Offshore Production, Student Effort, and Income Inequality

Transcription

1 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, Offshore Production, Student Effort, nd Income Inequlity Chu Ping Lo Ntionl iwn University, ipei Abstrct his pper presents simple model to show tht, in ddition to technologicl development nd globl outsourcing, heterogeneous lifetime preferences for leisure re essentil to income inequlity. Allocting leisure time to studying is viewed s suffering in exchnge for future comfort, nd income inequlity in country rises when some prefer the leisure trde-off but others prefer smooth leisure over the lifetime. I lso rgue tht, in country where residents feel less suffering in the leisure trde-off, greter income inequlity is incurred. his model implies tht low ltitude countries generlly suffer from greter income inequlity thn high ltitude countries. Keywords: Globl outsourcing, offshore production, student effort JEL Clssifiction: F4, F6, I20, J3. Introduction Feenstr nd Hnson (997) hve rgued tht the North pplies more offshore production to shift the lbor-intensive prts of processing to the low-cost South, fter the South dopts n open policy. Offshore production reshuffles the fctor composition nd leds to n increse in the reltive demnd for skilled lbor in the North, worsening its income inequlity between skilled nd unskilled lborers. Feenstr nd Hnson (996) further estimted tht expenditures on high-technology cpitl ( proxy of technologicl improvement) led to n increse in the reltive income of skilled lbor, explining bout 60% of the ctul chnge of the rising income inequlity from 979 to 990 in the U.S., while offshore production cn ccount for 3%-5% of the increse. heir findings support the rgument tht technology improvement nd offshore production re underlying contributors to income inequlity between skilled nd unskilled lbor in country.

2 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, Nevertheless, I rgue tht, in ddition to technologicl improvement nd offshore production, heterogeneous preference for leisure is essentil to income inequlity between skilled nd unskilled workers within ntion. In corresponding to technologicl improvements, students re required to chieve greter eductionl ttinments to process incresingly sophisticted skills, dvnced knowledge, nd deep know-how in order to become productive skilled workers lter. he exchnge of leisure for study is scrifice mde in hopes of lter gin. Students ccumulte skill premium by llotting prt of their current leisure time to studying in exchnge for future comfort (i.e., more leisure time), which is dditionl to endowed leisure time nd is mesured in equivlent service goods. ht is, they feel more comfort with more leisure time lter (e.g., more vction or more entertinments) during their incresed spre time sved in service cretion fter completing their eduction, while unskilled workers must provide such services themselves. However, the trde-off between current nd future leisure is bsed on individuls leisure preferences long the lifetime spectrum. Heterogeneous leisure sensitivity gives rise to lifetime income inequlity. Findly nd Kierzkowski (983) ssumed tht ech individul in n economy hs equl potentil to become skilled worker if he or she is determined to receive n eduction. hey pplied skill-formtion function to trnsform students into skilled workers fter fixed length of school time, with the only required input being country-endowed eductionl resources. In long-run competitive equilibrium, their model implies identity of rel income over the life cycle between skilled nd unskilled workers. However, in the rel world, it hs been widely observed tht significnt lifetime skill premium exists. 2 An dditionl importnt fctor, which is overlooked in their rgument, is student cdemic effort. 3 he Americn inventor homs Edison once sid, "Genius is one percent inspirtion nd ninety-nine percent perspirtion." 4 Empiriclly, Loury nd Grmn (995) documented tht, in ddition to school yers, student performnce nd college selectivity significntly ffect ernings inequlity. Mrtins nd Pereir (2004) found tht schooling hs positive impct on within-group income inequlity, with tht impct conditioned on students bility nd schools qulity. Generlly, both student performnce nd college selectivity re highly determined by student bility nd effort, where the ltter generlly plys lrger role. 5 Given tht skill returns depend not so much on ingenuity s on hrd work, I tke only the ltter into ccount in my nlysis. 6 ht is, the return on eduction depends not only on eductionl intensity, s mentioned in Findly nd Kierzkowski (983), but lso on student study effort, s

3 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, implied in both Loury nd Grmn (995) nd Mrtins nd Pereir (2004). Students must devote themselves to studying in order to effectively bsorb the eduction leding to skill formtion. his pper serves to develop simple model, where student study effort is dded to Findly nd Kierzkowski s (983) skill formtion function, to show how heterogeneous preferences for leisure long the lifetime spectrum re essentil to lifetime income inequlity between skilled nd unskilled workers. A service sector is introduced into this model. Skilled workers tke dvntge of their rising skill premium to purchse services insted of relying on self-production, creting incresed demnd for vrious personl services. he service sector bsorbs unskilled mnufcturing workers mde redundnt from offshore production, which in turn provides incresed support for the expnsion of personl services to meet the rising demnd from skilled workers in the mnufctured sector. 7 However, the scrifice of current leisure to studying is viewed s study suffering, requiring skill premium for returns to exchnge for future comfort tht mesured in equivlent service good. I rgue tht income inequlity in country rises when some individuls prefer to the leisure trdeoff but the others prefer to smooth leisure over lifetime. his model lso implies tht low ltitude country generlly suffers from greter income inequlity thn high ltitude country. Endowed with longer sunshine hours, inhbitnts in the low ltitude countries find less suffering to llocte time to studying while inhbitnts in high ltitude countries re reltively inctive (e.g., sleeping is kind of leisure) for keeping wrm in their wintery climte. Climte ffects leisure preferences becuse frigid wether generlly dds suffering of studying, deterring trde of current leisure for future comfort. By contrst, wrm climte generlly leves more brething spce to the leisure trdeoff. I then rgue tht individuls in low ltitude country, ll other things equl, who re reltively inclined to llocte more leisure to studying, form higher skill premium thn tht in high ltitude country. his gives rise to greter income inequlity in the low ltitude country. he rgument finds support from the dt in Figure, which shows tht, where Jpn is n outlier, the Gini coefficients of OECD countries re negtively relted to their cpitl cities ltitudes. 8 ht is, low ltitude country generlly suffers greter income inequlity thn high ltitude country. In Section 2, I present the ide tht those who prefer to scrifice current leisure for future comfort enroll s students, while the remining become unskilled workers herefter. he trde-off between current leisure nd future comfort is deterred while the required length of lerning for ccumulting sufficient skills increses due to technologicl improvements, dding suffering of studying. However, n increse in eductionl resources cn reduce the threshold

4 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, of student cdemic efforts, helping students upgrde their productivity with less study suffering. Further, I show how the skill premium s purchsing power for service goods ffects the optiml study effort, determining the extent of income inequlity. In Section 3, I conclude with discussion of these findings. 2. he Model he seminl model of Findly nd Kierzkowski (983), who defined skill formtion s function of eductionl resources, provides the bsic structure for my nlysis. Bsed on the empiricl work from both Loury nd Grmn (995) nd Mrtins nd Pereir (2004), I prtition the returns to eduction into two prts: returns to eductionl resources nd returns to student study effort. 2. Augmented Skill Formtion Function Suppose tht in world of North nd South, the North consists of unskilled workers, skilled workers, nd students. Let the unit fctor returns to n unskilled worker be, nd the returns to skilled worker be, where w. Ech individul consumes mnufctured goods, w w2 services, nd leisure. he mnufctured goods re interntionlly trdble, while service is trdble only domesticlly, nd leisure cnnot be trded t ll. Individuls hve similr preferences for both goods nd services, but preferences for leisure differ between unskilled nd skilled workers. All individuls, lboring to ern their goods nd services, re ssumed to be similrly endowed with unit of leisure time. Students must llot some prt of their leisure time to studying while other workers relx during their leisure time. People feel more comfort with more leisure time. w 2 he popultion is sttionry becuse the birth of N new individuls is exctly offset by the deth of N old individuls t ech instnt of time. Every individul hs lifetime, choosing whether she/he goes to school for fixed length of time θ, to cquire skills right fter his/her birth. Suppose tht E individuls, mong the N new-born individuls, tend to cquire eduction, while the remining L= N E individuls remin unskilled for the durtion of their lives. hose who choose to cquire eduction become skilled workers fter the θ length of lerning t school, erning n income rte with skill premium for the rest of their lifetime. 9 here is fixed exogenous supply of eduction inputs, denoted s K, which cn be thought of s

5 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, composite of techers, lbortories, librries, nd so on (Findly nd Kierzkowski, 983), which mke up the eductionl resources. In my model, the output of skills of ech student, K mesured in efficiency, is not only positively relted to eductionl resource intensity, k =, E but is lso positively relted to student study effort,. Here denotes the shre of student s leisure time tht hs been llotted to studying: the more time student llots to studying, the more skills the student cn cquire. Supposed tht the birth of N new individuls hs been rndomly ssigned leisure preference so tht individuls study effort hs probbility density function g() nd corresponding cumultive distribution function is given by G(). A student s skill formtion function is then given by q = f (, k; θ ), () where f ( k) > 0, nd f ( k) < 0; tht is, with given θ length of lerning t school, skill formtion increses with eductionl resource intensity in diminishing returns. A student cn prolong the length of lerning to upgrde his skill level, tht is, f ( θ ) > 0. It is widely believed tht n individul hs the potentil to become skilled lborer if she intends to becuse humns hve n unlimited cpcity to lern (Eccles, 974). 0 Clinicl findings support this theory nd lso show tht humn intellectul development remins constnt up to deth (Lewin, 980). herefore, it is fesible to presume tht returns for student efforts in skill formtion re not diminished with study effort within finite lifetime, tht is f ( ) > 0 nd f ( ) 0. 2 Given the fct tht the powers of humn brin re endless (Eccles, 974), I rgue tht it is heterogeneous preferences for leisure rther thn heterogeneous students cpbilities tht re essentil to skill formtion. With exogenous eductionl resources K nd t given θ length of lerning t school, student cnnot cquire sufficient skills without minimum ggregte study efforts. ht is, with given length of lerning θ nd with exogenous eductionl resources, there exists threshold of study effort, sy, which I will derive in Section 2.3. he N new born individuls re then detched into two groups: individuls with enroll s students nd the others remin unskilled workers. Being reltively scrce in eductionl resources, individuls in the South must llocte more study efforts in comprison to those in the North to obtin equivlent skills. As in Feenstr nd Hnson s (997) outsourcing model, there is one single mnufctured good

6 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, which is ssembled with continuum of intermedite inputs, indexed by z [0.]. he production fctor of ech of the intermedite inputs is composition of unskilled nd skilled workers. he lbor-intensive intermedite inputs require more unskilled workers, nd the skill-intensive intermedite inputs require more skilled workers. he intermedite input z employs S ( units of skilled lbor nd U ( units of unskilled lbor, with productivity q(, k) nd one, respectively. Supposed tht the composition of production fctors increses with, indicting tht the input is more lbor-intensive thn input if z z z2 he production function of ech intermedite input is given by ( min{ U (, qs( } S( U ( z <. z 2 x =. he only mnufctured finl-good hs Cobb-Dougls production function s lny = 0 finl good. α(ln x(dz with α( dz =, incorporting ll of these intermedite inputs into one 0 he unskilled workers in the North cn be substituted for through outsourcing trde, while reltively lbor-intensive inputs re relocted to the South. his leds to n oversupply of Northern unskilled lbor becuse this reloction leds to reltively less demnd for unskilled lbor in the North, nd then reshuffles the fctor composition in fvor of the skilled lbor in terms of both the wge rte nd employment. In my model, I rgue tht, with no skill brriers, unskilled workers cn move freely between service nd mnufctured-goods sectors. A full employment condition holds, so this oversupply of unskilled workers drops out of the mnufctured goods sector, shifting frictionlessly to the service sector. 2.2 Student Effort nd Leisure Alloction All individuls re ssumed to hve the sme preferences nd tstes for both goods nd services, but skilled workers cn purchse services from mrket nd enjoy more leisure time (i.e., more comfort) in lter life from their incresed spre time sved in service cretion. Unskilled workers, in contrst, must perform such services for themselves. For exmple, persons with high income cn fford to buy vrious services, such s eting out t resturnts, hving lundry done, child cre, sp, nd housekeeping, nd so on. Comprtively, persons with low income must generlly perform these necessry tsks on their own. I rgue tht, without hving ernings during the period of study, students lso perform ll services for themselves. Students use student lons from the cpitl mrket in order to py for their eduction nd bsic living (i.e., the mnufctured goods). he cpitl mrket chrges n

7 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, interest rte r nd expects full returns with good confidence fter the students become skilled lborer. If students like to use consumer lons to purchse services, the risk relted to students bility to fulfill future pyments increses becuse the students might llocte the incresed spre time sved in service cretion to leisure rther thn to study. Considering the uncertinty tht students might not ccumulte sufficient skills to fulfill the lon pyment, the cpitl mrket will chrge risk premium in ddition to the underlying interest rte r for the consumer lon. Since the risk premium is substntilly high, I presume tht ll students use only the student lon in this model. 3 While both skilled nd unskilled workers enjoy their leisure time, students must study hrd fter clss during the eduction period in order to productively bsorb the specilized knowledge. A student s scrifice in studying is rewrded with f (, k) skill premium units. One cn rgue tht the price of leisure hour is for skilled workers, nd w for both unskilled workers w 2 nd students (Owen, 97). he opportunity cost of llotting leisure time to studying is then equivlent to the vlue of the corresponding incresed skill premium from studying: = θ C( ) w f e dt, (2) which lso indictes the suffering of study tht cn be rewritten s C ) = [ f r ( w ( e rθ e r )]. As unskilled workers, students must lso crete services for themselves, while skilled workers use the ssocited rise in the skill premium derived from their previous study efforts, to purchse personl services in the mrket. hen the benefit of studying, nmely, future comfort, is given by s(, L B( ) = E( θ ) 0 w e 2 dt, which is the vlue of services skilled worker cn purchse by his/her skill premium. 4 Here, s(, denotes the shre of unskilled lbor work in the service sector, incresing with interntionl outsourcing trde, tht is, (, > 0. Obviously, greter skill premium is creted when more study effort is devoted, which then leds to lrger demnd for services. It is fesible to ssume tht the s(, increses with the study effort but with diminishing returns, tht is, s (, > 0 nd s (, < 0. s z However, with given length of lerning θ nd exogenous eductionl resources K, there exists threshold of study effort,, required for student to become n effective skilled worker.

8 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, Without reching the ggregte threshold of study effort for certin period of lerning, student cnnot successfully become skilled worker. Generlly, with given eductionl resources K, student must llot not less thn shre of leisure to studying for period of time θ in order to productively bsorb the ssocited skills. In the rel world, few, if ny, students cn pss exms nd grdute with degree without minimum level of study, nd the forml degree usully functions s signl to employers tht whether the cndidte is qulified skilled worker. herefore, with exogenous K nd θ, I ssume s (,0) = 0 if 0 <. In equilibrium, the net rewrd of future comfort from studying equls the skill premium creted from studying. Mrket clers in the service sector: θ w f e s(, L dt = E( θ ) 0 w e 2 dt, when, (3) As mentioned bove, the births of N new individuls re rndomly ssigned student cdemic effort, which hs probbility density function g(). Let s presume further tht people djust their leisure preferences right fter birth with the cost nd benefit nlysis in ccordnce with the leisure trde-off long the lifetime spectrum, following & = B( ) C( ). his eqution of motion for leds to two stedy sttes s in Figure 2: = 0 nd = respectively. Here denotes the ctul study effort student mkes. Here, we exclude the condition of > becuse it will led to C ( ) > B( ) for ll, implying tht there is no student who devotes to ny level of studying (with stedy stte = 0 only), contrdicting observtions in the rel world. I lso rule out the condition where B() > C() for ll, becuse this implies tht ll students will llot ll of their leisure time to studying (with stedy stte = only), which lso contrdicts ctul observtions. herefore, the study effort threshold in ccumulting skills hs reltion to ctul study effort s 0 < <. Individuls with < < find it beneficil to increse their study effort to nd individuls with lso find study effort of be the optiml choice; however, individuls with < reduce their study effort to 0. In the former stedy stte, ll students llocte shre of leisure time to studying. In the ltter stedy stte, individuls prefer to consume leisure smoot hly throughout the lifetime. he number of students is then given by E = N g( e) de = N( G( )), which grows with reduce in the study effort threshold. In

9 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, the next section, I will derive the optiml study effort threshold in order to obtin the equilibrium school enrollment E ( ). 2.3 Equilibrium he productivity of ech skilled worker is ugmented by fctor of f (, k), so tht the income of skilled worker is given b y w f (, ). We obtin the present vlue of the gross lifetime income of skilled worker s k θ w f (, k) e w dt = r f (, k)( e rθ e r ). (4) he ggregte cost of eduction in present vlue from (2) nd (3) is given by θ θ ke dt + w2e dt + w2e θ 0 w kf = [ w ( kf r k + f )( e rθ e r ) + w 2 dt + C( ) ( e r )]. (5) he first term in (5) is the tuition pyment from 0 to θ for ech student, which equls the sum of the skill formtion function s discounted mrginl products over the working life of the student (Findly nd Kierzkowski, 983). he second nd third terms re the opportunity costs of enrolling s student. he fourth term is ggregte study suffering for the θ length of time. Subtrcting (5) from (4), we obtin the net benefit from eduction s rθ r e π = [ w ( f kf k f )( e e ) w2 ( e )]. (6) r Individuls like to cquire eduction until the mrginl benefit of eduction equls the mrginl cost of eduction. he optiml solution is derived from the first order condition of (6) with respect to eductionl resource intensity (i.e., k ): π [ ( rθ r = w kf kk + f k )( e e )] = k r 0. (7)

10 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, From (7), we obtin the equilibrium k fkk (, k ) = f k (, k ). Rewrite it s k fk (, k ) =, (8) f (, k ) kk where effort K th e optiml eductionl intensity k = reduces with n increse in the ctul study E. 5 his is becu kill formtion function is non-diminish returns to study se tht the s effort. he term k f (, k ) in (8) represents the mrginl benefit of educt ion resource kk intensity net of mrginl costs of eductionl resource intensity. he term f (, k ) represents the mrginl cost of studying suffering with respect to eductionl resource intensity. Corportely, (8) implies substitution between eductionl resource intensity nd study effort in skill formtion. k With the exogenous eductionl resources K, we cn find optiml student enrollment s E ( ) = N( G( )), where school enrollment is encourged with lower threshold of study effort (i.e., ) so tht E ( ) < 0. Howeve r, these enrolled students will llocte shre of leisure time to studying bsed on the leisure trde-off between current leisure nd future comfort s discussed in the previous section, where. In the long run, the net benefit from eduction must be zero. We obtin long-term equilibrium from (3) nd (6) s ω = s(, ( N E ( )) +, (9) E ( )( θ ) where ω denotes the lifetime income inequlity between skilled nd unskilled workers s w ( f ω = kfk )( e w ( e 2 rθ r e ) r ). Here the nomintor term is the lifetime income of skilled worker net of the tuition pyment nd the denomintor term is the lifetime income of n unskilled worker. he equilibrium (9) illustrtes how heterogeneous preferences on leisure ffect the lifetime income inequlity between skilled nd unskilled lbor, which not only increses with n

11 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, increse in the ctul study effort study effort. 6 but lso increses with n increse in the threshold of he model of Findly nd Kierzkowski (983) implies identity of lifetime income becuse they ssume tht ech individul hs equl potentil to become skilled. However, I rgue tht the lifetime income inequlity rises while there re heterogeneous preferences for leisure long lifetime spectrum. he study effort threshold is bsiclly dependent on country s eductionl resources while the ctul study effort is minly determined by country s lifetime leisure preferences. Individuls with leisure preference enroll s students but the others remin unskilled throughout the lifetime. hese enrolled students devote shre of leisure time for θ periods to studying in exchnge for more future leisure. Note tht we rgue bove tht s (, = 0, tht is, there is no lifetime income inequlity while =. It is interesting to note th t if z = 0 nd =, we return to Findly nd Kierzkowski s model, where the skill premium is just sufficient to py bck the student lons. 2.4 echnologicl Development, Offshore Production, nd Eductionl Resources An exogenous improvement in technology requires upgrdes in lbor skills to pressure students to prolong the length of studying (i.e., lrger θ ) in order to ccrue the ssocited dvnced skills. his increses opportunity costs while prolonging the length of studying, requiring higher skill premium for compenstion. 7 As shown in Figure 3, the increse in the length of studying θ gives rise to ctul s tudy effort, implying widening income inequlity s implied in equilibrium (9). An incresed in leds to smller, which ttrcts more school enrollment (i.e., lrger E ). I rgue tht the increse in school enrollment is to prtilly but not fully compenste for the rising income inequlity in the long run. his is becuse tht n incresing in ctul study effort gives rise to not only skill premium but lso study suffering, the former encourges school enrollment but the ltter discourges school enrollment. On net, technologicl development leds to greter income inequlity, lthough giving rise to supply of skilled lborers. Note tht in this model, the lifetime income exists provided tht >. 8 Next, we come to the discussion on offshore production, which hs never been so frequently covered in the medi with growing concerns on rising inequlity between skilled nd unskilled

12 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, lborers. Offshore production expnds the service sector while the North reloctes its reltively lbor-intensive inputs to the South, mking some unskilled lborers oversupply. he remining intermedite inputs reshuffle their production fctors compositions so tht ech intermedite input becomes more skilled-lbor intensive herefter. Offshore production then reduces the reltive demnd for unskilled lbor, resulting in higher skill premium. he rising skill premium gives rise to more demnd for services. Note tht the rising skill premium in turn ttrcts more student enrollment s implied in (8): n increse in z leds to n increse in nd n reduction in the threshold of study effort. he illustrtion is shown s in Figure 4. While the oversupply of unskilled workers in the mnufcturing-good sector freely shifts to the service sector, mrkets cler in the service sector. It requires tht the rising demnd for services should equl the rising supply of services mde by the oversupply of unskilled workers due to offshore production. Similr to technologicl development, n increse in offshore production genertes disproportionte inequlity, fvoring skilled workers in terms of not only income but lso employment. he rgument is in line with tht of Feenstr nd Hnson (997). However, lthough offshore production excerbtes income inequlity, governmentl efforts should focus on n eductionl infrstructure tht cn trin sufficient qulified lbor force rther thn resorting to trde protectionism. In equilibrium (8), with the optiml k nd, n increse in eductionl resources encourges more student enrollment (i.e., lrger E ) s illustrted in Figure 5. his incresing student enrollment leds to n increse in the reltive supply of skilled workers, exerting downwrd pressure on the skill premium. his implies tht country with higher levels of eductionl resources suffers less wge inequlity thn countries with fewer eductionl resources s implied in (9). he policy impliction for n economy cught in the rising tide of offshore production is cler. As Aln Greenspn (2000) hs urged, government should plce incresed emphsis on eduction nd offer more retrining progrms for creer chnges, rther thn to fight outsourcing through protectionism. In the er of globliztion, worsening income inequlity is not necessrily inescpble if sufficient eductionl investment exists. Miller (200) lso rgued tht, in the cse of the U.S., hlting globliztion would do little to offset the rising wge inequlity tht hs lredy occurred, unless U.S. firms could reverse the outsourcing ctivities bck home, however, which would do much more dmge thn trde restriction lone.

13 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, Current Leisure versus Future Comfort In my model, skilled workers use θ w( f kf k ) e dt to py for trdble mnufctured goods nd use s(, L E( θ ) θ w f 2 0 w e rt e dt to exchnge for untrdeble service goods s mesured by dt, nd the lter constructs the lifetime income inequlity. he mrket clering for the untrdeble service goods requires the equilibrium (3) s s(, L wf e rt dt = w2e θ E( θ ) 0 ds dt, where country-specific prmeter ε = s > 0 d exists to equte the demnd nd supply sites of the service mrket. he prmeter ε cn be considered s n elsticity of purchsing power for future comfort with respect to current study effort. he less inclined the individuls in n economy re to mke the leisure trdeoff for future comfort, the lower is the reltive demnd for the supply of service goods, nd the lrger the purchsing power elsticity is. Allocting leisure time to studying is viewed s suffering by ll individuls, where the lrger is this purchsing power for service goods, the less current leisure students need to scrifice in order to exchnge for sufficient future comfort, tht is, ˆ < if ˆ ε > ε, s illustrted in Figure 6. 9 he decrese in ctul study effort, lthough leding to higher threshold of study effort s ˆ >, reduces income inequlity, s implied in (9). 20 On the contrry, the more inclined the individuls in n economy re to mke the leisure trdeoff for future comfort, the lrger is the reltive demnd for the supply of service goods, nd the lower purchsing power elsticity is. In country where the skill premium s purchsing power for service goods is lower, more study efforts re required to ccumulte sufficient skill premium, leding to higher income inequlity. his model implies tht climte ffects the preferences for leisure becuse frigid climte, compring to wrm climte, generlly mkes llocting leisure time to studying more pinful, deterring the leisure trdeoff. Low ltitude countries enjoy longer hours of sunshine thn the high ltitude countries, 2 nd humn re more ctive in wrm climtes compred to in frigid climtes. Endowed with longer hours of sunshine, inhbitnts in low ltitude countries find less suffering in llocting more time to studying while inhbitnts in high ltitude countries re reltively inctive to keep wrm in their wintery climte. I therefore rgue tht wrm wether

14 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, leves low ltitude country, ll other things equl, with more brething spce to llocte to studying in comprison to high ltitude country. As illustrted in Figure 6, this leds to lrger study effort nd lower study threshold in the low ltitude country thn high ltitude country. As implied in (9), low ltitude countries should hve higher income inequlity thn high ltitude countries. he dt in Figure supports this rgument. 3. Conclusions In line with Feenstr nd Hnson (997), both n increse in technologicl development nd n increse in globl outsourcing led to worsening inequlity in terms of income nd employment between skilled nd unskilled workers within ntion. However, my model lso shows tht n increse in eductionl resources cn reduce threshold of cdemic efforts, encourging school enrollment. In turn, this increse leds to greter supply of skilled workers, exerting downwrd pressure on income inequlity, which gives rise to the policy impliction tht government should llocte more resources to its eductionl infrstructure rther thn resorting to trde protection. In my model, there re heterogeneous preferences for leisure long the lifetime spectrum of people, where some individuls enroll s students nd llocte prt of their leisure time to studying to exchnge for future comfort. However, those who prefer to consume leisure smoothly over the lifetime become unskilled workers. After eduction, the students become skilled workers with skill premium nd use the skill premium to purchse service goods from mrkets but unskilled workers must crete service goods themselves. he more leisure time is llocted to studying, the lrger the skill premium skilled lborers cquire, nd the greter the income inequlity in country is. his is the min conclusion in this pper: in ddition to technologicl development nd globl outsourcing, heterogeneous lifetime preferences for leisure led to income inequlity between skilled nd unskilled workers within ntion. I lso rgue tht the more residents in country like to mke the leisure trdeoff for future comfort, the smller is the elsticity of purchsing power for service goods, leding to greter lifetime income inequlity in this country. his model implies tht low ltitude countries generlly suffer from greter income inequlity thn high ltitude countries. Endowed with longer hours of sunshine, residents of low ltitude country, ll other things equl, generlly feel less suffering in llocting leisure time to studying in comprison to those in high ltitude country, incurring lrger income inequlity. his is the second conclusion in this model.

15 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, Endnotes Contct informtion correspondence to Chu Ping Lo. E-Mil: Phone: (886) ; Fx: (886) Deprtment of Agriculturl Economics, Ntionl iwn University, ipei, iwn. In Section 2., I presume ech individul is endowed with unit of leisure time t ny instnt of time. 2. Investigting the welfre consequences of income inequlity in the U.S., over the lst 30 yers, Krueger nd Perri (2003) found tht skilled workers enjoy sizble welfre gins t similr mgnitude to the losses of the poorest gents. 3. Blnkenu nd Cmer (200) rgued tht student effort is necessry for skill formtion during eduction. 4. Quote from he New Dictionry of Culturl Litercy, 3rd edition Brron et l. (2003) found tht more thn hlf of the return on college eduction is ssocited with student effort rther thn student bility. 6. Among others, Shin (2004) found tht students respond to lower tuition policy in the U.S. by decresing their level of study effort, resulting in potentil loss of humn cpitl of round 5%. Her findings confirm the conventionl wisdom tht student cdemic effort plys key role in future ernings. 7. While foreign direct investment leds to bis effect fvorble to smll indigenous elite in less developed countries, Fil (983) rgued tht such n increse in equlity, in the context of restricted occuptionl opportunities for industril employment, would crete n environment fvorble to the expnsion of service nd informl sector employment. 8. he Gini coefficient dt re vilble for 28 OECD countries, which re estimted by the United Ntion in different yers. Most re surveyed in one of the following yers: 2000, 200, 2002, nd he remining re surveyed in ny yer from 993 to 999. he ltitude informtion is from Look-up Ltitude nd Longitude found t 9. he popultion census of the economy is given by L + Eθ + E( θ ) = N (Findly nd Kierzkowski, 983). 0. John Eccles ws n Austrlin neurophysiologist, who won the 963 Neurology Nobel Lurete.. Lewin (980) reported the studies of Dr. John Lorber on cerebrl cortex losses. Among

16 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, severl similr cses, young Sheffield University student hd n IQ of 26 nd first clss honors degree in mthemtics, lthough over 90% of his brin tissue ws lost. In other words, he hd virtully no brin. 2. As implied in Figure 2, it wouldn t lter my results if f () is slightly less thn zero. 3. In the U.S. s in mny other countries, the interest rtes of consumer lons (e.g., credit crd rtes) re currently bout three times lrger thn student lon interest rtes, where the former re bout 8% nd the ltter re bout 6%. 4. he nomintor term of B() denotes the entire service sector size nd the denomintor term represents the totl supply of skilled workers. Note tht the lifetime incomes of unskilled workers, either working in the goods or service sector, must be the sme, while unskilled workers cn freely move cross the two sectors. 5. ke derivtive of the right-hnd site of (8) with respect to, we obtin f k[ f k + f k f kk ] f kk f k < 0, where f > 0, 2 k f kk = < 0 nd f k = 0. f kk 6. Note tht s > 0 nd < 0. E 7. As Greenspn (2000) hs put it, he completion of high school used to equip the verge worker with sufficient skills to lst lifetime. ht is no longer true, s evidenced by community colleges being inundted with workers returning to school to cquire new skills nd on-the-job trining being expnded nd upgrded by lrge proportion of Americn business. 8. See the definition of S(, in Section It is becuse of > > Agin, higher threshold of study effort, lthough increses in the school enrollment, is to prtilly compenste for the reducing income inequlity. 2. I rgue tht people generlly llocte time to fight ginst not only bitter cold but lso boiling hot climte, so tht the ltter lso deters the leisure trdeoff in this model. However, the OECD countries in Figure re sufficient wy from the Equtor.

17 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, References Brron, J. M., K. H. Blnchrd, V. Brencic, nd A. Shin Perspirtion or Inspirtion: Wht Do Grdes Mesure? Purdue University. Blnkenu, W. nd G. Cmer Productive Eduction or Mrketble Degree? Mimeo. Knss Stte University nd Purdue University. Chusseu, N., M. Dumont, nd J. Hellier Explining Rising Inequlity: Skill-Bised echnicl Chnge nd North-South rde, Journl of Economic Surveys, 22, Dvis, D. R. nd A. Reeve Humn Cpitl, Unemployment, nd Reltive Incomes in Globl Economy, Hrvrd Institute of Economic Reserch Pper, No Eccles, J. C he brin indictes its powers re endless, lecture t University of Colordo, University Memoril Center Boulder. Feenstr, R. C. nd G. H. Hnson Globliztion, Outsourcing nd Income Inequlity, Americn Economic Review, 86, Issue 2. Feenstr, R. nd G. H. Hnson Foreign Direct Investment And Reltive Wges: Evidence From Mexico's Mquildors, Journl of Interntionl Economics, 42, Fil, R Inequlity nd the Service Sector in Less Developed Countries: Renlysis nd Respecifiction, Annul Review of Sociology, 48, Findly R. nd H. Kierzkowski Interntionl rde nd Humn Cpitl: A Simple Generl Equilibrium Model, Journl of Politicl Economy, 9, Greenspn. A he New Economy, Boston College s Finncil Conference 2000, Krueger, D. nd F. Perri On the Welfre Consequences of the Increse in Inequlity in the United Sttes, NBER Mcroeconomics Annul, 2003, Lewin, R Is Your Brin Relly Necessry? SCIENCE 20:

18 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, Loury G. C. nd D. Grmn College Selectivity nd Ernings, Journl of Lbor Economics, 3, Miller,. C Impct of Globliztion on U.S. Wge Inequlity: Implictions for Policy, North Americn Journl of Economics nd Finnce, 2, Mrtins, S. P. nd P.. Pedro Does Eduction Reduce Income Inequlity? Quntile Regression Evidence from 6 Countries, Journl of Lbor Economics,, Owen, D. J. 97. he Demnd for Leisure, Journl of Politicl Economy, 79, Shin, A he Incentive Effects of Higher Eduction Subsidies on Student Effort, Federl Reserve Bnk of New York Stff Reports, No. 92. Syek, S. nd F. Sener Outsourcing nd Wge Inequlity in Dynmic Product Cycle Model, Review of Development Economics, 0, -9.

19 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, Percentge 50 Figure. Gini Index v.s. Ltitude 45 GINI Coefficient Jpn Dollr LA IUDE Figure 2. Leisure Alloction Equilibrium

20 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, 92-2 Figure 3. With echnologicl Development Figure 4. When Outsourcing Increses

21 Lo, Journl of Interntionl nd Globl Economic Studies, (2), December 2008, Figure 5. When Eductionl Resource Increses Figure 6. Current Leisure v.s. Future Comfort