Migration and Local Labor Markets
Stephan J. Goetz, Penn State University


1. Migration–Definition and Illustration

Why do people live, work, and play where they do? How much, why, and between what points do people migrate, both temporarily and permanently? These are rather simple questions, but the answers are much less clear and do not fall neatly into one field of human knowledge.   Graves and Clawson 1981, 363.

Migration is defined here as the movement of people across county (and state) lines within the United States for the purpose of establishing a new place of residence. There is no universally agreed-upon, single definition of migration, and the definition used in this chapter is a compromise between including only migration across state lines and excluding changes in residence within a single county to define migration (see Figure I.1). Defining migration in this manner is largely a matter of convenience. The definition is consistent with that used by the U.S. Census Bureau to define migrants, and it allows us to take advantage of a current and comprehensive database on migration compiled each year by the federal government. Later on, we will revisit this question of defining migration in more detail.

Over the period March 1996 to March 1997, nearly one in six Americans moved, or 42.1 million people in one year. However, two-thirds of all movers stayed within the same county, while 34% moved across a county line–either within the same state (19%) or across state lines (15%). These numbers and more current data are available from the Census Bureau Web site. There are many different types of migration and factors motivating individuals’ migration decisions, as we will see later in this chapter.

Migration behavior is often studied within the context of job searches and, in fact, for many people it is an integral element of upward mobility in pursuit of the "American Dream." When people take new jobs, they may need to move to a different part of the country. Some researchers define migration as occurring whenever the distance involved in a residential move (change in homes) is so large that it is no longer possible for the mover to commute to the old place of work (Clark 1986). This definition can be turned around to argue that migration has occurred when a change in jobs involves such a large distance that the individual also has to move to a new home closer to the new job. This highlights the interdependence between places of residence and employment for working citizens. Yet other authors argue that migration has occurred whenever the move takes a worker into a new labor market. Local labor market areas are discussed in Section V.

The matrix shown in Table 1 allows us to separate commuting from changes in residences and formal migration. In fact, commuting can be viewed as a substitute for migration, which can also be regarded as a nontemporary change in residence to an area that is not in close proximity to the original residence. These important definitions are revisited later in the section on empirical issues.

Table 1: Types of Migration–Change in Residence vs. Job Location
Note: click on the linked items for further explanations


What happens to the location of the residence?

No change

It changes

What happens to

the location of

the job?

No change


® Residential migration

It changes

Change of job
­ location

Intra-regional migration

Inter-regional migration




®Labor market®

Source: Adapted from Klaassen and Drewe, 1973, p. 22.
Note: * refers to the local submarkets of a region; ** the regional submarkets of the nation.

The migration decision can be viewed as evolving in three stages. Rossi (1980) proposed the taxonomy shown in Table 2 for that purpose. The first stage involves the decision whether or not to move, perhaps in response to unhappiness with the present job or home, or in response to a layoff forcing a search for new employment locally or nonlocally. The second stage involves gathering information about other places, including information about job and housing opportunities. In the last stage, one place is selected out of the various places considered in the second stage. Whether the migrant considers one migration target or many potential targets–including the current location–is an important consideration, and it has implications for the modeling of migration decisions.

Table 2: Migration Decision Stages

Stage I

Decision to move away from present place of employment and residence

Stage II

Search for a new residence and location (simultaneous)

Stage III

Selection of a new residence and job site from many

Reason for changing:

Relevant considerations:

Reasons for choice:

External circumstance: e.g., job loss, employer relocates

Information channels (formal and informal): head hunters, realtors, the WWW

Only one alternative, which meets migrant’s expectations, is considered

Dissatisfaction with current residence or job

Characteristics of new residence and job location

Several alternatives are considered; one is selected based on its relative attractiveness

Source: Adapted from Rossi (1980, 222)

Historically, one of the most significant migrations of population in the United States has been the movement of people from rural to urban areas. To appreciate the magnitude of this shift, consider the fact that in 1800, 94% of the U.S. population lived in a rural area, compared with 60% in 1890 and only 25% in 1990! Greenwood (1985, 922) refers to this phenomenon as "metropolization" and cites a number of studies on this subject. The heartland states have been hard hit by population out-migration since the turn of the century, and they were particularly affected by the farm financial crisis of the 1980s. Perhaps ironically, the appeal of living in rural areas is very high today in developed countries, but it continues to be low in developing countries. In those countries, rural-to-urban population migration causes enormous problems, and cities are ill-equipped to deal with the influx of people, for example, in terms of providing basic sanitation infrastructure and public services.

Migration patterns vary by sociodemographic characteristics such as race and age of the population. A significant migration pattern for a particular racial group has been the movement of African Americans from the southern United States to northern states such as Michigan in the twentieth century. As the population ages, an increasingly important migration pattern is the movement of retirees from New England and the upper Midwest to sunshine states such as Florida and Arizona. As the leading edge of the baby boom generation approaches retirement age, this trend is likely to become more pronounced. Another important form of migration in the United States has been the movement of people from downtown urban cores to suburbs (which is known as suburbanization or urban flight).

The history of U.S. migration is efficiently summarized in the adjacent map, showing the geographic centers of population in each decade starting in 1790 (obtained from 1998 Statistical Abstract, Section 1 on Population). This geographic center shows "the point at which an imaginary flat, weightless, and rigid map of the United States would balance if weights of identical value were placed on it so that each weight represented the location of one person on the date of the census" (U.S. Census Bureau 1998, 27). The map shows the mean center of population located in Kent County, Maryland (23 miles east of Baltimore) in 1790, and a migration of this center toward the West and Southwest in more recent history, resulting in a location in Crawford County, Missouri (10 miles southeast of Steelville) in 1990. Of course, this map shows not only migration by U.S. residents over time, but also the effect of immigration from foreign countries, including China and Mexico, and natural population growth (births minus deaths).

The recent Current Population Report–one of the most up-to-date sources of migration data–provides the following additional information about characteristics of migrants. Individuals in the 25-29 year old category are the most likely to migrate (16% of all migrants as defined above), followed by the 20-24 year old age group (13.7%). Characteristics of migrants are discussed in greater detail in Section III. Between March 1996 and March 1997, the population of the South grew by 391,000 people due to in-migration from the three other regions, and this region was the only one "with a significant net change due to internal migration" (Faber 1998, 1). Over the same period, "central cities lost 3.0 million people due to migration, while the suburbs gained 2.8 million, [and, in comparison,] 1.3 million people moved to the United States from abroad. Around 92% of these movers went to metropolitan areas" (1).

Table 3 provides information about population movement in 1996 among the four major census divisions of the United States (Hansen 1997). The Midwest, West and South each experienced a net gain of people due to migration, with the South gaining the largest number (151,000). The Northeast lost a substantial number of residents: almost 250,000 people migrated out of the Northeast and to other parts of the nation. Note that the table does not include immigration from foreign countries or natural population increases, so that these numbers do not add up to total population change in a region.

Table 3: Population Migration Among Four U.S. Census Regions, 1996, in Thousands

¯From      \        To®





Gross Loss

























Gross gain






Net gain (loss)






Source: Hansen 1997 plus author’s calculations.

Although relatively simple, Table 3 illustrates a few important points related to internal migration in a country. First, regions gain and lose people at the same time. Obviously, the in-migrants are not disturbed by whatever conditions caused the out-migrants to leave. Conversely, whatever factors attracted the in-migrants were not important enough to keep those who decided to leave in the region, at least not relative to whatever it was that attracted them to the new region.

A second key insight from Table 3 is that the net gains and losses of migrants across the four regions have to sum to zero. This is an important check for internal consistency in studying migration patterns across regions. Also, what matters in the end for regional pride is the net number of migrants: the difference between those coming and those going. For current migration data across U.S. geographic areas, visit William H. Frey’s Web site .

2. Migration–its Effects and Why it Matters

Because it is a selective process, migration can have important effects on both the supplying and receiving regions. Immigration and emigration frequently alter the age composition, sex ratios, literacy rates, and demography of the affected areas; they can also create social and economic problems.      Berry, Conkling and Ray 1987, 57

Population migration is important for many reasons–both public and private. Most notably, it is a key determinant of population change in an area, together with vital statistics such as the number of people who are born and die in the area over a specified period of time, such as a year:

Population Change =Births – Deaths + In-Migrants – Out-Migrants
=Net Births – Net Migration

For the sake of simplicity, we are ignoring foreign immigrants in this equation. However, migration is important not only because it affects total population numbers of a community but also because it can change the makeup or composition of the communities involved, as implied in the above quote. We revisit this important point later in the section discussing the characteristics of migrants.

Table 4: Vital Statistics for Selected States, 7/1/97 to 7/1/98


Population Change



Net Domestic Migration







New York

North Dakota


West Virginia









































Note: Births, deaths and net domestic migration do not add up to population change because net international migration is excluded.

In the above table, note that even rapidly growing states–such as California–can still be losing population to domestic migration. Also, more people died in Florida over the period shown than migrated in, on balance, from other states. Kentucky and Louisiana had roughly comparable numbers of births and deaths, but Louisiana lost migrants to other states while Kentucky gained migrants. Gross out- and in-migration numbers are presented below in Section IV.

Economically booming areas usually experience an increase in population, while declining areas experience population outflows (see discussion question 4 at the end of this section). In the latter situation, net births in the community (the excess of the number of births over the number of deaths) are not sufficient to stem the net rate of out-migration from the community. Births and deaths are referred to as "vital statistics" by the government, and migration statistics are clearly also "vital" in the sense that they influence population change over time, and thus the economic fortunes of regions.

Americans were reminded recently of one of the most profound and far-reaching population migrations affecting this country in the twentieth century. The widely-circulated Sunday newspaper supplement, Parade, carried the headline "Where Have All the Small Towns Gone?" in its December 13, 1998 issue. The subtitle of the article was: "Communities throughout rural America have been disappearing." San Juan County, Colorado, has the distinction of having lost the highest percentage of population of all counties between 1990 and 1997: 25.2%, according to Parade magazine. This is perhaps ironic, because Colorado recently also boasted the county with the most rapid population growth of all U.S. counties. The county with the next-highest decline in the share of population between 1990 and 1997 was Burke County, North Dakota (–22.3%), followed by Geary County, Kansas (–16.9%). Students of migration behavior attempt to identify systematic economic, social and demographic reasons why some places gain new populations at high rates, and why others lose residents.

When people move, they do so for private or personal reasons. For example, they may move to new and better jobs, as we saw earlier, or to live in an area with a better climate or more pleasant environmental amenities. However, the collective decisions of individuals to move have public or community-wide effects both on the places from where they come and on the places to which they move. Obvious examples include a declining tax base to fund minimal levels of public services such as schools, along with falling property values, in places rapidly losing population. This is explored below in more detail.

Local businesses lose revenues as people move away from an area, and business owners may themselves be forced to shut down their operations or move elsewhere. In addition, city and regional planners have to have some idea of population change over time to plan for supplies of public infrastructure and services. Planning for adequate school facilities construction, so that there is neither an oversupply nor an undersupply of buildings, is a difficult challenge facing public planners in both urban and rural communities.

Economic costs associated with congestion can escalate rapidly in communities receiving in-migrants at high rates. Or, property values can rise rapidly in downtown communities as wealthy in-migrants follow the current fad of moving out of suburbs and back into cities. Often the long-term residents of the cities, who may have limited incomes or fixed incomes if they are living on pensions, find their property values and tax burdens rise significantly, and perhaps to levels they can no longer afford to pay. This process is known as gentrification. Fitchen (1994) argued that a similar process is occurring in rural areas located on the fringes of metropolitan centers. Wealthy individuals, attracted by the rural amenities, are driving up land and housing costs, in the process pushing low-income families further into remote areas that become enclaves of the poor. Conversely, merchants and shopkeepers located in "booming" communities may become wealthy, at least in the short term, simply because they happen to be in the right place at the right time.

The fact that an individual’s decision to migrate has public and private consequences is one reason why migration is a fascinating topic of study. Population migration may accelerate in the information age as individual entrepreneurs become increasingly able to set up their businesses wherever they want by using cellular phones or satellite dishes to connect to the World Wide Web. In turn, entrepreneurs who correctly predict which areas of the United States will gain population, and which will lose, can convert such knowledge into profits.

One manifestation of the collective impact of individuals’ decisions to migrate is population change in different states over time. Figure 1.2 shows that the band of states stretching from Washington to Texas stands out remarkably for having experienced above-average gains (of at least 14%) in population between 1990 and 1997. Figure I.3 (available at suggests that virtually all of the population growth in the 10 fastest-growing counties between 1997 and 1998 was driven by net (domestic) in-migration. Migration patterns in the West tend to reflect the attractiveness of natural amenities appealing to decision-makers in the high-tech industries, on the one hand, and the desire to leave overpopulated areas of California, on the other. Georgia is a population growth pole in the South, with Atlanta experiencing rapid increases both in residents (see below) and problems associated with urban sprawl. Many people do not realize that hidden federal subsidies–primarily related to road construction–tend to encourage urban sprawl (Longman 1998).

Nevada led this group of fastest-growing states, with a phenomenal population growth rate of 39.5% between 1990 and 1997. This translates into a doubling of population approximately every 15 years! In Utah, rapid population growth (19.5% between 1990 and 1997) is determined not only by a higher-than-average propensity of families to have children in that state but also by natural amenities. The population moving into these areas creates tremendous opportunities for local businesses–and also attracts new businesses to take advantage of the rapidly growing population. An important research question in the area of population migration is whether people follow jobs across the country, whether jobs follow people, or whether both events occur simultaneously.

Often migration is driven by differences in employment opportunities across regions, which in turn are caused by differences in how business cycles and federal macroeconomic policies affect local economies (Chisholm 1990). This is a relatively underresearched area in regional economics and the study of migration (Muth 536 of Greenwood), although recent studies suggest there is increasing interest among researchers to examine this subject (e.g., Hervey and Strauss 1998).

2.1. Public Sector Issues

2.1.1. Net Out-Migration

Loss of population is perhaps the single-most widely recognized indicator that a community is in economic decline. Movers leave behind "large service infrastructures and capital investment that are increasingly difficult to maintain with declining population and tax bases" (Maynard et al. 1997, 131). Newspaper reports on school consolidations often lament the loss of smaller schools where students are less anonymous, receive more attention, and may have more opportunities to develop leadership skills. According to Department of Education Statistics, there were 370,000 schools and 119,000 school districts prior to World War II. Today only 87,000 schools and 14,900 school districts remain (Wall Street Journal, Nov. 13, 1998). Furthermore, according to the same article (A1),

...when rural schools close, the little towns that support them often do, too. Families move; the town’s identity fades. "The greatest loss is a high school," says George Chambers, who teaches school finance at the University of Iowa in Iowa City. A high school is usually the geographic, social and political center of town. When it closes, he says, "the focal point shifts"–to another town.

A more subtle, although equally important, public sector impact in the educational arena is that many communities, particularly in rural areas, are deprived of any returns on their investments in their young people if high school graduates leave the area for more remunerative employment elsewhere. This phenomenon has long been recognized in the international migration literature as a "brain drain." In this case, the returns to the human capital investment that occurred between kindergarten and grade 12, and that was at least in part funded out of local taxes, accrues to the community or area to which the individual has migrated.

Some observers have argued that communities (often urban) which receive the benefit of a youth trained elsewhere should compensate the community losing the youth. An objective measure of the benefit transfer is the discounted cumulative value of local income taxes and any occupational taxes paid by the individual. Of course, in a society in which individuals repeatedly move over their lifetimes, it would be rather expensive to track precisely to which communities compensatory payments should be made. Even so, it is important to note the existence of these implicit benefit transfers as one public consequence of private migration decisions.

Public policymakers argue that sometimes it is more desirable for people to move away from economically declining areas. This type of strategy necessitates the consideration of important issues related to those who leave and those who stay. First, one should consider whether those who leave the impoverished area have the skills needed to compete with job seekers elsewhere, or if it makes more sense to try and create local jobs for these potential migrants, perhaps with appropriate job training. Second, one should realize that to the extent that those leaving an impoverished area (rural or urban) are more mobile and have more marketable job skills than those who remain behind, migration tends to lead to a concentration of poverty. In particular, the people staying behind are–by definition–less mobile, and this lack of mobility tends to be correlated with a lack of skills and greater poverty. One study found that only 2% of workers who lost their jobs due to structural economic change (often these are the elderly and those with few skills) actually moved away to find employment elsewhere (Flaim and Seghal 1985).

When residents leave rural areas, the ability of local government agencies to provide services such as police and fire protection, and ambulances, can also decline along with the ability to fund education. The reason for this is that the relationship between population size and the quantity of services delivered is not constant (or linear). More specifically, the provision of a service such as police protection entails both fixed costs (police buildings, including jail cells, secretarial staff, 911 operators, etc.), which tend not to vary with population size – at least over certain ranges – and variable costs (police officers, uniforms, cruisers, etc.), which do vary with population size.

The key problem for communities losing population is that even though they can reduce or adjust variable costs (expenditures), by such strategies as reducing the number of police officers, they are not able to readily reduce their fixed costs. These fixed costs have to be spread among fewer and fewer taxpayers as a community loses population. This increases the relative fiscal burden on those left behind, and it is one of the major economic costs associated with the loss of residents from declining areas. Loveridge and Schmid (1993, 1162) have pointed to the possibility of encouraging foreign immigrants to move into rural areas to offset the negative consequences of population change.

2.1.2. Net In-Migration

In places receiving in-migrants, problems can arise when newcomers demand more of certain types of public services or new services than are available. This is especially a problem when newcomers have socioeconomic characteristics that differ from those of the resident population. For example, newcomers may on average have more small children, leading to pressure to build costly new schools. Another example is increased demand for social or medical services. Even if the fixed costs of providing these services are spread over a larger population base, taxes paid by longtime residents will likely go up, and this can create a resentment if these residents feel they should not be taxed to provide the new services.

Certain regions of the country have experienced very rapid increases in population over time. For example, in describing the experience of San Jose, California in Silicon Valley, Eberts and Stone write (1992, 5):

Much of the adjustment San Jose has faced has been in accommodating the streams of newcomers and the congestion and social problems that followed them. The area constantly confronted shortages–in workers, housing, and in adequate infrastructure.

As corroborating evidence, Inc Magazine’s twentieth anniversary issue reported (33) that a "single-story home, [with] brick exterior, two bedrooms, two baths, 1,688 square feet," which sold for $75,000 in 1975, sold for $900,000 in 1999. Even considering inflation, this is a phenomenal increase that illustrates one of the costs of rapid in-migration to an area, particularly one that has limited space for expansion.

In the summer of 1998, the city of Tooele in Tooele County, Utah, was attempting to forecast population growth into the future under three different scenarios of high (3%), medium (2.167%) and low (1.5%) annual rates of increase (Figure I.4). The medium scenario represents the actual rate of population growth in Tooele County over the period 1990 to 1996; the other two percentages are hypothetical. Figure 3 shows very clearly the vastly different impacts in the long run on total population in Tooele County 25 years into the future of different assumptions about this growth rate. Each has dramatically different implications for the demand for land and housing, schools and other infrastructure needs in the community. Even under the medium growth scenario, which is based on recent growth rates, population is forecast to double over the next 25 years.

Migration decisions also have aggregate implications for the average level of incomes both in the destination and the origination community. For example, Cromartie and Nord (1997, 40) report that "[d]uring the post-1990 nonmetro population growth spurt, the higher income of inmigrants compared with outmigrants increased overall nonmetro per capita income by an estimated $30 per year" and "[m]igration increased per capita income in roughly half of all non-metro counties."

2.2. Private Sector Decision Makers

2.2.1. Cost and Benefit Calculations

When deciding whether to move to a new place, individuals weigh the costs of moving against the benefits of moving. Most theoretical models of migration posit utility maximization as the central behavioral criterion. Costs and benefits of migration include both pecuniary and nonpecuniary or psychic factors. As we will see in subsequent chapters, many migration models compare the earnings of potential migrants at the destination and the current residence. If the earnings differential (wi – w0) is positive, where w is the hourly wage while subscript 0 denotes the current residence and i the destination, then the migrant, in principle, will make the move (seeSection II.2).

In reality, however, a number of other intervening factors will enter into the migrant’s decision. These include all of the costs associated with moving a household, finding a new job at the destination (if one has not already been secured), and finding a new home. Some of these costs are falling as the Internet makes it quite easy to find homes in cities far across the country from the convenience of a home computer. Other factors to consider include the probability of finding work – that is, earning a wage. In some studies this is modeled by including an unemployment rate (ui) at the destination, which can be used to calculate an expected wage: E(wi)=wi(1– ui), where the unemployment rate is expressed as a decimal. For example, a wage of $25 per hour, together with an unemployment rate of 5%, translates into an expected wage of only ($25.00*0.95=) $23.75.

Nonpecuniary factors can raise or lower the net benefit associated with moving. For example, family considerations may force an individual to forgo a better job opportunity in a distant city. This may involve the spouse having a good job at the present location, or the desire to remain close to parents and siblings. An important advance in migration economics is the insight that individuals do not make migration decisions on their own (Stark 1991; Greenwood 1985; DaVanzo 1977). Instead, intrahousehold factors come into play, such as the preferences of spouses and children, when these decisions are made.

The notion of living in a large city with its own Major League baseball team may be enough of an inducement for a baseball fan to even take a pay cut. Likewise, a pleasant climate or natural amenities can have the same effect on an individual’s decision to relocate by changing the net benefits derived. We will see later how these effects have been built into economic models of migration.

2.2.2. Labor Mobility

Perfect mobility of resources, including labor, over space is a key assumption of neoclassical economic theory. One reason why American labor markets are generally believed to function well – in comparison to European labor markets, for example – is that labor is highly mobile geographically. Workers in areas with high unemployment rates are apt to move to areas where unemployment rates are low. In this manner, labor market equilibrium is restored and maintained over time (exceptions occur in depressed areas with persistent long-term unemployment). In contrast, many European workers are simply unwilling to relocate even when they face long-term unemployment prospects in their local areas. This rigidity is at least in part responsible for a less smoothly functioning labor market in Europe.

2.2.3. Impacts on Local Businesses

The back-and-forth fluctuations of community populations can create serious problems for building stores and providing optimal amounts of retail space. This has given rise to a field of research on population "threshold analysis," which estimates the number of retail stores–such as building materials, clothing and furniture stores–that are supported by different population sizes (see, e.g., Shonkwiler and Harris 1996, Deller and Harris 1993, Harris and Shonkwiler 1993 and Schuler and Leistritz 1990). The estimates derived from this type of research are particularly useful for entrepreneurs in communities rapidly gaining new residents. They are also helpful for communities experiencing decline, because they allow them to better anticipate necessary adjustments in retail and related services.

Table 5: Population Thresholds for Selected Retail Functions

Function Number of Stores






Food store

Gas station

Feed store


Insurance agency




Liquor store

Women’s clothing store


Shoe store





















































Source: Foust and Pickett, 1974, Table 1, as reported in Shaffer 1989.

2.3. Summary and Conclusion

Migration can be vitally important to the economic well-being of communities. It affects the level of economic activity within a community, determining flows of incomes. Migration also affects costs facing longtime and new residents. These costs can affect both the destination community to which the migrant moves, and the place from which the migrant came. In particular, costs associated with congestion arise in communities receiving migrants at above-sustainable rates, while all sorts of public finance problems can arise in communities rapidly losing population to out-migration.

Accurate migration forecasts are essential for planners in both the public and private sectors. Indeed, both current and projected data on migration have economic value. A number of public and private companies report timely migration data on the Internet for various geographic entities. For example, the Colorado Governor’s Office reports net population in-migration numbers for various regions of the state; the Big Sandy Area Development District in Kentucky reports net migration numbers for each of the five counties in the district; the private company, the Dismal ScientistÔ (also known as Regional Financial Associates) reports net migration rates for populations of metro areas and states as part of its rankings of different geographic entities using key economic statistics; and Gordon Pattison’s Investor’s Property Services likewise maintains migration data from the Census Bureau on its Web site. For example, data for Pima County in Arizona are available. Clearly, much is at stake for real estate and investment companies when people migrate!

Despite the important effect that migration has on both receiving and sending communities, there is a relative lack of systematic and quantitative information–that goes beyond anecdotes–about the impacts of migration. In fact, Greenwood (1985, 521) points out that most migration research has focused on the causes of migration as opposed to its impacts or consequences (retirement migration, discussed below, is an exception). This theme was echoed recently by Plane and Bitter (1997, 149-150) who write, "[w]ith the wave of current interest in applied regional science, it might be an auspicious time for the migration research community to channel attention towards better understanding how local areas, regions, and nations are affected by migration. ... high mobility is one of the defining characteristics of modern economies and ... yet its repercussions remain relatively little examined or understood."

3. American’s Living Preferences

Americans on average prefer to live in smaller, less densely-populated places. This fact comes out repeatedly in opinion surveys, and it is remarkable if only because there has been such a massive out-movement of people from rural areas to urban places. This out-movement is contrary to expressed preferences. In a Roper poll of 1,210 adults conducted a few years ago, 53% of respondents indicated that people living in small towns and rural areas were likely to "get fun out of life." Over 60% thought these people would be "warm and friendly" and 58% would be "active in their communities." (The basic data in this section were reported in the Lexington Herald Leader on June 11, 1992, A3). One manifestation of the appeal of rural living is the fact that movie stars are buying up large ranches in less densely populated, rural states such as Wyoming and Montana.

With this background it is not surprising that, given a choice, 32% of the respondents to the Roper poll indicated that they would prefer to live in a small, rural town; 21% would live in a midsize city; 16% would live in a small city; but only 9% would prefer a big city and 21% the suburbs (2% had no opinion or "did not know"). Thus, a fundamental discordance exists between where people would like to live, and where they actually live. As we will see later on, the Internet and revolutions in telecommunications technologies may give people more flexibility in choosing where they live.

Rowley (1997) explores the values Americans attach to rural life and rural areas. Quoting William Howarth, an English professor at Princeton University, he writes (3) "nostalgia for rural roots increases during periods of rapid social and economic change [and] ... that expressing rural values is a mechanism used to stem fears of cultural loss." Further, "...rural values are tied to the land as symbols of social and natural stability" and, according to the urban sociologist John Logan, "...rural America has the special advantage of being the place where most of us don’t live any more, which frees us to reconstruct it in our imagination."

4. Rural-Urban Migration Trends in the United States

Rural to urban, and urban to rural migration trends have varied dramatically over the decades of U.S. history. The publication Understanding Rural America (Rowley 1995), which is available on-line, contains two maps that summarize much of the history of migration between rural and urban areas of the United States. These maps are reproduced here as Figure I.5. The larger map shows in green the number of counties that were farming-dependent in 1989. The smaller map (inset) shows farming-dependent counties back in 1940, also using a green color. A phenomenal decline can be seen in the number of farming-dependent counties over this period, particularly in the center of the country. See

Figures from the United States Department of Agriculture reveal that in the 1960s, the nonmetro population grew at an anemic 2.5% compared with a 17.1% rate of growth in the metro areas. In the 1970s and 1980s the numbers observed in the 1960s reversed themselves twice, as discussed in the next two sections.

4.1. The Rural Renaissance: Part I (the 1970s)

During the 1970s, a significant regional population shift occurred in the United States, with the South and West gaining about 90% of the nation’s additional population (Greenwood 1985). According to Greenwood, this shift was more pronounced than any previous regional population shift in U.S. history, including that between 1930 and 1940, when 65% of the nation’s additional population growth accrued to these areas (522). Between 1970 and 1980, nonmetropolitan population growth roared ahead at a 14.4% rate, compared with only a 10.6% rate of growth in the metro areas.

To a large extent, diseconomies of agglomeration in urban areas explained the backlash against metropolitan living in the 1970s. Agglomeration diseconomies are the increasing costs associated with living in more and more densely populated areas or cities. These costs include rush hour traffic jams, exorbitant downtown rents, parking problems and other forms of congestion. When the costs of agglomeration offset the benefits of agglomeration (essentially, living in the highest-order place of a central place hierarchy, where all goods and services are supplied), people will prefer to move into less-densely settled areas.

Other landmark events occurred in the 1970s to increase the appeal of rural living. These were reductions in the relative costs of overcoming distances, such as reduced commuting costs due to cheaper gasoline (albeit only later in the decade, as the energy crisis subsided), and cheaper means of telecommunications (e.g., through fax machines), which reduced the degree of isolation of rural areas. As a result, the 1970s became known as the "rural renaissance" decade. Popular magazines carried headlines such as "Out of the Cities, Back to the Country" or "What Lures Americans Back to the Land." In this period, Los Angeles Times reporter Judy Pastornak wrote (published in the Lexington Herald Leader, June 28, 1996, A3):

[r]eporters comb[ing] small towns in Kentucky, West Virginia, Illinois, Michigan, Idaho [and] New York [found] retirees heading back to their birthplaces, long-distance commuters looking for safety and cleanliness, entrepreneurs escaping the fast pace of the city and corporate life. Thousands of Americans, it seemed, were getting fed up with the metropolis. Today, the amount of prestige associated with living in rural areas is high in developed countries, but it remains low in developing countries.

By way of summary, Greenwood (1985, 525) presents the five following categories of reasons for the rural renaissance: "(1) Changing relative costs of doing business in older urban centers; (2) growth of resource-based of the 1970s industries in nonmetropolitan areas [e.g., Beale (1977) and McCarthy and Morrison (1977)]; (3) rising income and wealth and increasing demand for location-specific amenities [e.g., Graves (1979)]; (4) changing demographic structure of the population and the labor force [e.g., Alonso (1978b), and Greenwood (1984a, 1984b)]; and (5) government policy [e.g., Coleman (1978)]."

However, it was also in this period that the seeds of the farm financial crisis of the 1980s were sown. In particular, high rates of inflation during President Jimmy Carter’s four-year term in office forced a fundamental policy change in the early 1980s.

4.2. The Bicoastal Economy (the 1980s)

In the early 1980s it became clear that drastic macroeconomic policy changes were needed to control inflation. Ronald Reagan had just become president, and Paul Volker was chairman of the Federal Reserve Board. The Fed cut back the rate of money supply growth, which drove up real interest rates sharply, leading to a reduction in economic activity and a subsequent recession. In addition, the Reagan administration introduced tax cuts (The Economic Recovery Tax Act of 1981) while at the same time increasing military spending to thwart the Soviet empire. This placed further pressure on interest rates.

Because production agriculture is capital-intensive, the farm economy suffered the brunt of the macroeconomic adjustment of the early 1980s. In this period, family farms declared bankruptcies at unprecedented rates. Also, because the increased military spending was funneled largely into states on both coasts of the United States, job growth in those states soared and this attracted many people from the center of the country. In fact, at the time it appeared that the center of the country was becoming depopulated, while the coastal areas were booming. Hence the term "bi-coastal economy" was coined. Note how the population of Tooele County, Utah (Figure I.4) follows the growth patterns discussed here at the national level.

During the decade of the 1980s the population growth rate in nonmetropolitan areas dropped back to only 4.1%, compared with a more robust increase of 11.6% in metropolitan areas. This period also saw the onset of a retirement migration boom in some regions, as senior citizens moved to warmer climates, particularly to the southern United States.

4.3. The Rural Renaissance: Part II (the 1990s)

Perhaps more strongly reflecting citizens’ preferences for where to live, the period 1995-96 marked an interesting turnaround. In that year, 250,000 more people moved out of metropolitan areas than moved in. According to Hansen (1997, 1)

[d]uring most of the last decade, metro areas had about equal numbers of people moving in and out each year.... Exceptions occurred in the 1992-1993 period, when metro areas experienced a net loss of 317,000 people, and during the mid-1980s, when they had net gains of from 300,000 to nearly half a million.

The following county-level maps, in Figures I.7, I.8, I.9, and I.10, show the rate of net domestic migration between 1990 and 1998, and 1997 and 1998, and the percent change in total county population over the same two periods, respectively. The maps were obtained from the Census Bureaus’ Web site and are reproduced here for the sake of convenience (updated maps may be made available over time). These maps show that, even if rural areas are on balance gaining population, there are wide differences in the distributions of those gains over place. In other words, some rural areas continue to lose significant population numbers.

Business Week reported that "The Boonies are Booming" in its October 9, 1995 issue (104-112), while the Wall Street Journal suggested that "Small Town Life Lures Young Professionals" in its September 29, 1995 issue (B10). That same year, Johnson and Beale published a paper examining some of the causes for the renewed interest in rural America. They list the following "causes of rebirth" (from an e-mail summary prepared in 1995 by Tim Walters of the Aspen Institute):

The December 6, 1997 issue of Time magazine contained an extensive article under the headline, "The Great Escape: Americans are fleeing suburbia for small towns. Do their new lives equal their dreams?" (52-64). At least in part, this ongoing trend of migration into rural areas is also driven by a desire of individuals to own hobby or "leisure" farms. The August 22, 1997 issue of the Wall Street Journal carried an article titled, "Chick Chic: ‘Hobby Farming’ Catches On." Two years later, U.S. News and World Report (April 12, 1999, 31) reported on a "Code of Country Living" booklet prepared by the Farm Bureau in response to "[t]he outsized expectations of all those city slickers moving back to the land [which] have gotten so out of hand...". The report suggests that upgrades in telecommunications and roads increasingly allow businesses and individuals to locate in the "country."

The pronounced movement of people from metro to nonmetro areas, according to another observer, was due to the fact that "...the recession of the early '90s hit cities harder, and the trend toward just-in-time delivery systems, adopted to catch up with the Japanese, made dispersal of assembly lines and warehouses more attractive. ... The rise of ex-urbs, or 'edge cities,' at the very periphery of the suburbs brought more metropolitan jobs within commuting range of once-isolated small towns as well. ... The migration to the metropolitan areas is immigrants from overseas [while] for native-born Americans, ‘the attraction is the non-metro areas’" (LHL, June 28, 1996, A3). This theme is discussed in more detail in subsection 11 below.

Another manifestation of population flight from congested areas is the recent movement of people from California into the Western Mountain States. In the 1990s, largely as a result of in-migration from California, property values in some parts of Salt Lake City, Utah shot up by more than 100%. According to Frey (1998, 3),

[t]he fastest growing counties via domestic migration are located in the Southeast and Rocky Mountain West, and in smaller and nonmetropolitan areas. The latter counties tend to attract itinerant professionals and the soon-to-be burgeoning elderly population, but many of them also attract "would-be suburbanites". The latter have shown especially strong tendencies to leave both inner and outer suburbs of densely populated "high immigration metropolises." ... The 30 counties with the highest domestic migration rates in the 1990s are emblematic of new destinations: smaller places and nonmetropolitan counties in fast-growing states like Colorado, Utah, Texas, and Nevada. On the list are also suburban counties of metropolitan areas that lie in "domestic migrant magnet" regions.

This leads us to the topic of urban or intermetropolitan migration.

5. Urban and Suburban Migration

If rural areas have been subjected to major population shifts due to migration, the same is true of urban areas. However, the experience of individual cities varies dramatically in this regard, as the following table illustrates. Clearly, the benefits and costs of population growth have not been distributed evenly across urban areas, and migration is one of the primary mechanisms through which these differences have been created. For purposes of comparison, the total U.S. population grew by 29.5% between 1970 and 1996.

Table 6: Populations of Selected Urban Areas, 1970 and 1996, in Thousands, and Change

Metropolitan Area*

1970 Population

1996 Population

Change (%)

Atlanta, GA MSA

Buffalo-Niagara Falls, NY MSA

Dallas-Fort Worth, TX CMSA

Detroit, MI PMSA

Los Angeles-Long Beach, CA PMSA

Miami-Fort Lauderdale, FL CMSA

New York, NY PMSA

Philadelphia, PA-NJ PMSA

Pittsburgh, PA MSA

St. Louis, MO-IL MSA

San Francisco-Oakland-San Jose, CA CMSA

Utica-Rome, NY MSA





































Source: Author’s calculations using census data.
*Note: MSA is a metropolitan statistical area; CMSA is a consolidated MSA; and PMSA is a primary MSA.

Depending on the time period covered, differences among cities can be even more pronounced. For example, over the course of the 1980s, the city of Pittsburgh lost 7% of its population. These population shifts can be attributed to two primary forces. First, with the economy shifting to a services-oriented base, the former capitals of industrial production, particularly in the Midwest, have declined in relative (and even absolute) importance. Second, new economic activity–particularly when it is footloose–has often emerged in places with milder and more pleasant climates, such as the West and Southwest. Some authors refer to the advent of air conditioning in explaining the increasing prominence of the South, along with lower-cost labor. As mentioned above and discussed in more detail in Section II, foreign immigrants show a strong preference for locating in metropolitan rather than in rural areas. In part this is believed to reflect the desire to keep social and economic ties with former countrymen and -women who are already in major cities.

Another form of urban migration–sometimes referred to as intra-urban migration or suburbanization–has been the "flight" of inner-city residents to the suburbs. Often this migration has consisted of white racial groups, which left behind minorities in the inner cities. According to the spatial mismatch hypothesis, discussed at length in Section V, low-skill jobs often also migrated out of the inner cities, leaving few employment opportunities for those staying behind.

One manifestation of this urban flight is the population growth in counties on the periphery of large metro areas such as New York. In some of these, including "southern and eastern New Jersey, as well as Pike County, Pennsylvania, ... [the] domestic migration gains overshadow [foreign] immigrant gains" (Frey 1998, 3). Factors motivating migration patterns of foreign immigrants are discussed below. More generally, primary urban migration targets vary for different races, as suggested by Table 7. Section III includes an examination of differences in migration characteristics according to race.

Table 7: Metro Areas with Greatest Population Gains, 1990-96 for Hispanics,
Asians, Blacks and Whites


Metropolitan Area

1990-96 change



1 Los Angeles-Riverside-Orange County, CA CMSA 1,028,141
2 New-York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA 447,867
3 San Francisco-Oakland-San Jose, CA CMSA 250,747
4 Houston-Galveston-Brazoria, TX CMSA 222,144
5 Chicago-Gary-Kenosha, IL-IN-WI CMSA 221,308
1 Los Angeles-Riverside-Orange County, CA CMSA 305,860
2 New-York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA 294,485
3 San Francisco-Oakland-San Jose, CA CMSA 240,969
4 Washington-Baltimore, DC-MD-VA-WV CMSA 87,208
5 Chicago-Gary-Kenosha, IL-IN-WI CMSA 70,966
1 Atlanta, GA MSA 159,830
2 New-York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA 154,446
3 Washington-Baltimore, DC-MD-VA-WV CMSA 129,909
4 Houston-Galveston-Brazoria, TX CMSA 97,163
5 Miami-Fort Lauderdale, FL CMSA 86,812
1 Atlanta, GA MSA 320,841
2 Phoenix-Mesa, AZ MSA 301,505
3 Dallas-Fort Worth, TX CMSA 245,672
4 Las Vegas, NV-AZ MSA 202,944
5 Portland-Salem, OR-WA CMSA 198,702

Source: Frey 1998, Table 3.
*Non-Hispanic Whites. MSA=Metropolitan Statistical Area; CMSA=Consolidated MSA

An interesting consequence of intermetropolitan migration is that the benefits of economic growth are not necessarily passed on to the impoverished local residents who most urgently need those jobs. A recent study of intermetropolitan migration to Atlanta using the PUMS data set (Section IV.5.2) finds that in-migrants from other metro areas responding to new job opportunities crowd out resident minority and young workers for whom the jobs were created in the first place (Sawicki and Moody 1997; see also Bartik 1993). This also raises serious challenges for economic developers in rural areas, who have to deal with the consequences of return migration.

Although an individual’s locational preferences are often circumscribed by economic necessity (such as declining rural job opportunities during the 1980s), they also tend to follow fads. One such current fad is living in downtown areas of large cities. An article in Housing Facts & Findings reports that the downtown populations of major cities are projected to grow by 75% between 1998 and 2010 (Sohmer and Lang 1999). Houston and Cleveland are forecast to lead all cities, with population growth of over 225% each. In part this trend is expected to be driven by a booming number of "empty-nesters." Another segment predicted to prefer life in downtown areas is that of young professionals without families. These trends are likely to partially counteract the pervasive urban sprawl of the recent past.

6. Barriers to Migration

Barriers to migration impede the efficient operation of labor markets in articulating labor demand and supply, and they prevent individuals from maximizing utility. In some countries the government imposes strict controls on the movement of people over space, and especially between rural and urban areas (e.g., China). In the former Soviet Union, residents were also severely restricted in where they could live.

In free market economies, however, there are no government-imposed restrictions on the movements of people within national borders. In the United States, citizens have a constitutional right to choose the states and counties in which they live. Local zoning ordinances and restrictions are perhaps the only tools available to government officials to reduce in-migration in communities that attract migrants at very high rates, but often zoning regulations are relaxed under pressure from various interest groups, especially if they are collectively viewed as too restrictive. In addition, urban professionals, such as lawyers and teachers, face licensing requirements in each state that constitute a barrier to migration.

6.1. A Taxonomy of Barriers

Klaassen and Drewe (1973, 28-32) provide a useful taxonomy of barriers that arise in free-market economies. These barriers include perceived lack of opportunities, information barriers, physical capability barriers, financial capability barriers and social capability barriers, in addition to occupational barriers discussed separately in the next section. Other barriers that become more and then less restrictive during an individual’s life cycle, such as having children, are discussed in the later sections, which review empirical studies of migration behavior.

The potential migrant’s perception of economic opportunities at a distant location is usually developed in a highly subjective manner (Klaassen and Drewe, 28). This subjective assessment can be made more objective through comprehensive economic development programs that enhance the attractiveness of places, or by persuading individuals to reevaluate their subjective assessments. In the age of the Internet and with significantly greater information flows, the gap between perceived and real opportunities in different places is rapidly shrinking.

Webbox: Falling Barriers to Migration: the Role of the Web
While personal relationships and contacts with migrants from earlier periods are critical in breaking down information barriers, the Internet is also fundamentally changing the flows of information between places and consumers. This includes local newspapers that are on line, job-vacancy announcements and home listings posted by national real-estate brokerages. For the latter, check out (the National Association of Realtors home page);;; or It is interesting to compare the amounts of information available today, and the speed with which the information is conveyed, to a foreign national considering a move to the United States with the amount of information that was available to a mid nineteenth-century settler who relied on personal contacts and mail conveyed by boat.

In general and traditionally, the greater the distance between two places, the more limited the flows of information between them (Berry, Conkling and Ray 1987). Information barriers, as their name implies, may be more easily dealt with than a negative perception about a place to the extent that mere dissemination of knowledge can break down the barriers. In contrast, it may be more difficult to change a person’s negative feelings about a place. The migration studies reviewed later on in this chapter focus on the information that was available to the migrants, but they necessarily fail to consider the information that was not available to them.

Previous studies have shown the following to be important sources of information about distant places that are potential migration targets, in the order presented: friends and relatives, "special trips," employer representatives, "other methods," newspaper advertisements, private employment agencies, unions and state employment agencies. Clearly, some of these involve more cost than others, and not all sources of information are available to all individuals.

People who are in ill health or not physically mobile are unlikely to move, as is confirmed by migration data. For them, moving to a distant place where they do not have the support of family members or other relatives and friends poses a high risk. These physical barriers can be very difficult to overcome.

Moving an entire household is a costly proposition, and financial limitations can constitute a real barrier to migration. Moving can be costly in itself, and a low-income migrant from a depressed economic area may have great difficulty finding reasonable housing in a major city, let alone comparable housing. If friends or relatives previously provided transportation or baby-sitting services, then having to pay for these in a new place represents an additional financial barrier to the potential migrant.

When people move out of their familiar neighborhood and into a new region they face a possible challenge or question of social adaptability. Klaassen and Drewe use the term "attitudinal mobility" to describe whether or not an individual is sufficiently innovative and open-minded to live under new social relationships. Psychic costs associated with new and unfamiliar dialects, customs and social arrangements in faraway places can also pose formidable barriers to migration (Berry, Conkling and Ray 1987, 56). The question of differences in socioeconomic class and the acceptability of individuals within communities has also been viewed as a barrier to migrant workers in agricultural areas and in meat processing facilities. Here the issue is viewed from the other end, i.e., from the perspective of the community receiving the in-migrant (Simon 1999; Eastman 1998; Debertin 1993).

6.2. Social Capital Considerations

People make investments in their communities by getting to know or helping their neighbors. These investments result in the formation of "social capital," which has been defined as the "glue that holds society together" (e.g., Flora 1998; Debertin and Goetz 1997). The fact that migrants will lose this capital, which they may have developed at some cost over time, constitutes a further barrier to moving. In the migration literature this is also known as cumulative inertia: the longer are the ties that people have to a community, the less likely they are to move because they are place-bound (Huff and Clark 1978; Boyle, Halfacree and Robinson 1998). Family, cultural and social ties binding people to certain areas, such as the Appalachian mountains or the Ozarks region, have also been cited in the literature as reducing the propensity of individuals and their families to migrate out of economically depressed places (e.g., Berry, Conkling and Ray 1987, 56).

Schiff (1992) goes one step further and argues that out-migrants impose a negative externality on those they leave behind because they destroy some of the existing social capital, which he views as a "common property resource" not unlike a public park or beach. He questions whether the welfare benefits associated with mobility of workers are large enough to compensate for these costs, and is concerned that the great ease with which individuals can migrate across states contributes to the feeling of isolation that people sometimes have.

6.3. The Changing Importance of Barriers

As indicated above, changes in transportation infrastructure and communications facilities are dramatically reducing the cost for potential migrants of obtaining information about distant cities and regions. White and Mueser (1994) argue that these developments affect the propensity of individuals to migrate. In addition to transportation and communications advances, they write (248) "...integration of regions within the United States would help to lower barriers to movement by reducing the cultural costs of relocation and by providing a national job market."

7. Occupational Mobility

A change of occupation–occupational mobility–can be related to geographical labour migration in a variety of ways. ... Inter-regional migration and occupational mobility are in part associated: a change in job location and residence can imply a change of occupation, too. When people are released from declining and stagnating activities [industries] ... but remain in the labour force, a prerequisite of getting a job in another sector is occupational mobility.        (Klaassen and Drewe, 1973, 32)

Historically, farmers have been one of the most mobile occupational groups. As the United States has developed from an agrarian society to one built on industries and services, the number of people employed in production agriculture has plummeted dramatically. This drop would not have been possible without a dramatic increase in labor productivity in agriculture, brought about through science and research. At the same time, the rapid industrialization of this country could not have occurred without the ready availability of workers coming out of agriculture to work in factories. Barkley (1990), whose work is reviewed later, discusses factors influencing the occupational mobility of farmers over time.

One of the chief concerns surrounding the 1996 Welfare Reform Act (the Personal Responsibility and Work Opportunities Act) is how former welfare recipients in economically depressed areas will fare in finding new employment. A question mark surrounds the issue of whether these former recipients will move to areas with greater economic opportunities, and whether they can successfully compete for jobs in those areas with existing residents, particularly if the economy enters a recession. This also relates to the larger question of whether people should be moved to jobs, or jobs moved to people, to accomplish economic development.

Different occupations are affected in varying degrees by the business cycle. For example, blue-collar manufacturing workers in the automobile industry are more susceptible to cutbacks in employment in a recession than are white-collar professional workers in health care or education. Thus, regions or pockets of the country with greater concentrations of workers in particular occupations may be harder hit by an economic downturn than other regions, and the question arises whether workers in different occupations are more or less likely to move, i.e., to have varying migration propensities. For example, Hill and Bier (1989) showed the spillover impact of varying neighborhood-level concentrations of different types of workers (e.g., blue-collar auto workers) in Cleveland on property values and income over time.

Recent data from the Current Population Report for the period March 1996 to March 1997 show that workers in different occupations did have varying propensities to migrate. In fact, 7.3% of all technicians and related support workers migrated into a different county, compared with only 3.5% of the sales occupations, which was the lowest category. In comparison, 5.8% of the executive, administrative and managerial occupations migrated, as did 6.5% of the professional specialty occupations.

Webbox: The Spatial Mobility of U.S. Versus European Academic Economists

How does the training of academic economists in the U.S. differ from that in Europe? An article in the Journal of Economic Perspectives a few years ago points out that U.S. economists are trained in economic theories and principles that are "universally valid." In other words, their training allows these economists to ply their trade anywhere in the nation. Indeed, the job market for academic economists in this country is a national one.

In the European university-institute system, in contrast, economists study the local government structure and institutions of the cities in which their institute is located in great detail. They develop strong ties with local leaders and businesses and, unlike their U.S. counterparts, are "locally specialized" and not as readily employable elsewhere in Europe (or the world).

Thus, U.S. economists are much more mobile than their European counterparts and, therefore, they are also much more likely to migrate across the country, away from their alma maters, in search of employment.

Migration propensities for occupations requiring the least amount of skills or formal training were: 4.8% for transportation and materials moving workers, and 6.5% for handlers, equipment cleaners and general labor. For all sales occupations combined, the percentage was 6.7. Thus, although some differences exist across occupations, it is not entirely obvious that the occupations most likely to absorb unskilled former welfare recipients are necessarily more or less mobile.

For declining areas where economic activity is concentrated in only a few industries, a few questions are in order: What type of training should laid-off workers receive, and for which occupations should they be prepared? Should the workers receive general training or specific training for jobs that may be available only in other regions, which would entail private and social costs when the workers subsequently move? These are important and pressing questions facing community and government leaders, particularly in rural areas that have just lost their dominant employers, such as textile manufacturing plants.

8. Return Migration

An important phenomenon is that of return or reverse migration. If people were forced out of certain regions, such as declining rural places, early in their working lives because that was the only way in which they could make a living, then we might expect them to return to their birthplace once they enter the retirement phase of their lives. They may also return if economic conditions improve over time in their hometowns, or if they have acquired a new set of skills elsewhere that they can now apply in their home communities. Alternatively, what if people made a mistake with their migration choice (e.g., Mueser 1997)? What if the destination turns out not to be as attractive as they initially thought? In this case we may expect return migration to people’s original homes.

Reverse migration can present an economic development headache. Suppose a certain community has a high unemployment rate, and that numerous residents have migrated away from this community. Despite the out-migration, the unemployment rate remains high and the local chamber of commerce pursues various economic development strategies to assist those who remain in the community.

Suppose further that the chamber succeeds in bringing new jobs to the community. The problem that may well develop is that the former residents who migrated out of the community decide to return once they hear (through informal or other communication channels) that job prospects have improved in their former home county. Thus, they return-migrate and, because they are likely to have gained new skills in the places to which they had moved, they are likely to be more competitive candidates for the new jobs than the residents who never left. The upshot is that the economic development effort fails to reach or assist those it was intended to help: those who never left the community in the first place. This can make it very difficult to develop certain communities of the country. More generally, it raises a question: What is the relevant (local) labor market of a community? Moreover, it once again raises the question for depressed communities of whether people should be moved to jobs, jobs to people, or both?

Notice in the above example that some feature or characteristic originally distinguished those who left the community from those who stayed behind. Perhaps those who left had a different outlook toward work. They certainly were more willing or more able to explore new job opportunities than those who remained behind. This and other characteristics may intrinsically make return migrants more attractive to employers.

Eldridge (1965) was one of the first researchers to develop a taxonomy of migrants that included return migration. He defined primary migrants as anyone leaving a home region for the first time; secondary migrants as anyone moving subsequent to the first move and to any place other than the place of birth; and return migrants as those returning to a place of birth. Results of a study of interstate migrations reported in Long (1988) sheds light on how many migrants are primary migrants, return migrants or onward (secondary) migrants. These numbers are shown in Table 8 both as a share of all migrants and as a rate of people considered "at risk" to migrate.

Table 8: U.S. Interstate Migration Characteristics, 1975–1980

Total interstate migration

Primary migrants

Return migrants

Onward migrants

Percent of total





Rate/100 persons





Source: Long 1988.

According to Table 8, return migrants make up fewer than one in five migrants. However, when the rate of return migrants per 100 persons is considered on an age-specific base, it turns out to be 6.2. As Long writes (1988, 105-106), "if persons living outside their state of birth picked their next state of residence at random, the rate of onward migration would be 49 times the rate of return since there are 49 possible destinations." Not all retirees return to their roots, however. The next section examines the increasingly important subject of retirement migration.

9. Retirement Migration

As noted above, retirement decisions have a profound influence on migration patterns across the United States. Senior citizens often have considerable freedom in choosing where they live, since many are no longer bound to a particular location to earn an income. Thus they are free to choose the most desirable locations not only in terms of climate but also natural amenities and entertainment, as is the case with the recent retiree-entertainment phenomenon of Branson, Missouri. Some retirees do not settle down at all, preferring instead to travel across the country in mobile homes.

As life expectancies increase and people have many years to look forward to in their retirement years, decisions by retirees in terms of where to locate will have increasingly important economic implications. Of course, some retirees move back to the rural areas they were once forced to leave in pursuit of better economic opportunities. This is discussed in more detail under subsection 8 of this section (Return Migration). Frey, Liaw and Lin (1998) identify states that appeal to elderly migrants with different socioeconomic characteristics.

Webbox: Where do Retirees End Up?

The following Web site contains a map of nonmetro retirement destination counties in 1990: Retirement-destination counties are those counties which "experienced 15 percent or more in-migration of people age 60 and older in the 1980s. These counties are generally more rural than other nonmetro counties. In addition to being located near amenities, these counties also tend to be near military bases, reflecting the desire of military retirees to be near medical and shopping facilities located on the bases" (Source: This map is reproduced here as Figure I.12.

Although a limited number of studies exist on economic impacts of migration, the effect of retirement migration into communities has been studied by researchers. Siegel and Leuthold (1993) identified the economic and fiscal impacts of the Tellico Village retirement/recreation community in Tennessee. They found, for example that even though the residents of Tellico Village made up only 3.5% of the community residents, they nevertheless accounted for 12% of the property taxes paid in the county in which the village was located (Loudon County, Tennessee). In contrast, Tellico Village residents required below-average services from the local government, in part because most of them did not have school-age children in their households.

Kelsey (1993) discusses conceptually some of the fiscal impacts and related issues associated with migration of the elderly. Citing Drabbenstott and Welch (1992), he reports that in the 1980s only those rural areas which attracted retirees performed well economically. Further, "[e]lderly in-migrants have relatively stable incomes, isolated from the economic downturns that befall other industries, providing a steady stream to the economic base of the area" (1170). Another study of the economic impact of retirement migration is that by Deller (1995). Using the REMI simulation model, he finds that retiree in-migration has significant economic effects on a community, with the biggest effects occurring in sectors related to health, retailing and construction.

10. U.S. Migration Policy

Although the United States has no formal internal migration policy – and indeed, residents have a constitutional right to locate in whichever state they please – elected officials in individual states sometimes express preferences with respect to migration into (or out of) their states. A few years ago, for example, the governor of Oregon invited residents of other states to visit Oregon, but he also asked them not to stay in the state indefinitely.

Other states, such as California, have had long-standing concerns that their relatively generous welfare payments attract recipients from states in which benefit levels are not as high. For example, the average AFDC payment per family in California in 1996 was $549, compared with $118 in Mississippi and $148 in Alabama. In fact, a high correlation exists between per capita income in a state and the size of the average benefit level. In a landmark decision, the Supreme Court ruled in Spring 1999 that individual states, such as California, could not discriminate against in-migrants by paying those migrants the amount of benefits they were receiving in their original state.

Research by Frey et al. (1995) suggests that the concerns expressed in states such as California are misplaced, to the extent that no statistical evidence exists that higher grant payments attract welfare recipients. With the sweeping changes brought about by the 1996 Welfare Reform Act, migration pressures due to differences in the level of benefits may diminish. However, Frey’s study also suggests that foreign immigrants may displace low-skill U.S. residents from certain large metro cities (see subsection 11). This would imply that an important linkage exists between federal immigration policy and domestic migration.

11. Migration Patterns of Foreign Immigrants

The United States is, of course, known as a melting pot of civilization, having attracted people from all over the world in its relatively young history. Data on the rate of immigration into the United States (the number of immigrants divided by the total existing population) are available in the Statistical Abstract at (see Table 5 in the 1998 Abstract). In the period 1901-1910, for example, the rate of immigration was 10.4%. The immigration rate fell to a low of 0.4% during the 1931-1940 period, and recently it rose to a postwar high of 7.2% in 1991. Since then, the rate has fallen to a more moderate 3.5% in 1996, the most recent year for which data are available. If you are interested in the country of birth of recent immigrants, the data are reported in Table 8 of the above-referenced publication.

International immigration is briefly discussed in this section, especially as it relates to or affects internal migration. Emigration (expatriation) of U.S. citizens and re-emigration of foreign nationals (repatriation) who previously migrated into the United States are not discussed, because little is known about these subjects beyond anecdotal evidence (although formal research is starting to appear on the latter topic). As was already noted in Section 4.3., the locational preferences and hence migration patterns of foreign immigrants once they reach U.S. shores differ from those of footloose U.S. citizens. This section reviews the migration targets of foreign migrants within the nation. An important national debate is taking place over whether immigrants take jobs away from citizens, or whether immigrants on balance create net new jobs. Knowing where these migrants locate in the country is an important first step in assessing the consequences of immigration.

Pisarski (1996, 22) points out that the United States' 8.7 million foreign immigrants during the 1980s accounted for a whopping 40% of the total population growth in the nation over that period. Many immigrants settle in central cities, thereby counteracting the outflow of existing residents from those areas. Also (ibid., 23):

[f]oreign immigrants have had direct impact on the growth patterns of many states. Some traditionally rapidly growing states, such as California and Florida, have had their growth expanded–in the case of California, dramatically so. Other states, such as Texas, Pennsylvania, and Massachusetts, have had their population losses reversed by immigrants. Others, such as New York and Ohio, have had their declines reduced, but not reversed.

Data for the period 7/1/97 to 7/1/98 (at show the top five states receiving (net) international migrants were California (268,685), New York (124,484), Texas (100,175), Florida (80,559) and New Jersey (55,161). The bottom five states over this period were Wyoming (338), West Virginia (345), Montana (366), Maine (619) and South Dakota (812). Over this period, net international immigrants accounted for 37.3% of total population growth in the U.S., compared with only 29.8% over the period 1990 to 1997.

The migration patterns of domestic residents differ profoundly from those of foreign immigrants. Frey (1998) presents the top migration magnets (targets) for domestic and foreign migrants, and finds that only the Dallas-Fort Worth metro area appears on the top-10 list of both types of migrants (Table 9). With the exception of this single area, the top 10 magnets for foreign migrants is completely different from that of domestic migrants. Also noteworthy is the fact that nearly every one of the top targets of foreign immigrants experienced a net loss of population to net domestic migration over this period (that is, on balance more citizens moved out of than into these areas). The only exceptions to this rule are the two metro areas in Texas.

Frey explains the continued trend toward concentration of foreign immigrants in specific metro areas as follows (1998, 1):

This concentration is influenced by the strong family reunification provisions of our immigration law, and the change toward Latin America and Asia as dominant origins for immigrants over the past several decades. Family reunification immigration tends to occur in "chains" that link family members and friends to common destinations. This is especially the case for lower-skilled immigrants since they are more dependent on kinship ties for assistance in gaining entry to informal job networks that exist in port-of-entry areas.
Table 9: Top 10 Magnets for Foreign Immigrants and Domestic Migrants, 1990-1997


Metropolitan Area*


Net Domestic Migration























High Immigration Metros

New York CMSA

Los Angeles CMSA

San Francisco CMSA

Chicago CMSA

Miami CMSA

Washington DC CMSA

Houston CMSA

Dallas-Fort Worth CMSA

San Diego MSA

Boston NECMA

High Domestic Migration Metros

Atlanta MSA

Las Vegas MSA

Phoenix MSA

Portland, OR MSA

Denver CMSA

Dallas-Fort Worth CMSA

Seattle CMSA

Austin MSA

Orlando MSA

Raleigh-Durham MSA













































Source: Frey 1998, Table 1.
NECMA is a New England Consolidated Metropolitan Area.

Immigrants from the same national backgrounds have long concentrated in the same neighborhoods of cities. For example, Toronto has a "Chinatown" and a "Greektown," New York has "Little Italy" and cities such as Detroit have clusters of former Iranians. Likewise, many rural farming communities originally consisted of families from the same national backgrounds, such as German, Norwegian or Danish. Knowledgeable observers have also pointed out that nationals from certain countries were more likely to be assimilated into the national melting pot than nationals from other countries. Along these lines, Frey (1998) argues that (1):

The latest migration statistics for the 1990s reinforce a new regional division that ... is occurring because of the continued clustering of foreign-born immigrants into a few multi-ethnic urban areas, as native-born and longer-term mostly white and black residents disperse to new employment opportunities in other parts of the country. These separate migration processes are creating a demographic divide across space that could be just as monumental as well-known past demographic divides: rural versus urban, city versus suburb, snow belt versus sun belt. The new one will separate the regions of the country which serve as "immigrant gateways" from the remainder of the national territory, and ... [t]he single melting pot image might be supplemented by "multiple melting pots" in the context of a less diverse Middle America. ... Our contention that a new demographic divide is emerging ... holds important implications for regional social and political cleavages, [for] the economies in high immigration labor markets, and for the upward mobility and assimilation of immigrants.

Readers interested in the economic impact of immigration are referred to the survey article by Borjas (1994). For a review of current issues in the area of migrant agricultural labor see, for example, Perloff, Lynch and Gabbard (1998), Jobes (1998), Martin and Taylor (1998), Ise and Perloff (1995), Taylor (1992) or Emerson (1989).

12. Migration Issues in Foreign Countries

People have migrated across borders since the dawn of civilization. In the twentieth century, the United States, Canada and Australia were primary targets of transnational migrants: these three countries absorbed about one-half of all emigrants worldwide (Borjas 1994, 1667). Even countries losing citizens to this "New World" attracted immigrants from other parts of the world, however. For example, the foreign-born make up nonnegligible shares of the total population in the following countries (ibid.): France: 11% of the total population; Switzerland: 17% and the United Kingdom: 9%.

International migrants tend to be motivated by the same factors as domestic U.S. migrants: they want to improve their economic condition. In addition, however, they may also be trying to escape oppressive political systems. The collapse of the former Soviet Union has triggered significant population movement from Eastern to Western Europe, even as government control of economic activity has ostensibly been relaxed. Another key trigger of international population migration is the growing income gap between rich and poor parts of the world, such as the United States and Mexico, Europe and Africa, and Russia-Central Asia (Diamantides nd).

Other triggers of involuntary migration include natural disasters (such as Hurricane Hugo in Central America in 1998) or civil and ethnic war (such as in Rwanda in the mid-1990s), which can create large populations of refugees. Governments sometimes engage in forced population resettlements (e.g., Indonesia), which tend to fail if they occur without regard to the infrastructure available to migrants in the new regions.

Within more-developed countries, Plane and Rogerson (1994, 101) maintain that:

[m]uch of the total volume of migration ... now consists of interchange between pairs of the most economically advanced regions, rather than of the more demographically effective movement into these areas from less well-to-do places.

Indeed, migration has been referred to as leading to (and resulting from) a certain "elitism" in advanced countries, since the better-educated (and wealthier) populations tend to be more mobile.

The International Handbook on Internal Migration, edited by Nam, Serow and Sly (1990), is an excellent starting point for readers interested in internal migration issues in specific countries, such as Botswana, Canada, Ecuador, Egypt, France, Indonesia, Israel, Japan, Poland, the Soviet Union as well as others. Readers can also consult Castles and Miller (1993) (The Age of Migration: International Population Movements in the Modern World), which is reviewed by Waldorf (1996).

Another important resource for readers interested in the subject of international migration is the Web site maintained by Tom Kuecker. His site has information on migration-related data sets (including the Inter-University Consortium for Political and Social Research or IUCPSR); government-based sites with immigration info/data; immigration readings; listservs for discussion groups; research centers and foundations; refugee-related sites; and journals devoted to the study of migration.

International organizations with Web sites related to immigration include the Organization for Economic Cooperation and Development (OECD) and the International Labor Office (ILO).

Other sites of interest include: a migration-related site in Australia; and the Wellesley College Library’s site for Internet research resources related to population, migration and demography. This site also contains addresses of Web sites with data as well as of other international agencies involved in some aspect of migration. The site,, is devoted to issues of integration and immigration around the world.

13. The State of Migration Research: Recent Areas of Emphasis

To conclude this review of contemporary migration patterns and issues, this section presents a summary of recent areas of research emphases as determined by topics published in five regional science journals (Plane and Bitter 1997). Although this does not cover all migration research–many economics journals also publish such studies, for example–the review does suggest where scholars in the field have and have not devoted their energies.

First, in terms of geographic scale of the analysis, the majority of studies (81.5% out of a total of 248 studies reviewed) are concerned with interregional or internal migration within a country. Seven percent address the question of residential mobility, or local migration, 3.2% focus on international migration or immigration, and 8.1% cannot be classified using this criterion (ibid., 145). Most of the studies reviewed were model-oriented (53.2%), using econometric or similar statistical techniques, 24.2% were purely theoretical, 16.5% descriptive and 2.4% consisted of literature reviews (ibid., 147).

In terms of the major topics addressed, the following classification emerged (Plane and Bitter, 149). The vast majority of papers (65.9%) examined determinants of migration. This was followed, in descending order of importance by impacts of migration (15.7%); role of wages and income in migration (15.7); decision-making process and perceptions (7.6); Third World movement patterns (7.6); nonmetropolitan movement patterns (6.0); composition and role of age in migration (5.6); interregional migration equilibrium (5.2); rural-to-urban movement patterns (4.8); composition and role of race in migration (4.0); role of amenities (3.2); alternative opportunities (3.2); influence of past migration (2.8); role of costs of movement (2.4); role of unemployment (2.0); role of occupation (2.0); migration patterns of young adults (2.0); role of space or distance (2.0); role of gender (1.6) and education (1.6); movement patterns of the elderly (0.8) and role of place ties (0.8). As mentioned earlier, a considerable need exists for further research on the impacts and consequences of migration in different types of communities.

14. Summary of Section I

This section has reviewed contemporary migration issues and trends, with a particular emphasis on the United States. The U.S. population is more mobile than the population of virtually any other developed country, with over 5% of residents moving across county lines in an average year. Migration in pursuit of better wages or employment opportunities is a key mechanism whereby workers move up the economic ladder. Predicting migration flows and understanding the determinants of migration are important areas of ongoing scholarly inquiry, but more effort needs to be devoted to understanding the aggregate effects of migration on both sending and receiving communities. Beneficiaries and users of information about local migration patterns range from elected officials to private entrepreneurs.

Exercises and Discussion Questions for Section I

  1. Explain how migration relates to an individual’s job search and location of residence. How does migration differ from commuting? Have you migrated from your place of birth? If so, how?

  2. What does the map showing the geographical center of population tell us about the "exploration and discovery" of the West, and historical U.S. population migration patterns in general?

  3. Describe recent migration patterns among the four U.S. census regions. Which region is experiencing the largest net gain of residents through migration? Which is losing the most people to migration? Can you explain why this is happening? What accounts for the migration boom in the West?

  4. The text claims that "[e]conomically booming areas usually experience an increase in population, while declining areas experience population outflows." How would you test this proposition statistically? To what extent is this a chicken-and-egg question, and how would you handle the simultaneity between these two variables in a regression model?

  5. Discuss public sector-related issues that can result from migration. What do you see as the economic advantages and disadvantages of a government policy that restricts where people live?

  6. Discuss private sector-related issues that can result from migration. As an entrepreneur, how could you benefit from the knowledge that an area is about to lose/gain a large number of people?

  7. What are "population thresholds" and how can they be used in the context of migration studies? How do you explain the difference in population needed to support the different types of retail functions shown in Table 5? Which factors other than the size of the local population might affect the number of shoppers at a grocery store or a gas station?

  8. Outline some of the factors that account for the rural renaissance of the 1970s and 1990s. What is the origin of the term "bicoastal economy?" How would you go about studying the prospects for recruiting foreign immigrants into declining rural areas?

  9. In Table 6, which factors do you suppose account for the rate of population change between 1970 and 1996 in cities such as Detroit, Michigan and Miami-Fort Lauderdale, Florida?

  10. List three types of barriers to migration, and give an example of each type. How does the advent of the information age affect these barriers? Give an example. Would you expect the level of social capital to be high or low in a community with high in- or out-migration rates?

  11. What is meant by occupational mobility? Which group of workers, do you think, is the least occupationally mobile? Do you believe that the national market for highly skilled workers is more or less "thin" than the national market for low-skilled workers?

  12. What is "return migration?" Is it desirable to move unemployed people away from depressed communities rather than help then find local employment? What are the disadvantages of either strategy?

  13. Consider the metro area ranked first under the four ethnic groups in Table 7. For which group is the migration most likely to be driven by "return migration?"

  14. Kelsey (1993) claims that little is known about the factors that make some communities more desirable places than others for retirees. Can you suggest a list of variables that may be important for retirees? Which factors would you consider as you approach retirement age?

  15. Outline an econometric model that would allow you to test the role of national "affinity" in attracting international migrants to a large city. Assume that you have individual migrant-level data to carry out your study.

  16. Compare and contrast the factor motivating international as opposed to domestic migrants when they make their decisions about where to take up a new residence.

  17. Develop a research proposal for estimating the impact of net migration on a local community. Which endogenous variables would you include in your model, and which explanatory variables would you use? Be specific.

Back to Contents

Next section

Back to contributions