Skilled Worker Migration and Trade: Inequality and Welfare

Transcription

1 Skilled Worker igration and Trade: Inequality and Welfare Siros ougeas University of Nottingam Douglas R. Nelson Tulane University and University of Nottingam ay 011 We develo a two-sector, two-country model were trade is driven by tecnological differences. Eac country is oulated by large number of eterogeneous workers, distinguised by teir level of skills. Te countries are identical but for tecnological sueriority in te roduction of one of te goods. We sow tat, in tis model, migration of skilled workers and trade are comlements. Ten we analyze te welfare non-movers in bot countries and of migrants, sowing tat migration and trade are welfare imroving. We ten roceed to analyze te distributional effects as an inut to an analysis of te olitical economy of migration and redistribution olicy. JEL: F Keywords: Skilled Labor, igration, Welfare, Political Economy We would like to tank articiants at te ETSG (Warsaw, 008) and te ank of England-F-GEP acroeconomics and International igration Conference (London 010) for elful comments and suggestions. We also acknowledge financial suort from te Leverulme Trust under Programme Grant F/00/114/.

2 1. INTRODUCTION For all te ublic concern wit unskilled immigration, it is notable tat a sizable sare of immigration involves eole tat are skilled--relative to bot te ome and ost country labor markets (Docquier and arfouk, 006, Defoort and Rodgers, 008). Wile suc migrations ave long been a stale of te literature on te "brain drain" (e.g. agwati and Rodriguez, 1975, Grubel and Scott, 1966, Kaur and chale, 009, Kwok and Leland, 198, Commander et al., 003), te focus of researc on ost countries as rimarily been unskilled migration (Etier, 1986). Furtermore, most of tis literature as emasized single-country analysis, not te global equilibrium of suc flows. Similarly, most of tis literature tends to consider economies tat are closed to all global flows but migration (or small, oen economies wit relative rices fixed by international trade). Tis was a sensible researc strategy in an era caracterized by ig trade restrictions in Nort-Sout trade, but it seems roblematic in an increasingly globalized world were international trade constitutes a sizable sare of world roduct. In addition, as long as trade and factor movements take lace between countries wit democratically elected governments we also need to distinguis welfare outcomes from outcomes tat are olitically feasible. To address tese issues, we introduce migration of skilled workers in a two-country, twosector trade model. s in te Roy selection model (Roy, 1951), workers in bot countries are differentiated by skill and, in autarky or under trade (but witout migration), tey will sort temselves among sectors based on te sector tat offers te iger income. 1 s in te Roy- orjas model (orjas, 1987), wen resented wit te oortunity to migrate, tey will coose weter or not to migrate based on comarison of incomes net of migration costs. It is wellknown tat in models of te Heckscer-Olin-Samuelson tye trade and migration are substitutes (Etier, 1996, arkusen, 1983, undell, 1957). In line wit muc of te recent literature on migration, we will assume tat tere is a tecnology ga between countries. 3 Secifically, wile te two countries use te same tecnology in te rimary sector, one country as a more efficient tecnology in te ig-tec sector. Te two countries ave identical endowments of skills (uman caital). In te rimary sector worker roductivity is indeendent of te level of skills. Skills owever matter in te ig-tec sector. fter we derive te equilibrium under autarky for eac country, we solve for te free trade equilibrium. Next, we do not allow trade but we ermit migration. Wen migration is costless we arrive at a full integration equilibrium were everybody lives in te country tat is tecnologically suerior. We also derive te equilibrium wen migration is costly and we find 1 ougeas and Riezman (007), Davidson and atusz (006) and Davidson, atusz and Nelson (007) use te same framework to analyze a ost of trade issues, but not migration. Te Roy-orjas model as roved to be an extremely useful framework for organizing bot teoretical and emirical work on migration (e.g. elot and Hatton, 008, rücker and Defoort, 006, Grogger and Hanson, 008, Ortega and Peri, 009, Ciquiar and Hanson, 005). It is notable tat all of te aers just cited find suort for ositive selection. 3 In is arsall lecture, Lucas (1988, g. 38) noted in assing tat atterns of immigration seem inconsistent wit atterns of caital flow in standard growt model. Davis and Weinstein (00) and Iranzo and Peri (009) build models of trade and migration wit tecnological differences in wic te tecnologically suerior country not only exorts te skilled labor-intensive good, but also imorts skilled labor. uilding on te same emirical observation, Rauc (Rauc, 1991) develos an alternative model wit eterogeneous workers in wic roduction requires managerial labor as well as skilled and unskilled labor, only skilled workers ban be managers. Rauc's model also generates te observed atterns of mobility and trade.

3 tat wen only skilled workers move te ga between te autarky rices widens. Te intuition is tat as skilled workers move away from te country tat is tecnologically disadvantaged it increases te relative cost of roducing te ig tec-good and tus increases its relative rice wic in turn deteriorates te incomes of te roducers of te rimary commodity wo also are te relatively oor. igration as exactly te oosite effects on te tecnologically advanced country. Wen we allow for bot trade and migration we derive two main results. Te first is tat trade and migration are comlements. Put differently, migration is iger under trade tan under autarky. Te second result is tat welfare is iger wen bot trade and migration are allowed. ore imortantly, wen we consider searately migrants, and te citizens of eac country witout including te migrants we find tat migration and trade increase te welfare of eac of tese tree grous. Finally, we turn to te olitical economy imlications of our model. Wen citizens are allowed to vote for weter or not to allow migration and on weter or not to comensate tose wo lose as a consequence of migration we find tat te sequence of referenda in te source country and te citizensi status of migrants in te ost country can bot matter. Tere is now a sizable literature develoing on te olitical economy of migration. 4 In an imortant early contribution, enabib (1996) develoed a small, oen, one-sector economy wit eterogeneous agents (every agent ossesses one unit of labor and some non-negative quantity of caital), were migration olicy is decided by referendum. Greter, Deelo and uller (001) dro te agent eterogeneity, but consider a small, oen, -sector economy, wit a referendum mecanism. In our analysis, we model a large, -sector economy wit eterogeneous agents. In our -country world, we can follow te literature on selection, wile still retaining te tractability to resent a simle intuitive model ermitting welfare and olitical economy issues.. THE ODEL Tere are two countries and eac oulated by a continuum of workers of unit measure. Eac worker is endowed wit one unit of indivisible labor and some level of uman caital. In bot countries uman caital is uniformly distributed on te interval [0,1]. In eac country tere are two cometitive sectors. Sector X roduces a ig-tec roduct wile sector Y roduces a rimary commodity. Te roductivity of workers deends on teir level of uman caital and teir sector of emloyment. Let zx ( ) denote te roductivity of a worker wit uman caital wo works in country in sector X (oter tyes of emloyment are similarly defined). ll four tecnologies are linear, but bot sector and country matter. Te roductivity of any worker in te rimary sector, anywere in te world, is te same (v = z Y = z Y ). y contrast, worker skills () matter in te ig tec sector, and skills matter more in country tan in country. Tus: 4 uc of tis literature deals wit te way redistributive olicy interacts wit migration (e.g. azza and van Winden, 1996, Razin et al., 00, Wellisc and Walz, 1998). Tat is not te focus of our aer, and we will not comment on it any furter. Similarly, a number of aers ave used lobbying models to analyze te olitical economy of immigration (e.g. Estein and Nitzan, 006, Faccini and Willmann, 005). s our aer focuses on te referendum mecanism, we will not discuss tese aers eiter.

4 zx = z = k; k >1 X zy = zy = v; 0 < v < 1; ll workers ave identical Cobb-Douglas references given by U = X α Y β ; 0< αβ, <1; α + β =1 Since all tecnologies are constant returns to scale and all markets are cometitive, worker incomes equal teir roductivities and eac workers send a fraction α of er income on roduct X. a. utarky Trougout we use good X as te numeraire and let and denote te autarky rices in countries and resectively. In wat follows we are going to concentrate our analysis on country since by setting k =1 we can obtain te corresonding solutions for country. Workers coose teir sector of emloyment by comaring wages. worker wit uman caital will receive income k if emloyed in sector X wile te same worker will receive income v if emloyed in sector Y. Tis imlies tat all workers wit uman caital iger v tan will be emloyed in te ig-tec sector wile workers wit uman caital below k tis tresold will be emloyed in te rimary sector. Given our secification of references te equilibrium autarky rice is roortional to te ratio of aggregate roduction in sector X divided by te aggregate roduction in sector Y : 1 k d β = α v Te corresonding closed form solution for te autarky rice is given by: k β = v α + β Notice tat te above solution imlies tat > wic follows from te fact tat country as a suerior tecnology for roducing te ig-tec roduct. It also follows tat =. --Figure 1 about ere-- Figure 1 sows te derivation of te roduction frontier from te roduction functions and te labor constraint. Te SW quadrant sows te labor constraint common to bot countries and te SE quadrant sows te linear roduction function tat transforms labor into good Y (also common to bot countries). Te NW quadrant is a bit trickier. s in Figure 1, workers are arranged suc tat te worker wit te igest uman caital, and tus igest roductivity in sector X, is at te origin, and uman caital dereases monotonically as one moves out along te L X axis. s a result, outut increases at a decreasing rate, wic we reflect wit a te roduction

5 functions in te NW quadrant. Furtermore, country s roduction function lies above country s at every oint excet te origin. Using tis information, we can trace out te two roduction frontiers in te usual fasion. Finally, note tat te autarky equilibrium is caracterized by bot countries allocating te same sare of L to te roduction of good Y, so te equilibrium outut of Y is te same in bot countries. Given te tecnological difference between countries, altoug bot countries allocate te same amount of L (and uman K) to X roduction, country roduces more X in autarky. Finally, it sould be clear tat te relative rice of good Y in country must be greater tan tat in country. --Figure about ere-- Wile Figure 1 as te virtue of illustrating te autarky equilibrium in a standard fasion, it will rove convenient for our uroses to use a different graical aaratus. Tus, consider te autarky equilibrium as illustrated in Figure. Te orizontal axis sows te (uniform) distribution of skill from 0 to 1. Since bot countries sare te same distribution of skill, bot can be sown on te same gra. Te left vertical axis sows value marginal roduct in Y. Since te marginal ysical roduct of labor in Y is assumed to be constant across all workers, and te same in bot countries (i.e. v), te value marginal roduct in eac country will be determined by te rice of Y. ecause X is te numeraire, te rigt vertical axis sows te roduct of eac worker emloyed in te X sector. ny worker wose roductivity in X exceeds tat in Y will coose to work in X, wit te cutoff at (were te V X curve intersects te V Y curve). Our assumtion tat k > 1 imlies tat te value marginal roduct curve for X in country lies above tat in country. Finally, we see bot > and = (wic we simly denote in te diagram). b. Trade From Figure 1, we know tat country as te lower autarky relative rice of rimary commodity Y. Tus, wen te two economies trade, we exect country to exort good Y and T country to exort te ig-tec commodity X. Te world rice will lie between te two autarky rices and be given by β = α 1 1 d + T T T T T ( + ) v k d T T T v T were = v> = > =. Te corresonding closed form solution for te world k rice is given by: T k β = v α + β --Figure 3 about ere-- ecause bot countries sare te same omotetic utility function, wen trade equalizes

6 commodity rices, te relative quantities of te two goods consumed will be te same. Tat is, tey will lie on a common ray from te origin assing between te two autarky consumtion oints in Figure 1 and teir sares of eac good in national income will also be te same (given by te arameters α and β). Te national income lines, wit sloes given by te common relative rice, will suort consumtion levels not available under autarky, so trade is mutually beneficial. We can use Figure 3, wic extends Figure, to illustrate te equilibrium wit trade. Free trade equalizes rices, so bot countries face T v, wic roduces new, and internationally different, cutoff values identifying te marginal skilled worker in eac of te two countries. T Tus, we see country secializing in te roduction of good X, as to workers sift from Y T roduction to X roduction. Similarly, in country, to workers sift from X roduction to Y roduction. Te income distribution effects are also straigtforward. For country, workers wit skill greater tan exerience an increase in real income greater tan te cange in relative T rice, wile workers wit skill less tan exerience a fall in income roortionally equal to T te fall in te rice of good Y (country s imortable). Workers wit skill between and, by switcing sectors and alying teir uman caital, exerience a smaller fall in income, or even gain (toug not by as muc as re-cange workers in X). Tus, te income sare of re-cange X workers rises and tat of re-cange Y workers falls. Te case of country is exactly te reverse. c. igration Suose tat te two economies do not trade but migration is allowed. Costless igration Given tat te two countries use te same tecnology to roduce te rimary commodity but country as a suerior tecnology for roducing te ig-tec roduct every worker in country will emigrate in country. Te integration equilibrium rice is equal to, i.e. country 's autarky rice. Costly migration Now suose tat migration entails a cost γ measured in numeraire units. We are going to assume tat tis cost is sufficiently ig so tat workers in te rimary sector do not wis to emigrate. 5 Let and denote te two new autarky rices. Te utility of a worker wo is a citizen of country, emloyed in sector X and does not emigrate is equal to 5 Wen we introduce trade, tese workers will not ave an incentive to emigrate. Tus, by restricting teir movement in te absence of trade allows for clearer comarisons between te two cases.

7 α β β ( ) ( ) = ( ). 6 If te same worker emigrates to country er new utility will be α k γ β β ( k γ) ( ) =( k γ )( ). Ten tere exists a tresold level of uman caital = k β γ ( ) β ( ) ( ) β suc tat all workers wit uman caital above it and wo initially were citizens in country emigrate to country. Letting and denote te new tresold levels of uman caital tat searates tose emloyed in sector X from tose emloyed in sector Y, in countries and resectively, te new autarky rices are given by te system of equations: and = β α = β α v d 1 1 ( k γ ) d + k d Notice tat te two equations need to be solved simultaneously since deends on bot new autarky rices. Tis comlication imlies tat it is not ossible to derive closed-form solutions for te two rices. Neverteless, te effect of migration is to increase te ga between te autarky rices, i.e. > > >. Te intuition is tat in country te old tresold level of uman caital tat searates tose emloyed in sector X from tose emloyed in sector Y is too ig, given tat only ig ability workers ave migrated wile te corresonding tresold in country is too low. 7 v --Figure 4 about ere-- Figure 4 sows diagrammatically te costly migration equilibrium. Wile te rincile is straigtforward, our graical aaratus gets messier ere. Since all workers wit uman caital above migrate from to, we truncate te distribution at and aend te µ skilled workers to te distribution. 8 It is easy to see tat, wit uncaged references and no international trade, te migration of workers tat roduce only good X results in too little X roduction in country and too muc in country. Tis results in an increase in te relative rice of X in country and a reduction country (i.e. a fall in and a rise in ) and, of course, an increase in te allocation of labor to X roduction in and to Y roduction in. 6 α β ll utilities ave been divided by α β. 7 Let µ denote te roortion of migrants. Given tat bot oulations originally were of unit measure te new oulation of country is of measure 1 µ and uman caital is uniformly distributed on te interval [0,1 µ ] 8 It sould be clear now wy we refer te graical aaratus alied in Figures -5 to tat in Figure 1..

8 3. IGRTION ND TRDE a. Costless igration Wen migration is costless it is otimal tat te wole roduction of te ig-tec roduct takes lace in country. In te absence of trade costs, te roduction location of te rimary roduct is inconsequential. In te resence of trade costs it is otimal tat all workers move to country FI. In eiter case we ave full integration. Te full integration equilibrium rice is given as te solution of FI v β = α 1 FI FI FI k d FI were =. It is clear tat te equilibrium world trade rice and te critical tresold are k te same as tose for te autarky case for country. v b. Costly igration Once again suose tat migration entails a cost γ measured in numeraire units. Given tat wen te two countries trade all workers face te same rice, te only workers tat move across borders are some of tose workers wo were citizens of country and emloyed in sector X and now move to country and are emloyed in te same sector. For tese workers teir remigration income was equal to wile teir ost-migration income is equal to k γ. Tis γ imlies tat tere exists a tresold level of uman caital suc tat all workers wit k 1 uman caital above it and wo initially were citizens in country emigrate to country. 9 Letting and denote te new tresold levels of uman caital tat searates tose emloyed in sector X from tose emloyed in sector Y, te new equilibrium rice is given by β = α Te corresonding closed form solution is: Given tat k > 1 1 d + ( k γ ) d + k d = k v ( + ) v k γ β 1+ k α + β k γ we ave >. In addition, te inequality γ < k 1 imlies tat 1+ k 9 Notice tat ere it is sufficient to comare incomes because te rices in te two countries are equal. Earlier we examined te case were tere is no trade and terefore te rices in te two countries were different.

9 > T wic in turn imlies tat > T and > T. Tis is because migration as allowed a more efficient allocation of resources by moving ig skilled workers to country wic as a suerior tecnology for roducing te skill-intensive roduct X. Comaring te tresold levels of uman caital tat searate tose wo emigrate from tose wo do not for te cases of migration witout trade and migration wit trade we get te following result Proosition 1 igration and trade are comlements. Proof Te roosition follows from te inequality > wic ensures tat <. Figure 5 illustrates te equilibrium wit trade and migration. --Figure 5 about ere-- Tere is more migration under free trade and te skilled migration to te skill abundant country suorts increased comarative advantage based trade. Te intuition is tat, under trade, rices converge wic furter boosts te incomes of te migrants and tus te incentive to migrate. Of course tis is not a new result. arkusen (1983) and Razin and Sadka (1994) ave, witin more traditional trade modes, demonstrated te comlementarity between trade and migration wen countries differ in tecnologies. 4. WELFRE We measure grou welfare as te sum of utilities of its individual members. We are focusing on tree grous: Te citizens of country (excluding migrants), te migrants and te citizens of country excluding migrants. elow we demonstrate tat migration increases te welfare of all tree grous and tus global welfare. 10 a. Citizens of Country (Excluding igrants) Tose workers wo were emloyed in te rimary sector during te re-migration eriod and are still emloyed in te same sector during te ost-migration eriod are better off in te later eriod. Tis is because migration increases te rice of te rimary commodity and ence teir income. For exactly te same reason tose workers tat are emloyed in te ig-tec sector during bot eriods exerience a loss in utility. In addition, te inequality > T imlies tat tere are some workers wo before migration were emloyed in te ig-tec sector and after migration tey moved to te rimary sector. mong tese workers some are better off and some are worse off. ore secifically, tere exists a tresold level of uman caital,, suc tat all 10 ll utilities ave been divided by α β α β. In addition, by we denote te equilibrium rice were for notational simlicity we ave droed te suerscrit wen confusion can be avoided..

10 workers wit uman caital below tat level are better off after migration wile all tose workers wit uman caital above tat level are worse off after migration. Lemma 1 igration increases te welfare of te citizens of country. Proof Te ost migration utility of an agent working in sector X is equal to α β β ( k) ( k ) = k and te corresonding utility of an agent working in sector Y is equal to α β α ( v) v = v Ten, te ost-migration welfare,, of te citizens of country is given by W v 1 k α β + 0 v k 1+ α β 1 β 1 W = vd k d = v v = + k = k k 1 v 1 (1 ) = + α k α +. k Next consider te effect on welfare of a small cange in te migration cost γ. dw v α (1 α) k d sign = sign (1 α) α dγ + k dγ Substituting for and rearranging we get dw 1+ k d sign = sign 1 dγ k γ dγ d Given tat γ < k 1 and <0 a decrease in migration costs will increase welfare. dγ b. igrants Lemma igration increases te welfare of te migrants. Proof Te re-migration utility of a migrant is equal to T β ( ) wile te ost-migration utility of a migrant is equal to β ( ) k T Substituting for and we find tat migration increases te welfare of migrants if and only if

11 1 ( 1+ k ) ( k γ ) α 1 α k >1 s γ 0 te left-and side of te above exression aroaces ( ) 1 α 1 1+ k k 1 α + α Te above exression is increasing in k and is equal to 1 for k =1 wic, given tat k >1, comletes te roof. c. Citizens of Country (Excluding igrants) For exactly te same reasons as tose analyzed for country, tose workers wo were emloyed in te rimary sector during te re-migration eriod and are still emloyed in te same sector during te ost-migration eriod are better off in te later eriod. Once more, te inequality > T imlies tat tere are some workers wo before migration were emloyed in te ig-tec sector and after migration tey moved to te rimary sector. mong tese workers some are better off and some are worse off. gain, tere exists a tresold level of uman caital,, suc tat all workers wit uman caital below tat level are better off after migration wile all tose workers wit uman caital above tat level are worse off after migration. Lemma 3 igration increases te welfare of te citizens of country excluding te migrants. Proof Te ost migration utility of an agent working in sector X is equal to α β β ( ) = and te corresonding utility of an agent working in sector Y is equal to α β α ( v) v = v Ten, te ost-migration welfare, W, of te citizens of country is given by γ v α k 1 β W = 0 vd + v d = 1+ α 1 β γ 1 β = v + v = k 1 = 1 1+ α γ v + ( 1) (1 α ) k Next consider te effect on welfare of a small cange in te migration cost γ. dw α (1 α) kγ d γ sign = sign (1 α) v ( α) 1 α dγ + + ( k 1) dγ ( k 1)

12 Notice tat a cange in migration costs will affect te number of migrants and te last term catures te cange in welfare as a result of te cange in te size of te grou. We know tat a decrease in migration costs will increase te number of migrants and we know tat teir welfare will increase. Here, we need to concentrate only on te first term. Given tat γ γ < k 1, <1 and te roof is comleted by following te same stes as tose used in k 1 te roof of roosition 1. d. Global Welfare Te following roosition is a direct imlication of te tree lemmas. Proosition igration increases global welfare. 5. POLITICL ECONOY Te results of te revious section suggest tat migration increases welfare in bot countries even wen migrants (wo definitely gain) are ignored. Tis imlies tat te gains of tose wo benefit are more tan sufficient to cover te losses of tose workers wo are worse off as a result of migration. However, for countries wit democratically elected governments were majority voting decides key olicy issues eole migt vote against migration and even if tey decide to allow it losers migt not be comensated. In tis section, we identify te conditions under wic eac of tese outcomes materializes.we begin by examining te ossible outcomes of a referendum on migration ignoring, for te moment, any comensation olicies. a. Referendum on igration Consider te voting outcome in country. Te roortion of eole wo would vote in favor of migration, ϕ, is equal to γ ϕ = + 1 k 1 were te second term catures te votes of te migrants. Te median voter's re-migration T utility is equal to β ( ) wile te same voter's ost-migration utility is ( ) α v; keeing in mind tat in te re-migration eriod is emloyed in sector X and ten moves to sector Y. Equating te two utilities, solving for te median voter's uman caital and substituting in te above equality, we get α T k γ γ ϕ = v γ + k 1 Next consider te voting outcome in country. Te roortion of eole wo would vote in

13 ϕ favor of migration,, is equal to ϕ = = k Te roosition below follows from te inequality ϕ < ϕ and te fact tat bot countries ave to vote in favor of migration. Proosition 3 In te absence of any comensation olicy migration will take lace if and only if ϕ > 0.5. lso, notice tat a decline in migration costs increases te likeliood tat te voting outcome will favor migration. b. Referenda on igration and Comensation Suose tat citizens in eac country ave te oortunity to vote on two referenda. In one referendum, as in te revious section, tey vote on weter or not to allow migration (emigration in country and immigration in country ). In te oter referendum tey vote on weter or not tere will be comensation for tose wo exerience a loss in income as a result of migration. We demonstrate tat outcomes would deend on weter or not migrants are allowed to vote on te comensation referendum in te ost country ( ) and on te sequence of referenda in te source country ( ). We begin wit country. We assume tat if te comensation referendum takes lace first te comensation sceme can include a tax on migrants wose revenues can be used for comensating te losers. Te roosition below identifies conditions suc tat te sequence of referenda matters. Proosition 4: Suose tat γ γ + 1 > >. Ten country will vote for k 1 k 1 comensation if and only if te migration referendum takes lace first.. Proof Te first inequality states tat te winners togeter wit te migrants are a majority grou. Te second inequality states tat te roortion of losers is iger tan te roortion of winners. Suose tat te comensation referendum takes lace first. Te migrants do not ave an incentive to vote for comensation given tat togeter wit te winner will vote in favor of migration. In contrast, wen te migration referendum takes lace first tere will be an unaminous vote in favor of migration. Te winners are still better off and te losers know tat tey can get comensation after te next referendum. Now consider te referenda in country. Proosition 5 Suose tat γ + 1 >1 >. Ten country will vote for k 1 comensation if and only if migrants are not allowed to vote on te comensation referendum. Proof Clearly, if migrants are allowed to vote on te comensation referendum tey would vote against comensation. Te first inequality states tat te winners togeter wit te migrants are a majority grou. Te second inequality states tat te roortion of losers is

14 iger tan te roortion of winners. Notice tat if migrants were allowed to vote and given tat tey would vote against comensation tere would be a vote against migration given tat te losers (in te absence of migrant vore) are a majority grou. 6. CONCLUSION We ave resented a two-country model of trade and skilled labor migration. Trade and migration are induced by differences in tecnologies and tus are comlements. Cross-border movements of workers as significant welfare imlications for bot source and ost countries. In our model, were only skilled workers ave an incetive to migrate tose workers wo are emloyed in skill-intensive sectors (oter tan te migrants) suffer welfare losses. Neverteless, we ave sown tat, in bot countries, te welfare gains of tose workers wo benfit from migration exceed (in absolute terms) te losses of of tose workers wo suffer losses. However, we ave also sown tat in countries were bot migration and any comensation olicies are decided by majority vote, tere is no guarantee tat neiter te loses will be comensated sould migration take lace nor tere will be a vote in favor of migration. Tis framework can be easily extended to account for unskilled worker migration by aving te two countries using different tecnologies for te roduction of te rimary commodity. For tat matter a general version can allow for migration of bot tyes of workers wic can be eiter in te same or in te oosite direction. 11 REFERENCES elot. V. and Hatton T. J. (008), 'Immigrant Selection in te OECD', CEPR Working Paer no (Centre for Economic Policy Researc: London). enabib J. (1996), 'On te olitical economy of immigration', Euroean Economic Review, 40, 9, agwati J. and Rodriguez C. (1975), 'Welfare-Teoretical nalyses of te rain Drain', Journal of Develoment Economics,, orjas G. J. (1987), 'Self-Selection and te Earnings of Immigrants', merican Economic Review, 77, 4, ougeas S. and Riezman R. (007), 'Trade and te Distribution of Human Caital', Journal of International Economics, 73, rücker H. and Defoort C. (006), 'Te (Self) Selection of International igrants Reconsidered', IZ Discussion Paer no. 05 (IZ: onn). Ciquiar D. and Hanson G. H. (005), 'International igration, Self-Selection, and te Distribution of Wages: Evidence from exico and te United States', Journal of Political Economy, 113,, Commander S., Kangasniemi. and Winters L.. (003), 'Te rain Drain: Curse or oon? Survey of te Literature', in R. E. aldwin and L.. Winters (eds.), Callenges to 11 For examle, Isikawa (1996) offers a model of trade and migration, were te two countries differ in teir endowments, and were bot tyes of workers move in te same direction.

15 Globalization: nalyzing te Economics, (University of Cicago Press: Cicago), Davidson C. and atusz S. (006), 'Trade liberalization and comensation', International Economic Review, 47, 3, Davidson C., atusz S. and Nelson D. R. (007), 'Can comensation save free trade?', Journal of International Economics, 71, 1, Davis D. R. and Weinstein D. E. (00), 'Tecnological Sueriority and te Losses from igration', NER Working Paers no (National ureau of Economic Researc: oston). Defoort C. and Rodgers G. (008), 'Long-term Trends in International igration: n nalysis of te Six ain Receiving Countries', Poulation, 63,, Docquier F. and arfouk. (006), 'International migration by educational attainment ( )', in C. Ozden and. Sciff (eds.), International igration, Remittances and Develoment, (Palgrave acmillan: New York). Estein G. S. and Nitzan S. (006), 'Te struggle over migration olicy', Journal of Poulation Economics, 19, 4, Etier W. J. (1986), 'International Trade Teory and International igration', in O. Stark (ed.), igration, Human Caital and Develoment, (JI Press: Greenwic), Etier W. J. (1996), 'Teories about Trade Liberalisation and igration: Substitutes or Comlements?', in P. J. Lloyd and L. Williams (eds.), International Trade and igration in te PEC Region, (Oxford University Press: elbourne), Faccini G. and Willmann G. (005), 'Te olitical economy of international factor mobility', Journal of International Economics, 67, 1, Greter J.-., de elo J. and uller T. (001), 'Te olitical economy of migration in a Ricardo-Viner model"', in S. Djajic (ed.), International igration: Trends, Policy, Imact, (Routledge: London), Grogger J. and Hanson G. H. (008), 'Income aximization and te Selection and Sorting of International igrants', NER Working Paer no (National ureau of Economic Researc: oston). Grubel H.. and Scott. D. (1966), 'International Flow of Human Caital', merican Economic Review, 56,, Iranzo S. and Peri G. (009), 'igration and Trade: Teory wit an alication to te Eastern Western Euroean integration', Journal of International Economics, 79, 1, Isikawa J. (1996), 'Scale Economies in Factor Sulies, International Trade, and igration', Canadian Journal of Economics, 9, Kaur D. and chale J. (009), 'International igration and te World Income Distribution', Journal of International Develoment, 1, 8, Kwok V. and Leland H. (198), 'n Economic-odel of te rain-drain', merican Economic Review, 7, 1, Lucas R. E. (1988), 'On te ecanics of Economic Develoment', Journal of onetary Economics,, 1, 3-4. arkusen J. R. (1983), 'Factor ovements and Commodity Trade as Comlements', Journal of International Economics, 14, azza I. and van Winden F. (1996), ' Political Economic nalysis of Labor igration and Income Redistribution', Public Coice, 88, 3-4, undell R.. (1957), 'International Trade and Factor obility', merican Economic Review,

16 47, 3, Ortega F. and Peri G. (009), 'Te Causes and effects of International igrations: Evidence from OECD Countries ', Human Develoment Researc Paer no. 009/06 (United Nations Develoment Programme: New York). Rauc J. E. (1991), 'Reconciling te Pattern of Trade wit te Pattern of igration', merican Economic Review, 81, Razin. and Sadka E. (1994), 'International igration and International Trade', in. R. Rosenzweig and O. Stark (eds.), Handbook of Poulation and Family Economics, Vol. 1, Nort-Holland Publisers, (Nort-Holland: msterdam). Razin., Sadka E. and Swagel P. (00), 'Tax urden and igration: Political Economy Teory and Evidence', Journal of Public Economics, 85,, Roy. D. (1951), 'Some Tougts on te Distribution of Earnings', Oxford Economic Paers, 3,, Wellisc D. and Walz U. (1998), 'Wy do ric countries refer free trade over free migration? Te role of te modern welfare state', Euroean Economic Review, 4, 8,

17 Figure 1: utarky Equilibrium I

18 Figure : utarky Equilibrium II

19 Figure 3: International Trade witout igration

20 Figure 4: igration witout International Trade

21 Figure 5: International Trade and igration