Migration transitions and the migration hump A theoretical inquiry and empirical test

1.  Migration transitions and the migration hump  A theoretical inquiry and empirical test   Hein de Haas hein.dehaas@qeh.ox.ac.uk www.imi.ox.ac.uk

2. The core issue Will development in poor countries lead to less migration ? The relation between development and net out-migration is non-linear. Human development leads to increased emigration in the short to medium term. Development is associated with generally higher levels of mobility

3. Background Concerns on (regular and irregular) migration of (low) skilled migrants from poor countries to Europe and North America Generally perceived as a problem in need of control (in contrast to high skilled migration). Policy response: Restrictive policies and increasing border controls However, given the sustained labour market demand, migration has not stopped, but rather changed its face Paradox: Increased tendency towards permanent settlement Reliance of de facto labor migrants on other legal migration channels Increasing irregular character of migration Diversifying migration methods and migration routes

7. Frontier of the Spanish enclave of Melilla with Morocco in 2006, the day after many African migrants tried to scale the fence � six were killed. Frontier of Melilla (Spanish Enclave in Morocco). About April 2006. The army guards the wall, after an attempt at mass crossing by Africans desperate to get into the European Union.Frontier of the Spanish enclave of Melilla with Morocco in 2006, the day after many African migrants tried to scale the fence � six were killed. Frontier of Melilla (Spanish Enclave in Morocco). About April 2006. The army guards the wall, after an attempt at mass crossing by Africans desperate to get into the European Union.

9. Received Knowledge on African Migration to Europe Image of �desperate migrants� fleeing African misery Poverty, war and climate change as the main cause Alarmist discourse: Europe and North Africa risk being �inundated� by �waves� of sub-Saharan immigrants (aquatic metaphors) Challenge all these assumptions Challenge all these assumptions

10. Doomsday scenarios African hordes are �on their way�

11. Shaky empirical foundations Media and policy discourses tends to be alarmist and anecdotal Exaggeration of the actual magnitude; apocalyptic representations are highly misleading Lack of historical contextualisation of the phenomenon (e.g., how �new� is this migration?) Lack of theoretical undertanding of the phenomenon (is poverty the main cause of migration?)

12. Smart solutions? The argument: The only way to reduce migration from poor countries is to address its �root causes� by promoting development. �if you don�t help the third world . . . then you will have these poor people in our society� (Prime Minister Rasmussen of Denmark, 1995) �Broad-based and rapid development will induce potential poverty migrants to stay at home� (IOM and UNCTAD, 1996). �The problem of immigration, the dramatic consequences of which we are witnessing, can only be addressed effectively [. . .] through an ambitious and coordinated development [plan] to fight its root causes� (Jos� Manuel Barrosso, 2005)

13. Smart solutions? Perceived policy instruments Aid (�Marshall plan for Africa�) Trade Circular migration as a �win-win-win� strategy (remittances, co-d�veloppement) Return migration and �Diaspora involvement�

14. Practical concerns Aid Limited scope and effectiveness of aid. Limited credibility of �aid instead of migration� policies Trade Protectionist policies are often inconsistent with development aims. Return migration and remittances as a medicine against migration? Enforcing return? (�Safeguards which prevent overstaying and ensure return are key elements� � EC Council Nov 07) Are temporary migrants better �development actors�? Reinvention of guest workers ? Back to the future (Castles) Chicken is to egg as development is to migration (Agunias) One can question the very credibility of certain �aid instead of migration� policies.Their stated development intentions often seem to camouflage a hidden agenda of forcibly returning irregular immigrants or rejected asylum seekers, after providing them some modest financial assistance or rapid and often ineffective professional training (Weil, 2002).Notwithstanding their official references to development objectives, receiving countries have not so far shown any serious commitment to a genuine policy that addresses the linkages between migration and development beyond the narrow perspective of stimulating the return of unwanted migrants. Moreover, the resources and funds allocated to such programmes are generally too modest to warrant any expectations as regards significantly diminishing immigration and increasing return migration, let alone �bringing development� and addressing the root causes of migration (cf. Olesen, 2002). For instance, is the dissemination of a brochure on business opportunities in Senegal or information campaigns aimed at �revers[ing] the exodus of the Senegalese� and convincing emigrants to return and stay at home (Diatta and Bbow, 1999: 251), really likely to have any significant effect without altering the structural conditions which make people want to leave the country? The new Dutch government policy on migration and development has been proposed without allocating any additional budget for this purpose (DGIS, 2004). One can question the very credibility of certain �aid instead of migration� policies.Their stated development intentions often seem to camouflage a hidden agenda of forcibly returning irregular immigrants or rejected asylum seekers, after providing them some modest financial assistance or rapid and often ineffective professional training (Weil, 2002).Notwithstanding their official references to development objectives, receiving countries have not so far shown any serious commitment to a genuine policy that addresses the linkages between migration and development beyond the narrow perspective of stimulating the return of unwanted migrants. Moreover, the resources and funds allocated to such programmes are generally too modest to warrant any expectations as regards significantly diminishing immigration and increasing return migration, let alone �bringing development� and addressing the root causes of migration (cf. Olesen, 2002). For instance, is the dissemination of a brochure on business opportunities in Senegal or information campaigns aimed at �revers[ing] the exodus of the Senegalese� and convincing emigrants to return and stay at home (Diatta and Bbow, 1999: 251), really likely to have any significant effect without altering the structural conditions which make people want to leave the country? The new Dutch government policy on migration and development has been proposed without allocating any additional budget for this purpose (DGIS, 2004).

15. Fundamental concerns Underlying assumption �Development will lead to less migration� Migration as development failure Primary cause of migration is poverty and human misery Migration is primarily a function of income gaps Theoretical foundations: neo-classical, place utility, gravity and �push-pull� models Development and economic growth (whether through aid, trade, remittances, or otherwise) is a substitute to migration More Challenge all these assumptions More Challenge all these assumptions

16. Fundamental concerns Sedentary assumptions are fundamentally flawed (immobility as natural state; mobility as an anomaly). Non-randomness of real-world migration patterns Limited understanding of the relation between development and migration: the process of capitalist development is conditional on transfer of labour from rural to urban sectors. Failure to see migration as integral part of �development� rather than as a problem to be solved. Casual empirical observations: Micro: most migrants are not amongst the poorest Macro: most non-OECD migrants in OECD countries from middle income countries

17.  Origins of African migrants in OECD countries  Leading emigration countries are typically not among the least developed This alone is no proof of causality � for instance, one might also argue that this reflects proximity rather than levels of socio-economic development To further determine the correlation between levels of socio-economic development and migration, one needs to study this into a multivariate settingLeading emigration countries are typically not among the least developed This alone is no proof of causality � for instance, one might also argue that this reflects proximity rather than levels of socio-economic development To further determine the correlation between levels of socio-economic development and migration, one needs to study this into a multivariate setting

18. �Non-linear� migration theories Non-linear migration theories conceptualise migration as constituent part of more broader processes of economic and demographic change and restructuring Spatio-temporal models (long term): Migration systems theory (Mabogunje 1970) mobility transition (Zelinsky 1971), further elaborated by Skeldon (1997); Hatton and Williamson (1998; 2002); Draw links between (parallel) demographic and economic transitions on the one hand and migration trends on the other Migration hump (short term) (Martin and Taylor 1996) in wake of trade and other structural political economic reforms Migration is a constituent part of broader development processes and economic restructuring, because of the accompanying disruptions and factor movements between economic sectors and from rural to urban areas (Massey, 1991; Weintraub and D�iaz-Briquets, 1994) ? this notion is not new Important however, to distinguish short term from long term effects. In an attempt to mitigate exaggerated hopes of �trade instead of migration�, Martin and Taylor (1996) put forward several compelling arguments why trade and migration can be complementary in the short to medium run, beyond the fact the economic development materially enables more people to migrate abroad. Higher productivity and efficiency, technological advantages, and economies of scale in the North may harm the competitiveness of the South even in the production of labour-intensive goods. Under such circumstances, trade liberalization can lead to concentrations of highly productive economic activities in the North along with more immigration of labourers to support them. Furthermore, adjustment to new market and policy environments is never instantaneous. While the negative impacts of trade liberalization (on protected sectors) are often immediate, the expansion of production in sectors potentially favoured by trade reforms always takes time. There may be a long lag between investment and the creation of new jobs, which seems to be a recipe for a migration hump in the wake of trade reforms (Martin and Taylor, 1996: 52). It is not entirely clear whether this pertains to income gaps based on absolute income or income adjusted to purchasing power. Instead of using relative income gaps, some studies from the mid-1990s suggested an average real per capita income threshold of around US$ 4,000 at which migration transitions would occur (Stalker, 2002: 171). Migration is a constituent part of broader development processes and economic restructuring, because of the accompanying disruptions and factor movements between economic sectors and from rural to urban areas (Massey, 1991; Weintraub and D�iaz-Briquets, 1994) ? this notion is not new Important however, to distinguish short term from long term effects. In an attempt to mitigate exaggerated hopes of �trade instead of migration�, Martin and Taylor (1996) put forward several compelling arguments why trade and migration can be complementary in the short to medium run, beyond the fact the economic development materially enables more people to migrate abroad. Higher productivity and efficiency, technological advantages, and economies of scale in the North may harm the competitiveness of the South even in the production of labour-intensive goods. Under such circumstances, trade liberalization can lead to concentrations of highly productive economic activities in the North along with more immigration of labourers to support them. Furthermore, adjustment to new market and policy environments is never instantaneous. While the negative impacts of trade liberalization (on protected sectors) are often immediate, the expansion of production in sectors potentially favoured by trade reforms always takes time. There may be a long lag between investment and the creation of new jobs, which seems to be a recipe for a migration hump in the wake of trade reforms (Martin and Taylor, 1996: 52). It is not entirely clear whether this pertains to income gaps based on absolute income or income adjusted to purchasing power. Instead of using relative income gaps, some studies from the mid-1990s suggested an average real per capita income threshold of around US$ 4,000 at which migration transitions would occur (Stalker, 2002: 171).

19. Migration transitions   �Take-off� development in the poorest countries may well result in increasing migration to Europe and elsewhere. �Take-off� development in the poorest countries may well result in increasing migration to Europe and elsewhere.

20. Migration and development as substitutes? The relationship between development and net emigration is neither linear nor inversely proportional. Social and economic development is generally associated with increased mobility and migration because it increases people�s (1) capability and (2) aspirations to migrate Beyond a hypothesized threshold level, migration gains its own momentum due to �internal processes� (e.g., networks, remittances, inequalities) ? accelerates migration between specific places and countries despite declining wage gaps Capabilities and relative deprivation Only after income gaps narrow significantly and networks disintegrate, migration tends to decrease and immigration to increase Beside their limited scope, consistency and credibility, the more fundamental reason why �development instead of migration� policies, even if they would encourage development, are unlikely to curb migration is that they are based on an inaccurate analysis of the developmental causes of migration. It is often implicitly or explicitly assumed that development has the effect of linearly decreasing emigration, which tends to be seen as the outflow of poverty, crises and general misery. However, the paradox is that the process of social and economic development in its broadest sense tends to be associated with generally higher levels of mobility and more migration, at least in the short to medium term The argument that promoting development in sending countries is an effective means to reduce immigration is based on the underlying assumption that migration and development are negatively and linearly correlated processes and, hence, substitutes for each other. Following this logic, migration can be reduced by addressing the alleged �root causes� of migration, such as economic underdevelopment, poverty and unemployment. Boosting �development� in sending societies is seen as the key, to quote the title of B�ohning�s (1994) article, to �helping migrants to stay at home�. However, there is no empirical evidence that aid and trade policies have had any significant effect on reducing people�s propensity to migrate (cf. Nyberg-S�rensen et al., 2002a; Stalker, 2002). Indeed, there are a number of practical and more fundamental reasons why �development instead of migration� programmes are bound to fail; these will be addressed in the following sections. Beside their limited scope, consistency and credibility, the more fundamental reason why �development instead of migration� policies, even if they would encourage development, are unlikely to curb migration is that they are based on an inaccurate analysis of the developmental causes of migration. It is often implicitly or explicitly assumed that development has the effect of linearly decreasing emigration, which tends to be seen as the outflow of poverty, crises and general misery. However, the paradox is that the process of social and economic development in its broadest sense tends to be associated with generally higher levels of mobility and more migration, at least in the short to medium term The argument that promoting development in sending countries is an effective means to reduce immigration is based on the underlying assumption that migration and development are negatively and linearly correlated processes and, hence, substitutes for each other. Following this logic, migration can be reduced by addressing the alleged �root causes� of migration, such as economic underdevelopment, poverty and unemployment. Boosting �development� in sending societies is seen as the key, to quote the title of B�ohning�s (1994) article, to �helping migrants to stay at home�. However, there is no empirical evidence that aid and trade policies have had any significant effect on reducing people�s propensity to migrate (cf. Nyberg-S�rensen et al., 2002a; Stalker, 2002). Indeed, there are a number of practical and more fundamental reasons why �development instead of migration� programmes are bound to fail; these will be addressed in the following sections.

21. Critique on prior theories Based on rather narrow definitions of development Focus on income and demographic variables, and non-inclusion of theoretically relevant variables like education and political freedoms Unilinear character (reversibility, relative poverty and opportunity gaps) Questionable causality assumptions (e.g. portrayal of demographic factors as �cause�) Lack of formal empirical tests

22. An empirical test Lack of empirical tests, partly due to lack of appropriate data. Recent advances - OECD database on immigrants and expatriates Bilateral migration matrix created by the University of Sussex and the World Bank. Aim: To provide a multivariate test for migration transition theory by assessing the effect of the levels of economic and human development on emigration and immigration patterns. Data: Bilateral migration matrix (N=226), World Bank, UNDP, Freedom House

23. Variables Dependent Emigration, Immigration, net migration, migration to OECD countries (proxy values based on bilateral stock data � Sussex/WorldBank database) Main independent variables Log GDP per capita PPP Fertility (average 20-40 years ago) GDP per capita Growth over the last 20 years Literacy (and squared value) Island state (dummy) Border OECD country (dummy) Political rights (Freedom House) Human Development Index (UNDP) Considered, but excluded from regression analysis (due to multicollinearity) Life expectancy School enrolment

27. ANOVA analysis: Stronger correlations with HDI compared to GDPANOVA analysis: Stronger correlations with HDI compared to GDP

36. Conclusions First results seem to confirm hypotheses of transitional migration theory Inclusion of HDI increases correlations and statistical significance Apparent dominance of economic factors Significant, but complex role of literacy and schooling (disentangling problem) Caveat: Migration transitions neither inevitable nor unidirectional (relative deprivation!)

37. Limitations Caution is warranted in making claims on causality based on limited, cross-sectional data Correlations on macro-level do not provide a behavioral explanation at the micro-level (e.g., how to measure aspirations?). Difficulty of disentangling the effects of different components of human development (e.g., unclear role of education)

38. Migration transitions revisited  �Take-off� development in the poorest countries may well result in increasing migration to Europe and elsewhere. �Take-off� development in the poorest countries may well result in increasing migration to Europe and elsewhere.

39. Further improvements (ongoing) Multi-level empirical assessment of macro and meso (migrant group or country-to-country) determinants of migration Disaggregating analysis to the bilateral level Simultaneous assessment of sending and receiving country and country-to-country migrant group variables (development gaps, role of distance, networks, policies) Improved operationalisation of political and policy variables Analyse time series data to study bilateral migration flows instead of stocks

40. Policy implications Take-off development in the least developed countries is likely to lead to take-off emigration (e.g., sub-Saharan Africa) Rather than a result of development failure, migration is integral part of broader development processes, increasing (1) capabilities, (2) aspirations and (3) opportunities to migrate Learn to see migration as an integral part of development processes rather than as problem to be �solved� The limited effects of migration policies compared to more general policies affecting human development (macro-economic, trade, education, labour market, health).

41.  Migration transitions and the migration hump  A theoretical inquiry and empirical test   Hein de Haas hein.dehaas@qeh.ox.ac.uk www.imi.ox.ac.uk

46. www.worldmapper.org