factor iput temporary immigrats i the productio process. The combied effect of these two opposite forces may icrease, decrease or may ot chage the demad for foreig capital. Give the fixed supply of 13 The value of margial physical product of a factor iput ca be iterpreted as the demad of that factor iput, i the presece of the perfect competitio i fial cosumptio good market. 14 Value of margial physical product of foreig capital= P K.

13 MMGRATN VERSUS UTSURCNG 121 foreig capital at K, the equilibrium retal o foreig capital may icrease, decrease or remai uchaged. This result is also i cotrast with Badyopadhyay ad Wall (2005), eve though i both the models capital ad immigrat workers are complemetary factors of productio. ifferetiatig Equatio (6) with respect to gives, k dr / d Q Q ( ). Ki Sice Q 0, ( ) 0 ad K 0, dr k / d 0. K Propositio 4: A icrease i temporary immigratio quota has ucertai impact o the wage rate of the domestic workers, but it defiitely reduces the equilibrium retal o domestic capital ivolved i domestic outsourcig sector. The equilibrium price of itermediate iput reduces with a icrease i. This reduces the value of margial physical product 15 (or derived demad) of domestic workers. the other side, the supply of labour to the domestic outsourcig sector falls as the domestic workers migrate to the foreig coutry. The equilibrium domestic wage may thus icrease, decrease or remai uchaged, ulike Badyopadhyay ad Wall (2005) showig a rise i immigratio quota raises domestic wage. ifferetiatig Equatio (16) with respect to gives, d( w ) / d Q [ ] P. Sice ( ) 0 ad Q, 0, d ( w )/ d 0. A icrease i immigratio quota reduces the labour availability to the domestic outsourcig sector. As a result, the productivity of the complemetary factor, domestic capital, also reduces. However, the price of itermediate iput falls. These two effects joitly reduce the value of margial physical product of domestic capital leadig to a fall i equilibrium domestic retal i the face of a fixed supply of domestic capital at K. ifferetiatig Equatio (17) with respect to gives, 15 Value of margial physical product of domestic workers= P, ad Value of margial physical product of domestic capital= P K. These ca be iterpreted as the respective demad of the domestic labour ad domestic capital.

14 122 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY d( r ) / d Q [ ] P. K K Sice Q 0, K, ( ) 0 ad K 0, d ( r )/ d 0. Lemma: The wage differece betwee temporary immigrats ad domestic workers falls o accout of a icrease i temporary immigratio quota. A icrease i immigratio quota reduces margial productivity of the immigrats, but icreases that of the domestic workers. The equilibrium price of the itermediate iput also falls. This reduces the absolute differece betwee the value of margial physical product of immigrats ad domestic workers, eve i the face of a ucertai chage i equilibrium domestic wage, give fall i the immigrats wage. Absolute wage gap betwee immigrats ad domestic workers is measured by ( W w ). Now ( W w ) P ( ). After differetiatig the above equatio with respect to we obtai, d( W 2 w ) / d Q ( ) P( ). (26) Sice Q,, 0, ( ) 0, d( W w )/ d 0. Propositio 5: A icrease i temporary immigratio quota icreases world welfare (measured i terms of fial cosumptio good productio). the regio where, a icrease i temporary immigratio quota icreases the total productio as well as cosumptio (i the absece of savig ad taxatio) of the fial cosumptio good. ce the world reaches, there will be o further immigratio from the domestic coutry to the foreig coutry. The wage differece will also disappear at. The productio of the fial cosumptio good is maximised at. ifferetiatig twice Q Q(, N ) with respect to we obtai, dq / d 2 Q [ ]. (27) 2 d Q / d 2 Q [ ] Q [ ].

15 MMGRATN VERSUS UTSURCNG 123 With ( ) 0, dq / d 0 ad d Q / d 0. This shows that the curve is positively sloped ad cocave to the origi. 2 2 Q Propositio 6: The foreig atioal icome (measured i terms of fial cosumptio good) icreases due to a icrease i temporary immigratio quota. The foreig atioal icome has two compoets: total wage icome eared by foreig workers ad total ret icome eared by foreig capital owers. With a icrease i immigratio quota, equilibrium foreig wage icreases, but equilibrium foreig retal may icrease, decrease or remai uchaged. The first effect is a domiatig oe, which outweighs the secod effect. ifferetiatig Equatio (20) with respect to gives, dy / d Q [ ][ ] Q. (28) Sice ( ) 0, ad Q, 0, dy / d 0. Equatio (28) shows that the foreig atioal icome is a mootoically icreasig fuctio of provided that. Lemma: Sice Y C, dc / d dy / d 0 accordig as. Propositio 7: A icrease i has ucertai impact o the domestic atioal icome (measured i terms of fial cosumptio good) ad domestic atioal cosumptio level. omestic atioal icome comprises of the value of the total outsourcig product ( P ) 16 produced i the domestic coutry ad total wage earig by the immigrats i the foreig coutry ( W ). With a icrease i the immigratio quota, the equilibrium supply of outsourced itermediate iput as well as the equilibrium price of the itermediate iput falls. As a result, P decreases. the other had, relaxatio of foreig immigratio policy defiitely reduces the equilibrium wage of immigrats ( W ). The total wage eared by the immigrats ( W ) may icrease, decrease or remai uchaged. The total wage earig by the immigrats would fall, provided the ow wage elasticity of the immigrats is greater tha oe. this situatio, temporary 16 P cosists of domestic wage compoet ad domestic retal compoet; with the icrease i domestic retal icreases but domestic wage compoet may icrease, decrease or remai uchaged.,

16 124 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY immigratio quota ad domestic atioal icome are iversely related. therwise, the impact of a icrease i the immigratio quota o the domestic atioal icome is ucertai. ifferetiatig Equatio 22 with respect to we obtai, dy / d [ ][ Q P Q ] Q. (29) Sice ( ) 0 ad Q, 0, dy / d 0. Lemma: Sice C Y, dc / d dy / d 0 accordig as. this model, rise i immigratio quota icreases world welfare ad developed coutry welfare. However it may reduce developig coutry welfare. Badyopadhyay ad Wall (2005) show that atioal icome of the developed coutry mootoically icreases with immigratio quota, but i the presece of moopsoy power of developed coutry i iteratioal labour market, atioal icome of the developed coutry may fall with rise i immigratio quota. our model, there is o such possibility. t also eeds to be metioed that Badyopadhyay ad Wall (2005) do ot aalyse the impact of rise i immigratio quota o developig coutry welfare. 3. EMPRCAL ANALYSS The theoretical model, detailed out i the earlier sectio, has established a iverse relatioship betwee temporary immigratio ad product outsourcig. However, the impact of a icrease i the temporary immigratio quota o the welfare level of the developig coutry remais ambiguous. These propositios ecourage us to do a empirical aalysis to determie the ature of impact of temporary immigratio ad product outsourcig o welfare of developig ecoomies. or this purpose, three reduced form equatios, derived i the theoretical model (after some modificatio), are estimated. The empirical aalysis is carried out i a partial equilibrium framework eve though applied geeral equilibrium welfare calculatio is a possibility Estimatig Model The whole empirical aalysis is divided ito three stages; i stage, we estimate

17 MMGRATN VERSUS UTSURCNG 125 relatio betwee the real product outsourcig ( VA ) 17 ad temporary immigratio quota (). The sig of d( VA ) / d at the mea value of regressor () will show the relatioship betwee product outsourcig ad temporary immigratio; i stage, relatio betwee icome of the immigrats ( W ) ad the umber of immigratio quota ( ) is estimated. The sig of d( W 18 )/d shows the ature of the relatioship betwee icome of the immigrat workers of i th coutry ad the total umber of immigrats; ad i stage, the impact of the real value of the product outsourcig o the real cosumptio level (which is a represetative measure of the welfare level) of a low-icome coutry is estimated. the geeral equilibrium model, the developig coutry has oly oe sector; the outsourcig sector. omestic labour ad domestic capital are egaged i the outsourcig sector. omestic labour has the optio to migrate to the developed coutry. There are may sectors i the developig coutry, which cotribute to atioal icome. or istace if we assume that there are three sectors i a developig coutry, amely agriculture, trasport ad commuicatio, o-outsourced maufacturig sectors apart from outsourced maufacturig sector. The atioal icome equatio for the developig coutry is give by, Y VA A VA T VA NM VA W, where VAA value added by agriculture sector, VAT value added by trasportatio ad commuicatio sector, VANM value added by o-outsourced maufacturig sector, VA value added by outsourced sector ad W icome of the immigrat workers. Cosumptio level ( C ) of the developig coutry ca be cosidered as the proxy for the coutry s welfare. 19 the theoretical model, cosumptio was equivalet to 17 VA ( ) it real value of product outsourced to i th coutry by a developed coutry i period t, it = umber of people temporarily migrated from the i th couty to the same developed coutry i t th year. 18 (W ) it the curret trasfers by migrat workers ad wages ad salaries eared by oresidet workers to i th coutry from abroad i period t. 19 the absece of savigs, cosumptio is equivalet to icome. Therefore o problem occurs. But i reality there are savigs ad dis-savigs, so there is a choice betwee cosumptio ad icome for the measuremet of welfare. the idustrialised coutries (like U.S.A), icome data is cosidered as a proxy measure of welfare. the other had, i less developed coutries (like Asia coutries), where data o wealth is ot available, cosumptio is a better choice tha icome for welfare measuremet. Apart from this, cosumptio also provides more stable measure of welfare tha icome. eato (1980), eato ad Zaidi (1999), Pradha (2001), ad others are i favour of cosumptio as a measure of welfare. Amog them,

18 126 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY atioal icome i absece of savig ad taxatio. Here a liear cosumptio fuctio is used for the purpose. C ( VA ) ( VA ) ( VA ) ( VA ) ( W ). 0 A A T T NM this equatio, cosumptio is made to deped o value-added i differet sectors. The estimatig equatio will thus be C NM ( VA ) ( VA ) ( VA ) ( VA ) ( W ) e, (30) 0 A A T T NM NM UT UT where e error term. Partial differetiatio of Equatio (30) with respect to VA gives, C UT VA. The sig of UT determies the partial effect of real value of outsourcig o the domestic cosumptio level, whereas total effect of real value of outsourcig o the domestic cosumptio level is measured by dc. dva Total differetiatio of Equatio (30) with respect to VA gives, dc dva The sig of d( W ) d UT. (31) d d( VA ) dc depeds o the relative sigs of UT dva dva. Give the sigs of d Equatio (30) Methodology ad ata dva ad d,, dw ad d dw, UT, will be estimated from the d eato (1980, 1999) strogly supported cosumptio measure sice cosumptio is the basis of other measures of welfare. fact, Livig Stadard Measuremet Study (LSMS) surveys are i favour of cosumptio measuremet.

19 MMGRATN VERSUS UTSURCNG 127 The Equatio (30) is estimated i a multi-coutry framework. U.S.A. is cosidered here as the foreig coutry ad seve low-icome coutries (Bagladesh, Ghaa, Haiti, Keya, Myamar, Nepal ad Seegal) 20 are cosidered as domestic coutries. Pael ata Aalysis is used to estimate Equatio (30) Pael ata Aalysis Here we have 7 less developed coutries (cross-sectioal uits) idexed by i 1,2,...,7 for the time period 1997 to 2009, 21 idexed by t 1997,1998,..., The liear relatioship, give by the Equatio (30), for i th coutry i t th period ca be writte as ( C ) it 0t A( VAA) it T ( VAT ) it NM ( VANM ) it ( VA ) ( W ) e. UT it it it There are two tests, used for appropriate model selectio. Lagrage multiplier test helps to distiguish betwee simple pooled regressio ad radom effect model, while Hausma s specificatio test helps to distiguish betwee fixed effect model 22 ad radom effect model. The descriptio of the data used i the estimatio of the ecoometric model is provided i what follows. Curret value of product outsourcig (VA ): ata o imports of goods (HS-9802) ito U.S.A is cosidered as the statistical iformatio o product outsourcig of the importig coutry (Sweso, 2005). 23 Uited States teratioal Trade Commissio 20 Here dia, Pakista ad Vietam are excluded from the coutry set. ata o the immigrats icome of Vietam is ot available for some years. Scatter diagram shows that the patters of data for dia ad Pakista are differet from the rest of seve coutries. these two coutries, temporary immigratio quota ad product outsourcig (i curret US$) are positively related. our theoretical model, supply side costrait determies the iverse relatioship betwee immigratio quota ad product outsourcig. With the rise i immigratio quota, labour availability to the domestic outsourcig sector decreases. This reduces the productio of domestic outsourcig sector. dia ad Pakista are two major seders of h1b visa. The supply side costrait may ot be valid for these two coutries, as they have surplus. 21 No H-1B visa was issued by U.S.A govermet i 2001, due to the terrorist attack o the Twi Tower. Therefore 2001 is ot icluded i the aalysis. 22 Here fixed effect model is estimated with time dummy istead of cross-sectio dummy. As our theoretical model is fixed edowmet model, ad immigratio policy of U.S.A. chages over time, fixed effect model with time dummy gives more meaigful result. 23 some studies, is cosidered as the proxy measure of service outsourcig (Khalifa ad Megova, 2010).

20 128 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY (USTC) reports the time series data o HS or our study, the cif import value uder HS-9802 is cosidered. The uit of measuremet is i curret US $. This is referred to as the offshore assembly programme (AP). Uder AP, exporters of the foreig coutry ejoy the tariff beefits, if their exported products cotai parts, compoets or materials made i the U.S.A. There is o tariff o the part of the value added produced withi the U.S.A. The dutiable AP import, as a result, is the value added produced abroad ad hece, is a part of product beig outsourced. Uder HS-9802 product group there are six tariff lies, which are stated as follows: : Photographic films ad dry plates maufactured i U.S.(except commercial motio-picture film) ad exposed abroad, whether developed or ot, : Articles retured to the U.S. after havig bee exported for repairs or alteratios, made pursuat to a warraty, : Articles retured to the U.S. after havig bee exported for repairs or alteratios, es, : U.S. articles of specific metals exported for further processig ad retured for further processig, : U.S. articles assembled abroad, which have ot lost their physical idetity or have ot advaced i value or improved i coditio abroad, : Textile ad apparel goods, assembled i Mexico i which all fabric compoets were wholly formed ad cut i the Uited States, etc. The data o these are electroically documeted for 154 coutries, icludig high icome, upper middle icome, lower middle icome ad low icome group, over the period 1989 to The data cosidered here are oly a part of total outsourcig activity. fact, this dataset does ot iclude the service outsourcig (software outsourcig). GP deflator: GP deflator (base year 2005=100) is used to covert the curret value of product outsourcig i US $ ito costat price estimates. These data are collected from the official web page of Uited Natios Statistics ivisio. Usig GP deflator, product outsourcig by i th coutry to a developed coutry i real terms is arrived at. Temporary immigratio quota ( it ) : H-1B issued by U.S.A Govermet i a year, categorised by coutry of birth, is cosidered as a measure of temporary immigratio quota. However, o H-1B visa was issued by U.S.A govermet i U.S. epartmet of Homelad Security provides with this data. The real cosumptio data of i th coutry i period t (C ) it : The cosumptio expediture data at costat price are collected from the official web page of Uited Natios Statistics ivisio.

21 MMGRATN VERSUS UTSURCNG 129 Value added by agriculture sector of i th coutry i period t ( VA A) it : The data o value added by agriculture sector at costat price are collected from the official web page of Uited Natios Statistics ivisio. Value added by trasportatio ad commuicatio sector of i th coutry i period t ( VA T ) it : The data o value added by trasportatio ad commuicatio sector at costat price are collected from the official web page of Uited Natios Statistics ivisio. Value added by o-outsourced maufacturig sector of i th coutry i period t ( VA NM ) it : The data o value added by o-outsourced maufacturig sector at costat price are obtaied by subtractig the value of product outsourcig from the value added by maufacturig sector. come of the immigrat workers i t th period ( W ) it : come of immigrat workers icludes workers' remittaces, compesatio of employees, ad migrats trasfers. The data o these variables available i curret U.S. dollars are deflated by GP deflator to arrive at real icome of the immigrat workers Stage Estimatio 4. ESTMATN RESULTS The empirical estimatio is carried out to fid the impact of real product outsourcig o welfare of developig coutries from where itermediate products are beig outsourced by developed coutry U.S.A. doig so, it is ecessary to arrive at the estimates of the relatios betwee real product outsourcig ad temporary immigratio quota. order to estimate the relatio betwee real product outsourcig ad temporary immigratio quota, a simple pooled regressio is carried out o the data of 7 low-icome coutries over the time period 1997 to The bivariate equatio estimates real product outsourcig ( VA ) as the depedet variable ad temporary immigratio () as the explaatory variable. The results are show i Table 1. The results show that the iverse fuctio is most appropriate for describig the relatioship betwee real product outsourcig ad the temporary immigratio quota, sice R 2 ( 0.164) is highest for iverse form. fact statistic also idicates that oly logarithmic fuctio ad iverse fuctio are appropriate to describe the relatioship betwee real product outsourcig ad the temporary immigratio quota. both the fuctios, real product outsourcig is iversely related with immigratio quota. We also estimate liear fuctio, quadratic fuctio, cubic fuctio, ad quartic (4 th order)

22 130 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY fuctio. 24 However, oe of these is appropriate. We choose iverse fuctio betwee 2 logarithmic ad iverse fuctios, sice R is highest for this fuctioal form. The estimated iverse fuctioal form ca be writte as, VA ( ) 1. d( VA ) d ( ) 1 2, where Therefore, d 2 ( ) ( ) VA d at mea value of. 265 ; mea =265. Table 1. uctioal orm Relatioship betwee Real Value of Product utsourcig ad Temporary mmigratio Policy across Low-icome Coutries (epedet variable= VA ; depedet Variable=) Liear Quadratic Cubic Quartic (4 th order) Logarithmic verse R Sigificace Level β β (0.133) (0.003) (0.463) β (0.730) β 3 β (0.011) (0.451) (0.591) -1.E-06 (0.630) (0.009) (0.22) 1.305E-02 (0.275) E-05 (0.303) 1.105E-08 (0.33) (0.003) (0.028) (0.453) Notes: igures i bracket idicate the sigificace level. idicates sigificat at 10 percet level, idicates sigificat at 5 percet level ad idicates sigificat at 1 percet level. The sig of { d ( VA ) d 265} shows that there is a iverse relatio betwee real 24 Quitic fuctio (5 th order) ca ot be estimated, sice oe variable is dropped due to multicolliearity. Here, compoud fuctio, power fuctio, S fuctio, growth fuctio, expoetial fuctio ad logistic fuctio caot be estimated due to the presece of o-positive values of depedet variables.

23 MMGRATN VERSUS UTSURCNG 131 product outsourcig ad the umber of temporary migrats i developig coutries at the mea value of regressor (). The results show that if a developed coutry (say, the U.S.A) icreases temporary immigratio quota for low icome coutries, the the product outsourcig from that low icome coutry decreases. Shortage of specialised labour (ivolved i outsourcig sector ad i migratio process) of these low-icome coutries may cause this Stage Estimatio order to estimate the relatioship betwee total wage icome eared by immigrats ad temporary immigratio quota, a simple pooled regressio is ru o the data of seve low-icome coutries over the time period 1997 to The result for the appropriate model selectio is give i Table 2. Selectio of the appropriate fuctioal form: Whe the umber of the regressor icreases, the the justificatio of itroductio of the ew variables ca be checked through test. Null Hypothesis; H 0 : Coefficiets of ew regressors = 0. Alterative Hypothesis; H A : Coefficiets of ew regressors 0. 2 NEW 2 L 2 L Test statistic is, ( q, k) ( R R / q)/(1 R / k), where, q=umber of ew regressors, =umber of observatios, k=umber of parameter i the ew model. Uder ull hypothesis, test statistic follows distributio with degrees of freedom (q, - k). The results show that the quartic fuctio (4 th order) 26 is a sigificat improvemet over liear, quadratic ad cubic fuctios. Table 2 shows that quartic fuctio is most appropriate for describig the relatioship betwee total wage icome eared by immigrats ad temporary immigratio quota. 25 There are two problems regardig empirical aalysis. There is uavailability of data o the product outsourcig ad o the temporary immigratio quota. Number of H-1B visa, issued by U.S.A govermet, is cosidered as the measure of the temporary immigratio quota. But the sector wise classificatio of H-1B visa is ot available for all years. This makes it difficult to get the sector wise sceario. So i the empirical part, aggregate level relatioship betwee the temporary immigratio policy ad the real value of product outsourcig is aalysed. Secod problem is that there is a o-liear relatioship betwee these two variables. Therefore, the iverse relatioship (measured by the slope of the iverse fuctio) betwee two variables is estimated at the mea value of the regressor. 26 Quitic fuctio (5 th order) ca ot be estimated, sice oe variable is dropped due to multicolliearity.

24 132 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY Table 2. Appropriate Model Selectio for Seve Low-come Coutries Test Hypothesis Test Statistic Value Commet Liear vs. Quadratic Liear vs. Cubic H 0 : β 2 =0 (1, 80) 8.66 H 0 is rejected. So quadratic fuctio is better tha liear fuctio. H 0 : β 2 =β 3 =0 (2, 79) 4.47 H 0 is rejected. So cubic fuctio is better tha liear fuctio. Liear vs. Quartic Quadratic vs. Cubic Quadratic Vs. Quartic H 0 : β 2 =β 3 =β 4 =0 (3, 78) 4.91 H 0 is rejected. So quartic fuctio is better tha liear fuctio H 0 : β 3 =0 (1, 79) 0.43 H 0 is accepted. So quadratic fuctio is better tha cubic fuctio H 0 : β 3 =β 4 =0 (2, 78) 3.53 H 0 is rejected. So quartic fuctio is better tha quadratic fuctio Cubic vs. Quartic H 0 : β 4 =0 (1, 78) 6.72 H 0 is rejected. So quartic fuctio is better tha cubic fuctio Notes: H 0 is rejected at 10% sigificace level; H 0 is rejected at 5% sigificace level; H 0 is rejected at 1% sigificace level The results, give i Table 3, show that R ( 0.326) is highest for quartic fuctio with a -statistic of 9.42 beig sigificat at 1 percet level. The estimated relatio betwee the two variables at the mea value of the regressor () for the quartic equatio is give by, ( ) E, W 0 where E= radom error term ( ) 33( ) 44( ), d( W )/ d d( W )/ d (265) 33(265) 44(265) , where , , ad

25 MMGRATN VERSUS UTSURCNG 133 Table 3. Relatioship betwee Real mmigrats come from Abroad ad Temporary mmigratio Policy for 7 Low-come Coutries (epedet variable=w ; depedet Variable=) uctioal orm R 2.. β 0 β 1 β 2 β 3 β 4 Liear Quadratic Cubic Quartic (4 th order) Logarithmic verse (0.018) Compoud Power S-curve (0.056) Growth Expoetial Logistic (0.084) (0.461) (0.878) (0.142) (0.009) (0.101) (0.212) (0.131) (0.018) (0.056) (0.003) (0.985) (0.012) (0.546) (0.007) (0.009) Notes: igures i bracket idicate the sigificace level. idicates sigificat at 10 percet level, idicates sigificat at 5 percet level ad idicates sigificat at 1 percet level. The above results also show that the icome of immigrat workers from abroad ad immigratio quota are directly related i quartic fuctioal forms Stage Estimatio The impact of real product outsourcig o welfare of low-icome coutries is estimated usig Equatio (30). Here real cosumptio is used as a measure of developig coutry welfare. The five explaatory variables, as i Equatio (30), are value added by agriculture, value added by trasport ad commuicatio, value added

26 134 SMNTN AS, AJTAVA RAYCHAUHUR AN SAKAT S. RY by o-outsourced maufacturig sector, value added by outsourcig sector ad the icome of the immigrats from the abroad. Pael data method is used for the purpose. or the purpose of this estimatio, the results of the earlier two estimates are take ito accout. The results of the empirical exercise are preseted i Table 4. The results show that all the explaatory variables, except the icome of the immigrats from the abroad, have positive effect o real cosumptio level i all three models. However, the coefficiet of value added by outsourcig sector is statistically sigificat (at 1% level) oly i radom effect model. the other had, icome of the immigrats from abroad has egative impact o real cosumptio i all three models. This idicates that immigrats may sped their icome o foreig cosumptio istead of domestic cosumptio. The coefficiets of icome of the immigrats are statistically sigificat (at 1% level) i pooled regressio model ad fixed effect model. However, as the estimated coefficiet correspodig to value added by outsourcig sector is greater tha uity so the icrease i domestic cosumptio will be greater tha the icrease i value added by outsourcig sector. Comparig three models, it is oted that the total effect of value added by outsourcig o real cosumptio ( dc / dva ) is positive i all three models. rom Equatio (31), we have ( dc / dva ) UT ( d( W ) / d)( d / d( VA )). Stage ad Stage estimatios give that d( VA )/ d 0 ad d( W )/ d 0 respectively. Pael data estimatio gives that UT 0 ad 0 i all three models. Pooled Regressio Model, ( dc / dva ) ixed Effect Model, ( dc / dva ) Radom Effect Model, ( dc / dva ) t is oted that total effect of value added by outsourcig o real cosumptio is positive irrespective of model specificatio. fact, the partial effect of value added by outsourcig o real cosumptio ( C / VA ) is also positive i all three models; however it is statistically sigificat oly i radom effect model. These results show that give the egative impact of temporary immigratio quota o real product outsourcig, real cosumptio i developig coutries icreases with rise i real product outsourcig. The partial as well as total effects of real product outsourcig o welfare of developig coutries are thus positive. Lagrage Multiplier test ad Hausma test show that fixed effect model with time specific dummy is suitable for this aalysis. Choice of fixed effect model implies that immigratio policy of developed coutry (here USA) over the years, year wise labour edowmet of developig coutries ad other time specific factors ifluece the real cosumptio of developig coutries, apart from other explicit factors, like value added by agriculture, value added by trasport ad commuicatio, value added by