INTERNAL MIGRATION

Migration is the relatively permanent movement of individuals or groups over varying distances to change places of residence; permanence and distance are its major defining dimensions. Internal migration occurs within the boundaries of a given country. (International migrants, not considered here, are called immigrants.) Internal migration, therefore, is a type of geographic mobility status.

DEFINITIONS

The following definitions are standard in the field of social demography (Bogue 1985):

Mobility status. A classification of the population based on a comparison between the place of residence (destination) of each individual in a census enumeration or survey and the place of residence (origin) at some specified earlier date. Mobility status in terms of the distance of the move falls into four main categories: nonmovers, local movers, intrastate migrants, and interstate migrants.

They may be examined more specifically in the list below:

I. Nonmovers, or nonmobile persons, live in the same house at the time of the census as at the date of origin.

II. Movers, or mobile persons, live in a different house and are further classified as to where they were living at the earlier date.

a. Local movers are mobile persons who live in the same county at census time as at the date of origin.

b. Internal migrants are mobile persons who live in a different county at census time than at the date of origin. Internal migrants may be further subclassified:

1. Intrastate migrants live in a different county but within the same state.

2. Interstate migrants live in a different state.

3. Interregional migrants live in a different geographic division or census geographic region; they are also interstate migrants.

Mobility interval. The lapsed time between the date specified for previous residence and the date of enumeration is usually either one year or five years. Recent census enumerations specify five years, and the Current Population Surveys have specified intervals of one, two, three, four, and five years.

Metropolitan mobility. A system of subdividing mobile persons into categories by place of residence at the beginning and the end of the mobility interval and, according to metropolitan statistical areas (MSAs), is as follows:

Mobility rates. The number of persons in a specified mobility status per 100 or 1,000 in the population of the area in which they resided at the end of the mobility interval is a mobility rate. Such rates may refer to any of the categories of nonmobile or mobile persons specified above. Mobility rates may be specific for age, race, sex, or other traits. The denominator may also be the origin date or the midpoint of the migration interval.

Migration flows. The key distinction of flows is that either the origin or the destination is unknown. There are two types of flows:

1. In-migration is comprised of migrants arriving at a particular place of destination, with no reference to the place of origin. In-flows could also arrive at specified types of places, such as central cities or metropolitan areas.

2. Out-migration is comprised of migrants departing from a particular area, with no reference to the place of destination. Outflows may also depart from specified types of places, such as places outside MSAs or suburban metropolitan rings of MSAs.

Migrations streams. These connect an origin to a destination. There are three types of migration streams:

1. Specific streams. Streams that connect particular places within a category, such as streams between specific cities, counties, states, or regions. This is the major use of the term.

2. Typological streams. Streams that connect types of places, such as streams between all central cities and suburbs in a state or the nation.

3. Counterstreams. When a stream between two places endures, it usually generates a counterstream, a smaller stream in the opposite direction. The stream and counterstream are referred to as an exchange.

Net migration. This is the difference obtained when the number of out-migrants is subtracted from the number of in-migrants in a particular place or type of place. A location that experiences a loss of population through migration is said to have a negative net migration; one that gains population through migration has a positive net migration. Because of its birth and death rates, an area may have a negative net migration and continue to have a growing population. There is no such thing as a net migrant, however.

Return migration. The census contains an item that identifies the state of birth. Return migrants are those persons who return to their state of birth during the mobility interval. There is no way of knowing how long they have been away from their state of birth when they return.

WHY STUDY MIGRATION?

Migration is important to social scientists because an increase or decrease in the size of a population, due to excess in- or out-migration, causes many social conditions to change. Community infrastructure, such as highways and schools, may become overburdened due to population growth, while public services may become difficult to maintain when population declines. Furthermore, social scientists study the equilibrating effects of population movement on national and regional economic systems. Growth or decline in the local economy is an incentive for people to move, which redistributes the population to balance the system.

The ability to predict the impacts of population growth or decline on the institutional sectors of a community and the ability to understand regional population dynamics, of course, provide many practical benefits to government and business planners.

MIGRATION RESEARCH

Net migration rates before 1940 were estimated using a survival-rate method. This method takes the population in one census as its base. It adjusts the number by adding births and subtracting deaths during the next decade. The amount of population change not accounted for is attributed to migration (Bogue and Beale 1961). The 1940 census was the first to include a mobility item. It asked where persons lived five years earlier. In 1950, after World War II, there was so much population movement that a one-year interval was substituted in the census. In 1960 the five-year mobility interval was restored and has been retained in subsequent decades. Because of these measurement changes, the 1960 and 1970 censuses were the first from which decade changes could be derived. Thus several landmark studies appeared in the 1960s, breaking new ground and setting patterns for future migration research (Long 1988). Shryock’s (1964) work showed the importance of studying gross migration flows in addition to the prevailing dependence on net migration. Lowry (1966) introduced econometric modeling to migration research. Finally, Lansing and Mueller (1967) helped introduce survey approaches to analyzing internal migration.

U.S. MOBILITY

Americans are unusually mobile (Bogue 1985). Only Canada and Australia have populations as mobile as that of the United States. In a single year, from March 1995 to March 1996, 17 percent of U.S. inhabitants moved from one domicile to another and about 6 percent changed their county of residence. At current mobility rates, average Americans live at fourteen different addresses during their lifetimes. Of these thirteen moves, three are as a dependent moving with parents and ten are of their own volition. People who have lived their entire lives at the same address account for no more than 2 or 3 percent of the adult population. Perhaps no more than 10 to 15 percent of people spend their entire lives in their county of birth.

When using the five-year mobility interval, mobility rates are not five times as large as those for a single year because persons who move several times within the interval are counted only once. Nearly one-half of the population is mobile over a five-year period, and more than one-fifth are migrants. Since 1980 there appears to have been no diminution in the tendency to migrate, but there has been an apparent reduction in local mobility.

One can discover contradictory findings in mobility literature. These contradictions are often due to the specific databases under analysis. Some databases use mortgage data and leave out renters; others, such as the Annual Housing Survey, use households; and some, such as most census publications, use individuals as units of analysis, each database giving somewhat different results. In addition, some data sources offer little information on the characteristics of migrants. The individual master file of the Internal Revenue Service includes state and county migration data but no personal characteristics, and several large moving companies provide data on their customers also without personal characteristics (Kahley 1990).

Reasons for Migration. Migration may occur in response to changing economic, social, or political conditions. Push factors are conditions in the sending population that impel or stimulate migration. Conditions that attract in-migrants are classified as pull factors (Ravenstein 1889).

Declining economic opportunities, political instability, or the weakening of place ties may stimulate out-migration. Expanding economic opportunities, potential for advancement, the presence of family members and friends, or previous vacationing or residential experience tend to attract migrants. Not surprisingly, rural communities with high birthrates and regions with limited opportunities are areas of high out-migration, whereas urban, industrial regions and communities with expanding opportunities tend to have high in-migration (Prehn 1986). Marriage, divorce, increasing or decreasing family size, and housing adequacy top the list in surveys. A sizable majority of respondents to the Annual Housing Survey reported housing or family dynamics as reasons for their move (Gober 1993).

The average age at which young adults leave home declined from the low twenties to the upper teens between 1920 and 1980, and then the median age began to rise again. These trends mirror another one; for the Vietnam cohort of young adults forward, those who return to live at home at some time holds at about 40 percent. About 25 percent had moved back in earlier cohorts. The expectation of a permanently empty nest for parents of young adults now seems a less certain one (Goldscheider and Goldscheider 1994).

Zelinski (1971) proposed a macro-level three-stage model of national internal migration. First, with the onset of modernization, the overall level of migration increases, primarily in the form of rural-to-urban moves. Second, as industrialization and modernization spread to more regions, migration may continue to increase; improved transportation and communication increase the availability of information and decrease the uncertainty of moving. Interurban moves become the majority of all moves. Finally, at advanced stages, when lev-el-of-living differences among areas have diminished, there may be more urban-to-rural movement and more ”consumer-oriented” migration toward warm climates or locations with other amenities (Long 1988).

Differential Migration. What population characteristics predict migration? Characteristics that indicate less entanglement with social obligations, greater need for employment, and higher job skills are good predictors. Men are more mobile with respect to residence than women, although the difference is small. The single migrate at higher rates than the married. For several decades, blacks have been more mobile than whites. However, in 1980 whites migrated at higher rates than blacks, although blacks continued to be more mobile locally. Hispanics migrated internally at a rate between those of the black and white populations. Persons with higher levels of education are more likely to migrate than those who are less well educated.

Age and Mobility. The shape of the age profile of migrants in the United States has been consistent for decades, changing only gradually over time. The younger children are, the more likely they are to migrate. The migration rate of children bottoms out in the early teens and does not increase rapidly until the late teens. More than one-third of Americans in their young adult years, ages twenty to twenty-four, the peak migration years during the life course, moved at least once between 1982 and 1983, and nearly one-half of this mobility was migratory. Not surprisingly, this age corresponds with college graduation and marriage for many. The increasing age of children in the home, particularly once they begin their formal schooling, dampens the attractiveness of migration for parents. The age-specific migration rate declines slowly at first, then more steeply until age thirty-five, after which it slowly declines throughout the middle years to a life-course low point just before the retirement years. The retirement migration hump between ages sixty and seventy is small by comparison to the early adulthood migration bulge. The final increase in age-specific migration rises at the end of life and relates largely to health issues. The elderly as a broad category are only about one-half as mobile as the general population.

MIGRATION AND REGIONAL DISTRIBUTION OF POPULATION

Three large interregional flows of internal migration have been occurring in the United States for many decades.

Westward Movement. For a long time, there was a high-volume flow of persons into the Pacific region, principally California, as well as a high-volume flow into the mountainous southwestern states. The 1970-1980 decade had a higher volume of westward movement than any previous one. Mountain states that previously suffered losses all made positive gains, and Colorado, Nevada, and Arizona continued the large gains of the previous decade. In the 1990s, there was a net flow out of California, largely to other western states, reversing a long-term trend for that state.

Northward Movement from the South. The southern region lost population heavily between the close of the Civil War and 1950. Industrial centers in the northeast and east-north-central regions absorbed a very large share of the migrating population. Both white and black migrants flowed along these channels in great numbers. Some southern states, however, particularly Florida and Texas, were exceptions. Between 1970 and 1980 the net outflow from the South completely disappeared. Those who left the South preferred the West to the North as a destination, and in-migrants to the South balanced the out-migrants. Every state in the northeast and north-central regions suffered a net migration loss during the decade, resulting in a major regional migration turnaround (Bogue 1985). By 1990 there were no net flows out of the South to other regions, but the Northeast, Midwest, and West all contributed to the southern region (Gober 1993).

The Southward Movement to the Gulf Coast and the Southern Atlantic Seaboard. The entire Gulf Coast, from the mouth of the Rio Grande in Texas across the coastal portions of lower Louisiana, Mississippi, and Alabama and on to include all of Florida, experienced much more rapid and intensive economic development than the southern and southeastern parts of the United States lying away from the coast. Although this trend is a very old one, it accelerated rapidly in the 1970s.

As of1980 there were only two regional migration streams instead of three: movement toward the South and Southwest and movement toward the West. The northeast and the north-central regions are the sources from which these migrants came (Bogue 1985). But in the 1980s, the South gained more through net migration than did the western states (Weeks 1996), a trend that accelerated by 1990. The geographic redistribution of knowledge-based industries of the information age carries in its wake a college-educated work force to the Sunbelt, including the South (Frey 1995).

Metropolitan Deconcentration. One of the macro-level processes that affects geographic mobility in our time is metropolitan deconcentration. Many nonmetropolitan counties in the United States experienced a slowing of population decline in the 1960s, and in the 1970s their net migration rates climbed above the break-even point, which signaled a genuine and widespread ”rural-urban turnaround.” Older people seem to have been in the vanguard of migration to nonmetropolitan counties; the turnaround for them happened in the 1960s rather than the 1970s. This reversal of a long-term trend of rural-to-urban migration is of great interest to demographers. Mounting evidence now indicates that although deconcentration continues in nonmetropolitan America as a whole, by the late 1980s metropolitan counties were outgrowing nonmetropolitan ones (Long and DeAre 1988). In the 1990s there is an uneven urban revival, with a few metropolitan areas with more flexible and diverse economies, mostly outside the Northeast and Midwest, gaining migrants. The new dominance of the suburbs over the central city is key to metropolitan deconcentration in the 1980s and 1990s. During this period, the suburbs are capturing the bulk of employment and occupational growth (Frey 1995).

RETIREMENT MIGRATION

Demography traditionally tends to focus on youthful migration, and labor-force migration in particular. Increasing attention, however, is being given to non-labor-force-motivated migration, particularly to the migration of persons of retirement age (Longino 1996). For the elderly, interstate flows are highly channelized—that is, half of the interstate migrants, regardless of their origin, flow into only eight of the fifty states. Florida dominates the scene, having received about one-quarter of all interstate migrants aged sixty or over in the five years preceding the 1960, 1970, 1980, and 1990 censuses. Although Florida, California, Arizona, and North Carolina have different major recruitment areas, they are the only states that attract several unusually large streams from outside their regions. Florida and North Carolina draw primarily from east of the Mississippi River, and Arizona and California draw from west of it. Among the elderly, the special characteristics of the destination tend to be more important than the distance. Warm climate, economic growth, and lower cost of living are still important pull factors.

Distance selectivity of elderly migration has been studied. Local movers are generally not as economically and socially well-off as nonmovers, and migrants are more well-off. Interstate migrants tend to have the most positive characteristics.

Permanence is an important but difficult dimension of migration to study. The census assumes that one’s ”usual place of residence” is not temporary. In reality, however, much of the migration among older people may be temporary. So far, studies of elderly seasonal migrants show them to be relatively advantaged, attracted by non-labor-force issues such as climate, cost of living, and the locations of family members and friends.

Metropolitan-to-metropolitan migration predominates among the elderly. Of the one-third who changed environmental types, no increase occurred between the 1960 and 1980 censuses in the proportion moving out of metropolitan areas in each decade. However, the movement in the opposite direction, up the metropolitan hierarchy, declined, both among older intrastate and interstate migrants. The net difference made it appear as though the flow from cities increased. Metropolitan-to-metropolitan migrants, especially those moving longer distances, tend to have more income, to be married, and to live in their own homes. A higher proportion of nonmetro-to-met-ro migrants is older, widowed, and living dependently, especially with their children. Coding revisions in the 1990 census made updates of these comparisons impossible to calculate.

The cycle of migration for a job when one is young and returning to one’s roots after retirement is an appealing notion to theorists. In contrast, Rogers (1990) demonstrated that elderly persons are not any more likely to return home than are the nonelderly; in fact, the probabilities of return migration by the elderly are lower than those of the general population, even after controlling for the different mobility levels of the two populations. There is wide state variability, however. The southeastern region is unusually attractive to older return migrants, and return migration is uncommonly high among the older black population moving into that region. Some evidence from the 1990 census shows that regional return migration patterns are shifting away from the Sunbelt states. Some migrants, apparently, return to their home states after an earlier retirement move (Longino 1995).

Some dubbed retirement migration as the growth industry of the 1990s. The amount of income transferred between states through retirement migration is quite substantial. Not surprisingly, economic development agencies are mounting efforts to attract mature migrants. This is leading to sharp competition among destinations for these migrants as new residents. The impact of elderly migration as a social phenomenon has yet to generate enough research to provide definite statements.

INTERNATIONAL COMPARISONS OF INTERNAL MIGRATION

Little research exists to compare countries on internal migration because measures, data sources, and units of analysis differ widely among nations. Consequently, international organizations have not published compendia of national comparative data on migration as they have on fertility and mortality. In addition, certain types of cultures conceive internal migration differently. In some small countries, such as England, lack of new housing stock limits residential movement. Migration is also limited in countries such as France, where transportation routes primarily connect the peripheral towns to a central national capital for historical reasons. Conversely, internal migration is amplified and culturally expected in nations of immigrants with widely dispersed regional centers and major cities, such as the United States, Canada, and Australia.

Nonetheless, existing studies provide some tentative generalizations that compare internal migration in the United States with that in other countries (Long 1988). The U.S. national average for moves is higher than that of most other countries because (1) cities in the South and West are growing; (2) a relatively large minority of people who repeatedly move elevates the U.S. average for lifetime moves above that for most other countries; and (3) during the 1980s and 1990s the baby boom generation in the United States has moved through the life-cycle stages that have the highest rates of geographic mobility.

Comparative studies also give attention to older migrants, although their mobility rates are lower than for the young. Rogers (1989) argues that as the populations of industrialized nations age, the internal migration patterns of elderly persons will change. Elderly migration levels are low in countries in the first stage of this population transition. In the second stage of the transition, large, long-distance flows to particular principal destination regions appear. The third stage continues to exhibit large numbers of elderly migrants, but their moves now include a significant number of short-distance moves to more dispersed inland regions. Rogers and colleagues (1990) argue from comparative data that England is in the third stage, the United States is transitioning between the second and the third stages, Italy is well into the second stage, and Japan is in the first stage.

Since 1970, for most developed countries, population aging has brought declining national rates of internal migration (Long 1988). For the United States, the decline appears to be greater for local moves than for long-distance movement. Urbanization was the dominant redistribution trend in the 1950s in fourteen European countries studied by Fielding (1989). The relationship between net migration and settlement size began to break down, however, in the 1960s—first in the countries of northwestern Europe in the mid-1960s, then in the countries and regions of the southern and western European periphery through the 1960s, and in the case of Spain, into the 1970s. By the 1970s, most of the countries of western Europe were recording counterurbanization, where the net flow was away from cities and toward small settlements. That counterurbanization became less dominant in the early 1980s but was not replaced by urbanization. Only in West Germany and Italy did the counterurbanization relationship persist. The United States experienced a similar pattern of long-term urbanization, reversed in the 1970s and then nearly reversed again in the 1980s (Frey 1990).

PREDICTING FUTURE MIGRATION

Migration rooted in labor movement will change in the future as the geographical basis of the economy changes. Robust new industry will attract migrants. Such developments in the southern region may extend several more decades into the future. On the other hand, migration not related to the labor force, such as retirement migration, is more sensitive to lifestyle issues. Eventual overcrowding, which results in a decline in the quality of life of local residents, will tend to discourage retirement migration. Florida’s dominance of retirement destinations in 1990 lost 2 percent of the migrant retiree market.

Migration for better jobs increases in times of economic expansion. Thus, studies of the late 1990s may find migration to have increased in response to an improved economy. Other trends could also increase migration rates. First, the age composition is always shifting. In the 1980s, there were more persons in the twenty-to-thirty age range, their prime mobility years. The baby boom generation has a lower migration rate for long moves than others. However, because of its large size, a large number of baby boomers migrated. In the 1990s, the incidence of moving will likely dampen as baby boomers age and leave their prime mobility years. Second, the rising level of education may increase migration. Each new cohort of adults has a higher level of education than its predecessor. The third factor, household change, contains counterindicators. Married couples are increasingly likely to divorce, a situation favoring migration, but at the same time there are more dual-career couples in the population, a situation favoring nonmobility (Long 1988).

As we have seen, many factors motivate migration. These factors need further study, which will certainly generate new research hypotheses to be tested by migration researchers in the twenty-first century.