divergent patterns in the ethnic transformation of societies

POPULATION AND DEVELOPMENT REVIEW 35(3 ) : 449–478 (SEPTEMBER 2009) 449

Divergent Patterns in the Ethnic Transformation of Societies

DAVID COLEMAN

DEVELOPED-COUNTRY POPULATIONS that began their fertility transitions rela-tively early are becoming increasingly diverse with respect to the ancestry, ethnic origins, and religions of their inhabitants. This is primarily a result of high levels of immigration, particularly migration from the less developed countries. To a lesser extent, higher birth rates of some of the immigrant pop-ulations further amplify diversity. This article considers whether the growing diversity of the origins of the populations of Western countries is inevitable, what its social, economic, and political consequences might be, and whether increasing diversity may eventually become a universal phenomenon or instead might remain confined primarily to countries currently classed as developed. Globalization of increasing population diversity may be much less marked in many of those countries that began their fertility transitions later and are now progressing toward full economic development. If so, the later that those countries develop, the less heterogeneous they will become through migration. That divergence between countries, if it arose, could have important political and social consequences.

The assumption of demographic globalization

Many demographically relevant characteristics of the developed world—low death and birth rates and possibly the new attitudes to marriage and living arrangements common to those countries—are becoming universal. That may be regarded as a facet of globalization. Some long-term future scenarios en-visage a world where economic prosperity is more universally shared, where all populations share similarly low birth and death rates, where population growth has ended and the socioeconomic and demographic compositions of populations are convergent. Critics of this view point to the lack of theoreti-

450 E T H N I C T R A N S F O R M A T I O N O F S O C I E T I E S

cal justification for convergence on a conveniently replacement-level fertility rate (Lutz 2007), noting the increasing numbers of countries, including poor countries, where the net reproduction rate has fallen below one (Demeny 2003). Their population momentum exhausted, natural demographic change has already become negative in some European countries (Lutz, O’Neill, and Scherbov 2003). In Germany and in some Eastern European countries popu-lation is actually declining, with or without net immigration.

The drift into population decline will not be shared by all countries: if differential growth rates continue, then a new era of divergence in population size will unfold. In the nineteenth century, protracted divergence in growth rates reduced France from its former superpower status to mere numerical equivalence with other European powers. In the twenty-first century, the latest United Nations projections expect the United Kingdom to become the most populous country in Western Europe by mid-century, surpassing Germany and all the others. Further afield, the UN long-range projections of 2004 expected population growth in countries such as Congo and Uganda to continue until around 2100, by which time their combined populations would exceed that of today’s Northern Europe, Southern Europe, and West-ern Europe put together. Nonetheless, demographic convergence is assumed even for the most disadvantaged states with laggard fertility transitions, if only in the very long term. Population distribution in the world will then be very different from that of today.

Demographic transition and international migration

Large-scale migration from poor countries to the developed world is pow-ered, in part, by the century-long gap between the onset of their respective fertility transitions. In the developed world, societies have moderated birth and death rates for more than a century and are relatively old. Many have ceased to grow. With the exception of some former Communist countries, they have become rich. They send relatively few emigrants. Poor countries, mostly with incomplete regulation of fertility, still grow rapidly, have more youthful populations, often send many emigrants, and endure varying lev-els of poverty. As those countries develop, they are expected to undergo a ”migration transition,” turning from emigration countries into immigration countries (Zelinsky 1971; Skeldon 1997).

Many migrants from those poorer countries move to the developed world. Except in a few countries such as France, their immigration now comprises the greater part of developed-country population growth (Euro-stat 2008a) and most future projected growth to mid-century. Nine Western European countries are projected to grow by 15 percent or more by 2060, primarily from immigration. Seventeen former Communist countries, plus

D A V I D C O L E M A N 451

Germany, are projected to decline by up to 28 percent by the same year (Eurostat 2008b). Many countries in Central Europe also receive immigrants from low- and middle-income countries, although those immigrants do not at present compensate numerically for population loss from chronic low fertility (OECD 2007). Germany’s population decline is powerfully underwritten by negative demographic momentum despite relatively high immigration. High-er levels of immigration relative to population size maintain the population of Italy roughly in balance, while the huge inflows to Spain have reversed its demographic trends and prospects, from substantial projected decline to substantial growth. Migration flows can change rapidly and are consequently precarious to project.

Transformation of population composition

Aside from its strictly demographic effects in driving population growth or mitigating its decline, this migration has begun to make substantial, and presumably permanent, changes to the ethnic and religious make-up of the Western receiving countries. Populations of recent immigrants and their descendants increase, progressively diminishing the share of the native or indigenous population over time. According to the projections by race and Hispanic origin of the population of the United States (US Census Bureau 2008), the minority populations together will comprise a majority of the US population by 2042, mostly through the direct and indirect effects of im-migration. The black, American Indian, and other indigenous populations are, however, mostly not of recent immigrant origin. Setting aside those populations (in all projected as 67 million by 2050), the Hispanic and Asian populations, whose most rapid growth dates from the acceleration of im-migration in the 1960s, are projected to comprise 36 percent of the total US population in 2050.

For ten other developed countries, mostly in Northwest Europe and the English-speaking world, projections point to similar trends and outcomes to mid-century and beyond, although on somewhat different definitions, on the assumption of continued net immigration at the base year of the projec-tion. Most do not use ethnic or racial criteria and include only the first (im-migrant) generation and second generation born in the country of destina-tion (see Coleman 2006). Together they are defined as the foreign-origin or foreign-background population. The third generation is generally assumed to be assimilated to the national population and to become Dutch, Swedish, and so on, thus disappearing from statistical view. Potentially permanent, usually self-ascribed ethnic categories are employed in the projections for the English-speaking countries. Ethnic identity can also fade over generations, of course, be changed or rejected by individuals, or become merely symbolic (Gans 1979).


On these projections, between 15 percent and 30 percent of the total population of these Northwest European and English-speaking countries is expected to be of first- or second-generation origin by mid-century, in some cases not far short of the equivalent projected proportion in the United States. In most of the European projections, up to one-third of the projected foreign-origin totals originate from other European or from other developed countries. In all cases the projected proportion of foreign-origin population increases in nearly linear fashion up to the end of the projection period (mostly 2050 or 2060) (see Figure 1). Among the foreign-origin populations, the proportion with non-European backgrounds is expected to increase from about one-half today to about two-thirds by mid-century, or up to 20 percent of the total population. The composition of the population by origin does not necessarily determine the degree of social or economic salience of these differences. That would depend upon the origins of the immigrants, the suc-cess of policy, and the degree of integration and social assimilation, including inter-ethnic union and changes in ethnic self-identification.

In general, immigrant and minority populations are expected to inte-grate progressively into the host economy and society, especially the second

2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 20505

10

15

20

25

30

35

Per

cen

t of

tota

l popu

lati

on

FIGURE 1 Projected growth of the population of immigrant or foreignorigin 2000–50, selected countries, as percent of total population

NOTE ON THE PROJECTIONS: Because not all the sources provided data by single-year periods, five-yearperiods had to be used instead. Some of the projections did not begin with years ending in 0 or 5. In those casesthe starting point was taken to be the nearest year ending in 0 or 5 (either 2000 or 2005). United States:medium variant 1999-based (excludes non-immigrant minorities); Germany: medium variant 2000-based;Netherlands: medium variant 2004-based; Austria: Restriction of immigration, no naturalization 1999-based;Sweden: foreign background 2004-based; Denmark: 2004 medium variant; Norway: medium variant 2005-based.SOURCES: US Census Bureau 2000; Ulrich 2001; Alders 2005 and personal communication; Lebhart and Münz2003 and personal communication; Statistics Sweden 2004; Statistics Denmark 2004; Statistics Norway 2005.

Germany

Netherlands

United States

AustriaSweden

Denmark

Norway


and subsequent generations born in the host country. This is evident to an uneven degree in different groups and countries. In most European countries, evaluation of economic integration and marital assimilation is hampered by data that depend on statistics of nationality, not birthplace or ethnic origin. The former data give misleading impressions because of the high level of naturalization of persons of foreign immigrant origin. OECD reports (e.g., 2003) show that the situation of the children of immigrants in some OECD countries is unfavorable, with a high percentage at the margin of the labor market (OECD 2008: 6). Persistent difficulties of integration into the labor market are particularly notable in Germany (Liebig 2007: 56, 57).

Levels of intermarriage or inter-ethnic union can also be important indicators of integration, although not always in step with indicators on education and employment. In general, mixed unions between the major-ity population and persons of Caribbean origin are relatively frequent in the UK and the Netherlands, and also unions between whites and Arabs and Africans, especially among persons who are themselves of mixed origin. In the UK and elsewhere, mixed unions with persons of Asian origin, especially Muslims, are markedly less frequent and most show little sign of increasing over time (Berrington 1996; Coleman 2004; Lucassen and Larman 2009), although such unions are more frequent among second-generation Turks and North Africans in Belgium and France. A Belgian study showed that small size of ethnic group, urban origin, and high educational attainment promoted unions of Turks and Moroccans with whites in Belgium (Lievens 1998). The likely importance of the growth of populations of mixed origin is discussed later in the article.

Relative importance of migration and of differential fertility

Migration assumptions are more important than fertility assumptions in determining the demographic outcomes projected in Figure 1. Taking all minority groups together, ethnic-minority or foreign-origin total fertility in most European countries is not high: for example in France in 1991–98, 2.5 compared with 1.65 for women born in France. Corrected for distortions in childbearing arising from the migration process, these figures become 2.16 and 1.70 respectively (Toulemon 2004: Tables 2 and 3). The position is similar in the UK (ONS 2008: FM1 Table 9.5; Coleman and Scherbov 2005). Youthful age-structure of new entrants ensures a disproportionate contribution from births of immigrant origin, but for the most part birth rates are declining. Most projections assume eventual convergence of the fertility of ethnic groups on or near the national average.

Migration flows, by contrast, were very high in some countries compared with the annual total of births (see Table 1). In Spain, Switzerland, and Italy around 2008, net foreign immigration either exceeded the annual number of


births or was close to it. Projections of foreign populations are not available for many of these countries; for example France, Spain, Italy, Belgium. But the data in Table 1 suggest that transitions in population composition in those countries must be proceeding at a pace similar to those depicted in Figure 1.

An unavoidable destiny?

Migration trends are the most volatile of the three components of demo-graphic change, driven by often unconnected economic, demographic, and political developments in the numerous sending countries and in the receiv-ing country itself. Migration is much more subject to government policy than are birth or death rates, all the more so because the direction of migration policy—restrictive or promotional—can change radically between successive governments (Hollifield 2000).

The projected ethnic changes, based on unchanged migration flows, are therefore by no means inevitable but depend on migration policy as well as on many other unpredictable developments. Especially in liberal democra-cies, migration is much easier to encourage than to control (Freeman 1994). Western governments have limited their freedom of action by subscribing to

TABLE 1 Comparisons of yearly live births, net foreign immigration, and natural increase around 2008 in selected Western countries

Population Live Natural Net Immigration at start of births increase immigration as percent year (000) (000) (000) (000) of births

Spain 44,475 488 107 702 144Switzerland 7,509 74 13 69 93Italy 59,131 563 –7 494 88Norway 4,681 58 17 40 68Belgium 10,585 121 20 62 52Austria 8,299 76 2 31 41Greece 11,172 110 2 41 37Denmark 5,447 64 8 20 32United Kingdom 60,817 771 195 175 23France 61,538 784 268 70 9Germany 82,315 683 –141 48 7

11-country total 355,968 3,792 483 1,752 46

Australia 21,015 285 145 213 75Canada 33,311 357 127 204 57New Zealand 4,263 64 35 4 11United States 298,363 4,217 1,840 844 20

NOTE: US data refer to 2006; Australia, Canada, and New Zealand to 2007. Net immigration data not available for US. Admission for permanent settlement in US in 2006 was 1,266,264; figure reduced by one-third to allow for emigration. SOURCES: Websites of Eurostat, Australian Bureau of Statistics, Statistics Canada, Statistics New Zealand, US Census Bureau, National Center for Health Statistics, US Dept of Homeland Security.


international conventions on human rights and asylum, and by facilitating the migration of spouses, other relatives, and dependents that has dominated legal migration streams to the developed world in recent decades. The ef-fectiveness of state attempts to restrict immigration is uneven and depends on many factors (Weiner and Teitelbaum 2001: Ch. 8). But intentionally or otherwise, migration can go down as well as up, as the recent experience of Germany and the Netherlands has shown. Given the volatility of migration flows, projections can only show the consequences of specified assumptions: usually the continuation of current rates for want of any demonstrably pref-erable alternative. Reality may change radically; however, immigration to Western Europe and overseas countries of European origin has mostly been high and rising for some years and has been encouraged by government policy in the United States, Canada, Australia, and the United Kingdom, although recently the UK government has somewhat reined in its previous enthusiasm.

Until late 2008 it seemed reasonable to assume that high levels of migration and a positive approach to immigration were likely to continue among the governments of the English-speaking world, although policy in Continental Europe has been more volatile and more often reversed by political change. Flows of would-be migrants, legal and otherwise, continue at a high level. Despite opposition from electorates, upward pressures are strong—from the growing population of immigrant voters and their pressure groups, from the demands of employers for easy access to labor, from hopes that migration might avert demographic and workforce deficiencies, from the difficulty of stopping illegal immigration—all supported by pro-immigration sympathies of liberal elites. In addition, demand for overseas labor continues to be underwritten by the European social model of welfare, by retirement policies and labor protection that help to maintain relatively high levels of youth unemployment, and by low levels of workforce participation among the native and (in most countries) the immigrant populations.

Since the economic downturn of 2008–09 the picture looks somewhat different. Economic trends may not affect spouse and dependent migration and may enhance asylum-claiming. But labor migration would be expected to fall, especially if rising unemployment were to sharpen native resentment of immigrant workers. Published migration data, often a year late, have hardly begun to reflect the new economic environment. But some early figures show a downturn: registrations of new workers in the UK from the eight Eastern European countries that became members of the European Union in 2004 fell to 26,815 in Q4 2008, almost halved from the 50,820 in Q4 2007 (Home Office/UK Border Agency 2009: Table 1). Despite that decline, recent research has thrown cold water on this “buffer theory.” Long-term data from the UK, at least, suggest that the various inflows and outflows during economic downturns will fairly quickly return to levels that will keep net immigration more nearly stable than expected (Dobson et al. 2009).


Inter-ethnic unions

Tidy demographic categories may not long survive untidy human emotions.The growth of the number of children from mixed unions, and of children with mothers or fathers themselves of mixed origin, has begun to complicate and blur these ethnic or foreign-origin distinctions. In statistics, however, persons of more than one origin do not—indeed cannot—feature in most of the projections by origin made in those European countries where (as in most cases) these are based on registration data. Persons must be one thing or the other. Nonetheless children of mixed parentage comprise an increasing proportion of all minority populations where it is possible to record them, and their diversity is beginning to erode the usefulness of conventional clas-sifications of origin (Smith and Edmonston 1997; Edmonston, Lee, and Pas-sel 2002; Goldstein and Morning 2002; US Census Bureau 2008). In the US census of 2000 7.3 million people, 2.6 percent of the total population, claimed multiple ancestry (Jones 2005; Hollmann and Kingcade 2005). In the British census of 2001, 661,200 people identified themselves as mixed (1.1 percent of the UK population), or were so identified by their parents. Eight percent of children born from 1990 to 2001 were recorded as having a different ethnic group from their mothers, as were 22 percent of children of non-white moth-ers (ONS Commissioned Table CO431 from the 2001 census).

Taking those patterns into account, the mixed populations in the UK have been projected to increase from 661,200 in 2001 (1.1 percent) to 3.7 mil-lion by mid-century (5 percent of the projected total), by that time becoming the largest ethnic minority group (Coleman 2007). A different, probabilistic projection of the growth of the mixed population in the UK gave a median value of 8 percent of the total population mixed by 2050, with 26 percent of infants mixed, and 29 percent mixed overall by the end of the century (Cole-man and Scherbov 2005). A high prevalence of inter-ethnic unions would moderate the growth of immigrant-origin populations strictly defined, while creating new identities of mixed origin. These may eventually defy categoriza-tion and become a very numerous, if diffuse, population. The 2000 US census showed how individuals are developing their own simplified non-bureaucratic categories of identity (rather than ancestry), suggesting that “if intermarriage should continue to blur group boundaries, the reporting of ancestral roots will likely become more flexible, symbolic, and situational” and blur contemporary divisions and boundaries (Perez and Hirshman 2009: 44).

A universal globalized trend?

Data from a wider range of countries than in Table 1 suggest that diversifica-tion is a near-universal phenomenon in Western Europe and is beginning to make itself felt in Eastern Europe as well. Diversification is further advanced


in the English-speaking world outside Europe but less advanced in the larger developed countries in the Far East, for example Korea and especially Japan. Across the world, the proportion of immigrants in population totals has become highly divergent. The proportions of immigrants in the developed realms of Oceania (mostly Australia and New Zealand), North America, and Europe have shown substantial increases since the 1960s, especially of popu-lations of non-European origin. In the less developed regions the proportions of immigrants are only a fraction of those and, overall, have diminished (Table 2). Altogether, in 1960, according to the United Nations (2005), 43 percent of the world’s international migrants were in the more developed countries, 57 percent in the less developed. By 2005 those proportions had almost reversed: to 61 percent and 39 percent respectively. Furthermore, a substantial proportion of immigration to the three major regions of the more developed realm has been from outside the region, while most of the migration in the less developed world is within the region and often between neighboring countries. The data presented in Table 2, from official sources, cannot include most illegal or clandestine immigration. That clearly operates on a large scale; the data cited must underestimate flows and stocks but to an unknown extent.

Foreign immigrant populations in the developing world

Populations in the past that belonged politically to any kind of state, rather than to an autonomous tribe, were likely to be components of multiethnic empires, not of ethnically homogeneous nation-states in the modern sense

TABLE 2 Estimated percent of migrants within populations of major world regions 1960–2005

More Less Northern Latin World developed developed Oceania America Europe Africa Asia America

1960 2.5 3.4 2.1 13.4 6.1 3.4 3.2 1.7 2.81965 2.4 3.5 1.8 14.3 5.8 3.8 3.0 1.5 2.31970 2.2 3.6 1.6 15.4 5.6 4.1 2.7 1.3 2.01975 2.1 3.9 1.5 15.8 6.3 4.3 2.7 1.2 1.81980 2.2 4.2 1.6 16.4 7.1 4.5 2.9 1.3 1.71985 2.3 4.6 1.6 17.0 8.2 4.8 2.6 1.3 1.61990 2.9 7.2 1.8 17.8 9.7 6.9 2.6 1.6 1.61995 2.9 8.1 1.6 17.5 11.2 7.6 2.5 1.4 1.32000 2.9 8.8 1.5 16.3 12.8 8.0 2.0 1.4 1.22005 3.0 9.5 1.4 15.2 13.5 8.8 1.9 1.4 1.2

NOTE: These official figures underestimate the true level of migration, especially in the less-developed regions. SOURCE: United Nations Population Division, Trends in Total Migrant Stock: The 2005 Revision.


of the word (McNeill 1986). Despite the fragmentation of many empires, many of their smaller independent modern successor states retain some of that diversity in ethnic origin, language, and religion, while some very large modern states may be regarded as multiethnic empires in all but name (In-dia, Russia, Indonesia) even if run on democratic, not autocratic lines: the cumulated result of a history of invasions, expansions, migrations, and wars. For example, the 1991 census of India listed 114 languages each spoken by over 10,000 people, of which 18 were given recognized (“scheduled”) status. A more exhaustive estimate gives much higher numbers of languages: India 427, Indonesia 742, Malaysia 140, Nigeria 516, Thailand 83. Many of those are spoken only by a few people. The great majority of the languages so specified are of indigenous, not of immigrant origin. In the countries of the developed world the reverse is true. Linguistic diversity is lower, and the great majority of languages spoken are of immigrant, not indigenous origin, including also the former Soviet states of Central Asia (data from Lewis 2009: Table 5). The least diverse countries on this measure were Japan and the two Koreas.

This article is not concerned with the existing, long-standing diversity of so many of the world’s populations, to which their society, law, and at-titudes may have become well adapted or at least tolerant unless destabilized by the focus on numbers and diversity arising from democratic change and differential demographic growth. Instead it is concerned with the additional diversity, especially from new origins, that would ensue if the rest of the world followed the example of the contemporary Western countries in attracting new large-scale migration flows from novel origins.

In 1960 about 60 percent of the world’s stock of international migrants had moved from one less developed country to another, although by 2005 that fraction had fallen to 40 percent (UN 2005). Many less developed coun-tries have large numbers of foreign residents from neighboring states, even when all are relatively, though not equally poor and where there are few regional ”champions.” Numbers can fluctuate according to economic boom and bust, policy change, and political crisis, and are not well recorded.

In Africa, many migrants are refugees in the broad sense. Two percent of the population overall are estimated to be international migrants, and the proportion has not increased in a decade. In sub-Saharan Africa, 69 percent of migrants stayed within the region in 2005, while 80 percent of migration from North Africa was to Europe. Much of the migration into North Africa was of people from sub-Saharan Africa, many in transit to Europe. Not all succeed, leading to the possibility that those North African countries might become settlement countries (de Haas 2007). A few African countries have immigrant proportions approaching European levels, originating mostly from neighboring countries: Ivory Coast 13 percent in 2005, Ghana 7.5 percent. Except for small countries most of the rest have small migrant populations: 2 percent or less. In Latin America the larger countries have large numbers, but relatively small proportions, of immigrants: Argentina 3.9 percent, Venezu-


ela 3.6 percent, Chile 1.4 percent, Brazil 0.3 percent. Smaller countries, as is frequently the case, often record higher proportions: for example, Costa Rica 9.8 percent (data for 2005, from IOM 2008). Migration within Asia is now a significant influence on its economic, social, and demographic development. These flows, according to Hugo (2005: 62), are likely to continue regardless of political and economic development and the interventions of government. The smaller Asian developed countries now have significant numbers of im-migrants; the major Asian sending countries with their big populations still have small proportions of immigrants (although large absolute numbers): India 0.5 percent, Bangladesh 0.7 percent, and Pakistan 2.1 percent, the last including many refugees from Afghanistan.

Trends tending to diminish global ethnic diversity

The process of population and ethnic change has not been symmetrical. While the composition of the population of the developed world, especially of its cities, more closely resembles that of the developing world through migration, settled populations of European and many other origins are diminishing in areas outside Europe, some to the point of extinction. The developed world becomes more heterogeneous, parts of the developing world less so. In ear-lier centuries, European emigration had radical and sometimes catastrophic consequences for the indigenous populations of the lands that they settled: almost eliminating them in Australia and the United States; creating new mestizo populations in South America; introducing new variety by their own migration and by those of their African slaves and Asian indentured laborers throughout the world, from Trinidad to Fiji to East Africa. Empires tend to facilitate and often to encourage migration; the creation of Israel shortly after World War II was perhaps the last example. Now, with the end of empire, the tendencies are reversed.

In the calamitous first half of the twentieth century, ethnic diversity diminished in many of the countries of Europe. Re-drawing the map at Ver-sailles dismembered multi-ethnic empires in Europe, although some of the treaty creations were themselves multi-ethnic nation-states (Czechoslovakia, Yugoslavia) that proved to be fragile. The reduction of Hungary largely elimi-nated Romanian and other minorities in the drastically reduced Hungarian territory by creating new Hungarian minorities in Romania, Czechoslovakia, and elsewhere. But overall the effect was of ethnic simplification, albeit within more constricted national boundaries. The destruction of most of Europe’s Jews in the 1940s, and the severe diminution of its gypsy and other minori-ties, was altogether of a more savage kind. Further horrors would have been in store for Poles and other Slavs had the longer-term demographic aims of Hitler’s General Plan East been realized (Mazower 2008: Ch. 7). At the end


of the war and after it, forced migration removed most of the German popu-lations from Mitteleuropa—about 12 million people—a process now more or less completed in peacetime by the opportunities of return given to ethnic Germans from Eastern Europe and Russia under the German Basic Law. The borders of Poland were moved 200 miles further west; Poland’s dimin-ished population, formerly diverse, was 95 percent Polish by the late 1940s (Kosinski 1970). After the break-up of the Soviet Union in 1991, about 25 million people of Russian nationality living in the non-Russian constituent republics found themselves defined as foreigners (that accounts for most of the discontinuity in trend in Table 2). At least a third has emigrated to Russia, some (especially from the Central Asian republics) under duress. On a smaller scale, the demographic composition of the former republics and autonomous regions of Yugoslavia after 1991 has been greatly simplified by forced migra-tion and atrocities that have re-arranged about a third of the population and created “states and statelets that are over 70 percent monoethnic” (Mann 2005: Chs. 12, 13).

Looking further afield, at various times in the last century European populations, both ancient and of more recent immigrant origin, have almost completely disappeared from many regions outside Europe, leaving behind a more homogeneous population. Today the 70 million people of Turkey include substantial numbers of Kurds, Circassians, and Georgians. But most Armenians and Greeks have been removed, one way or another. Of the 1914 Armenian population in Anatolia of about 2 million, only about 10 percent remained by 1922 after the genocide of 1915 (Mann 2005: Chs. 5, 6; Bloxham 2007). The population exchange with Greece from 1923 and earlier mas-sacres and forced migration removed about 1.5 million Greeks from Turkey. Later measures in the 1930s and 1950s further reduced those permitted to remain in Istanbul. Thus Turkey has lost most of its (aboriginal) Greeks and Armenians. As a counterpart, about 500,000 non-Greeks, mostly Turks, were removed from Greece as part of the 1923 population exchange. Num-bers are uncertain but there may remain today between 15,000 and 20,000 Greeks and between 40,000 and 60,000 Armenians in Turkey, compared with perhaps 2–3 million of each at the beginning of the twentieth century. In the early Middle Ages Anatolia had been predominantly Greek-speaking. Following the creation of the state of Israel, the Jews of most Middle Eastern countries and from North Africa have departed or were forced out, as have been most of the Arabs from Israel. Christians in the Middle East, mostly of ancient—but not European—origin, are fast diminishing (Dalrymple 1997). There were estimated to be up to 1.4 million Christians in Iraq in 2004; in 2008 about 700,000 remained.

Permanently resident European-origin populations, mostly of colonial origin dating from the nineteenth and early twentieth centuries, have almost completely disappeared from North Africa and Egypt, and from most of sub-Saharan Africa except for Southern Africa. Altogether these populations once


numbered several millions. Thus in the French former territories around 1935, there were 1.48 million Europeans, in the British (not including the Dominions such as India and South Africa) about 735,000; in all about 4.16 million of European origin (Kuczynski 1937: Ch. 2). By 1950 that total had increased to over 5 million people. Thirty years later their number had dimin-ished to perhaps a tenth of that total. In Algeria there remained only about 75,000 Europeans, just 45,000 French, mostly expatriates. Between 800,000 and 1 million people of Portuguese origin left Angola and Mozambique after the Portuguese Revolution ended colonial wars in 1974. Southern Africa still has a numerous but diminishing white native-born population, mostly of British and Dutch origin, the latter dating from the seventeenth century; in Zimbabwe, the white population has fallen to a fraction of its peak number of about 250,000. According to the South African Institute of Race Relations, the white population of South Africa fell by 840,000 from 1995 (5.2 million) to 2005 (4.3 million), mostly because of the emigration of young adults and their children. Most other former colonies worldwide have lost their (often native-born) white minorities, perhaps the most numerous of which were the Dutch population who fled what is now Indonesia in 1950. Except for Southern Africa, those who remain are mostly expatriates, not long-term residents or native-born.

This process of diminishing ethnic diversity is not limited to European-origin populations. The Muslim and Hindu populations of India partly sepa-rated from each other through forced migration and massacre at the time of the partition that accompanied independence in 1947. Between 14 million and 18 million people left their homes, between 1 and 2 million died. Hindus comprised about 2 percent of the population of Pakistan at the 1998 census, compared with up to 20 percent before partition. The 2001 census of India reported that 13.5 percent of the population was Muslim, compared with about 20 percent before partition. In a number of less developed countries, non-European “market-dominant” ethnic minorities have found themselves expelled by newly independent governments, such as the South Asian popu-lations of Kenya and Uganda in 1973, or forced to flee as in the recurrent episodes of ethnic violence directed at the Chinese minority in Indonesia. The massacres of Tutsis in Rwanda, particularly in 1994, are also claimed to be of this kind. But various processes provoke migration crises and expulsions. Governments may expel relatively recent immigrants in economic downturns or because their numbers are deemed to be a threat. Two million clandestine workers, mostly Ghanaian, were expelled from Nigeria in 1983–85, several hundred thousand Palestinians and Yemenis from Kuwait and Saudi Arabia after the first Gulf War. Other new nation-states turn on old minorities, de-nying their right to citizenship: for example the quarter of a million Muslim Rohingyas of Burma who fled in 1978 and 1991–92, and the Nepalis who fled from Bhutan (Van Hear 1998: 63–110). Notions of representative government in newly independent countries, often with nationalist governments, have


sharpened such trends by focusing attention upon relative numbers (Chua 2002: Ch. 7).

Various formerly cosmopolitan cities such as Alexandria, Istanbul, and Sarajevo have accordingly become much more culturally and racially ho-mogeneous with the departure of Greeks, Jews, and others (Courbage and Fargues 1997), while a reverse process of diversification through new immi-gration transforms Los Angeles, London, Paris, New York, Sydney, Singapore, and other major cities.

How far will diversification through migration continue?

Studies of globalization have had little if anything direct to say about ethnic diversity per se. Strong accounts of the causes of globalization give primacy to economic processes: the rise of multinational corporations, the ease with which capital can migrate, the increase of economic interdependency. These processes weaken national boundaries and nation-state sovereignty and identity and bring in their train a more globalized or hybridized international culture facilitated by electronic communication, tourism, and migration (Ohmae 1996; Tomlinson 1999; Nederveen Pieterse 2003; reviewed in Martell 2007). While ethnic diversity is seldom mentioned in these accounts, it would seem to follow inevitably from the processes discussed, especially openness to migration. However, whatever the implications of globalization theory may be for the future, ethnic diversification of the world population is far from uniform and in some cases proceeds in the opposite direction.

Can the countries that currently send migrants expect to experience im-migration in their turn, and the diversity resulting from it, as they themselves complete the demographic transition, run out of population momentum, and begin to age, decline, and become economically attractive to migrants? To some extent, the experience of Western Europe and the United States is being reproduced in countries outside Europe as their economies have de-veloped and their surplus population is absorbed by their growing demand for labor. They too have attracted asymmetrical immigration streams from poorer neighboring countries, for example of Indonesians to Malaysia. In a few cases wealth has preceded demographic transition, particularly in the oil-revenue rentier economies of the Gulf, where an estimated 8.7 million temporary contractual workers from Asian countries were working around 2000 (IOM 2008: 443). There, oil money flowed in independently of indig-enous economic transition, permitting a high standard of living in parallel with only a slow and late decline in fertility and high levels of population growth. Manual and white-collar work is done predominantly by expatriates, who in some cases outnumber the local population. Although the presence of so many Indians, Pakistanis, Bangladeshis, Filipinos, Europeans, Americans, and others makes the populations currently highly diverse, diversity is not


intended to be permanent. Citizens and foreigners comprise quite separate populations: expatriates neither wish to nor are able to integrate, naturalize, or enjoy permanent residence or family reunion.

More interesting and more typical are those formerly developing-world populations that have been through more conventional demographic and economic transitions and social modernization: first and foremost Japan, fol-lowed by South Korea, Taiwan, Singapore, and Hong Kong. China, Malaysia, and Thailand are less advanced examples. All sent numerous emigrants to richer countries before and during their economic and demographic transi-tions, and the latter three still do. Now with demographic transition complete or well underway and rapid economic growth, immigration is increasing with concomitant increases in the proportion of immigrant workers in the labor force.

Apart from a strong overseas Asian diaspora, much of this migration is intra-regional, some of it between culturally and linguistically similar countries. In small, rich city-states about half the population was immigrant around 2000: for example Hong Kong (43 percent in 2005), Macau (56 percent), and Singapore (43 percent). Larger countries have much smaller proportions: Japan 1.6 percent, South Korea 1.2 percent. Net foreign im-migration to Japan and to South Korea in 2006 was 108,000 and 183,000 respectively, less than 0.1 percent and 0.3 percent of the population (OECD 2009). Japan, with only just over one million foreign immigrants, remains exceptional as a rich, aging country with relatively low immigration. In Japan, employers’ demands for labor conflict with a popular and political desire to protect Japanese society from the diversity that characterizes other old rich countries. In attempting to resolve this dilemma, Japanese policy attracted back to Japan descendants of the Japanese who emigrated to Brazil at the beginning of the twentieth century and who now number about 1.4 million. The 302,000 Brazilians in Japan by 2005 were the third-largest immigrant group. But they were welcomed for a Japanese origin that they had mostly lost: they had adopted Brazilian culture and spoke Portuguese (Higuchi 2005). Looking Japanese is one thing; behaving Japanese is quite another. They are now being paid to leave. Thailand, with 1.6 percent of its population im-migrant, is shifting from a migrant-sending to a migrant-receiving country. Malaysia, where immigrants (65 percent of whom are from Indonesia) ac-count for 6.5 percent of the population, has already done so (all data from International Organisation for Migration 2008).

Divergent destinies

This uneven timing of the demographic transition in different countries and regions of the world seems likely to create significant divergence in the level of national ethnic and religious homogeneity. Globalization and the migra-tion that goes with it will affect the composition of all populations. But the


growing diversity, mostly from immigration, currently transforming most of the Western countries may not be repeated among those countries that join the rich countries’ low-fertility club much later than others. Countries in the developing world have traditionally comprised many ancient minority groups. But the larger countries at least seem unlikely to acquire much ad-ditional diversity from new large-scale migration. The later their transition, the more likely they are to retain the basic pattern of their original popula-tion composition. However, their internal diversity may shift as indigenous minorities or culturally distinct regions (e.g., the states of North versus South India) follow the fertility transition at different speeds (Dyson et al. 2005) or stall along the path.

Five obstacles to the globalization of diversity

Five possible obstacles stand in the way of a globalization of diversity at the lev-el currently occurring in Europe and the United States. First is the sequential nature of demographic transition and economic development across the world. The second is relative population size. The third is past and current politics: the lack of imperial connections linking future potential receiving countries with even poorer potential sending areas, and the greater capacity of countries that are not democratic to resist unwanted forms of migration. The fourth is the resource constraint that might prevent current Western standards of consump-tion from being shared across the globe. The last is climate change.

The first set of economically developed post-transitional societies, with less than a billion people, comprise about 15 percent of the world’s total. By comparison, potential migrant-supplying populations are very large. The total population of the migrant-sending regions is about ten times that of the European Economic Area that receives them. In the case of the Indian sub-continent vis à vis the UK, the ratio is about 20 to 1. While those countries remain poor, the potential supply of immigrants is effectively inexhaustible. In the long run, though, the pool of potential migrants may become relatively small in relation to the population size, and the demand for migrants, of the destination countries. In addition, the remaining potential sending countries may be geographically remote from the potential destination countries, and not connected to them by any historical ties. The combined population of the next group to develop, led by Japan and then the four Asian dragons, is less than 200 million, or about 300 million if Thailand and Malaysia are included. That is still relatively small in relation to even regional, let alone to the global population of less developed status. But as other, larger countries, especially China, move only slightly later from population growth to stability and population aging, from labor surplus to labor shortage, the situation may begin to look rather different.


The future size of many currently less developed countries as they be-come rich, in conjunction with economic development in potential sending countries, may make it demographically more difficult for European-type minority proportions to arise. It is not easy to imagine immigration streams large enough to transform the populations of India (projected to peak at 1.557 billion in 2065, according to United Nations 2004: Table 8) or of China (peak-ing at about 1.451 billion in 2030), nor of Bangladesh, Brazil, Egypt, Mexico, Nigeria, Pakistan, Vietnam, or Yemen (the last with a peak population of 147 million shortly after 2100). They may people other countries, but they are less likely to be transformed by immigration themselves.

Table 3 presents a very partial selection of larger countries currently at different stages of economic and demographic development. The columns show the projected population in 2050 and the years at which population will reach its maximum before declining, will attain replacement-level fertility, zero natural increase, and an aged potential support ratio (that is, the number of persons aged 15–64 for every person aged 65 and over) of four, and esti-mates of GDP per head in 2030 where available. The projected demographic indicators are similar to those of developed, immigrant-receiving countries today, although the projected values of GDP per head in 2030 still lag consid-erably behind. Other economic forecasts are somewhat different: Goldman Sachs expects the 2030 GDP per head for Brazil to be $9,828, more than in Table 3, but less for India and China ($3,473 and $9,369 respectively in 2003 US dollars; Wilson and Purushothaman 2003: 9). Growth rates can be very high: Vietnam, with GDP per capita of only $723 in 2006, is expected to grow at 8 percent per year to reach $2,834 by 2020 in 2006 US dollars (Qiao 2008). The projected demography, anyway, could indicate the earliest approximate period when those countries become attractive to immigrants from poorer countries and when domestic employers will begin demanding them. Most have high levels of economic inequality, however, and considerable sources of potential labor in the traditional agricultural sector. Three non-European developed countries, which have surpassed all these indicators, are shown at the head of the table for comparative purposes.

Together with China, seven other large countries totaling 900 million people, 2.3 billion in all, have already reached replacement fertility or will do so within a few years. Demographic momentum will maintain popula-tion growth for longer; most except China are still projected to show modest natural increase by 2050. Before the recent economic downturn at least, all were projected to become richer. Their aged potential support ratio is pro-jected to fall to about four within about 30 years, in China much sooner. That is the current level in most European and other developed countries today, at which population aging has become a serious concern and the bulges in the domestic workforce are about to retire into elderly dependence. GDP per head is expected to be around $10,000 per year or more, still only a half to


a third of that of the current rich world but four times or more that of many remaining countries that are less developed.

Poorer countries such as the Philippines, Indonesia, and Egypt, and the less developed Latin American countries will no doubt remain countries of outmigration. But closer to mid-century, their population surpluses may also be running out. Of itself, that does not necessarily imply that emigration will cease. It is possible for emigration to continue, at least for a while, from a

TABLE 3 Selected basic data for assessing future potential migration into larger non-European countries currently at different stages of economic and demographic transition, arranged by projected year of attaining replacement-level total fertility rate (TFR = 2.1)

Year of projected Year of GDP per Estimated Projected RNI = 0 projected capita population population Year of Year of (or per PSR = 4 projected in 2008 in 2050 maximum projected 1000 (or in in 2030 (millions) (millions) population TFR = 2.1 in 2050) 2050) ($000)

Japan 126.9 99.4 2005 1969 2005 1997 30.072Taiwan 23.0 20.2 2020 1984 2024 2022 33.666South Korea 48.2 44.1 2023 1989 2027 2022 30.643

Thailand 67.3 73.1 2050 1992 2030 2032 14.014China 1,336.9 1,416.3 2032 1997 2037 2027 15.763Iran 73.3 97.0 2050 2002 2.8 2047 10.789Brazil 192.0 218.5 2045 2007 2.1 2032 8.316Turkey 73.9 97.4 2050 2007 1.7 2042 13.111Indonesia 227.2 287.6 2050 2012 1.6 2042 6.924Mexico 108.6 129.0 2043 2012 1.6 2037 10.668Vietnam 87.8 118.5 2060 2012 2.4 2042 8.292Argentina 38.7 50.9 2050 2017 2.5 2040 Bangladesh 160.1 222.5 2075 2017 5.9 4.5 8.292India 1,181.5 1,613.7 2065 2022 3.3 4.5 7.089Malaysia 27.0 39.5 2070 2022 3.8 2047 8.292Philippines 90.3 146.2 2075 2027 6.3 4.6 8.292South Africa 48.1 56.8 2075 2032 1.1 8.5 Egypt 80.1 129.5 2075 2037 6.2 4.5 Congo 64.3 147.5 2100 2050 11.0 7.8 Nigeria 151.2 289.1 2095 2050 10.3 9.3 2.027Pakistan 176.9 335.5 2090 2050 8.2 5.9 8.292Uganda 31.7 91.3 2115 2050 17.0 15.6 2.027

NOTE: Japan, Taiwan, and Korea are included because they are already economically and demographically developed but do not yet admit large immigration streams. SOURCES: GDP data from Maddison, 2007, table 7.9; GDP values in italics are from Maddison’s regional values in table 7.3 and are not country-specific. Population size, TFR, RNI (rate of natural increase), PSR (potential support ratio: persons aged 15–64/65 and older) data from UN World Population Prospects Medium Variant (2008). Taiwan data from US Census Bureau and Council for Economic Planning and Development, Taiwan «http://www.cepd.gov.tw/encontent/dn.aspx?uid=5829». Italics in the “year of maximum population” column indicates that the data are from UN 2004, table 8, not from the 2008-based projections, which do not extend beyond 2050.


society that is aging or even declining in numbers. Some poor countries that are approaching low population growth levels, or even Eastern European countries that are in a state of acute population decline, such as Bulgaria, still send emigrants. So far, however, that combination of circumstances can only be found in the transitional economies of Eastern Europe with their proximity to Western Europe and their recent EU membership, particularly from the unreformed economies of Bulgaria and Romania.

Emigration to the UK from the more prosperous new EU countries such as Poland was already in decline before the 2008–09 economic downturn, with rapid growth in Polish GDP and corresponding changes in currency ex-change rates altering the balance of attraction. Institutionalized emigration, not the balance of births and deaths, may become the main demographic driver of population stabilization or decline, as in nineteenth-century Ireland. Unreformed economies and attractive neighbors may maintain or even ac-celerate departures, as in today’s Bulgaria. But the examples of Spain, Italy, Greece, Hong Kong, Japan, and South Korea, which were transformed from emigrant-sending to immigrant-receiving counties as they developed demo-graphically and economically, have been more typical.

Demographic patterns of migration can change rapidly, of course. The Mediterranean countries shifted postwar from sending migrants (mostly guestworkers to richer European countries) to receiving migrants (mostly from poorer non-European countries) in just a few decades. Spain offers the most dramatic example in the last decade, although the apparent abruptness of the change arises in part from the official recognition of previously hid-den large-scale clandestine immigration. The recent mass immigration to the Mediterranean countries has progressed to some extent despite government policy, not because of it. The rapid expansion in international migration to the UK, from about 50,000 net entry in the 1990s to over 200,000 today, is largely the result of a specific change of immigration policy by the UK govern-ment. Such rapid switches—so far always from emigration to immigration as development proceeds—may occur elsewhere.

Beyond that, the remaining sources of potential large-scale migration would have to come primarily from sub-Saharan Africa and from a few large Asian countries with delayed transitions, notably Pakistan. By mid-century many currently less developed populations should be approaching demographic maturity. Population aging will impose new demands on their domestic population resources, and further economic growth should have moderated the attractions of emigration. Looking at the poorer countries, the vast gulf in earnings, opportunities, and lifestyle between them and the developed world, which shows little sign of narrowing, implies a continuing potential for migration. But in the very long term the world will start to run out of potential migrants. Demand for labor arising from population aging will have to be met more often from domestic sources. When they become


economically developed, therefore, the future population composition of some large currently developing countries may not have changed much from that of the present, except where their national territory already includes re-gional populations or indigenous minority groups at different stages in fertility transition. In China, these number only about 5 percent of the population. India, however, is a complex mosaic where internal differentials in fertility transition will alter the future ethnic and religious population balance even without substantial immigration.

Political contrasts

Political considerations also militate against a European-style experience being closely repeated in all other countries. Most immigration into Europe since the 1970s has been of dependents, spouses, students, asylum seekers, or illegals. Many in those categories may work but they did not enter as part of a formal labor recruitment stream. This inflow was greatly facilitated by the adoption of human rights and asylum conventions by liberal democra-cies. Such democracies do not find it easy to control immigration, even when pubic opinion, as it usually does, favors control. Such states, characterized as postmodern by theorists such as Cooper (2003), are willing to transfer some of their sovereignty to international conventions and supranational organi-zations such as the European Union and the European Court of Justice and the European Court of Human Rights. Postmodern states have “a much more open and tolerant attitude toward cultural, economic and political interaction, and define a much narrower range of things as threats to national security” (Cooper 2003: 221–222). Many non-European countries do not adhere to such principles or conventions and, pace Fukuyama (1992), may be unlikely to develop Western-type liberal systems or adopt human rights conventions in the foreseeable future. Political postmodernity may not follow in the train of globalization but may remain restricted to a “few groups of like-minded states” (Demeny and McNicoll 2006: 268) increasingly marginalized by their low fertility and by successful economic competition from emerging states with very different histories and traditions.

Indeed some argue that the instant acquisition of forms of universal democracy in poor countries with a weak middle class, without qualifica-tions for the franchise, and without established liberal institutional structures accentuates hostility to minorities and immigrants by focusing attention on relative numbers and ethnically based interests (Chua 2002). Middle East-ern and African states and some Asian ones can and do control migration effectively, prevent the settlement of dependents, and remove en masse mi-grants whose presence is not required. The constrained position of the very large migrant labor populations in the Gulf States was noted earlier. Highly developed non-Western countries where the rule of law is fully established,


notably Japan, but where liberal sentiments have not necessarily developed to advantage, have so far also controlled immigration and kept immigrants constitutionally at arm’s length more effectively than European countries have wished or been able to do. As Castles (2000: 3) notes, “The prospects of ethnic community formation, growing cultural diversity and emergence of multicultural societies are anathema in Asian labor-importing countries, while the notion of turning immigrants into citizens is unthinkable.” Their view of the experiences of Western countries with ethnic diversity may be decisive, softening or hardening controls. Nevertheless, governments may be losing control there, although more to employer pressure and illegal entry than to liberal sentiment.

The checks of resource limits and climate change

Serious obstacles may derail the economic forecasts of Maddison and other analysts, such as issued by Goldman Sachs, and upset the patterns of migration expected from uninterrupted globalization of wealth and demographic transi-tion. The current economic downturn will slow the whole process; all players will be affected, though unequally. That will pass. A longer-term objection to the achievement of the rapid and, it is assumed, unlimited growth posited in economic forecasts is that assumptions of the globalization of consumption will collide with resource limits. In part, this is a consequence both of the very large population size of the emerging economies and of their late and in some cases sluggish fertility transition. We are accustomed to unrealized forecasts of environmental and resource crises, but the logic of recent trends affecting energy, water, and food is difficult to avoid. Many trends are adverse: decreas-ing yields from fishing, the persistently large and even growing numbers of persons in the world who are malnourished, the degradation of agricultural land and diminishing usable reserves, and perhaps most crucially the over-use of fresh water (Millennium Ecosystem Assessment 2005; Schiermeier 2008). Some of the biggest populations and fastest-growing economies, no-tably China’s, are particularly vulnerable to increasing water shortages, with supply constrained by pollution and water availability per head already only one-quarter of the world average (Xie et al. 2009). This and other problems of population and economic growth have considerable potential for disrupting China’s return to world economic hegemony (Smil 2005).

Resource and environment problems would be bad enough by them-selves. Their interaction with climate change may push economic trends and migration patterns onto very different paths from those discussed above. Initially, climate change was believed to be progressing slowly, with its dire effects measurable only on a century-long scale (Manabe et al. 2004). While demographers recognized the effects of population on climate change (Meyer-


son 1998), effects of climate on population seemed far off. Climate change was not mentioned by the United Nations as a possible constraint in a projection to 2300 made in 2002 (United Nations 2004), although commentators on those projections drew attention to the likely constraints of climate change on the projected growth (Dyson 2004). Ecological constraints, while mentioned in the UN projections to 2150 of 11 years ago, were not incorporated in them. Even without global warming, sustainable population is notoriously difficult to estimate: the middle half of Cohen’s range of long-term carrying capacity estimates lay between 5 billion and 30 billion (Cohen 1996; United Nations 1999: 36). Other projections, however, have attempted to incorporate nega-tive feedbacks (Lutz, Sanderson, and Scherbov 2004). Limiting factors may of course be more difficult to project than demographic trends themselves (Cohen 1998).

Recent analyses point to a faster pace of climate change and to weaker prospects of averting it (Dyson 2005). Important (mostly damaging) climate changes are now expected within the conventional 50-year time horizon of population projections and are therefore within range of the speculations presented here. Some models forecast a temperature rise of 4 degrees C by mid-century. Even with a rise of over 2 degrees but less than 4, serious cli-mate-related problems are likely to arise. Furthermore they may do so in a sudden, non-linear fashion (Lenton et al. 2008). The loss of Arctic summer ice is considered to be nearly certain; more remote in time, and with a lower probability, is the loss of the Greenland Ice Sheet and the West Antarctic Ice Sheet. Other serious possibilities include the conversion to savannah of the Amazon rainforest. Rising sea levels would affect millions of people in African and Asian delta areas and Pacific islands, and higher temperatures would damage agriculture (Battisti et al. 2009), making some areas increas-ingly uninhabitable. Most scenarios suggest that, while some high Northern latitudes and other areas would benefit, severe environmental damage will be inflicted on some of the poorest and most marginal tropical and subtropical areas, notably in sub-Saharan Africa and in the Indian subcontinent (Parry et al. 2007), rendering human populations unsustainable and similarly affecting parts of Australia, Southern Europe, and some other developed areas.

The double uncertainties of climate warming and of demographic re-sponse to it make the magnitude of numbers of people potentially displaced by any future climate change very difficult to estimate. Already 25 million people are claimed to be ”environmental refugees,” although by no means necessar-ily as a result of global warming. ”Forced climate migrant” is a better term in the present context. A widely cited estimate of the number of persons likely to be displaced by future climate change is 200 million by 2050 (Myers 2001); other estimates range from 25 million to 1 billion. The Intergovernmental Panel on Climate Change has prepared a number of scenarios of migration effects, of varying degrees of severity, based on assumptions about future gas


emissions, population growth, meteorological response, and effectiveness of adaptation. In the mildest scenario, forced climate migration would increase by 5–10 percent, on existing routes. In the worst, Myers’s estimate might be exceeded (Brown 2008).

Climate change has already been cited as contributing to the flight of about one million Africans across the Mediterranean to Southern Europe in recent years (Myers 2001, 2005). But climate change is nothing new; nor are its effects on human migration. Cold periods facilitated the invasions of Goths and Huns across river barriers in antiquity; dry periods are thought to have facilitated the Arab expansion (Brown 2008) and movements in Africa, and in earlier times the spread of agriculture. The drier climate of the Pliocene is thought to have forced our ancestors out of the trees and into the spreading African savannah (Potts 2007). Those, however, were natural changes. The current episode is not.

Most people fleeing problems in poor countries end up in the poor coun-try next door. But it is difficult to see how large numbers of people could be accommodated in poor neighboring African or Asian countries if the scale of environmental damage were great and if many parts of each region could no longer sustain large populations; all would be suffering to some extent. The most attractive option to many of these people would be movement to the developed world, with its greater resources, through previously established migration channels and networks.

Forced climate migration would disrupt in ways impossible to simulate the sequence of migrations sketched above if the globalization assumptions were realized, and would also severely hinder the assumed economic progress that underlies the demographic and economic projections. Migration might continue to be centered on Western Europe and North America with their liberal asylum policies, not on more vulnerable and less welcoming areas whose economic growth might be stifled. Small low-lying populations are already preparing to move. In 1987 the President of the Maldives, where the population of 300,000 lives on average 2 meters above sea level, pleaded with the United Nations to save his country from elimination. The Maldives are still there, but the government of Tuvalu, where flooding and salination from the sea is already troublesome, aimed to abandon the homeland and move the entire population (11,000) (Brown 2001). The population of the Carteret Islands, expected to become uninhabitable by 2020, began to be evacuated to Papua New Guinea in 2008.

These presumed miner’s canaries of climate change are demographi-cally insignificant: their total population is about 500,000 and other factors contribute to their difficulties. The real challenges to conventional expecta-tions about migration will come from elsewhere. A one-meter rise in sea level would submerge one-half of the riceland of Bangladesh and displace about 70 million people, according to the World Bank (2000: 111–112). In that case


India, which would also lose productive land, would surely find itself hosting a large part of the population of Bangladesh. Bangladeshis already represent a big component of the estimated 5.7 million immigrants in India.

Long-term consequences of ethnic transformation

Setting aside resource constraints and the unpredictable consequences of severe climate change, what would be the consequences of ethnic transfor-mation and its divergent levels between regions should they occur? As noted above, continuation for just a few more decades of current levels of immigra-tion into developed countries would raise the foreign-origin proportion of the population to between 20 and 30 percent of the projected mid-century total. In the UK projection, for example, while 80 percent of the 2051 population would be of British origin, only 54 percent of the 0–14 age group and just 52 percent of 0–4-year-olds would be of indigenous origin. If continued to the end of the century and beyond, that trend would eventually displace the original population from its majority position. In that event, the character, identity, and cohesion of countries that formerly regarded themselves as na-tion-states could be profoundly altered.

The prospect of progressive ethnic transformation, which I have elsewhere labeled a third demographic transition (Coleman 2006), raises political and philosophical considerations untouched by earlier demographic transitions. The issue is controversial. Opinions are divergent, often bound up with strong political and moral sentiments. Depending on the level of adaptation and the pace of immigration, a pessimistic outlook would be that small-scale, urban daily life outside the home could be conducted more and more in the com-pany of people who might be regarded as strangers, and the older indigenous population would be increasingly concentrated in suburban and rural areas. In Europe, unless integration became more successful than it has been to date, national identity and the perception of national interests could become more complex and contentious, and older national histories and formerly shared myths, values, and symbols would become increasingly irrelevant or even offensive to newer populations. The culture to be taught in schools, and the choice of whose history, if any, is to be covered are already vexed questions. Popular opinion, reflected in opinion polls, expresses unease at such develop-ments and resentment at the equivalence with which the demands, needs, and entitlements of established populations and of immigrants are treated (Dench 2003). However, response to the prospective drift of the white, non-Hispanic population in the United States into minority status before mid-century has, thus far at least, been restrained (see Perez and Hirschman 2009).

Postmodern elite opinion is more nuanced and optimistic, believing that society should have evolved beyond identities defined by religion and nation and its past (Fukuyama 2007). Some observers, perhaps lacking confidence in


the merits of their own society, find it difficult to articulate acceptable reasons for objecting to the prospect of ethnic change, while others actively welcome a more diverse society on various liberal and humanitarian grounds or be-cause of a distaste for the current nature of their society. Consensualist views deny that there can be real divergences of interest between minority groups and their hosts (Dench 2003: 251). In the longer run, the rise of inter-ethnic unions would be both symbol and cause of the creation of a new society, al-beit with new identities. The extent to which the presence of people of varied origins is increasingly taken for granted in neighborhoods, in the workplace, and among circles of friends, and the (mostly) peaceful nature of the process, are encouraging. The last-mentioned effect, however, may be accomplished primarily by the peaceful if reluctant withdrawal of the original inhabitants to suburban and rural locations, or overseas.

Whether minority religions, notably Islam, retain current high levels of adherence, orthodoxy, and ethnic associations and customs—or instead develop a new form in which religious identity loses its priority—would be especially important in gauging social outcomes. Much depends on the origins of the immigrant populations, their desire to integrate or otherwise, and the policies put in place on integration and immigration. These factors all vary con-siderably. But with such a major transition in origins occurring in a relatively short time, the rapidly shifting numerical balance would make it difficult to in-sist on (former) minority populations adapting to (former) majority identities or values. The former majority may have to do the adapting. Slower inflows might facilitate a gradual accommodation. The introduction of new cultures into different environments has produced a series of unintended natural ex-periments on a very large scale that will shed new light on these issues. The final outcomes are yet unclear; the results are unlikely to be uniform.

There might be critical consequences for domestic and foreign policy. Already European countries with many Asian and Middle Eastern voters find themselves drawn into overseas conflicts where influential and growing sec-tions of electorates have specific ancestral loyalties or grievances (Armenia, Kurdistan, the Punjab, Kashmir, Israel, the Muslim world in general). The special interests of these populations have a claim to be incorporated into national policy, which thereby takes into account concerns that might be thought more proper to foreign governments. Many modern migrant popu-lations have become transnational populations, preserving links with their country of origin. That may require or encourage the governments of those countries to create and formalize links with migrants (such as the Mexican IME; Cano and Delgado 2007) for their welfare and possibly to benefit in turn from their leverage. The governments of Mexico and Turkey encourage the further growth and political involvement of their emigrant populations in the United States and Germany respectively. On a visit to Germany in February 2008, Turkey’s Prime Minister Erdogan claimed some rights of representation over Germany’s 2.7 million Turks, calling upon them to reject assimilation


as a “crime against humanity” while integrating into the German economy (Spiegel Online 2008). Immigration can have security consequences and pro-voke international complications for sending countries as well as for receiving countries (Weiner and Teitelbaum 2001: Ch. 9). Parliamentary representation and possibly the outcome of elections may also depend on the position taken by candidates on such matters. Defining the national interest could become more difficult, and the orientation of foreign policy altered. For example, it has been suggested that the traditional orientation of US policy toward Europe will make less sense to future American electorates, a high propor-tion of whom will have origins in Latin America or Asia, quite apart from global shifts in the balance of power and realpolitik. Lacking foreign policy pressures from immigrant minorities, sending countries might find it easier than receiving countries to retain a reasonably coherent national identity and foreign policy.

On the other hand, the ethnic diversity of the Western powers could be regarded as a considerable asset in foreign policy, permitting a more sen-sitive appreciation of the concerns of other governments and peoples than otherwise might be possible. If diversity inhibited overseas military or other adventures, some would regard that as a benefit, and in general it might en-courage awareness of the interests of others. In many parts of the world the history, actions, and aims of the West are regarded with deep suspicion. Rep-resentatives of Western powers who are obviously not of traditional Western origin, but of similar origin to those with whom they are dealing, may help to replace suspicion with confidence and empathy. The selection of some recent Western ambassadors suggests that this consideration is already in play; and the election of President Obama is claimed to have favorably changed the attitudes and expectations of African governments and people.

This article has suggested that the ethnic diversity becoming typical of Western countries may not become a globalized phenomenon. Today’s send-ing counties, by virtue of their very large populations, their stricter policies on admission or integration, or some eventual global shortage of suitable labor migrants, may retain their traditional ethnic and religious population composition. Those countries, coming late to demographic and economic transition, might not share, at least for a long time, the new ethnic diversifi-cation characteristic of the pioneers in transition, while the original cultures of the latter would be at least modified and possibly transformed. Were that to be the case, the third demographic transition mentioned earlier would not be a transition at all. It would remain incomplete at the global level, confined mostly to today’s rich countries with their future diminished position on the world stage. Nothing is certain; much depends upon the development of migration patterns and the success of integration policies. But however dif-ficult these outcomes might be to predict, either way for good or ill they are unlikely to be without consequences for the internal coherence of nations and for the relations between them.


Note

I am grateful to Robert Rowthorn for helpful comments.

References

Alders, M. 2005. “Allochtenprognose 2004–2050: belangrijkste uitkomsten” (Projections of the foreign-origin population in the Netherlands), Bevolkingstrends: Statistisch kwartaalblad over de demografie van Nederland. 53, 1e qwartaal 2005. (Population Trends: Statistical Quarterly on the Demography of the Netherlands, 53, 1st quarter 2005): 32–41.

Battisti, D. S. and R. L. Naylor. 2009. “Historical warnings of future food insecurity with un-precedented seasonal heat,” Science 323(5911): 240–244.

Berrington, A. 1996. “Marriage patterns and inter-ethnic unions,” in D. A. Coleman and J. Salt (eds.), Ethnicity in the 1991 Census. Volume 1: Demographic Characteristics of the Ethnic Minority Populations. London: HMSO, pp. 178– 212.

Bloxham, D. 2007. The Great Game of Genocide: Imperialism, Nationalism and the Destruction of the Ottoman Armenians. Oxford: Oxford University Press.

Brown, L. R. 2001. Rising Sea Level Forcing Evacuation of Island Country. Washington, DC: Earth Policy Institute.

Brown, O. 2008. “Migration and climate change,” IOM Migration Research Series No. 31, International Organisation for Migration, Geneva.

Cano, G. and A. Delano. 2007. “The Mexican government and organized Mexican immigrants in the United States: A historical analysis of political transnationalism, 1848–2005,” Working Paper 148, The Center for Comparative Immigration Studies, San Diego.

Castles, S. 2000. “Migration as a factor in social transformation in East Asia,” presented at the Conference on Migration and Development, Princeton University, 4–6 May. University of Wollongong, Centre for Asia Pacific Social Transformation Studies «http://www.the-globalsite.ac.uk/press/010castles.pdf».

Chua, A. 2002. The World on Fire: How Exporting Free Market Democracy Breeds Ethnic Hatred and Global Instability. New York: Doubleday.

Cohen, J. E. 1996. How Many People Can the Earth Support? New York: W.W. Norton.———. 1998. “Should population projections consider ‘limiting factors’—and if so, how?,”

Population and Development Review 24(Supp.): 118–138.Coleman, D. A. 2004. “Partner choice and the growth of ethnic minority populations,” Bevolking

en Gezin 33(2) (Special Issue for the 75th Jubilee Conference of the Dutch Demographic Society, “Het huwelijk in de multiculturele samenleving/Marriage in multicultural so-ciety”): 7–33.

———. 2006. “Immigration and ethnic change in low-fertility countries: A third demographic transition,” Population and Development Review 32(3): 401–446.

———. 2007. “Ethnic change in the populations of the developed world,” presented at the An-nual Conference of the British Society for Population Studies, St. Andrews, 12 September «http://www2.lse.ac.uk/socialPolicy/BSPS/ppt/2007_Coleman.ppt».

Coleman, D. A. and S. Scherbov. 2005. “Immigration and ethnic change in low-fertility coun-tries—Towards a new demographic transition?,” presented at the Annual Meeting of the Population Association of America, Philadelphia, PA.

Cooper, R. 2003. The Breaking of Nations: Order and Chaos in the Twenty-first Century. New York: Atlantic Monthly Press.

Courbage, Y. and P. Fargues. 1997. Christians and Jews under Islam. London: I. B. Tauris.Dalrymple, W. 1997. From the Holy Mountain: A Journey in the Shadow of Byzantium. London:

HarperCollins.de Haas, H. 2007. “North African migration systems: Evolution, transformations and develop-

ment linkages,” Working Paper no. 6, International Migration Institute, James Martin 21st Century School, University of Oxford.


Dench, G. 2003. Minorities in the Open Society. New Brunswick, NJ: Transaction.Demeny, P. 2003. “Population policy dilemmas in Europe at the dawn of the twenty-first cen-

tury,” Population and Development Review 29(1): 1–28.Demeny, P. and G. McNicoll 2006. “The political demography of the world system, 2000–2050,”

Population and Development Review 32(Supp.): 254–287.Dobson, J., A. Latham, and J. Salt. 2009. “On the move? Labour migration in times of reces-

sion: What can we learn from the past?,” Policy Network Paper, July: 23.Dyson, T. 2004. “Why the world’s population will probably be less than 9 billion in 2300,” in

United Nations (ed.), World Population to 2300. New York: United Nations, pp. 145–150.———. 2005. “On development, demography and climate change: The end of the world as we

know it?,” Population and Environment 27(2): 117–149.Dyson, T., R. Cassen, and L. Visaria. 2005. Twenty-First Century India. Oxford: Oxford University

Press.Edmonston, B., S. M. Lee, and J. S. Passel. 2002. “Recent trends in intermarriage and im-

migration and their effects on the future racial composition of the US population,” in J. Perlmann and M. C. Waters (eds.), The New Race Question: How the Census Counts Multiracial Individuals. New York: Russell Sage Foundation, pp. 227–255.

Eurostat. 2008a. Population Projections of the Member States of the EU 2004-based. Luxembourg: Eurostat.

———. 2008b. “Population in Europe 2007: First results,” Statistics in Focus 81. Luxembourg: Eurostat.

Freeman, G. S. 1994. “Can liberal states control unwanted migration?,” Annals of the American Academy of Political and Social Science 534: 17–30.

Fukuyama, F. 1992. The End of History and The Last Man. London: Hamish Hamilton.———. 2007. “Identity and migration,” Prospect, February: 26–31.Gans, H. J. 1979. “Symbolic ethnicity: The future of ethnic groups and cultures in America,”

Ethnic and Racial Studies 2(1): 1–20.Goldstein, J. R. and A. J. Morning. 2002. “Back in the box: The dilemma of using multiple-race

data for single-race laws,” in M. Waters and J. Perlmann (eds.), The New Race Question: How the Census Counts Multiracial Individuals. New York: Russell Sage Foundation, pp. 119–136.

Higuchi, N. 2005. “Brazilian migration to Japan: Trends, modalities and impact,” presented at the Expert Group Meeting on International Migration and Development in Latin America and the Caribbean, 30 November–2 December, Mexico City.

Hollifield, J. F. 2000. “The politics of international migration: How can we ‘bring the state back in’?,” in C. Brettell and J. F. Hollifield (eds.), Migration Theory: Talking Across Disciplines. London: Routledge, pp. 137–185.

Hollmann, F. W. and W. W. Kingcade. 2005. Impact of racial and ethnic exogamy and interna-tional migration on forecast population distributions for the United States in 2030: Results of a macro-simulation,” 25th International Population Conference of the International Union for the Scientific Study of Population, Tours.

Home Office/UK Border Agency. 2009. Accession Monitoring Report May 2004–December 2008 A8 Countries. London: Home Office/UK Border Agency.

Hugo, G. 2005. “Migration in the Asia-Pacific region,” paper prepared for the Policy Analysis and Research Programme of the Global Commission on International Migration.

International Organisation for Migration (IOM). 2008. World Migration 2008: Managing Labour Mobility in the Evolving Global Economy. Geneva: International Organisation for Migration.

Jones, N. A. 2005. We the People of More Than One Race in the United States. Census 2000 Special Reports CENSR-22. Washington, DC: US Census Bureau.

Kosinski, L. 1970. The Population of Europe: a Geographical Perspective. London: Longman.Kuczynski, R. R. 1937. Colonial Population. London: Oxford University Press for the Royal

Institute of International Affairs.Lebhart, G. and R. Münz. 2003. Szenarien zur Bevölkerungsentwicklung in Österreich nach Staats-


angehörigkeit. Forschungsbericht Nr. 26. Vienna: Vienna Institute of Demography. (Scenarios for population development in Austria by citizenship. Research Report No. 26).

Lenton, T. M. et al. 2008. “Tipping elements in the Earth’s climate system,” Proceedings of the National Academy of Sciences 105(6): 1786–1793.

Lewis, M. P. 2009. Ethnologue: Languages of the World, 16th Edition. Dallas: SIL International.Liebig, T. 2007. The Labour Market Integration of Immigrants in Germany. Paris: OECD.Lievens, J. 1998. “Interethnic marriage: Bringing in the context through multilevel modelling,”

European Journal of Population 14(2): 117–155.Lucassen, L. and C. Larman. 2009. “Immigration, intermarriage and the changing face of Eu-

rope in the post war period,” The History of the Family 14(1): 52–68.Lutz, W. 2007. “The future of human reproduction: Will birth rates recover or continue to

fall,” Ageing Horizons 7: 15–21 «http://www.ageing.ox.ac.uk/system/files/ageing_hori-zons_7_lutz_fd.pdf».

Lutz, W., B. C. O’Neill, and S. Scherbov. 2003. “Europe’s population at a turning point,” Science 299(28): 1991–1992.

Lutz, W., W. C. Sanderson, and S. Scherbov (eds.). 2004. The End of World Population Growth in the 21st Century: New Challenges for Human Capital and Sustainable Development. London: Earthscan.

Maddison, A. 2007. Contours of the World Economy 1–2030 AD: Essays in Macro-Economic History. Oxford: Oxford University Press.

Mann, M. 2005. The Dark Side of Democracy: Explaining Ethnic Cleansing. Cambridge: Cambridge University Press.

Manabe, S. et al. 2004. “Century-scale change in water availability: CO2-quadrupling experi-

ment,” Climatic Change 64(1-2).Martell, L. 2007. “The third wave in globalisation theory,” International Studies Review 9(2):

173–196.Mazower, M. 2008. Hitler’s Empire: Nazi Rule in Occupied Europe. London: Allen Lane.McNeill, W. H. 1986. Polyethnicity and National Unity in World History. Buffalo: University of

Toronto Press.Meyerson, F. A. B. 1998. “Population, carbon emissions, and global warming: The forgotten

relationship at Kyoto,” Population and Development Review 24(1): 115–130.Millennium Ecosystem Assessment. 2005. Ecosystems and Human Well-being. Synthesis. A Report

of the Millennium Ecosystem Assessment. Washington, DC: Island Press.Myers, N. 2001. “Environmental refugees: Our latest understanding,” Philosophical Transactions

of the Royal Society B 356(16): 16.1–16.5.———. 2005. Environmental refugees: An emergent security issue. OCSE 13th Economic Forum,

Session III – Environment and Migration. Prague, OCSE.Nederveen Pieterse, J. 2003. Globalization and Culture: Global Mélange. Lanham, MD: Rowman

& Littlefield.OECD. 2003. Trends in International Migration. SOPEMI 2002. Paris: OECD.———. 2007. International Migration Outlook 2007. Paris: OECD.———. 2008. “Chapter 1. Summary of the main findings on the labour market integration of

immigrants and their children,” OECD Social Issues/Migration/Health 2008(18): 18–42.———. 2009 “A profile of immigrant populations in the 21st century: Data from OECD coun-

tries,” Database on Immigrants in OECD Countries (DIOC).Ohmae, K. 1996. The End of the Nation State. London: HarperCollins.ONS. 2008. Birth Statistics: Review of the National Statistician of Births and Patterns of Family Building

England and Wales. Series FM1 no. 36, 2007. London: Office for National Statistics.Parry, M. et al. 2007. “Climate change, global food supply and risk of hunger,” Philosophical

Transactions of the Royal Society 360(1463): 2125–2136.Perez, A. D. and C. Hirschman. 2009. “The changing racial and ethnic composition of the US

population: Emerging American identities,” Population and Development Review 35(1): 1–51.


Potts, R. 2007. “Environmental hypotheses of Pliocene human evolution,” in R. Bobe, Z. Alem-seged, and A. K. Behrensmeyer (eds.), Hominin Environments in the East African Pliocene: An Assessment of the Faunal Evidence. Springer, pp. 25–49.

Qiao, H. 2008. “Vietnam—The next Asian tiger in the making,” Goldman Sachs Global Eco-nomics Paper No. 165, Goldman Sachs.

Schiermeier, Q. 2008. “Water: A long dry summer,” Nature 452: 270–273.Skeldon, R. 1997. Migration and Development: A Global Perspective. Essex: Longman.Smil, V. 2005. “The next 50 years: Unfolding trends,” Population and Development Review 31(4):

605–643.Smith, J. P. and B. Edmonston (eds.). 1997. The New Americans: Economic, Demographic and Fiscal

Effects of Immigration. Washington, DC: National Academy Press.Spiegel Online. 2008. 12 February «http://www.spiegel.de/international/germany/0,1518,

534724,00.html»Statistics Denmark. 2004. Statistiske Efterretninger. Befolkning og valg (Population and elections).

Copenhagen, Statistics Denmark, p. 12.Statistics Norway. 2005. Framskriving av innvandrerbefolkningen, 2005–2060 Venter sterk vekst i

innvandrerbefolkningen (in Norwegian, with English summary: Projection of the immi-grant population 2005–2060: strong growth of immigrant population ) «http://www.ssb.no/emner/02/03/innvfram/».

Statistics Sweden. 2004. Sveriges framtida befolkning 2004–2050. reviderad befolkningsprognos frãn SCB. (Population projections 2004–2050: Revised projections from Statistics Sweden) BE 18 SM 0401. Stockholm: Sveriges Officiella Statistik. Statistik Meddelanden.

Tomlinson, J. 1999. Globalisation and Culture. Cambridge: Polity Press.Toulemon, L. 2004. “The fertility of immigrants: New facts, new approach,” Population and

Societies 400(April): 1–4.Ulrich, R. E. 2001. Die zukünftige Bevölkerungsstruktur Deutschlands nach Staatsangehörigkeit, Ge-

burtsort und ethnischer Herkunft: Modellrechnung bis 2050. Berlin/Windhoek (typescript), Unabhängigen Kommission “Zuwanderung” (The future population structure of Ger-many according to citizenship, place of birth and ethnic origin). Independent Commission on Migration, p 38.

United Nations. 1999. Long-Range World Population Projections: Based on the 1998 Revision. New York: United Nations.

———. 2004. World Population to 2300. New York: United Nations.———. 2005. Trends in Total Migrant Stock: The 2005 Revision. New York: United Nations.———. 2009. World Population Prospects: The 2008 Revision. New York: United Nations.US Census Bureau. 2000. Population Projections of the United States by Age, Sex, Race, Hispanic Origin

and Nativity, 1999–2100. Washington, DC: US Census Bureau.———. 2008. 2008 National Population Projections. Washington, DC: US Census Bureau. «http://

www.census.gov/population/www/projections/2008projections.html».van Hear, N. 1998. New Diasporas: The Mass Exodus, Dispersal and Regrouping of Migrant Communi-

ties. London: UCL Press.Weiner, M. and M. S. Teitelbaum 2001. Political Demography, Demographic Engineering. Oxford:

Berghahn.Wilson, D. and R. Purushothaman. 2003. Dreaming with BRICs: The Path to 2050, Goldman Sachs

Global Economics Paper No: 99, Goldman Sachs.World Bank. 2000. World Development Report 1999/2000: Entering the 21st Century: The Changing

Development Landscape. Oxford: Oxford University Press.Xie, J. et al. 2009. Addressing China’s Water Scarcity: Recommendations for Selected Water Resource

Management Issues. Report no. 47111. Washington, DC: The International Bank for Recon-struction and Development / The World Bank.

Zelinsky, W. 1971. “The hypothesis of the mobility transition,” Geographical Review 61(2): 219–249.

divergent patterns in the ethnic transformation of societies

Documents