INTERNAL MIGRATION

Internal migration is defined as a change in permanent residence, typically of a year or more in duration, within the boundaries of a country.

Long-Distance Migration and Residential Mobility

A distinction is made between long-distance migration and short-distance migration in which the latter is referred to by the more specialized term residential mobility. Long-distance moves typically are operationalized as movement across broader areas, such as metropolitan areas or states and in some cases counties. They reflect movement across labor markets and often are associated with changes in economic conditions in those larger areas. Residential mobility typically occurs within the same labor market and is associated more frequently with neighborhood and housing considerations. The distinction is important because the two types of internal migration occur at different frequencies, are associated with different kinds of explanations, and often display different selectivities with respect to individual social and economic attributes (Speare, Goldstein, and Frey 1975; Long 1988).

Data Sources

In the United States and in most developed countries the basic sources for measuring both kinds of migration derive from nationwide censuses, population registers, and large surveys. In several European countries with population register traditions (e.g., the Netherlands and the Scandinavian countries) annual migration measures for localities can be traced back for many decades.

In the United States the decennial census provides the most comprehensive source of migration data based on the five-year fixed-interval migration question ("Where did you live five years ago?") that has been included in censuses since 1960 and also was included in the 1940 census. This question allows the calculation of a variety of migration measures, such as mobility-incidence rates, and measures of in-migration, out-migration, net migration, and migration streams for places, counties, metropolitan areas, and states (Shryock 1964). These measures can be cross-classified by an array of social and economic attributes available from the census and can be used to ascertain selective migration patterns.

These census data are limited by the restricted reference period of five years before each decennial census. For example, the 2000 Census permits the assessment of migration over the 1995–2000 interval but not over the 1990–1995 interval. Another U.S. Census item relevant to measuring internal migration over longer historical periods is the question on the respondent's state of birth. This question has appeared in every decennial census since 1850 and can be used to assess long-term and current internal moves across states as well as "return migration" when cross-classified by current residence and residence five years before the census (Long 1988).

The U.S. Census Bureau's Current Population Survey is another important source of migration data. It contains a fixed-interval one-year migration item and is used to assess time-series patterns of migration frequency and selectivity with respect to social and economic differentials. Time-series estimates of internal migration at the county and state level also are produced by the U.S. Census Bureau. Beyond these government sources, several national panel surveys conducted by universities and research organizations have been used to infer migration patterns for specific groups. In addition, administrative records collected for other purposes (e.g., Internal Revenue Service data) can be employed to examine migration patterns.

Reasons for Moving: The U.S. Case

The long-distance/short-distance migration dichotomy is reflected clearly in the frequency of movement in the United States Because long-distance, inter—labor market migration occurs only a few times during the life course (e.g., the move to college, the first job, retirement), the annual rate of interstate migration is relatively low (3.3 percent in the period 1999–2000) in comparison to the annual rate of movement within counties (9%). Overall, about 16 percent of the U.S. population moves in a given year, and only about one in five of those moves is across state lines.

The preponderance of local moves is a response to household and family changes as well as changing needs in regard to neighborhoods and homes. Among within-county moves in 1999–2000, 64 percent were for housing-related reasons and 26.5 percent were for family-related reasons. Only 6 percent of those moves were motivated by work and job-transfer considerations. In contrast, 47 percent of moves between counties undertaken by persons with postcollege training are related to work (see Figure1) (Schacter 2001). Climate and natural amenities are becoming increasingly important in motivating long-distance moves, especially among the retired population (Gober 1993).

By far the strongest selectivity differential associated with both long-distance migration and residential mobility is a person's age. The incidence of making each kind of move is highest for persons in their early to middle twenties and then declines precipitously during the thirties and forties, with a sometimes small upturn in the early retirement years. Among persons age 20 to 24 in the United States about one-third make a move of some kind in any given year (see Figure 2) This is twice the overall migration rate for all ages combined and reflects important life-cycle transitions such as moving to a college or to a new job (among long-distance migrants), marriage, and moving out of the parental home (among local movers).

Most other selectivity differentials are more specific to either long-distance or short-distance movers. Among long-distance movers there is strong educational selectivity in movement. College graduates, who are likely to be in a national labor market, show higher rates of movement than do those with lesser educational attainment. Among local movers there is a large difference between homeowners and renters: Homeowners tend to form long-term economic bonds to a particular location.

The recent influx of foreign-born populations to the United States has had both direct and indirect effects on internal migration. Locally, foreign-born residents tend to make several additional internal moves after they initially settle. Indirectly, the influx of foreign-born populations in selected "port of entry" metropolitan areas has tended to precipitate out-migration among established native-born residents (Frey 2003).

Explanations for Migration

Explanations for migration can be divided into two essential classes: those which explain individual decision making and those which explain aggregate migration patterns across geographic regions. Individual decision-making models of long-distance migration tend to be formulated around economists' cost-benefit model. This model assumes a rational decision-making process that weighs the economic and noneconomic costs and benefits of making a move. It has to be modified for particular population groups such as retirees and the college-bound population and for individuals in particular statuses, such as single-earner husband—wife families. These mobility models of decision making make a distinction

FIGURE 1

between "the decision to move" and "the choice of destination" in which the former decision implies some kind of disruption in family status or housing need (Speare, Goldstein, and Frey 1975).

Aggregate models of long-distance migration can be used to explain net migration levels for specific areas, migration streams across pairs, and matrices of areas. The latter models tend to have a strong geographic base related to the classic "gravity model," in which migration is directly related to the number of opportunities at a destination but inversely related to the distance between the origin and the destination (Speare, Goldstein, and Frey 1975). This model has been modified to take into account various economic and noneconomic opportunities. Aggregate models that explain net migration as a dependent variable take both cross-sectional and time-series forms (Greenwood 1981).

Consequences of Internal Migration

The consequences of internal migration have been addressed from a variety of disciplinary perspectives. The demographic consequences of internal migration processes for spatial population change are treated in a formal demographic model developed by Andrei Rogers (Rogers 1995). Rogers and his associates at the International Institute for Applied Systems Analysis also have developed techniques, including model age—migration schedules, for making population projections of subnational areas that take explicit account of internal migration streams.

INTERNAL MIGRATION

Long-Distance Migration and Residential Mobility

Data Sources

Reasons for Moving: The U.S. Case

Explanations for Migration

Consequences of Internal Migration

BIBLIOGRAPHY