《国际政治经济学》文献资料（Political Economy of Migration）A Political Economy of International Migration, 1815-1914

The Manchester School Vol LX No.4 December 1992 0025-2034$2.50359-376 A POLITICAL ECONOMY OF INTERNATIONAL MIGRATION,1815-1914* by JAMES FOREMAN-PECK St.Antony's College,Oxford "Give me...your huddled masses,yearning to breathe free" I INTRODUCTION Why was Emma Lazarus's sentiment,engraved on the base of the Statue of Liberty,so much more widely embodied in policy in the nineteenth century than in recent decades?During the earlier period,the enormous inter- continental migrations that peopled Siberia,the mid-west of the United States and other territories often laid the foundations for international reconfigurations of economic and political power.These population move- ments certainly established ethnic mixes that have greatly influenced sub- sequent domestic policies,from Fiji to South Africa.Despite the magnitude of the changes,the nineteenth-century pattern of intercontinental migration left open to Europeans their principal destinations:North America,Australasia and Latin America.Equally notable was the segregated stream of Asian migrants to mines and plantations in Malaya,the Dutch East Indies,Ceylon, Mauritius and the Caribbean. The years between the World Wars marked a new phase of forced migration in Europe as a consequence of the redrawing of frontiers.Immi- gration restrictions in destinations of European migration,abandonment of indentured labour and substantial net return migration during the 1930s further distinguished this period from the nineteenth century.After 1945, north western Europe hosted immigrants from southern Europe and from current and former colonial possessions in Africa,Asia and the Caribbean. The U.S.increasingly drew in migrants,legal and illegal,from the south, especially from Mexico.All higher-income countries after two decades began to conclude that unskilled immigrants were undesirable and imposed tight restrictions.But highly-trained professionals in scarcity fields remained eli- gible,at least for temporary permits:the "brain drain"flowed faster. The present paper investigates how much a simple economic theory of .Manuscript received 27.2.91;final version received 4.12.91. tMy thoughts on this subject have been clarified by anonymous referees and by seminar and conference participants at Leeds,Oxford,Liverpool,Manchester and Santander.The Hallsworth Fellowship in Political Economy at the University of Manchester allowed me to complete the project. 359

The Manchester School Vol LX No. 4 December 1992 0025-2034 $2.50 359-376 A POLITICAL ECONOMY OF INTERNATIONAL MIGRATION, 1815-1914* by JAMES FOREMAN-PECKT St. Antony’s College, Oxford “Give me.. . your huddled masses, yearning to breathe free” I INTRODUCTION Why was Emma Lazarus’s sentiment, engraved on the base of the Statue of Liberty, so much more widely embodied in policy in the nineteenth century than in recent decades? During the earlier period, the enormous inter￾continental migrations that peopled Siberia, the mid-west of the United States and other territories often laid the foundations for international reconfigurations of economic and political power. These population move￾ments certainly established ethnic mixes that have greatly influenced sub￾sequent domestic policies, from Fiji to South Africa. Despite the magnitude of the changes, the nineteenth-century pattern of intercontinental migration left open to Europeans their principal destinations: North America, Australasia and Latin America. Equally notable was the segregated stream of Asian migrants to mines and plantations in Malaya, the Dutch East Indies, Ceylon, Mauritius and the Caribbean. The years between the World Wars marked a new phase of forced migration in Europe as a consequence of the redrawing of frontiers. Immi￾gration restrictions in destinations of European migration, abandonment of indentured labour and substantial net return migration during the 1930s further distinguished this period from the nineteenth century. After 1945, north western Europe hosted immigrants from southern Europe and from current and former colonial possessions in Africa, Asia and the Caribbean. The U.S. increasingly drew in migrants, legal and illegal, from the south, especially from Mexico. All higher-income countries after two decades began to conclude that unskilled immigrants were undesirable and imposed tight restrictions. But highly-trained professionals in scarcity fields remained eli￾gible, at least for temporary permits: the “brain drain” flowed faster. The present paper investigates how much a simple economic theory of * Manuscript received 27.2.91; final version received 4.12.91. t My thoughts on this subject have been clarified by anonymous referees and by seminar and conference participants at Leeds, Oxford, Liverpool, Manchester and Santander. The HaHsworth Fellowship in Political Economy at the University of Manchester allowed me to complete the project. 359

360 The Manchester School policy can contribute to an understanding of the distinctive nineteenth- century international migration policy.At first sight,the explanation is twofold.First,European immigration was not considered a problem for state policy in host countries,or perhaps was to be encouraged,because the railway and the steamship augmented the effective supply of land so that there was plenty of space.Second,racism was responsible for the restriction of Asian migrants to tropical areas,where Europeans had no interest in settling. If this is correct,a theoretical approach to nineteenth-century international migration policy is unnecessary.However the simple explanations fail to account for the variety of migration policies,do not address Arthur Lewis's (1978,p.192)economic analysis of discrimination against Asian migration to the regions of recent European settlement,and are silent about why even host country electors who were comparatively well-off should have accepted lower incomes by allowing in migrant labour.(Williamson,1974,p.249,calculated U.S.real annual earnings in Eastern industries were 11 per cent lower than they would have been in 1910 if no immigration had been permitted.) A simple model cannot hope to explain all policies but it may provide a classification of cases and an account of why the nineteenth-century world economy was so open to labour movements.Section II outlines a political economy of international migration based upon factor income maximization, factor endowments,income distribution and constitutions,Section III shows how the model fits a large component of nineteenth-century experience and Section Iv focuses on the,at first sight,anomalous policy of the most important destination for European immigration,the United States. II POLITICAL ECONOMY The two key questions of any political economy of international migration are(1)who gains and who loses from migration?and(2)who is in a position to do something about it?The answer to the first question identifies a demand for policies and that to the second reveals a policy supply.The model assumes,at first,economies closed to external trade in goods and self- interested persons who receive their income primarily from one factor.Then the demand for policies will be aimed at preventing the reduction of returns to the owner's factor by increases in supply of that factor.Factor owners prefer those policies that expand the supply of other factors.A pro-labour state aims to protect and augment labour income,which might require the encourage- ment of emigration and discouraging of immigration. The simplest case to analyse is where migrant and domestic labour are perfect substitutes and the policy objective is only to maximize domestic wages (analogously with the insider-outsider behaviour of the neoclassical monopoly union).Then,in Fig.1,the expansion path as the demand for labour moves outward from L is L shaped.When L<L*,there is domestic unemployment,which is likely to deter immigration and encourage emigra-

360 The Manchester School policy can contribute to an understanding of the distinctive nineteenth￾century international migration policy. At first sight, the explanation is twofold. First, European immigration was not considered a problem for state policy in host countries, or perhaps was to be encouraged, because the railway and the steamship augmented the effective supply of land so that there was plenty of space. Second, racism was responsible for the restriction of Asian migrants to tropical areas, where Europeans had no interest in settling. If this is correct, a theoretical approach to nineteenth-century international migration policy is unnecessary. However the simple explanations fail to account for the variety of migration policies, do not address Arthur Lewis’s (1978, p. 192) economic analysis of discrimination against Asian migration to the regions of recent European settlement, and are silent about why even host country electors who were comparatively well-off should have accepted lower incomes by allowing in migrant labour. (Williamson, 1974, p. 249, calculated U.S. real annual earnings in Eastern industries were 11 per cent lower than they would have been in 1910 if no immigration had been permitted.) A simple model cannot hope to explain all policies but it may provide a classification of cases and an account of why the nineteenth-century world economy was so open to labour movements. Section I1 outlines a political economy of international migration based upon factor income maximization, factor endowments, income distribution and constitutions, Section I11 shows how the model fits a large component of nineteenth-century experience and Section IV focuses on the, at first sight, anomalous policy of the most important destination for European immigration, the United States. I1 POLITICAL ECONOMY The two key questions of any political economy of international migration are (1) who gains and who loses from migration? and (2) who is in a position to do something about it? The answer to the first question identifies a demand for policies and that to the second reveals a policy supply. The model assumes, at first, economies closed to external trade in goods and self￾interested persons who receive their income primarily from one factor. Then the demand for policies will be aimed at preventing the reduction of returns to the owner’s factor by increases in supply of that factor. Factor owners prefer those policies that expand the supply of other factors. A pro-labour state aims to protect and augment labour income, which might require the encourage￾ment of emigration and discouraging of immigration. The simplest case to analyse is where migrant and domestic labour are perfect substitutes and the policy objective is only to maximize domestic wages (analogously with the insider-outsider behaviour of the neoclassical monopoly union). Then, in Fig. 1, the expansion path as the demand for labour moves outward from L* is L shaped. When L < L*, there is domestic unemployment, which is likely to deter immigration and encourage emigra-

A Political Economy of International Migration,1815-1914 361 Wages(w)↑ Labour↑ V(w.r) Compromise 一声 L Land Production Function 务 Rent(r) FIG.1 Biases in Immigration Policy tion policies,but the wage is determined by market forces.Once labour demand is strong enough for L>L,immigration controls are introduced so that all of the increased demand is channelled into higher wages and none into increased employment(for immigrants)(Labour in Fig.1).Where other factor owners influence immigration policy,the expansion path will allow for some immigration,for it raises the returns to other factors (Compromise in Fig.1).An entirely landowner-dominated policy would aim to encourage immigration sufficiently to prevent any rise in wages(Land in Fig.1). The supply of policies then comes from governments in effect choosing the numbers of migrants.They do so by maximizing an objective function,V, with factor prices as arguments,which will best satisfy the government's principals.The precise form of function then depends upon the distribution of wealth,factor endowments and the (de facto)constitution.If all land and capital were distributed equally and the government was elected by universal adult suffrage,then the objective function would depend only upon national factor endowments.A labour-abundant country in these circumstances would place the greatest weight upon wages,whereas electors in a land- abundant country under the same conditions would favour emphasizing rents. The more common historical experience is that the bulk of land and capital is owned by a few.Under a democratic franchise the objective function then only includes wages,regardless of national endowments,because the majority have no income from land or from capital.A fully democratic franchise in the nineteenth century,given difficulties of registration and limited powers of elected institutions,was rather rare.In practice,even constitutions with universal adult suffrage were prone to be swayed more by the wealthy and,therefore,non-wage incomes could be of interest to the government.Many constitutions were still effectively controlled by a rela-

A Political Economy of International Migration, 181 5-1 91 4 I Production Function 361 Rent (r) FIG. 1 Biases in Immigration Policy tion policies, but the wage is determined by market forces. Once labour demand is strong enough for L > L*, immigration controls are introduced so that all of the increased demand is channelled into higher wages and none into increased employment (for immigrants) (Labour in Fig. 1). Where other factor owners influence immigration policy, the expansion path will allow for some immigration, for it raises the returns to other factors (Compromise in Fig. 1). An entirely landowner-dominated policy would aim to encourage immigration sufficiently to prevent any rise in wages (Land in Fig. 1). The supply of policies then comes from governments in effect choosing the numbers of migrants. They do so by maximizing an objective function, V, with factor prices as arguments, which will best satisfy the government’s principals. The precise form of function then depends upon the distribution of wealth, factor endowments and the (de facto) constitution. If all land and capital were distributed equally and the government was elected by universal adult suffrage, then the objective function would depend only upon national factor endowments. A labour-abundant country in these circumstances would place the greatest weight upon wages, whereas electors in a land￾abundant country under the same conditions would favour emphasizing rents. The more common historical experience is that the bulk of land and capital is owned by a few. Under a democratic franchise the objective function then only includes wages, regardless of national endowments, because the majority have no income from land or from capital. A fully democratic franchise in the nineteenth century, given difficulties of registration and limited powers of elected institutions, was rather rare. In practice, even constitutions with universal adult suffrage were prone to be swayed more by the wealthy and, therefore, non-wage incomes could be of interest to the government. Many constitutions were still effectively controlled by a rela-

362 The Manchester School tively small number of landholders who owned most of the wealth.In these territories the objective function included only rents,regardless of endow- ments. The constraint on the objective function is the productive capacity of the economy,determined by technology,institutions and endowments.A second constraint is that domestic labour(L)is fully employed.Factors are assumed to be paid their marginal products.Where L2 are migrants,K is capital,T is land,O is output,wi and r are(real prices)respectively of domestic labour and land,and x is the rental of capital: max y=aw1+Bx+(1-a-B)r (1) s.t. Q=f(L1,L2,T;K) (2) f1L2+fxL2+(1-&-)frL2-f2=0 (3) The equilibrium policy is attained when the marginal product of immi- grant labour is a weighted average of the effects of an increment of immigrant labour on domestic factor prices.Where domestic and immigrant labour are (q into p)substitutes,f>0.Other things being equal,the larger is a(the more pro-domestic labour is immigration policy),the larger are the positive equilibrium marginal effects of migrant workers on other factor prices.The immigrant labour target stock is smaller,to hold up these positive marginal effects,when a is large.Where immigrants are complements to domestic labour,larger a's encourage policies that favour more migrants. Where land ownership and political control are highly concentrated(B is large),an increase in the scarcity of labour relative to land creates both a demand,and a supply from the landowner-controlled government,for poli- cies to encourage immigration,either by direct subsidy to immigrant travel to plantations and latifundia or through administering or recognizing indenture contracts.A neutral state policy,by contrast,sets the rules to maximize a weighted average of all factor prices,the weights reflecting the current factor distribution of income,rather than interests of owners of particular factors. The resulting migrant worker stock is that of the "free market",the operations of which policy is intended to facilitate by the provision of defence,law and order and the elimination of harmful externalities,such as communicable diseases."Free market"economies do not exist in cultural and institutional vacuums;even their migration stances will differ with assimilation and franchise policy,for these influence when a member of L2 transfers to Lt,and therefore affect the size of the Lz target. From(2)and the assumption that factors are paid their marginal products,(1)may be rewritten, V=或1+附x+(1一a-阴fr (3)is then obtained by differentiating V-@with respect to L2

362 The Manchester School tively small number of landholders who owned most of the wealth. In these territories the objective function included only rents, regardless of endow￾ments. The constraint on the objective function is the productive capacity of the economy, determined by technology, institutions and endowments. A second constraint is that domestic labour (L,) is fully employed. Factors are assumed to be paid their marginal products. Where L2 are migrants, K is capital, T is land, Q is output, w1 and r are (real prices) respectively of domestic labour and land, and n is the rental of capital: s.t. Q = fW1, ~52, T K) 4.lLZ + BfXL.2 +u -.-P)fTL2-L.2 = 0 (2) (3) The equilibrium policy is attained when the marginal product of immi￾grant labour is a weighted average of the effects of an increment of immigrant labour on domestic factor prices.’ Where domestic and immigrant labour are (q into p) substitutes, fLlL2 > 0. Other things being equal, the larger is a (the more pro-domestic labour is immigration policy), the larger are the positive equilibrium marginal effects of migrant workers on other factor prices. The immigrant labour target stock is smaller, to hold up these positive marginal effects, when o! is large. Where immigrants are complements to domestic labour, larger a’s encourage policies that favour more migrants. Where land ownership and political control are highly concentrated (j is large), an increase in the scarcity of labour relative to land creates both a demand, and a supply from the landowner-controlled government, for poli￾cies to encourage immigration, either by direct subsidy to immigrant travel to plantations and latifundia or through administering or recognizing indenture contracts. A neutral state policy, by contrast, sets the rules to maximize a weighted average of all factor prices, the weights reflecting the current factor distribution of income, rather than interests of owners of particular factors. The resulting migrant worker stock is that of the “free market”, the operations of which policy is intended to facilitate by the provision of defence, law and order and the elimination of harmful externalities, such as communicable diseases. “Free market” economies do not exist in cultural and institutional vacuums; even their migration stances will differ with assimilation and franchise policy, for these influence when a member of Lz transfers to L1, and therefore affect the size of the L2 target. ‘From (2) and the assumption that factors are paid their marginal products, (1) may be rewritten, = afLl + PfK -.-fi)fP (3) is then obtained by differentiating V-Q with respect to L2

A Political Economy of International Migration,1815-1914 363 TABLE 1 INTERNATIONAL MIGRATION REGIMES:A FIRST APPROXIMATION Pro-Labour Pro-Land/Capital Empire Source Country policy encourage discourage encourage Destination Country policy discourage encourage encourage The supply of policies may be subject to external constraints,such as treaty obligations.Nineteenth-century empires introduce another constraint and perhaps policy motivation,if not type of state.In at least one respect,a concern with security,the empire is analogous to the neutral or coalition state.But because the empire spans continents or partly self-governing boundaries,migration policy may appear different.Perhaps more important is that the geographical range of empire places a greater policy emphasis on security and sometimes on the related goal of "nation building"(ensuring cultural homogeneity).In principle,empires could be any of the three previous types of state,neutral/coalition,labour or land/capital-dominated. Typically land/capital empires can be expected to have been more concerned to control emigration so as to prevent the escape of conscripts and sub- versives. For international migration to take place not only must economic conditions be favourable but accommodating policies must be adopted in both source and destination territories.The analogy between immigration policy and slavery and serfdom is clearest when both source and host state are run by landowners.The source state will never encourage migration and will generally discourage it,to hold down the cost of labour.Potential source countries must be run either by labour or neutral ruling groups or the land- owning rulers must be less than fully effective if there is to be much inter- national migration (Table 1). Relations between states could be developed by agreements that allowed a closer meshing of their desired policies:by labour conditions treaties and by the regulation of indenture contracts.Indenture contracts facilitated the migration of labour unable to afford or to borrow fares.The employer advanced the fares in return for an agreement to work for a specified period at a specified wage.2 As transport costs fell,the period of the contract also should have shortened,if the contracts were a solution to a market imperfection. Treaties between nations to regulate labour conditions could have prevented immigrant labour threatening native labour and made sense for pro-labour The practical difficulty was that enforcement of the indenture contract was too close to a new form of slavery for many people's liking

A Political Economy of International Migration, 1815-1 91 4 363 TABLE 1 INTERNATIONAL MIGRATION REGIMES: A FIRST APPROXIMATION Fro-Labour Pro-LandlCapital Empire Source Country pohcy encourage discourage encourage Destination Country policy discourage encourage encourage The supply of policies may be subject to external constraints, such as treaty obligations. Nineteenth-century empires introduce another constraint and perhaps policy motivation, if not type of state. In at least one respect, a concern with security, the empire is analogous to the neutral or coalition state. But because the empire spans continents or partly self-governing boundaries, migration policy may appear different. Perhaps more important is that the geographical range of empire places a greater policy emphasis on security and sometimes on the related goal of “nation building” (ensuring cultural homogeneity). In principle, empires could be any of the three previous types of state, neutral/coalition, labour or land/capital-dominated. Typically landlcapital empires can be expected to have been more concerned to control emigration so as to prevent the escape of conscripts and sub￾versives. For international migration to take place not only must economic conditions be favourable but accommodating policies must be adopted in both source and destination territories. The analogy between immigration policy and slavery and serfdom is clearest when both source and host state are run by landowners. The source state will never encourage migration and will generally discourage it, to hold down the cost of labour. Potential source countries must be run either by labour or neutral ruling groups or the land￾owning rulers must be less than fully effective if there is to be much inter￾national migration (Table 1). Relations between states could be developed by agreements that allowed a closer meshing of their desired policies: by labour conditions treaties and by the regulation of indenture contracts. Indenture contracts facilitated the migration of labour unable to afford or to borrow fares. The employer advanced the fares in return for an agreement to work for a specified period at a specified wage.2 As transport costs fell, the period of the contract also should have shortened, if the contracts were a solution to a market imperfection. Treaties between nations to regulate labour conditions could have prevented immigrant labour threatening native labour and made sense for pro-labour 2The practical difficulty was that enforcement of the indenture contract was too close to a new form of slavery for many people’s liking

364 The Manchester School governments,where there was a flow in both directions.When the flow was in one direction,the treaty would have been acceptable to a labour host and a land-or capital-dominated source.In both cases,the treaties were a less hostile alternative to migration restrictions. Discrimination between migrants by origin is optimal for a labour- influenced destination state when the reservation wage varies by origin.High- wage labour is much less of a threat to the host labour force than low.Closely connected is the skill or human capital issue.It may become worth a nation's while poaching the investment that has been made by other countries,by admitting the highly-trained.The formal model captures this characteristic by introducing a third category of labour,L3,skilled immigrants,fL3>0, L<0.For administrative reasons,in earlier periods admission rules by origin might be simpler than identifying relevant human capital. With sufficient discrimination within host labour markets,immigrant labour could be transformed from substitutes to complements to host workers.An increase in unskilled labour (through immigration)might enhance the marginal productivity of,and therefore the demand for,skilled labour.A labour-dominated host government or a trade union organization which can discriminate has an incentive to do so if immigrants can be admitted only to those occupations which enhance the marginal productivity of other labour and which are not politically influential enough to block the policy.A variation on this theme is the crowding of immigrants into occupa- tions without job security so that in a recession immigrants can be "dumped" back in their countries of origin,thereby enhancing the job security of native, enfranchised labour(Ethier,1985). Surprisingly,the introduction of trade in goods does not affect the predictions of the model.In an autarkic economy,implicitly assumed in the discussion based on slavery and serfdom,and in the migration analyses of Williamson (1974)and Greenwood and McDowell (1986),trade does not equalize factor prices,and therefore(intra-marginal)migrants motivated by economic gain will be more productive in the host country than in the source (governed by absolute advantage).However,most migrations in the nine- teenth century were between areas that participated in substantial inter- national trade.In such circumstances,movement of goods and factors are substitutes.With the free international movement of goods there is a strong tendency for one price to emerge on world markets,as Harley (1980)has graphically shown for the nineteenth-century wheat trade.If these goods were made with the same technology,then factor productivity and rewards would have converged in trading countries.With similar technologies available in the trading countries,and incomplete specialization,more migration is likely to reduce comparative advantage and,therefore,exports of the migration host countries'labour-scarce goods.In turn,this has consequences for wages and rents in host and source countries.In the limit,when tariffs and transport costs are negligible,factor price equalization takes place,once allowance has

364 The Manchester School governments, where there was a flow in both directions. When the flow was in one direction, the treaty would have been acceptable to a labour host and a land- or capital-dominated source. In both cases, the treaties were a less hostile alternative to migration restrictions. Discrimination between migrants by origin is optimal for a labour￾influenced destination state when the reservation wage varies by origin. High￾wage labour is much less of a threat to the host labour force than low. Closely connected is the skill or human capital issue. It may become worth a nation’s while poaching the investment that has been made by other countries, by admitting the highly-trained. The formal model captures this characteristic by introducing a third category of labour, Lf, skilled immigrants, fLlL3 > 0, jLIL1 < 0. For administrative reasons, in earlier periods admission rules by origin might be simpler than identifying relevant human capital. With sufficient discrimination within host labour markets, immigrant labour could be transformed from substitutes to complements to host workers. An increase in unskilled labour (through immigration) might enhance the marginal productivity of, and therefore the demand for, skilled labour. A labour-dominated host government or a trade union organization which can discriminate has an incentive to do so if immigrants can be admitted only to those occupations which enhance the marginal productivity of other labour and which are not politically influential enough to block the policy. A variation on this theme is the crowding of immigrants into occupa￾tions without job security so that in a recession immigrants can be “dumped” back in their countries of origin, thereby enhancing the job security of native, enfranchised Iabour (Ethier, 1985). Surprisingly, the introduction of trade in goods does not affect the predictions of the model. In an autarkic economy, implicitly assumed in the discussion based on slavery and serfdom, and in the migration analyses of Williamson (1974) and Greenwood and McDowell (1986), trade does not equalize factor prices, and therefore (intra-marginal) migrants motivated by economic gain will be more productive in the host country than in the source (governed by absolute advantage). However, most migrations in the nine￾teenth century were between areas that participated in substantial inter￾national trade. In such circumstances, movement of goods and factors are substitutes. With the free international movement of goods there is a strong tendency for one price to emerge on world markets, as Harley (1980) has graphically shown for the nineteenth-century wheat trade. If these goods were made with the same technology, then factor productivity and rewards would have converged in trading countries. With similar technologies available in the trading countries, and incomplete specialization, more migration is likely to reduce comparative advantage and, therefore, exports of the migration host countries’ labour-scarce goods. In turn, this has consequences for wages and rents in host and source countries. In the limit, when tariffs and transport costs are negligible, factor price equalization takes place, once allowance has

A Political Economy of International Migration,1815-1914 365 been made for non-pecuniary advantages.Immigration is then unnecessary to enhance incomes and immigration controls alone are ineffective in raising or protecting labour income.Higher tariffs would have to accompany such measures.Equally,trade restrictions would have to offset falling freight and passenger transport costs merely to maintain a wage differential(and there- fore an incentive to migrate).3 In the absence of capital market imperfections,migration takes place so long as the difference between the capitalized value of future expected earnings in the source and destination countries are greater than the costs of moving between the two labour markets.Wages in migrant source and destination countries even without trade should differ by the travel time, money and emotional costs of migration. In an open world economy the effects of migration on factor prices will be stronger than the impact of movements of goods.+Migration can be a once-and-for-all movement whereas goods must be continually traded for factor prices to be influenced.Even,say,a 5 per cent difference between similar products in two countries,induced by tariff and transport costs,in turn raising wages in the host economy 5 per cent above those of the source, would have justified the spending of perhaps up to 50 per cent of a young migrant's annual wage on the move.Much less than that was necessary to cross the Atlantic by the end of the nineteenth century.The results of the simple(no trade)model,therefore,are applicable for the trade in goods case, within the limits of wage variation allowed by transport costs and trade restrictions. III EXPLAINING POLICY The behaviour of Latin America and the tropical dependencies,land- abundant with landowner-dominated governments,corresponds with the model predictions of encouraging immigration.The more representatively governed land-abundant dominions and Boer republics were less enthusi- astic but were constrained,one way or the other,by the international regime. Generally labour-abundant Europe exhibited a spectrum of policies from West to East,varying with the degree of democratization and distribution of income.European policies ranged from low-cost minimal encouragement of 3The imposition of the MeKinley and Dingley U.S.tariffs in 1890 and 1897,nominally intended to protect U.S.workers from low-wage European competition,may have received an impetus from falling transport costs.Such tariffs merely provided a stronger incentive for immigration to lower U.S.wages,yet paradoxically (if the tariff justification is believed), European immigration was left unrestricted. Information flows between source and destination territories were very effective and migration was often financed by remittances from the host economy.See,for example,Baines(1985, pp.27-31).By 1875,the Atlantic could be crossed in less than 10 days.Consequently,wage arbitrage was quite efficient

A Political Economy of International Migration, 1815-1 91 4 365 been made for non-pecuniary advantages. Immigration is then unnecessary to enhance incomes and immigration controls alone are ineffective in raising or protecting labour income. Higher tariffs would have to accompany such measures. Equally, trade restrictions would have to offset falling freight and passenger transport costs merely to maintain a wage differentiai (and there￾fore an incentive to migrate).3 In the absence of capital market imperfections, migration takes place so long as the difference between the capitalized value of future expected earnings in the source and destination countries are greater than the costs of moving between the two labour markets. Wages in migrant source and destination countries even without trade should differ by the travel time, money and emotional costs of migration. In an open world economy the effects of migration on factor prices will be stronger than the impact of movements of goods4 Migration can be a once-and-for-all movement whereas goods must be continually traded for factor prices to be influenced. Even, say, a 5 per cent difference between similar products in two countries, induced by tariff and transport costs, in turn raising wages in the host economy 5 per cent above those of the source, would have justified the spending of perhaps up to 50 per cent of a young migrant’s annual wage on the move. Much less than that was necessary to cross the Atlantic by the end of the nineteenth century. The results of the simple (no trade) model, therefore, are applicable for the trade in goods case, within the limits of wage variation allowed by transport costs and trade restrictions. 111 EXPLAINING POLICY The behaviour of Latin America and the tropical dependencies, land￾abundant with landowner-dominated governments, corresponds with the model predictions of encouraging immigration. The more representatively governed land-abundant dominions and Boer republics were less enthusi￾astic but were constrained, one way or the other, by the international regime. Generally labour-abundant Europe exhibited a spectrum of policies from West to East, varying with the degree of democratization and distribution of income. European policies ranged from low-cost minimal encouragement of 3The imposition of the McKinley and Dingley U.S. tariffs in 1890 and 1897, nominally intended to protect U.S. workers from low-wage European competition, may have received an impetus from falling transport costs. Such tariffs merely provided a stronger incentive for immigration to lower US. wages, yet paradoxically (if the tariff justification is believed), European immigration was left unrestricted. 41nformation flows between source and destination territories were very effective and migration was often financed by remittances from the host economy. See, for example, Baines (1985, pp. 27-31). By 1875, the Atlantic could be crossed in less than 10 days. Consequently, wage arbitrage was quite efficient

366 The Manchester School emigration in Britain to bureaucratic obstruction in Russia.Host policy on non-white migration to the land-abundant regions was discriminatory by the end of the century as a means of maintaining host wages. In Brazil and Argentina,land-owning oligarchies recognized that sub- sidies to labour could permanently expand the labour supply by the once- and-for-all payment of a transport grant.Their subsidy policies of the 1880s were prompted by the opening up of the interior,which enhanced the effective supply of land.Almost at the same time that the U.S.was increasingly pressing for restrictions on immigration,between 1885 and 1913,fl1 million was spent on immigration subsidies by the Sao Paolo and federal govern- ments(Leff,1982).Only in the 1880s did Argentine military and political organization become adequate to defeat the Indians and drive them off their land (di Tella,1985).When that expropriation was releasing new land for commercial exploitation,Argentina even experimented with providing free passages to immigrants during the 1880s(Martinez and Lewandowski,1911, pp.118-121).Most Russian emigration was illegal,especially since many young men wished to avoid conscription(British Parliamentary Papers,1903, p.464).Collective responsibility through the mir for repayment of the emanci- pation loan was another great deterrent to peasant migration.Authoritarian empires like Russia were also interested in ridding themselves of minorities, such as Jews,to promote national security,or "nation building". With the abolition of slavery within the British Empire between 1834 and 1838,the indenture contract,regulated by the(British)Indian govern- ment,became an increasingly popular means of attracting Asian plantation labour.Between 1842 and 1870,Mauritius took about 350,000 indentured workers (B.P.P.,1910).Half a million indentured migrants,80 per cent from India,went to the British West Indies between 1834 and 1918. Unable to obtain a legally-enforceable claim on the extra income gener- ated by migration,capital markets were unwilling to advance loans for migration,thereby restricting labour movements to workers wealthy enough to finance their own movement.By enforcing an obligation of labour to work for the landowner who had paid the passage money,the law ensured that such advances were forthcoming.More low-income labour could move and earn higher wages when-indenture contracts were enforced.5 The land-abundant dominions were constrained in their immigration policy by membership of the British Empire.But other influences operated as well.Despite somewhat similar endowments,Canada pursued policies less favourable to immigration than Argentina because of a different income distribution and franchise(Solberg,1985).Australia,with a fairly representat- sWith the abolition of slavery in Jamaica,for example,former slaves abandoned plantation work and became self-sufficient peasant proprietors in the formerly unoccupied central highlands (Green,1984).Other emancipations reinforced labour shortage for landlords throughout the nineteenth century:France,1848,Russia,1861,the Netherlands,1863,the United States,1863-5 and Brazil,1888

366 The Manchester School emigration in Britain to bureaucratic obstruction in Russia. Host policy on non-white migration to the land-abundant regions was discriminatory by the end of the century as a means of maintaining host wages. In Brazil and Argentina, land-owning oligarchies recognized that sub￾sidies to labour could permanently expand the labour supply by the once￾and-for-all payment of a transport grant. Their subsidy policies of the 1880s were prompted by the opening up of the interior, which enhanced the effective supply of land. Almost at the same time that the U.S. was increasingly pressing for restrictions on immigration, between 1885 and 1913, €11 million was spent on immigration subsidies by the Sgo Paolo and federal govern￾ments (Leff, 1982). Only in the 1880s did Argentine military and political organization become adequate to defeat the Indians and drive them off their land (di Tella, 1985). When that expropriation was releasing new land for commercial exploitation, Argentina even experimented with providing free passages to immigrants during the 1880s (Martinez and Lewandowski, 191 1, pp. 118-121). Most Russian emigration was illegal, especially since many young men wished to avoid conscription (British Parliamentary Papers, 1903, p. 464). Collective responsibility through the mir for repayment of she emanci￾pation loan was another great deterrent to peasant migration. Authoritarian empires like Russia were also interested in ridding themselves of minorities, such as Jews, to promote national security, or “nation building”. With the abolition of slavery within the British Empire between 1834 and 1838, the indenture contract, regulated by the (British) Indian govern￾ment, became an increasingly popular means of attracting Asian plantation labour. Between 1842 and 1870, Mauritius took about 350,000 indentured workers (B.P.P., 1910). Half a million indentured migrants, 80 per cent from India, went to the British West Indies between 1834 and 1918. Unable to obtain a legally-enforceable claim on the extra income gener￾ated by migration, capital markets were unwilling to advance loans for migration, thereby restricting labour movements to workers wealthy enough to finance their own movement. By enforcing an obligation of labour to work for the landowner who had paid the passage money, the law ensured that such advances were forthcoming. More low-income labour could move and earn higher wages whenindenture contracts were enforced.’ The land-abundant dominions were constrained in their immigration policy by membership of the British Empire. But other influences operated as well. Despite somewhat similar endowments, Canada pursued policies less favourable to immigration than Argentina because of a different income distribution and franchise (Solberg, 1985). Australia, with a fairly representat- ’With the abolition of slavery in Jamaica, for example, former slaves abandoned plantation work and became self-sufficient peasant proprietors in the formerly unoccupied central highlands (Green, 1984). Other emancipations reinforced labour shortage for landlords throughout the nineteenth century: France, 1848, Russia, 1861, the Netherlands, 1863, the United States, 1863-5 and Brazil. 1888

A Political Economy of International Migration,1815-1914 367 ive government,did subsidize white immigration but aimed to discriminate in the payment of the fare subsidy in favour of immigrants who would obtain rural employment(Condliffe,1930,p.32).Rural labour was complementary to the majority of Australians who lived in towns and held secondary or tertiary sector jobs.Moreover,the subsidy varied with the pace of expansion of the Australian economy.Fifty per cent of arrivals in the 1870s were assisted but only 10 per cent in the 1890s(Kelley,1965).In 1878-9,New South Wales labour temporarily lobbied effectively against supported immigration (Malchow,1979,p.262). The Boer republics thought their immigration policy was less con- strained than that of the dominions.After the 1886 discovery of gold in Witwatersrand,immigrant gold miners raised the marginal product of dom- estic labour and their exploitation by the state dynamite monopoly and exorbitant railway rates provided additional benefits.However,Transvaalers saw immigrant miners as a threat to their way of life,while the migrants objected to the denial of their right to vote.British intervention on behalf of its "uitlander"subjects culminated in the Boer War (Houghton,1971, pp.13-15;Thompson,1971). Europe was mainly a source of international migration,the countries of origin being typically labour-abundant and governed more representatively than Russia.British Parliamentary Select Committees of 1826 and 1827, greatly exercised by the consequences for English wages of Irish immigration, pronounced the U.K.overpopulated and advocated emigration to the col- onies.Parliament regularly made grants in aid of emigration and,in 1834, allowed parishes to mortgage their rates to subsidize emigration as a form of Poor Relief (Johnson,1913).In 1840,the Colonial Land and Emigration Department was established with responsibility for colonial land sales, encouraging emigration and providing information.The British response to Russian and Polish immigration after 1882,the 1905 Act,is consistent with the model predictions for a democratic government in a labour-abundant economy.On three occasions during the 1890s the British Trade Union Congress passed motions objecting to immigration (Holmes,1982,p.181). By contrast,the French willingness to absorb migrants may be related to the more equal distribution of land,thanks to the Revolution,and a low natural rate of population increase,further retarded by the impact of the Franco-Prussian War (Williamson and Lindert,1980b).France was in the unusual position of being a European net destination for migrants.Despite about fifty legislative proposals between 1883 and 1914 to restrict or tax immigrants,the only substantial restriction was the limitation in 1899 of the proportion employable on public works(Green,1985).Elsewhere in Europe, emigration was a matter of concern by the end of the century but policy measures were ineffective.Even in Belgium,emigration strictly was not merely a matter of individual decision:an Emigration Commission was the sole judge of whether emigration was permissible.The Swiss went as far as to

A Political Economy of International Migration, 1815-1914 367 ive government, did subsidize white immigration but aimed to discriminate in the payment of the fare subsidy in favour of immigrants who would obtain rural employment (Condliffe, 1930, p. 32). Rural labour was complementary to the majority of Australians who lived in towns and held secondary or tertiary sector jobs. Moreover, the subsidy varied with the pace of expansion of the Australian economy. Fifty per cent of arrivals in the 1870s were assisted but only 10 per cent in the 1890s (Kelley, 1965). In 1878-9, New South Wales labour temporarily lobbied effectively against supported immigration (Malchow, 1979, p. 262). The Boer republics thought their immigration policy was less con￾strained than that of the dominions. After the 1886 discovery of gold in Witwatersrand, immigrant gold miners raised the marginal product of dom￾estic labour and their exploitation by the state dynamite monopoly and exorbitant railway rates provided additional benefits. However, Transvaalers saw immigrant miners as a threat to their way of life, while the migrants objected to the denial of their right to vote. British intervention on behalf of its “uitlander” subjects culminated in the Boer War (Houghton, 1971, pp. 13-15; Thompson, 1971). Europe was mainly a source of international migration, the countries of origin being typically labour-abundant and governed more representatively than Russia. British Parliamentary Select Committees of 1826 and 1827, greatly exercised by the consequences for English wages of Irish immigration, pronounced the U.K. overpopulated and advocated emigration to the col￾onies. Parliament regularly made grants in aid of emigration and, in 1834, allowed parishes to mortgage their rates to subsidize emigration as a form of Poor Relief (Johnson, 1913). In 1840, the Colonial Land and Emigration Department was established with responsibility for colonial land sales, encouraging emigration and providing information. The British response to Russian and Polish immigration after 1882, the 1905 Act, is consistent with the model predictions for a democratic government in a labour-abundant economy. On three occasions during the 1890s the British Trade Union Congress passed motions objecting to immigration (Holmes, 1982, p. 181). By contrast, the French willingness to absorb migrants may be related to the more equal distribution of land, thanks to the Revolution, and a low natural rate of population increase, further retarded by the impact of the Franco-Prussian War (Williamson and Lindert, 1980b). France was in the unusual position of being a European net destination for migrants. Despite about fifty legislative proposals between 1883 and 1914 to restrict or tax immigrants, the only substantial restriction was the limitation in 1899 of the proportion employable on public works (Green, 1985). Elsewhere in Europe, emigration was a matter of concern by the end of the century but policy measures were ineffective. Even in Belgium, emigration strictly was not merely a matter of individual decision: an Emigration Commission was the sole judge of whether emigration was permissible. The Swiss went as far as to

368 The Manchester School create the unenforceable offence of urging a man to leave his native land (Johnson,1913,p.336. If pure racism had been the sole reason for anti-Asian immigrant legislation in temperate zones,Asians would have been excluded altogether from the dominions rather than subject to special taxes or literacy tests.True, the United States was unusually exclusive.Following an inquiry in 1876, which falsely represented Chinese immigration to California as contract labour,Chinese workers were largely excluded in 1882(Cloud and Galenson, 1987;McClain,1990).But the United States attempted to keep out contract labour from all sources.The Chinese were charged a special poll tax in Australia,New Zealand and British Columbia(B.P.P.,1903,p.828).Literacy tests were also introduced at the end of the nineteenth century by Natal,and later by Australia,to limit non-European migration.Australian attempts to restrict specifically Indian immigration were vetoed by the London govern- ment on the grounds that India was an Empire country (Bevan,1986,p.67) Both dominion special taxes and literacy tests and the objections to contract labour in the U.S.which motivated Chinese exclusion there,are entirely consistent with the labour objective of maintaining wages at levels deter- mined by the productivity of temperate zone agriculture,as Lewis (1978, p.192)suggested.In the model of Section II,two immigrant labour-supply functions must be added to account for this discrimination.The South African 1911 Mines and Works Act similarly is more consistent with a policy of "insider"wage maximization rather than with pure racism.African immi- gration was not at issue but African employment threatened white wages.The act,therefore,restricted employment of African labour in skilled and semi- skilled jobs. IV THE AMERICAN ANOMALY Some questions remain over the policy of the principal destination of Euro- pean migration,the United States.The U.K.reacted politically to a very small volume of immigration yet the U.S.maintained a virtual open house to Europeans for the entire period despite political agitation for controls.Had land abundance been the sole reason we would expect the closing of the frontier around 1890 to mark the beginning of the restriction period.The slumps of 1893 or 1907 would have precipitated legislation. Nor does the explanation advanced for French policy seem so promising for the U.S..Though the Homestead Act's offer of 160 acres of iand to western settlers might have been expected to encourage a more equal distribution of income,industrializing tendencies predominated.By the end of our period, the United States's distribution was considerably more unequal than France's and only a little less unequal than Britain's. American policy towards migration was as limited as British at the mid- century;the first legislation was concerned only with public health.New York

368 The Manchester School create the unenforceable offence of urging a man to leave his native land (Johnson, 1913, p. 336). If pure racism had been the sole reason for anti-Asian immigrant legislation in temperate zones, Asians would have been excluded altogether from the dominions rather than subject to special taxes or literacy tests. True, the United States was unusually exclusive. Following an inquiry in 1876, which falsely represented Chinese immigration to California as contract labour, Chinese workers were largely excluded in 1882 (Cloud and Galenson, 1987; McClain, 1990). But the United States attempted to keep out contract labour from all sources. The Chinese were charged a special poll tax in Australia, New Zealand and British Columbia (B.P.P., 1903, p. 828). Literacy tests were also introduced at the end of the nineteenth century by Natal, and later by Australia, to limit non-European migration. Australian attempts to restrict specifically Indian immigration were vetoed by the London govern￾ment on the grounds that India was an Empire country (Bevan, 1986, p. 67). Both dominion special taxes and literacy tests and the objections to contract labour in the U.S. which motivated Chinese exclusion there, are entirely consistent with the labour objective of maintaining wages at levels deter￾mined by the productivity of temperate zone agriculture, as Lewis (1978, p. 192) suggested. In the model of Section 11, two immigrant labour-supply functions must be added to account for this discrimination. The South African 191 1 Mines and Works Act similarly is more consistent with a policy of “insider” wage maximization rather than with pure racism. African immi￾gration was not at issue but African employment threatened white wages. The act, therefore, restricted employment of African labour in skilled and semi￾skilled jobs. IV THE AMERICAN ANOMALY Some questions remain over the policy of the principal destination of Euro￾pean migration, the United States. The U.K. reacted politically to a very small volume of immigration yet the U.S. maintained a virtual open house to Europeans for the entire period despite political agitation for controls. Had land abundance been the sole reason we would expect the closing of the frontier around 1890 to mark the beginning of the restriction period. The slumps of 1893 or 1907 would have precipitated legislation. Nor does the explanation advanced for French policy seem so promising for the U.S.. Though the Homestead Act’s offer of 160 acres of land to western settlers might have been expected to encourage a more equal distribution of income, industrializing tendencies predominated. By the end of our period, the United States’s distribution was considerably more unequal than France’s and only a little less unequal than Britain’s. American policy towards migration was as limited as British at the mid￾century; the first legislation was concerned only with public health. New York