Stages of Globalization, Inequality and Unemployment

Similar documents
Trade liberalization and labour markets:

The Policy Performance of NFSF and Slippage in Futures Markets

Oil Discovery, Real Exchange Rate Appreciation and Poverty in Ghana

Inventory Decision Model of Single-echelon and Two-indenture Repairable Spares

Consumer Price Indices

Exploring ice content on partially saturated frozen soils using dielectric permittivity and bulk electrical conductivity measurements

Understanding the Performance of Cooperative Web Caching Systems

Demand for meat quantitu and quality in Malaysia: Implications to Australia

THE ROLE OF ENVIRONMENT IN A REGION S SUSTAINABLE DEVELOPMENT AS DESCRIBED BY A BUTTERFLY CATASTROPHE MODEL

Evaluation Method of Banking System Stability Based on the Volume of Subsystems

Geometry Supplement for Math 60 Perimeter, Area, and Volume

Impacts of U.S. Sugar Policy and the North American Free Trade Agreement on Trade in North American Sugar Containing Products

Spatiotemporal Analysis of Marriage and Marital Fertility in Japan: Using Geographically Weighted Regression

OPTIMIZATION OF TWO-DIMENSIONAL WING IN GROUND EFFECT CONSIDERING AERODYNAMIC CENTER OF HEIGHT AND LIFT

Can Survey Bootstrap Replicates Be Used for Cross-Validation?

The Exchange Rate and the Performance of Japanese Firms: A Preliminary Analysis Using Firm-level Panel Data

LABOUR UNIONS AND WAGE INEQUALITY AMONG AFRICAN MEN IN SOUTH AFRICA

Optimization Design of a Piezoelectric Actuator with Orthogonal Theory

Lecture 15: Effect modification, and confounding in logistic regression

Food Marketing Policy Center

Overall stability of multi-span portal sheds at right-angles to the portal spans

CALIBRATION ALGORITHM FOR CURRENT-OUTPUT R-2R LADDERS

Practical design approach for trapezoidal modulation of a radio-frequency quadrupole

Heat Spreading Revisited Effective Heat Spreading Angle

Faculty Research Working Papers Series

Coffee Differentiation: Demand Analysis at Retail Level in the US Market

Optimization Model of Oil-Volume Marking with Tilted Oil Tank

Weight Gain During the Transition to Adulthood among Children of Immigrants: Is Parental Co-residence Important? Elizabeth Baker

The Pennsylvania State University. The Graduate School. College of Agricultural Sciences ESSAYS ON WELFARE USE, THE WAGE GAP AND UNEMPLOYMENT

The Flower of Paradise: Substitution or Income Effect? Sara Borelli University of Illinois at Chicago

The Rise of Obesity in Transition Economies: Theory and Evidence from the Russian Longitudinal Monitoring Survey

epub WU Institutional Repository

Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

Coffeemar. Installation and maintenance guide. Contents: Ctra, Marcilla, Km PERALTA, Navarra SPAIN

AN EVALUATION OF TRAINING

Balanced Binary Trees

Red Green Black Trees: Extension to Red Black Trees

Resource Allocation for Cocoyam and Coffee Production in Momo, North West Region of Cameroon

William C. Hunter. Julapa Jagtiani

Midterm Economics 181 International Trade Fall 2005

Demand Analysis of Non-Alcoholic Beverages in Japan

Ethnic Sorting in the Netherlands

Oil spill forecasting in the Mediterranean Sea

Calculation of Theoretical Torque and Displacement in an Internal Gear Pump

Experimental and Numerical Studies on Flocculation of Sand-Mud Suspensions

Dominance Testing for Pro-Poor Growth with an Application to European Growth

QUICK START GUIDE Armonia

Prediction of steel plate deformation due to triangle heating using the inherent strain method

Math GPS. 2. Art projects include structures made with straws this week.

234 The National Strategies Secondary Mathematics exemplification: Y7

The labour market impacts of adult education and training in Canada

16.1 Volume of Prisms and Cylinders

Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model. Pearson Education Limited All rights reserved.

Migration and Fertility: Competing Hypotheses Re-examined

The responsibility for the contents of this CPB Discussion Paper remains with the author(s)

Development and application of a rural water supply assessment tool in Brazil

Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

Preview. Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

Fixation effects: do they exist in design problem solving?

MARKET NEWSLETTER No 93 April 2015

Do Regional Trade Pacts Benefit the Poor?

The Optimal Wine. A Study in Design Optimization. April 26, Erin MacDonald Alexis Lubensky Bryon Sohns

1/17/manufacturing-jobs-used-to-pay-really-well-notanymore-e/

The collision avoidance control algorithm of the UAV formation flight

Answer Key 1 Comparative Advantage

Conjoint Analysis: A Study of Canned Coffee in Taiwan

International Trade CHAPTER 3: THE CLASSICAL WORL OF DAVID RICARDO AND COMPARATIVE ADVANTAGE

EXTERNAL SHOCKS AND POVERTY: HOW RECESSION IN EUROPE, JAPAN, AND CHINA AFFECTS THE INDONESIAN POOR

Interannual Herbaceous Biomass Response to Increasing Honey Mesquite Cover on Two Soils

Concession Contract Renegotiations: Some Efficiency vs. Equity Dilemmas

Recent U.S. Trade Patterns (2000-9) PP542. World Trade 1929 versus U.S. Top Trading Partners (Nov 2009) Why Do Countries Trade?

László Mester. The new physical-mechanical theory of granular materials

Point Pollution Sources Dimensioning

Preview. Introduction. Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

School Breakfast and Lunch Costs: Are There Economies of Scale? Authors. Michael Ollinger, Katherine Ralston, and Joanne Guthrie

Russell James Department of Scientific and Industrial Research Taupo-ldairakei, New Zealand

Catching up or falling behind in Eastern European agriculture the case of milk production

ANALYSIS OF WORK ROLL THERMAL BEHAVIOR FOR 1450MM HOT STRIP MILL WITH GENETIC ALGORITHM

Further Evidence on Finance-Growth Causality: A Panel Data Analysis

Cardiff Economics Working Papers

Chapter 3: Labor Productivity and Comparative Advantage: The Ricardian Model

Goal: Measure the pump curve(s)

'""' USAFA/ Coord.{!tr lv~ ""' DFCE... ~A.., USAFA/ DFER. Sign C:.dl A:>.-').l'. 23 \,;'~ rs- 7 USAFA-DF-PA- CJ

AN ATTRACTIVENESS-BASED MODEL FOR SHOPPING TRIPS IN URBAN AREAS

To find the volume of a pyramid and of a cone

Description of Danish Practices in Retail Trade Statistics.

Preview. Introduction. Chapter 3. Labor Productivity and Comparative Advantage: The Ricardian Model

Dumping on Free Trade, Optimal Antidumping Duties, and Price Undertakings: Welfare Implications in a Two-Market Equilibrium Analysis

Physics Engineering PC 1431 Experiment P2 Heat Engine. Section B: Brief Theory (condensed from Serway & Jewett)

Preview. Introduction (cont.) Introduction. Comparative Advantage and Opportunity Cost (cont.) Comparative Advantage and Opportunity Cost

Entomology ABSTRACT INTRODUCTION MATERIALS AND METHODS. D. Samoedi. Indonesian Sugar Research Institute, Pasuruan, Indonesia

Background. Sample design

QUESTIONS FOR REFLECTION: VISUAL 4.1 WHY DID THE COLONISTS PROSPER BETWEEN 1585 AND 1763?

I - 1 The IBPGR was requested to: 1. recognize the two designated ISSCT world collections; 2. establish seed repositories; and

Chapter 3 Labor Productivity and Comparative Advantage: The Ricardian Model

Homework 7. problems: 9.33, 9.40, 9.65, 9.78

6 λk 0gh 3 þ 1 6 k. þ 1. z ¼ ð1 z ÞH p þ b qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

Study of Steam Export Transients in a Combined Cycle Power Plant

Modelling Beta Risk for New Zealand Industry Portfolios

Factors Affecting Frequency of Fast Food Consumption

Transcription:

tages of Gobazaton, Inequaty and Unempoyment Joë eer* Abstract To anayse te mpacts of gobazaton upon nequaty and unempoyment n bot advanced and emergng countres, we bud a ort-out O mode wt cency wages based on reatve deprvaton. Gobazaton s caractersed by te out growng n sze and sgnfcant dfferences n sk endowments between te two areas. e generate tree stages of gobazaton dependng on te sze of te out and sowng substanta dvergences n terms of nequaty, unempoyment and productvty. Te ort s caractersed by growng nequaty and unempoyment and by a decrease n productvty at te eary stages of gobazaton. Te out sows a decrease n nequaty and unempoyment at te frst stage of gobazaton and growng nequaty and/or unempoyment at te ater stage. Key ords: Effcency wage; Gobazaton; Inequaty; Reatve deprvaton; Unempoyment. JE Cassfcaton: E24, E25, F6. *Joë eer EQUIPPE, Unversty of e and EMA, Unversty of antes Pers. address: 28 rue de évgné 75004 Pars FRACE +33 42 77 57 40 / joe.eer@wanadoo.fr / Joe.eer@unv-nantes.fr I ws to tank te Frenc Researc atona Agency (AR) for ts fnanca support.

2. Introducton Ts artce anayses te mpact of gobazaton upon nequaty and unempoyment n bot advanced and emergng countres. As regards advanced countres (te ort), te mpact of ort-out trade upon growng nequaty between sked and unsked workers was ntay dsputed because of ts teoretca sortfas (Krugman and awrence, 993; awrence and augter, 993) and ts ack of emprca evdence (Borjas et a, 992; Katz & Murpy, 992). Ts frst dagnoss as subsequenty been reconsdered on te bass of bot new emprca works and new teoretca approaces (Cusseau et a., 2008, for a revew). In ts respect, one of te man arguments n favour of te mpact of gobazaton upon ortern nequaty s tat, wt te advent of Cna and Inda, te out as become a major actor n nternatona trade and producton, wc was not te case n te egtes and eary nnetes (Krugman, 2007). In te case of emergng countres (te out), te resuts of te terature on te mpact of openness upon nequaty are rater ambguous (Anderson, 2005, and Cusseau & eer, fortcomng, for revews). Unt te eary nnetes, te reducton n nequaty n East Asan countres (Krongkaew, 994) appeared to confrm te eckscer-on-amueson (O) predcton of a rse n te reatve prce of te abundant factor,.e., unsked abour n soutern countres. Ts resut as subsequenty been dsputed wen consderng () te deveopments n atn Amerca (Feenstra & anson, 997; Green & Dckerson, 200; Esquve & Rodrguez-opez, 2003; Gaan & angunett, 2003; opez & Mer, 2008) and () te fact tat nequaty as ncreased n most of te Asan countres snce te md-nnetes (Ragaya, 2005). A number of recent emprca works concude tat openness as ncreased nequaty n emergng countres (Berman & Macn, 2004; Manovc, 2005; Conte & Vvare, 2007; Mesc & Vvare, 2009). evera teoretca expanatons can account for tese evoutons. en te soutern country s comparatve advantage s based on natura resources, openness may ncrease nequaty wen tese resources are n te ands of a mnorty (eamer, 987). en tere s a contnuum (or a arge number) of goods wt dfferent sk ntensty, te cornerng of new more sk-ntensve goods by te out prompts an ncrease n te demand for sk, and tereby n nequaty (Feenstra & anson, 996; Zu & Trefer, 2005; u 2003). Fnay, f openness resuts n te adopton by te out of nortern tecnooges tat requre te use of more sked abour, ts can ncrease te sk premum and nequaty n te out (Pssardes, 997; Berman & Macn, 2000). Te mpact of openness upon unempoyment as been anaysed wtn a O framework by Davs (998) so as to expan te dvergence between Europe and te U. In Davs mode, Europe sets a mnmum wage tat moves te sk premum beow ts fu empoyment vaue wereas te U ets market forces work. Ts sfts te word sk premum towards te European sk premum, tereby generatng unempoyment of te unsked, a of wc s ocated n Europe. Davs expanaton can easy be extended to te case of ort-out trade f nortern countres prevent te ncrease of ter sk premum troug abour market nsttutons (mnmum wage, coectve barganng etc.). Ts mode neverteess fas to account for stuatons n wc te ort and te out do not produce te same goods,.e., wen te two countres are not stuated wtn te dversfcaton cone. Te far wage ypotess provdes anoter way to generate trade-drven unempoyment n te ort. Age and undborg (995) generate unempoyment by ntroducng far wage nto a 2 2 2 O mode. Insertng an cency wage ypotess nsde a eckser-onan mode wt capta and abour, Abert & Merck (200) sow tat most of te O resuts are Dreer and Gaston (2008) aso fnd tat economc gobazaton as ad a postve mpact upon wage nequaty n deveopng countres, but ts mpact s non sgnfcant.

3 preserved. owever, neter of tese approaces makes any dstncton between sked and unsked abour, wc makes tem unsutabe for anaysng te unempoyment of unsked workers due to ort-out openness. Kreckemeer & eson (2006) propose a ort-ort O mode wt a far wage-drven ort for bot sked and unsked workers, te mode beng subsequenty extended by te ntroducton of an unsked abour-abundant out. Tese autors focus owever on te ntra-ort dspartes and nteractons and not on te ort-out dfferences. Egger & Kreckemeer (2009) anayse te ects of trade n a mode wt frm-specfc far wages. owever, ter mode does not account for ort-out dfferences and sk dfferences. In te approaces wt two types of workers, te sked and te unsked, far wages are typcay defned foowng Akerof & Yeen (990) as a combnaton of te wage of te oter group and te wage te worker expects to receve f s/e resgns and searces for anoter job 2 (Kreckemeer & eson, 2006; Kreckemeer, 2008). owever, anoter means of defnng far wage s to assocate t wt reatve deprvaton. In te case of advanced countres, te reatve deprvaton ypotess rests upon a arge body of emprca evdence and s used n a arge number of works on nequaty, mgraton and appness (Cark & Oswad, 996, ecton 2 for a revew). evera recent works ave sown tat t aso appes to emergng countres (Rao, 200; arayan & Petesc, 2002; enk, 2004; Ravaon & oksn, 200). Consequenty, n contrast wt mnmum wages, matcng and barganng stores tat are better taored to mode mperfectons n nortern abour markets, reatve deprvaton based cency wage can account for wage rgdty n bot advanced and emergng countres. en reatve deprvaton s based on te frustraton provded by te stuaton of tose wo are better treated, ts ypotess nks te reference wage tat a worker consders to determne er/s ort to te deprvaton s/e suffers from bot te ncomes and te number of tose wo earn more tan s/e does. Fnay, te emergence of an nequaty-unempoyment trade-off n ortern countres as been anaysed to expan te emprca observaton of jobess Europe versus penness Amerca (Krugman, 994). For Krugman (994, 995) ts trade-off comes from te dfferent responses tat ave been gven to te same ncrease n te reatve demand for sked workers. By preventng te reated rse n te sk premum (nequaty) troug abour markets nsttutons, Europe as created unempoyment wereas te U as aowed nequaty to worsen. From a pane of ortern countres, Cecc & Garca-Penaosa (2008) fnd tat a te abour market nsttutons except te tax wedge generate an nequaty unempoyment trade-off, wt abet very dfferent ntenstes. tn a O mode wt a contnuum of goods, eer & Cusseau (200) provde teoretca bases for te gobazaton-drven nequaty unempoyment trade-off and tey sow tat ts trade-off s more ntense n tose countres tat were nequaty-orented at te outset of te gobazaton process. Tese anayses are owever centred on ortern countres. e ntroduce cency wages based on reatve deprvaton nto a ort-out O mode so as to anayse te mpact of gobazaton upon nequaty and unempoyment n bot te ort and te out. Gobazaton s defned by te growng sze of te out and a sgnfcant dfference n te reatve endowments of sked abour between te two areas. Tree stages of gobazaton are put forward tat correspond to te out beng successvey sma, medum-szed and arge. e fnd tat te deveopments n te ort and te out n terms of nequaty, unempoyment and productvty crtcay cange accordng to te pase of gobazaton. In partcuar, nequaty and/or unempoyment decrease n te out and ncrease n te ort durng te frst stage wereas tey constanty ncrease n te out and reman uncanged n te ort durng te trd stage. Productvty decreases n te ort and 2 Akerof & Yeen (990) defne te reference wage as a combnaton of te wage of te oter group and te market cearng wage of te group te ndvdua beongs to.

4 ncrease n te out durng te frst stage(s), and t decreases n te out and remans uncanged n te ort over te ast stage. Te artce s orgna n severa respects. It frsty anayses te mpacts of gobazaton upon nequaty and unempoyment n bot te ort and te out wtn a unfed framework. By ntroducng te growng sze of te out as a key eement of gobazaton, t secondy dstnguses tree stages of te gobazaton process. It trdy provdes an anayss of te canges n nequaty, unempoyment and productvty n eac of tese tree stages. It fnay generates severa outcomes tat are consstent wt certan observed deveopments. Tese ncude te ncrease n bot nequaty and unempoyment and te sow down n productvty n te ort as and wen te out becomes arge enoug, and te reversa of te pro-equaty trend at te ater stages of gobazaton n te out. ecton 2 presents te mode and ts possbe equbra n autarky. ecton 3 determnes te tree stages of gobasaton. Te canges n nequaty, unempoyment and productvty n eac country at eac stage are anaysed n secton 4. e dscuss tese fndngs and concude n ecton 5. 2. Te mode e construct a ort-out O mode wt cency wages based on reatve deprvaton. 2.. Genera framework Tere are two factors, sked abour and unsked abour, two goods and tat are respectvey -ntensve and -ntensve, and two countres, te ort () and te out (), te former beng reatvey better endowed wt sked abour and te atter wt unsked abour. Markets for goods and factors are compettve. Tecnooges are dentca and te reatve endowments = /, =,, are constant n bot te ort and te out. Tese reatve endowments are suc tat te wage of te unsked w s aways ower tan te wage of te sked w, wc ndcates tat te sk premum w / w tat measures nequaty s aways ger tan. Te workng popuaton s constant n te ort. Conversey, we aow for an ncrease n te workng popuaton n te out, and tereby n te out s endowments of bot and. In bot sectors, output s determned by te amount of sked and unsked abour utsed n producton and by te ort made by workers. Ts ort depends on te dfference between te wage a worker receves and a reference wage tat denotes reatve deprvaton. Consequenty, a worker decdes on er/s ort by consderng bot te wages and te proporton n te workng popuaton of tose wo earn more tat s/e does. In addton, a worker suppes one unt of abour tme watever er/s sk. In ne wt te usua cency wage framework, we assume tat frms cannot contro te workers orts but know ter ort functon. Tey can tereby determne te cency wage tat maxmses ter proft. If te fu empoyment wage s ger tan te cency wage, te former stands out. In contrast, f te fu empoyment wage s ower tan te cency wage, frms enforce te atter and ts resuts n unempoyment.

5 2.2. Effcency wage and cency sk premum Te producton functons n sectors and are j j j j ( j j ) ( j j ) Y = A E E, j =,, were E j depcts te ort of -workers, =,, n j-ndustry. Te utty functon s smar to tat presented by Kreckemeer and eson (2006), except n te defnton of te reference ort. In bot countres, workers maxmse te same utty functon u = v ( x, x ) v2 ( E), were x j s ter consumpton of good j, w ter wage and E te ort tey provde, subject to te ncome constrant w px + p x, wt p j te prce of good j, and to te non-negatvty of ort 0 β E. e assume ( ) ( ) v = x x so tat eac good accounts for a constant sare of tota expendture n bot countres and at te word eve,.e., β for good and ( β ) for good. In addton, ( ) 2 β v 2 ( E) = E E * depcts w aqw te dsutty of ort above te reference vaue E*. e suppose E* = f ( aq) w w aqw and E* = 0 f w < aqw, wt w te average wage of te workers wo are better pad (and w = w f nobody s better pad), q ter proporton n te workng popuaton, 0 < γ <, and cocent a, 0 < a < / q, depctng te deprvaton ntensty. It s assumed tat a s te same n bot countres. Te maxmsaton of utty resuts n te foowng ort functon: γ w aqw f w aqw E( w) = ( aq) w 0 f w < aqw Te nterpretaton of ts resut s as foows. Te ndvdua consders te wage of tose wo are te better pad. If s/e beongs to te better pad ( w = w ), s/e makes te maxmum E( w) = w / w γ =. In te opposte case ( w < w ), s/e reduces er/s ort ort ( ) γ γ w aqw w w E( w) = = by an amount tat ncreases wt te dfference ( aq) w ( aq) w w w between te wage of te better pad and er/s wage. Ts denotes te punsment te worker nfcts upon te empoyer for beng unfar to er/m, wc es at te very core of te far wage ypotess. If w aqw, te pay s consdered so unfar tat te worker provdes zero ort. Fnay, functon E( w ) vares between 0 and and decreases n bot w and q, wc estabses a reatve deprvaton beavour. It can be easy verfed tat sked workers provde te maxmum ort 3 : E( w ) =. Te ort of te unsked E( w ) s determned by: γ () γ w aq w f E( w ) = ( aq ) w 0 f w w aq w < aq w (2) wt q = / ( + ) <, wc sows tat q w s aways ower tan w. 3 For te sked workers, q = 0 and w = w = w.

6 Te frms n country and sector j frsty determne te unsked workers cency wage w (equa n bot ndustres, see ereafter) tat maxmses ter proft p Y w w j j Y = A E ( w ) and π j = j j j j suc tat j j j ( j ) w ( ) aq w E w = ( aq ) w w γ. If te market forces determne a fu empoyment wage tat s ger tan, ten te fu empoyment wage stands out. If te fu empoyment wage s ower tan te cency wage, ten te frms enforce te atter. Defnton: Te cency sk premum s te rato of te sked workers wage cency wage determned by te frms. w to te In country and sector j, te frms maxmsaton programme generates te foowng vaues of te cency wage optma ort w E Remarks: w, te cency sk premum = w / w and te E (cacuatons n Appendx A): a = q w γ γ γ = = ( + ) (4) aq a γ γ aq γ a = = γ aq γ a +. Te cency wage, te cency sk premum and te correspondng ort are dentca n bot sectors. 2. e encefort suppose tat q < ( γ ) / a wc s te condton for >. 3. nce frms never pay unsked workers beow te cency wage, ten te sk premum never exceeds ts cency vaue γ. As a consequence: (3) (5) emma : et be country ' s fu empoyment sk premum. Ten, country s sk premum at te equbrum s * mn{, } =. 4. Te ger te proporton of sked workers nsde a country, te ower ts country s cency sk premum ( / q < 0 ), and snce q > q : emma 2: Te cency sk premum s ger n te out tan n te ort: >.

7 2.3. Equbra n autarky In autarky te mode generates te foowng fu empoyment sk premum (Appendx B): A q = =, =, (6) q wt β ( β ). Gven te comparatve advantages >, te fu empoyment sk premum s ger n te out tan n te ort n autarky: A A > emma 3: Consder country n autarky wt te proporton q of sked abour n ts workng popuaton. Ts country s at ts fu empoyment sk premum A f q q *, and t s at ts cency sk premum and exbts unempoyment of unsked abour f ( γ ) q < q *, wt q * =. a( ) Proof. see Appendx C. As usua n cency wage modes, wen te fu empoyment sk premum A s ger tan te cency sk premum workers. From emmas, 2, 3 and nequaty q, ts generates unempoyment of unsked > q, we derve te foowng proposton: A Proposton : In autarky, country s sk premum s * mn{, } tree stuatons ony are possbe: =, =,, and ) Fu empoyment n bot te ort and te out wt * = and * = f * > q q q 2) Fu empoyment n te ort and underempoyment of te unsked n te out wt A * * = and * = f q > q > q. 3) Underempoyment of te unsked n bot te ort and te out wt * = * and * = f q > q > q. A A Bot countres beng n autarky, Proposton sows tat f te out s at fu empoyment, ten te ort s at fu empoyment as we, wereas te ort beng at fu empoyment can come wt fu empoyment or under empoyment n te out.

8 3. Te stages of Gobazaton e ntroduce gobazaton wt free trade between te ort and te out. 3.. Gobazaton Te gobazaton process s caractersed by: ) An ncrease wt tme n te sze (workng popuaton) of te out wereas te sze of te ort remans uncanged. Assumng a growng sze of te out ams at portrayng te fact tat new soutern regons and countres contnuousy enter te gobased economy, and tat ts process comes to an end once te woe word as joned ts economy. e aso suppose tat te sze of te out s neggbe at te outset of te gobazaton process, and tat te ort becomes mnute compared to te out at ts concuson. Ts resuts n te word factors endowment beng tat of te ort at te begnnng, and tat of te out at te fna stage of gobazaton. 2) Te dfference n te reatve factor endowments s arge enoug to pace te out outsde te dversfcaton cone (encefort D-cone) for te word fu empoyment sk premum at te begnnng of te gobazaton process, and to pace te ort outsde te cone for te word fu empoyment sk premum at te concuson of gobazaton 4. Te condton for ts s / > max, (see Appendx D, ecton2). Ts ams at portrayng te arge dvergence n sk reatve endowments between advanced and emergng countres tat caracterses te gobasaton process. e aso assume tat:. Te out s at fu empoyment wen producng good ony, wc mpes tat ts fu empoyment sk premum wen producng ony s ower tan ts cency sk premum: < (7) Ts does not mean tat tere s no unempoyment n deveopng countres wen te out produces ony, but tat te regons tat are nserted n te word economy and tereby partcpate n te gobasaton process are at fu empoyment wen tey ony produce good. 2. Te ort s at fu empoyment at te outset of te gobazaton process, wc ndcates tat ts autarkc fu empoyment sk premum s beow ts cency sk premum. In contrast te out can, eter be at fu empoyment, or suffer unempoyment of te unsked at te start of gobazaton (Proposton ). 3. Te ort suffers unempoyment wen producng good ony, and unempoyment appears for a dstrbuton of ts producton between te two goods wt a non-zero producton of. Ts ndcates tat ts cency sk premum s ower tan ts fu empoyment sk premum wen producng good ony. Assumptons 2 and 3 enta: A < < (8) Te reease of tese assumptons s dscussed n ecton 5 e fnay suppose tat reatve deprvaton s domestcay determned,.e., tat unsked workers consder te wage and te proporton of sked workers nsde ter own country wen determnng ter ort. Ts ndcates tat te cency sk prema, =,, as 4 ee te defnton of te dversfcaton cone n Appendx D, ecton.

9 determned by equaton (4) od n autarky as we as n ort-out openness. Consequenty, cency sk prema are constant durng te woe gobazaton process (as s constant). Te fu empoyment sk premum at te word eve s (Appendx B): = (9) wt = /, = + and = +. As / < / < /, ten: A A < <, wt A beng te autarkc fu empoyment sk premum of country as defned by (6). ereafter, country s fu empoyment sk premum s denoted, te word fu empoyment sk premum wen bot countres produce bot good, and country s fu empoyment sk premum wen produces good j ony j wt: j j =, =,, j =,. (0) It must be noted tat: j ) Te fu empoyment sk prema, =,,, are cangng over tme wt gobasaton. 2) Te vaues, A and j are constant over tme and fuy determned by te mode parameters and te gven reatve endowments. A 3) < <, =,, snce bot countres produce bot goods n autarky. At te outset of gobazaton, te sze of te out neggbe and tus A =. At te end of te gobazaton process, te ort becomes mnute compared to te out and tus A =. As a consequence, te word fu empoyment sk premum moves from A up to A trougout te gobazaton process. Fgure 2 depcts ts deveopment wen te ort s at fu empoyment and te out suffers underempoyment n autarky. A A x Gobazaton x Fgure. Te gobazaton process Te outern reatve endowment and te ortern endowments and beng constant, an ncrease n te wegt of te out resuts n an ncrease n te word reatve endowment / because te out s te unsked abour abundant country. As te dversfcaton cone s determned by te two nes = and =

0 (Appendx D, ecton ), and snce = (Reaton 9), te ncreasng sze of te out entas a rse n and tereby a rotaton to te eft of te dversfcaton cone (Fgure ). In addton, snce te outern reatve endowment remans constant, te growng sze of, aong te te out may be represented by an upward transaton of ts endowments ( ) ne = (Fgure 2). = = = out ort Dversfcaton cone = Fgure 2. Gobazaton, te dversfcaton cone and factor endowments In Fgure 2, te gobazaton process s tus depcted () by an upward dspacement of te out aong te ne =, () by a rotaton to te eft of te dversfcaton cone, and () by an mmobty of te factor endowments of te ort (, ) 3.2. Te tree stages. Proposton 2: Gobazaton s caractersed by tree successve stages suc tat: ) Te ort produces bot goods and te out good ony durng te frst stage. 2) Durng te second stage, eter bot countres produce bot goods, or te ort produces ony and te out ony. 3) Te ort produces good ony and te out bot goods durng te trd stage. Te way te ncreasng sze of te out generates te tree successve stages as we as te dfferent possbe confguratons of eac stage are anaysed and dscussed n Appendx E. nce te successve stages are generated by te ncrease n te sze of te out, we ca te frst stage ma out, te second stage Medum-szed out and te trd stage arge out. Eac stage s tus caractersed by bot te sze of te out and te correspondng specasaton of eac country n te producton of goods. Fgures 3-5 depct te poston of eac country n reaton to te dversfcaton cone at eac stage of te gobazaton process.

= = = = ort = out ort out Dversfcaton cone (a) Dversfcaton cone (b) Fgure 3. ma out out ort out ort Dversfcaton cone Dversfcaton cone (a) (b) Fgure 4. Medum-szed out out Dversfcaton cone ort Fgure 5. arge out

2 tage frsty corresponds to te ort beng nsde te cone and producng bot goods, wereas te out stands outsde te cone, producng tereby good ony (Fgure 3a). But te frst stage can contnue even wen bot countres are nsde te D-cone (Fgure 3b). Ts s wen te ort s at ts cency sk premum and te out cannot attan fu empoyment at ts sk premum even wen producng good ony (te out s stuated above ne = / ( ) and tus < ). Ten, te sk prema of te ort and te out cannot equase and bot countres cannot produce bot goods smutaneousy. Tus, as ong as te out s not suffcenty arge for ts producton to meet te word tota demand for good, te out produces ony and te ort bot goods. At te second stage, te out becomes arge enoug so tat, eter bot te ort and te out produce bot goods, or te out produces good ony and te ort good ony. In te former case, bot te ort and te out are nsde te cone and tey sare te same sk premum (Fgure 4a). Te atter case typcay occurs wen bot countres are outsde te D- cone (Fgure 4b). Tere are owever two cases n wc te second stage wt eac country producng one good ony s mted to one pont of tme correspondng to te passage from stage to stage 3. Ts frsty appens wen nes = and = merge at te same tme as nes = and =, wc s unkey. It aso appens n te case aready mentoned were bot countres are st n te frst stage wen beng bot nsde te D-cone because tey cannot produce bot goods at te same tme (see above). t te out growng n sze, te ort producton of dmnses and a moment comes wen te ort produces ony and te out bot goods,.e., stage 3. Ten, te passage from stage to stage 3 goes troug te moment (pont of tme) wen te ort produces ony and te out ony. At te trd stage of gobazaton, te out becomes suffcenty arge to produce bot goods wt te ort producng good ony. ne = s nsde te cone and te ortern endowments (, ) outsde (Fgure 5). Proposton 3. Durng te gobazaton process: ) If < <, ten () te ort remans at ts fu empoyment sk premum durng stage and () bot countres produce bot goods durng stage 2, wt te word sk premum frsty ncreasng and subsequenty remanng at vaue. 2) If < <, ten te ort attans and subsequenty remans at ts cency sk premum durng stage, and tere s one soe pont of tme, correspondng to te passage from stage to stage 3, wen te ort produces ony and te out ony. 3) If < <, ten te ort sk premum ncreases to attan ts cency vaue and subsequenty remans at ts vaue durng stage, and te ort produces good ony and te out good ony durng stage 2. 4) Durng stage 3, te ort remans at ts cency sk premum and te soutern sk premum ncreases from to A A f >, and from to A f > Proof. Appendx E. Proposton 3 provdes te condtons for eac possbe confguraton of gobazaton.

3 4. Inequaty, unempoyment and productvty e anayse te canges n nequaty, unempoyment and productvty n bot te ort and te out durng te tree stages of gobasaton. e start from te dfferent confguratons defned n Proposton 3 and we appy te foowng features determned above: ) en a country s at ts cency sk premum, te atter beng ower tan ts fu empoyment sk premum, ts entas unempoyment of te unsked n te country, and an ncrease n te dfference between te fu empoyment and te cency sk premum rases unempoyment. 2) An ncrease (decrease) n te sk premum entas a decrease (ncrease) n unsked workers ort. 3) An ncrease (decrease) n unsked workers ort entas an ncrease (decrease) n productvty 5. Tabes -3 depct te deveopments n nequaty, unempoyment and productvty durng te frst two stages of gobazaton for eac possbe confguraton. Tabe 4 depcts tese deveopments at stage 3 snce tey are te same n te tree confguratons. tages Te ort produces bot goods and te out ony 2 Bot countres produce bot goods Tabe. Te case < < Inequaty (sk premum) ort. Contnuousy ncreasng. out. Constant, ower tan n autarky. ort. Increasng at frst and constant afterwards. out. Increasng at frst, and constant afterwards Empoyment / Unempoyment ort. Fu empoyment out. Fu empoyment ort. Fu empoyment at frst; contnuousy rsng unempoyment as soon as s attaned. out. Fu empoyment Productvty (ort) ort. Contnuousy decreasng. out. Constant, ger tan n autarky (gest possbe for ) ort. Decreasng at frst and constant after. out. Decreasng at frst and constant afterwards. tage Te ort produces bot goods and te out ony Tabe 2. Te case < < Inequaty (sk premum) ort. Increasng at frst and constant afterwards. out. Constant ower tan n autarky., Empoyment / Unempoyment ort. Fu empoyment at frst; contnuousy rsng unempoyment as soon as s attaned. out. Fu empoyment Productvty (ort) ort. Decreasng at frst and constant after. out. Constant, ger tan n autarky (gest possbe for ) tage2: Te ort produces ony and te out produces ony. tage 2 corresponds to te pont of tme wen te word economy passes from stage to stage 3 of gobasaton 5 e tus consder apparent tota factor productvty.

4 tage Te ort produces bot goods and te out ony 2 Te ort produces ony and te out produces ony Tabe 3. Te case < < Inequaty (sk premum) ort. Increasng at frst and constant afterwards. out. Constant ower tan n autarky. ort. Constant out. Constant, Empoyment / Unempoyment ort. Fu empoyment at frst; contnuousy rsng unempoyment as soon as s attaned. out. Fu empoyment ort. Constant gest possbe unempoyment for. out. Fu empoyment. Productvty (ort) ort. Decreasng at frst and constant after. out. Constant and ger tan n autarky (gest possbe for ) ort. Constant wt ort E ( ) out. Constant. tage 3 Te ort produces good and te out bot goods Tabe 4. Te deveopments n stage 3 Inequaty Empoyment / (sk premum) Unempoyment ort. Constant ort. Constant. unempoyment. out. Fu empoyment out. Contnuous A A ncrease f > f ; > ; fu empoyment foowed by ncrease at frst and constant after f A < rsng unempoyment as soon as s attaned f A < Productvty (Effort) ort. Constant wt ort E ( ) out. Contnuous A ncrease f > and ncrease at frst and constant as soon as s attaned f A >. Te deveopments n nequaty at eac stage of gobazaton and for eac confguraton drecty stem from Proposton 3 (see te anayses n Appendx E). Te deveopments n unempoyment and productvty are stragtforward gven te features recaed at te begnnng of secton 4. owever, one pont deserves to be dscussed,.e., te fact tat te out remans at fu empoyment n te second stage of gobazaton wen bot countres produce bot goods (case < <, Tabe ) even wen ts sk premum s, wc s ower tan te word fu empoyment sk premum. In fact, t s te ort tat mposes ts ower sk premum, wereas te sk premum can ncrease n te out snce <. e are tus n te same stuaton as tat anaysed by Davs (998), except tat t s te nortern frms tat enforce te sk premum ere wereas ts s enforced by te government n Davs mode. Consequenty, and for te same reason as n Davs (998), te word unempoyment generated by te sk premum < s totay ocated n te country tat mposes te ower sk premum,.e., te ort. From te deveopments depcted n Tabes -4, t s cear tat: ) Durng te frst stage(s) of gobazaton, te ort suffers growng nequaty, growng unempoyment as we as a decne n productvty (ort), wereas te out benefts from ower nequaty (compared to autarky), fu empoyment and a rse n productvty.

5 2) Te ast stage of gobazaton s caractersed by steady nequaty, unempoyment and ort n te ort, and by growng nequaty and/or unempoyment and decnng productvty n te out. 5. Dscusson and concuson Te man fndngs of te mode are frsty dscussed n te gt of te observed deveopments. e subsequenty reax te assumpton of constant sk endowments and expore te mpacts of sk upgradng. e fnay come back to certan smpfyng assumptons and examne te ects of ter wavng, wc opens new feds of researc. 5.. Te man fndngs facng observed facts e compare te mode s predctons wt te observed deveopments n te forty ast years. A frst dagnoss must be made concernng te begnnng and te present stage of te gobasaton process. e suppose tat gobasaton as begun n te md or ate sxtes, wen te frst wave of ICs (ew Industrased Countres) decded to ean ter economc deveopment on exports to advanced countres. Ten came a second wave of ICs n te seventes/egtes. Gobazaton as acceerated n te nnetes and 2000s wt te deveopment of Cna and Inda as major actors of te word economy. Fnay, consderng () tat unsked-ntensve manufactured goods are no onger produced n nortern countres (Japan, te U and estern Europe), and () tat soutern countres do not produce sk ntensve goods yet 6, we can suppose tat we are now n te second stage of gobasaton wt eac country producng one good ony. Certan predctons of te mode ceary ft wt te observed deveopments, wereas oters appear more controversa. Te confrmed predctons comprse growng nequaty and growng unempoyment n te ort durng te frst stage(s) of gobasaton. Te fa n productvty due to ower ort can provde an addtona expanaton to te productvty (oow) paradox tat occurred n te egtes and eary nnetes n advanced countres. Te mode aso predcts te substanta ncrease n productvty observed n emergng countres wen tey jon te gobazed economy. Certan resuts concernng te out are more dsputabe. Frsty, unempoyment st exsts n emergng countres. owever, as aready mentoned, te out must not be understood as a sum of countres, but rater as an aggregate of regons tat are nserted nto te gobased economy. Ts means tat te areas tat are st specased n tradtona sectors (subsstence agrcuture, tradtona craft ndustres) do not beong to te out as defned ere. Tere s owever a stuaton n wc unempoyment can persst n certan soutern regons even wen tey are fuy gobazed and tey ony produce good. Ts s wen tere s a g fow of ow sked workers wo mgrate to certan gobazed areas 7. In ts case, te nduced fa n sk endowment can move te soutern fu empoyment sk premum above te soutern cency sk premum, and create tereby unempoyment n te out. Ts case s not anaysed ere because we draw attenton to te out as a woe and do not account for mgratons. 6 By assumng ts, we consder tat te frst wave of ICs (Korea, ong Kong, ngapore and Tawan) no onger beongs to te out. In addton, f Cna s present n sk ntensve ndustres (ICT) ts producton s st concentrated n unsked ntensve segments (Krugman, 2008). 7 arrs and Todaro (970) ave sown tat mmgraton can create unempoyment wen te beneft from mgratng s g enoug to offset te decrease n te expected return resutng from unempoyment.

6 Anoter questonabe resut s te decrease n nequaty n te out at te frst stage(s) of gobasaton. Ts s consstent wt East Asa n te egtes and eary nnetes, but not wt te experences of atn Amerca, Cna and even East Asa from te md-nnetes. Bot te teoretca and emprca teratures pont to te gobazaton-nduced tecnoogca transfers from te ort to expan tese ncreases n nequaty. By nsertng suc transfers n te mode, we coud easy generate two stuatons. Inequaty augments n te out wen te mported tecnoogy n sector s suffcenty sk-ntensve to ncrease te demand for sked worker after ort-out openness (te out producng ony wt te nortern tecnoogy) compared to autarky (te out producng bot goods wt ts own tecnoogy). In te opposte case, nequaty decreases n te out at te frst stage of gobazaton. As a consequence, te mode can easy be modfed to generate te two cases observed n te reaty. Fnay, te mode provdes severa predctons for te ast stage of gobazaton. Ts concerns future deveopments. owever, as certan eary emerged economes ave gone up te sk ntensty adder, some of tem avng even ntegrated te ort (Korea, Tawan, ngapore), one can draw on tese experences to dagnose te man caracterstcs of stage 3. In ts respect, tree observed deveopments are n ne wt te mode s resuts. Frsty, even wt a arge ncrease n ter sk endowments, tese countres ave experenced ncreasng nequaty snce te md-nnetes. econdy, te rytm of growt of ter productvty as decreased compared to te seventes and egtes, even f ts growt s st eevated. Trdy, tese countres producton structures ave moved from purey unsked-ntensve sectors to ncreasngy sk ntensve ones. Tese features can be seen as ndcatons tat te mecansms generated by te mode ave operated, even f oter factors may aso expan tese deveopment (pubc poces, sk upgradng, tecnoogca catc-up etc.). 5.2. k upgradng e ave assumed uncanged sk endowments n bot areas. For te out, ts assumpton s justfed by te fact tat new areas wt a g proporton of ow sked abour contnuousy jon te gobazed economy. owever, a nortern countres ave known sgnfcant ncreases n ter sk endowments for te trty ast years. mary, te educaton pocy mpemented n Cna snce te ate nnetes as substantay ncreased te number of students n tertary educaton, wc w resut n sk upgradng of te workng popuaton. Gobazaton tsef can generate endogenous canges n sk endowments. nce te semna work of Fnday and Kerzkowsk (983), te mpact of trade upon sk accumuaton as been abundanty anaysed. Two opposte mecansms ave been put forward. Frsty, ort-out trade nduces a rse n te sk premum n te ort and a decrease n te out, wc fosters sk accumuaton n te former and dscourages t n te atter, renforcng tereby te nta comparatve advantages (Fnday and Kerzkowsk, 983; Borsook, 986). On te oter and, trade favour te ort-out convergence n sk endowments because sked abour s te man factor used n educaton. By ncreasng te sk premum, openness rases te cost of educaton and nders educaton n te ort, and te opposte mecansm operates n te out (Cartga, 997; Ecer, 999). Fnay, gobazaton can encourage nortern governments to support educaton so as to adapt ter economes. Te nfuence of gobazaton on endogenous sk accumuaton s not anaysed ere. e ony assume exogenous rses n te countres sk endowments and dscuss ter mpact upon nequaty and unempoyment.

7 An ncrease n te sk endowments n te ort and/or n te out decreases rato, = and/or, and tereby te word reatve endowment = /. Ts reduces te word fu empoyment sk premum. It aso reduces te cency sk premum as we as te fu empoyment sk prema n te country n wc sk upgradng occurs. Fnay, te ncrease n te sk endowment of te ort and/or te out aways reduces nequaty n bot countres, except n te case of a country tat () produces one good ony and () knows no cange n ts sk endowment. Te mpact of sk upgradng upon unempoyment s ess obvous because a decrease n, =,, owers bot te cency sk premum and te fu empoyment sk premum n country. Te possbe cases are many and tey depend on te stage and sape of gobazaton, on te country(es) tat experence(s) sk upgradng, and on ts ntensty. It can owever been sown tat te ncrease n one country s sk endowment can ead to tree possbe outcomes: () a decrease n unempoyment n bot countres; () a decrease n unempoyment n te country tat experence sk upgradng wtout mpact on te oter country s empoyment; () no cange n (un)empoyment n bot countres. 5.3. Assumptons and furter researces e now reconsder certan smpfyng assumptons. e ave assumed tat te out s at fu empoyment wen producng good ony ( < ). In te opposte case, te out remans at ts cency sk premum and suffers unempoyment trougout te gobazaton process. Te owest unempoyment occurs n te frst stage(s) of gobasaton, and unempoyment subsequenty ncreases wen and as te out begns to produce good. e ave aso supposed tat te ort s at fu empoyment at te outset of te gobazaton process and suffers unempoyment wen t produces good ony A A ( < < ). If tere s unempoyment n te ort n autarky ( < ), ten te ort remans at ts cency sk premum wt unempoyment trougout te gobazaton process. Unempoyment frsty ncreases as ong as te ort produces bot goods, and subsequenty remans uncanged and g once te ort produces good ony. In contrast, f te ort s at fu empoyment wen producng good ony ( < ), ten te ort s aways at fu empoyment and ts nequaty rses unt t reaces. Fnay, even wen consderng cangng sk endowments, we ave mpcty assumed tat te dfference n endowments between te ort and te out was arge enoug to make te ort qut te dversfcaton cone at a moment durng te gobasaton process,.e., to generate te trd stage of gobasaton. Anoter possbty s tat te out sk endowment catces up te ort sk endowment from a certan tme. Tat s wat appened wt te frst wave of ICs tat can now be consdered as nortern countres. Ts catc-up dynamcs, te condtons for soutern countres to enter te nortern area, and te nduced canges n nequaty and unempoyment are mportant ssues to anayse. But ts s anoter story. APPEDI A In sector j and country te frms maxmse ter proft π = p Y w w suc tat j ( ) ( ( ) ) j Yj Aj j E w w = j and ( ) aq w E w =. ( aq ) w j j j j j γ

8 j j j j j j π / = 0 p Y = w / (A) ( ) j π j / w = 0 p jyj = w aq w (A2) γ j j j j j j j π / = 0 p Y = w / ( ) (A3) (A) and (A3) determne te usua reaton between j j = j j j and j (wt = w / w ): w Combnng (A) and (A2), te frm s optmum s suc tat ( w aq w ) wc determnes w aq w = γ, a (A4) j j γ γ = = ( + ), and E aq APPEDI B. Determnaton of te sk premum and empoyment To smpfy, te subscrpt ndcatng te country s omtted. Because of te utty functon, te tota demand for ( Y ) and ( Y d s p Y ( Y and d j =, γ γ = γ d ) are γ d p Y. j = β I and = ( β ) I, wt I = w + w beng te country s tota ncome. Equasng suppy s Y ) and demand on bot markets yeds: s p Y = β ( w + w ) (A5) s p Y = ( β )( w + w ) (A6) Because of te producton functons, te demands for unsked abour n bot sectors at s s te frms optmum are = py / w and = py / w, and tereby at te country s eve: s s = + = p Y / w + p Y / w (A7) = β + ( β ) w ( w + w ). ence Insertng (A5) and (A6) nto (A7) yeds ( ) w =, wt β ( β ). Te fu empoyment sk premum s tus: w = (A8) Fnay, te demand for unsked workers resutng from any < and fu empoyment of te sked s: = < (A9) Reatons (A8) and (A9) appy for eac country beng n autarky as we as at te word eve wen bot te ort and te out are nsde te dversfcaton cone. Consequenty Te word fu empoyment sk premum s =, wt = + and = +.

9 APPEDI C en te fu empoyment sk premum s ger tan te cency sk premum, ts generates unempoyment of unsked workers n country. In fact, as >, ten te frms enforce. Te reated empoyment of unsked workers s =, wc generates unempoyment ( < ) snce = and >. γ q e know tat = and =. Tus: aq q > ( ) γ > q > = = q = γ q * < aq < q < a( ) > > ence: = q = q *. < < Te two possbe cases are: ) q q * = Country s at ts fu empoyment equbrum 2) q < q * < = Country suffers unempoyment of te unsked. Ts estabses emma 3. Fnay, te condton for q * > 0 s APPEDI D.. Te dversfcaton cone ( γ ) a >. Te dversfcaton cone (encefort D-cone) s te set of a te coupes (,) of factor utsatons correspondng to a te producton mx of goods and for a gven vaue of te word fu empoyment sk premum. In te quadrant (,), te D-cone s tus te set n-between te nes = and =. For country to reac fu empoyment at te O equbrum wt te word fu empoyment sk premum, ts factor endowment (, ) must tereby beong to te correspondng D-cone. 2. Te condton for te out to be outsde te D-cone at te start of gobazaton and te ort outsde te D-cone at ts concuson A Te word fu empoyment sk premum s equa to = ( ) / at te begnnng A of te gobazaton process and to = ( ) / at ts concuson. For te out to be outsde te D-cone at te start of gobasaton, we must ave >,.e., >. For te ort to be outsde te D-cone at te concuson of te

20 gobasaton process, we must ave <,.e., <. Fnay, for te out to be outsde te cone at te start of gobazaton and te ort to be outsde at ts concuson, we must ave > max,. APPEDI E e frsty estabs te foowng emma tat w be utsed beow: emma 4. If bot countres produce bot goods, ten <. Proof: en bot countres produce bot goods n free trade, tey sare te same sk premum. As < (nequaty 7) te owest possbe sk premum n te out s. In addton, te ortern sk premum cannot be ger tan. Consequenty, te set of te possbe sk prema of te ort and te set of te possbe sk prema of te out do not overap for <. ence, f <, bot countres never sare te same sk premum and tey cannot produce bot goods at te same tme. out = = ort Dversfcaton cone = = Fgure A. Te D-cone and te countres endowments trougout gobazaton Te gobazaton process conssts n () an upward dspacement of te out aong te ne =, () a rotaton to te eft of te D-cone, and () an mmobty of te nortern endowments (, ) on te ne =. Fnay, te D-cone s demted by nes = and = (Fgure A). At te begnnng of gobazaton, te ort s nsde te D-cone and te out outsde, wt te word fu empoyment sk premum beng A = <. At te concuson of te gobazaton process, te out s nsde te D-cone and te ort outsde, wt

2 A =. Consequenty, () te ne = stands n-between nes = and =, and te ne = above te ne = at te outset of te gobazaton process, and () te ne = stands n-between nes = and =, and te ne = under te ne = at te concuson of te process. To anayse te caracterstcs of stages and 2, we frsty make a dstncton between te two ways by wc te word economy can eave te stuaton n wc te ort s nsde and te out outsde te D-cone: Case. en nes = and = merge before nes = and =, te ort quts te D-cone before te out enters te D-cone, and bot te ort and te out are outsde te D-cone from te moment wen = = (correspondng to te mergng of = and = ) unt te moment wen = = (mergng of = and = ). Te condton for beng n ts confguraton s < (because ne = s ocated above ne =, ence = < wen = and = merge,.e. wen = = ). In ts case, tere s a perod of tme, correspondng to stage 2, durng wc eac country produces one good ony ( for te out and for te ort). In addton, snce < by assumpton, we ave < < and te ort as tus aready attaned ts cency sk premum before te mergng of = and =. ence, te ort attans and remans at te sk premum before enterng te second stage of gobazaton. Case 2. en nes = and = merge before nes = and =, te out enters te D-cone before te ort eaves t, and bot te ort and te out are nsde te D-cone from te moment wen = (mergng of = and = ), unt te moment wen = (mergng of = and = ). Ts corresponds to <. Two cases can ten be dstngused:

22 Case 2a. If < <, ten te ort s at ts cency sk premum wen te out enters te cone (at ts moment = and > ). In addton, as <, ten bot countres cannot produce bot goods even f tey are n te cone (emma 4). As a consequence, te ort and te out are nsde te D-cone ( < ) but tey cannot produce bot goods ( < ). en te out enters te cone, ts sk premum s tus and te out contnues to produce ony as ong as te ort s arge enoug to provde te word wt ts tota demand for. Durng ts tme te word economy tus remans n stage of gobazaton (snce te ort produces bot goods and te out ony) even wt bot countres beng nsde te D-cone. t te ncrease n te sze of te out, tere s a moment from wc te ort becomes too sma to meet te word demand for good. From ten, () te ort produces goof ony wt te sk premum (snce < ), and () te out produces bot goods wt a sk premum tat contnuousy ncreases from to A f <, and to n te opposte case. Ts s stage 3 of gobasaton. Tere s tus one pont of tme tat corresponds to te passage from stage to stage 3, wen te out produces ony and te ort ony. Ts s te moment wen te szes of te ort and te out are suc tat, for te respectve sk prema and n eac country, te ort s just arge enoug to meet te word tota demand for, and te out arge enoug to meet te word tota demand for. Case 2b. If < <, ten te ort s not at ts cency sk premum wen te out enters te cone (at ts moment = and < ). Consequenty, bot countres are n te D-cone ( < ) and sare te same sk premum as ong as. Tus, bot countres produce bot goods durng te second stage. As bot countres are n te D-cone for, and snce <, ten tere s a moment durng te tme wen bot countres are n te cone wen reaces. From ten, te ort remans at ts cency sk premum (t cannot go beyond) and tere s aways a producton mx of and tat ensures fu empoyment n te out for ts vaue (snce < < ). Te cases presented above are summarsed n Tabe A, wc estabses Proposton 3. Condton < < < < < < Tabe A. Te sk prema durng te frst two stages of gobazaton ape of te gobazaton process - Te ort remans at ts fu empoyment sk premum trougout te frst stage of gobazaton and attans ts cency sk premum durng te second stage. - Durng stage 2, bot te ort and te out ave te same sk premum and produce bot goods. - Te ort attans ts cency sk premum durng te frst stage of gobazaton at remans at ts sk premum vaue ater on. - Tere s one soe pont of tme wen te ort produces good ony and te out good ony and wc corresponds to te passage from stage to stage 3. tage 2 s tus mted to ts pont of tme. - Te ort attans ts cency sk premum durng te frst stage of gobazaton and remans at ts sk premum vaue ater on. - Tere s one perod of tme correspondng to stage 2 durng wc te ort produces good ony and te out good ony.

23 References Age, J. and undborg, P. (995), Far wages n te open economy, Economca 62: 325 35. Akerof G.A. and J.. Yeen (990), Te far wage ort ypotess and unempoyment, Quartery Journa of Economcs 05(2): 255-283. Abert, M. and Merck J. (200), Effcency wage, unempoyment and ntersectora wage dfferentas n a eckscer-on mode, German Economc Revew 2: 287-30. Anderson, E. (2005), Openness and Inequaty n Deveopng Countres: A Revew of Teory and Recent Evdence, ord Deveopment 33(7): 045 063. Berman, E., and Macn,. (2000), k-based tecnoogy transfer around te word, Oxford Revew of Economc Pocy 6(3): 2 22. Berman, E., and Macn,. (2004). Gobazaton, sk-based tecnoogca cange and abour demand, n E. ee & M. Vvare (Eds.), Understandng Gobazaton, Empoyment and Poverty Reducton (pp. 39 66). ew York: Pagrave Macman. Borjas, G.J., Freeman, R.B. and Katz.F. (992), On te abor Market Effects of Immgraton and Trade, n: G.J. Borjas, R.B. Freeman (Eds), Immgraton and te work force: Economc consequences for te Unted tates and source areas, Ccago, U. of Ccago Press, pp. 23-244. Borsook, I. (986) Earnngs abty and nternatona trade, Journa of Internatona Economcs 22, 28-295. Cartga, F. (997), Credt constrants and uman capta accumuaton n te open economy. Journa of Internatona Economcs 43: 22 236. Cecc D. and C. García-Peñaosa (2008), abour market nsttutons and ncome nequaty, Economc Pocy 23: 60 649. Cusseau., Dumont M. and eer J. (2008) Expanng Rsng Inequaty: k-based Tecnoogca Cange and ort-out Trade, Journa of Economc urveys 22(3): 409-457. Cusseau. and eer J. (fortcomng), Inequaty n Emergng Countres, fortcomng n: J. eer and. Cusseau Eds., Growng Income Inequates, Pagrave Macman. Cark A.E. and Oswad A.J. (996), atsfacton and comparson ncome, Journa of Pubc Economcs 6: 359 38. Conte, A., and Vvare, M. (2007). Gobazaton and empoyment: Imported sk based tecnoogca cange n deveopng countres, IZA Dscusson Paper o. 2797, IZA, Bonn. Davs, D.R. (998), Does European Unempoyment Prop up Amercan ages? atona abor Markets and Goba Trade, Amercan Economc Revew 88: 478-494. Deer A. and. Gaston (2008), as Gobazaton Increased Inequaty? Revew of Internatona Economcs 6(3): 56 536. Egger,. and U. Kreckemeer (2009), Frm eterogenety and te abor market ects of berazaton, Internatona Economc Revew 50: 87-26. Ecer, T.(999). Trade, deveopment and convergng growt rates: dynamc gans from trade reconsdered, Journa of Internatona Economcs 48: 79 98. Esquve, G. and Rodrguez-opez, J.A. (2003), Tecnoogy, trade, and wage nequaty n Mexco before and after AFTA, Journa of Deveopment Economcs 72: 543 565 Feenstra, R., & anson, G. (996). Gobazaton, outsourcng, and wage Inequaty, Amercan Economc Revew 86(2): 240 245. Feenstra, R., & anson, G. (997). Foregn drect nvestment and reatve wages: Evdence from Mexco s maquadoras, Journa of Internatona Economcs 42: 37 393. Fnday, R., Kerzkowsk,. (983), Internatona trade and uman capta: a smpe genera equbrum mode, Journa of Potca Economy 9: 957 978. Gaan,, and angunett, P. (2003) Te mpact of trade berazaton on wage nequaty: evdence from Argentna, Journa of Deveopment Economcs 72: 497 53. Green, F. and Dckerson, A (200), A pcture of wage nequaty and te aocaton of abor troug a perod of trade berazaton: Te case of Braz, ord Deveopment 29(): 923-939. arrs, J.R. and Todaro, M.P. (970), Mgraton, unempoyment and deveopment: A two-sector anayss, Amercan Economc Revew 60 (): 26-42. eer, J. and Cusseau,. (200) Gobazaton and te Inequaty-Unempoyment Trade-off, Revew of Internatona Economcs 5(): 028-043..

Katz.F. and Murpy K. (992), Canges n reatve wages, 963-987: suppy and demand factors, Quartery Journa of Economcs.February (992): 36-78. Krongkaew, M. (994), Income Dstrbuton n East Asan Deveopng Countres: an update, Asan- Pacfc Economc terature 8(2): 58-73. Kreckemeer, U. (2008), Unempoyment n modes of nternatona trade, n: Greenaway D, Upward R. and rgt P. Eds., Gobazaton and abour market adjustment, Pagrave, McMan, 73-96. Kreckemeer, U. and eson, D. (2006) Far wages, unempoyment and tecnoogca cange n a goba economy. Journa of Internatona Economcs 70: 45 469. Krugman, P. (994), Europe Jobess, Amerca Penness, Foregn Pocy 95 (ummer): 9-34. Krugman, P. (995), Growng ord Trade: Causes and Consequences, Brookngs Papers on Economc Actvty (Macroeconomcs) :327-377. Krugman P. (2007), Trade and Inequaty revsted, VO ste, CEPR, 5 June. Krugman P. (2008), Trade and ages, Reconsdered, Brookng Papers on Economc Actvty, (prng): 03-37. Krugman, P. and awrence, R.Z. (993), Trade, Jobs and ages, BER orkng Paper n 4478. awrence, R.Z and augter, M.J. (993), Internatona Trade and Amercan ages n te 980s: Gant uckng ound or ma ccup?, Brookng Papers on Economc Actvty: Mcroeconomcs 2: 6-226. eamer, E. (987), Pats of deveopment n te tree-factor, n-good genera equbrum mode, Journa of Potca Economy 95(5): 96 999. opez, R. and Mer,.J. (2008), Ce: Te Unbearabe Burden of Inequaty, ord Deveopment, 36(2): 2679 2695 Mesc, E. and Vvare, M. (2009), Trade and Income Inequaty n Deveopng Countres, ord Deveopment 37(2): 287 302. Manovc, B. (2005), Can we dscern te ect of ncome dstrbuton? Evdence From ouseod budget surveys ord Bank Economc Revew 9: 2 44. arayan, D., Petesc, P. Eds., (2002) Voces of te Poor from Many ands. ord Bank, asngton, DC. Pssardes, C. (997), earnng by tradng and te returns to uman capta n deveopng countres, ord Bank Economc Revew (): 7 32. Ragaya. M. Z. (2005), Income Dstrbuton n East Asan Deveopng Countres: recent trends, Asan-Pacfc Economc terature, 9(2): 36-54. Rao, V. (200), Poverty and pubc ceebratons n Rura Inda. Annas of te Amercan Academy of Potca and oca cence 573, 85 04. Ravaon, M. and oksn, M. (200), o cares about reatve deprvaton?, Journa of Economc Beavor & Organzaton 73: 7 85 enk, C. (2004), en nformaton domnates comparson: earnng from Russan subjectve pane data. Journa of Pubc Economcs 88: 2099 223. u, B. (2003). Trade berazaton, wage nequaty, and endogenousy determned non-traded goods, Journa of Internatona Economcs 60: 47 43. Zu,.C., and Trefer, D. (2005). Trade and nequaty n deveopng countres: A genera equbrum anayss. Journa of Internatona Economcs 65(): 2 24. 24