A quarterly magazine of the IMF

September 2006, Volume 43, Number 3

Global Migration
Jeffrey G. Williamson

Two centuries of mass migration offers insights into the future of global movements of people

World migration has been going on for centuries, and free mass migration—of those not coerced, like slaves and indentured servants—has been going on for the past two. The reasons people move are no big mystery: they do it today, as they did two centuries ago, to improve their lives. What has changed is who is migrating and where they come from.

Both the demand for long-distance moves from poor to rich countries and the ability of the potential migrants to finance those moves have soared over the past two centuries. As the gap in living standards between the third world and the first world widened in the 20th century, the incentive to move increased. At the same time, improved educational levels and living standards in poor parts of the world—and falling transport costs globally, thanks to new technologies—have made it increasingly possible for potential emigrants to finance the move.

Thus, over time, poorer and poorer potential migrants, those who live the farthest from high-wage labor markets, have escaped the poverty trap. This emigration fact implies an immigration corollary that has important political backlash implications: relative to native-born host country populations, world immigrants have declined in "quality" over time—at least as judged by the way host country markets value their labor.

Adding to the rising demand for emigration, the population pool of the most mobile young adults increased as poor countries started the long process of economic modernization. Every country passes through a demographic transition as modern development unfolds: improved nutrition and health conditions cause child mortality rates to fall, thereby raising the share of surviving children in the population. After a couple of decades, this glut of children becomes a glut of young adults, exactly those who are most responsive to emigration incentives.

These demographic events were important in pushing poor Europeans overseas in swelling numbers in the late 19th century and even more important in pushing poor third-world workers to the first world in the late 20th century. At the other end of this demographic transition are the rich industrial countries, where population aging contributes to a scarcity of working adults and thus to a first-world immigration pull that reinforces the third-world emigration push.

Thus, the dramatic rise in world mass migration after the 1960s should have come as no surprise to any observer who has paid attention to history. But to truly understand world mass migration—and what might lie ahead—it is not enough to look at only the past few decades. We must assess the present relative to a past that stretches back over two centuries.

The first wave

The discovery of the Americas stimulated a steady stream of voluntary migration from Europe. High transport costs and big risks ensured that only the richest and most fearless made the move. Furthermore, distance mattered: the longer the move, the bigger the cost, and the greater the positive selection. These voluntary migrants, however, were dwarfed by those who came under contract and coercion. About 11.3 million journeyed to the New World before 1820, of whom 8.7 million were African slaves. Another large European emigrant group consisted of indentured servants and convicts, whose migration costs were financed by others. Thus, coercion and contracts were the chief means by which the labor-scarce New World recruited workers before the 19th century.

However, once started, the transition to free migration—which marks a decisive shift in the history of intercontinental migration—was spectacular: the share of free migrants in the total jumped from 20 percent in the 1820s to 80 percent by the 1840s. The combination of incentives, constraints, and policies that underlie the transition speak directly to the global migrations of today.

In the first three decades after 1846, the numbers of emigrants averaged about 300,000 a year; they more than doubled in the next two decades; and, after the turn of the century, they rose to over a million a year (see chart). Their countries of origin also changed dramatically. In the first half of the century, emigrants came predominantly from the richer parts of Europe—the British Isles, followed by Germany. By mid-century, they were joined by a rising tide of Scandinavian and other northwestern European emigrants and, in the 1880s, by southern and eastern Europeans.

The overwhelming majority of these emigrants headed for the Americas, the United States in particular. U.S. immigration from 1846 until the imposition of quotas in the 1920s follows closely the total European emigration pattern. After the mid-1880s, significant numbers of emigrants also went to South America, primarily Argentina and Brazil, and to Canada after the turn of the century. Another stream linked the United Kingdom to Australia, New Zealand, and South Africa. Still, between 1906 and 1910, the United States absorbed 64 percent of all emigration to the Americas (the main competitor being Argentina, which took in 17 percent).

Important migrations also took place within Europe. Spurred by the first industrial revolution, the Irish migration into Britain yielded an Irish-born share of almost 9 percent in British cities by 1851. In the 1890s, more than half of all Italian emigrants went to European destinations, chiefly to France and Germany. A third example is offered by the movement from eastern Europe to Germany, and from east Germany to west, patterns repeated even today. As the cost of migration from rural Europe to the gateway cities of the U.S. east coast fell, return migration soared. The U.S. authorities estimated that, between 1890 and 1914, return migration amounted to 30 percent of the gross inflow.

How large were these mass migrations for the sending and receiving countries? Rates exceeding 50 per 1,000 each decade were common for Britain, Ireland, and Norway throughout the late 19th century, and for Italy, Portugal, and Spain by the end of the century. The lower rates achieved by other countries are still high by modern standards. New World immigration rates were even larger than the European emigration rates, an inevitable arithmetic consequence of the fact that the sending populations were bigger than the receiving populations. In every New World country except Brazil, immigration rates far exceeded 50 per 1,000 in the decade of the 1900s.

Migration rates of this size imply significant economic effects on sending and receiving labor markets. This is especially so when we recognize that migrations tended to self-select those who had most to gain from the move, namely young adult males. Thus, the migrants had far higher labor participation rates than did either the populations they left or the ones they joined. It follows that the labor migration rates were even higher than the already-high population migration rates.

What was the resulting foreign-born share of Europe's and the New World's population in the late 19th century? Just before World War I, the highest foreign-born shares were for Argentina and New Zealand, about 30 percent, while the share was 14.7 percent for the biggest immigrant economy, the United States (see Table 1). These proportions are considerably higher than today, with migrant stocks now much more evenly spread around the greater Atlantic economy (Europe, the Americas, Australia, New Zealand, and South Africa). And Western Europe and Latin America are changing roles.

The mass emigration life cycle

Most of the 60 million Europeans who emigrated to the New World in the century after 1820 did so to escape poverty, and they did it without government assistance or guest worker status. Famine and revolution may have helped push the first great mass migration in the 1840s, but it was the underlying economic and demographic fundamentals that made each subsequent surge bigger than the previous one. If our only purpose were to explain why so many Europeans emigrated in the first global century, this essay would be very short indeed: after all, living standards were a lot higher in labor-scarce host countries.

But why did emigrating countries typically trace out a life cycle pattern? That is, emigration rates typically rose steeply from low levels as economic development took place in poor sending countries, after which the rise began to slow, and emigration rates peaked and subsequently fell off. This stylized fact—an emigration life cycle—has been documented in the first global century again and again. What accounts for it?

In preindustrial episodes, low emigration rates and low home wages coexisted: those who had the most to gain from migrating were trapped in poverty. Thus, enormous wage gaps between industrializing, resource-rich, high-wage countries and agrarian, resource-poor, low-wage countries were quite consistent with low emigration rates. As industrialization took place in the poor sending countries, real wages rose and the supply constraints on emigration were gradually released: more and more potential emigrants could finance the move. As this trend continued, the backlog of potential migrants was slowly exhausted.

The demographic transition also played a role. The fall in infant mortality rates tended, after a 15- or 20-year lag, to create a fatter cohort of mobile young adults, thus contributing even more to the emigration boom. In addition, remittances from previous emigrants helped finance the move of family members left behind. When the demographic transition reached a crescendo, when remittances leveled off, and when industrialization at home had raised wages and unlocked the migration poverty trap, further increases in the real wage at home caused the emigration rate to decline from the peak.

The first global century thus shows us that country emigration histories typically pass through two regimes—the first constrained by the supply of emigrants, and the second constrained by the demand for emigrants. The first regime was consistent with rising emigration and rising home wages. But, at some point, home wages were high enough that financial constraints became less binding: further increases in the home wage relative to the foreign wage then reduced the incentive to emigrate, the emigration rate fell, and the demand-constrained regime prevailed.

The emigrant life cycle implies that the source and quality of immigrants change over time. The spread of the transport and industrial revolutions, which reduced the cost of long-distance moves and the ratio of migration cost to annual income at home, extended the reach of global migration. More potential emigrants from the hinterland of western Europe and from distant parts of eastern and southern Europe could make the move. Thus, migrant origins shifted toward the countries that came late to modern economic growth. In addition, as each of these countries went through its own emigration life cycle, the share coming from the poor countries soared.

The powerful positive self-selection that had characterized the global migrations early in the century disappeared, and negative selection began to emerge. This dramatic shift obeyed economic and demographic laws of motion and implied a decline in the quality of immigrants—that is, the value of their skills in host country labor markets—and an even bigger decline in the quality of immigrants relative to the native born, who were accumulating human capital at a fast pace. There has never been any unambiguous evidence to suggest that immigrants face discrimination in U.S. labor markets, but they earned less than the native born before 1913 and, again, since 1970 have earned less. Why? On average, immigrants have less formal schooling and on-the-job training and poorer English-language skills and knowledge about jobs than do the host country native born. The number of immigrants increased markedly in the decades up to 1913 and has increased even more markedly since 1950.

This deterioration in relative and sometimes absolute immigrant quality had a great deal to do with rising negative attitudes toward immigration in the United States. Rising immigration also helped reinforce anti-immigrant feelings, with the native born feeling crowded out by the newcomers, and anti-immigrant sentiment intensified in the 1890s. Responding to constituent complaints, the House of Representatives proposed the Literacy Act to filter out immigrants from poor source countries; the Literacy Act finally became law in 1917. After the Great War ended, it was an easy matter for Congress to add the more restrictive Quota Acts of 1921, 1924, and 1927, as well as the ban on Asians. Other high-wage immigrant countries followed the U.S. lead, and the first global migration century came to an end.

The second wave

Annual immigration to North America and Oceania rose gradually after World War II until the mid-1970s before surging to a million a year in the 1990s. The absolute numbers were, by then, similar to those of a century earlier but were smaller relative to the population and labor force that had to absorb them. Thus, the annual U.S. immigration rate fell from 11.6 per 1,000 in the 1900s to 0.4 per 1,000 in the 1940s, before rising again to 4 immigrants per 1,000 in the 1990s. The proportion of the foreign-born U.S. population fell from a 1910 peak of 15 percent to an all-century low of 4.7 percent in 1970. The postwar immigration boom increased the foreign-born share to more than 8 percent in 1990 and more than 10 percent in 2000. After a half-century retreat, the United States has reclaimed the title "a nation of immigrants."

What happened to the United States after World War II also happened worldwide. The foreign-born share increased by about a third in Oceania between 1965 and 2000 (from 14.4 to 19.1 percent), more than doubled in North America (from 6 to 13 percent), and more than tripled in Europe (from 2.2 to 7.7 percent). Of course, the addition of undocumented immigrants would raise these foreign-born shares and would probably even raise their measured increase over time.

What is amazing about this modern boom in world mass migration is that it has taken place in such a hostile policy environment. Before World War I, most mass migrations took place without visas, quotas, asylum status, smuggled illegals, or security barriers. Since World War II, all mass migrations have taken place under those circumstances. Imagine how much bigger world mass migration would be today were we still living in the age of unrestricted migration that characterized the first global century before 1913. Twice as big? Three times? Five times?

While world migration has surged, the labor market quality of these immigrants has declined. For example, U.S. immigrant males earned 4.1 percent more than native-born males in 1960 but 16.3 percent less in 1990. Immigrants always suffer an earnings disadvantage before they assimilate, but their initial wage (relative to that of the native born) deteriorated by 24 percentage points between 1960 and 1990. Although the average educational attainment of immigrants improved, it did not increase as rapidly as that of the native born. The percentage of newly arrived immigrants with only a high school education or even less schooling was 5.6 percentage points higher than the better-educated native born in 1970, but 20.4 percentage points higher in 1990, an almost fourfold increase.

Most of this decline in immigrant quality was due to changes in the source country composition (Table 2), and it reflects four massive shifts in world migration patterns over the half century since World War II. The first shift involved the decline of European emigrants, part of which can be explained by the resurgence of migration within Europe (including Turkey): foreign European nationals increased from 1.3 percent of the western European population in 1950 to 10.3 percent in 2000. The rise would be even higher if it included the foreign born who had become naturalized.

More recently, western and southern Europe have become destinations for immigrants from Asia, the Middle East, and Africa. And, since the collapse of the Soviet Union in the 1990s, western Europe has also absorbed immigrants from the east, including from the former Soviet republics. As a result, annual net immigration into the European Union has soared since the 1980s: it now surpasses that of the United States and would exceed it by even more if illegal immigrants were included.

The second shift involved emigration from eastern Europe. This traditional east-west European flow has a long history but was stopped cold by postwar emigration policy in the centrally planned economies. Things changed dramatically in the 1980s when Poland and Romania opened up and even more dramatically when the Berlin Wall fell in 1989. Emigration from these transition economies increased fivefold between 1985 and 1989 and exceeded a million a year until 1993, when it eased a bit. In any case, Europe seems to have reestablished its old east-west migration tradition.

The third shift involved the transformation of Latin America from a major emigrant destination to a major immigrant source. The first global century leads us to expect that poor, low-wage, agrarian countries should send out more emigrants as they industrialize, but at some point they should start to retain their own and receive immigrants as they continue to industrialize and wages rise. Latin America is an exception to the rule: in 1960, it hosted 1.8 million (net) immigrants; in 1980, it sent abroad 1.8 million (net) of its own. The explanation for this unique regime switch appears to be Latin America's much richer and faster-growing northern neighbor.

The fourth and biggest postwar shift—which repeats the migration life-cycle experience of the first global century—involved Asian, African, and Middle Eastern immigrants, whose numbers rose from a trickle to a flood. Early industrializations and demographic transitions unlock the migration poverty trap and unleash a surge of emigration. Thus, the East Asian "miracle" first fostered an emigration surge, which then slowed, peaked, and subsequently declined as modern development ensued. The Middle Eastern life cycle has been delayed, as has the region's development. In Africa, where per capita income growth over the past half century has been so disappointing, the life cycle has been delayed even more.

In the first global century, demographic booms and early industrial revolutions generated an emigration surge from poor countries, and demographic busts and mature industrial revolutions generated a fall in emigration from now-richer countries. Falling transport costs and the effect of remittances amplified these forces, which slowly reduced positive selection: the really poor could finance the move only late in the first global century, as their incomes at home rose and as the cost of passage fell. Exactly the same forces have also been at work in the modern era although they have been strengthened by policies. In the United States, policies included the 1965 abolition of the country-of-origin quotas (and Asian bans), the shift to a worldwide quota, and the emphasis on family reunification as a key criterion for admission. Australia, Canada, and other industrial countries also leveled the source country playing field, but the effects on immigrant composition were not quite as dramatic as in the United States.

What can we expect?

Do the two global centuries of migration—before 1913 and since 1950—offer any insights into the future of global migration? It seems to me that they do.

While the poorest have never been part of any mass migration, it is clear that the European emigration of the 19th century diminished poverty there. Indeed, the living standards of host and sending countries converged during those decades, and the mass migrations did most of the convergence work. That is, world mass migration was much more important in contributing to convergence than were booming world trade and booming world capital markets in the first global century. If the same cannot be said of modern Asia, Africa, the Middle East, and Latin America, it is not because the impact of world capital markets and world trade are any more powerful, but rather because the emigrations are so much smaller relative to the huge populations that send their citizens to high-wage host countries. That is, compared with host countries, third-world sending countries have vastly bigger populations than did sending European countries before 1913. Thus, the same host country immigration rates today imply much smaller sending country emigration rates than they did a century ago.

In the first global century, emigration raised living standards in poor sending countries a lot. In the second global century, emigration could raise living standards in poor sending countries, but typically it does not. Why? First, successful development in poor countries today depends far more on rapid productivity growth and catch-up at home; second, today's rich countries no longer have open borders.

If more can be gained from world mass migration today than in the first global century, why are so many potential migrants kept out of industrial countries? In large part, the answer has to do with economic adjustment in the host countries and who is doing the adjusting. Thus, it has to do with the economic damage done to low-skilled native-born workers and their political clout. These factors played a central role when the United States, Australia, Argentina, and other overseas high-wage countries retreated from unrestricted immigration before World War I. They play the same role today. Modern immigration restriction also has to do with immigrants' net fiscal impact, who pays for it, and the political clout of those taxed. This issue did not arise during the immigration debates of the first global century because the welfare state did not yet exist.

Still, migrant demand for entrance into high-wage economies will not grow unabated. Indeed, it is unlikely to grow as fast over the next quarter century as it did in the previous quarter century. As the underlying transitional forces that have driven the surge in third-world emigration—their demographic and industrial revolutions—die out, so will the pressure to emigrate. That stage has already been reached in most of East Asia and much of Southeast Asia, regions that have completed their growth miracles. And spectacular growth in China and India ensures that it will soon be reached in Asia's two most populous countries. I believe that this stage will soon be reached even in slower-growing Latin America and the Middle East. Africa has yet to release a mass emigration on world markets and remains a wild card. Population aging in the postindustrial part of the world may increase the demand for immigrant labor, but a growth slowdown in host countries is likely to offset it.

My guess is that the next major shift in global migration will be a pronounced relative rise in migration within the third world (south-south migration) and a pronounced relative fall in migration between the third world and the west (south-north migration).